content
stringlengths 1
15.9M
|
|---|
\section{Introduction}
The launch of fifth-generation (5G) wireless communication networks leads to an explosive increment of internet of things (IoT) devices, which will generate massive data that need to be processed timely either through local or remote computing facilities. Limited Bandwidth and limited device computation capability have gradually become a bottleneck to meet delay requirement for the massive amount of data.
Meanwhile, the research on sixth-generation (6G) wireless networks has taken off and has been attracting ever increasing attention from academia and industry \cite{6g}. Future wireless networks aim for realizing ultra-high spectrum efficiency (SE) and energy efficiency (EE), ultra-dense user connectivity, and ultra low latency. EE is critically important in future wireless networks since massive energy constrained IoT devices need to be reliably connected to support smart city, smart factory, etc. Moreover, it is imperative to address the finite computation capacity issue of IoT devices.
Mobile edge computing (MEC) is envisioned as a promising paradigm by exploiting the edge servers that are deployed close to the end-users. In wireless networks, edge nodes, such as base stations and edge routers, can be equipped with high computing and storage capabilities. MEC enables users to offload their tasks to edge servers for processing. It is important to have efficient offloading in MEC as massive local devices may need to offload tasks to edge servers for computing. Towards that end, NOMA can be applied to further enhance EE and SE of MEC networks. By exploiting superposition coding at transmitter and successive interference cancellation (SIC) at receiver, NOMA allows multiple users to share the same bandwidth simultaneously in either power domain or code domain to improve spectral usage \cite{hj1}.
Extensive works have formulated EE optimization frameworks and designed energy-efficient resource allocation schemes in NOMA enabled MEC networks \cite{ee1}-\cite{ee3}.
In order to maximize EE under the delay constraints in MEC, a joint radio and computation resource allocation problem was formulated in \cite{ee1}, which takes both intra and inter cell interference into consideration.
In \cite{ee2}, the authors proposed a scheme to maximize the system computation EE of a wireless power transfer enabled NOMA based MEC network by jointly optimizing the computing frequencies, execution time, and transmit power.
In \cite{ee3}, the task offloading and resource allocation problem in a NOMA assisted MEC network with security and EE was investigated. A dynamic task offloading and resource allocation algorithm was proposed based on Lyapunov optimization theory.
However, complex wireless environment and devices' limited transmit power can render it very challenging for the NOMA based MEC network to achieve high offloading rate and high EE. IRS recently has emerged as a promising technique to tackle wireless capacity and energy issues. IRS consists of a large number of low-cost passive reflecting elements with the adjustable phase shifts \cite{irs1}. By properly adjusting the phase shifts of the elements, the reflected signals from various paths can be combined coherently to enhance the link achievable rate at the MEC receiver \cite{irsqun}. Through this, IRS can compensate for the pathloss and fading and allow users to use a relatively lower power to achieve a higher data rate. Since IRS does not employ any transmit radio frequency (RF) chains, energy consumption solely comes from reflective element phase adjustment and is usually very low. It becomes a very promising technology to improve EE in future wireless networks.
The following studies have investigated the performance of using IRS with NOMA.
In \cite{nomairs1}, an IRS-assisted uplink NOMA system was considered to maximize the sum rate of all the users under the individual power constraint. The considered problem requires a joint power control at the users and beamforming design at the IRS, and an SDR-based solution has been developed.
In \cite{nomairs2}, the problem of joint user association, subchannel assignment, power allocation, phase shifts design, and decoding order determination was formulated for maximizing the sum-rate for an IRS-assisted NOMA network.
In \cite{nomairs3}, an EE algorithm was proposed to yield a tradeoff between the rate maximization and power minimization for an IRS-assisted NOMA network. The authors aimed to maximize the system EE by jointly optimizing the transmit beamforming and the reflecting beamforming. It was shown that NOMA can improve EE compared to OMA.
Furthermore, recently application of IRS into NOMA-based MEC networks has been studied.
In \cite{nomairs4}, the authors investigated an IRS-aided MEC system with NOMA. By jointly optimizing the passive phase shifters, the size of transmission data, transmission rate, power, and time, as well as the decoding order, they aimed to minimize the sum energy consumption. A block coordinate descent method was developed to alternately optimize two separated subproblems.
In \cite{nomairs5}, an IRS-aided MEC system was considered and a flexible time-sharing NOMA scheme was proposed to allow users to divide their data into two parts that are transmitted via NOMA and TDMA respectively. By designing the IRS passive reflection and users’ computation-offloading scheduling, the delay was minimized.
However, neither \cite{nomairs4} nor \cite{nomairs5} considered the EE performance of IRS-assisted MEC networks with NOMA, which is very important for system design to obtain the optimal trade-off between achievable rate and consumed power.
Motivated by the above-mentioned observations, in this paper, the EE maximization problem is studied in an IRS-assisted MEC network with NOMA. To the authors' best knowledge, this is the first work that focuses on EE performance for applying both NOMA and IRS in the MEC network. The major contributions of this paper are summarized as follows
We investigate the joint design of the receiver beamforming, offloading power, phase shift matrix, and local computing frequency to maximize the EE in an IRS-assisted MEC network with NOMA. The problem is challenging to solve due to its non-convexity fractional objective function and coupling of the beamforming vector with the IRS phase shift matrix. An alternating optimization algorithm is proposed to solve the non-convex fractional problem by using semidefinite programming relaxation (SDR).
The simulation results show that the proposed method can achieve the highest EE among all the benchmarks
The rest of the paper is organized as follows. The system model is provided in Section II. The EE maximization problem and its solution are presented in Section III. Simulation results are given in Section IV. The paper is concluded in Section V.
\textit{Notation:} $\mathbb{C}^{M\times N}$ denotes the $M \times N $ complex-valued matrices space. $\mathcal{CN}(\mu, \sigma^2)$ denotes the distribution of complex Gaussian random variable with mean $\mu$ and variance $\sigma^2$. For a square matrix $\mathbf{X}$, the trace of $\mathbf{X}$ is denoted as $\text{Tr}(\mathbf{X})$ and rank($\mathbf{X}$) denotes the rank of matrix $\mathbf{X}$. $\angle(x)$ denotes the phase of complex number $x$. Matrices and vectors are denoted by boldface capital letters and boldface lower case letters.
\section{System Model
As shown in Fig. \ref{model}, an IRS-assisted MEC system is considered. There are $K$ single-antenna user equipments (UEs) in the system, which can do both local computing and data offloading. The access point (AP) with an MEC server is equipped with $N$ antennas and the IRS has $M$ reflecting elements.
\begin{figure}[h]
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\centering
\includegraphics[width=3.0in]{sm0330.eps}
\caption{An IRS-aided MEC system with NOMA.\label{model}}
\end{figure}
\subsection{Offloading Model}
The baseband equivalent channel from UE $k$ to IRS, IRS to AP, and UE $k$ to AP are denoted as ${\bf{h}}_{I,U,k} \in \mathbb{C}^{1 \times M}$, ${\bf{H}}_{B,I,k} \in \mathbb{C}^{M \times N}$, and ${\bf{h}}_{B,U,k} \in \mathbb{C}^{1 \times N}$, respectively.
In this paper, IRS adjusts its elements to maximize the combined incident signal from each UE to the AP. The diagonal phase-shift matrix can be denoted as $\mathbf{\Theta} = \text{diag}(\exp({j\theta_{1}}), \exp({j\theta_{2}}), $ $\cdots, \exp({j\theta_{M}}))$, wherein its main diagonal, $\theta_{m}\in [0,2\pi)$, denotes the phase shift on the combined incident signal by its $m$th element, $m=1,2,...,M$ \cite{cwh3}.
The transmitted signal from UE $k$ is given as $\sqrt{p_k} s_k$, where $\sqrt{p_k}$ denotes the transmit power and $s_k$ denotes the independent information. $\mathbf{m}_{B,k}\in \mathbb{C}^{N \times 1}$ denotes the receive beam vectors with unit norm, i.e., $\left\| \mathbf{m}_{B,k} \right\|^2=1$ \cite{nomairs1}.
Therefore, the signal received at AP can be given as
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\mathbf{y}_{B,U}=\sum_{k=1}^K({\bf{h}}_{B,U,k}^H + {\bf{h}}_{I,U,k}^H\Theta {{\bf{H}}_{B,I,k}}){{\bf{m}}_{B,k}}\sqrt {{p_k}} {s_k} + n_{B,U,k},
\end{equation}
where $n_{B,U,k} \sim \mathcal{CN}(0,\sigma^2)$ is the complex additive white Gaussian noise (AWGN) \cite{nomairs3}, \cite{chanel1}.
NOMA is used to improve SE and mitigate the interference between different UEs. By exploiting the SIC techniques, the received signal at AP is sequentially decoded and the UE with the best channel conditions is firstly decoded. The channel of each UE includes a direct link and a reflect link. Since the reflect link depends on the unknown parameters $\bf{\Theta}$, the effective channels cannot be used to order the users at the receiver side. Similar to \cite{nomairs1}, we simply remove unknown reflect matrix by considering it as an identity matrix $\mathbf{I}$. UEs are then sorted based on this channel gain $|(\mathbf{h}_{B,U,k}^{H}+\mathbf{h}_{I,U,k}^{H}\mathbf{I}\mathbf{H}_{B,I,k})|$. Without loss of generality, we assume that UEs are sorted in an increasing order, i.e., $|(\mathbf{h}_{B,U,1}^{H}+\mathbf{h}_{I,U,1}^{H}\mathbf{I}\mathbf{H}_{B,I,1})|\le |(\mathbf{h}_{B,U,2}^{H}+\mathbf{h}_{I,U,2}^{H}\mathbf{I}\mathbf{H}_{B,I,2})|\le \cdots \le|(\mathbf{h}_{B,U,K}^{H}+\mathbf{h}_{I,U,K}^{H}\mathbf{I}\mathbf{H}_{B,I,K})|$. When decoding the signal for UE $k$, the signals from $i=1,2,\cdots, k-1$ are treated as interference.
Thus, the signal to interference plus noise ratio (SINR) for UE $k$ is expressed as
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\gamma_{B,k} = \frac{p_k|(\mathbf{h}_{B,U,k}^{H}+\mathbf{h}_{I,U,k}^{H}\mathbf{\Theta}\mathbf{H}_{B,I,k})\mathbf{m}_{B,k}|^2}{\sum_{i=1}^{k-1}p_i|(\mathbf{h}_{B,U,i}^{H}+\mathbf{h}_{I,U,i}^{H}\mathbf{\Theta}\mathbf{H}_{B,I,i})\mathbf{m}_{B,i}|^2+\sigma^2}.
\end{equation}
The achievable offloading rate is
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
R^{off}_{k}=B\log_2(1+\gamma_{B,k}).
\end{equation}
\subsection{Local Processing Model}
Let $C_{k} $ be the number of computation cycles required to process one bit of data for UE $k$ locally. UE can compute and transmit simultaneously. Let $f_{k}$ denote the computing frequency of the processor (cycles/second) \cite{hj1}. Therefore, the local computing rate can be given a
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
R_{k}^{loc}=\frac{f_{k}}{C_{k}}.
\end{equation}
The power consumption of local computing is modeled as a function of processor speed $f_{k}$. It can be given as $p_{k}^{loc}= \epsilon f_{k}^3$, where $\epsilon$ is effective capacitance coefficient of processor chip.
\subsection{Energy Efficiency}
The energy consumed by each UE consists of transmit power, local computing power, and circuit power consumption.
Thus, the total power consumed by each UE is given as
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
P^{tot}_k= p_k+\epsilon f_{k}^3+P^{cn}_k,
\end{equation}
where $P^{cn}_k$ denotes the constant circuit power consumed for signal processing and it is assumed to be the same for all UEs.
The total achievable rate for each UE is
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
R_k^{tot}=R_k^{off}+R_k^{loc}.
\end{equation}
According to \cite{cwh3}, EE is defined as
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\eta =\frac{\sum_{k=1}^{K}R_k^{tot}}{\sum_{k=1}^{K}P_k^{tot}}.
\end{equation}
In order to maximize the EE, the local CPU frequency, offloading power, decoding vectors, and the phase shift matrix need to be jointly optimized.
\section{Resource Optimization}
In this section, the EE maximization problem is studied by jointly optimizing the local CPU frequency, offloading power, decoding vectors, and phase shift matrix. An alternating algorithm is further proposed to tackle the formulated problem
\vspace{-0.03in}
\subsection{Problem Formulation}
The EE maximization problem is formulated as
\begin{subequations}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\label{P0}
\begin{alignat}{5}
\textbf{P}{_1:}~ &\max_{p_k,f_k,\mathbf{m}_{B,k},\mathbf{ \Theta}}~\eta\nonumber\\
s.t.~~&P_k^{tot} \le P^{th}_k,\\
&R_k^{tot} \ge R_{th},\\
&|\exp({j\theta_{m}})|=1,\\
&\left\| \mathbf{m}_{B,k} \right\|^2=1,
\end{alignat}
\end{subequations}
where $R_{th}$ is the minimum required rate threshold. $P_k^{th}$ is the maximum available power of each UE. It is evident that problem $\textbf{P}_1$ is non-convex due to the fractional structure of the objective function and the non-convex constraints. In order to tackle it, an alternating algorithm is proposed.
By introducing ${\mathbf{w}^H}=[w_1, w_2, \cdots, w_M]$, one has $\mathbf{h}_{I,U,k}^{H}\mathbf{\Theta}\mathbf{H}_{B,I,k}={\mathbf{w}^H}\mathbf{H}{_{B,k}}$, where $w_m=\exp({j\theta_{m}})$, $\mathbf{H}_{B,k}=\text{diag}(\mathbf{h}_{I,U,k}^H)\mathbf{H}_{B,I,k}$.
Thus, the SINR of UE $k$ is given as $\gamma_{B,k}=\frac{a_0p_k|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,k}\mathbf{m}_{B,k}|^2}{a_0\sum_{i=1}^{k-1}p_i|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,i}\mathbf{m}_{B,i}|^2+1}$, where $a_0=1/\sigma^2$, $\overline{\mathbf{H}}_{B,k}=\left[ {\begin{array}{*{20}{c}}
{{\mathbf{H}_{B,k}}} \\
{{\mathbf{h}_{B,U,k}}} \\
\end{array}} \right]$, $\overline{\mathbf{w}}^H=\exp({j\overline{w}})[\mathbf{w}^H,1]$, and $\overline{w}$ is an arbitrary phase rotation.
The objective of the optimization problem can be transformed into
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\frac{\frac{B}{\ln2}(\ln(1+\sum_{k=1}^K{a_0p_k|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,k}\mathbf{m}_{B,k}|^2}))+\sum_{k=1}^KR_k^{loc}}{\sum_{k=1}^KP_k^{tot}}.
\label{10}
\end{equation}
To tackle the complexity introduced by the logarithmic function of $R_k^{off}$ in (\ref{P0}\text{b}), Lemma 1 is introduced. First, we have
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{split}
R_k^{off}&=\frac{B}{\ln2}[\ln(a_0\sum_{i=1}^{k}p_i|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,i}\mathbf{m}_{B,i}|^2+1)\\
&-\ln({a_0\sum_{i=1}^{k-1}p_i|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,i}\mathbf{m}_{B,i}|^2+1})].
\end{split}
\end{equation}
$\mathbf{Lemma~1}$: By introducing the function $\phi(t)=-tx+\ln t+1$ for any $x > 0$, one has
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
-\ln x=\max_{t>0}\phi(t).
\end{equation}
The optimal solution can be achieved at $t=1/x$.
By setting $x={a_0}\sum\limits_{i = 1}^{k - 1} {{p_i}} |{\overline {\bf{w}} ^H}{\overline {\bf{H}} _{B,i}}{{\bf{m}}_{B,i}}{|^2} + 1$, and $t=t_{B,k}$, one has
\begin{equation}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{split}
R_k^{off}&=\frac{B}{\ln2}\max_{t_{B,k}>0}\phi_{B,k}(p_k,f_k,\mathbf{m}_{B,k},\overline{\mathbf{w}},t_{B,k})\\
&=\frac{B}{\ln2}[\ln(a_0\sum_{i=1}^{k}p_i|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,i}\mathbf{m}_{B,i}|^2+1)+ \ln ({t_{B,k}}) + 1\\
& - {t_{B,k}}({a_0}\sum\limits_{i = 1}^{k - 1} {{p_i}} |{\overline {\bf{w}} ^H}{\overline {\bf{H}}_{B,i}}{{\bf{m}}_{B,i}}{|^2} + 1) ].
\end{split}
\end{equation}
By further introducing a variable $\eta_1$ to deal the fractional structure of (\ref{10}), $\textbf{P}{_{1}}$ can be transformed into
\begin{subequations}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{alignat}{5}
\label{p3}
\begin{split}
\textbf{P}_{2}:~ &\max_{p_k,f_k,\mathbf{m}_{B,k},t_{B,k},\overline{\mathbf{w}}}~ [\sum_{k=1}^KR_k^{loc}+\frac{B}{\ln2}(\ln(1+\\
&\sum_{k=1}^K{a_0p_k|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,k}\mathbf{m}_{B,k}|^2}))]-\eta_{1}{\sum_{k=1}^KP_k^{tot}}
\end{split}\nonumber\\
s.t.~~& (\ref{P0}\text{a}), (\ref{P0}\text{c}),(\ref{P0}\text{d}),\nonumber\\\
\begin{split}
&\frac{B}{\ln2}\phi_{B,k}(p_k,f_k,\mathbf{m}_{B,k},\overline{\mathbf{w}},t_{B,k})+R_k^{loc} \ge R_{th}.
\end{split}
\end{alignat}
\end{subequations}
$\textbf{P}_{2}$ is still non-convex due to the coupling of variables.
An alternating algorithm is proposed. To be specific, $p_k$ and $f_k$ are first optimized with a given $\mathbf{m}_{B,k}$, and $\overline{\mathbf{w}}$. $\mathbf{m}_{B,k}$ can then be optimized with the obtained $p_k,f_k$, and $\overline{\mathbf{w}}$. Further $\overline{\mathbf{w}}$ can be optimized with the obtained $p_k,f_k$, and $\mathbf{m}_{B,k}$. This process iteratively continues until convergence.
\subsection{CPU Frequency and Offloading Power Optimization}
With the given $\mathbf{m}_{B,k}$ and $\overline{\mathbf{w}}$, let $A_{B,k}=a_0|\overline{\mathbf{w}}^H\overline{\mathbf{H}}_{B,k}\mathbf{m}_{B,k}|^2$, the problem can be transformed into
\begin{subequations}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{alignat}{5}
\begin{split}
\textbf{P}{_{3.1}:}~ &\max_{p_k,f_k}~\frac{B}{\ln2}(\ln(\sum_{k=1}^{K}p_kA_{B,k}+1))\\
&+\sum_{k=1}^{K}\frac{f_k}{C_k}-\eta_1 \sum_{k=1}^K(\zeta p_k+\epsilon f_{k}^3+P^{cn}_k)
\end{split}\nonumber\\
s.t.~~& p_k+\epsilon f_{k}^3+P^{cn}_k \le P_k^{th},\\
\begin{split}
&\frac{B}{\ln2}\phi_{B,k}(p_k,f_k)+\frac{f_k}{C_k} \ge R_{th}.
\end{split}
\end{alignat}
\end{subequations}
Problem $\textbf{P}_{3.1}$ is convex with respect to $f_k$ and $p_k$, therefore, it can be solved by using a standard convex optimization tool.
\subsection{Optimizing the Receiving Beamforming}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
In this section, we solve the problem $\textbf{P}_{2}$ to achieve the receive beamforming vector $\mathbf{m}_{B,k}$ for a given $\overline{\mathbf{w}}$, $p_k$, and $f_k$.
Let $\mathbf{\overline{h}}_{B,k}^H=\mathbf{\overline{w}}^H\mathbf{\overline{H}}_{B,k}$, $\textbf{P}_{2}$ can be transformed into
\begin{subequations}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{alignat}{5}
\label{p3.2}
\begin{split}
\textbf{P}_{3.2}:~ &\max_{\mathbf{m}_{B,k}}~\frac{\ln2}{B}\ln(a_0\sum_{k=1}^{K}p_k|\mathbf{\overline{h}}_{B,k}^H\mathbf{m}_{B,k}|^2\\
&+1)+\sum_{k=1}^KR_k^{loc}-\eta_{1}\sum_{k=1}^KP_k^{tot}
\end{split}\nonumber\\
\begin{split}
s.t.~~&\frac{B}{\ln2}\ln(a_0\sum_{i=1}^{k}p_i|\mathbf{\overline{h}}_{B,k}^H\mathbf{m}_{B,i}|^2+1)+ \ln ({t_{B,k}}) + 1\\
& - {t_{B,k}}({a_0}\sum\limits_{i = 1}^{k - 1} {{p_i}} |{\mathbf{\overline{h}}_{B,i}^H{{\bf{m}}_{B,i}}}{|^2} + 1) +R_k^{loc}\ge R_{th},
\end{split}\\
&\left\| \mathbf{m}_{B,k} \right\|^2=1.
\end{alignat}
\end{subequations}
Let $|\overline{\mathbf{h}}_{B,k}^H\mathbf{m}_{B,k}|^2 = \rm{Tr}(\tilde{\mathbf{H}}_{B,k}\mathbf{m}_{B,k}\mathbf{m}_{B,k}^H)$. By defining $\tilde{\mathbf{H}}_{B,k}=\mathbf{\overline{h}}_{B,k}\mathbf{\overline{h}}_{B,k}^H$, $\mathbf{M}_{B,k} = \mathbf{m}_{B,k}\mathbf{m}_{B,k}^H$, one has $\mathbf{M}_{B,k}\succeq0$ and $\text{rank}(\mathbf{{M}}_{B,k})=1$.
The rank-$1$ constraint makes the problem difficult to solve. Thus, we apply the SDR method to relax the constraints \cite{irsqun}. $\textbf{P}_{3.2}$ is then expressed as
\begin{subequations}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{alignat}{5}
\label{p3.3}
\begin{split}
\textbf{P}_{3.3}:~ &\max_{\mathbf{M}_{B,k}}~[\frac{B}{\ln2}\ln(a_0\sum_{k=1}^{K}p_k\rm{Tr}(\tilde{\mathbf{H}}_{B,k}{\mathbf{M}_{B,k}})\\
&+1) +\sum_{k=1}^KR_k^{loc}]-\eta_{1}\sum_{k=1}^KP_k^{tot}
\end{split}\nonumber\\
\begin{split}
s.t.~~&\frac{B}{\ln2}\ln(a_0\sum_{i=1}^{k}p_i\rm{Tr}(\tilde{\mathbf{H}}_{B,i}{\mathbf{M}_{B,i}})+1)+ \ln ({t_{B,k}}) + 1\\
& - {t_{B,k}}({a_0}\sum\limits_{i = 1}^{k - 1} {{p_i}} \rm{Tr}(\tilde{\mathbf{H}}_{B,i}{\mathbf{M}_{B,i}}) + 1) +R_k^{loc}\ge R_{th},
\end{split}\\
&\rm{Tr}({\mathbf{M}_{B,k}})=1.
\end{alignat}
\end{subequations}
$\textbf{P}_{3.3}$ is convex and can be solved by using a standard convex optimization tool \cite{wqqirs}.
After $\mathbf{M}_{B,k}$ is obtained, if $\text{rank}(\mathbf{M}_{B,k})=1$, $\mathbf{m}_{B,k}$ can be obtained from $\mathbf{M}_{B,k}=\mathbf{m}_{B,k}\mathbf{m}_{B,k}^H$ by performing the eigenvalue decomposition. Otherwise, the Gaussian randomization can be used for recovering $\mathbf{m}_{B,k}$ \cite{wqqirs}.
\subsection{Optimizing the IRS Reflecting Shifts $\overline{\mathbf{w}}$}
After obtaining the beamforming vectors $\mathbf{m}_{B,k}$, by setting $\mathbf{h}_{W,B,k}=\mathbf{\overline{H}}_{B,k}\mathbf{m}_{B,k}$, problem $\textbf{P}_{2}$ can be transformed into
\begin{subequations}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{alignat}{5}
\begin{split}
\textbf{P}{_{3.4}:}~ &\max_{\overline{\mathbf{w}}}~\frac{B}{\ln2}\ln(a_0\sum_{k=1}^{K}p_k(\overline{\mathbf{w}}^H\mathbf{h}_{W,B,k})+1)\\
& +\sum_{k=1}^K\frac{f_k}{C_k} -\eta_{1}\sum_{k=1}^KP_k^{tot}
\end{split}\nonumber\\
s.t.~~& |w_{m}|=1,~m=1,2,...,M,\\
\begin{split}
&\ln(a_0\sum_{i=1}^{k}p_i(\overline{\mathbf{w}}^H\mathbf{h}_{W,B,i})+1)+ \ln ({t_{B,k}}) + 1 \\
&- {t_{B,k}}({a_0}\sum\limits_{i = 1}^{k - 1} {{p_i}} (\overline{\mathbf{w}}^H\mathbf{h}_{W,B,i}) + 1) +\frac{f_k}{C_k} \ge R_{th}.
\end{split}
\end{alignat}
\end{subequations}
Similar to the previous section, let $\mathbf{W}=\overline{\mathbf{w}}\overline{\mathbf{w}}^H$, $\mathbf{H}_{W,B,k}=\mathbf{h}_{W,B,k}\mathbf{h}_{W,B,k}^H$. By applying the SDR method, we have
\begin{subequations}
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\begin{alignat}{5}
\begin{split}
\textbf{P}{_{3.5}:}~ &\max_{\mathbf{W}}~\frac{B}{\ln2}\ln(a_0\sum_{k=1}^{K}p_k\text{Tr}(\mathbf{H}_{W,B,k}\mathbf{W})+1)\\
& +\sum_{k=1}^K\frac{f_k}{C_k} -\eta_{1}\sum_{k=1}^KP_k^{tot}
\end{split}\nonumber\\
s.t.~~& \mathbf{W} \succeq 0, \mathbf{W}_{mm}=1,~m=1,2,...,M,\\
\begin{split}
&\ln(a_0\sum_{i=1}^{k}p_i\text{Tr}(\mathbf{H}_{W,B,k}\mathbf{W})+1)+ \ln ({t_{B,k}}) + 1 \\
&- {t_{B,k}}({a_0}\sum\limits_{i = 1}^{k - 1} {{p_i}} \text{Tr}(\mathbf{H}_{W,B,k}\mathbf{W}) + 1)+\frac{f_k}{C_k} \ge R_{th}.
\end{split}
\end{alignat}
\end{subequations}
The problem $\textbf{P}_{3.5}$ is a convex problem and can be solved by using a standard convex optimization tool.
After obtaining $\mathbf{W}$, $\overline{\mathbf{w}}$ can be given by eigenvalue decomposition if $\text{rank}({\mathbf{W}})=1$; otherwise, the Gaussian randomization can be used for recovering the approximate $\overline{\mathbf{w}}$ \cite{wqqirs}. The reflection coefficients can be given by $w_m=\angle(\frac{\overline{{w}}_m}{\overline{{w}}_{M+1}}),~m=1,2,..,M$.
The overall optimization algorithm is summarized in Algorithm 1, where $\delta$ is the threshold and $T$ is the maximum number of iterations
\begin{table}[htbp]
\setlength{\abovedisplayskip}{3pt}
\setlength{\belowdisplayskip}{3pt}
\centering
\begin{center}
\begin{tabular}{lcl}
\\\toprule
{ $\textbf{Algorithm 1}$: Alternating Algorithm for Solving $\textbf{P}_{1}$}\\ \midrule
\ 1) \textbf{Input settings:}\\
\ \ \ \ \ \ \ $\delta$, $R_{th},P_k^{th} >0$, and $T$.\\
\ 2) \textbf{Initialization:}\\
\ \ \ \ \ \ \ $t_{B,k}(0)$, $\mathbf{\overline{w}}(0)$,$\mathbf{m}_{B,k}(0)$, and $\eta_1(0)$;\\
\ 3) \textbf{Optimization:}\\
\ \ \ \ \ \textbf{$\pmb{\unrhd} $ for \emph{$\tau_1$}=1:T }\\
\ \ \ \ \ \ \ \ \ \ solve $\textbf{P}_{3.1}$ with $\mathbf{\overline{w}}^{*}(\tau_1-1)$,$\mathbf{m}_{B,k}^*(\tau_1-1)$;\\
\ \ \ \ \ \ \ \ \ \ obtain the solution $p_k^*(\tau_1)$, $f_k^*(\tau_1)$;\\
\ \ \ \ \ \ \ \ \ \ solve $\textbf{P}_{3.3}$ with $p_k^*(\tau_1)$, $f_k^*(\tau_1)$, and $\mathbf{\overline{w}}^{*}(\tau_1-1)$;\\
\ \ \ \ \ \ \ \ \ \ obtain the solution $\mathbf{m}_{B,k}^*(\tau_1)$;\\
\ \ \ \ \ \ \ \ \ \ solve $\textbf{P}_{3.5}$ with $p_k^*(\tau_1)$, $f_k^*(\tau_1)$, and $\mathbf{m}_{B,k}^*(\tau_1)$;\\
\ \ \ \ \ \ \ \ \ \ obtain the solution $\mathbf{\overline{w}}^{*}(\tau_1)$;\\
\ \ \ \ \ \ \ \ \ \ calculate EE $\eta(\tau_1)$ and update $t_{B,k}(\tau_1)$ and $\eta_1(\tau_1)$;\\
\ \ \ \ \ \ \ \ \ \ \textbf{if} $| \frac{\eta(\tau_1)-\eta(\tau_1-1)}{\eta(\tau_1)}| \le \delta$; \\
\ \ \ \ \ \ \ \ \ \ \ \ \ \ the optimal EE $\eta^*$ is obtained;\\
\ \ \ \ \ \ \ \ \ \ \textbf{end}\\
\ \ \ \ \ \textbf{$\pmb{\unrhd} $ end} \\
\ 4) \textbf{Output:}\\
\ \ \ \ \ \ \ $p_k^*$, $f_k^*$, $\mathbf{m}_{B,k}^*$, and $\mathbf{w}^{*}$ and EE $\eta^*$.\\
\bottomrule
\end{tabular}
\end{center}
\end{table}
\section{Simulation Results}
In this section, simulation results are provided to evaluate the performance of the proposed algorithms. We consider a three-dimensional Cartesian coordinate system. The simulation settings are based on those used in \cite{irs1}, \cite{wqqirs}.
We consider a $2$-UE case and it can be readily extend to multiple UE cases. The locations of the MEC, the IRS, UE1, and UE2 are set as $(5,0,20)$, $(0,50,2)$, $(5,75,5)$ and $(5,50,10)$, respectively \cite{wqqirs}.
The channels are generated by $h_{i,j}=\sqrt{G_0d_{i,j}^{-c_{i,j}}}g_{i,j}$, where $G_0=-30$ dB is the path loss at the reference point. $d_{i,j}$, $c_{i,j}$ and $g_{i,j}$ denote the distance, path loss exponent, and fading between $i$ and $j$, respectively, where $i \in \{B, I\}$ and $j \in \{U,k\}$. The path loss exponents are set as $c_{B,U,k}=5$, $c_{B,I}=3.5$, and $c_{I,U,k}=2$.
The bandwidth $B$ is set to $1$ Mhz. Other parameters are set as $\sigma^2= -105$ dBm, $P_k^{th}=31$ dBm, $P_k^{cn}=23$ dBm, $C_k=10^3$ cycles/bit, and $\epsilon=10^{-28}$.
The proposed scheme is marked as `Efficiency-IRS'. We consider three other cases as benchmarks to compare with the proposed method. The first benchmark, marked as `OMA-IRS', uses FDMA with equally allocated bandwidth to all the users. The second benchmark, marked as `OnlyOff-IRS', has no local computing and all the tasks are offloaded. The third benchmark, marked as `Efficiency-NoIRS", aims to investigate the performance without IRS.
\begin{figure}[h]
\setlength{\abovecaptionskip}{-0.2cm}
\setlength{\belowcaptionskip}{-1cm}
\centering
\includegraphics[width=3.0in]{etarth.eps}
\caption{Energy efficiency versus the minimum rate threshold.\label{etarth}}
\end{figure}
Fig. \ref{etarth} shows EE versus the minimum rate threshold $R_{th}$.
The EE achieved by the proposed method is the best among all the schemes. This indicates that the IRS assisted MEC with NOMA can help improve the system rate and achieve high EE.
With the increase of $R_{th}$, all the curves are decreasing. The system has to consume excessive energy to increase the rate in order to meet the minimum rate constraint, which decreases EE.
\begin{figure}[h]
\setlength{\abovecaptionskip}{-0.2cm}
\setlength{\belowcaptionskip}{-1cm}
\centering
\includegraphics[width=3.0in]{raterth.eps}
\caption{Achievable rate versus the minimum rate threshold.\label{raterth}}
\end{figure}
A comparison of the system rate versus the rate threshold $R_{th}$ is presented in Fig. \ref{raterth}. All the curves increase with $R_{th}$ in order to meet the service requirement, which verifies the observation in Fig. \ref{etarth}.
The system rates obtained by the proposed method and the `OnlyOff-IRS' method are higher than those of the other two methods, which indicates that combining IRS with NOMA can significantly help the system to achieve a higher rate. It is worth noting that even though the `OnlyOff-IRS' method can achieve the highest rate when $R_{th}$ is low, its efficiency is lower than the proposed method. This indicates that the overall efficiency performance degrades when there is no local computing.
\begin{figure}[h]
\setlength{\abovecaptionskip}{-0.2cm}
\setlength{\belowcaptionskip}{-1cm}
\centering
\includegraphics[width=3.0in]{powerrth.eps}
\caption{Power consumption versus the minimum rate threshold.\label{powerrth}}
\end{figure}
\begin{figure}[h]
\setlength{\abovecaptionskip}{-0.2cm}
\setlength{\belowcaptionskip}{-1cm}
\centering
\includegraphics[width=3.0in]{irsdis.eps}
\caption{Energy efficiency versus the relative distance of UE-IRS.\label{etairs}}
\end{figure}
\begin{figure}[h]
\setlength{\abovecaptionskip}{-0.2cm}
\setlength{\belowcaptionskip}{-1cm}
\centering
\includegraphics[width=3.0in]{iters.eps}
\caption{Convergence with Iteration.\label{iters}}
\end{figure}
Fig. \ref{powerrth} presents the power consumption versus $R_{th}$ for different methods. The results of all the methods in Fig. \ref{powerrth} are consistent with what are shown in Fig. \ref{etarth} and Fig. \ref{raterth}. It is worth noting that the power consumption by the proposed method is quite low. So UEs can use less energy to achieve a higher rate, which demonstrates the advantage of combining NOMA and IRS to MEC network in improving EE.
Fig. \ref{etairs} shows EE versus the distance between UEs and IRS. The distance is the relative increased amount compared with UEs' original position. The curves for all the methods with IRS decrease with the increase of the distance except `Efficiency-NoIRS'. This is because the increase of the distance results in the increase of the path loss and the reduction of the power gain from the reflecting path through the IRS. Therefore, the achievable rate and EE are both decreased. It can also be seen that the `Efficiency-IRS' method still has the highest performance among all the methods, which validates the superiority of the proposed design.
In Fig. \ref{iters}, the coverage of proposed methods based on different $R_{th}$ setting are investigated. It can be observed from Fig. \ref{iters} that only several iterations are needed for the proposed algorithms to converge, showing the computation efficiency of the proposed algorithm.
\section{Conclusions}
In this paper, an IRS-assisted MEC network with NOMA was considered. EE was maximized by jointly optimizing the offloading power, local computing frequency, beamforming vectors, and IRS phase shift matrix. An alternating algorithm was proposed to solve the challenging non-convex fractional optimization problems. The numerical results showed that our proposed method outperforms other benchmark schemes in terms of EE. It was proved that NOMA and IRS could help the MEC network to achieve a higher rate with a lower power. The convergence of the proposed algorithm was also verified.
|
\section{#1}}
\renewcommand{\theequation}{\thesection.\arabic{equation}}
\newcommand{\reff}[1]{(\ref{#1})}
\newcommand{\vl}[2]{\rule[#1]{0.1mm}{#2}}
\newcommand{\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}}{\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}}
\newcommand{\!\stackrel{\textstyle{\hst{0ex} \atop \circ}}{\hst{0ex}}\!}{\!\stackrel{\textstyle{\hst{0ex} \atop \circ}}{\hst{0ex}}\!}
\newcommand{\Gamma}{\Gamma}
\newcommand{\Delta}{\Delta}
\newcommand{\alpha}{\alpha}
\newcommand{\beta}{\beta}
\newcommand{\gamma}{\gamma}
\newcommand{\delta}{\delta}
\newcommand{\varepsilon}{\varepsilon}
\newcommand{\varrho}{\varrho}
\newcommand{\sigma}{\sigma}
\newcommand{\varphi}{\varphi}
\newcommand{\omega}{\omega}
\newcommand{\Gamma^{\ast}}{\Gamma^{\ast}}
\newcommand{\Co\Gas}{\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}\Gamma^{\ast}}
\newcommand{\Co\Ga}{\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}\Gamma}
\newcommand{\langle}{\langle}
\newcommand{\rangle}{\rangle}
\newcommand{\rightarrow}{\rightarrow}
\newcommand{\longrightarrow}{\longrightarrow}
\newcommand{\otimes}{\otimes}
\begin{document}
\begin{titlepage}
\hspace*{\fill} \parbox{8em}{BONN--TH--95--16 \\
August 1995 }
\vspace{18mm}
\begin{center}
{\LARGE\bf Universal $R$--matrices \\[-1.5ex]
for finite Abelian groups --- \\[1ex]
a new look at graded multilinear algebra} \\
\vspace{11mm}
{\large M. Scheunert} \\
\vspace{5mm}
Physikalisches Institut der Universit\"{a}t Bonn \\
Nu{\ss}allee 12, D--53115 Bonn, Germany
\end{center}
\vspace{20mm}
\begin{abstract}
\noindent
The universal $R$--matrices and, dually, the coquasitriangular structures
of the group Hopf algebra of a finite Abelian group (resp. of an arbitrary
Abelian group) are determined. This is used to formulate graded multilinear
algebra in terms of triangular or cotriangular Hopf algebras. For the
convenience of the reader, in a separate section the definitions and basic
properties of quasitriangular and coquasitriangular Hopf algebras are
recalled.
\end{abstract}
\vspace{\fill}
q-alg/9508016
\end{titlepage}
\setcounter{page}{2}
\section{Introduction}
In the present work I am going to discuss the universal $R$--matrices [1]
for finite Abelian groups, or, more precisely, the universal $R$--matrices
for the group Hopf algebras of these groups [2]. To explain how I became
interested in this problem I remind the reader of the graded multilinear
algebra described in Ref. [3]. As shown in this reference, given an
(arbitrary) Abelian group $\Gamma$ and a commutation factor $\sigma$ on $\Gamma$,
the basic constructions of classical multilinear algebra can easily be
generalized to the $\Gamma$--graded setting. In modern terms, this means that
the $\Gamma$--graded vector spaces form a tensor category (depending on $\sigma$)
with an internal hom, and this category is rigid if we consider only
finite--dimensional vector spaces. It is this category where the so--called
Lie colour algebras (also called $\sigma$ Lie algebras) live [4,\,5], and that
is why I have investigated these structures.
Since the present discussion is only meant to serve as a motivation, I do not
want to explain the various category theoretical terms used above, but refer
to Refs. [6,\,7,\,8], where more details are given. However, I would like to
emphasize that the properties of the categories in question are closely
related to the axioms for a triangular Hopf algebra [1]. In fact, given any
triangular Hopf algebra $H$, the (left) $H$--modules form a category of this
type, and conversely, under suitable assumptions, such a category is
equivalent to the category of $H$--modules for some triangular Hopf algebra
$H$ (this is the content of various so--called reconstruction theorems).
In the simple case under consideration we do not want to invoke one of the
reconstruction theorems in order to find the appropriate triangular Hopf
algebra. Instead, we recall that, for a finite Abelian group $\Gamma$, a
$\Gamma$--graded vector space is nothing but a vector space endowed with a
representation of $\Gamma^{\ast}$, the character group of $\Gamma$ (see Section 2). Thus
the group Hopf algebra $\Co\Gas$ of $\Gamma^{\ast}$ is the natural candidate for $H$,
and we have to determine the triangular universal $R$--matrices for $\Co\Gas$.
(Recall that, for a finite Abelian group $\Gamma$, the groups $\Gamma$ and $\Gamma^{\ast}$
are isomorphic.) Actually, we do more and determine all the universal
$R$--matrices for $\Co\Gas$. From the category theoretical point of view this
means the transition from (symmetric) tensor categories to quasitensor or
braided monoidal categories. The outcome is completely satisfactory: There
exists a simple bijection of the set of bicharacters on $\Gamma$ onto the set of
universal $R$--matrices for $\Co\Gas$, and this bijection maps the commutation
factors onto the triangular universal $R$--matrices.
The proof given in Section 2 is elementary, it only uses some simple
properties of finite Abelian groups and their characters. Nevertheless,
it is worth--while to consider the problem from a dual point of view. It is
easy to see that the group Hopf algebras $\Co\Ga$ and $\Co\Gas$ are dual to each
other. Thus the $\Gamma$--graded vector spaces, i.e., the $\Co\Gas$--modules, can
be identified with the $\Co\Ga$--comodules, and the universal $R$--matrices for
$\Co\Gas$ correspond to coquasitriangular structures on $\Co\Ga$. Once this has
been realized, the results summarized above follow immediately. Moreover,
in the dual language, these results hold also for not necessarily finite
Abelian groups. All this will be shown in Section 4, after some of the
notions used there (and above) have been explained in the preparatory
Section 3. The final Section 5 contains some concluding remarks and draws
the reader's attention to related work by other authors.
In closing I would like to mention some notations and conventions which will
be used throughout this work. The base field will always be the field $\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$
of complex numbers, and the multiplicative group of non--zero complex
numbers will be denoted by $\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}_{\ast}$\,. Moreover, $\Gamma$ will be an Abelian
group which, for convenience, will be written {\it multiplicatively}. A
character of $\Gamma$ is a homomorphism of $\Gamma$ into $\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}_{\ast}$\,. The
characters of $\Gamma$ form a (multiplicative) group $\Gamma^{\ast}$, and the canonical
pairing of $\Gamma^{\ast}$ and $\Gamma$ will be denoted by a pointed bracket:
\be \langle\,\;,\;\rangle : \Gamma^{\ast} \times \Gamma \longrightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}_{\ast} \end{equation}
\be \langle\gamma',\gamma\rangle = \gamma'(\gamma)
\;\;\mbox{for all}\;\; \gamma'\in\Gamma^{\ast}\,,\;\gamma\in\Gamma. \end{equation}
Of course, if $\Gamma$ is finite, its characters take their values in the unit
circle $U(1)$.
\section{Universal $R$--matrices for the group
\newline
Hopf algebra of a finite Abelian group}
\setcounter{equation}{0}
{\it In the present section the group} $\Gamma$ {\it is assumed to be finite.}
Let us first show that there exists a bijective correspondence between
$\Gamma$--graded vector spaces and $\Gamma^{\ast}$--modules, i.e., vector spaces endowed
with a representation of $\Gamma^{\ast}$. In fact, suppose $V$ is a $\Gamma^{\ast}$--module
and let $\gamma'\cdot\,x$ denote the action of an element $\gamma'\in\Gamma^{\ast}$ on an
element $x\in V$. Since $\Gamma^{\ast}$ is finite, any representation of $\Gamma^{\ast}$ is
completely reducible, and since $\Gamma^{\ast}$ is Abelian, its irreducible
representations are one--dimensional, hence given by its characters, i.e.,
by the elements of $\Gamma$. This implies that $V$ has the decomposition
\be V = \bigoplus_{\gamma\in\Gamma}V_{\gamma}\:, \label{dec} \end{equation}
where the subspaces $V_{\gamma}$ of $V$ are defined by
\be V_{\gamma} = \{x \in V\,|\,\gamma'\cdot\,x = \langle\gamma',\gamma\rangle\,x
\;\;\mbox{for all}\;\;\gamma'\in\Gamma^{\ast}\}\,. \end{equation}
Conversely, suppose that $V$ is a $\Gamma$--graded vector space, i.e., suppose
that a decomposition like \reff{dec} is given. Then the action of $\Gamma^{\ast}$ on
$V_{\gamma}$ defined by
\be \gamma'\cdot\,x = \langle\gamma',\gamma\rangle\,x\;\;\mbox{for all}\;\;\gamma'\in\Gamma^{\ast}\,,\;
x\in V_{\gamma} \end{equation}
makes each of the subspaces $V_{\gamma}$ of $V$ into a $\Gamma^{\ast}$--module, and
$V$ can be considered as the direct sum of these modules. It is easy to see
that these transitions from $\Gamma^{\ast}$--modules to $\Gamma$--graded vector spaces
and vice versa are inverse to each other. Moreover, the homomorphisms of
$\Gamma^{\ast}$--modules are exactly the homomorphisms of $\Gamma$--graded vector spaces,
i.e., the linear mappings which are homogeneous of degree zero.
Now let $\Co\Gas$ be the group algebra of $\Gamma^{\ast}$. Recall that $\Co\Gas$ is the
associative algebra consisting of the formal linear combinations of elements
of $\Gamma^{\ast}$, with the multiplication inherited from $\Gamma^{\ast}$, recall also that
$\Gamma^{\ast}$--modules and $\Co\Gas$--modules are essentially the same thing. It is
well--known [2] that $\Co\Gas$ is a Hopf algebra, where the coproduct $\Delta$,
the counit $\varepsilon$, and the antipode $S$ are given (on the basis
$\Gamma^{\ast}$ of $\Co\Gas$) by
\begin{eqnarray} \Delta(\gamma') \!\! & = & \!\! \gamma'\otimes\gamma' \\
\varepsilon(\gamma') \!\! & = & \!\! 1 \\
S(\gamma') \!\! & = & \!\! \gamma'^{-1} \end{eqnarray}
for all $\gamma'\in\Gamma^{\ast}$.
Let us now proceed to the determination of the universal $R$--matrices for
$\Co\Gas$. To find a good Ansatz, let $\sigma$ be a commutation factor on $\Gamma$
and let $V$ and $W$ be two $\Gamma$--graded vector spaces. Then, according to
Ref. [3], the canonical isomorphism
\be \psi_{V,W} : V \otimes W \longrightarrow W \otimes V \end{equation}
is given by
\be \psi_{V,W}(x \otimes y) = \sigma(\alpha,\beta)\,y \otimes x \label{col} \end{equation}
for all $x \in V_{\alpha}$\,, $y \in W_{\beta}$\,; $\alpha,\beta \in \Gamma$.
On the other hand, for a triangular Hopf algebra $H$ with universal
$R$--matrix
\be R = \sum_{i}R^1_i \otimes R^2_i \end{equation}
(with $R^1_i, R^2_i \in H$) and for any two $H$--modules $V$ and $W$, the
corresponding braiding isomorphism is fixed by
\be \psi_{V,W}(x \otimes y) = \sum_{i}R^2_i\cdot y \otimes R^1_i\cdot x
\label{braid} \end{equation}
for all $x \in V$, $y \in W$, where, as before, the dot denotes the action of
an element of $H$ on some element of an $H$--module.
In our case, we have $H = \Co\Gas$, and a general element of $\Co\Gas \otimes \Co\Gas$
can be uniquely written in the form
\be R = \sum_{\alpha',\,\beta'\in \Gamma^{\ast}}c_{\alpha',\,\beta'}\,\alpha'\otimes \beta' \; , \end{equation}
with some coefficients $c_{\alpha',\,\beta'} \in \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$\,. Inserting this expression
into Eq. \reff{braid} and equating the right hand sides of Eqs. \reff{col}
and \reff{braid} we obtain the condition
\be \sigma(\alpha,\beta) = \sum_{\alpha',\,\beta'}c_{\alpha',\,\beta'}\, \langle\alpha',\alpha\rangle
\langle\beta',\beta\rangle \label{cond} \end{equation}
for all $\alpha,\beta \in \Gamma$.
Using the well--known orthogonality and completeness relations for the
characters, i.e.,
\be \sum_{\alpha \in \Gamma}\langle\alpha',\alpha\rangle\langle\beta',\alpha\rangle^{\ast}
= n \delta_{\alpha',\,\beta'} \label{orth} \end{equation}
\be \sum_{\alpha' \in \Gamma^{\ast}}\langle\alpha',\alpha\rangle\langle\alpha',\beta\rangle^{\ast}
= n \delta_{\alpha,\,\beta}\;, \label{compl} \end{equation}
where $\alpha',\beta' \in \Gamma^{\ast}$ and $\alpha,\beta \in \Gamma$ and where $n$ denotes the
order of $\Gamma$, we can solve Eq. \reff{cond} for the coefficients
$c_{\alpha',\,\beta'}$ and obtain the following Ansatz for the universal
$R$--matrix
correseponding to $\sigma$\,:
\be R_{\sigma} = \frac{1}{n^2}\sum_{\alpha,\beta \in \Gamma \atop \alpha',\,\beta' \in \Gamma^{\ast}}
\sigma(\alpha,\beta)\langle\alpha',\alpha\rangle^{\ast}\langle\beta',\beta\rangle^{\ast}\,
\alpha' \otimes \beta'\,. \label{ans} \end{equation}
It turns out that Eq. \reff{ans} gives a nice parametrization of the elements
of $\Co\Gas \otimes \Co\Gas$. Let ${\cal F}(\Gamma \times \Gamma)$ be the algebra of all
complex--valued functions on $\Gamma \times \Gamma$. Then we can prove the
following proposition. \\
\noindent
{\bf Proposition 2.1} \\
{\it 1) The mapping $\sigma \rightarrow R_{\sigma}$ of ${\cal F}(\Gamma \times \Gamma)$ into}
$\Co\Gas \otimes \Co\Gas$ {\it given by Eq. \reff{ans} is an algebra isomorphism.}
\noindent
In the subsequent statements, $\sigma$ denotes an arbitrary element of
${\cal F}(\Gamma \times \Gamma)$. \\
{\it 2) We have
\be (\Delta \otimes id)(R_{\sigma}) = (R_{\sigma})_{13}(R_{\sigma})_{23} \label{UR1} \end{equation}
if and only if
\be \sigma(\alpha\beta,\gamma) = \sigma(\alpha,\gamma)\sigma(\beta,\gamma)\;\;\mbox{\it for all}\;\;
\alpha,\beta,\gamma \in \Gamma\,. \label{bich1} \end{equation}
3) We have
\be (id \otimes \Delta)(R_{\sigma}) = (R_{\sigma})_{13}(R_{\sigma})_{12} \label{UR2} \end{equation}
if and only if
\be \sigma(\alpha,\beta\gamma) = \sigma(\alpha,\beta)\sigma(\alpha,\gamma)\;\;\mbox{\it for all}\;\;
\alpha,\beta,\gamma \in \Gamma\,. \label{bich2} \end{equation}
4) Let} $T:\Co\Gas \otimes \Co\Gas \longrightarrow \Co\Gas \otimes \Co\Gas$ {\it be the usual twist mapping
and define the function $\sigma^{T} \in {\cal F}(\Gamma \times \Gamma)$ by
\be \sigma^{T}(\alpha,\beta) = \sigma(\beta,\alpha)\;\;\mbox{\it for all}\;\;
\alpha,\beta \in \Gamma\,. \end{equation}
Then we have
\be T(R_{\sigma}) = R_{\sigma^{T}}\,. \end{equation}}
Proof \\
By direct calculations, using Eqs. \reff{orth} and \reff{compl}. (The
notation $(R_{\sigma})_{12}$ etc. is standard, see Section 3.) \\
Finally, we note that
\be R_{\sigma}\Delta(h) = (T\Delta(h))R_{\sigma} \label{UR3} \end{equation}
for all $h \in \Co\Gas$ and all $\sigma \in {\cal F}(\Gamma \times \Gamma)$. This is
obvious since $\Co\Gas$ is commutative and cocommutative.
The foregoing results immediately imply the assertion made in the
introduction. \\
\noindent
{\bf Corollary 2.1} \\
{\it Let $\sigma$ be an arbitrary complex--valued function on $\Gamma \times \Gamma$.
Then $R_{\sigma}$ is a universal $R$--matrix for} $\Co\Gas$ {\it if and only if
$\sigma$ is a bicharacter on $\Gamma$. Any universal $R$--matrix for} $\Co\Gas$
{\it can uniquely be written in this form, and $R_{\sigma}$ is triangular if
and only if $\sigma$ is skew--symmetric in the sense that
\be \sigma(\alpha,\beta)\sigma(\beta,\alpha) = 1\;\;\mbox{for all}
\;\;\alpha,\beta \in \Gamma\,. \label{skew} \end{equation}
The braiding isomorphisms corresponding to a bicharacter $\sigma$ and to the
associated universal $R$--matrix $R_{\sigma}$ coincide.}
Proof \\
By definition, $R_{\sigma}$ is a universal $R$--matrix if it has an inverse and
satisfies the Eqs. \reff{UR1}, \reff{UR2}, and \reff{UR3}. This is the case
if and only if $\sigma$ takes its values in $\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}_{\ast}$ and satisfies the Eqs.
\reff{bich1} and \reff{bich2}, i.e., if and only if $\sigma$ is a bicharacter
on $\Gamma$. Moreover, $R_{\sigma}$ is triangular if and only if
\be T(R_{\sigma}) = R^{-1}_{\sigma}\,, \end{equation}
i.e., if and only if
\be \sigma^{T} = \sigma^{-1}\,. \end{equation}
This is exactly Eq. \reff{skew}. The last statement of the corollary has
been used to find the correct Ansatz for $R_{\sigma}$\,. \\
{}From a calculational point of view, the parametrization \reff{ans} of the
universal $R$--matrices of $\Co\Gas$ is not quite convenient, for it involves
the detour over $\Gamma$ and hence a fourfold sum. This drawback can easily be
avoided, as follows. \\
\noindent
{\bf Proposition 2.2} \\
{\it Let $\Delta_1$ and $\Delta_2$ be two subgroups of $\Gamma^{\ast}$ such that there
exists a non--degenerate pairing}
\be \tau : \Delta_1 \times \Delta_2 \longrightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}_{\ast}\,. \end{equation}
{\it (Since $\Gamma^{\ast}$ is finite this is the case if and only if $\Delta_1$ and
$\Delta_2$
are isomorphic.) Let $m$ be the common order of $\Delta_1$ and $\Delta_2$\,. Then
the equation
\be \sigma(\alpha,\beta) = \frac{1}{m}\sum_{\alpha_i \in \Delta_i}\tau(\alpha_1,\alpha_2)
\langle\alpha_1,\alpha\rangle\langle\alpha_2,\beta\rangle \end{equation}
defines a bicharacter $\sigma$ on $\Gamma$ and
\be R^{\tau} = \frac{1}{m}\sum_{\alpha_i \in \Delta_i}\tau(\alpha_1,\alpha_2)\,
\alpha_1 \otimes \alpha_2 \label{Rtau} \end{equation}
is the corresponding universal $R$--matrix $R_{\sigma}$ for} $\Co\Gas$. {\it Every
bicharacter on $\Gamma$ and hence every universal $R$--matrix for} $\Co\Gas$ {\it
can uniquely be obtained in this way.}
Proof \\
Needless to say, a non--degenerate pairing $\tau$ of two Abelian groups
$\Delta_1$ and $\Delta_2$ is a bicharacter
\be \tau : \Delta_1 \times \Delta_2 \longrightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}_{\ast} \end{equation}
such that, for any $\alpha_1 \in \Delta_1$\,, the condition
\be \tau(\alpha_1,\alpha_2) = 1\;\;\mbox{for all}\;\;\alpha_2 \in \Delta_2 \end{equation}
implies that $\alpha_1$ is the unit element of $\Delta_1$\,, and likewise with the
roles of $\Delta_1$ and $\Delta_2$ interchanged.
The proof of the proposition is quite straightforward, it follows from the
standard duality theory for finite Abelian groups. Thus we shall only give
a few hints.
Let $\sigma$ be a bicharacter on $\Gamma$. Obviously, our task is to investigate
the function on $\Gamma^{\ast} \times \Gamma^{\ast}$ defined by
\be (\alpha',\beta') \longrightarrow \sum_{\alpha,\beta \in \Gamma}\sigma(\alpha,\beta)
\langle\alpha',\alpha\rangle^{\ast}\langle\beta',\beta\rangle^{\ast}\,. \end{equation}
Define the (left and right) kernels of $\sigma$ by
\begin{eqnarray} N_1 \!\! & = & \!\! \{\alpha \in \Gamma\,|\, \sigma(\alpha,\beta) = 1\;\;\mbox{for all}\;\;
\beta \in \Gamma \} \\
N_2 \!\! & = & \!\! \{\beta \in \Gamma\,|\, \sigma(\alpha,\beta) = 1\;\;\mbox{for all}\;\;
\alpha \in \Gamma \}\,. \end{eqnarray}
Then $\sigma$ induces a non--degenerate pairing $\tilde{\sigma}$ of $\Gamma/N_1$ and
$\Gamma/N_2$\,, hence the orders of these groups coincide. We denote them by $m$.
Define the subgroups $\Delta_1$ and $\Delta_2$ of $\Gamma^{\ast}$ by
\be \Delta_i = \{\gamma' \in \Gamma^{\ast}\,|\, \langle\gamma',\gamma_i\rangle = 1\;\;\mbox{for all}\;\;
\gamma_i \in N_i \}\,. \end{equation}
It is well--known that $\Delta_i$ can be identified with the character group of
$\Gamma/N_i$\,, thus $\tilde{\sigma}$ induces a non--degenerate pairing $\tau$ of
$\Delta_1$ and $\Delta_2$\,. This is the one we are looking for.
To be more explicit, define the complex--valued function $\sigma'$ on
$\Gamma^{\ast} \times \Gamma^{\ast}$ by
\be \sigma'(\alpha',\beta') = \frac{m}{n^2}\sum_{\alpha,\beta \in \Gamma}\sigma(\alpha,\beta)
\langle\alpha',\alpha\rangle^{\ast}\langle\beta',\beta\rangle^{\ast}\,, \end{equation}
where the normalization factor
\be \frac{m}{n^2} = \frac{1}{m}\frac{1}{(n/m)^2} \end{equation}
has been chosen conveniently. Using the duality theory for finite Abelian
groups it can be shown that $\sigma'(\alpha',\beta')$ vanishes whenever
$\alpha' \notin \Delta_1$ or $\beta' \notin \Delta_2$\,. Let $\tau$ be the restriction
of $\sigma'$ onto $\Delta_1 \times \Delta_2$\,. Then $\tau$ is the pairing described
abstractly above and has all the properties stated in the proposition. \\
\noindent
{\bf Remark} \\
The universal $R$--matrices given by Eq. \reff{Rtau} with
$\Delta_1 = \Delta_2 = \Gamma^{\ast}$ and $\tau$ a non--degenerate commutation factor on
$\Gamma^{\ast}$ have also been obtained in Ref. [9]. \\
\noindent
{\bf Example} \\
Let $n \ge 1$ be an integer, let $\Gamma_n$ be the (multiplicative) group of
all $n$th roots of unity, and let $Z_n$ be the ring of integers modulo $n$\,.
Obviously, for any $\alpha \in \Gamma_n$ and $r \in Z_n$\,, the power $\alpha^r$ has
a well--defined meaning. Define
\be \omega = e^{2\pi i/n} \end{equation}
and let $\gamma$ be a primitive $n$th root of unity, i.e., a generator of the
group $\Gamma_n$ (for example, $\gamma = \omega$). Let $\chi : \Gamma_n \to \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}_{\ast}$
be the character defined by
\be \langle\chi,\gamma^r\rangle = \omega^r\;\;\mbox{for all}\;\;r \in Z_n\,. \end{equation}
Then there exists a unique isomorphism of $\Gamma_n$ onto $\Gamma_{n}^{\ast}$
which maps $\gamma$ onto $\chi$\,, and we could use this isomorphism to
identify $\Gamma_{n}^{\ast}$ with $\Gamma_n$\,. The canonical pairing is given by
\be \langle\chi^r,\gamma^s\rangle = \omega^{rs}\;\;\mbox{for all}\;\;r,s \in Z_n\,, \end{equation}
and an arbitrary bicharacter on $\Gamma_n$ has the form
\be \sigma_k(\gamma^r,\gamma^s) = \omega^{krs}\;\;\mbox{for all}\;\;r,s \in Z_n\,, \end{equation}
with some element $k \in Z_n$\,. According to Eq. \reff{ans}, the universal
$R$--matrix $R_k \in \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}\Gamma_{n}^{\ast} \otimes \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}\Gamma_{n}^{\ast}$ corresponding
to $\sigma_k$ is given by
\be R_k = \frac{1}{n^2}\sum_{p,q,r,s \in Z_n}\omega^{kpq}\omega^{-pr}\omega^{-qs}
\chi^r \otimes \chi^s\,. \end{equation}
Obviously, part of this sum can be carried out. Consider, for a moment, $k$
as an element of $\{1,2,\ldots,n\}$ and let $d$ be the greatest common
divisor of $k$ and $n$\,. Let $\ell$ be the inverse of $k/d$ in the ring
$Z_{n/d}$\,. Then
\be R_k = \frac{d}{n}\sum_{x,\,y \in Z_{n/d}}(\omega^d)^{-\ell xy}
(\chi^d)^x \otimes (\chi^d)^y\,. \label{Rex} \end{equation}
Note that in this case the subgroups $\Delta_1$ and $\Delta_2$ of $\Gamma_{n}^{\ast}$
introduced in Proposition 2.2 are equal to
\be \Delta_1 = \Delta_2 = \{(\chi^d)^x\,|\, x \in Z_{n/d}\} \end{equation}
and are isomorphic to $\Gamma_{n/d}$\,. The universal $R$--matrix \reff{Rex}
with $k=1$ (and hence $d=1$, $\ell=1$) has already been used in Ref. [10],
the special case $n=2$ (related to supersymmetry) is contained in
Ref. [11]. \\
We would now like to interpret our results from a dual point of view.
However, this necessitates some preparation. Consequently, we shall first
comment on the duality of Hopf algebras and return to our main topic not
until Section 4.
\section{Duality of Hopf algebras and
\newline
coquasitriangular structures}
\setcounter{equation}{0}
The results of the present section are well--known to experts in Hopf algebra
theory. They are included for the convenience of the reader and for later
reference.
Let $A$ and $H$ be two bialgebras [2] and let
\be \varphi : A \times H \longrightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}} \end{equation}
be a bilinear form on $A \times H$. Then there exists a unique bilinear
form
\be \varphi^{\ot2}: (A \otimes A) \times (H \otimes H) \longrightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}} \end{equation}
such that
\be \varphi^{\ot2}(a \otimes a',h \otimes h') = \varphi(a,h)\,\varphi(a',h') \label{quad} \end{equation}
for all $a,a' \in A$ and $h,h' \in H$.
The form $\varphi$ is called a Hopf (or bialgebra) pairing of $A$ and $H$ if the
following relations are satisfied:
\begin{eqnarray} \varphi^{\ot2}(\Delta_{A}(a),h \otimes h') \!\! & = & \!\! \varphi(a,hh') \\
\varphi^{\ot2}(a \otimes a',\Delta_{H}(h)) \!\! & = & \!\! \varphi(aa',h) \end{eqnarray}
\vspace{-6ex}
\begin{eqnarray} \varphi(a,1_{H}) \!\! & = & \!\! \varepsilon_{A}(a) \\
\varphi(1_{A},h) \!\! & = & \!\! \varepsilon_{H}(h) \end{eqnarray}
for all $a,a' \in A$ and all $h,h' \in H$. Here, $\Delta_A$, $\varepsilon_A$, and
$1_A$ denote the coproduct, the counit, and the unit element of $A$\,, and
$\Delta_H$, $\varepsilon_H$, and $1_H$ have the analogous meaning.
{}From the point of view of dual Hopf algebras (see Ref. [2] and below) the
concept of a Hopf pairing is very natural and has been used by many authors.
The first reference I know of is Ref. [12].
The prototype of Hopf pairings can be constructed as follows. Let $H$ be a
bialgebra, let $H^{\ast}$ be the vector space dual to $H$, and let
$H^{\circ} \subset H^{\ast}$ be the set of all linear forms on $H$ which
vanish on an ideal of $H$ of finite codimension. It is easy to see that
$H^{\circ}$ is a subspace of $H^{\ast}$. Ref. [2] contains various
characterizations of the elements of $H^{\circ}$. In particular,
let $m:H \otimes H \rightarrow H$ be the product mapping of $H$ and let
$m^{t}: H^{\ast} \rightarrow (H \otimes H)^{\ast}$ be its transpose. Recall that
$H^{\ast} \otimes H^{\ast}$ is canonically embedded in $(H \otimes H)^{\ast}$. Then
an element $f \in H^{\ast}$ belongs to $H^{\circ}$ if and only if
$m^t(f) \in H^{\ast} \otimes H^{\ast}$.
Now let $\varphi$ be the restriction of the canonical pairing
\be \langle\,\;,\;\rangle:H^{\ast} \times H \longrightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}} \end{equation}
onto $H^{\circ} \times H$. Then there exists a unique bialgebra structure
on $H^{\circ}$ such that $\varphi$ is a Hopf pairing of $H^{\circ}$ and $H$ (for
a proof, see Ref. [2]). We call the bialgebra $H^{\circ}$ the Hopf dual
of $H$.
Using this result, the Hopf pairings can be characterized as follows. Let us
use the notation introduced at the beginning of this section. It is
well--known that, for any bilinear form $\varphi$ on $A \times H$, there exist
two linear mappings
\be \varphi_{\ell}:A \longrightarrow H^{\ast} \end{equation}
\be \varphi_{r}:H \longrightarrow A^{\ast} \end{equation}
such that
\be \langle\varphi_{\ell}(a),h\rangle = \langle\varphi_{r}(h),a\rangle = \varphi(a,h) \end{equation}
for all $a \in A$ and $h \in H$ (obviously, they are uniquely determined by
this equation). Then we can prove the following proposition. \\
\noindent
{\bf Proposition 3.1} \\
{\it We use the notation introduced above. Then the following statements are
equivalent: \\
1) The bilinear form $\varphi$ is a Hopf pairing of $A$ and $H$. \\
2) We have $\varphi_{\ell}(A) \subset H^{\circ}$, and the map of $A$ into
$H^{\circ}$ induced by $\varphi_{\ell}$ is a bialgebra homomorphism. \\
3) We have $\varphi_{r}(H) \subset A^{\circ}$, and the map of $H$ into
$A^{\circ}$ induced by $\varphi_{r}$ is a bialgebra homomorphism.}
Proof \\
The only non--trivial part of this proposition is that, for any Hopf pairing,
we have
\be \varphi_{\ell}(A) \subset H^{\circ}\;\;\;
\mbox{and}\;\;\;\varphi_{r}(H) \subset A^{\circ}\,, \end{equation}
and this follows from the characterization of $H^{\circ}$ and $A^{\circ}$
recalled above. The rest can be checked by straightforward calculations. \\
Suppose now that $H$ is even a Hopf algebra, and let $S_H$ be the antipode
of $H$. Then it is easy to see that the transpose $S_{H}^{t}$ of $S_H$ maps
$H^{\circ}$ into itself and that the map of $H^{\circ}$ into itself induced
by $S_{H}^{t}$ is an antipode for $H^{\circ}$. Thus if $H$ is a Hopf algebra,
then $H^{\circ}$ is likewise.
We now can prove the following corollary to Proposition 3.1. \\
\noindent
{\bf Corollary 3.1} \\
{\it We use the notation introduced at the beginning of this section, but
suppose in addition that $A$ and $H$ are Hopf algebras, with antipodes $S_A$
and $S_H$, respectively. If $\varphi$ is a Hopf pairing of $A$ and $H$, then
\be \varphi(S_{A}(a),h) = \varphi(a,S_{H}(h)) \end{equation}
for all $a \in A$ and $h \in H$.}
Proof \\
It is well--known [2] that a bialgebra homomorphism of a Hopf algebra into
another one is, in fact, a Hopf algebra homomorphism, i.e., it intertwines
the antipodes. Thus the corollary follows from Proposition 3.1. \\
Our next task is to dualize the notion of a universal $R$--matrix [1]. Let
$H$ be a bialgebra. An element $R \in H \otimes H$ is called a universal
$R$--matrix for $H$ if it has the following four properties:
\be R\;\;\mbox{is invertible in}\;\;H \otimes H \label{URM1} \end{equation}
\vspace{-5.5ex}
\be (T\Delta_{H}(h))R = R\,\Delta_{H}(h)\;\;\mbox{for all}\;\;h \in H \label{URM2}\end{equation}
\vspace{-7ex}
\begin{eqnarray} (\Delta_{H} \otimes id_{H})(R) \!\! & = & \!\! R_{13}R_{23} \label{URM3} \\
(id_{H} \otimes \Delta_{H})(R) \!\! & = & \!\! R_{13}R_{12}\,. \label{URM4} \end{eqnarray}
Our notation is standard:
\be T:H \otimes H \longrightarrow H \otimes H \end{equation}
is the linear twist mapping given by
\be T(h \otimes h') = h' \otimes h\;\;\mbox{for all}\;\;h,h' \in H \end{equation}
and the elements
\be R_{12}\,,\:R_{23}\,,\:R_{13} \in H \otimes H \otimes H \end{equation}
are defined by
\be R_{12} = R \otimes 1_H \end{equation}
\be R_{23} = 1_H \otimes R \end{equation}
\be R_{13} = (T \otimes id_H)(R_{23}) = (id_H \otimes T)(R_{12})\,, \end{equation}
where $1_H$ denotes the unit element of $H$.
In order to dualize these four properties, we assume that $H$ is the Hopf
dual of a bialgebra $A$\,,
\be H = A^{\circ}\,. \end{equation}
Since $A$ and hence $A \otimes A$ are coalgebras, their duals $A^{\ast}$ and
$(A \otimes A)^{\ast}$ are algebras, and $A^{\circ}$ is a subalgebra of
$A^{\ast}$. Actually, we have the following chain of canonical injective
algebra homomorphisms:
\be A^{\circ} \otimes A^{\circ} \longrightarrow A^{\ast} \otimes A^{\ast}
\longrightarrow (A \otimes A)^{\ast}\,. \label{inj} \end{equation}
Thus, since $R$ is an element of $A^{\circ} \otimes A^{\circ}$, it can also be
considered as an element of $(A \otimes A)^{\ast}$. On the other hand, due to the
familiar properties of tensor products of vector spaces, $(A \otimes A)^{\ast}$
can be canonically identified with the space $B(A,A)$ of all bilinear forms
on $A \times A$\,. In particular, $B(A,A)$ inherits from $(A \otimes A)^{\ast}$
the structure of an associative algebra with a unit element. The
multiplication is the so--called convolution. It is denoted by an asterisk
$\ast$ and is defined as follows. Let $\psi,\,\psi'$ be two elements of
$B(A,A)$ and let $a,b \in A$\,. If $\Delta_A$ is the coproduct of $A$ and if
\be \Delta_{A}(a) = \sum_{i}a^1_i \otimes a^2_i \label{dela} \end{equation}
\be \Delta_{A}(b) = \sum_{j}b^1_j \otimes b^2_j \label{delb} \end{equation}
with $a^1_i,a^2_i,b^1_j,b^2_j \in A$\,, we have
\be(\psi \ast \psi')(a,b) = \sum_{i,j}\psi(a^1_i,b^1_j)
\,\psi'(a^2_i,b^2_j) \,. \label{conv} \end{equation}
The unit element is the bilinear form $\varepsilon_B$ on $A \times A$ defined by
\be \varepsilon_B(a,b) = \varepsilon_A(a)\varepsilon_A(b) = \varepsilon_A(ab) \end{equation}
for all $a,b \in A$\,, where $\varepsilon_A$ denotes the counit of $A$\,.
Now if
\be R = \sum_{i}R^{1}_{i} \otimes R^{2}_{i} \end{equation}
is an arbitrary element of $A^{\circ} \otimes A^{\circ}$, the corresponding
bilinear form $\varrho \in B(A,A)$ is given by
\be \varrho(a,b) = \sum_{i}\langle R^{1}_{i},a\rangle\langle R^{2}_{i},b\rangle \end{equation}
for all $a,b \in A$\,, where $\langle\,\;,\;\rangle$ denotes the canonical pairing of
$A^{\ast}$ and $A$\,.
If $R$ is invertible in $A^{\circ} \otimes A^{\circ}$, then all the more in
$B(A,A)$. This means that there exists a bilinear form $\varrho' \in B(A,A)$
such that
\be \varrho \ast \varrho' = \varrho' \ast \varrho = \varepsilon_B \,. \label{inv} \end{equation}
Note that if $A$ is finite--dimensional, then the homomorphisms in Eq.
\reff{inj} are bijective, and hence $R$ is invertible in
$A^{\circ} \otimes A^{\circ}$ if and only if a bilinear form $\varrho'$ of the type
described above exists. In general, however, we cannot expect that the
invertibility of $\varrho$ in $B(A,A)$ implies the invertibility of $R$ in
$A^{\circ} \otimes A^{\circ}$.
The remaining conditions \reff{URM2} -- \reff{URM4} can also be reformulated
in terms of certain convolution products involving $\varrho$\,. Because of lack
of space, we prefer to present the results more explicitly, as follows.
Condition \reff{URM2} is satisfied if (and, if $A^{\circ}$ separates the
elements of $A$\,, only if) the following equation holds for all
$a,b \in A$\,:
\be \sum_{i,j}b^{1}_{j} a^{1}_{i} \varrho(a^{2}_{i},b^{2}_{j}) =
\sum_{i,j}\varrho(a^{1}_{i},b^{1}_{j}) a^{2}_{i} b^{2}_{j} \label{comm} \end{equation}
(here and in the subsequent equations we are using the notation introduced
in Eqs. \reff{dela}, \reff{delb}). Moreover, Eq. \reff{URM3} is satisfied
if and only if
\be \varrho(bc,a) = \sum_{i}\varrho(b,a^1_i)\varrho(c,a^2_i) \label{rho3} \end{equation}
for all $a,b,c \in A$\,, and Eq. \reff{URM4} holds if and only if
\be \varrho(a,bc) = \sum_{i}\varrho(a^1_i,c)\varrho(a^2_i,b) \label{rho4} \end{equation}
for all $a,b,c \in A$\,.
The foregoing results motivate the following definition. \\
\noindent
{\bf Definition 3.1} \\
{\it Let $A$ be a bialgebra. A bilinear form $\varrho$ on $A \times A$ is said
to define a coquasitriangular (or dual quasitriangular, or braiding)
structure on $A$ if it is convolution invertible (i.e., if it has an inverse
in the sense of Eq. \reff{inv}) and if the relations \reff{comm}, \reff{rho3},
and \reff{rho4} are satisfied. A bialgebra endowed with a coquasitriangular
structure will be called coquasitriangular.} \\
Apparently, dual quasitriangular structures have first been introduced in
Ref. [13] and shortly afterwards by various authors. In particular, I refer
the reader to Ref. [14] (note that the coquasitriangular structures discussed
here are the right braiding structures defined there).
Visibly, Eq. \reff{comm} is some sort of generalized commutation relation.
In fact, the famous RTT relations of Ref. [15] are of this type, and the
corresponding bi\-algebras $A(R)$ (with $R$ an invertible solution of the
Yang--Baxter equation) are coquasitriangular.
The Eqs. \reff{rho3} and \reff{rho4} can be rewritten as follows. Let
$\varrho^{\otimes 2}$ be the bilinear form on $A \otimes A$ defined by
\be \varrho^{\otimes 2}(a \otimes a',b \otimes b') = \varrho(a,b)\varrho(a',b') \end{equation}
for all $a,a',b,b' \in A$ (see Eq. \reff{quad}). Then Eq. \reff{rho3} is
equivalent to
\be \varrho^{\otimes 2}(b \otimes c, \Delta_{A}(a)) = \varrho(bc,a)\,, \label{rho3mod} \end{equation}
and Eq. \reff{rho4} is equivalent to
\be \varrho^{\otimes 2}(T\Delta_{A}(a), b \otimes c) = \varrho(a, bc)\,, \label {rho4mod} \end{equation}
(where, once again, $T$ denotes the twist mapping) or, what amounts to the
same, to
\be \varrho^{\otimes 2}(\Delta_{A}(a), c \otimes b) = \varrho(a,bc)\,. \label{rho4modd} \end{equation}
As is well--known, there is a real host of properties and relations satisfied
by a universal $R$--matrix. Thus one should expect the same for
coquasitriangular structures. Of course, due to the well--known problems
with duality for infinite--dimensional vector spaces, we cannot simply prove
these latter results by duality. Rather, we use duality to find out which
properties should be true, and then try to prove these directly.
Let us give some examples. In the following, $R$ denotes a universal
$R$--matrix for a bialgebra $H$ and $\varrho$ denotes a coquasitriangular
structure for a bialgebra $A$\,. It is obvious that with $R$ also
$T(R^{-1}) = T(R)^{-1}$ is a universal $R$--matrix for $H$ (as before, $T$
denotes the appropriate twist mapping, here the one of $H \otimes H$ onto
itself). Dually, with $\varrho$ also $\varrho' \!\stackrel{\textstyle{\hst{0ex} \atop \circ}}{\hst{0ex}}\! T = (\varrho \!\stackrel{\textstyle{\hst{0ex} \atop \circ}}{\hst{0ex}}\! T)'$ is a
coquasitriangular structure for $A$\,, where the prime denotes the convolution
inverse. A coquasitriangular structure $\varrho$ is said to be cotriangular if
$\varrho' \!\stackrel{\textstyle{\hst{0ex} \atop \circ}}{\hst{0ex}}\! T = \varrho$\,.
It is known that
\be (\varepsilon_H \otimes id_H)(R) = 1_H \end{equation}
\be (id_H \otimes \varepsilon_H)(R) = 1_H \end{equation}
(in these equations, $\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}} \otimes H$ and $H \otimes \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$ have been canonically
identified with $H$). Dually, it can be shown that
\be \varrho(1_A,a) = \varrho(a,1_A) = \varepsilon_A(a) \label{rhounit} \end{equation}
for all $a \in A$ (for example, apply Eq. \reff{rho4} with $c = 1_A$ and use
that $\varrho$ has a convolution inverse).
Finally, let us assume that $H$ and $A$ are Hopf algebras. Then it is known
that
\begin{eqnarray} (S_H \otimes id_H)(R) \!\! & = & \!\! R^{-1} \\
(id_H \otimes S_H)(R^{-1}) \!\! & = & \!\! R\,. \end{eqnarray}
Dually, it can be shown that
\begin{eqnarray} \varrho(S_A(a),b) \!\! & = & \!\! \varrho'(a,b) \label{rhoS1} \\
\varrho'(a,S_A(b)) \!\! & = & \!\! \varrho(a,b) \end{eqnarray}
for all $a,b \in A$\,, where, as before, $\varrho'$ is the convolution inverse
of $\varrho$\,. In particular, we have
\be \varrho(S_A(a),S_A(b)) = \varrho(a,b) \end{equation}
for all $a,b \in A$\,.
Conversely, if $\varrho$ is a bilinear form on a Hopf algebra $A$ satisfying
the Eqs. \reff{comm} -- \reff{rho4} and \reff{rhounit}, then the bilinear
form $\varrho'$ defined by Eq. \reff{rhoS1} is a convolution inverse of $\varrho$
and hence $\varrho$ defines a coquasitriangular structure on $A$\,.
Let us next make contact with our discussion of Hopf pairings. Let $A$ be a
bi\-algebra and let $A^{cop}$ (resp. $A^{aop}$) be the bialgebra which,
considered as an algebra (resp. a coalgebra) coincides with $A$\,, but which,
considered as a coalgebra (resp. an algebra) has the structure opposite to
that of $A$\,. Then the Eqs. \reff{rho3mod} -- \reff{rho4modd} and
\reff{rhounit} show that a coquasitriangular structure $\varrho$ of $A$ can be
defined to be a Hopf pairing $A^{cop} \times A \rightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$ (or, equivalently, a
Hopf pairing $A \times A^{aop} \rightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$) which has a convolution inverse in
the sense of Eqs. \reff{conv}, \reff{inv} and satisfies the generalized
commutation relations \reff{comm}. In particular, it follows from
Proposition 3.1 that, for any $b \in A$\,, the linear forms $a \rightarrow \varrho(a,b)$
and $a \rightarrow \varrho(b,a)$ on $A$ belong to $A^{\circ}$.
This result is highly welcome. In our dualization process, we started from
a universal $R$--matrix $R \in A^{\circ} \otimes A^{\circ}$ and embedded
$A^{\circ} \otimes A^{\circ}$ into $(A \otimes A)^{\ast} \simeq B(A,A)$. For
infinite--dimensional bialgebras, $(A \otimes A)^{\ast}$ is terribly much larger
than $A^{\circ} \otimes A^{\circ}$. On the other hand, a certain enlargement of
$A^{\circ} \otimes A^{\circ}$ is reasonable, since for the most interesting
examples of infinite--dimensional quasitriangular Hopf algebras $H$, the
universal $R$--matrix does not belong to $H \otimes H$ but to a certain
completion of it. Our last result shows that $\varrho$ seems to be not as far
from $A^{\circ} \otimes A^{\circ}$ as one might have feared. Thus one could hope
to find a subbialgebra $U$ of $A^{\circ}$ associated to $\varrho$ and a suitable
completion of $U \otimes U$ in $B(A,A)$ containing $\varrho$\,. Such a programme has
been carried out in Ref. [14].
Finally, let us recall that, for any coalgebra $C$\,, the right $C$--comodules
may also be regarded as left $C^{\ast}$--modules (where $C^{\ast}$ denotes
the associative algebra dual to $C$)[2]. In view of the discussion of the
present section we expect that a coquasitriangular structure on a bialgebra
$A$ can be used to convert the class of right $A$--comodules into a braided
monoidal category. This can in fact be done (actually, this is the starting
point of Ref. [14]). We do not want to go into detail here but only show how
the corresponding braiding isomorphisms are defined.
Thus let $V$ and $W$ be two right $A$--comodules, with structure maps
\begin{eqnarray} \delta_{V} \!\!\!& : &\!\!\! V \longrightarrow V \otimes A \\
\delta_{W} \!\!\!& : &\!\!\! W \longrightarrow W \otimes A \,. \end{eqnarray}
To define the braiding isomorphism
\be \psi_{V,W} : V \otimes W \longrightarrow W \otimes V \:, \end{equation}
choose $x \in V$, $y \in W$, and write
\begin{eqnarray} \delta_{V}(x) \!\! & = & \!\! \sum_{i}x_i \otimes a_i \\
\delta_{W}(y) \!\! & = & \!\! \sum_{j}y_j \otimes b_j \end{eqnarray}
with $x_i \in V$, $y_j \in W$\,; $a_{i},b_{j} \in A$\,. Then we set
\be \psi_{V,W}(x \otimes y) = \sum_{i,j}\varrho(a_{i},b_{j})\, y_j \otimes x_i \;.
\label{cobraid} \end{equation}
We close this section by two remarks. For reasons of uniformity, we have
formulated the results of this section over the base field $\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$\,. Of
course, everything holds for arbitrary base fields of any characteristic.
Moreover, by systematically using the techniques supplied by Ref. [3], these
results can easily be generalized to the graded case, with arbitrary Abelian
groups of degrees and arbitrary commutation factors.
\section{Coquasitriangular structures for the group
\newline
Hopf algebra of an Abelian group}
\setcounter{equation}{0}
We now are ready to rederive the results of Section 2 from the dual point of
view. Moreover, we are going to see that the dual picture is applicable also
to infinite Abelian groups.
To begin with we still assume that the group $\Gamma$ is finite. Then the group
Hopf algebras $\Co\Ga$ and $\Co\Gas$ are (Hopf) dual to each other. In fact, since
$\Gamma$ is a basis of $\Co\Ga$ and $\Gamma^{\ast}$ a basis of $\Co\Gas$, there is a unique
bilinear form $\Co\Gas \times \Co\Ga \rightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$ whose restriction to
$\Gamma^{\ast} \times \Gamma$ is equal to the group pairing $\Gamma^{\ast} \times \Gamma \rightarrow \mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}}$\,,
and it is easy to check that this bilinear form is a non--degenerate Hopf
pairing. Thus there is no problem in denoting these two pairings by the same
symbol $\langle\,\;,\;\rangle$.
The result above is well--known, it agrees with the fact that the dual of
the group Hopf algebra $\mbox{C\hst{-0.44em}\vl{0.1ex}{1.4ex}\hst{0.44em}} G$ of a finite group $G$ is canonically isomorphic
to the Hopf algebra ${\cal F}(G)$ of complex--valued functions on $G$\,. In
fact, the characters of $\Gamma$ form a basis of ${\cal F}(\Gamma)$. Of course, all
this is closely related to Fourier transformation on $\Gamma$ and $\Gamma^{\ast}$.
According to Section 3 the foregoing result implies that the universal
$R$--matrices for $\Co\Gas$ are in bijective correspondence with the
coquasitriangular structures on $\Co\Ga$. More precisely, consider the chain
\reff{inj} of injective algebra homomorphisms, with $A = \Co\Ga$. In the present
case these are, in fact, algebra isomorphisms, and according to the foregoing
result we can (and will) identify $(\Co\Ga)^{\ast}$ with $\Co\Gas$. On the other
hand $(\Co\Ga \otimes \Co\Ga)^{\ast}$ is canonically isomorphic to $B(\Co\Ga,\Co\Ga)$, the
algebra of bilinear forms on $\Co\Ga \times \Co\Ga$. Since $\Gamma$ is a basis of
$\Co\Ga$, the bilinear forms on $\Co\Ga \times \Co\Ga$ are uniquely determined by
their values on $\Gamma \times \Gamma$, which shows that the vector space
$B(\Co\Ga,\Co\Ga)$ can be identified with ${\cal F}(\Gamma \times \Gamma)$, the vector
space of complex--valued functions on $\Gamma \times \Gamma$. Summarizing, we have
the following chain of vector space isomorphisms:
\be \Co\Gas \otimes \Co\Gas \longrightarrow (\Co\Ga \otimes \Co\Ga)^{\ast} \longrightarrow B(\Co\Ga,\Co\Ga) \longrightarrow
{\cal F}(\Gamma \times \Gamma)\,. \end{equation}
We know that the first of these maps is an algebra homomorphism and (by
definition of the convolution) so is the second. A look at Eq. \reff{conv}
shows that the third map is an algebra homomorphism as well. Furthermore, it
is easy to check that the inverse mapping
\be {\cal F}(\Gamma \times \Gamma) \longrightarrow \Co\Gas \otimes \Co\Gas \end{equation}
is just the mapping $\sigma \rightarrow R_{\sigma}$ mentioned in part 1) of
Proposition 2.1.
Thus all we have to do is to determine those bilinear forms $\varrho$ on
$\Co\Ga \times \Co\Ga$ which define a coquasitriangular structure on $\Co\Ga$. Since
$\Gamma$ is Abelian, the condition \reff{comm} is trivially satisfied.
Obviously, the remaining conditions (convolution invertibility and Eqs.
\reff{rho3}, \reff{rho4}) are fulfilled if and only if the restriction of
$\varrho$ to $\Gamma \times \Gamma$ (i.e., the element of ${\cal F}(\Gamma \times \Gamma)$
corresponding to $\varrho$) is a bicharacter on $\Gamma$. Thus we have rederived
part 1) of Proposition 2.1 and the non--trivial parts of Corollary 2.1.
{\it Let us now drop the assumption that $\Gamma$ is finite.} Then the
coquasitriangular structures on $\Co\Ga$ can be characterized as before: A
bilinear form $\varrho$ on $\Co\Ga \times \Co\Ga$ defines a coquasitriangular structure
on $\Co\Ga$ if and only if its restriction $\sigma$ to $\Gamma \times \Gamma$ is a
bicharacter of $\Gamma$, and $\varrho$ is cotriangular if and only if $\sigma$ is a
commutation factor on $\Gamma$.
Of course, we should now answer the question of how graded multilinear
algebra (which has been the point of departure for our round trip) can be
formulated in the present setting. The answer is suggested by the remarks
at the end of Section 3: The $\Co\Ga$--comodules should take the role of the
$\Co\Gas$--modules.
In fact, it is well--known that $\Gamma$--graded vector spaces and
$\Co\Ga$--comodules are essentially the same thing. Indeed, let $V$ be a right
$\Co\Ga$--comodule and let
\be \delta : V \longrightarrow V \otimes \Co\Ga \end{equation}
be its structure mapping. If we define, for all $\gamma \in \Gamma$,
\be V_{\gamma} = \{x \in V \,|\, \delta(x) = x \otimes \gamma \}\,, \end{equation}
then, obviously, $V_{\gamma}$ is a subspace of $V$, and it is easy to see that
these subspaces form a $\Gamma$--gradation of $V$\,:
\be V = \bigoplus_{\gamma \in \Gamma} V_{\gamma}\,. \end{equation}
Conversely, if $(V_{\gamma})_{\gamma \in \Gamma}$ is a $\Gamma$--gradation of a vector
space $V$, there exists a unique linear mapping
\be \delta : V \longrightarrow V \otimes \Co\Ga \end{equation}
such that
\be \delta(x) = x \otimes \xi \end{equation}
for all $x \in V_{\xi}$\,, $\xi \in \Gamma$, and this mapping defines on $V$
the structure of a right $\Co\Ga$--comodule. Obviously, these transitions from
right $\Co\Ga$--comodules to $\Gamma$--graded vector spaces and vice versa are
inverse to each other. Moreover, the homomorphisms of $\Co\Ga$--comodules are
exactly the homomorphisms of $\Gamma$--graded vector spaces.
Now let $\varrho$ define a coquasitriangular structure on $\Co\Ga$ and let $\sigma$
be the restriction of $\varrho$ to $\Gamma \times \Gamma$. Consider two $\Gamma$--graded
vector spaces, i.e., right $\Co\Ga$--comodules, denoted by $V$ and $W$, say.
Then the braiding isomorphism
\be \psi_{V,W} : V \otimes W \longrightarrow W \otimes V \end{equation}
corresponding to $\varrho$ (see Eq. \reff{cobraid}) is given by
\be \psi_{V,W}(x \otimes y) = \sigma(\xi,\eta)\,y \otimes x \end{equation}
for all $x \in V_{\xi}$\,, $y \in W_{\eta}$\,; $\xi,\eta \in \Gamma$. This is
exactly the formula familiar from graded multilinear algebra [3].
Needless to say, there is a lot of other details which should be checked
before we can be sure that the graded multilinear algebra as developed in
Ref. [3] and the comodule picture described here (with $\varrho$ cotriangular)
are really equivalent. However, I do not want to embark on this boring
exercise. Instead, I would like to close this section by the remark that we
have come across one more example of a well--known fact: In the
infinite--dimensional case, coalgebras and comodules are easier to handle
than algebras and modules.
\section{Discussion}
\setcounter{equation}{0}
In the present work we have described how the graded multilinear algebra as
developed in Ref. [3] can be understood in the language of triangular or
cotriangular Hopf algebras. We may now proceed to apply the theory of Hopf
algebras to the study of algebras living in the graded multilinear category.
In particular, we could apply Majid's theory of bosonization [16] and try
to reduce (or, at least, to relate) the theory of Lie colour algebras to the
theory of usual Lie algebras. (Roughly speaking, bosonization is a procedure
which converts a Hopf algebra living in a category with non--trivial braiding
into a usual Hopf algebra.) The idea to try this is not new. It is used in
supersymmetry by introducing Grassmann variables, and it is also applied in
Ref. [5], where the present author has reduced the theory of Lie colour
algebras to the theory of graded Lie (super)algebras.
More recent and closer to the present work and to the bosonization procedure
are the Refs. [17,\,9]. In Ref. [17] the authors prove, in the context of
coquasitriangular Hopf algebras, a general bosonization theorem and use it
to derive Schur's double centralizer theorem for cotriangular Hopf algebras
(hence, in particular, for Lie colour algebras). In Ref. [9] the author
constructs a bosonized version of the enveloping algebra of a Lie colour
algebra and determines its universal $R$--matrix.
Of course, there is no need to restrict attention to the triangular case,
i.e., to commutation factors. In fact, allowing for arbitrary bicharacters,
the graded vector spaces are the objects of a really braided monoidal
category, i.e., one in which the braid groups really come into play. On the
other hand, these categories are still very close to the classical case,
hence they might be used to become acquainted with the braided situation.
For example, one could try to define generalized Lie algebras living in these
categories. This has in fact been done in a recent work by Pareigis [18].
Using the tensor operator techniques developed in Refs. [19,\,20] one might
then hope to construct interesting quantum spin chain Hamiltonians which
are invariant under these algebras.\\[2ex]
\noindent
{\bf Acknowledgements}\,\,
A major part of the present work has been carried out during two visits of
the author to the Erwin Schr\"{o}dinger Institut in Vienna. The kind
invitations by Harald Grosse and the hospitality extended to the author in
the ESI are gratefully acknowledged. \\
\newpage
\noindent
{\large\bf References}
\begin{enumerate}
\item V. Drinfeld, J. Sov. Math. {\bf 41}, 898 (1988) (expanded version of a
report to the International Congress of Mathematicians, Berkeley 1986).
\item M.E. Sweedler, {\it Hopf algebras}, W.A. Benjamin, New York (1969).
\item M. Scheunert, J. Math. Phys. {\bf 24}, 2658 (1983).
\item V. Rittenberg and D. Wyler, Nucl. Phys. B {\bf 139}, 189 (1978).
\item M. Scheunert, J. Math. Phys. {\bf 20}, 712 (1979).
\item S. Mac Lane, {\it Categories for the working mathematician}, Springer,
New York, Heidelberg, Berlin (1971).
\item A. Joyal and R. Street, Braided monoidal categories, Mathematics
Reports 860081, Macquarie University (1986), and Adv. in Math.
{\bf 102}, 20 (1993).
\item S. Majid, Int. J. Mod. Phys. A {\bf 5}, 1 (1990).
\item D.S. McAnally, Lett. Math. Phys. {\bf 33}, 249 (1995).
\item S. Majid, in {\it Spinors, twistors, Clifford algebras, and quantum
deformations,} Proceedings of the 2\,nd Max Born Symposium, Wroclaw
1992, edited by Z. Oziewicz et al., Kluwer, pp. 327--336.
\item D. Radford, J. Algebra {\bf 141}, 354 (1991).
\item M. Takeuchi, {\it The \#--product of group sheaf extensions applied
to Long's theory of dimodule algebras,} Algebra--Berichte,
Mathematisches Institut der Universit\"{a}t M\"{u}nchen, Nr. 34,
Verlag Uni--Druck, M\"{u}nchen (1977).
\item S. Majid, in {\it Quantum probability and related topics VI,} Trento
1989, proceedings, edited by L. Accardi et al., World Scientific,
Singapore (1991), pp. 333--358.
\item R.G. Larson and J. Towber, Comm. Algebra {\bf 19}, 3295 (1991).
\item N.Yu. Reshetikhin, L.A. Takhtadzhyan, and L.D. Faddeev, Leningrad
Math. J. {\bf 1}, 193 (1990).
\item S. Majid, J. Algebra, {\bf 163}, 165 (1994).
\item D. Fischman and S. Montgomery, J. Algebra {\bf 168}, 594 (1994).
\item B. Pareigis, On Lie algebras in braided categories, Mathematisches
Institut der Universit\"{a}t M\"{u}nchen, preprint, April 1995.
\item M. Scheunert, in {\it Generalized symmetries in physics,} Clausthal
1993, proceedings, edited by H.--D. Doebner, V.K. Dobrev, and A.G.
Ushveridze, World Scientific (1994), pp. 77--89.
\item M. Scheunert, Int. J. Mod. Phys. B {\bf 8}, 3655 (1994).
\end{enumerate}
\end{document}
|
\section{Introduction}
Few-body problems of interacting particles have vital importance in all
branches of physics from hadron to celestial levels. The main interest
in the few-body problems lies in, e.g., finding an accurate solution for the
system,
testing the equation of motion and the conservation laws and
symmetries, or looking for unknown interactions governing the system.
\par\indent
The investigation of few-nucleon systems interacting via realistic
forces has always been in the center of the interest. Considerable effort
has been exerted to obtain
accurate ground-state properties of the few-nucleon systems with
Faddeev-Yakubovsky (FY) \cite{payne,faddeev,glockb}, variational
\cite{atms,hypers,crcbgv}, variational Monte Carlo (VMC)
\cite{mcv,vmc} and Green's function Monte
Carlo (GFMC) methods \cite{gfma,gfmb}. Most of these approaches has focused
on three- or four-body problems.
\par\indent
To treat an $N$-particle system, one needs to cope with a large number of
variables
required to specify the wave function. By using $(N\!\!-\!\!1)$ relative
coordinates to
describe the system, for example, the discretization on a mesh with
$p$ points, or the expansion of the function of the relative motion between
the particles in terms
of $p$ suitably chosen functions leads to $p^{(N-1)}$ mesh points or basis
functions, which becomes prohibitively large with increasing $N$.
All but the Monte Carlo methods face this difficulty
as the number of particles increases. The VMC and GFMC
methods have proved to be most successful by being able to go beyond
the four-nucleon problem \cite{prl95}. The secret of the efficiency of the
Monte Carlo methods is the use of an importance sampling
of the most relevant parts of the configuration space.
This fact naturally raises a question: Even if the wave function of the
$N$-particle system is expanded into an (excessively) large number of basis
functions, can't one reduce the problem to a tractable one by selecting
``the most important'' basis functions ?
\par\indent
The aim of this paper is to present an alternative variational approach, the
stochastic variational method (SVM) \cite{SVM,VSL}, by using the correlated
Gaussians as basis functions \cite{corrgauss,temper}. Examples whose
solutions were known before are used to demonstrate the
performance of the method in treating nuclear as well as Coulomb
interactions.
To highlight some new physics, we have also included problems that have
been hitherto unsolved. We give the formulation and some details of the
method of the calculation and show applications to $N=2-7$-particle systems.
\par\indent
The variational foundation for the time-independent Schr\"odinger equation
provides a solid and arbitrarily improvable framework for the solution of
bound-state problems. The crucial point of the variational approach is the
choice of the trial function. There are two widely applied strategies:
(1) to select the most appropriate functional form to describe the
short-range as well as long-range correlations and to compute the matrix
elements by Monte Carlo technique, or (2) to use a number, possibly a great
number, of simple terms, which facilitate the analytical calculation of the
matrix elements. We
follow the second course by using an expansion over a correlated Gaussian
``basis".
\par\indent
To solve the $N$-particle problem, it is of prime importance to describe the
correlation between the
particles properly. The correlation is conveniently represented by a
correlation factor, $F=\prod_{i<j}^{N}f_{ij}$
\cite{atms,hypers,mcv,vmc,etbm}.
Most calculations have used this form of $F$ directly to evaluate
the matrix elements. Such calculations are, however, fairly
involved beyond the three-particle system and performed by
Monte Carlo integrations. An alternative way to incorporate the
correlation is to approximate $f_{ij}$ as a linear combination of Gaussians
${\rm exp}(-\alpha_{ij} ({\bf r}_i-{\bf r}_j)^2)$.
The $N$-particle basis function then contains product of these
Gaussians:
$\prod_{i<j}^{N}{\rm exp}(-\alpha_{ij}({\bf r}_i-{\bf r}_j)^2)=
{\rm exp}(-\sum_{i<j}^{N}\alpha_{ij}({\bf r}_i-{\bf r}_j)^2)$.
These Gaussian functions are widely used in variational
calculations (see,
for example, \cite{temper,etbm}). We will apply a more general form of the
correlated
Gaussian functions which allow for nonzero orbital angular momentum, and
will use the more convenient Jacobi relative coordinates instead of
the relative distance vectors.
The correlated Gaussians have an important advantage. Their
Hamiltonian matrix elements can be analytically calculated
in a unified framework, thus enabling one to avoid the formidable
calculation involving the correlation factor $F$.
\par\indent
The variational approximation, however, may run into difficulties for the
following reasons:
(i) if the nonlinear parameters specifying the basis functions are varied,
it is difficult to optimize them,
(ii) if they are not, then the number of terms required may be excessively
large, and, in both cases,
(iii) the trial function of proper symmetry becomes extremely involved. For
example, conventional methods \cite{crcbgv,temper}
for the choice of the Gaussian parameters lead to prohibitively large bases for
more than 3 or 4 particles, which has limited the applicability of the
Gaussian basis to few-body problems.
\par\indent
One can circumvent the optimization problem including large number of
nonlinear parameters or the diagonalization of huge matrices by using the
SVM. The SVM attempts to set up the most appropriate basis functions by the
following stepwise procedure: One generates a would-be
basis function by choosing the nonlinear parameters randomly, judges its
utility by the energy gained by including it in the basis, and
either keeps or discards it. One
repeats this ``trial and error'' procedure until the basis set up leads to
convergence. The original procedure of the SVM, proposed in \cite{SVM},
has recently been developed further and successfully applied to multicluster
descriptions of light exotic nuclei, such as $^6$He=$\alpha+ n + n$,
$^8$He=$\alpha+ n + n + n + n$, $^9$Li=$\alpha + t + n + n$,
and $^9$C=$\alpha+^3$He$+p + p$ \cite{VSL,SVAO}. Learning from these
applications, we have now generalized and refined the method further to
encompass diverse systems emerging in nuclear and atomic physics.
\par\indent
Besides the large number of nonlinear parameters, the treatment
of the increasing number of partial waves in the expansion of the wave
function would also pose a formidable task. We propose here an alternative
formulation to cope with this problem.
Instead of using the partial wave expansion, the angular dependence of the
wave function is represented by a single solid spherical harmonics
whose argument contains additional variational parameters. This form
makes the calculation of the matrix elements for nonzero orbital angular
momentum much simpler than other methods.
\par\indent
It will be demonstrated that the present method has several unique
features: It is based on a fully analytical calculation for most types of
interactions and thus ensures high accuracy and speed. Its calculational
scheme is quite universal and needs no change depending on whether the
system contains nuclear or Coulombic or other
interactions. It has no difficulty in
treating the system of particles of unequal masses. More importantly, the
wave function is obtained in a compact, analytical form and thereby can be
readily used in calculations of physical properties.
\par\indent
As you will see later, the present method has turned out to be
very accurate, and we think it is worth while to make the method and the
results
easily available and reproducible for interested readers. We collect all
the needed ingredients of our method in order.
Some of the formulae are our original developments or generalizations
of known relations to $N$-particle matrix elements, and some others are
collected here to make the paper self-contained. The calculation of the
matrix elements presented here is different from the one of Refs.
\cite{KA91,epep} in many aspects: The motion of the centre-of-mass is
removed from both the Hamiltonian and the wave function. Two-particle
potential matrix elements
of arbitrary radial form factor are evaluated in a unified way by reducing
them to the calculation of appropriate correlation functions corresponding
to the interaction.
The calculation of the matrix elements is extended to nonzero
orbital angular momentum as well. The symmetrization postulate is imposed
on the wave function at the single-particle level, which provides several
advantages, especially in evaluating the matrix elements of state-dependent
realistic nuclear interactions.
\par\indent
The organization of the paper is as follows. Section II defines the
correlated Gaussian basis functions and gives the details of the stochastic
procedure of selecting the basis set. Section III contains the method of
calculating the matrix elements. The main steps are the calculation of
matrix elements in Slater determinants (or permanents for bosons)
consisting of single-particle Gaussian wave packets, the elimination of the
center-of-mass motion with a very simple manipulation, and the
transformation to the correlated Gaussian basis. This section also presents
the modifications needed for treating systems of particles of unequal
masses. Section IV presents numerical results for various systems of
particles which interact via nuclear potentials or power-law potentials.
Section V gives a brief summary. In the appendices the most important
auxiliary formulae are collected to facilitate any future use of the
formulation.
\section{The correlated Gaussians and the stochastic variational method}
\subsection{Basis functions}
Since the variational method is always limited by the form chosen as a
trial function, the trial
function must be flexible enough to be able to describe
the full variety of correlations between the nucleons, e.g., the short-range
correlation due to the strong repulsive force, the $\alpha$-clustering
typical in some light nuclei, or the long-range
correlation at large distances in light halo nuclei. The correlation
between the nucleons can
be described by functions of appropriate relative coordinates.
\par\indent
Any square-integrable function with angular momentum
$lm$ can be approximated, to any desired accuracy, by a linear combination
of nodeless harmonic-oscillator functions (Gaussians) of continuous size
parameter $a$:
\begin{equation}
\Gamma_{lm}({\bf r}) \sim {\rm e}^{-{\frac{1}{2}} a r^2}
{\cal Y}_{lm}({\bf r}),\ \ \ {\rm with}\ \ \ {\cal Y}_{lm}({\bf r})=r^l
Y_{lm}(\hat{\bf r}).
\end{equation}
A generalization
of this to $N$-nucleon systems contains a product of the Gaussians as
mentioned in the previous section. It is convenient to use a set of the
Jacobi coordinates ${\bf x}=({\bf x}_1,...,{\bf x}_{N-1})$, instead of
$N(N\!\!-\!\!1)/2$ relative distance vectors $({\bf r}_i-{\bf r}_j)$.
An $N$-nucleon basis function, a so-called correlated Gaussian,
then looks like
\begin{equation}
\psi_{(LS)JMTM_T}({\bf x}, A)={\cal A} \bigl\{
{\rm e}^{-{1 \over 2}\tilde{{\bf x}}A {\bf x}} \left[ \theta_{L}({\bf x})
\chi_{S}\right]_{JM} {\cal X}_{TM_T}\bigr\},
\label{corrGauss}
\end{equation}
where $\tilde{{\bf x}}$, the transpose of ${\bf x}$, stands for the row
vector comprising the Jacobi coordinates. $\chi$ and ${\cal X}$ are the
spin and isospin functions.
$A$ is an $(N\!\!-\!\!1)\!\!\times\!\!(N\!\!-\!\!1)$ positive-definite,
symmetric matrix of nonlinear parameters, specific to each basis element,
and the quadratic form, $\tilde{\bf x}A{\bf x}$, involves scalar products
of the Cartesian vectors:
\begin{equation}
\tilde{{\bf x}} A {\bf x}=\sum_{i=1}^{N-1} \sum_{j=1}^{N-1}
A_{ij}
{\bf x}_i\cdot{\bf x}_j .
\end{equation}
The operator ${\cal A}$ is an antisymmetrizer defined by
\begin{equation}
{\cal A}={1 \over \sqrt{N!} }\sum_P^{N!} {\rm sign}(P) P,
\end{equation}
where the sum runs over all permutations of the $N$ nucleon indices and
sign($P$) stands for the parity of the permutation $P$. For a system of
identical bosons, the antisymmetrizer is to be replaced with
a symmetrizer. For a
general case ${\cal A}$ is to represent the operator that imposes the proper
symmetry on the wave function.
\par\indent
The function $\theta_{LM_L}({\bf x})$ in Eq. (2), which represents the
angular part of the wave function, is a generalization of
$\cal Y$ and can be chosen as a vector-coupled
product of solid spherical harmonics of the Jacobi coordinates
\begin{equation}
\theta_{LM_L}({\bf x})=\left[[[{\cal Y}_{l_1}({\bf x}_1){\cal Y}_{l_2}
({\bf x}_2)]_{L_{12}}
{\cal Y}_{l_3}({\bf x}_3)]_{L_{123}}
,...\right]_{LM_L} .
\end{equation}
Each relative motion has a definite angular
momentum in Eq. (5). It may be important, however, to include several sets of
angular
momenta $(l_1,l_2,...,l_{N-1};L_{12},L_{123},...)$ for a realistic
description. The various possible
partial wave contributions increase the basis dimension; moreover, the
calculation of matrix elements for this choice of $\theta_{LM_L}({\bf x})$
becomes too complicated. This choice is apparently
inconvenient especially as the number of nucleons increases.
To avoid this, we propose a different choice as the generalization of
$\cal Y$:
\begin{equation}
\theta_{LM_L}({\bf x})=\eta_{KLM_L}({\bf u},{\bf x})=v^{2K+L} Y_{LM_L}
({\hat{\bf v}}),\ \ \ \ {\rm with}\ \ \
{\bf v}=\sum_{i=1}^{N-1} u_i {\bf x}_i.
\end{equation}
Only the total orbital angular momentum appears in this expression and it
contains a parameter $\tilde{\bf u}=(u_1,...,u_{N-1})$. The vector {\bf u}
may be considered as a variational parameter and one may try to minimize
the energy functional with respect to it. It defines a linear combination
of the Jacobi coordinates, ${\bf v}$, and the wave function of the system is
expanded in terms of its angle $\hat{\bf v}$. The minimization amounts to
finding the most suitable angle or a linear combination of angles. The
factor of $v^{2K+L}$ plays an important role in improving the short-range
behavior of the wave function. A remarkable advantage of this form of
$\theta_{LM_L}({\bf x})$ is that the calculation of matrix elements becomes
much simpler than in the former case because the coupling of $(N\!-\!1)$
angular momenta is completely avoided.
\par\indent
The two forms of $\theta_{LM_L}({\bf x})$ are in fact closely related to
each other. Any of the functions of Eq. (5) may be expressed in terms of
a linear combination of the terms, $v^{2K+L} Y_{LM_L}
({\hat{\bf v}})$, by using some appropriate sets of $\bf u$
values provided that each term satisfies the condition
$2K\!+\!L{\le}l_1\!\!+\cdot\cdot\cdot+\!l_{N-1}$ and contains a monomial of
degree $l_1+\cdot\cdot\cdot+l_{N-1}\!-\!2K\!-\!L$ in the variables,
${{\bf x}_1}^2,...,{{\bf x}_{N-1}}^2$. Therefore, if one can calculate
the matrix elements using $\theta_{LM_L}({\bf x})$ defined in Eq. (6), then
those with the previous form of $\theta_{LM_L}({\bf x})$ can be obtained
readily.
\par\indent
The correlated Gaussian basis with the function
$\theta_{LM_L}({\bf x})$ of Eq. (6) has parity $(-1)^L$. To construct
a function with parity $(-1)^{L+1}$, Eq. (6) must be slightly
generalized, e.g., to
\begin{equation}
\theta_{LM_L}({\bf x})=[\eta_{KL}({\bf u},{\bf x})\eta_{01}
({\bf u}',{\bf x})]_{LM_L}.
\end{equation}
\par\indent
To assure positive definiteness, the matrix $A$ in Eq. (2) is in general
expressed as $A =\tilde{G}A'G$, where $G$ is an $(N\!-\!1)\!\times\!
(N\!-\!1)$
orthogonal matrix containing $(N\!-\!1)(N\!-\!2)/2$
parameters and $A'$ is a diagonal matrix, $(A')_{ij}=a'_{i}
\delta_{ij}$, including $(N\!-\!1)$ positive parameters $a'_i$. Although no
restriction on the parameters of the matrix $G$
is in principle necessary, it is advisable to avoid too many variables
if possible. The most naive choice would be to take $G$ as a
unit matrix, which is equivalent to using only a single set of the
Jacobi coordinates, and then to try to reach
convergence by including higher partial waves successively. Many examples
show \cite{crcbgv,SVAO}, however, that this does not work well because
the convergence is generally slow and moreover the computational cost of
using high partial waves is quite expensive.
\par\indent
The matrix $G$ can also be chosen as one of the rotation matrices that
connect the set of the Jacobi coordinates to other sets of independent
relative coordinates. Figures
1a-1f show all topologically different sets of independent relative
coordinates for a system of six identical particles. The set of
coordinates in Fig. 1a is what we call the set of the Jacobi coordinates
${\bf x}$. A correlation conforming to a specific set of relative
coordinates $\tilde{\bf x}'=({\bf x}'_1,...,{\bf x}'_{N-1})$ can be most
efficiently described by tailoring the form of the basis function to this
set of relative coordinates, that is, by using the form,
exp$\lbrace -\frac{1}{2}\sum_{i=1}^{N-1}a'_i{\bf x}'_i\cdot{\bf x}'_i
\rbrace$. Since the coordinates ${\bf x}'$ can be obtained by an
appropriate rotation ${\cal R}$ of the Jacobi coordinates as ${\bf x}'=
{\cal R}{\bf x}$, the basis function of such type can be clearly
encompassed in the trial function of Eq. (2) by choosing $G={\cal R}$.
The correlated Gaussian basis thereby allows for
various correlations between the nucleons and different asymptotics at large
distances flexibly. Depending on the character of the problem a more general
choice of $G$ might be necessary.
\par\indent
By selecting a set of basis functions $\lbrace \psi_i; i=1,...,{\cal K}\rbrace$
[$\psi_i\equiv \psi_{(L_iS_i)JMTM_T}({\bf x}, A_i)$]
that adequately spans the state space, the wave function of the
$N$-nucleon system can be expanded as
\begin{equation}
\Psi=\sum_{i=1}^{\cal K} c_i \psi_i ,
\end{equation}
where $\tilde{\bf c}=(c_1,...,c_{\cal K})$ is the set of linear
variational parameters. The Ritz variational method defined by this trial
function
reduces to the generalized algebraic eigenvalue problem
\begin{equation}
{\cal H} {\bf c} = E {\cal N} {\bf c},
\end{equation}
where ${\cal H}$ and ${\cal N}$ are, respectively, the matrices of the
Hamiltonian and of the overlap
\begin{equation}
{\cal H}_{ij}=\langle \psi_i \vert H \vert \psi_j \rangle, \ \ {\rm and} \ \ \
\
{\cal N}_{ij}=\langle \psi_i \vert \psi_j \rangle \ \ \ \ (i,j=1,...,{\cal
K}).
\end{equation}
\subsection{Stochastic selection of parameters and solution of the
eigenvalue problem}
Since a linear combination of the correlated Gaussians forms a
dense set, there are different sets of $A$ that
represent the wave function equally well. This enables one to
select the most appropriate parameters randomly. We set up the basis
stepwise by choosing $A$ from a preset domain of the parameter space and
increase the basis dimension one by one.
In the first step we select a number of parameter sets $A$
randomly, and keep the one that gives the lowest energy. Next we generate
a new random set and calculate the energy with this two-element basis.
As one more basis state always lowers the energy, we quantify its
``utility'' by the energy gained by including it in the
basis. If
the energy gain is larger than a preset value, $\epsilon$, then we admit this
state to the basis, otherwise we discard it and try a new random candidate.
This is repeated until the energy converges.
The rate of convergence can be controlled by dynamically decreasing the
value of $\epsilon$ during the search.
This procedure is more advantageous than the earlier versions
\cite{SVM,VSL} and, although
not a full optimization, results in very good and relatively small bases.
A similar procedure, called ``stochastic diagonalization'' has been used to
determine the smallest eigenvalue of extremely large matrices \cite{SD}.
\par\indent
To have an economical algorithm for setting up the basis by a trial and
error method,
one has to find an efficient way to solve the eigenvalue problem, Eq. (9).
The full diagonalization is rather time
consuming and in fact unnecessary because (i) in the $({\cal K}\!\!+\!\!1)$th
step of the procedure
we can use the result of the ${\cal K}$th step and (ii) to judge the
usefulness
of a would-be basis
state, only the lowest eigenvalue is needed. Let us
assume that in the ${\cal K}$th step the
Hamiltonian matrix ${\cal H}$ is diagonalized; its eigenenergies are $E_1 \le
E_2 \le ...
\le E_{\cal K}$ and its
corresponding normalized eigenfunctions are
${\Psi_1},...,{\Psi_{\cal K}}$. The eigenvalue problem
in the $({\cal K}\!\!+\!\!1)$th step takes the form
\begin{eqnarray}
& &\left(
\begin{array}{ccccc}
E_1 & 0 & ... & 0& \langle {\Psi_1} \vert H
\vert \psi_{{\cal K}+1} \rangle \\
0 & & & & \\
\vdots & & & \vdots & \vdots \\
0 & & ... & E_{\cal K} & \langle {\Psi_{\cal K}} \vert H \vert
\psi_{{\cal K}+1} \rangle \\
\langle {\psi_{{\cal K}+1}} \vert H \vert \Psi_{1} \rangle & & ... &
\langle {\psi_{{\cal K}+1} }\vert H \vert { \Psi_{\cal K}}\rangle
&
\langle {\psi_{{\cal K}+1} }\vert H \vert { \psi_{{\cal K}+1}}\rangle \\
\end{array}
\right)
\left(
\begin{array}{c}
c_1 \\
\vdots \\
\vdots \\
c_{\cal K} \\
c_{{\cal K}+1} \\
\end{array}
\right)
\nonumber \\
& &=E
\left(
\begin{array}{ccccc}
1 &0 & ... & 0 & \langle {\Psi_1} \vert \psi_{{\cal K}+1} \rangle \\
0 & & & & \\
\vdots & & & \vdots & \vdots \\
0 & & ... & 1 & \langle {\Psi_{\cal K}} \vert \psi_{{\cal K}+1}
\rangle \\
\langle \psi_{{\cal K}+1} \vert {\Psi_{1}} \rangle & & ... &
\langle { \psi_{{\cal K}+1} } \vert \Psi_{\cal K}\rangle
&
\langle { \psi_{{\cal K}+1} } \vert \psi_{{\cal K}+1}\rangle \\
\end{array}
\right)
\left(
\begin{array}{c}
c_1 \\
\vdots \\
\vdots \\
c_{\cal K} \\
c_{{\cal K}+1}
\end{array}
\right).
\end{eqnarray}
By using the Gram-Schmidt orthogonalization method, that is, by defining
\begin{equation}
\vert {\bar\psi_{{\cal K}+1}} \rangle={\vert \psi_{{\cal K}+1}\rangle
-\sum_{i=1}^{\cal K} \vert {\Psi_{i} } \rangle \langle {\Psi_{i}} \vert
\psi_{{\cal K}+1} \rangle \over
( \langle \psi_{{\cal K}+1} \vert\psi_{{\cal K}+1} \rangle - \sum_{i=1}^{\cal
K}
\langle \psi_{{\cal K}+1} \vert{\Psi_{i}} \rangle
\langle{\Psi_{i}} \vert\psi_{{\cal K}+1} \rangle )^{1/2}} ,
\end {equation}
this generalized eigenvalue equation can be reduced to the conventional
form
\begin{equation}
\left(
\begin{array}{ccccc}
E_1 & 0 & ... & 0& q_1 \\
0 & & & & \\
\vdots & & & & \vdots \\
0 & & ... & E_{\cal K} & q_{\cal K} \\
q_1 & & ... & q_{\cal K} & a \\
\end{array}
\right)
\left(
\begin{array}{c}
c_1 \\
\vdots \\
\vdots \\
c_{\cal K} \\
c_{{\cal K}+1} \\
\end{array}
\right)
=E
\left(
\begin{array}{c}
c_1 \\
\vdots \\
\vdots \\
c_{\cal K} \\
c_{{\cal K}+1} \\
\end{array}
\right) ,
\end{equation}
where
\begin{equation}
q_i=\langle {\Psi_i} \vert H \vert {\bar\psi_{{\cal K}+1}} \rangle,
\ \ \ \ \ \
a=\langle {\bar\psi_{{\cal K}+1}} \vert H \vert {\bar\psi_{{\cal K}+1}}
\rangle .
\end{equation}
The eigenvalues are easily obtained
by finding the roots of the secular equation
\begin{equation}
\lambda(E)\equiv\prod_{i=1}^{\cal K} (E_{\cal K}-E)\left((a-E)-\sum_{j=1}^{\cal
K} {q_j^2\over E_j - E }
\right)=0 .
\end{equation}
This secular equation has $({\cal K}\!\!+\!\!1)$ roots $\lbrace
{E'}_i;i=1,...,{\cal K}\!\!+\!\!1\rbrace$
fulfilling the inequalities
${E'}_1\le E_1\le{E'}_2\le E_2\le ... \le E_{\cal K}\le {E'}_{{\cal K}+1}$.
The eigenvectors are readily obtained after substituting the eigenvalues
$\lbrace {E'}_i;i=1,...,{\cal K}\!\!+\!\!1\rbrace$ into Eq. (13).
Note that one has to determine only the lowest eigenvalue ${E'}_1$ for the
admittance criterion.
\section{Calculation of the matrix elements}
In this section we will give the details of the method of calculating the
matrix elements between
the basis function of Eq. (2). The calculation consists of three steps:
(A)The calculation of the matrix elements between the Slater
determinants of the Gaussian wave-packet single-particle functions, (B)A
transformation from the single-particle coordinate representation to the
relative and center-of-mass coordinate representation, (C)An integral
transformation from the Gaussian
wave-packet functions to the correlated Gaussian basis. A procedure similar
to steps (A) and (B) was used to manipulate algebraically the
antisymmetrization operation and the transformation of the coordinates for
complex cluster systems \cite{suzuki}. In step (A) the
Slater determinant for the $N$-nucleon wave function is constructed by
distributing the nucleons at positions $({\bf s}_1,...,{\bf s}_N)$.
These position vectors serve as the generator coordinates. The Slater
determinant of the
Gaussian wave packets is often used in nuclear theory, e.g., in cluster
model \cite{suzuki,Brink,wild,supple} and fermionic or antisymmetrized
molecular dynamics
\cite{feldmeier,horiuchi}. The Hamiltonian matrix elements
are analytically evaluated with the use of technique of the Slater
determinants \cite{Lowden,Brink},
and can be expressed as a function of the generator coordinates. In step (B)
the center-of-mass motion is completely separated from the intrinsic motion,
and thus the trial wave function acquires the translational invariance. The
separation of the center-of-mass motion is particularly simple in this
formulation. In the last step (C) the matrix elements expressed in terms of
the intrinsic generator coordinates are transformed to those between the
correlated Gaussian basis functions with a definite angular momentum. Some
of the essential parts of the calculational scheme is our original
development, and some of them is a
generalization of the technique used in the nuclear cluster model
(see, for example, \cite{Kami}). We also show in subsection III.D those
modifications which are needed to treat the system of particles of
unequal masses.
\subsection{Slater determinants of Gaussian wave packets}
The $i$th nucleon with mass $m$, spin $\sigma_i$ and isospin
$\tau_i$ is to be put in the single-particle Gaussian wave packet
\begin{equation}
{\hat \varphi}_{{{\bf s}_i}{\sigma_i}{\tau_i}}^{\nu}({\bf r}_i)=
\varphi_{{\bf s}_i}^{\nu}({\bf r}_i)
\chi_{{1\over 2}\sigma_i}{\cal X}_{{1\over 2}\tau_i},
\end{equation}
with
\begin{equation}
\varphi_{{\bf s}_i}^{\nu}({\bf r}_i)=\left({2 \nu \over \pi}
\right)^{3/4} {\rm e}^{-\nu ({\bf r}_i-{\bf s}_i)^2},\ \ \ \ \ {\rm and}
\ \ \ \nu={m \omega \over 2 \hbar},
\end{equation}
where ${\bf r}_i$ is the position vector of the nucleon,
$\chi_{{1\over 2}\sigma_i}$ and ${\cal X}_{{1\over 2}\tau_i}$ are its spin
and isospin function. The angular frequency $\omega$ is
not a variational parameter and may be taken an arbitrary constant. The
${\bf s}_i$ parameter or ``generator''
coordinate will be used in an integral transformation to derive the matrix
elements between the Gaussian basis functions.
A Slater determinant of these Gaussian packets is defined by
\begin{equation}
\phi_{\kappa}({\bf s}_1,...,{\bf s}_N)={\cal A} \Biggl\{
\prod_{i=1}^{N}{\hat \varphi}_{{{\bf s}_i}{\sigma_i}{\tau_i}}^{\nu}
({\bf r}_i) \Biggr\} ,
\end{equation}
where $\kappa=(\sigma_1\tau_1,...,\sigma_N\tau_N)$ is the set of the
spin-isospin quantum numbers of the
nucleons.
The spins and the isospins of the nucleons are successively coupled to
add up, respectively, to the
total spin $S M_S$ and isospin $T M_T$ of the $N$-nucleon system:
\begin{equation}
{\chi}_{SM_S}=
\left[[[\chi_{1\over 2}\chi_{1\over 2}]_{S_{12}}
\chi_{1\over 2} ]_{S_{123}}
,...\right]_{SM_S} ,\ \ \ \
{\cal X}_{TM_T}=
\left[[[{\cal X}_{1\over 2}{\cal X}_{1\over 2}]_{T_{12}}
{\cal X}_{1\over 2}]_{T_{123}}
,...\right]_{TM_T} .
\end{equation}
To simplify the notation, the intermediate quantum numbers are suppressed
in the following.
The wave function in the ``generator coordinate space'' with the
definite spin and isospin quantum numbers is a linear
combination of the Slater determinants of the Gaussian packets:
\begin{equation}
\Phi_{SM_STT_z}({\bf s}_1,...,{\bf s}_N)=
{\cal A} \lbrace
\varphi_{{\bf s}_1}^{\nu}({\bf r}_1)...
\varphi_{{\bf s}_N}^{\nu}({\bf r}_N) {\chi}_{SM_S}{\cal X}_{TM_T}\rbrace
=\sum_{\kappa} c_{\kappa} \phi_{\kappa}({\bf s}_1,...,{\bf s}_N) ,
\end{equation}
where $c_{\kappa}$ is a product of the Clebsch-Gordan coefficients needed
to couple the spin and isospin as defined in Eq. (19).
\par\indent
The Hamiltonian of the $N$-nucleon system reads as
\begin{equation}
H=\sum_{i=1}^{N} {{\bf p}_i^2 \over 2 m} +\sum_{i<j}^{N} V_{ij} .
\end{equation}
The matrix elements of the Slater determinants can easily be evaluated using
the well-known rules \cite{Lowden,Brink}. To make this paper
self-contained, we have collected all the needed ingredients in Appendices
A, B and C. The overlap of the Slater determinants is found to take the form
\begin{equation}
\langle
\Phi_{SM_STM_T}({\bf s}_1,...,{\bf s}_N)
\vert
\Phi_{SM_STM_T}({{\bf s}'}_1,...,{{\bf s}'}_N)
\rangle=
\sum_{i=1}^{n_{o}} C_{i}^{(o)}
{\rm e}^{-{1\over 2} \tilde{{\bf s}} A_{i}^{(o)} {\bf s}} ,
\end{equation}
where $A_{i}^{(o)}$ is a $2N\!\times\!2N$ real, symmetric matrix and
$\tilde{{\bf s}}$ stands for the $2N$-dimensional row vector
comprising the single-particle generator coordinates,
$({\bf s}_1,...,{\bf s}_N,{\bf s}'_1,...,{\bf s}'_N)$.
To simplify the notation, we refer to the set of the vectors
$({\bf s}'_1,...,{\bf s}'_N)$ alternatively as
$({\bf s}_{N+1},...,{\bf s}_{2N})$. Note, therefore, that
the quadratic form, $\tilde{{\bf s}} A_{i}^{(o)} {\bf s}$, reads as
\begin{equation}
\tilde{{\bf s}} A_{i}^{(o)} {\bf s}=\sum_{j=1}^{2N} \sum_{k=1}^{2N}
\left(A_{i}^{(o)}\right)_{jk}
{\bf s}_j\cdot{\bf s}_k .
\end{equation}
The matrix elements of the kinetic energy
operator can also be expressed in terms of the same $C_{i}^{(o)}$'s and
$A_{i}^{(o)}$'s as
\begin{eqnarray}
& & \langle
\Phi_{SM_STM_T}({\bf s}_1,...,{\bf s}_N)
\vert \sum_{i=1}^{N} {{\bf p}_i^2 \over 2 m} \vert
\Phi_{SM_STM_T}({{\bf s}'}_1,...,{{\bf s}'}_N)
\rangle
\nonumber \\
& &={\hbar \omega \over 2}
\sum_{i}^{n_o} C_{i}^{(o)} \left({3 \over 2}N-{1\over 2}
\tilde{{\bf s}} A_{i}^{(o)} {\bf s} \right)
{\rm e}^{-{1\over 2} \tilde{{\bf s}} A_{i}^{(o)} {\bf s}}.
\end{eqnarray}
The matrix elements of any term of the two-body interaction
can be expressed as an integral of the two-particle correlation function
multiplied by the radial form factor, $V(r)$, of the term $V_{ij}$ as below:
\begin{eqnarray}
& &\langle
\Phi_{SM_STM_T}({\bf s}_1,...,{\bf s}_N)
\vert \sum_{i<j}^{N} V_{ij} \vert
\Phi_{S'M_S'T'M_T'}({{\bf s}'}_1,...,{{\bf s}'}_N)
\rangle
\nonumber \\
& & =\int d{\bf r} V(r) {\rm e}^{-\nu r^2}
\sum_{i=1}^{n_{p}} C_{i}^{(p)}
P_{i}({\bf s},{\bf r})
{\rm e}^{-{1\over 2} {\tilde{\bf s}} A_{i}^{(p)} {\bf s}+{\bf d}_i \cdot
{\bf r}},
\end{eqnarray}
where
$P_{i}({\bf s},{\bf r})$ is a polynomial of ${\bf s}$ and ${\bf r}$,
$A_{i}^{(p)}$ a $2N\!\times\!2N$ symmetric matrix, and
${\bf d}_i\cdot{\bf r}$ takes the form
\begin{equation}
{\bf d}_i\cdot{\bf r}=\sum_{j=1}^{2N} {\cal D}_{(i)j}{\bf s}_j
\cdot{\bf r}.
\end{equation}
The polynomial part reduces to unity
($P_{i}({\bf s},{\bf r})=1$) in the case of pure central forces,
but it has a rather simple form for spin-orbit, tensor, and other
interactions as well. See Appendices B and C. The $C_{i}^{(o)}$'s,
$A_{i}^{(o)}$'s,
$C_{i}^{(p)}$'s, and $A_{i}^{(p)}$'s, etc. are obtained with the use of
mathematical manipulation languages or fortran programs. See Appendix C.
\subsection{Transformation to relative and center-of-mass coordinates}
To eliminate the center-of-mass motion, we transform the single-particle
coordinates to the relative and center-of-mass coordinates. For this purpose
we choose one particular set of relative coordinates, the Jacobi
coordinates, which is expressed in terms of the single-particle
coordinates ${\bf r}_i$ as
\begin{equation}
{\bf x}_i=\sum_{k=1}^{N} U_{ik} {\bf r}_k \ \ \ \ (i=1,...,N),
\end{equation}
where the transformation matrix $U$ is defined by
\begin{equation}
U=
\left(
\begin{array}{ccccc}
-1 & 1 & 0 &... & 0 \\
-{1 \over 2} & -{1 \over 2} & 1 &... & 0 \\
\vdots & & & &\vdots \\
-{1 \over N-1}&-{1\over N-1} & ... &... & 1 \\
{1\over N} &{1\over N} & ... &... &{1 \over N}
\end{array}
\right),
\ \ \ \ \ \ {\rm and}\ \ \ \ \
U^{-1}=
\left(
\begin{array}{ccccc}
-{1 \over 2} & -{1 \over 3} &... & -{1 \over N} & 1 \\
{1 \over 2} & -{1 \over 3} & & -{1 \over N} & 1 \\
0 & {2 \over 3} & & \vdots & \vdots \\
\vdots & \vdots & & \vdots & \vdots \\
0 & 0 & ... & {N-1 \over N} & 1
\end{array}
\right).
\end{equation}
Similarly, the single-particle generator coordinates are transformed to
the relative and center-of-mass generator coordinates:
\begin{equation}
{\bf S}_i=\sum_{k=1}^{N} U_{ik} {\bf s}_k, \ \ \ \ \ \
{\bf S}_i'=\sum_{k=1}^{N} U_{ik} {\bf s}_k' \ \ \ \ \ \ (i=1,...,N).
\end{equation}
The reduced masses corresponding to the transformation $U$ are given by
\begin{equation}
\mu_i={i \over i+1} m \ \ \ \ (i=1,...,N\!\!-\!\!1),
\ \ \ \ \ \ \ {\rm and}
\ \ \ \ \ \mu_N=N m.
\end{equation}
The product of the Gaussian single-particle
wave packets can then be written as a product of Gaussians depending on the
relative and center-of-mass coordinates:
\begin{equation}
\prod_{i=1}^{N} \varphi_{{\bf s}_i}^{\nu}({\bf r}_i)=
\prod_{i=1}^{N} \varphi_{{\bf S}_i}^{\gamma_i}({\bf x}_i) ,
\end{equation}
with
\begin{equation}
\gamma_i={\mu_i \omega \over 2 \hbar}.
\end{equation}
By using Eq. (31) and noting that the last factor of the product depends
only on the
center-of-mass coordinate, which is symmetric under the exchange of nucleons,
the $N$-nucleon wave function can be rewritten as
\begin{equation}
\Phi_{SM_STM_T}({\bf s}_1,...,{\bf s}_N)=
\Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1})
\varphi_{{\bf S}_N}^{\gamma_N}({\bf x}_N) ,
\end{equation}
which defines the intrinsic function that depends solely
on the relative coordinates,
\begin{equation}
\Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1})=
{\cal A} \lbrace
\varphi_{{\bf S}_1}^{\gamma_1}({\bf x}_1) ,...,
\varphi_{{\bf S}_{N-1}}^{\gamma_{N-1}}({\bf x}_{N-1})
{\chi}_{SM_S}{\cal X}_{TM_T}\rbrace .
\end{equation}
As will be shown in the next subsection, this function serves as a
generating function of the correlated Gaussian basis.
\par\indent
The Hamiltonian of Eq. (21) can be recast to the relative plus
center-of-mass terms as
\begin{equation}
H=\left( \sum_{i=1}^{N-1} {{{\bf P}_i}^2 \over 2 \mu_i}+
\sum_{i<j}^{N} V_{ij}
\right)+{{{\bf P}_N}^2 \over 2 \mu_N}\ \ {\equiv}\ \ H_{rel}+T_{cm} ,
\end{equation}
where ${\bf P}_i$ is the momentum canonically conjugate to the Jacobi
coordinate ${\bf x}_i$. The matrix elements of the Hamiltonian are then
\begin{eqnarray}
& & \langle \Phi_{SM_STM_T}({\bf s}_1,...,{\bf s}_N)
\vert H \vert \Phi_{S'M_S'T'M_T'} ({{\bf s}'}_1,...,{{\bf s}'}_N) \rangle
\nonumber \\
& &=\langle \Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1})
\vert H_{rel} \vert
\Psi_{S'M_S'T'M_T'}({{\bf S}'}_1,...,{{\bf S}'}_{N-1}) \rangle
\langle\varphi_{{\bf S}_N}^{\gamma_N} \vert
\varphi_{{{\bf S}'}_N}^{\gamma_N}\rangle \nonumber\\
& &
+\langle \Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1}) \vert
\Psi_{S'M_S'T'M_T'}({{\bf S}'}_1,...,{{\bf S}'}_{N-1}) \rangle
\langle\varphi_{{\bf S}_N}^{\gamma_N} \vert
T_{cm} \vert
\varphi_{{{\bf S}'}_N}^{\gamma_N}\rangle .
\end{eqnarray}
The matrix elements of the Hamiltonian $H_{rel}$ can be expressed by
integrating
over ${\bf S}_N$
\begin{eqnarray}
& &\langle \Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1}) \vert H_{rel} \vert
\Psi_{S'M_S'T'M_T'}({{\bf S}'}_1,...,{{\bf S}'}_{N-1}) \rangle
\nonumber \\
& & = \left({{\gamma_N} \over 2\pi}\right)^{3/2}
\int d{\bf S}_N
\langle \Phi_{SM_STM_T}({\bf s}_1,...,{\bf s}_N)
\vert H \vert
\Phi_{S'M_S'T'M_T'} ({{\bf s}'}_1,...,{{\bf s}'}_N) \rangle ,
\end{eqnarray}
where use has been made of the single-particle matrix elements in
Appendix A and the formula
\begin{equation}
\int d{\bf S}_N \left(-{{\gamma_N} \over 2}({\bf S}_N -{{\bf S}'}_N)^2
\right)^n
{\rm e}^{-{{\gamma_N} \over 2}({\bf S}_N -{{\bf S}'}_N)^2}=
\left(-{1\over 2}\right)^{n}(2n+1)!! \left({2\pi \over
{\gamma_N}}\right)^{3\over2}.
\end{equation}
\par\indent
Equation (37) shows that the matrix elements between the intrinsic
function can be obtained in a simple way by factorizing
the ${\bf S}_N$-dependent terms from the matrix elements in the
single-particle basis:
The quadratic forms in Eqs. (22), (24) and (25), with the help of
Eq. (36), should take the form
\begin{equation}
\tilde{{\bf s}} A_i^{(k)} {\bf s}=\tilde{{\bf S}} B_i^{(k)} {\bf S}+
{\gamma_N} ({\bf S}_N-{{\bf S}'}_N)^2 \ \ \ \ (k=o,p).
\end{equation}
Here the $(2N-2)$-element column vector ${\bf S}$ is an abbreviation for the
set of the Jacobi generator-coordinate vectors
$({\bf S}_1,...,{\bf S}_{N-1},{{\bf S}'}_1,...,{{\bf S}'}_{N-1})$.
The matrix $B_i^{(k)}$ is the $(2N\!\!-\!\!2)\!\!\times\!\!(2N\!\!-\!\!2)$
symmetric
matrix defined by dropping
the $N$th and $2N$th rows and columns of the matrix $\tilde{T} A_i^{(k)} T$,
where
\begin{equation}
T=
\left(
\begin{array}{cc}
U^{-1} & 0 \\
0 & U^{-1}
\end{array}
\right) .
\end{equation}
It is clear from Eq. (36) that the polynomials $P_{i}({\bf s},{\bf r})$
and the vectors ${\bf d}_i$
defined in Eq. (25) can depend only on the relative generator coordinates.
The dependence of the matrix elements on the center-of-mass variables,
${\bf S}_N$ and ${\bf S}'_N$, can, therefore, be factorized and the
integration in Eq. (37) reduces to such a simple form as the one in Eq. (38).
After the integration over ${\bf S}_N$ the matrix elements can be
expressed in terms of the Jacobi generator coordinates in the form:
\begin{equation}
\langle
\Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1})
\vert
\Psi_{SM_STM_T}({{\bf S}'}_1,...,{{\bf S}'}_{N-1})
\rangle=
\sum_{i=1}^{n_{o}} C_{i}^{(o)}
{\rm e}^{-{1\over 2} \tilde{\bf S} B_{i}^{(o)} {\bf S}},
\end{equation}
\begin{eqnarray}
& & \langle
\Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1})
\vert \sum_{i=1}^{N} {{\bf P}_i^2 \over 2 \mu_i} \vert
\Psi_{SM_STM_T}({{\bf S}'}_1,...,{{\bf S}'}_{N-1})
\rangle
\nonumber \\
& &={\hbar \omega \over 2}
\sum_{i}^{n_o} C_{i}^{(o)} \left({3 \over 2}(N-1)-{1\over 2}
\tilde{\bf S} B_{i}^{(o)} {\bf S}\right)
{\rm e}^{-{1\over 2} \tilde{\bf S} B_{i}^{(o)} {\bf S}},
\end{eqnarray}
\begin{eqnarray}
& &\langle
\Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1})
\vert \sum_{i<j}^{N} V_{ij} \vert
\Psi_{S'M_S'T'M_T'}({{\bf S}'}_1,...,{{\bf S}'}_{N-1})
\rangle
\nonumber \\
& & =\int d{\bf r} V(r) {\rm e}^{-\nu r^2}
\sum_{i=1}^{n_{p}} C_{i}^{(p)}
P_{i}({\bf S},{\bf r})
{\rm e}^{-{1\over 2} \tilde{\bf S} B_{i}^{(p)} {\bf S}+{\bf D}_i
\cdot {\bf r}}.
\end{eqnarray}
Here the convention of renumbering the set of the vectors
$({{\bf S}'}_1,...,{{\bf S}'}_{N-1})$ of the ket as $({\bf S}_N,...,
{\bf S}_{2N-2})$ is used to simplify the notation and thus ${\bf S}$ stands
for the set of the vectors $({\bf S}_1,...,{\bf S}_{2N-2})$. The vector
${\bf D}_i$ is given by $\sum_{j=1}^{2N-2} {\hat{D}}_{(i)j}
{\bf S}_j $, where the ${\hat D}_{(i)j}$'s are formed from the elements
of the row vector
$\tilde{\cal D}_{(i)}T$ by omitting
its $N$th and $2N$th columns. The column vector
${\cal D}_{(i)}$ is defined in Eq. (26).
\subsection{Integral transformation to the correlated Gaussian basis}
\par\indent
Here we show how to evaluate the matrix elements in the correlated Gaussian
basis. Let us choose the correlated Gaussians with the form of Eq. (6) and
introduce the function
\begin{equation}
f_{KLM}({\bf u},{\bf x},A)=\eta_{KLM} ({\bf u},{\bf x}){\rm e}^{-{1\over 2}
\tilde{\bf x} A {\bf x}},
\end{equation}
where
\begin{equation}
\eta_{KLM}({\bf u},{\bf x})=v^{2K+L} Y_{LM}({\hat{\bf v}}),\ \ \ \ {\rm with}
\ \ \ \ \ {\bf v}=\sum_{i=1}^{N-1} u_i {\bf x}_i=\tilde{\bf u}{\bf x},
\end{equation}
and where $\tilde{\bf u}=(u_1,...,u_{N-1})$. Note that $\bf u$ is a set
of $(N\!-\!1)$ real numbers, whereas $\bf x$ are the $(N\!-\!1)$
Jacobi coordinates. The calculation of the matrix elements
becomes simple if one uses a generating function of the correlated Gaussian.
In fact, the following function $g$ is found to
be most convenient to generate the function $f$:
\begin{equation}
f_{KLM}({\bf u},{\bf x},A)={1\over B_{KL}}
\int d{\hat {\bf t}} Y_{LM}({\hat {\bf t}})
\left( {d^{2K+L} \over d\alpha^{2K+L}}
g(\alpha,{\bf t};{\bf u},{\bf x},A)
\right)_{\alpha=0 \atop t=|{\bf t}|=1},
\end{equation}
where
\begin{equation}
g(\alpha,{\bf t};{\bf u},{\bf x},A)=
{\rm e}^{-{1\over 2}\tilde{\bf x} A {\bf x} +\alpha {\bf v}\cdot{\bf t}},
\end{equation}
\begin{equation}
B_{nl}={4 \pi (2n+l)! \over 2^n n! (2n+2l+1)!!}.
\end{equation}
Equation (46) is easily proved by using the simple formula
\begin{equation}
({\bf v}\cdot{\bf t})^{k}=
v^k t^k \sum_{n,l \ge 0 \atop 2n+l=k} B_{nl} \sum_{m=-l}^{l}
Y_{lm}({\hat {\bf v}})
Y_{lm}({\hat {\bf t}})^{\ast}.
\end{equation}
For a case where the function $\theta_{LM_L}({\bf x})$ of Eq. (7) is needed,
the generating function $g$ of Eq. (47) must be generalized to include
another factor $\alpha'{\bf v}'\cdot{\bf t}'$ . Since the following derivation
remains essentially unaffected by this generalization, we will assume
Eq. (6) as $\theta_{LM_L}({\bf x})$.
\par\indent
The generating function $g$ can be related to the product of the Gaussians
centered around $\{{\bf S}_i;i=1,...,N\!-\!1\}$ through an integral
transformation. To show this, we express the product of the Gaussian
wave packets as
\begin{equation}
\prod_{i=1}^{N-1}
\varphi_{{\bf S}_i}^{\gamma_i}({\bf x}_i)=\left(\frac{{\rm det}\Gamma}{\pi^
{N-1}}\right)^{3/4}{\rm e}^{-\frac{1}{2}\tilde{\bf x}\Gamma{\bf x}+
\tilde{\bf x}
\Gamma{{\bf S}_H}-\frac{1}{2}{{\tilde{\bf S}}_H}
\Gamma{{\bf S}_H}}
\end{equation}
with an $(N\!\!-\!\!1)\!\times\!(N\!\!-\!\!1)$ diagonal matrix
\begin{eqnarray}
\Gamma=
\left(
\begin{array}{cccc}
2\gamma_1 & 0 & ...& 0 \\
0 & 2\gamma_2 & & \vdots \\
\vdots & & & \vdots \\
0 &... & ... & 2\gamma_{N-1}
\end{array}
\right),
\end{eqnarray}
where ${{\bf S}_H}$ stands for the set of the generator coordinate
vectors
$({\bf S}_1,...,{\bf S}_{N-1})$.
By using the
familiar formula of the $n$-dimensional Gaussian integration
\begin{equation}
\int d{\bf x} {\rm e}^{-{1\over 2}\tilde{\bf x}A {\bf x}+\tilde{\bf T}
{\bf x}}=
\left({(2\pi)^{n} \over {\rm det} A }\right)^{3/2}
{\rm e}^{{1\over 2}\tilde{\bf T} A^{-1} {\bf T}},
\end{equation}
it is easy to prove the following equation by a direct calculation
\begin{equation}
g(\alpha,{\bf t};{\bf u},{\bf x},A)= \left(\frac{({\rm det}\Gamma)^{3/2}}
{(4\pi)^{(N-1)/2}{\rm det}(\Gamma-A)}
\right)^{3/2}
{\rm e}^{-{1\over 2} \tilde{\bf T} C {\bf T}}
\int d{{\bf S}_H}
{\rm e}^{-{1\over 2} {{\tilde{\bf S}}_H} Q
{{{\bf S}_H}}+\tilde{\bf T}{{\bf S}_H}}
\left(\prod_{i=1}^{N-1}
\varphi_{{\bf S}_i}^{\gamma_i}({\bf x}_i)
\right),
\end{equation}
where $\tilde{\bf T}=({\bf T}_1,...,{\bf T}_{N-1})$, and
\begin{equation}
C=\Gamma^{-1}(\Gamma-A)\Gamma^{-1} ,\ \ \ \ \
Q=C^{-1}-\Gamma ,
\end{equation}
\begin{equation}
{\bf T}_i=\alpha {\bf t}\sum_{j=1}^{N-1}(\Gamma C)^{-1}_{ij}
u_j \ \ \ \ (i=1,...,N\!\!-\!\!1).
\end{equation}
By combining Eqs. (46) and (53), the correlated Gaussian basis is found to
be generated from the intrinsic state given in Eq. (34) by the integral
transformation
\begin{eqnarray}
& & {\cal A} \{f_{KLM}({\bf u},{\bf x},A){\chi}_{SM_S}{\cal X}_{TM_T}\}
={1\over B_{KL} }\left(\frac{({\rm det}\Gamma)^{3/2}}
{(4\pi)^{(N-1)/2}{\rm det}(\Gamma-A)}
\right)^{3/2}
\nonumber \\
&\times& \int d{\hat {\bf t}}
Y_{LM}({\hat {\bf t}})
\left( {d^{2K+L} \over d\alpha^{2K+L}}
{\rm e}^{-{1\over 2} \tilde{\bf T} C
{\bf T}}
\int d{{{\bf S}_H}}
{\rm e}^{-{1\over 2} {{\tilde{\bf S}}_H} Q
{{{\bf S}_H}}+\tilde{\bf T}{{{\bf S}_H}}}
\Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1})
\right)_{\alpha=0 \atop t=1} .
\end{eqnarray}
\par\indent
The matrix elements between the correlated Gaussians are now easily
obtained
by the integral transformation from those expressed in terms of the relative
generator coordinates ${\bf S}$. Using Eq. (56) gives a general formula to
calculate a matrix element for any translation-invariant operator
$\cal O$
\begin{eqnarray}
& &\langle {\cal A} \{f_{KLM}({\bf u},{\bf x},A){\chi}_{SM_S}{\cal X}_{TM_T}\}
\vert {\cal O} \vert {\cal A'} \{f_{K'L'M'}({\bf u}',{\bf x},A')
{\chi}_{S'M'_S}{\cal X}_{T'M'_T}\}\rangle
\nonumber \\
&=&\frac{1}{B_{KL}B'_{K'L'}}\left(\frac{({\rm det}\Gamma)^3}
{(4\pi)^{(N-1)}{\rm det}(\Gamma-A){\rm det}(\Gamma-A')} \right)^{3/2}
\nonumber \\
&\times& \int\!\int d{\hat {\bf t}}d{\hat {\bf t}}'
Y_{LM}({\hat {\bf t}})^{*} Y_{L'M'}({\hat {\bf t}}')
\left( {d^{2K+L+2K'+L'} \over d\alpha^{2K+L}d\alpha'^{2K'+L'}}
{\rm e}^{-{1\over 2} \tilde{\bf T} C {\bf T}} \right.\\
&\times& \left. \int d{{\bf S}}
{\rm e}^{-{1\over 2} \tilde{{\bf S}} Q {{\bf S}}+\tilde{\bf T}{{\bf S}}}
\langle \Psi_{SM_STM_T}({\bf S}_1,...,{\bf S}_{N-1}) \vert {\cal O} \vert
\Psi_{S'M'_ST'M'_T}({{\bf S}'}_1,...,{{\bf S}'}_{N-1}) \rangle
\right)_{\alpha=\alpha'=0 \atop t=t'=1},
\nonumber
\end{eqnarray}
where $\tilde{\bf S}=({\bf S}_1,...,{\bf S}_{N-1},{{\bf S}'}_1={\bf S}_N,...,
{{\bf S}'}_{N-1}={\bf S}_{2N-2})$ and the matrices, $C$ and $Q$, and the
vectors ${\bf T}$ in Eq. (57), although the same notation is used as in
Eqs. (54) and (55), are extended to include the corresponding primed
quantities of the ket, that is,
\begin{equation}
C \longrightarrow \pmatrix{
C & 0 \cr
0 & C' \cr
},
\ \ \ \
Q \longrightarrow \pmatrix{
Q & 0 \cr
0 & Q' \cr
},
\ \ \ \
{\bf T} \longrightarrow \pmatrix{
{\bf T} \cr
{\bf T}' \cr
}.
\end{equation}
As is shown in Eqs. (41)-(43), the ${\bf S}$-dependence of the matrix
elements is rather simple and the integration over ${\bf S}$ is done
analytically. Since the variables $\alpha, \alpha', {\bf t}$, and
${\bf t}'$ appear only through the vector ${\bf T}$, those operations
with respect to them as implied in Eq.
(57) are performed systematically. An illustrative example is given in
Appendix D. The coupling of the orbital and spin angular momenta
causes no difficulty. It is very satisfactory aesthetically that matrix
elements between the basis functions with any sets of the relative
coordinates can be evaluated in a unified framework without
any extra transformation of the coordinates. The choice of the set
of the relative coordinates amounts to the choice of the matrix $A$.
\subsection{Extension to the system of particles of unequal masses}
\par\indent
In this subsection we remark the modifications needed to treat
few-particle systems containing particles of unequal masses. As you will
see, most of the formulation presented in subsections III.A$-$III.C remains
unchanged.
Suppose that the masses of the $N$ particles are $m_1,m_2,...,m_N$. The
width parameter of the Gaussian wave packet is to be changed to
\begin{equation}
\nu_i={m_i\omega \over {2\hbar}} \ \ \ \ (i=1,...,N).
\end{equation}
The overlap and Hamiltonian matrix elements are obtained as a function of
the generator coordinates, $({\bf s}_1,...,{\bf s}_N,{\bf s}'_1,...,
{\bf s}'_N)$, in a form similar to the previous case. In this general case
of unequal masses,
the matrix $U$ in Eq. (28) which defines a set of the Jacobi coordinates
must be generalized to
\begin{equation}
U=
\left(
\begin{array}{ccccc}
-1 & 1 & 0 &... & 0 \\
-{m_1 \over m_{12}} & -{m_2 \over m_{12}} & 1 &... & 0 \\
\vdots & & & &\vdots \\
-{m_1 \over m_{12\cdot\cdot\cdot{N-1}}} &-{m_2 \over
m_{12\cdot\cdot\cdot{N-1}}}
& ... &... & 1 \\
{m_1 \over m_{12\cdot\cdot\cdot{N}}} & {m_2\over m_{12\cdot\cdot\cdot{N}}}
& ... &... &{m_N \over m_{12\cdot\cdot\cdot{N}}}
\end{array}
\right),
\end{equation}
where $m_{12\cdot\cdot\cdot{i}}=m_1\!+\!m_2\!+\cdot\cdot\cdot+\!m_i$. The
reduced
masses corresponding to this $U$ are accordingly given, instead of
Eq. (30), by
\begin{equation}
\mu_i={m_{i+1}m_{12\cdot\cdot\cdot{i}} \over m_{12\cdot\cdot\cdot{i+1}}}
\ \ \ \ (i=1,...,N\!\!-\!\!1),
\ \ \ \ \ \ \ {\rm and}
\ \ \ \ \ \mu_N=m_{12\cdot\cdot\cdot{N}}.
\end{equation}
What is important is to realize that Eq. (31), most crucial in eliminating
the center-of-mass motion, still holds even with these modifications as
\begin{equation}
\prod_{i=1}^{N} \varphi_{{\bf s}_i}^{\nu_i}({\bf r}_i)=
\prod_{i=1}^{N} \varphi_{{\bf S}_i}^{\gamma_i}({\bf x}_i).
\end{equation}
It is then easy to see that the rest of all the formulae are exactly the
same as the case of equal masses. We can conclude that the needed
modifications noted above are rather trivial and simple but still assures
the elimination of the center-of-mass motion. As a simple example of unequal
masses, the system of $t+d+\mu^{-}-$molecule will be considered in
subsection IV.B.
\section{Numerical results}
This section is devoted to present the solutions of various
$N\!\!=\!\!2-7$-body
problems by applying the method described above. To test the method,
different potentials (Yukawa, Gauss and Coulomb) have been used for bosonic
and fermionic systems. Some of the examples shown here has its own physical
significance, and
some other solutions may be considered as benchmark test and might
be useful in comparison of various few-body methods. One can expect that,
besides the VMC\cite{mcv,vmc} and GFMC\cite{gfma,gfmb}, other methods
will also be extended to treat more than $N=4$-particle systems. As
only a few solutions are at present available for simple
potentials, the examples listed here may help to test other methods.
\par\indent
As was discussed in subsection II.A, there is no restriction on the choice
of the orthogonal matrix $G$. We have found, however, that those
special rotation matrices which connect different sets of the relative
coordinates especially suitable (see
also \cite{SVAO}) and will use them in what follows.
This greatly helps to reduce the number of parameters of $G$.
In the stochastic selection of the basis elements,
these special matrices $G$ and the parameters of the diagonal matrix $A'$
are randomly chosen.
The vector ${\bf u}$ in Eq. (6) is also a variational
parameter. To avoid an excessively large number of variational parameters we
limited ${\bf u}$-vector values to those which are needed to
generate the function ${\theta}_{LM_{L}}({\bf x})$ of Eq. (5) for a given
set of angular momenta. Comparison of our calculation with others confirms
that these limitations have not deteriorated the accuracy of the present
calculation, that is, our trial function is flexible enough. In the
calculation the sets of angular momenta (i.e., the sets of these special
${\bf u}$ vectors) are also randomly chosen. The main advantage of using
${\bf u}$ lies in the simple and systematic evaluation of the matrix
elements from the point of view of both analytical and numerical
calculations. Further test calculations will be needed to explore the
utility of ${\bf u}$ as a variational parameter.
\par\indent
Because the dependence of the matrix elements on the variational parameters
is known as is shown in Appendix D, one can organize the numerical
calculation involved in the
random search economically. A change of the values of the parameters does
not require a recalculation of the whole matrix element. Once they have been
calculated for one set of values, to calculate them for many more requires
virtually no time. The average computational time
is 10 minutes for a four-body and 2 hours for a six-body calculation
on the VPP500 computer of RIKEN.
\subsection{Few-nucleon systems}
We have performed model calculations adopting different central potentials
as nucleon-nucleon interaction. Some of these model problems have already
been solved to high accuracy by various methods and therefore we can
directly compare the solutions. The potentials used for comparison of
different methods are
(i) the Malfliet-Tjon (MT-V) potential \cite{mt}, which has been most
extensively used as benchmark test in few-body
calculations, (ii) the Volkov ``super-soft'' core potential \cite{Volkov},
(iii) the Afnan-Tang S3 (ATS3) potential \cite{ATS3}
which exhibits a strong repulsive core and incorporates
a difference between the spin-singlet and spin-triplet channels, (iv) and
the Minnesota potential \cite{minesota} which reproduces
the most important low-energy nucleon-nucleon phase shifts. The Volkov and
MT-V potentials are spin-independent, while the ATS3 and Minnesota
potentials are spin-dependent. The parameters of the interactions are
tabulated in Table I. We choose $\hbar^2/m=41.47$ MeV fm$^2$. The Coulomb
interaction is included only in calculations with the Minnesota potential
where point charges are assumed and $e^2=1.44$ MeV fm.
\par\indent
The spins (and isospins) of the nucleons are coupled through successive
intermediate couplings. The spin couplings up to $N\!\!=\!\!7$ nucleons are
tabulated in Table II. One naturally expects and test calculations show
that, without spin-isospin coupling, the energy convergence is much slower.
The number of spin-isospin configurations rapidly increases with $N$. In
the case
of $^6$He, for example, assuming $S=0$ and $T=1$, the wave function has
$5\!\times\! 9=45$ spin-isospin components. The number of components
becomes even
higher if the interaction has non-central spin-orbit, tensor, etc. parts.
The nonlinear parameters are not optimized with respect to spin-isospin
components, but rather, for each trial choice of the matrix $A$, we select
the spin-isospin component
that gives the lowest energy. Because the matrix elements between different
spin-isospin components differ only in linear factors ($c_{\kappa}$ in
Eq. (20)), the calculation
of the matrix elements of each spin-isospin component of the wave function
requires essentially the same computational effort as that needed by the
calculation for only one component.
\par\indent
Each calculation has been repeated several times starting
from different random points to check the energy convergence.
The energy as a function of the number of basis states
is shown in Fig. 2 for the case of $^6$Li with the Volkov potential.
The energies on different random paths, after a few initial steps,
approach to each other and converge to the final
solution. The energy difference between two random paths as well
as the tangent of the curves give us some information on the accuracy
of the method on a given size of the basis. The root mean square (rms)
radius of the few-nucleon system is calculated in each step and
found to be rapidly convergent to its final value. By increasing the
basis size the results can be arbitrarily improved when needed.
\par\indent
The number of basis states required to reach energy convergence increases
with the number of particles but it depends on the form of the interaction
as well. This latter property is illustrated in Fig. 3 for the case
of the $\alpha$-particle. The soft-core Volkov potential shows rapid
convergence, while the hard-core ATS3 interaction requires more basis states
to get an accurate solution. The relatively fast convergence for the MT-V
potential of a strong repulsion can be explained by the simplicity -- the
spin-independent nature -- of this interaction.
\par\indent
In the following we show tables for the ground-state energies $E$ and point
matter rms radii $\langle r^2\rangle^{1/2}$. The basis dimension ${\cal K}$
of the SVM listed in the tables is such that, beyond it, the energy and
the radius do not change in the digits shown. Table III shows our results
(SVM), together with results of others, for the
application of the spin-averaged MT-V potential \cite{mt} to $N$=2--7-nucleon
systems. For three-body systems, the solution of the Faddeev equation
is known to be the method of choice, but the SVM can
easily yield the same accurate energy. As the MT-V potential is a preferred
benchmark test of the few-body calculations, there are numerous solutions
available. Table III includes a few of the most accurate results.
The nice agreement for four-nucleon case corroborates that the SVM is as
accurate as the direct solution of the FY equations
\cite{glocka}, the method of
the Amalgamation of Two-body correlations into Multiple Scattering (ATMS)
\cite{atms} process or the VMC \cite{wiringa} and GFMC \cite{gfma} method.
The basis used
in the Coupled-Rearrangement-Channel Gaussian-basis (CRCG) variational
method \cite{crcbgv} is similar to that of the SVM, but
the Gaussian parameters follow geometric progressions.
The fact that the basis size needed in the SVM is much smaller
proves the efficiency of the selection procedure.
The results of the VMC calculation for the five- and six-nucleon systems
are also in good agreement with the results of SVM. The MT-V potential
has no exchange term; therefore,
unlike the nature, it renders the five-nucleon system bound,
and the nucleus tends to collapse as the binding energy increases with the
number of particles.
\par\indent
The next example is the Volkov potential which, due to its very soft core,
is the most readily solvable case. This simple potential is widely used in
model calculations for light nuclei. As one sees in Table IV, the results
of SVM agree with those of other calculations, especially with the one
using hyperspherical harmonics (HH) functions. The number of basis states
needed to reach convergence is remarkably smaller than in the case of
the MT-V potential. Without the Majorana exchange term
($M=0$) this potential also leads to a collapsing
system. By setting the Majorana parameter to $M$=0.6, a commonly used value
to get the correct binding energy, one may obtain more reasonable energies.
The Volkov potential with $M$=0.6 does not change the energies of $N$=2$-$4
nuclei, does not bind $^5$Li in accordance with the nature, but does bind
the $^6$Li ground state ($E$=$-$31.82 MeV, ${\langle r^2 \rangle}^{1/2}$
=2.69 fm).
\par\indent
Another potential that is often used in test calculation is the ATS3
potential. We have challenged the SVM to get solution for this case
because, unlike the Volkov, this spin-dependent potential has a relatively
strong repulsive core (see Table I). The solution, although on a somewhat
larger basis, can easily be obtained, and
it is in good agreement with those of other methods in the $N$=3 system
as shown in Table V. We note, however, that the energy of the SVM is
significantly lower than the ones of other methods for the
$\alpha$-particle. This may be due to the strong repulsion of this
potential. For example, the FY calculation \cite{glocka} agrees
with ours for the
MT-V potential, but shows a noticeable difference in the case of the
ATS3 potential. The variational calculation \cite{etbm} using a
correlation factor also misses considerable energy for the ATS3 potential.
Surprisingly, inspite of its exchange part, this potential also binds the
five-nucleon system and overbinds the six-nucleon systems very much.
\par\indent
The last example for the few-nucleon system uses the more realistic
Minnesota potential\cite{minesota}, which is a central interaction of
Gaussian form,
containing space-, spin-, and isospin-exchange operators (see Table I).
The Minnesota potential has often been used in cluster-model
calculations of light nuclei. Table VI shows results with this potential,
where the Coulomb interaction between protons is also
included. All possible spin and isospin configurations
are allowed for and all spherical harmonics that give
non-negligible contribution are included in $\theta_{LM_L}({\bf x})$.
Since the method has proved to be accurate and reliable
for other potentials, it is justifiable to view
these results as testing the interaction rather than the method,
and hence the results are compared with experimental data.
The energy and the radius of triton and $\alpha$-particle converge, with
small
bases, to realistic values. The Minnesota potential, correctly, does not
bind the $N$=5 system, but it binds $^6$He and
slightly overbinds $^6$Li. The radius of $^6$He is found to be much larger
than that of $^4$He, consistently with the halo structure of $^6$He
\cite{SVAO}. It is for the first time that
the Minnesota force is tested without assuming any cluster structure or
restricting the model space by any other bias. The agreement is surprisingly
good not only with experiment but also with cluster-model calculations for all
nuclei \cite{LKBD}.
\par\indent
It is interesting to note that none of these simple potentials binds the
four-neutron system. The Volkov potential, for example, is so strong that it
binds the singlet
two-neutron system, but it does not allow the neutrons to form a
four-neutron bound state due to the Pauli principle.
\par\indent
As an example for bosonic nuclear few-body system, we consider the
case of structureless $\alpha$-particles interacting via the state-independent
potential;
\begin{equation}
V(r)=500\,\,{\rm exp}\,[-(0.7r)^2\,]-130\,\,{\rm exp}\,[-(0.475r)^2\,] \ \ \ \
({\rm MeV}),
\end{equation}
where $r$ is in fm. This potential is taken from Ali-Bodmer's $S$-state
potential\cite{AB}.
It has a repulsive core which is about 370 MeV high
and extends up to 2 fm. The repulsive core prevents the $\alpha$-particles from
collapsing.
The results are compared in Table VII. Our calculation agrees with the
ATMS result for the $N=$3 and 4 systems.
\subsection{Coulombic systems}
The results for the long range $1/r$ potential are collected in this
subsection. The first example is the polyelectric system $(me^+,ne^-)$.
The possibility that $m$ positrons and $n$ electrons form a bound system
was originally suggested by
Wheeler \cite{wheeler} and this question has been extensively
studied since then. Besides the trivial and analytically solvable $m=1$
and $n=1$
case, the existence of the positronium negative ion ($m=1,n=2$) was also
predicted by Wheeler \cite{wheeler}.
Dozens of works have attempted to solve the $e^+,e^-,e^-$ Coulombic
three-body problem, continuously
refining the accuracy of the calculated binding energy \cite{epe,ho}.
Despite of numerous attempts, no one has obtained bound states for the
polyelectric system of more than four particles.
The positronium negative ion has experimentally been observed \cite{mills}.
\par\indent
The binding energy of the positronium molecule $(2e^+,2e^-)$ was first
calculated by Hylleraas and Ore \cite{hyll}. To date, the positronium
molecule has not been directly observed, and this fact intensifies the
theoretical interest to solve this Coulombic four-body problem.
This molecule is short-lived because the electron and positron may
annihilate. Unlike the positronium ion, the positronium molecule is
neutral, and therefore the best
chance to distinguish it from the positronium itself is related to their
different lifetimes.
The QED formulae to determine the probability of a pair annihilation in
the positronium
molecule through a $k$-photon process ($k=0$,1,2,...) would require a highly
accurate wave function \cite{frolov95}.
\par\indent
In Table VIII we compare our results to the most precise calculations found in
the literature.
The correlated Gaussian function without the polynomial part
($K=0$ in Eq. (6))
is known to poorly represent the Coulomb cusps \cite{KA91}. To improve the cusp
properties
the trial function with $K=0,1,2$ polynomials has been used.
\par\indent
As is shown in Table VIII, our calculation reproduces the first six digits
of the variational calculation of Ref.~\cite{epe} for the ground state of
(2e$^+$,e$^-$), and the rms radius also agrees with it.
There are two recent variational calculations \cite{epep,KA93} for the
positronium molecule
using the correlated Gaussian functions. In these works the nonlinear
parameters were determined by optimatization. To compare our calculations
to theirs directly, the value of $K=0$ was chosen and the same basis size
(${\cal K}$=300) was used. Our result is slightly better than the energy
obtained by them
and this reinforces the reliability and powerfulness of the random selection of
the nonlinear
parameters.
The number of nonlinear parameters of this case is ${\cal K}\!\!\times
\!\!(4\!\!\times \!\!3)/2=1800$. The complete optimatization of the parameters
is, of course, superior to the SVM. Test calculations show that, provided
the number of parameters is low, that is, a
full optimatization is feasible, the optimatization finds lower energy.
But when the number of parameters becomes high, the full optimatization
becomes
less and less practical partly because it fails to find true minimum and
partly because the computation becomes too excessive.
\par\indent
We found no bound states for the ($3e^+,2e^-$) and ($3e^+,3e^-$)
systems. The energy of ($3e^+,3e^-$), for example,
converges to the sum of the energy of a dipositronium molecule and of
a positronium (0.515989 a.u.+0.25 a.u.=0.765989 a.u.). Allowing the selection
of the parameters from a larger region increases the rms radius, which is
typical of an unbound state. The system of
a negative and a positive positronium ion thus forms no bound state but
dissociates into a dipositronium molecule and a positronium.
\par\indent
Calculations for $L=1$ state also fails to find a bound system.
This result entails that the Coulomb force cannot bind more than four
particles out of identical charged fermions and their antiparticles.
\par\indent
To examine the role of the Pauli principle in preventing five-electron-positron
system from forming bound states, we repeated the same calculation
replacing the fermions by bosonic equivalents.
On a different scale, these systems may be identified, e.g., as the
systems of $\pi^-$ and $\pi^+$ with their strong interaction neglected
\cite{boson}. Such bosons turn out to form bound states even for $N=5$.
As is expected, the radius of the charged boson
system decreases as the number of particles increases. It is interesting
to note that the energies of bosonic and fermionic systems are
equal for $N=3$ and for $N=4$. The reason is that the energy minimum belongs
to the same spatial configurations, that is, to a triangular pyramid for
$N=4$, for example \cite{ho86}.
\par\indent
In Table IX we show results for bosonic and fermionic systems with a
purely attractive
$Gm^2/r$ (``gravitational") interaction.
An $N$-body system of identical particles bound together by attractive
pair potentials always collapses in large-$N$ limit
(the binding energy per particle rises with $N$ to infinity),
even if the particles are fermions. Self-gravitating boson systems have
recently attracted some interest \cite{selfgrav}. For these systems,
both variational lower and upper bounds are available.
In this case even the five-fermion system
is bound. Thus the lack of bound states in
five-electron-positron systems is a joint effect of the antisymmetry
and of the repulsion between identical particles.
\par\indent
Finally, we mention an example involving an excited state. With ${\cal K}=500$,
the SVM gives the energies of the ground and first excited states of the
$t+d+\mu^-$ system as $-$111.3640 and $-$100.9121 a.u., which are respectively
compared to $-$111.364342 and $-$100.916421 a.u. of the CRCG result
\cite{dtmu} with ${\cal K}=1442$, while the configuration-space Faddeev
calculation \cite{faddeev} gives $-$111.36 a.u. for the ground state but no
information for the excited state.
\section{Summary}
\par\indent
We have formulated a variational calculation for few-body systems using the
stochastic variational method on the
correlated Gaussian basis. We have demonstrated the versatility of the
correlated Gaussians and the efficiency of the stochastic variation by
various numerical examples for $N=2-7$-particle systems. All the details of
both formulation and calculational procedure are included to make this
paper self-contained and easily reproducible.
\par\indent
The comparison with other calculations has corroborated the accuracy and
efficiency of the method. In none of the test cases has the present method
proved to be inferior to any of the alternative methods, and yet the method
does not require excessive computational effort.
\par\indent
The correlation between the particles plays an important role in describing
the few-body system realistically. It has been taken into account in the
framework of the correlated Gaussian functions. The correlated Gaussians
are constructed from
products of the Gaussian wave-packet single-particle functions through an
integral transformation, which has enabled us to evaluate the
center-of-mass motion free matrix elements analytically starting from the
single-particle level. The nonlinear parameters of the correlated Gaussians
have been selected by the stochastic variational method with a trial and
error procedure. The success of the method using the correlated Gaussian
basis is probably due to the fact that none of the Gaussians is
indispensable, that is,
there are different sets of the Gaussian parameters that represent the wave
function equally well.
\par\indent
The method presented in this paper can be useful to solve few-body problems
in diverse fields of physics such as description of microclusters,
non-relativistic quark model, and halo nuclei. Among others, the most
important application is the
solution of the nuclear few-body problem, that is, a description of light
nuclei by using ${\it realistic}$ nucleon-nucleon potentials. In this case
one has to take into account both short-range repulsion and higher orbital
angular momenta required by the non-central components. Our test examples
show that the correlated Gaussian basis function might be a suitable
candidate to cope with these requirements. As is explained
in Appendix B, the evaluation of the matrix elements for the non-central
potentials poses no serious problem, and calculations including such
potentials for nuclear few-body systems are under way.
\par\indent
The limitations of
the present method are those implied by the basis size and by the computer
memory to store the matrix elements in the generator coordinate space. The
limitations may become excessive as the number of particles and/or spin
and isospin configurations become large.
\par\indent
To extend the method to nuclei of larger mass number in an approximate
way, one can
freeze part of the model space for a group of nucleons (cluster). One can
omit, for example, some of the spin or isospin channels. It might be a
good approximation to consider only those spin channels where the spins of
the like nucleon pairs are coupled to zero. One can also restrict the intrinsic
spatial motion of a cluster by fixing the nonlinear parameters to some
appropriate values. One can introduce $N$ clusters and place the nucleons
of each cluster into a common harmonic oscillator well, for example. The
microscopic multicluster model is based on this approximation. The matrix
elements needed in this multicluster model are given as a special case of those
presented
in this paper. In fact, one only needs to choose the single-particle
generator coordinates, $({\bf s}_1,...,{\bf s}_A)$, such that
\begin{equation}
{\bf s}_{n_{i-1}+1}={\bf s}_{n_{i-1}+2}=\cdot\cdot\cdot=
{\bf s}_{n_{i-1}+n_i} \ \ \ \ (i=1,...,N),
\end{equation}
where $n_i$ is the number of nucleons in
the ${\it i}$th cluster ($n_0=0,\ \ n_1+\cdot\cdot\cdot+n_N=A$), and needs
to couple appropriately the spins and isospins of the nucleons in each
cluster to the spin and isospin of the cluster. The microscopic
multicluster model has been successfully applied for description of the
structure of light nuclei (see, for example, \cite{SVAO,Brink,wild}).
\par\indent
\par\indent
Finally, we summarize some merits of our method in the following:
\begin{description}
\item{(i)} Fully analytical calculational scheme; this plays a major
role in the high speed and accuracy of the calculation.
\item{(ii)} Universality of the scheme. One needs to introduce
no change, for example, between describing a multinucleon system
and a Coulombic few-body system. It is easily adaptable to identical
or non-identical particles, to fermions or bosons or mixed systems.
The masses of particles may be different, yet no problem with
the center-of-mass motion arises.
\item{(iii)} No expansion of the interaction is needed, and thus
no problems in partial-wave truncation arise.
\item{(iv)} The convergence of the energy is fast. If one needs just a
2--3-digit estimate of the energy, it is enough to use a very small basis.
\item{(v)} The method is also accurate for excited states, which are
obtained simultaneously with the same diagonalization (provided their
angular momenta and parities are the same as those of the ground state;
but only such excited states may be problematic).
\item{(vi)} The wave function is obtained in a compact, analytical form.
It is then easy to use it in calculations of physical properties.
It is ``portable'', reproducible, and easily testable.
\end{description}
\section*{Appendix A. Single-particle matrix elements}
The aim of this appendix is to list the single-particle matrix elements
between Gaussian wave packets (Eq. (17)).
The Gaussian packets are generalized in this appendix to have different width
parameters and the expressions are therefore slightly more general
than needed
in the formulae of the main text. These single-particle matrix elements are,
however, required for treating particles of unequal masses, in which
the width parameter $\nu$ belonging to the particle of mass $m$ is
to be chosen by Eq. (59). The overlap of two Gaussian wave packets is
\begin{equation}
\langle \varphi_{{\bf s}_1}^{\nu_1} \vert \varphi_{{\bf s}_2}^{\nu_2} \rangle
=\left({2 \sqrt{\nu_1 \nu_2} \over \nu_1 +\nu_2}\right)^{3/2}
{\rm exp}\,\left(-{\nu_1 \nu_2 \over \nu_1 +\nu_2} ({{\bf s}_1}-
{{\bf s}_2})^2\,\right) .
\end{equation}
The matrix element of the kinetic energy operator $(T=-\frac{{\hbar}^2}
{2M}\Delta)$ reads as
\begin{equation}
\langle \varphi_{{\bf s}_1}^{\nu_1} \vert T\vert\varphi_{{\bf s}_2}^{\nu_2}
\rangle
={\hbar^2 \over 2M }
{2\nu_1\nu_2 \over \nu_1 +\nu_2}
\left( 3-{2 \nu_1 \nu_2\over \nu_1+\nu_2} ({\bf s}_1-{\bf s}_2)^2 \right)
\langle \varphi_{{\bf s}_1}^{\nu_1} \vert
\varphi_{{\bf s}_2}^{\nu_2} \rangle .
\end{equation}
The matrix element of the square radius becomes
\begin{equation}
\langle \varphi_{{\bf s}_1}^{\nu_1} \vert {\bf r}_1^2 \vert
\varphi_{{\bf s}_2}^{\nu_2} \rangle
= {1 \over 2(\nu_1 +\nu_2)}
\left( 3+{2 \over \nu_1+\nu_2} (\nu_1{\bf s}_1+\nu_2{\bf s}_2)^2 \right)
\langle \varphi_{{\bf s}_1}^{\nu_1} \vert
\varphi_{{\bf s}_2}^{\nu_2} \rangle .
\end{equation}
The two-particle matrix element of a $\delta$-function is given by
\begin{eqnarray}
& &\langle \varphi_{{\bf s}_1}^{\nu_1}\varphi_{{\bf s}_2}^{\nu_2}
\vert \delta({\bf r}_1-{\bf r}_2-{\bf r})\vert
\varphi_{{\bf s}_3}^{\nu_3}\varphi_{{\bf s}_4}^{\nu_4} \rangle
=
\left({(\nu_1+\nu_3)(\nu_2+\nu_4) \over
(\nu_1+\nu_2+\nu_3+\nu_4) \pi }\right)^{3/2} \nonumber \\
&\times&
{\rm exp}\,\Biggl(-{(\nu_1+\nu_3)(\nu_2+\nu_4) \over
\nu_1+\nu_2+\nu_3+\nu_4}
\Bigl({\bf r}-{\nu_1{\bf s}_1+\nu_3{\bf s}_3\over \nu_1+\nu_3}
+{\nu_2{\bf s}_2+\nu_4{\bf s}_4\over \nu_2+\nu_4}\Bigr)^2\,\Biggr)
\langle \varphi_{{\bf s}_1}^{\nu_1} \vert \varphi_{{\bf s}_3}^{\nu_3}
\rangle
\langle \varphi_{{\bf s}_2}^{\nu_2} \vert \varphi_{{\bf s}_4}^{\nu_4}
\rangle .
\end{eqnarray}
\section*{Appendix B. Matrix elements of the two-body potentials}
The scope of this appendix is the calculation of the matrix elements
of the different ingredients of the two-nucleon interaction
between Gaussian wave packets. Most of the widely used
coordinate-space two-nucleon interactions consist of central- ($O^c_{12}$),
tensor- ($O^t_{12}$),
spin-orbit- ($O^b_{12}$), ${\bf L}^2$- (or ${\bf p}^2$) ($O^q_{12}$) and
quadratic spin-orbit ($O^{bb}_{12}$) type potentials.
(We follow the abbreviated notations
-- $c,t,b,q,bb$ -- invented by Urbana-Argonne group \cite{LP81,WSA84}. The
definition of these operators is given in Table X.)
These potential terms might be multiplied by the
\mbox{\boldmath{$\tau$}}$_1\cdot$\mbox{\boldmath{$\tau$}}$_2$ ($O^\tau_{12}$),
\mbox{\boldmath{$\sigma$}}$_1\cdot$\mbox{\boldmath{$\sigma$}}$_2$
($O^\sigma_{12}$),
or \mbox{\boldmath{$\tau$}}$_1\cdot$\mbox{\boldmath{$\tau$}}$_2$
\mbox{\boldmath{$\sigma$}}$_1\cdot$\mbox{\boldmath{$\sigma$}}$_2$
($O^{\tau \sigma}_{12}$) spin-isospin operators
and then one ends up with the general form
\begin{equation}
V_{ij}=
\sum_{p} \int d{\bf r} V^{p}(r) \delta({\bf r}_i-{\bf r}_j-{\bf r})
O_{ij}^{p} ,
\end{equation}
where $p=c,c\tau,c\sigma,c\tau\sigma,t,t\tau,b,b\tau,q,q\tau,q\sigma,
q\tau\sigma,
bb,bb\tau$ is the short-hand notation to specify the component and $V^p(r)$
is the corresponding radial form factor. By using Eqs. (16), (17), (68),
(69), and after some straightforward transformation, the matrix element of
this interaction can be written as
\begin{eqnarray}
& &\langle
{\hat \phi}^{\nu}_{{\bf s}_1\sigma_1\tau_1}
{\hat \phi}^{\nu}_{{\bf s}_2\sigma_2\tau_2}
\vert
V_{12}
\vert
{\hat \phi}^{\nu}_{{{\bf s}_1}'\sigma_1'\tau_1'}
{\hat \phi}^{\nu}_{{{\bf s}_2}'\sigma_2'\tau_2'}
\rangle =
\sum_{p} \int d{\bf r} V^{p}(r) f_{p}(r){\cal M}({\bf r})\nonumber \\
&\times&\langle\chi_{{1\over 2}\sigma_1}{\cal X}_{{1\over 2}\tau_1}\,
\chi_{{1\over 2}\sigma_2}{\cal X}_{{1\over 2}\tau_2}\vert\,
r^{2}B^{p}+\sum_{l=0}^{2}\sum_{m=-l}^{l}
(-1)^{m}r^{l}Y_{lm}({\hat {\bf r}}) C^{p}_{l\,-m}\,\vert\,
\chi_{{1\over 2}\sigma'_1}{\cal X}_{{1\over 2}\tau'_1}\,
\chi_{{1\over 2}\sigma'_2}{\cal X}_{{1\over 2}\tau'_2}\rangle,
\end{eqnarray}
with
\begin{equation}
{\cal M}({\bf r})=\Bigl({\nu \over \pi}\Bigr)^{3/2}
{\rm exp}\,\Biggl(-\nu r^2+\nu ({\bf s}_1\!\!-\!\!{\bf s}_2\!\!+\!\!
{\bf s}'_1\!\!-\!\!{\bf s}'_2)\cdot{\bf r}-{\nu \over 4}
({\bf s}_1\!\!-\!\!{\bf s}_2\!\!+\!\!{\bf s}'_1\!\!-\!\!{\bf s}'_2)^2\,
\Biggr)\,\langle \varphi_{{\bf s}_1}^{\nu} \vert \varphi_{{\bf s}'_1}
^{\nu} \rangle
\langle \varphi_{{\bf s}_2}^{\nu} \vert \varphi_{{\bf s}'_2}^{\nu}
\rangle .
\end{equation}
The $B^{p}$ and $C^{p}_{lm}$, independent of ${\bf r}$, are the operators
in spin-isospin space and are listed in Table X for the most important terms
($c,t,b,q,bb$). The remaining terms
can be easily derived by multiplying these operators by the appropriate
$\mbox{\boldmath $\sigma$}$ and $\mbox{\boldmath $\tau$}$ operators. The
function $f_p(r)$ has the simple form:
$f_p(r)=r^{-2}$ if $p=t$ or $t\tau$, and $f_p(r)=1$ otherwise.
\par\indent
The calculation of matrix elements of the operators
appearing in Table X in the spin part is easily done with the
Clebsch-Gordan coefficient. They are given below by suppressing the spin
function $\chi_{\frac{1}{2} \sigma_i}\,(m_i=\frac{1}{2} \sigma_i)$.
\begin{eqnarray}
& &\langle m_1\,m_2\,\vert\,
{\mbox{\boldmath $\sigma$}_1}\cdot{\mbox{\boldmath $\sigma$}_2}\,
\vert \,m'_1\,m'_2\,\rangle \nonumber \\
&=&3(-1)^{m_1-m'_1}\delta_{m_1+m_2,m'_1+m'_2}
\langle {1 \over 2}\,m'_1\,1\,m_1\!\!-\!\!m'_1\,\vert\,
{1 \over 2}\,m_1\rangle
\langle {1 \over 2}\,m'_2\,1\,m_2\!\!-\!\!m'_2\,\vert\,
{1 \over 2}\,m_2\rangle.
\end{eqnarray}
\begin{eqnarray}
& &\langle m_1\,m_2\,\vert\,
\left[{\mbox{\boldmath $\sigma$}_1}\times{\mbox{\boldmath $\sigma$}_2}
\right]^{(2)}_{m}\, \vert \,m'_1\,m'_2\,\rangle=3\,\delta_{m,m_1+m_2-m'_1-m'_2}
\nonumber \\
&\times&\langle {1 \over 2}\,m'_1\,1\,m_1\!\!-\!\!m'_1\,\vert\,
{1 \over 2}\,m_1\rangle
\langle {1 \over 2}\,m'_2\,1\,m_2\!\!-\!\!m'_2\,\vert\,
{1 \over 2}\,m_2\rangle
\langle 1\,m_1\!\!-\!\!m'_1\,1\,m_2\!\!-\!\!m'_2\,\vert\,2\,m\rangle.
\end{eqnarray}
\begin{eqnarray}
& &\langle m_1\,m_2\,\vert\,
\left[{\bf x}\times
({\mbox{\boldmath $\sigma$}_1}+{\mbox{\boldmath $\sigma$}_2})
\right]^{(1)}_{m}\, \vert\,m'_1\,m'_2\,\rangle=\sqrt{3}\, \sum_{q_1,q_2=-1}
^{1}({\bf x})_{q_1} \nonumber \\
&\times&\langle 1\,q_1\,1\,q_2\,\vert\,1\,m\rangle
\Bigl\{\delta_{m_2,m'_2}
\langle{1\over 2}\,m'_1\,1\,q_2\,\vert\,{1\over 2}\,m_1\rangle+
\delta_{m_1,m'_1}
\langle{1\over 2}\,m'_2\,1\,q_2\,\vert\,{1\over 2}\,m_2\rangle\Bigr\}.
\end{eqnarray}
\begin{eqnarray}
& &\langle m_1\,m_2\,\vert\,
({\bf x}\cdot{\mbox{\boldmath $\sigma$}_2})
{\mbox{\boldmath $\sigma$}_{1_{m}}}+
({\bf x}\cdot{\mbox{\boldmath $\sigma$}_1})
{\mbox{\boldmath $\sigma$}_{2_{m}}}
\, \vert \,m'_1\,m'_2\,\rangle \nonumber \\
&=&3\,\Bigl\{(-1)^{m_2-m'_2}\,\delta_{m,m_1-m'_1}({\bf x})_{m'_2-m_2}+
(-1)^{m_1-m'_1}\,\delta_{m,m_2-m'_2}({\bf x})_{m'_1-m_1}\Bigr\} \\
&\times&\langle {1 \over 2}\,m'_1\,1\,m_1\!\!-\!\!m'_1\,\vert\,
{1 \over 2}\,m_1\rangle
\langle {1 \over 2}\,m'_2\,1\,m_2\!\!-\!\!m'_2\,\vert\,
{1 \over 2}\,m_2\rangle . \nonumber
\end{eqnarray}
\begin{eqnarray}
& &\langle m_1\,m_2\,\vert\,
({\bf x}\cdot{\mbox{\boldmath $\sigma$}_1})
({\bf x}\cdot{\mbox{\boldmath $\sigma$}_2})
\, \vert \,m'_1\,m'_2 \rangle \,
=3\,(-1)^{m_1-m'_1+m_2-m'_2} \nonumber \\
&\times&({\bf x})_{m'_1-m_1}({\bf x})_{m'_2-m_2}
\langle {1 \over 2}\,m'_1\,1\,m_1\!\!-\!\!m'_1\,\vert\,
{1 \over 2}\,m_1\rangle
\langle {1 \over 2}\,m'_2\,1\,m_2\!\!-\!\!m'_2\,\vert\,
{1 \over 2}\,m_2\rangle .
\end{eqnarray}
\begin{eqnarray}
& &\langle m_1\,m_2\,\vert\,
\left[\,\left[{\bf x}\times{\mbox{\boldmath $\sigma$}_1}\right]^{(1)}
\times\left[{\bf x}\times{\mbox{\boldmath $\sigma$}_2}\right]^{(1)}
\,\right]^{(2)}_{m}
\, \vert \,m'_1\,m'_2\,\rangle \nonumber \\
&=&3\,\langle {1 \over 2}\,m'_1\,1\,m_1\!\!-\!\!m'_1\,\vert\,
{1 \over 2}\,m_1\rangle
\langle {1 \over 2}\,m'_2\,1\,m_2\!\!-\!\!m'_2\,\vert\,
{1 \over 2}\,m_2\rangle\,\nonumber \\
&\times& \sum_{q_1,q_2=-1}^{1}({\bf x})_{q_1}({\bf x})_{q_2}
\langle 1\,q_1\!\!+\!\!m_1\!\!-\!\!m'_1\,1\,q_2\!\!+\!\!m_2\!\!-\!\!
m'_2\,\vert\,2\,m\rangle \\
&\times&\langle
1\,q_1\,1\,m_1\!\!-\!\!m'_1\,\vert\,1\,q_1\!\!+\!\!m_1\!\!-\!\!m'_1
\rangle
\langle 1\,q_2\,1\,m_2\!\!-\!\!m'_2\,\vert\,1\,q_2\!\!+\!\!m_2\!\!-\!\!m'_2
\rangle. \nonumber
\end{eqnarray}
Here ${\bf x}$ is a 3-dimensional vector and $({\bf x})_m$ stands for its
spherical component $\sqrt{4\pi\over 3} x Y_{1m}({\hat{\bf x}})$.
\section*{Appendix C. Matrix elements of Slater determinants}
In this appendix we briefly outline the calculation of the matrix elements
between Slater determinants of Eq. (18) and show their concrete functional
form, that is,
the dependence on the generator coordinates ${\bf s}$. We assume that the
width parameter of the Gaussian wave packet is chosen to be the same for
all nucleons.
The overlap of two Slater determinants is equal to the determinant of
the matrix of the single-particle overlaps:
\begin{equation}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N) \vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle
={\rm det} \lbrace B \rbrace,
\end{equation}
where
\begin{equation}
B_{ij}=
\langle {\hat\varphi}_{{\bf s}_i\sigma_i\tau_i}^{\nu}
\vert
{\hat\varphi}_{{{\bf s}'}_j{\sigma}'_j{\tau}'_j}^{\nu} \rangle
\ \ \ \ \ \ \ (i,j=1,...,N).
\end{equation}
By using the definition of the determinant, this can be rewritten as
\begin{equation}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N) \vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle
= \sum_{P}^{N!} {\rm sign}(P)
\langle {\hat \varphi}_{{\bf s}_1\sigma_1\tau_1}^{\nu} \vert
{\hat\varphi}_{{{\bf s}'}_{p_{1}}{\sigma}'_{p_1}{\tau}'_{p_1}}^{\nu}
\rangle...
\langle {\hat\varphi}_{{\bf s}_N\sigma_N\tau_N}^{\nu} \vert
{\hat\varphi}_{{{\bf s}'}_{p_{N}}{\sigma}'_{p_{N}}{\tau}'_{p_{N}}}
^{\nu} \rangle ,
\end{equation}
where $(p_{1}\cdot \cdot \cdot p_{N})$ is the permutation $P$ of the set
$(1\cdot \cdot \cdot N)$.
Substituting the overlap of the single-particle overlaps of Eq. (65)
into Eq. (80) yields an explicit formula for the overlap of the
Slater determinants:
\begin{equation}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N) \vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle =
\sum_{P}^{N!} C_{P}
{\rm e}^{-{1\over 2} \tilde{\bf s} A_{P} {\bf s}} ,
\end{equation}
where the matrix $A_P$ is defined by
\begin{equation}
(A_P)_{ij}=(A_P)_{N+i,N+j}= \nu \delta_{ij},\ \ \ \
(A_P)_{i,N+j}=(A_P)_{N+j,i}=-\nu \delta_{j,p_{i}}
\ \ \ \ (1\le i,j \le N),
\end{equation}
and
\begin{equation}
C_{P}={\rm sign}(P)
\delta_{{\sigma_1}{\sigma'}_{p_1}} \delta_{{\tau_1}{\tau'}_{p_1}}
\cdot\cdot\cdot \delta_{{\sigma_N}{\sigma'}_{p_N}}
\delta_{{\tau_N}{\tau'}_{p_N}}.
\end{equation}
The orthogonality of the spin-isospin functions greatly reduces the number
of terms in the summation over $P$.
\par\indent
The matrix element of the kinetic energy operator is
\begin{equation}
\sum_{i=1}^{N}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N) \vert T_i \vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle =
\sum_{i=1}^{N}
\sum_{j=1}^{N}
\langle{\hat \varphi}_{{{\bf s}_i}{\sigma_i}{\tau_i}}^{\nu}
\vert T \vert {\hat \varphi}_{{{{\bf s}'}_j}{\sigma'_j}
{\tau'_j}}^{\nu} \rangle
(-1)^{i+j} {\det} \lbrace B^{ij}\rbrace,
\end{equation}
where $B^{ij}$ is obtained by omitting the $i$th row and the $j$th column
of the matrix $B$ defined in Eq. (79).
The substitution of the single-particle overlaps and the
single-particle matrix elements of the kinetic energy operator enables us
to obtain
\begin{equation}
\sum_{i=1}^{N}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N) \vert T_i \vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle =
{\hbar \omega \over 2}
\sum_{P}^{N!} C_{P} \left({3 \over 2}N -{1\over 2} \tilde{\bf s} A_{P}
{\bf s}\right){\rm e}^{-{1\over 2} \tilde{\bf s} A_{P} {\bf s}}.
\end{equation}
Note that the coefficients $C_{P}$ and the matrices $A_P$ in Eqs. (81) and
(85) are the same. The manipulation similar to this leads us to the
matrix element of the square radius as
\begin{equation}
\sum_{i=1}^{N}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N) \vert {\bf r}_i^2 \vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle =
{1 \over 2\nu}
\sum_{P}^{N!} C_{P} \left({3 \over 2}N -{1\over 2} \tilde{\bf s} A_{P}
{\bf s} + \nu\tilde{\bf s}{\bf s} \right){\rm e}^{-{1\over 2} \tilde{\bf s}
A_{P} {\bf s}}.
\end{equation}
\par\indent
As is explained in Appendix B, the matrix element of any two-body
interaction between the Slater determinants may be reduced to the
following:
\begin{equation}
\sum_{i<j}^{N}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N)
\vert
\delta({\bf r}_i-{\bf r}_j-{\bf r})O_{ij}^p
\vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle.
\end{equation}
The matrix element of Eq. (87) is called the correlation function of type
$p$ evaluated between the Slater determinants. The calculation of this matrix
element can be done with the use of the basic
two-body matrix elements of Appendix B and the single-particle overlaps. We
will show, as an example, the case of $O_{ij}^p = 1$, i.e., Wigner-type
$\delta$-function two-body interaction $(p=c)$. Then we have
\begin{eqnarray}
& &\sum_{i<j}^{N}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N)
\vert
\delta({\bf r}_i-{\bf r}_j-{\bf r})
\vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle
\nonumber \\
&=&\sum_{i<j}^{N} \sum_{k,l=1}^{N}
\langle{\hat \varphi}_{{{\bf s}_i}{\sigma_i}{\tau_i}}^{\nu}
{\hat\varphi}_{{{\bf s}_j}{\sigma_j}{\tau_j}}^{\nu}
\vert
\delta({\bf r}_i-{\bf r}_j-{\bf r})
\vert
{\hat\varphi}_{{{{\bf s}'}_k}{\sigma'_k}{\tau'_k}}^{\nu}
{\hat\varphi}_{{{{\bf s}'}_l}{\sigma'_l}{\tau'_l}}^{\nu} \rangle
(-1)^{i+j+k+l} {\det} \lbrace B^{ijkl}\rbrace,
\end{eqnarray}
where $B^{ijkl}$ is obtained by omitting the $i$th and $j$th
rows and the $k$th
and $l$th columns of the matrix $B$. By substituting the explicit formula
of the ingredients, we arrive at the same form as given in Eq. (25):
\begin{eqnarray}
& &\sum_{i<j}^{N}
\langle \phi_{\kappa} ({\bf s}_1,...,{\bf s}_N)
\vert
\delta({\bf r}_i-{\bf r}_j-{\bf r})
\vert
\phi_{\kappa'}({{\bf s}'}_1,...,{{\bf s}'_N}) \rangle
\nonumber \\
&=&\bigl(\frac{\nu}{\pi}\bigr)^{3/2}{\rm e}^{-\nu r^2}
\sum_{P}^{N!}C_P\sum_{i<j}^{N}{\rm e}^{-{1\over 2} \tilde{\bf s}( A_{P}
+ B_{P}^{(ij)}) {\bf s}+ {\bf d}_{P}^{(ij)}\cdot {\bf r}},
\end{eqnarray}
with
\begin{equation}
{\bf d}_{P}^{(ij)}=\nu({\bf s}_i-{\bf s}_j + {\bf s}_{N+p_i}-{\bf s}_{N+p_j}).
\end{equation}
The coefficients $C_P$ and the matrices $A_P$ are the same as those that
appear in Eq. (81). The matrix $B_P^{(ij)}$ is defined by
\begin{eqnarray}
{(B_P^{(ij)})}_{kl}={(B_P^{(ij)})}_{lk}=\left\{
\begin{array}{ll}
\frac{\nu}{2} &\ \ \ (kl)=(ii), (jj), (N+p_i,N+p_i), (N+p_j,N+p_j),\\
& \ \ \ \ \ \ \ \ \ \ \ \ (i,N+p_i), (j,N+p_j). \\
-\frac{\nu}{2} & \ \ \ (kl)=(ij), (i,N+p_j), (j,N+p_i), (N+p_i,N+p_j). \\
0 & \ \ \ {\rm otherwise}. \\
\end{array}
\right.
\end{eqnarray}
\section*{Appendix D. Evaluation of the overlap of the correlated Gaussians}
In this appendix we derive the overlap of the basis functions,
defined in Eqs. (2) and (6), to illustrate the calculation of the
matrix elements through the operations prescribed in Eq. (57).
For the sake of simplicity we work out the case
of overlap only, but the matrix elements of the kinetic and the
potential energy are not much more involved either and these
matrix elements can be calculated by repeating the steps
detailed here.
\par\indent
In Eq. (57) the matrix element of an operator ${\cal O}$ between the
correlated Gaussians are derived from the matrix elements
between the Gaussian packets by applying suitably chosen operations.
In the case of ${\cal O}=1$, upon substituting
the matrix elements in the generator coordinate space (Eq. (41)), Eq. (57)
reads as
\begin{eqnarray}
& &\langle {\cal A} \{f_{KLM}({\bf u},{\bf x},A){\chi}_{SM_S}{\cal X}_{TM_T}\}
\vert {\cal A'} \{f_{K'L'M'}({\bf u}',{\bf x},A')
{\chi}_{S'M'_S}{\cal X}_{T'M'_T}\}\rangle
\nonumber \\
&\sim & \int\!\int d{\hat {\bf t}}d{\hat {\bf t}}'
Y_{LM}({\hat {\bf t}})^{*} Y_{L'M'}({\hat {\bf t}}') \\
&\times& \left( {d^{\kappa+\kappa'} \over d\alpha^{\kappa}d\alpha'^{\kappa'}}
{\rm e}^{-{1\over 2} \tilde{\bf T} C {\bf T}}
\,\int d{{\bf S}}
{\rm e}^{-{1\over 2} \tilde{{\bf S}} Q {{\bf S}}+\tilde{\bf T}{{\bf S}}}
\sum_{i=1}^{n_o}C_i^{(o)} {\rm e}^{-{1\over 2}\tilde{\bf S} B_i^{(o)}{\bf S}}
\right)_{\alpha=\alpha'=0 \atop t=t'=1},
\nonumber
\end{eqnarray}
with
\begin{equation}
\kappa=2K+L, \ \ \ \ \kappa'=2K'+L',
\end{equation}
where a constant factor is omitted for a moment.
By integrating over the generator coordinates ${\bf S}$ with the use of
Eq. (52), one readily obtains
\begin{eqnarray}
& &\langle {\cal A} \{f_{KLM}({\bf u},{\bf x},A){\chi}_{SM_S}{\cal X}_{TM_T}\}
\vert {\cal A'} \{f_{K'L'M'}({\bf u}',{\bf x},A')
{\chi}_{S'M'_S}{\cal X}_{T'M'_T}\}\rangle
\nonumber \\
&\sim & \int\!\int d{\hat {\bf t}}d{\hat {\bf t}}'
Y_{LM}({\hat {\bf t}})^{*} Y_{L'M'}({\hat {\bf t}}') \\
&\times& \left( {d^{\kappa+\kappa'} \over d\alpha^{\kappa}d\alpha'^{\kappa'}}
\,\sum_{i=1}^{n_o}C_i^{(o)}
\left({(2\pi)^{(2N-2)} \over {\rm det}(B_i^{(o)}+Q)}\right)^{3/2}
{\rm e}^{p_i\alpha^2+{{p_i}'}\alpha'^2+q_i\alpha\alpha' {\bf t}\cdot{\bf t'}}
\right)_{\alpha=\alpha'=0 \atop t=t'=1},\nonumber
\end{eqnarray}
where, by using Eqs. (54) and (55), we introduced the abbreviations
\begin{eqnarray}
& &p_i={1\over 2}\sum_{j,k=1}^{N-1} w_j ((B_i^{(o)}+Q)^{-1}-C)_{jk} w_k,
\ \ \ {p_i}'={1\over 2}\sum_{j,k=N}^{2N-2} {w_j}'
((B_i^{(o)}+Q)^{-1}-C)_{jk} {w_k}' ,\nonumber \\
& &q_i=\sum_{j=1}^{N-1} \sum_{k=N}^{2N-2} w_j ((B_i^{(o)}+Q)^{-1}-C)_{jk}
{w_k}' ,
\end{eqnarray}
with
\begin{equation}
w_k=\sum_{j=1}^{N-1}(\Gamma C)^{-1}_{kj}u_j, \ \ \ \
{w_k}'=\sum_{j=N}^{2N-2}(\Gamma C')^{-1}_{kj}{u_j}' ,
\end{equation}
in order to emphasize the $\alpha$- and $\alpha'$-dependence of the
resulting expression.
The differentiation with respect to $\alpha$ and $\alpha'$ can be promptly
given by expanding the exponential functions into power series:
\begin{eqnarray}
& &{d^{\kappa+\kappa'} \over d\alpha^{\kappa}d{\alpha'}^{\kappa'}}
{\rm e}^{p_i\alpha^2+{{p_i}'}\alpha'^2+q_i\alpha\alpha'{\bf t}\cdot{\bf t'}}
\nonumber \\
& = & \sum_{j,j',k}^{\infty}
{ p_i^j{{p_i}'}^{j'}({q_i}{\bf t}\cdot{{\bf t}'})^k \over j!{j'}!k!}
{(2j+k)!\over (2j+k-\kappa)!}
{(2j'+k)!\over (2j'+k-\kappa')!}
\alpha^{2j+k-\kappa} {\alpha'}^{2j'+k-\kappa'},
\end{eqnarray}
and therefore, after putting $\alpha=\alpha'=0$, the calculation of
Eq. (94) can be continued as
\begin{eqnarray}
& &\langle {\cal A} \{f_{KLM}({\bf u},{\bf x},A){\chi}_{SM_S}{\cal X}_{TM_T}\}
\vert {\cal A'} \{f_{K'L'M'}({\bf u}',{\bf x},A')
{\chi}_{S'M'_S}{\cal X}_{T'M'_T}\}\rangle
\nonumber \\
&\sim & \int\!\int d{\hat {\bf t}}d{\hat {\bf t}}'
Y_{LM}({\hat {\bf t}})^{*} Y_{L'M'}({\hat {\bf t}}') \\
&\times& \left(\sum_{i=1}^{n_o}C_i^{(o)}
\left({(2\pi)^{2N-2} \over {\rm det}(B_i^{(o)}+Q)}\right)^{3/2}
\kappa!\kappa'! \sum_{k}
{{p_i^{\kappa-k\over 2}{{p_i}'}^{\kappa'-k\over 2}}{q_i}^k \over
({\kappa-k \over 2})!
({\kappa'-k\over 2})! k!}
\left({\bf t}\cdot{\bf t}'\right)^{k}_{t=t'=1}\right), \nonumber
\end{eqnarray}
where the summation over $k$ runs from $0$ to max$(\kappa,\kappa')$
for those values that fulfill $(-1)^{k+\kappa}=1$ and $(-1)^{k+\kappa'}=1$.
\par\indent
The last step, the integration over the angles of ${\bf t}$ and ${\bf t}'$,
can be accomplished by applying Eq. (49) to express the scalar product
${\bf t}\cdot {\bf t}'$, and then we get the final expression:
\begin{eqnarray}
& &\langle {\cal A} \{f_{KLM}({\bf u},{\bf x},A){\chi}_{SM_S}{\cal X}_{TM_T}\}
\vert {\cal A'} \{f_{K'L'M'}({\bf u}',{\bf x},A')
{\chi}_{S'M'_S}{\cal X}_{T'M'_T}\}\rangle
\nonumber \\
&=&\delta_{LL'}\delta_{MM'}
\frac{1}{B_{KL}B'_{K'L'}}\left(\frac
{({\rm det}\Gamma)^3}
{(4\pi)^{(N-1)}{\rm det}(
\Gamma-A){\rm det}(\Gamma-A')}
\right)^{3/2} \\
&\times& \left(\sum_{i=1}^{n_o}C_i^{(o)}
\left({(2\pi)^{2N-2} \over {\rm det}(B_i^{(o)}+Q)}\right)^{3/2}
\kappa!\kappa'!
\sum_{k}
{{p_i^{\kappa-k\over 2}{{p_i}'}^{\kappa'-k\over 2}}{q_i}^k \over
({\kappa-k \over 2})! ({\kappa'-k\over 2})! k!} B_{kL} \right), \nonumber
\end{eqnarray}
where the constant term previously suppressed is also included. The
orthogonality of the matrix element in the spin and isospin quantum numbers
is implicitly contained in the coefficients $C_i^{(o)}$.
\par\indent
The calculation of the matrix elements in the way described above is very
simple. This fact becomes especially evident if one compares the work
described above
to the formidable task of the calculation of the matrix elements
in the case where the function $\theta_{LM_L}({\bf x})$
is decomposed into partial waves of the relative coordinates as in Eq. (5).
In fact, in that latter
case one has to integrate over the angles of the relative
coordinates and one has to cope with complicated angular momentum algebra.
We note, however, that the calculation of the matrix element of the latter
type poses no problem if the function $\theta_{LM_L}({\bf x})$ of Eq. (5)
is expressed as a linear combination of the terms of Eq. (6) with
appropriate ${\bf u}$-vectors.
\par\indent
All the matrix elements can be given in a similar closed analytic
form and the numerical evaluation of the matrix elements as a function
of the nonlinear parameters is therefore straightforward.
The values $\kappa=2K\!\!+\!\!L$ and $\kappa'=2K'\!\!+\!\!L'$ are usually
small in practical cases and the summation over $k$ is limited to just a
few terms in Eq. (99).
\bigskip
\par\indent
\par\indent
This work was supported by OTKA grant No. T17298 (Hungary) and
by a Grant-in-Aid for Scientific Research (No. 05243102 and No. 06640381)
of the Ministry of Education, Science and Culture (Japan). K. V. gratefully
acknowledges the hospitality of the RIKEN LINAC Laboratory and the
support of the Science and Technology Agency of Japan, 1994. The authors
are grateful for the use of RIKEN's VPP500 computer which
made possible most of the calculations.
Thanks are also due to the support of both Japan Society for the Promotion
of Science and Hungarian Academy of Sciences, 1994-1995. The authors thank
Prof. R. G. Lovas for careful reading of the manuscript and useful
suggestions.
|
\section{Introduction}
The analysis of the octet baryon masses in the framework of chiral
perturbation theory already has a long history, see e.g.
\cite{lang,pagels,juerg,gl82,liz,dobo,bkmz,lelu,samir,mary,dent,buergi}.
In this paper, we present
the results of a first calculation including {\it all} terms of order
${\cal O}(m_q^2)$, where $m_q$ is a generic symbol for any one of the light
quark masses $m_{u,d,s}$. We work in the isospin limit $m_u = m_d$ and
neglect the electromagnetic corrections. Previous investigations only
considered mostly the so--called computable corrections of order
$m_q^2$ or included some of the finite terms at this order.
This, however, contradicts the spirit of chiral perturbation
theory (CHPT) in that all terms at a given order have to be retained, see
e.g. \cite{wein79,gl84,gl85}. In general, the quark mass expansion of
the octet baryon masses takes the form
\begin{equation}
m = \, \, \krig{m} + \sum_q \, B_q \, m_q + \sum_q \, C_q \, m_q^{3/2} +
\sum_q \, D_q \, m_q^2 + \ldots
\label{massform}
\end{equation}
modulo logs. Here, $\krig{m}$ is the mass in the chiral limit of
vanishing quark masses and the coefficients $B_q, C_q, D_q$ are
state--dependent. Furthermore, they include contributions proprotional
to some low--energy constants which appear beyond leading order in the
effective Lagrangian. In this letter, we evaluate these
coefficients for the octet baryons $N, \Lambda, \Sigma$ and $\Xi$.
In addition, we also calculate the pion--nucleon $\sigma$--term which
is intimately related to the quark mass expansion of the baryon masses
\cite{juerg,bkmz,jms}. For some comprehensive reviews, see e.g.
\cite{ulfrev,ecker,bkmrev,toni}.
\section{Effective Lagrangian}
To perform the calculations, we make use of the effective
meson--baryon Lagrangian. Our notation is identical to the one
used in \cite{bkmz} and we discuss here only the new terms.
Denoting by $\phi$ the pseudoscalar Goldstone fields ($\pi, K, \eta$) and
by $B$ the baryon octet, the effective Lagrangian takes the form
\begin{equation}
{\cal L}_{\rm eff} = {\cal L}_{\phi B}^{(1)} + {\cal L}_{\phi B}^{(2)} +
{\cal L}_{\phi B}^{(4)} + {\cal L}_{\phi}^{(2)}+ {\cal L}_{\phi}^{(4)}
\label{leff}
\end{equation}
where the chiral dimension $(i)$ counts the number of derivatives
and/or meson mass insertions. The baryons are treated in the extreme
non--relativistic limit \cite{jm,bkkm}, i.e. they are characterized
by a four--velocity $v_\mu$. In this approach, there is a one--to--one
correspondence between the expansion in small momenta and quark masses
and the expansion in Goldstone boson loops, i.e. a consistent power
counting scheme emerges. The form of the lowest order meson--baryon
Lagrangian is standard, see e.g. \cite{bkmz}, and the
meson Lagrangian is given in \cite{gl85}.
The dimension two meson--baryon Lagrangian can be written as
(we only enumerate the terms which contribute)
\begin{equation}
{\cal L}_{\phi B}^{(2)} = {\cal L}_{\phi B}^{(2, {\rm stand})} +
\sum_{i=1}^{10} \, b_i \, O_i^{(2)} \, \, ,
\label{leff2}
\end{equation}
with the $O_i^{(2)}$ monomials in the fields of chiral dimension two. Typical
forms are ${\rm Tr}(\bar{B} [ u_\mu , [ u^\mu,B]])$,
${\rm Tr}(\bar{B} [ v \cdot u , [ v \cdot u,B]])$
or $\bar{B} B {\rm Tr} (u_\mu u^\mu ) $, with $u_\mu = i
u^\dagger \partial_\mu U u^\dagger$, $u = \sqrt{U}$ and $U=\exp(i \phi
/ F_P)$ collects the pseudoscalars. The form
of ${\cal L}_{\phi B}^{(2, {\rm stand})}$ is \cite{bkmz},
\begin{equation}
{\cal L}_{\phi B}^{(2, {\rm stand})} = b_D \, {\rm Tr}(\bar B \lbrace
\chi_+ , B \rbrace ) + b_F \, {\rm Tr}(\bar B [\chi_+,B]) +
b_0 \, {\rm Tr}(\bar B
B) \, {\rm Tr}(\chi_+ ) \, \, ,
\label{leff2st}
\end{equation}
i.e. it contains three low--energy constants and $\chi_+ = u^\dagger
\chi u^\dagger + u \chi^\dagger u$ is proportional to the quark mass
matrix ${\cal M} ={\rm diag}(m_u,m_d,m_s)$ since $\chi = 2 B {\cal
M}$. Here, $B = - \langle 0 | \bar{q} q | 0 \rangle / F_P^2$ and $F_P$ is
the pseudoscalar decay constant.
All low--energy constants in ${\cal L}_{\phi B}^{(2)}$ are finite.
A subset of the $b_i$ has been estimated in \cite{norb} by
analyzing kaon--nucleon scattering data.
The splitting of the dimension two meson--baryon Lagrangian in
Eq.(\ref{leff2}) is motivated by the fact that while the first three
terms appear in tree and loop graphs, the latter ten only come in via
loops. There are seven terms contributing at dimension four,
\begin{equation}
{\cal L}_{\phi B}^{(4)} =
\sum_{i=1}^{7} \, d_i \, O_i^{(4)} \, \, ,
\label{leff4}
\end{equation}
with typical forms of the $O_i^{(4)}$ are $\bar{B}B {\rm Tr}
(\chi_+^2)$ or ${\rm Tr}(\bar{B}[\chi_+ , [\chi_+ , B]] )$.
At this stage,
we take $m_u = m_d \ne m_s$. For $m_u \ne m_d$, there is an additional
term in ${\cal L}_{\phi B}^{(4)}$. The
explicit expressions for the various terms in
Eqs.(\ref{leff2},\ref{leff4}) can be found in \cite{bora}.
We point out that there are 20 a priori unkown constants. In addition,
there are the $F$ and $D$ coupling constants (subject to the
constraint $F+D = g_A =1.25$)
from the lowest order Lagrangian
${\cal L}_{\phi B}^{(1)}$.
What we have to calculate are all one--loop
graphs with insertions from ${\cal L}_{\phi B}^{(1,2)}$ and tree
graphs from ${\cal L}_{\phi B}^{(2,4)}$. We stress that we do not
include the spin--3/2 decuplet in the effective field theory
\cite{dobo}, but rather use these fields to estimate the pertinent
low--energy constants (resonance saturation principle). We therefore
strictly count in small quark masses and external momenta with no
recourse to large $N_c$ arguments.
\section{Baryon masses and pion--nucleon $\sigma$--term}
The form of the terms $\sim m_q $ and $\sim m_q^{3/2}$
for the baryon masses and $\sigma_{\pi N} (0)$ is standard, we
use here the same notation as Ref.\cite{bkmz}. The $q^4$
contribution to any octet baryon mass $m_B$ takes the form
\begin{eqnarray}
m_B^{(4)} & = & \epsilon_{1,B}^P \, M_P^4 +
\epsilon_{2,B}^{PQ} \, M_P^2 \, M_Q^2 \nonumber \\
\quad & \, + &
\epsilon_{3,B}^P \, M_P^4 \, \ln ( \frac{M_P^2}{\lambda^2}) +
\epsilon_{4,B}^{PQ} \, M_P^2 \, M_Q^2 \, \ln ( \frac{
M_P^2}{\lambda^2} )
\, \, ,
\label{mB4}
\end{eqnarray}
with $P,Q = \pi , K, \eta$ and $\lambda$ the scale of dimensional
regularization. The explicit form of the state--dependent
prefactors $\epsilon_{i,B}$ can be found in Ref.\cite{bora}.
Notice the appearance of mixed terms $\sim M_P^2 \, M_Q^2$
which were not considered in most existing investigations. The fourth
order contribution to the baryon masses contains divergences
proportional to (using dimensional regularization)
\begin{equation}
L = \frac{\lambda^{d-4}}{16 \pi^2} \biggl\lbrace \frac{1}{d-4} -
\frac{1}{2}[\ln (4 \pi) +1 - \gamma_E] \biggr\rbrace
\, \, ,
\label{L}
\end{equation}
with $\gamma_E = 0.572215$.
These are renormalized by an appropriate choice of the low--energy
constants $d_i$,
\begin{equation}
d_i = d_i^r (\lambda) + \Gamma_i \, L
\, \, .
\label{divd}
\end{equation}
In fact, all seven $d_i$ are divergent. The appearance
of these divergences is in marked contrast to the $q^3$ calculation
which is completely finite (in the heavy fermion approach). In what
follows, we set $\lambda =1$~GeV.
Similarly, the fourth order contribution to the pion--nucleon $\sigma$--term
can be written as
\begin{equation}
\sigma_{\pi N}^{(4)}(0) =
M_\pi^2 \, \biggl[ \, \epsilon_{1,\sigma}^P \, M_P^2 +
\epsilon_{2,\sigma}^P \, M_P^2 \, \ln(\frac{M_P^2}{\lambda^2}) +
\epsilon_{3,\sigma}^{PQ} \, M_P^2 \, \ln(\frac{M_Q^2}{\lambda^2}) \, \biggr]
\, \, ,
\label{sigma4}
\end{equation}
with the $\epsilon_{i,\sigma}$ given in \cite{bora}.
Here, the renormalization is somewhat more tricky. It can most
efficiently performed in a basis of a linearly independent subset
of the operators $dO_i^{(4)}/ dm_q$, $q=(u,d,s)$, as detailed in
Ref.\cite{bora}. A good check on the rather lengthy expressions for
the nucleon mass and $\sigma_{\pi N} (0)$ is given by the
Feynman--Hellmann theorem, $\hat{m}(\partial m_N / \partial
\hat{m}) = \sigma_{\pi N} (0)$, with $\hat m$ the average light quark
mass, $\hat m = (m_u + m_d)/2$.
We remark here that in contrast to the order $q^3$
calculation, the shift to the Cheng--Dashen point, $\sigma_{\pi N}
(2M_\pi^2) - \sigma_{\pi N} (0)$, is no longer finite, i.e. there
appear $t$--dependent divergences. We therefore do not consider this
$\sigma$--term shift in what follows. We will also not discuss in
detail the two kaon--nucleon $\sigma$--terms,
$\sigma_{KN}^{(1,2)} (t)$, for similar reasons in this letter.
\section{Resonance saturation}
Clearly, we are not able to fix all the low--energy
constants appearing in ${\cal L}_{\phi B}^{(2,4)}$
from data, even if we would resort to large $N_c$ arguments.
We will therefore use the principle of resonance saturation to
estimate these constants. This works very accurately in the meson
sector \cite{reso,reso1,reso2}. In the baryon case, one has to account for
excitations of meson ($R$) and baryon ($N^*$) resonances. One writes
down the effective Lagrangian with these resonances chirally coupled to
the Goldstones and the baryon octet, calculates the pertinent Feynman
diagrams to the process under consideration and, finally,
lets the resonance masses go to infinity (with fixed ratios of
coupling constants to masses). This generates higher order terms in
the effective meson--baryon Lagrangian with coefficients expressed in
terms of a few known resonance parameters. Symbolically, we can write
\begin{equation}
\tilde{{\cal L}}_{\rm eff} [\, U,B,R,N^* \, ] \to
{\cal L}_{\rm eff} [\, U,B \, ] \, \, .
\end{equation}
Here, there are two relevant contributions. One comes from the excitation
of the spin-3/2 decuplet states and the second from t--channel scalar
and vector meson excitations, cf. Fig.\ref{fig1}. It is important to
stress that for the resonance contribution to the baryon masses, one
has to involve Goldstone boson loops. This is different from the
normal situation like e.g. in form factors or scattering processes.
\begin{figure}[t]
\hskip 1.2in
\epsfysize=1.5in
\epsffile{ma1.ps}
\caption{\label{fig1} Resonance saturation for the masses. In (a), a
baryon resonance (from the decuplet) is excited, whereas (b) represents
the t--channel meson (scalar or vector) excitation. Solid and dashed
lines denote the octet baryons and Goldstones, respectively.}
\end{figure}
Consider first the decuplet contribution. We treat these field
relativistically and only at the last stage let the mass become very
large. The pertinent interaction Lagrangian between the spin--3/2
fields (denoted by $\Delta$), the baryon octet and the Goldstones reads
(suppressing SU(3) indices)
\begin{equation}
{\cal L}_{\Delta B \phi} =
\frac{{\cal C}}{\sqrt{2} F_P} \, \bar{\Delta}^\nu \, T
\, \Theta_{\nu \mu} (Z) \, B \, \partial^\mu \phi + {\rm h.c.} \, \, ,
\label{lmbd}
\end{equation}
with $T$ the standard $\frac{1}{2} \to \frac{3}{2}$ isospin transition
operator, $F_P = 100$~MeV the (average) pseudeoscalar decay constant
and ${\cal C} =1.5$ determined from the decays $\Delta \to B \pi$. The Dirac
matrix operator $\Theta_{\mu \nu} (Z)$ is given by
\begin{equation}
\Theta_{\mu \nu} (Z) = g_{\mu \nu} - \biggl(Z + \frac{1}{2} \biggr) \,
\gamma_\mu \, \gamma_\nu \, \, \, \, .
\label{theta}
\end{equation}
For the off--shell parameter $Z$, we use $Z = -0.3$ from the
determination of the $\Delta$ contribution to the $\pi N$ scattering
volume $a_{33}$ \cite{armin}.
The tadpole graphs shown in Fig.~1b with an intermediate vector meson
only contribute at order $q^5$. This is evident in the conventional
vector field formulation. In the tensor field formulation, the vertex
${\rm Tr}(V_{\mu \nu}u^\mu u^\nu)$ seems to lead to a contribution at
order $q^4$. However, in a tadpole graph one needs the index
contraction $\mu =\nu$ and thus has no term at order $q^4$. Matters
are different for the scalars. Denoting by $S$ and $S_1$ the scalar
octet and singlet with $M_S \simeq M_{S_1} \simeq 1$~GeV, respectively,
the lowest order coupling to the baryon octet reads
\begin{equation}
{\cal L}_{SB}^{(0)} = D_S \, {\rm Tr}(\bar B \lbrace S,B \rbrace )
+ F_S \, {\rm Tr}(\bar B [ S,B ] ) + D_{S_1} \,S_1 \, {\rm Tr}(\bar B B )
\label{sb}
\end{equation}
where the coupling constants $D_S$, $F_S$ and $D_{S_1}$ are of the
order one.\footnote{In what follows, we neglect the singlet field.} In
fact, there are no empirical data to really pin them down. From the
decay pattern of the low--lying baryons one can, however, estimate
these numbers to be small. We will generously vary them between zero
and one. For the couplings of the scalars to the Goldstones, we use the
notation of \cite{reso} and the parameters determined therein.
Putting pieces together, all low--energy constants are expressed via
resonance parameters and the baryon masses take the form
\begin{eqnarray}
m_B & = & \krig{m} + m_B^{(3)}
+ \lambda_B \, ({\krig{m}})^{-1}
+ \beta \, D_B^\beta
+ \delta \, D_B^\delta
+ \epsilon \, D_B^\epsilon + D_B^S
\quad , \nonumber \\
\beta & = & -\frac{1}{96 \pi^2} \, \frac{m_\Delta^4}{{\krig{m}}^3} \, , \quad
\delta= -\frac{\beta}{4{\krig{m}}^2} \, , \quad
\epsilon = \frac{2}{3m_\Delta} (2Z^2 + Z -1 )
\, \, \, ,
\label{massres}
\end{eqnarray}
with $ m_B^{(3)}$ the contribution of ${\cal O}(m_q^{3/2})$ and
$m_\Delta = 1.455$~GeV is the average decuplet mass.
Notice that it is convenient to lump the ${\cal O}(m_q)$ and
the ${\cal O}(m_q^2)$ corrections togther as it was done
in Eq.(\ref{massres}) (in the $D_B^\beta$ and $D_B^S$).
We have kept explict the baryon mass in the chiral limit. Of course,
in the fourth order terms it could be substituted by the corresponding
physical values. At second order, however, we would get a
state--dependent shift, see e.g. \cite{bkmzas}, and we thus prefer to
work with $\krig m$. Alternative representations
for the baryon masses are given in \cite{bora}.
Let us briefly explain the origin of the
various terms in Eq.(\ref{massres}). The $\lambda$ contributions are
tadpoles with $1/m$ insertions from ${\cal L}_{\phi B}^{(2)}$. Similarly,
the $\beta$ and $\epsilon$ terms stem from tadpole graphs with
insertions proportional to the low--energy constants $b_{0,D,F}$ and
$b_i$, respectively. Note, however, that in the resonance exchange
approximation not all of the ten $b_i$ are contributing. The
$\delta$ terms subsume the contributions from ${\cal L}_{\phi B}^{(4)}$,
these are proportional to the low--energy constants $d_i$.
Finally, the terms of the type $D_B^S$ are the scalar meson
contributions to the mass. They amount to a constant, state--dependent shift.
These consist of terms of the types $\sim M_P^2$, $\sim M_P^4 \, \ln
M_P^4$ and $\sim M_P^2 M_Q^2 \, \ln M_P^4$, see \cite{bora}.
To be specific,
we give the coefficients $D^{\beta,\delta,\epsilon}$ and $\lambda$
for the nucleon,
\begin{eqnarray}
& \lambda_N & =
\frac{1}{8 \pi^2 F^2_P} \biggl\lbrace
-\frac{3}{32} (D+F)^2 M_\pi^4 \ln \frac{M_\pi^2}{\lambda^2}
- \frac{1}{96}(D - 3F)^2 M_\eta^4 \ln \frac{M_\eta^2}{\lambda^2}
\nonumber \\
& \, &
- \frac{1}{16}(D^2 - 2DF +3F^2) M_K^4 \ln \frac{M_K^2}{\lambda^2}
\biggr\rbrace \, \, , \nonumber \\
& D_N^\beta & = D_N^{\beta,(2)} + D_N^{\beta,(4)} =
\frac{{\cal C}^2 }{16 \pi^2 F^4_P} \biggl\lbrace
(-4\pi^2 F_P^2)(M_K^2 + 4M_\pi^2) \nonumber \\
& + & \frac{27}{16} M_\pi^4 \ln \frac{M_\pi^2}{\lambda^2}
+\frac{33}{24} M_K^4 \ln \frac{M_K^2}{\lambda^2}
+ \frac{7}{12} M_\eta^4 \ln \frac{M_\eta^2}{\lambda^2} \nonumber \\
& + & \biggl[ -\frac{11}{18}M_K^2 + \frac{43}{144}M_\pi^2
+\frac{1}{8}(D - 3F)^2 \bigl( \frac{M_K^2}{4}+M_\pi^2 \bigr) \biggr]
M_\eta^2 \ln \frac{M_\eta^2}{\lambda^2} \nonumber \\
& + & \biggl[
\frac{9}{8}(D + F)^2 \bigl( \frac{M_K^2}{4} +M_\pi^2 \bigr) \biggr]
M_\pi^2 \ln \frac{M_\pi^2}{\lambda^2} \nonumber \\
& + & \biggl[
-\frac{3}{4}(D - F)^2 M_K^2 + \frac{1}{96} (223D^2 - 318DF +351F^2) M_\pi^2
\biggr] M_K^2 \ln \frac{M_K^2}{\lambda^2}
\biggr\rbrace \, \, , \nonumber \\
& D_N^\delta & = \frac{{\cal C}^2}{2F_P^2} \biggl\lbrace
-\frac{415}{288} M_\pi^4 + \frac{83}{72} M_\pi^2 M_K^2 -
\frac{31}{72} M_K^4 \biggr\rbrace \, \, , \nonumber \\
& D_N^\epsilon & = \frac{{\cal C}^2}{8 \pi^2 F_P^2} \biggl\lbrace
- \frac{1}{2} M_\pi^4 \ln \frac{M_\pi^2}{\lambda^2}
- \frac{1}{8} M_K^4 \ln \frac{M_K^2}{\lambda^2} \biggr\rbrace
\, \, .
\label{DN}
\end{eqnarray}
The corresponding coefficients for the $\Lambda$, $\Sigma$ and $\Xi$
and also the $D_B^S$
can be found in \cite{bora}. The tree contribution from
${\cal L}_{\phi B}^{(2, {\rm stand})}$ is subsumed in the $D_B^\beta$.
The numerical values of the $\lambda_B$, $D^\beta_B, D^\delta_B$,
$D^\epsilon_B$ and $D_B^S$ are given in table~1
(using $F = 0.5$, $D=0.75$ and $c_{d,m}$ from \cite{reso}).
\begin{table}[h]
\begin{tabular}{|c|ccccc|}
\hline
B & $\lambda_B$ [GeV$^{2}$]
& $D^\beta_B$ & $D^\delta_B$ [GeV$^{2}$]
& $D^\epsilon_B$ [GeV$^{2}$] & $D_B^S$ [GeV] \\
\hline
$N$ & 0.0049 & $-$39.875 & $-$2.3617 & 0.0321
& $-0.017 \, D_S + 0.061 \, F_S$ \\
$\Lambda$ & 0.0179 & $-$68.044 & 1.0831 & 0.0617
& $-0.047 \, D_S - 0.010 \, F_S$ \\
$\Sigma $ & 0.0144 & $-$132.90 & $-$11.273 & 0.1422
& $+0.051 \, D_S + 0.010 \, F_S$ \\
$\Xi$ & 0.0216 & $-$126.46 & $-$5.4799 & 0.1324
& $-0.037 \, D_S - 0.082 \, F_S$ \\
\hline
\end{tabular}
\medskip
\caption{Numerical values of the state--dependent coefficients
in Eq.(\protect\ref{massres}). The $D_B^\beta$ are
dimensionless. The $D_B^S$ are for $c_d$, $c_m$ from
\protect\cite{reso} and $M_S =1$ GeV.}
\end{table}
We see that the dominant terms at ${\cal O}(m_q^2)$ are indeed the
tadpole graphs with an insertion from ${\cal L}_{\phi B}^{(2, {\rm
stand})}$ (this holds for the masses but not for $\sigma_{\pi N}(0)$).
It is also instructive to compare the values we find from resonance
exchange with the ones previously determined from KN scattering
data. We have transformed the results of
Ref.\cite{norb} into our notation. As can be seen from
table~2, most (but not all) coefficients agree in sign and magnitude.
\begin{table}[bht]
\begin{tabular}{|c|ccccccc|}
\hline
& $b_1$ & $b_2$ & $b_3$ & $b_8$ & $b_0$ & $b_F$ & $b_D$ \\
\hline
Reso. $(D_S=F_S=0)$ & 0.084 & $-$0.144 & 0.108 & $-$0.216 & $-$0.738
& $-$0.264 & 0.317 \\
Reso. $(D_S=F_S=1)$ & 0.100 & $-$0.111 & 0.125 & $-$0.239 & $-$0.733
& $-$0.208 & 0.345 \\
Ref.\cite{norb} & 0.044 & $-$0.145 & $-$0.054 & $-$0.165 & $-$0.493
& $-$0.213 & 0.066 \\
\hline
\end{tabular}
\medskip
\caption{Some low--energy constants from ${\cal L}_{\phi B}^{(2)}$
in GeV$^{-1}$. In the first row, only the decuplet
contribution is given. In the second row, scalar meson
exchange is added.
\end{table}
Note, however, that this is only a subset of the coefficients
considered in this work. The full list will be given in \cite{bora}.
We remark that the procedure used in \cite{norb} involves the summation of
arbitrary high orders via a Lippmann--Schwinger
equation and is thus afflicted with
some uncertainty not controled within CHPT.
\section{Results and discussion}
The only free parameter in the formula for the baryon masses,
Eq.(\ref{massres}), is the mass of the baryons in the chiral limit,
$\krig m$, since all low--energy constants are fixed in terms of
resonance parameters. In particular, in contrast to Ref.\cite{bkmz},
the parameters $b_0$, $b_D$ and $b_F$ are no longer free. Also, at
quadratic order in the quark masses the ambiguity between $\krig m$
and $b_0$ is resolved, it is not necessary to involve any one
of the $\sigma$--terms in the fitting procedure \cite{liz,bkmz}.
In fact, one can not find one single value
of $\krig m$ to fit all four octet masses, $m_N = 0.9389$~GeV, $m_\Lambda =
1.1156$~GeV, $m_\Sigma = 1.1931$~GeV and $m_\Xi = 1.3181$~GeV exactly. We
therefore fit these masses individually and average the corresponding
values for $\krig m$.
The contribution from scalar meson exchange only enters the
uncertainties of the numbers given. This is justified since the
numerical values for the couplings $F_S$ and $D_S$ are supposedly small.
A more thorough discussion on this point can be found in \cite{bora}.
We have
$\krig{m}_N = 711 \, \, {\rm MeV}$,
$\krig{m}_\Lambda = 679 \, \, {\rm MeV}$,
$\krig{m}_\Sigma = 877 \, \, {\rm MeV}$,
$\krig{m}_\Xi = 728 \, \, {\rm MeV}$,
with the following average
\begin{equation}
\krig{m} = 749 \pm 125 \, \, {\rm MeV}\,\, . \quad
\label{mkrig}
\end{equation}
This number is compatible with the one found in the analysis
of the pion--nucleon $\sigma$--term, where approximately 130 MeV to the
nucleon mass were attributed to the strange matrix element $m_s
\langle p | \bar s s|p \rangle$ (with a sizeable uncertainty) \cite{gls}.
The spread of the various values is a good measure of the
uncertainties related to this complete $q^4$ calculation.
Let us take a closer look at the quark mass expansion of the
nucleon mass, in the notation of Eq.(\ref{massform}),
\begin{equation}
m_N = 711 + 202 - 272 + 298 \, \, \, {\rm MeV}
= 939 \, \, {\rm MeV} \, \, .
\label{mnuc}
\end{equation}
This looks similar for the other octet baryons.
We conclude that the quark mass corrections of order $m_q$,
$m_q^{3/2}$ and $m_q^2$ are all of the same size.\footnote{Note that
for certain values of the scalar couplings, the fourth order term
$m_N^{(4)}$ can be significantly smaller. However, the nucleon mass
is much more sensitive to the scalar contribution than the other
octet masses. }
In Ref.\cite{jms},
it was argued that only the leading non--analytic corrections (LNAC)
$\sim m_q^{3/2}$ are large and that further terms like the ones $\sim m_q^2$
are modestly small, of the order of 100 MeV. This would amount to
an expansion in $M_K^2/(4 \pi F_P)^2 \sim 1/6$
with a large leading term. This expectation is not
borne out by our results, the next corrections are as large as the
LNACs. These findings agree with the meson cloud model calculation of
Gasser \cite{juerg}. A last remark about the baryon masses
concerns the deviation from the Gell-Mann-Okubo relation, $\Delta_{\rm
GMO} = (3m_\Lambda+m_\Sigma-2m_N-2m_\Xi)/4$ which empirically is
about 6.5 MeV. We find $\Delta_{\rm GMO} = 31$ MeV, which is larger
in magnitude than the value found in \cite{liz,dobo}. We remark
that in our case the decuplet contribution is contained in
the $m_q^2$ contributions and not in the $m_q^{3/2}$ as in \cite{liz}.
Therefore, in our case, $\Delta_{\rm GMO}$ is dominated by the
$m_q^2$ piece. The sizeable uncertainty in the chiral limit masses,
Eq.(\ref{mkrig}), does not allow for a very accurate statement about
this very small quantity. It is also very sensitive to the scalar couplings.
To get a better handle on this issue, one either has to be able to fix all
pertinent low--energy constants at order $m_q^2$ from data or improve upon the
resonance saturation estimate used here by including e.g. the mass splitting
within the decuplet and the SU(3) breaking of the
decuplet-octet-meson couplings. A better understanding of this topic
is, of course, at the heart of the determination of the quark mass
ratio $R = (m_d-m_u)/(m_s - \hat m )$ from the baryon masses (once the
electromagnetic corrections have been included).
The pion--nucleon $\sigma$--term is completely fixed. Using for
$\krig m$ its average, we find (no scalar resonance contribution)
\begin{equation}
\sigma_{\pi N} (0) = 48 \, \, {\rm MeV} \, \, ,
\label{sigval}
\end{equation}
with an uncertainty of about $\pm 10$ MeV due to the spread in $\krig
m$. This number compares favourably with the one found
in Ref.\cite{gls}, $\sigma_{\pi N} (0) = 44 \pm 9$~MeV. We stress that
this result reflects a very non--trivial consistency
for the complete calculation to quadratic order in the quark masses
using the
resonance saturation principle in the scalar sector.
The additional contribution from the scalar meson exchange is
accounted for in the $\pm 10$ MeV uncertainty. To be specific, we have
$\sigma_{\pi N} (0) = (48 + D_S \cdot 0.5 - F_S \cdot 2.0) \, \,
{\rm MeV}$. It is furthermore
instructive to disentangle the various contributions to $\sigma_{\pi
N} (0)$ of order $q^2$, $q^3$ and $q^4$, respectively,
\begin{equation}
\sigma_{\pi N} (0) = 54 - 33 + 27 \, \, \, {\rm MeV}
= 48 \, \, {\rm MeV} \, \, ,
\label{signo}
\end{equation}
which shows a moderate convergence, i.e. the terms of increasing order
become successively smaller. Still, the $q^4$ contribution is
important. Also, at this order no $\pi \pi$ rescattering effects are
included. We notice that using the values for $b_{0,D,F}$ and $b_i$
as determined in Ref.\cite{norb} leads to a much increased fourth
order contribution.
\section{Summary and outlook}
In this paper, we have used heavy baryon chiral perturbation theory to
calculate the octet baryon masses to quadratic order in the quark
masses, including 20 local operators with unknown coeffcients. These
low--energy constants were fixed by resonance exchange.
The dominant contribution comes indeed from the excitation of the
spin--3/2 decuplet fields. Tadpole graphs with scalar meson exchange
only lead to small corrections. This left us
with one free parameter, the baryon mass in the chiral limit, which
could be determined within 18\% accuracy, and is compatible with the
value inferred for the strange matrix--element $m_s \langle p | \bar
s s | p \rangle$ in Ref.\cite{gls}.
Furthermore, the pion--nucleon
$\sigma$--term comes out surprisingly close to its empirical
value, $\sigma_{\pi N} (0) = 48$~MeV, with an uncertainty of
about $\pm 10$ MeV.
This first exploratory $q^4$ study of the three flavor scalar sector
of baryon CHPT points towards a significant improvement compared to
previous investigations which were mostly confined to so--called
``computable'' corrections and/or fitted a few of the pertinent
low--energy constants.
However, the calculation is not yet accurate enough to
determine the quark mass ratio $R$ reliably from the octet masses.
Furthermore, we did not address the kaon--nucleon $\sigma$--terms,
the corresponding shifts to the pertinent Cheng-Dashen points
together with the strangeness content of the proton here.
We will come back to these topics in Ref.\cite{bora}.
\section*{Acknowledgements}
We are grateful to Daniel Wyler for a very useful remark.
|
\section{Introduction}
Self-avoiding random walk models appeared in chemical physics
as models for long polymer chains. Roughly speaking, a polymer is
composed of a large number of monomers which are linked together
randomly but cannot overlap. This feature is modelled by a self-repulsion
term. Let ${\bf A}$ be a d-dimensional hypercubic lattice, typically
the integer lattice ${\bf Z}^d$ (or a finite subset of ${\bf Z}^d$).
Its elements are called sites, oriented pairs of sites are called steps.
A walk $w$ on ${\bf A}$ is an ordered sequence $(w(0),w(1),...,w(k))$ of
sites in ${\bf A}$; $k\geq 0$, $k$ being the length of the walk \cite{Fer}.
Thus, a simple random walk on ${\bf A}$, starting at $w(0)\in {\bf Z}^d$,
is a stochastic process indexed by the non-negative integers.\
To state a self-repulsion term, Flory \cite{Fr} used the self-avoiding walk
(SAW). A self-avoiding walk of length $n$ is a simple random walk which
visits no site more than once. Although this simple model possesses some
qualitative features of polymers, turns out to be very difficult for obtaining
rigorous results. Instead of using SAW we take advantage of the
several measures on random walks which favor self-avoiding walks.\
In this paper we study weakly SARW (so called Domb-Joyce model or
self-repellent walk). This is a measure on the set of simple walks in which
self-intersections are discouraged but not forbidden. Here, double
intersections
of random walks are penalized by a factor
$e^{-\lambda \mbox{ (measure of (self-)intersection)}}$, $\lambda>0$ being a
small constant.
This factor is needed to make the process tend to avoid itself. Thus,
probabilities of the random walks are modified by a measure of
(self-)intersection inside the lattice. This measure is written in terms of
waiting (or local) times for the process ($t_i$). Here, as was done in
reference
[5], the measure is $\sum_{i<j}t_it_j$, for all local times in the\newline
(self-)intersections of the walks on the lattice, which can be either
hypercubic
or hierarchical. If in this model the lattice is the hypercubic lattice, then
the state space is ${\bf Z}^d$. To understand this model, we use a
hierarchical lattice which state space is defined in next section. We want to
stress that the only feature of the model that depends upon the lattice used
(hypercubic or hierarchical) is the state space in this. Thus, definitions
for the weakly SARW in section 4 (interaction energy in Theorem 4) are
equivalent in both cases, but with different state space in the lattice. We
develop our method on a hierarchical lattice because they have the feature
that the renormalization-group map is particulary simple, which is not the
case on the hypercubic lattice. We believe that the results of this procedure
extend to weakly SARW on the hypercubic lattice.\
The hierarchical models introduced by Dyson \cite{Dy} feature a simple \newline
renormalization-group transformation. This can easily be seen in related
literature $\cite{Ga}$. We would like to understand the logarithmic correction
that appears in the hypercubic lattice, d=4, for the end-to-end distance
of the weakly SARW. The real space renormalization-group map we develop
here, is factorizable only in terms of a hierarchical lattice. So, we present
a method in which $A$ is labeled in terms of a hierarchical metric space,
from this an easy realization of the map is followed. In the integer lattice
${\bf Z}^d$, this is not true and technical problems arise.\
We use a hierarchical lattice where the points are labeled by elements of a
countable, abelian group ${\it G}$ with ultrametric $\delta$; i.e. the metric
space $({\it G},\delta)$ is hierarchical. The hierarchical structure of this
metric space induces a renormalization-group map that is ``local"; i.e.
instead of studying the space of random functions on the whole lattice, we
can descend to the study of random functions on L-blocks (cosets of {\it G})
\cite{Ev}. This simplifying feature was used by Brydges, Evans and Imbrie
\cite{Ev} to prove (in the $\lambda \phi^4$ Grassmann valued field
representation for a weakly SARW that penalizes the (self-)intersection of
two random walks) that the introduction of a sufficiently weakly
self-avoidance interaction does not change the decay of the Green's function
for a particular L\'{e}vy process (continuous time random walk), \cite{Ev}
when d=4, provided the mass is introduced critically.\
A rigorous proof of the end-to-end distance for the weakly SARW, $d=4$, on
the hierarchical lattice, has recently been reported \cite{Im}. This was done
in
the field theoretical approach by means of controlling the interacting
Green's function and inverting the Laplace transform.\
Low dimensional models are the most interesting from physical viewpoint,
but rigorous results are difficult of being obtained. One major result is the
proof that in high dimensions the exponents of weakly SARW take the
``mean-field" value \cite{Ma}. In this context ``lace expansion" is one of
the most successful tools. Contrary to what has been done from this method,
in this paper we develop a map on real space which is full of the physical
intuition
needed for being applied on some other cases not yet solved.\
Weakly SARW exhibits logarithmic corrections for physically meaningful
magnitudes in the critical dimension of the model, i.e. d=4. We study the
end-to-end distance for weakly SARW that penalizes the double crossing of
walks in $d=4$. A probabilistic meaning is given to the exponent of this
logarithmic correction. In this paper we present an heuristic space-time
renormalization-group argument to show that the end-to-end distance of a
weakly self-avoiding random walk (SARW) on the hierarchical lattice, that
penalizes the (self-)intersection of two walks in $d=4$, is asymptotic to a
constant times $T^{\frac{1}{2}}log^{\frac{1}{8}}T$ as T tends to infinity,
T being the total time for the walk. This has already been conjectured
before \cite{Br}. This is the testing ground to check that our map reproduces
previously known results with physical improved intuition. The weakly
self-avoiding random walk model that penalizes the (self-)intersection of two
random
walks is in the same universality class as the perfect self-avoiding random
walk model; therefore, the same logarithmic correction for the end-to-end
distance is expected to hold for both cases regardless the details of the
state space used. Real space renormalization-group methods have proved to be
useful in the study of a wide class of phenomena.\
Since our method is intended to provide an alternative way, full of physical
intuition, to renormalize random walk models on a hierarchical lattice; we
study weakly SARW in $d=4$ just as a testing ground. We consider our method
suitable of being directly applied on kinetically growing measure models,
discrete version. Kinetically growing measure models are produced from
consistent measures and are the natural framework for random walks provided
these are seen like stochastic processes. Myopic self-avoiding walk (a model
for adsorption of linear polymers on surfaces \cite{Bu}), infinite
self-avoiding walk and Laplacian (or loop-erased) random walk \cite{La} are
examples of kinetically
growing measure models. These models have been studied mainly in the field
theoretical framework where has been shown \cite{De} that some contributions
neglected in the derivation of the continuum problem from the discrete version
of the model might be essential in determining the asymptotic behaviour of the
model analytically. In fact, usually the corresponding action (in the
continuum limit) is not fully renormalizable. We address this problem by
presenting a real space renormalization-group suitable of been applied on
discrete models. The metod can be used for obtaining both, rigorous and
heuristic results for these models.\
We hope this paper could be interesting for various readerships. We intend to
brydge a gap between probabilistic approaches and field theoretical ones,
thereby providing a new probabilistic meaning to critical and asymptotic
exponents. The method we develop in this paper can be used also as an
intuitive mean to search new exact results, which then remain to be proven
rigorously within the framework presented here, or by other means. The
renormalization scheme constructed here is not considered or known in the
literature.\
This paper is organized as follows; in Section 2 we present
the hierarchical lattice and the L\'{e}vy process we study, on the space
of simple random walks. In Section 3 we define the
renormalization-group map on the hierarchical lattice and prove that two
particular probabilities, functions of random walks, flow to fixed forms
after applying the map. In Section 4 we apply the renormalization-group map
to the weakly SARW model that penalizes the intersection of two random walks
on the hierarchical lattice. In Section 5 we present an heuristic
proof for the asymptotic behavior of the end-to-end distance for the weakly
SARW on the hierarchical lattice, d=4. Although this is an heuristic proof,
it helps in understanding the way the map can be used and gives a new
probabilistic meaning to the exponent of the logarithmic correction. Sections
3, 4 and 5 involve important results of this paper, being Theorem 4
the main result reported in the paper and an important step to obtain
heuristically the end-to-end distance for the weakly SARW. Even more, we claim
this Theorem to be a random walk version of the field theoretical
approach in reference [5] with improved physical intuition. Finally, we
summarize.\
\section{The hierarchical random walk.}
The hierarchical lattice used in this paper was recently introduced by
Brydges, Evans and Imbrie \cite{Ev} \cite{Im}. Here we are presenting a slight
variant of the model in reference [5].\
Fix an integer $L\geq 2$. Hereafter, the points of the lattice ${\bf A}$ are
labeled by elements of the countable abelian group
${\it G}=\oplus ^{\infty}_{k=0}{\bf Z}_{L^d}$, d being the dimension
of the lattice. Through the paper the abelian group
${\it G}=\oplus ^{\infty}_{k=0}{\bf Z}_{L^d}$ replaces ${\bf A}$.\
An element $X$ in ${\it G}$
is an infinite sequence
$$
X\equiv (...,X_k,...,X_2, X_1, X_0 )\;\; ;
\mbox{$X_i \in {\bf Z}_{L^d}$ thus $X\in {\it G}=\oplus^{\infty}_{k=0}
{\bf Z}_{L^d}$},
$$
where only finitely many $X_i$ are non-zero.\
Let us define subgroups
\begin{equation}
\{0\}={\it G}_0\subset {\it G}_1\subset ...\subset {\it G}
\;\;\;\mbox{where }
{\it G}_k=\{X\in {\it G}| X_i=0, i\geq k \},
\end{equation}
and the norm $|\cdot|$ as
\begin{equation}
|X|=\left\{\begin{array}{cc}
0 & \mbox{ if $X=0$} \\
L^p & \mbox{where $p=\inf\{k| X\in {\it G}_k\}$ if $X\neq 0$}.
\end{array}
\right.
\end{equation}
Then, the map $\delta:(X, Y)\rightarrow |X-Y|$ defines a metric on ${\it G}$.
In this metric the subgroups
${\it G}_k$ are balls $|X|\leq L^k$ containing $L^{dk}$ points.
Here the operation + (hence - as well) is defined componentwise.\
In Figure 1 we have described two examples for $L=2$, a one-dimensional
hierarchical lattice (Figure 1.a)) already presented in reference [5] and a
two-dimensional hierarchical lattice (Figure 1.b)). In these Figures we
depict ${\it G}_k$ cosets, a way to calculate distances among
points, and the concept of scales for each example.\
The metric defined by eq(2) satisfies a stronger condition than the
triangle inequality, namely
\begin{equation}
|X+Y|\leq \mbox{ Max}(|X|,|Y|).
\end{equation}
This {\it non-archimedean} property implies that every triangle is isosceles
and that every point interior to a ball can be considered its center.
Moreover, balls of radius L are the same as balls of diameter L, and are
the same as ${\it G}_1$ cosets. From inequality (3), it is clear that the
metric introduced is an ultrametric and confers the hierarchical structure.
Strictly speaking, it is only the metric space $({\it G},\delta)$ that is
hierarchical. Here, ultrametric appears naturaly as a property of polynomials.
It can be shown that ${\it G}_k$ represents polynomials of degree $k$ on a
formal basis.\
Let us now introduce the L\'{e}vy process we propose in this paper.
The elements of the lattice ${\it G}$ are called sites; unoriented pairs
$\{X,Y\}$ of sites in ${\it G}$ with $X\neq Y$ are called bonds; oriented
pairs $(X,Y)$ are called steps (or jumps) with initial site $X$ and final
site $Y$. Let us define the L\'{e}vy process \cite{Ev}($\equiv $ continuous
time random walk), $w$, as an ordered sequence of sites in {\it G};
\begin{equation}
(w(t_0),...,w(t_0+...+t_n))\;\;,w(t_0+...+t_i)=X_i\in {\it G},
\;\;T=\sum ^n_{i=0}t_i ,\; n\geq 0
\end{equation}
where $t_i$ is the time spent in $X_i\in {\it G}$ (waiting time at $X_i$)
and T, fixed, is the running time for the process. For convenience we take
$X_0=0$. The support of the walk $w$ is defined by
\begin{equation}
supp(w)=\{X\in {\it G} | w(t_0,...,t_j)=X \mbox{ for some j}\},
\end{equation}
for any $w$. The random walk we are dealing with is not the nearest
neighbor random walk on the lattice, provided we mean neighbourhood with
respect to the ultrametric distance $\delta$ previously defined.\
If we compare this L\'evy process with the simple random walk in the
hypercubic lattice defined in the Introduction, we see that in this Section we
construct a
different stochastic process. Here, time is continuous and the path of the
walk is given on ${\it G}$ (in the hypercubic lattice, the walk was indexed
by non-negative integers, instead of waiting times, with path on ${\bf Z}^d$).
Besides, here we fix the running time for the process. This corresponds to
walks of fixed length, feature that was not imposed on simple random walks on
the hypercubic lattice.\
We suppose the L\'{e}vy process in the hierarchical lattice has a probability
$rdt$ (r is the jumping rate) of making a step in time $(t, dt)$ and, given
that it jumps, the probability of jumping from X to Y is $q(X,Y)$. Thus, the
process, conditioned to $n$ jumps spaced by times $t_0,t_1,...,t_n$, has a
probability density
\begin{equation}
P(w)=r^ne^{-rT}\prod^{n-1}_{i=0}q(X_{i+1},X_{i})
\;\;\mbox{, where $T=\sum^{n}_{i=0}t_i$.}
\end{equation}
Define $Dw$ by
$$
\int(\cdot ) Dw=
\sum_{n}\;\;\sum_{\left[ X_{i}\right]^{n}_{i=0}}
\;\;\;\int^{T}_{t_i=0}\;\;\prod^{n}_{i=0}\;\;dt_{i}
\delta(\sum^{n}_{i=0}t_{i}-T)(\cdot ).
$$
{}From this and eq(6) it is straightforward to obtain
$$
\int P(w)Dw=
\left\langle \sum_{\left[ X_{i}\right]^{n-1}_{i=0}}
\prod^{n-1}_{i=0}q(X_{i+1},X_{i})
\right\rangle_{Poisson}=1,
$$
$\left\langle (\cdot) \right\rangle$ being the expectation of ($\cdot$).\
Here $\prod^{n-1}_{i=0}q(X_{i+1},X_{i})$ has been normalized on
${\it G}$ and we have used
$$
\int^T_0\prod^n_{i=0}r^ne^{-rt_i}\prod^n_{j=0}dt_j\delta
\left(\sum^n_{j=0}t_j-T\right)=
$$
$$
r^ne^{-rT}\int^T_0\prod^n_{i=0}dt_i\delta
\left(\sum^n_{i=0}t_i-T\right)=\frac{\left(rt^n\right)}{n!}e^{-rT}.
$$
\section{The renormalization-group map.}
To start with, let us introduce a renormalization-group map on the
lattice; $R(X_i)=LX'_i$ where $X_i\in {\it G}$ and $LX'_{i}\in {\it G}'$
=${\it G}/{\it G}_1\sim {\it G}$; i.e.
the renormalized lattice ${\it G}'$ is isomorphic to the original
lattice ${\it G}$.\
{}From this renormalization-group map we construct $R(w)=w'$, from $w$
above as defined, to $w'$. Here,
$w'$ is the following ordered sequence of sites in ${\it G}'=
{\it G}/{\it G}_1\approx {\it G}$;
\begin{equation}
(w'(t'_0),...,w'(t'_0+...+t'_k))\;\;,\mbox{ where}
\end{equation}
$$
w'(t'_0,...,t'_{i'})=X'_{i'}\in {\it G},\;\;T'=\sum ^k_{i'=0}t'_i ,
\;0\leq k\leq n,
\;\;T=L^{\beta}T'.
$$
R maps $w(t_0),w(t_0+t_1),...,w(t_0+...+t_n)$ to cosets
$w(t_0)+{\it G}_1,w(t_0+t_1)+{\it G}_1,...,w(t_0+...+t_n)+{\it G}_1$. If
two or more successive cosets in the image are the same, they are listed only
as one site in $w'(t'_0),...,w(t'_0+...+t'_k)$, and the times $t'_j$ are sums
of the corresponding $t_j$ for which successive cosets are the same, rescaled
by $L^{\beta}$. For $\beta=2$, we are dealing with normal diffusion,
in case $\beta<2$ with superdiffusion, and subdiffusion for $\beta>2$.
Additionally, the renormalization-group maps each ${\it G}_1$ coset to
the center of the ball and rescales by $L$. In reference [5], $\beta$ is set
to 2.\
The renormalization group map, applied to functions of the hierarchical
random walk, preserves locality \cite{Ev}. Thus, if
$F(w)=\prod_{X_{i}\in {\it G}}f(w(t)=X_{i})$, the effect of the
renormalization-group
map on $F(w)$ can be studied as the product,
for all elements of the group ${\it G}/{\it G}_1={\it G}'\sim {\it G}$, of
the images of the renormalization-group map of $f(w)$ in the ${\it G}_1$
coset \cite{Ev}. This can be seen as follows; $\prod_{X_{i}\in {\it G}}$
splits into two parts; the first part is $\prod_{X'_{i'}\in {\it G}/{\it G}_1}$
which corresponds to $\prod_{X'_{i'}\in {\it G}}$. The second part is
$\prod_{X_{i}\in ({\it G})_{X'_{i'}}}$ and stands for the $L$-block (${\it
G}_1$
coset) of the lattice which, under the renormalization-group transformation,
maps to $LX'_{i'}$, $(0\leq i'\leq k)$. The geometrical interpretation of this
is quite simple. The renormalization-group map applied on $F(w)$ in ${\it G}$
splits into that of $f(w(t)=X_{i})$ in the contracting ${\it G}_1$ cosets,
multiplied by the whole ${\it G}/{\it G}_1$ group. To study the
renormalization of $F(w)$ in ${\it G}$ we descend to analize the
renormalization-group action on $f(w(t)=X_{i})$ in the contracting ${\it G}_1$
coset.\
In Figure 2 we present the lattice ${\it G}$. On this, walks with fixed
topology in ${\it G}/{\it G_1}$ are depicted. In this example three different
types of local topologies inside the ${\it G}_1$ cosets are illustrated.
Once the renormalization-group map is applied, the renormalized lattice and
random walk are also shown. The renormalized random walk $w'$ visits each
type of ${\it G}_1$ cosets once, twice and three times respectively. In the
renormalized lattice we show the contracted ${\it G}_1$ cosets. The particular
fixed topology chosen for $w'$ is one in the class of the simplest cases
studied by our method. Here is clear what is meant by factorization and
locality of the renormalization-group map. The map factorizes into two terms.
Roughly speaking, the first term corresponds to events inside $L$-blocks or
${\it G}_1$ cosets and the second term corresponds to events outside the
$L$-block (${\it G}_1$ cosets); therefore in ${\it G}/{\it G}_1$=$
{\it G}'\sim {\it G}$. Moreover, for obtaining the flow of the interaction
constant, the map descends to study events in ${\it G}_1$ cosets; thereby
preserving locality.\
We can now work out probabilities at the $(p+1)^{th}$ stage in the
renormalization provided only that we know the probabilities at the $p^{th}$
stage. We sum over the probabilities of all the walks $w^{(p)}$ consistent
with a fixed walk $w^{(p+1)}$ in accordance with the following;\
{\bf Definition. Let $R(w)=w'$ be the renormalization-group
map, above as stated. Then
\begin{equation}
P'(w')=
L^{\beta k}\int Dw P(w)\chi (R(w)=w')
\end{equation}
for any probability $P(w)$ where the running time for the process
$T=\sum^{n}_{i=0}t_{i}$ is fixed}.\
In this definition, R is a renormalization-group transformation that maps
a density $P(w)$ to a new one, $P'(w')$, on rescaled coarse walks.
Besides, $\chi(c)$ is the characteristic function of the
condition c.\
Let $P(w)=\prod_{X\in {\it G}}p(w(t)=X)$ then\newline
$P'(w')$=$L^{\beta k}\int Dw\prod_{X\in {\it G}}p(w(t)=X)\chi (R(w)=w')$ .
{}From this and factorization properties in the hierarchical lattice follows
$$
P'(w')=\prod_{X'\in{\it G}}\left\{
L^{\beta k}\int Dw\prod_{X\in ({\it G}_1)_{X'}}p(w(t)=X)\chi(R(w)=w')\right\}
$$
$$
=
\prod_{X'\in {\it G}}p'(w'(t')=X'),
$$
that proves in this case the statement of preservation of locality as above
given for $F(w)$. Eq(17) and eq(21) are examples for suitable $P(w)$.\
Eq(8) corresponds precisely to
\begin{equation}
P'(w')=
L^{\beta k}\;\sum_{\left[ n_{i'}\right]^{k}_{i'=0}}
\;\sum_{\left[ X_{i}\right]^{n}_{i=0}}
\int \prod^{n}_{i=0}dt_{i}\;
\prod^{k}_{j'=0}\;
\delta (\sum^{m_{j'}}_{{i}=m_{j'-1}+1}\;\;t_{i}-
L^{\beta }t'_{j'})\times
\end{equation}
$$
\times
\prod^{k}_{j'=0}\;
\prod^{m_{j'}}_{{i}=m_{j'-1}+1}\;
\chi (X_{i}\in LX'_{j'}\;\;)P(w).
$$
Hereafter
\begin{equation}
m_{j'}=\sum^{j'}_{i'=0}n_{i'}+j' \;\;\mbox{ and}
\end{equation}
$$
n=\sum^{k}_{i'=0}n_{i'}+k\;\;\;\;\;\;0\leq j'\leq k.
$$
$n_{i'}$ is the number of steps (for walks $w$) in the ${\it G}_1$
coset which, once the renormalization-group map is applied, has the image
$LX'_{i'}$. Thus, the total number of steps (for walks $w$) on the lattice is
given by the sum of the steps within each L-block (${\it G}_1$ cosets) plus
$k$ times 1 (due to the step out of the corresponding block).\
{\bf Theorem 1. The probability in eq(6), where $q(X_{i+1},X_{i})$ is
chosen of the form $c|X_{i+1}-X_{i}|^{-\alpha}$ (c is a constant fixed up to
normalization and $\alpha$ another constant), is a fixed point of the
renormalization-group map $R(w)=w'$, provided $\beta =\alpha-d$}.\
{\bf Proof}. Using the definition of the renormalization-group
map on the probability given in eq(6) and doing some elementary
manipulations we arrive to the following expression
\begin{equation}
P'(w')=L^{\beta k}\;r^{k}\;e^{-rT}
\prod^{k-1}_{j'=0}\;\;(q(LX'_{j'}-
LX'_{j'+1})L^d)\times
\end{equation}
$$
\prod^{k}_{j'=0}\;\;\sum_{n_{j'}}\;\;
r^{n_{j'}}\;\;(q_{1}(L^d-1))^{n_{j'}}
\frac {(L^{\beta }t'_{j'})^{n_{j'}}}
{n_{j'}!}
$$
where $\prod^{n-1}_{i=0}q(X_{i+1}-X_{i})$ has been split into two factors; the
first factor
corresponding to jumps from one $L$-block to another $L$-block
(different ${\it G}_1$ cosets) and the second factor corresponding
to jumps inside the same $L$-block or ${\it G}_1$ coset. Function $q_1$ in
eq(11)
is the probability of jumping to a given point within the ${\it G}_1$ coset
that has the image $LX'_{j'}$
(hereafter $\left({\it G}_{1}\right)_{X'_{j'}}$).
There are $(L^d-1)$ possibilities with equal probability $q_1$ and
$n_{j'}$ steps.\
{}From normalization, i.e. $\sum_{X\in {\it G}}q(X)=1$,
we get $c=\frac {L^{\alpha -d}-1}{1-L^{-d}}$ and \newline
$q_{1}(L^d-1)=1-L^{d-\alpha }$.\
On the other hand, we know that $T=L^{\beta }T'$, therefore, eq (11) becomes
\begin{equation}
P'(w')=L^{(d+\beta -\alpha)k}\;\;r^{k}
\;\;\prod^{k-1}_{i'=0}q(X'_{i'+1}-X'_{i'})e^{-rL^{(d+\beta -\alpha)}T'}
\end{equation}
Provided $d+\beta -\alpha =0$, eq(12) is clearly a fixed point
of the renormalization-group map $R$, i.e.
$P'(w')=r^{k}\prod^{k-1}_{i'=0}q(X'_{i'+1}-X'_{i'})e^{-rT'}$.
Q. E. D.\
Theorem 1 corresponds to the case worked out by Brydges, Evans,
and Imbrie if we choose $\beta =2$, i.e. diffusive
behavior \cite{Ev}.\
{\bf Theorem 2. If the probability in eq(6) where
$q(X_{j},X_{j+1})$ is chosen of the form
\begin{equation}
q(X_j-X_{j+1})=c\left(\frac{1}{|X_j-X_{j+1}|^\alpha }+
\frac{1}{|X_j-X_{j+1}|^\gamma }\right)
\end{equation}
(c is a constant fixed up to normalization, $\alpha$ and $\gamma$ are
constants, $\alpha\neq \gamma$) then $P(w)$ flows to the fixed
point $P'(w')$ given in theorem 1, (i.e. $q(X_j-X_{j+1})$
is given as in theorem 1), under the
renormalization-group map $R(w)=w'$,
provided $\gamma >>\alpha$ (such that
$log \left(\frac{L^{-\alpha}-L^{-\gamma}-2L^{d-\gamma-\alpha}}{L^{-\alpha}-
2L^{d-\gamma}}\right)\rightarrow 0$) and $\beta =d+\alpha $}.
{\bf Proof}. Following the same ideas as in theorem 1, from normalization,
we obtain
$
c=\frac{(1-L^{d-\alpha})(1-L^{d-\gamma})}
{(L^d-1)(L^{-\alpha}+L^{-\gamma}-2L^{d-\gamma -\alpha})}
\;\;\mbox{ and }
q_{1}=c(L^{-\gamma}+L^{-\alpha}) .
$
Then, if $\gamma >> \alpha$, $P'(w')$
corresponds to
\begin{equation}
P'(w')=L^{(d+\beta -\alpha)k}\;\;r^{k}
\;\;e^{-rL^{\beta }T'}\times
\end{equation}
$$
\times
\prod^{k-1}_{j'=0}\frac
{(L^{\alpha -d}-1)}{(1-L^{-d})}
|X'_{j'}-X'_{j'+1}|^{-\alpha}
e^{r(1-L^{d-\alpha})L^{\beta }T'}
$$
If $d+\beta -\alpha=0$, eq(14) reduces to $P'(w')$ as given in
Theorem 1.\
Q.E.D.\
Theorem 2 is presented here in order to learn about the L\'{e}vy process we
are studying. This case is intended to answer the question about the
feasibility of introducing perturbations in the probability and the possible
results to be obtained. We are looking forward to studying of asymmetric and
``trapping" environments .\
\section{The renormalization-group map on \newline
the weakly SARW}
A configurational random walk model can be defined by assigning to every
n-tuple of walks
$w_1,...,w_n $ $ (n\geq 0)$ a statistical weight. For a simple random walk
model, this is the product of the statistical weights for each of the n walks
and can serve as a random walk representation of the Gaussian model
\cite{Fer}\cite{Br}. The best known mathematical model that involves a
self-repulsion term is the self-avoiding random walk. A self-avoiding
walk of length n is a simple random walk which visits no site more
than once.
Unfortunately, it turns out that it is extremely difficult to obtain rigorous
results from this model for $d\leq 4$ \cite{Ma}\cite{La1}
\cite{La2}. However, there are other
ways to include self-repulsion terms in random walk models \cite{La}.
These split naturally into two categories: configurational measures where
random walks are weighted by the number of (self-)intersections, and
kinetically growing measures where random walks are produced from
consistent measures that are generated by Markovian transition probabilities
on the states space of simple random walks (these measures are non-Markovian
on the state space of the lattice) \cite{La}. The weakly self-avoiding random
(or Domb-Joyce) model and the Edwards model are examples of the first category
\cite{La}. The ``true" (or ``myopic") self-avoiding walk and the Laplacian
random walk are examples of the second category \cite{La}.
In this paper we deal only with configurational measures.
A simple random walk model can be thought of as being endowed with a
configurational measure where the weight for the self-intersections of a
walk (and/or among the n-tuple of walks $w_1,...,w_n $ $ (n\geq 0)$) is null.
Configurational measures are measures on $\Lambda_n$, the space of simple
random walks of length n. Let $P_U(w)$ be a probability on $\Lambda_n$
such that
\begin{equation}
P_U(w)=\frac{U(w)P(w)}{Z}
\end{equation}
where $Z=\int U(w)P(w)Dw$. In this paper we consider $P(w)$ as the
probability on $\Lambda_n$ given in eq(6) and $U(w)$ as the interaction
energy of the walks \cite{Fer}. Thus, to study the effect of the
renormalization-group map on $P_U(w)$ we need to follow the trajectory of
$U(w)$ after applying several times the renormalization-group map.\
Therefore, from the definition of the renormalization-group
map in eq(8);
$$
P'_{U'}(w')=L^{\beta k}\int
P_{U}(w)
\chi (R(w)=w')Dw
$$
where $Z'=Z$, it follows that
\begin{equation}
U'(w')=
\frac{\int Dw P(w)\chi (R(w)=w')U(w)}
{\int Dw P(w)\chi (R(w)=w')}
\end{equation}
Note that eq(16) can be viewed as the expectation of $U(w)$
given that the renormalization-group map is imposed, calculated using
$P(w)$ on $\Lambda_{n}$ defined as in eq(6). Therefore, and hereafter,
to simplify notation, we write eq(16) as $U'(w')=< U(w) >_{w'}$.\
In this Section we deal with $U(w)$ factorizable in terms of the interaction
energy with null weight
for the (self-)intersection of $n$ $(n=2,3,...,etc.)$ random walks
(i.e. a simple random walk factor), and the
interaction energies that
weight the intersection of n random walks. Hereafter, as a hypothesis,
we assume all the factors in
$U(w)$ (functions of $w$) as independent, simple, random variables.
Thus, the conditional (given the $R(w)=w'$ map) expectation of
$U(w)$ is the product of conditional (given the $R(w)=w'$ map)
expectation of each factor in $U(w)$.
This hypothesis follows the same spirit as in the approach used
in polymer networks \cite{Du} and can be seen as a consequence of
factorizability and locality of the map. See Figure 2.\
To start with, we study the simple random walk model such that
\begin{equation}
U(w)=\prod_{X\in {\it G}} e^{-a\sum _{j\in J_{X}}\;\;t_j},
\end{equation}
where $J_{X}=\{j\in \{0,...,n\}|X_j=X \}$ for $w(t_0+,...,+t_j)=X_j$ and
$X\in {\it G}$.\
{\bf Theorem 3. The probability $P_U(w)$ for the simple random walk
model where
U(w) is given by eq(17), is a fixed form of the \newline
renormalization-group map
$R(w)=w'$ such that, after applying the renormalization-group map,
$a'=L^{\beta }a$}.\
{\bf Proof}. Let us split the product on sites in the lattice in
$ \prod_{X\in {\it G}} e^{-a\sum _{j\in J_{X}}\;\;t_j} $
into two parts. The first one, i.e.
$
\prod_{X'_{i'}\in {\it G}/{\it G_1}}
\mbox{ corresponds to }
\prod_{X'_{i'}\in {\it G}}
$
due to the hierarchical structure of the lattice. The second one, i.e.
$
\prod_{X_{i}\in ({\it G_1})_{X'_{i'}}}
$
stands for the L-block (${\it G_1}$ coset) of the lattice ${\it G}$
that, under the renormalization-group transformation, maps to
$LX'_{i'},\;\;(0\leq i'\leq k)$. There
are $k$ replicas of this. If we again split $\prod^{n-1}_{i=0}q(X_{i+1}-X_i)$
into two factors,
as was done in theorem 1. We obtain
\begin{equation}
U'(w')=\prod_{X'_{i'}\in {\it G}}
\left\{
\prod_{j'\in J_{X'_{i'}}}
e^{q_{1}(L^d-1)r(L^{\beta}t'_{j'})}
\right\}^{-1}\times
\end{equation}
$$
\left\{
\sum_{n_{i'}}\;\int\prod_{i\in I_{X'_{i'}}}
\;dt_{i}\;\prod_{j'\in J_{X'_{i'}}}
\;\delta (\sum^{m_{j'}}_{{i}=m_{j'-1}+1}\;\;t_{i}-
L^{\beta }t'_{j'})\times
\right.
$$
$$
\sum_{X_{i}\in ({\it G}_1)_{X'_{i'}}}
\prod^{m_{j'}}_{{i}=m_{j'-1}+1}\;\;\;
\chi (X_{i}\in LX'_{j'})\;\;
(q_{1}(L^d-1)r)^{n_{i'}}
\left.
e^{-a\sum_{\left(X_{i}\in {\it G_1}\right)_{X'_{i'}}}
\sum _{j\in J_{X_{i}}}\;\;
t_{j}}
\right\}
$$
where we have defined , for $X_{i}\in w$
and $X'_{i'}\in w'$;
\begin{equation}
I_{X'_{i'}}=\left\{i\;|\;X_{i}\in
LX'_{i'}\right\}=
\end{equation}
$$
\bigcup_{j'\in J_{X'_{i'}}}
\left\{i\;|\;m_{j'-1}+1\leq i\leq m_{j'}\right\}=
\bigcup_{j'\in J_{X'_{i'}}}
\left\{i\;|\; 0\leq i\leq n_{j'}\right\}.
$$
Rearranging the double sum in the exponential of eq(18), we obtain
\begin{equation}
U'(w')=\prod_{X'\in {\it G}'}
e^{-aL^\beta \sum_{j'\in J_{X'}}t'_{j'}}.
\end{equation}
Q.E.D.\
The next model we want to study is a weakly self-avoiding random walk (or
Domb-Joyce model), with a configurational measure in which double
intersections of walks are penalized by (roughly speaking) a factor of
$e^{-\lambda}$. As $\lambda \rightarrow \infty$, this reduces to random walks
with strict mutual avoidance. Recall that this weakly model (with
$\lambda >0$) and the perfect self-avoiding random walk, are in the same
universality class (this implies that the critical exponents are the same).
If $\lambda=0$, this corresponds to a simple random walk model. Next Theorem
is an important result of this paper; it is a random walk version of the
field theoretical approach \cite{Ev} with improved physical intuition and the
key stone in our method to obtain the end-to-end distance of the weakly SARW.
The renormalization-group map applied on a weakly SARW involves the
paramaters $\gamma_1$, $\gamma_2$, $\beta_1$, $\eta$, $A$, $B$, $C$. These
are some conditional expectations of local times for different topologies in
both, $w$ and $w'$ random walks and are precisely defined in Figure 3.\
In Figure 3, $\gamma_1$ and $\gamma_2$ correspond to $O(\lambda)$ and
$O(\lambda^2)$ contributions in Figure 2.b), respectively. $\beta_1$
corresponds to $O(\lambda^2)$ contribution in figure 2.c) and $\eta$
corresponds to $O(\lambda^2)$ contribution in Figure 2.d). Although $A$, $B$,
and $C$
are not depicted in Figure 2, it is straightforward to figure out the
corresponding pictures.\
{\bf Theorem 4. For the weakly SARW with interaction energy}
\begin{equation}
U(w)=\prod_{X_i\in {\it G}}
e^{-\xi\sum_{i\in J_{X_{i}}}t_{i}-\lambda \sum_{i<j\in J_{X_{i}}}
t_{i}t_{j}
{\bf 1}_{\left\{w(t_i)=w(t_j)\right\} }}
\end{equation}
{\bf $\xi_{2}<0$ and $\lambda >0$ being (small) constants, the probability
$P_U(w)$ flows to a fixed form after the renormalization-group
transformation is applied. This fixed form
is characterized by the interaction energy}
\begin{equation}
U'(w')=
\prod_{X'_{i'}\in{\it G}}
e^{
-\xi'
\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}-
\lambda'
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}}))}}
\times
\end{equation}
$$
\left\{1+\eta'_{1}
\sum_{
\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'< k_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}',
k_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
t'_{k_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}})
=w(t'_{k'_{\alpha_{1}}}))}+
\right.
$$
$$
+\left.
\eta'_{2}
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}}))}
+\eta'_{3}\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}\right\}+r'.
$$
{\bf Here}
\begin{eqnarray}
\xi' & = & L^{\beta}\xi+\xi'_2 \\
\xi'_{2} & = & \gamma _{1}\lambda-\gamma _{2}\lambda^2+O(\lambda^3)\\
\lambda' & = & L^{(2\beta-d)}\lambda-\beta_{1}\lambda^{2}+O(\lambda^3) \\
\eta'_{1} & = & \eta_{1}A+\eta \lambda^2 \\
\eta'_2 & = & \eta_1B+L^{(2\beta-d)}\eta_2 \\
\eta'_3 & = & \eta_1C+\eta_2\gamma_{1}+L^{\beta}\eta_{3} \\
r'& \sim & O\left(\lambda^3\right).
\end{eqnarray}
{\bf Proof}. From our initial hypothesis, follows
\begin{equation}
U'(w')=
\left\langle U_{\xi}(w)\right\rangle_{w'}
\left\langle U_{\lambda}(w)\right\rangle_{w'},
\end{equation}
to errors $O(\xi\lambda)$ at each site. Here we introduce $\xi=O(\lambda^2)$,
so this can be put in $r$. Besides the subindices $\xi$ and $\lambda $
identify the two factors in eq(21). We know, from theorem 3, the trajectory
of $\left\langle U_{\xi}\left(w\right)\right\rangle_{w'}$, so we only have to
study the trajectory of
$\left\langle U_{\lambda}\left(w\right)\right\rangle_{w'}$.\
We split the product in the interaction term
$\left\langle U_{\lambda}\left(w\right)\right\rangle_{w'}$
into two parts and identify
$
\prod_{X'_{i'}\in {\it G}/{\it G_1}}
\mbox{ with }
\prod_{X'_{i'}\in {\it G}}
$
, therefore;
\begin{equation}
\left\langle U_{\lambda}\left(w\right)\right\rangle_{w'}
=\prod_{X'_{i'}\in {\it G}}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}}_{w'}
\;\mbox{ where }
\end{equation}
\begin{equation}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}}_{w'}=
\left\{
\prod_{j'\in J_{X'_{i'}}}
e^{q_{1}(L^d-1)r(L^{\beta}t'_{j'})}
\right\}^{-1}\times
\end{equation}
$$
\left\{
\sum_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}\;
\sum_{n_{i'}}\int\prod_{i\in I_{X'_{i'}}}dt_{i}
\right.
\prod_{j'\in J_{X'_{i'}}}
\;\delta (\sum^{m_{j'}}_{{i}=m_{j'-1}+1}\;\;t_{i}-L^{\beta }t'_{j'})
\prod^{m_{j'}}_{{i}=m_{j'-1}+1}
\chi (X_{i}\in LX'_{j'})
$$
$$
\times
\prod^{n-1}_{i=0}
q(X_{i+1}-X_i)
\left.
r^{n_{i'}}\;
e^{-\lambda\sum_{X_{i}\in ({\it G_1})_{X'_{i'}}}
\sum _{i_{\alpha_{1}}< j_{\alpha_{1}}\in J_{X_{i}}}\;\;
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}}
\right\}
$$
Recall that $w'$ is fixed; i.e. the walk, after the
renormalization-group map is applied, visits each site
$X'_{i'}\in {\it G}$ a fixed number
of times $n'^{*}_{i'}$. Let us assume $w'$ is such that
$\left\{n'^{*}_{i'}\right\}>0$, and at least once
$n'^{*}_{i'}=1,2,3$ on ${\it G}$. See example in Figure 2. We ask for this
condition to hold, in order to learn about the flow of the interaction
constant in the (self-) intersection of 2 and 3 random walks. In other words,
we ask for a fixed and not totally arbitrary topology for the renormalized
random walk $w'$ on ${\it G}$. To make this condition explicit we rewrite
eq(31) as;
\begin{equation}
\left\langle U_{\lambda}\left(w\right)\right\rangle_{w'}
=\prod_{X'_{i'}\in {\it G}}
\prod_{\left(n'^{*}_{i'}\right)^{k}_{i'=0}}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}}
_{w'_{n'^{*}_{i'}}}
\end{equation}
Here,
$
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}}
_{w'_{n'^{*}_{i'}}}
$
is the renormalized interaction energy for all possible topologies of walks
$w$ inside the ${\left({\it G}_{1}\right)_{X'_{i'}}}$ coset that, once the
renormalization-group map is applied, corresponds to a fixed topology in
$LX'_{i'}\in {\it G}$.\
Let us introduce a formal Taylor series expansion in $\lambda$, then;
\begin{equation}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}}
_{w'_{n'^{*}_{i'}}}=
\sum^{\infty}_{s=1}\frac{\left(-\lambda \right)^s}{s!}
\end{equation}
$$
\left\langle
\sum_{X_i\in {({\it G}_1)}_{X'_{i'}}}\;
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}<;...;
<i_{\alpha_{s}}< j_{\alpha_{s}}}
{\left\{i_{\alpha_{1}},...,i_{\alpha_{s}}
; j_{\alpha_{1}},...,j_{\alpha_{s}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
\times ...
\right.
$$
$$
\left.
...\times t_{i_{\alpha_{s}}}t_{j_{\alpha_{s}}}
{\bf 1}_{(w(t_{i_{\alpha_{s}}})=w(t_{j_{\alpha_{s}}}))}
\chi\left(\left(i_{\alpha_{1}},...,i_{\alpha_{s}};
j_{\alpha_{1}},...,j_{\alpha_{s}}\right)\in (i'_{\alpha_{1}},...,
i'_{\alpha_{n'^{*}_{i'}}})
\right)
\right\rangle ^{l.c.}_{w'}
$$
In this formal series expansion, we are writing
$
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}}
_{w'_{n'^{*}_{i'}}}
$
in terms af all possible classes of topology for walks $w$ inside the
${\left({\it G}_{1}\right)_{X'_{i'}}}$ coset. Each class is an element
of this Taylor series and corresponds to a fix
number $s$ of double (self)-intersections, weighted by $\lambda^s$. Here, the
superscript $l.c.$ means linear contribution. We take into account only
linear contributions to conditional expectations. This approach is considered
to avoid double-counting sites in
walks.\
To start with, we analyze explicitly the case $n'^{*}_{i'}=1$,
(for some $0\leq i'\leq k$) up to 2nd order in $\lambda$;
\begin{equation}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}}
_{w'_{n'^{*}_{i'}=1}}=
1-\lambda\times
\end{equation}
$$
\left\langle
\sum_{X_i\in {({\it G}_1)}_{X'_{i'}}}\;
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},
j_{\alpha_{1}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
\left. {\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
\chi\left(\left(i_{\alpha_{1}},
j_{\alpha_{1}}\right)\in i'_{\alpha{1}}
\right)
\right\rangle ^{l.c.}_{w'}
\right.
$$
$$
+\frac{\lambda^{2}}
{2}
\left\langle
\sum_{\left(X_{i}\in {\it G}_{1}\right)_{X'_{i'}}}\;
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}<
i_{\alpha_{2}}< j_{\alpha_{2}}}
{\left\{i_{\alpha_{1}},i_{\alpha_{2}},
j_{\alpha_{1}},j_{\alpha_{2}}\right\}
\in J_{\left(X_{i}\right)}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}\times
\right.
$$
$$
\times\left.
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
t_{i_{\alpha_{2}}}t_{j_{\alpha_{2}}}
{\bf 1}_{(w(t_{i_{\alpha_{2}}})=w(t_{j_{\alpha_{2}}}))}
\chi\left(\left(i_{\alpha_{1}},i_{\alpha_{2}},
j_{\alpha_{1}},j_{\alpha_{2}}\right)\in i'_{\alpha_{1}}
\right)
\right\rangle ^{l.c.}_{w'}+ r'_{1},
$$
where $r'_1\sim O(\lambda^3)$.\
Thus, eq(35) is written as
\begin{equation}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}},l.c.}
_{w'_{n'^{*}_{i'}=1}}=
1-\left(\gamma_{1}\lambda-\gamma_{2}\lambda^{2}
+O(\lambda^3)\right)
\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}
\end{equation}
$$
\cong e^{-\xi'_{2}
\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}}.
$$
where $\xi'_{2}$ is given as in eq(24),
\begin{equation}
\gamma_{1}=
\left\langle
\sum_{X_i\in ({\it G}_1)_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}}\right)\in i'_{\alpha_{1}}
\right)
\right\rangle ^{l.c.}_{w'}\;\mbox{ and}
\end{equation}
\begin{equation}
\gamma_{2}=
\frac{1}{2}
\left\langle
\sum_{X_i\in ({\it G}_1)_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}<
i_{\alpha_{2}}< j_{\alpha_{2}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}\times
\right.
\end{equation}
$$
\left.\times
t_{i_{\alpha_{2}}}t_{j_{\alpha_{2}}}
{\bf 1}_{(w(t_{i_{\alpha_{2}}})=w(t_{j_{\alpha_{2}}}))}
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right)\in i'_{\alpha_{1}}
\right)
\right\rangle ^{l.c.}_{w'}.
$$
$\gamma_1$ and $\gamma_2$ are explained in Figure 3.\
In the same spirit we study the case $n'^{\ast}_{i'}=2$, (for some
$0\leq i'\leq k$) up to second order in $\lambda $;
\begin{equation}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}}lc}
_{w'_{n'^{*}_{i'}=2}}=1-\left(
L^{(2\beta-d)}\lambda-\beta_{1}\lambda^{2}+O(\lambda^3)\right)\times
\end{equation}
$$
\times
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}}))}
$$
$$
\cong
e^{-
\lambda'
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}}))}}.
$$
where $\lambda'$ is given by eq(25) and
\begin{equation}
\beta_1=\frac{1}{2}
\left\langle
\sum_{X_i\in ({\it G}_1)_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}<
i_{\alpha_{2}}< j_{\alpha_{2}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
\right. \times
\end{equation}
$$
\times\left.
t_{i_{\alpha_{2}}}t_{j_{\alpha_{2}}}
{\bf 1}_{(w(t_{i_{\alpha_{2}}})=w(t_{j_{\alpha_{2}}}))}
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right)\in \left(i'_{\alpha_{1}}<
j'_{\alpha_{1}}\right)
\right)
\right\rangle ^{l.c.}_{w'}.
$$
$\beta_1$ is explained in Figure 3.\
Finally, let us present the factor $n'^{*}_{i'}=3$,
(for some $0\leq i'\leq k$), up to 2nd order in $\lambda$;
\begin{equation}
\left\langle U_{\lambda}\left(w\right)
\right\rangle^{\left({\it G}_{1}\right)_{X'_{i'}},l.c.}
_{w'_{n'^{*}_{i'}=3}}=
\left(1+
\right.
\end{equation}
$$
+\eta\lambda^{2}
\sum_{
\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'< k_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}',
k_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
t'_{k_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}})
=w(t'_{k'_{\alpha_{1}}}))}
\left.
+r'_{3}\right)
$$
where $\eta$ is
\begin{equation}
\eta=\frac{1}{2}
\left\langle
\sum_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}<
i_{\alpha_{2}}< j_{\alpha_{2}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
\right. \times
\end{equation}
$$
\times\left.
t_{i_{\alpha_{2}}}t_{j_{\alpha_{2}}}
{\bf 1}_{(w(t_{i_{\alpha_{2}}})=w(t_{j_{\alpha_{2}}}))}
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right)\in \left(i'_{\alpha_{1}}<
j'_{\alpha_{1}}<k'_{\alpha_{1}}\right)
\right)
\right\rangle ^{l.c.}_{w'} .
$$
$\eta$ is explained in Figure 3.\
Note that in eq(41)
{\footnotesize \begin{equation}
\left\langle
\sum_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},
j_{\alpha_{1}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
\chi\left(\left(i_{\alpha_{1}},
j_{\alpha_{1}}\right)\in \left(i'_{\alpha_{1}}< j'_{\alpha_{1}}
< k'_{\alpha_{1}}\right)\right)
\right\rangle ^{l.c.}_{w'}=0
\end{equation}}
for all set of walks $\{w\}\in \Gamma_{n}$, being $\Gamma_n$ the space of
random walks with only double (self-)intersecting walks.\
Note that the case $n'^{*}_{i'}=1$, for some $0\leq i'\leq k$,
shows how the weakly (self-)avoiding random walk on the
$\left({\it G}_{1}\right)_{X'_{i'}}$ coset can lead to a ``mass"
term; i.e. a local time contribution, after the transformation is
applied. This shall be used in next section to obtain the asymptotic end-to-end
distance of a weakly SARW on a hierarchical lattice, $d=4$, heuristically.
Similarly, the case $n'^{*}_{j'}=3$, for some
$j'\neq i'$ and $0\leq j'\leq k$, shows how a different
realization on the ${\it G}_{1}$ cosets of the weakly SARW can render, a pair
of double intersections of random walks inside the $\left({\it G}_{1}\right)
_{X'_{j'}}$ coset, to the triple intersection of renormalized walks in
$X'_{j'}$ after the renormalization-group map is applied. See Figure 2.\
Writing together cases $n'^{*}_{i'}=1,2,3$ we obtain
\begin{equation}
U'(w')=
\prod_{X'_{i'}\in{\it G}}
\end{equation}
$$
e^{
-\left(L^{\beta}\xi+\xi'_{2}\right)
\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}-
\lambda'
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1} } })
=w(t'_{j'_{\alpha_{1}} })) } }\times
$$
$$
\left\{1+\eta\lambda^{2}
\sum_{
\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'< k_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}',
k_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
t'_{k_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}})
=w(t'_{k'_{\alpha_{1}}}))}
\right\}+r'.
$$
where $r'\sim O(\lambda^{3})$ bounds the cosets where
$n'^{\ast}_{i'}\geq 4$, for some
$0\leq i'\leq k$. Note that $\lambda^3$ is the leading
contribution to $r'$ if
$n'^{\ast }_{i'}= 4$. As $n'^{\ast }_{i'}$
increases the leading contribution decreases. Here
$$
\xi'=\xi'_{2}+L^{\beta}\xi
$$
$$
\eta'_1=\eta\lambda^2\;\;\;\;\;\;\;
\lambda'=L^{(2\beta-d)}\lambda-\beta'\lambda^2+O(\lambda^3)
$$
We apply once more the renormalization-group map to eq(44). In sake of
clarity let us supress primes in eq(44), so the primed terms always
correspond to the renormalized ones. From the initial hypothesis, approaching
the three walks intersection factor in eq(44) to an exponential, factorize
this as done in eq(33), then expanding the exponential up to the first order,
i.e. $\sim \left(\eta\lambda^2\right)$;
from the result in theorem 3, eq(36), eq(39) and eq(41), follows
\begin{equation}
U'(w')=
\end{equation}
$$
\prod_{X'_{i'}\in{\it G}'}
e^{
-\xi'
\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}-
\lambda'
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})
=w(t'_{j'_{\alpha_{1}}}))}}\times
$$
$$
\left\{1+\eta'_{1}
\sum_{
\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'< k_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}',
k_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
t'_{k_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}})
=w(t'_{k'_{\alpha_{1}}}))}
\right.
$$
$$
+\left.
\eta'_{2}
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})
=w(t'_{j'_{\alpha_{1}}}))}
+\eta'_{3}\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}\right\}+r'.
$$
where $r'\sim O(\lambda^3)$ and
$$
\xi'_{2}= \gamma_{1}\lambda-\gamma_{2}\lambda^2
+O(\lambda^3),
$$
$$
\xi'=L^{\beta}\xi+\xi'_2,
$$
$$
\lambda'=L^{(2\beta -d)}\lambda-\beta_{1}\lambda^2+
O(\lambda^3);
\;\;\;
\eta'_{1}=\eta_{1}A+\eta\lambda^{2}
$$
$$
\eta'_{2}=\eta_{1}B,
\;\;\;\; \eta'_{3}=\eta_{1}C
$$
In eq(45) $A$, $B$ and $C$ are respectively given as follows;
\begin{equation}
A=\left\langle
\sum_{X_i\in ({\it G}_1)_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}< k_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},k_{\alpha_{1}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
t_{k_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})
=w(t_{k_{\alpha_{1}}}))}
\right. \times
\end{equation}
$$
\times\left.
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}},k_{\alpha_{1}}\right)
\in \left(i'_{\alpha_{1}},j'_{\alpha_{1}},k'_{\alpha_{1}}\right)
\right)
\right\rangle ^{l.c.}_{w'},
$$
\begin{equation}
B=\left\langle
\sum_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}< k_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},k_{\alpha_{1}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
t_{k_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})
=w(t_{k_{\alpha_{1}}}))}
\right. \times
\end{equation}
$$
\times\left.
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}},k_{\alpha_{1}}\right)
\in \left(i'_{\alpha_{1}}<j'_{\alpha_{1}}\right)
\right)
\right\rangle ^{l.c.}_{w'} \;\;{\mbox and}
$$
\begin{equation}
C=\left\langle
\sum_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}<j_{\alpha_{1}}< k_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},k_{\alpha_{1}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
t_{k_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})
=w(t_{k_{\alpha_{1}}}))}
\right. \times
\end{equation}
$$
\times\left.
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}},k_{\alpha_{1}}\right)
\in \left(i'_{\alpha_{1}}\right)
\right)
\right\rangle ^{l.c.}_{w'}.
$$
$A$, $B$ and $C$ are explained in Figure 3.\
Note that, to get the result in eq(45), we have studied the
trajectory, under the renormalization-group map R, of
\begin{equation}
\left\{
\prod_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}\;
e^{
\eta'_1
\sum_{
\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}< k_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
k_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
t_{k_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})
=w(t_{k_{\alpha_{1}}}))}
}
\right\}.
\end{equation}
When we apply the renormalization-group map to eq(49) we end up with the
contributions in $A,B$ and $C$. In other words, we use the same procedure
above developed for the double (self-)intersection of random walks but for
the triple (self-)intersection of random walks up to order $\lambda^{2}$,
this corresponds to the first term in the corresponding Taylor
series expansion.\
Applying the renormalization-group map to eq(45) we obtain eq(22) (eq(44) and
eq(45) are particular cases of eq(22)). Then, the proof follows from
induction. We apply the renormalization-group map to $U'(w')$ (eq(22)).
Recall that we supress primes in eq(22), thus primed terms are the
renormalized ones. From our original hypotheses and due to the hierarchical
structure of the lattice;
\begin{equation}
U'(w')=
\end{equation}
$$
\prod_{X'_{i'}\in{\it G}}
\left\langle
\prod_{X_{i}\in ({\it G_1})_{X'_{i'}}}
e^{
-\xi
\sum_{i_{\alpha_{1}}\in J_{X_{i}}}
t_{i_{\alpha_{1}}}}
\right\rangle^{l.c}_{w'}\times
$$\
$$
\times
\left\langle
\prod_{X_{i}\in ({\it G_1})_{X'_{i'}}}
e^{
-\lambda
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}}
\right\rangle^{l.c}_{w'}\times
$$\
$$
\left\langle1+\eta_{1}
\sum_{X_{i}\in ({\it G_1})_{X'_{i'}}}
\sum_{
\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}< k_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
k_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
t_{k_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})
=w(t_{k_{\alpha_{1}}}))}+
\right.
$$\
$$
+\eta_{2}
\sum_{X_{i}\in ({\it G_1})_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}+
$$
$$
\left.
+\eta_{3}
\sum_{X_{i}\in ({\it G_1})_{X'_{i'}}}
\sum_{i_{\alpha_{1}}\in J_{X_{i}}}
t_{i_{\alpha_{1}}}\right\rangle^{l.c}_{w'}
+r'.
$$
Assuming ${n^{\ast}}'_{i'}=1,2,3$ and carrying out a Taylor
series expansion up to order $\lambda^{2}$, it is straightforward to
prove that eq(50) leads to eq(22), provided we use the same bookkeeping
device above explained. We just need
to apply theorem 3, eq(36), eq(39), eq(41) and
\begin{equation}
\left\langle
\prod_{X_{i}\in ({\it G_1})_{X'_{i'}}}
exp(
\eta_{1}
\sum_{
\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}< k_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
k_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
t_{k_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})
=w(t_{k_{\alpha_{1}}}))}+
\right.
\end{equation}
$$
\left.+
\eta_{2}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
+\eta_{3}
\sum_{i_{\alpha_{1}}\in J_{X_{i}}}
t_{i_{\alpha_{1}}})\right\rangle^{l.c}_{w'}=
$$
$$
\prod_{\left(n'^{*}_{i'}\right)^{n'}_{i'=0}}
\left\langle
\prod_{X_{i}\in ({\it G_1})_{X'_{i'}}}
e^{
\eta_{3}
\sum_{i_{\alpha_{1}}\in J_{X_{i}}}
t_{i_{\alpha_{1}}}}
\right\rangle^{l.c.}_{w'_{n'^{*}_{i'}}}
$$
$$
\times
\left\langle
\prod_{X_{i}\in ({\it G_1})_{X'_{i'}}}
e^{\eta_{1}
\sum_{
\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}< k_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
k_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
t_{k_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})
=w(t_{k_{\alpha_{1}}}))}}
\right\rangle^{l.c}_{w'_{{n^{*}}'_{i'}}}
$$
$$
\times
\left\langle
\prod_{X_{i}\in ({\it G_1})_{X'_{i'}}}
e^{
\eta_{2}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}}
\right\rangle^{l.c.}_{w'_{{n^{*}}'_{i'}}}
=
$$
$$
\left\{1+A\eta_{1}
\sum_{
\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'< k_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}',
k_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
t'_{k_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}})
=w(t'_{k'_{\alpha_{1}}}))}
+
\right.
$$
$$
+\eta'_2
\sum_
{\stackrel{i_{\alpha_{1}}'< j_{\alpha_{1}}'}
{\left\{i_{\alpha_{1}}',j_{\alpha_{1}}'\right\}\in J_{X'_{i'}}}}
t'_{i_{\alpha_{1}}'}t'_{j_{\alpha_{1}}'}
{\bf 1}_{(w(t'_{i'_{\alpha_{1}}})=w(t'_{j'_{\alpha_{1}}}))}
\left.
+\eta'_{3}\sum_{i_{\alpha_{1}}'\in J_{X'_{i'}}}
t'_{i_{\alpha_{1}}'}\right\}+r'_{3}
$$
where $A$, $\eta'_{2}$ and $\eta'_{3}$ are given by Figure 3, eq(27), and
eq(28); $r'\sim O(\lambda^{3})$. Note that $A\eta_1$ in eq(51) plus the
contribution from eq(41) defines $\eta'_{1}$ as done in eq(26). Eq(51) is
written up to $O(\lambda^2)$ terms.\
Q.E.D.\
We claim theorem 4 is the space-time renormalization-group trajectory of the
weakly SARW energy interaction studied by Brydges, Evans and Imbrie \cite{Ev},
provided $\beta=2$ and $d=4$. In reference [5] the trajectory of a
$\lambda\phi^{4}$ superalgebra valued interaction was studied (this can be
understood in terms of intersection of random walks due to the Mc Kane,
Parisi, Sourlas Theorem \cite{Mk}) using a field-theoretical version of the
renormalization-group map. The field theory is defined on the same
hierarchical lattice we are studying here. In this paper, we provide exact
probabilistic expressions for $\lambda'$ and $\xi'$ (which are not given in
reference [5]), these are crucial to propose an heuristic proof for the
asymptotic behavior of the end-to-end distance of a weakly SARW. To do so we
just need to calculate $\gamma_{1}$ from eq (37) and $\beta_{1}$ from eq(40).\
Finally, we summarize important features of our method; \newline
a) the conditional expectation of $U(w)$ can be approached in terms of the
product of conditional expectations.\
b) We take into account only linear contributions to conditional expectations
for probabilities on $\Lambda_{n}$.\
c) Formal Taylor series expansions are introduced.\
d) We assume our model to be such that each step the renormalization-group
map is applied, the number of times the renormalized walk visits any site of
the lattice is 1, 2 and 3 at least once (i.e. a fixed and not totally
arbitrary topology for the renormalized random walk). See Figure 2.
e) Finally, we take advantage of the hierarchical structure of the lattice.
Since ${\it G}'={\it G}/{\it G}_{1}\approx {\it G}$ and the
map is local, the renormalization-group transformation descends to the
study of walks in the ${\it G}_1$ cosets.\
{}From all of these, we obtain, after applying the
renormalization-group map, the fixed form for a weakly SARW that penalizes,
roughly speaking, the (self-)intersection of two random walks by a factor
($e^{-\lambda}$, $\lambda >0$ and small). Furthermore, this fixed form is the
random walk version (for $d=4$, $\beta=2$) of the one obtained from a
field-theoretical renormalization-group map for a $\lambda \phi^{4}$ model
recently reported by Brydges, Evans and Imbrie \cite{Ev}. We obtain an exact
probabilistic expression for the parameters that appear in the flow of the
interaction factor $\lambda$ which is not given in reference [5]. This shall
be used in next section for the heuristic study of the asymptotic behavior of
the end-to-end distance for a weakly SARW model that punishes the
(self-)intersection of two random walks.
\section{Asymptotic end-to-end distance of a weakly SARW on a hierarchical
lattice in dimension four. An heuristic example as a testing ground.}
The process of renormalizing the lattice is completed by reducing all
dimensions of the new lattice by a factor $L$ each step the
renormalization-group map is applied so we end up with exactly the same
lattice we start with. For a diffusive simple random walk model we reduce
waiting times
by $L^{2}$ each step we apply the renormalization-group transformation.
Moreover, when we iterate probabilities, the end-to-end distance shrinks by
a factor $L$ at each interaction, because in renormalizing the lattice we
divide every length, including the end-to-end distance, by $L$ \cite{Ma}
\cite{Bi}.
{}From this viewpoint we intend to understand, heuristically, the asymptotic
end-to-end distance of a weakly SARW on a hierarchical lattice in $d=4$,
thereby providing a new probabilistic meaning to this magnitude.\
For weakly SARW, we generalize the standard scaling factor for local times of
the renormalization transformation above as described by including, up to
$O(\lambda)$, the contribution of the self-repulsion term to renormalized
local times. Namely, from the renormalization-group map on weakly SARW,
renormalized local times are generated from the interaction. In the field
theoretical approach this corresponds to generating mass. Equivalently, we
can say that the interaction kills the process at a specific rate. If we take
into account only $O(\lambda)$ contributions to this and follow standard
thinking, the well known asymptotic end-to-end distance for the weakly SARW
in $d=4$ follows. By including higher order contributions in $\lambda$ to
renormalized local times (as we have already shown this is not the case for
weakly SARW on the hierarchical lattice, because these contributions are no
significant), and/or different dimension for the lattice, the functional form
of the end-to-end distance changes drastically. Moreover, from our method,
the exponent of the logarithmic correction involved is expressed in terms of
conditional expectations for random walks on the lattice, that upon
calculation, give the well known exponent. In Figure 4 the contributions to
the scaling factor proposed for the weakly SARW used to explain the
asymptotic end-to-end distance in the hierarchical lattice are depicted.\
We remark that the proposition presented in this section involves heuristic
considerations in order to understand, from a probabilistic real-space
viewpoint, the asymptotic end-to-end distance for the weakly SARW on a
hierarchical lattice in $d=4$. This has already been conjectured heuristically
before by other means. Recently a rigorous proof has been given, provided
properties of the Green function are known, in the field theoretical approach
\cite{Im}. Our proposition is anyway presented as a testing ground for
our method, and for giving probabilistic meaning to the exponent involved
in the logarithmic correction. Once the method shows to be useful for
explaining well known results (at least heurstically), we shall apply
this on more complicated cases, for example kinetically growing measure
model. These are renormalizable, in the field theoretical limit, only for
particular cases. In the process of taking the continuum limit of these
discrete models some memory is lost. Our method is suitable of being applied
on the discrete models. This can be done both, heuristically and rigorously.\
A final remark before introducing the main point of this section is about
the finiteness of moments for random walks on a hierarchical lattice. The
end-to-end distance for the weakly SARW, d=4, on the hierarchical lattice, is
independent of the moment used to obtain it, as should be, provided this is
finite. Let $\left\langle w^{\alpha}(T)\right\rangle$ be an $\alpha$-moment
of the random walk, it is known that the only finite moments for diffusive
random walks on a hierarchical lattice are $0<\alpha<2$ \cite{Im}. This range
of $\alpha$ values is used to obtain the end-to-end distance in the
following \newline
{\bf Proposition. For d=4, up to $O(\lambda)$, the generated renormalized local
times (mass for the field or killing rate for the process), from applying the
renormalization-group map on the interaction, is such that the asymptotic
behavior of the end-to-end distance for a weakly SARW that penalizes
the intersection of two random walks is $T^{1/2}log^{1/8}T$ as
T tends to infinity}.\
{\bf Proof}. After applying $(p)$ times the renormalization-group
transformation on $\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha}$ we
have
\begin{equation}
\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha}=
\frac{\left\langle w^{\alpha}(1)\right\rangle^{1/\alpha (0)}}{L^{p}}
\end{equation}
where we have chosen a system of units such that, for $p=0$, $T=1$. Hereafter,
$\left\langle w^{\alpha}(1)\right\rangle^{1/\alpha (0)}$=D, constant. Here,
we are following the standard procedure for scaling length type magnitudes
\cite{Bi} ( i.e. $\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha'}$=
$\frac{\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha}}{L}$). Moreover, by
$\left\langle w^{\alpha}(1)\right\rangle^{1/\alpha}$ we mean
$\left\langle w^{\alpha}(1)\right\rangle^{1/\alpha (p)}$. Since in
renormalizing the lattice we divide every length, including the end-to-end
distance by $L$, then, upon p iterations, eq(52) follows. This is exactly
what is done in scaling correlation lengths but used here on the end-to-end
distance, both length type magnitudes. So eq(52) becomes
\begin{equation}
\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha}=L^{-p}D
\end{equation}
On the other hand, from what we stated in theorem 5 we know that
\begin{equation}
T=\frac{1}
{L^{2p}\prod^{p}_{i=1}(1+\gamma^{\ast}_{1}\lambda^{(i)})}\;\;
\mbox{ ,where}
\end{equation}
$\gamma^*_1$=$\gamma_1/L^2$ and by $T$ we mean $T^{(p)}$. Here we have included
up to $O(\lambda)$ contributions to renormalized local times for scaling the
running time of the process. In this scaling factor of the renormalization
transformation, the $O(\lambda)$ contribution comes from the first term in
right-hand side of eq(24). See Figure 4.\
{}From eq(54) and eq(53) follows
\begin{equation}
\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha}=DT^{1/2}
\left(\prod^{p}_{i=1}(1+\gamma^{\ast}_{1}\lambda^{(i)})\right)^{1/2}
\end{equation}
or
\begin{equation}
\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha} \sim
DT^{1/2}\left(e^{\gamma^{\ast}_{1}\sum^{(p)}_{i=1}
\lambda^{(i)}}\right)^{1/2}.
\end{equation}
For $\beta=2$, $d=4$ and up to order $(\lambda^{(i)})^2$, follows
\begin{equation}
\lambda^{(i+1)}=\lambda^{(i)}-\beta_{1}(\lambda^{(i)})^{2}.
\end{equation}
Introducing the solution of the eq(57) recursion into eq(56) this becomes
\begin{equation}
\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha} \sim DT^{1/2}
e^{\frac{\gamma^{\ast}_{1}}{2\beta_{1}}ln p}
\end{equation}
or
\begin{equation}
\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha} \sim DT^{1/2}
(p)^{\frac{\gamma^{\ast}_{1}}{2\beta_{1}}}
\end{equation}
In eq(59) we have assumed p to be large enough so
$\lambda^{-1}<<\beta_{1}(p)$. Taking the asymptotic limit we rewrite eq(59)
as
\begin{equation}
\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha} \sim DT^{1/2}
log^{\frac{\gamma^{\ast}_{1}}{2\beta_{1}}}T,
\end{equation}
which is the asymptotic behavior of the end-to-end distance.\
It only remains to know the value of
$\left(\frac{\gamma^{\ast}_{1}}{\beta_{1}}\right)$. Actually we can
calculate $\gamma^{\ast}_{1}$ and $\beta_{1}$ from their definitions.\
Let us start with $\gamma_{1}$, from eq(37) we obtain
\begin{equation}
\gamma_{1}=\left\{
\left(
\prod_{i'_{\alpha_{1}}\in J_{X'_{i'}}}
e^{q_{1}(L^d-1)r(L^{\beta}t'_{i'_{\alpha_{1}}})}
\right)^{-1}\times
\right.
\end{equation}
$$
\times
\left(
\sum_{n_{i'}}\;L^{d}\int\prod_{i\in I_{X'_{i'}}}
\;dt_{i}\;\prod_{i'_{\alpha_{1}}\in J_{X'_{i'}}}
\;\delta
(\sum^{m_{i'_{\alpha_{1}}}}_{{i}=m_{i'_{\alpha_{1}}-1}+1}\;\;t_{i}-
L^{\beta }t'_{i'_{\alpha_{1}}})\times \right.
$$
$$
\times
\left(
\begin{array}{c}
n_{i'}+1 \\
2
\end{array}
\right)
(q_{1})^{\left(n_{i'}-1\right)}r^{\left(n_{i'}\right)}
(L^{d}-1)\times...\times
(L^{d}-(n_{i'}-1))
\times
$$
$$
\left.
\times
\sum_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}}\right\}\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}}))}
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}}\right)\in i'_{\alpha_{1}}
\right)
\right\}^{l.c.}
$$
Hereafter, we assume $L^{d}>>n_{i'}$, so
$$
\left(L^{d}-(n_{i'}-1)\right)\sim \left(L^{d}-1\right),
$$
i.e. the number of points inside each ${\it G}_{1}$ coset is larger than the
number
of steps the walk $w'$ spends inside each L-block. Thus, the numerator of
eq(61) can be written as
\begin{equation}
\frac{\left(\sum_{i'_{\alpha_{1}}\in J_{X'_{i'}}} L^{\beta}t'
_{i'_{\alpha_{1}}}\right)^{2} }
{2q_{1}(L^{d}-1)}
\prod_{i'_{\alpha_{1}}\in J_{X'_{i'}}}
\int^{1}_{0}\sum_{n_{i'}}\left(n_{i'}+1\right)
\left(n_{i'}\right)^{2}\times
\end{equation}
$$
\times
\left(n_{i'}-1\right)
\frac{ \left( rq_{1}(L^{d}-1)(L^{\beta}t'_{i'_{\alpha_{1} } })
(1-t-t^{\ast})\right)^{n_{i'}}}
{n_{i' }! }
tdtt^{\ast}dt^{\ast}.
$$
where we have taken $(1-t-t^{\ast})^{n_{i'}-2}$ $\sim $
$(1-t-t^{\ast})^{n_{i'}}$. To obtain an asymptotic estimate of eq(61), we
assume that the following holds
$$
\left(n_{i'}+1\right)\left(n_{i'}\right)^{2}
\left(n_{i'}-1\right)\sim
\frac{\left(n_{i'}\right)!}
{\left(n_{i'}-4\right)!}
$$
Althought from this follows $n_{i'}$ chosen to be large, we certainly assume
finite local times after the renormalization-group transformation is
applied.\
Substituting eq(62) in eq(61), with a jumping rate $r$ such that\newline
$rq_{1}(L^{d}-1)\sim 1$ (as done in reference[5]) and for $\beta =2$
we obtain, in the asymptotic limit, $\gamma^{\ast}_{1}\sim 8$ provided
$rq_{1}(L^{d}-1)\sim 1$.\
To calculate $\beta_{1}$ we use eq(40).
\begin{equation}
\beta_{1}=\frac{1}{2}
\left\{
\left(
\prod_{i'_{\alpha_{1}}\in J_{X'_{i'}}}
e^{q_{1}(L^d-1)r(L^{\beta}t'_{i'_{\alpha_{1}}})}
\right)^{-1}\times
\right.
\end{equation}
$$
\times
\left(L^{d}
\sum_{(n_{i'_{a}})}\;
\sum_{(n_{i'_{b}})}\;
\int
\prod_{i_{a}\in I_{X'_{i'}}}
\;dt_{i_{a}}\;
\prod_{i_{b}\in I_{X'_{i'}}}
\;dt_{i_{b}}\;
\right.
\prod_{i'_{\alpha_{1}}\in J_{X'_{i'}}}
\;\delta
(\sum^{m_{i'_{\alpha_{1}}}}_{{i_{a}}=m_{i'_{\alpha_{1}}-1}+1}\;\;
t_{i_{a}}-
L^{\beta }t'_{i'_{\alpha_{1}}})\times
$$
$$
\times
\prod_{j'_{\alpha_{1}}\in J_{X'_{i'}}}
\;\delta
(\sum^{m_{j'_{\alpha_{1}}}}_{{i_{b}}=m_{j'_{\alpha_{1}}-1}+1}\;\;
t_{i_{b}}-
L^{\beta }t'_{j'_{\alpha_{1}}})
\left(
\begin{array}{c}
n_{i'_{a}}+1 \\
2
\end{array}
\right)
\left(
\begin{array}{c}
n_{i'_{b}}+1 \\
2
\end{array}
\right)
(q_{1})^{\left(n_{i'_{a}}\right)}r^{\left(n_{i'_{a}}\right)}\times
$$
$$
\times
(q_{1})^{\left(n_{i'_{b}}-1\right)}r^{\left(n_{i'_{b}}\right)}
(L^{d}-1)\times...\times
(L^{d}-(n_{i'_{a}}+n_{i'_{b}}-1))
\times
$$
$$
\times
\sum_{X_{i}\in ({\it G}_{1})_{X'_{i'}}}
\sum_
{\stackrel{i_{\alpha_{1}}< j_{\alpha_{1}}<
i_{\alpha_{2}}< j_{\alpha_{2}}}
{\left\{i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right\}
\in J_{X_{i}}}}
t_{i_{\alpha_{1}}}t_{j_{\alpha_{1}}}
{\bf 1}_{(w(t_{i_{\alpha_{1}}})=w(t_{j_{\alpha_{1}}})|
i_{\alpha_{1}},j_{\alpha_{1}}\leq n_{i'_{a}})}
\times
$$
$$
\times
\left.
\left.
t_{i_{\alpha_{2}}}t_{j_{\alpha_{2}}}
{\bf 1}_{(w(t_{i_{\alpha_{2}}})=w(t_{j_{\alpha_{2}}})|
i_{\alpha_{1}},j_{\alpha_{1}}\leq n_{i'_{b}})}
\chi\left(\left(i_{\alpha_{1}},j_{\alpha_{1}},
i_{\alpha_{2}},j_{\alpha_{2}}\right)\in \left(i'_{\alpha_{1}}<
j'_{\alpha_{1}}\right)
\right)
\right)
\right\}^{l.c.}
$$
As we did for the calculation of $\gamma_{1}$, we assume $L^{d}>>n_{i'}$, so
\begin{equation}
\left(L^{d}-(n_{i'_{a}}-1)\right)\sim (L^{d}-1)\;\mbox{ and }\;
\left(L^{d}-(n_{i'_{b}})\right)\sim (L^{d}-1).
\end{equation}
Furthermore, we choose
$
n_{i'_{b}}\sim n_{i'_{b}}-1,
$
and approximations in eq(62) to hold for both
$n_{i'_{a}}$ and $n_{i'_{b}}$ with a jumping rate $r$ such that
$rq_{1}(L^{d}-1)\sim 1$. For $\beta=2$ in the asymptotic limit, we
obtain $\beta_{1}\sim 32$ provided $rq_{1}(L^{d}-1)\sim 1$.\
Finally eq(60) becomes
\begin{equation}
\left\langle
w^{\alpha}(T)\right\rangle^{1/\alpha}
\sim (DT)^{1/2}log^{1/8}T
\end{equation}
Q.E.D.\
Note that $d=4$ is the only choice that renders eq(25) (for $\beta=2$)
to a recursion as simple as eq(57) provided that $rq_{1}(L^{d}-1)\sim 1$.\
We want to remark that the heuristic study of the asymptotic end-to-end
distance of a weakly SARW on a hierarchical lattice, $d=4$, is independent of
hypotesis a) in the summary of former section. This is because we could have
obtained the $O(\lambda)$ contribution to renormalized local times without
introducing initial mass into the process.\
\section{Summary}
In this paper we present a real space renormalization-group map,
on the space of probabilities, to study weakly SARW that penalizes the
(self-)intersection of two random walks for a hierarchical lattice, in
dimension
four. This hierarchical lattice has been labeled by elements of a countable,
abelian group ${\it G}$. For any random function $F(w)$ on ${\it G}$ of the
form described in Section 3, i.e. factorizable on the lattice, ( see eq(17)
and eq(21) for examples of suitable $P(w)$) we can descend
from the study of the space of walks on the whole lattice to the trajectory
in the contracting ${\it G}_{1}$ cosets. Then we show how the L\'{e}vy
process studied in reference [5] is a particular case of the processes that
are (or flow to) fixed points of the renormalization-group map. We apply the
renormalization-group map on some random walk models with configurational
measure, working out explicitly the weakly SARW case. An heuristic proof of
the end-to-end distance for a weakly SARW on a hierarchical lattice is
derived. This gives a new probabilistic meaning to the exponent of the
logarithmic correction.\
In Section 4 we study a weakly SARW that penalizes the (self-)intersection of
two random walks.
The weakly SARW probability studied, involves a factor linear in local times,
i.e. a random walk representation of a field-theoretical
gaussian component that adds to the corresponding term produced for the
renormalization-group map applied on the weakly SARW. We
show how this probability flows to a fixed form (the random walk version of
the
field-theoretical result given in reference [5]) relying on;\
a) An hypothesis that assumes we can approach the expectation of the
interaction energy in terms of the product of expectations for each of its
factors, conditioned to applying the renormalization-group map.\
b) The hierarchical metric space used to label the lattice that allows,
for $F(w)$ of the form described in Section 3 (f.e. eq(17) and eq(21)), the
factorization in terms of the quotient group ${\it G}/{\it G}_{1}$ and the
image of the renormalization-group map on the cosets ${\it G}_{1}$, each step
the renormalization-group is applied.\
c) A class of realizations of the model such that, each step we apply
the renormalization-group transformation, the renormalized fixed walk
visits $1, 2, 3$ times different sites in the lattice
${\it G}/{\it G}_{1}$, at least once. Other realizations might not allow us
to study the flow of the (self-)intersecting coefficients that are interesting
for us.\
d) A formal Taylor series expansion in $\lambda$ from which, upon
renormalization, we use only the linear contributions. \
Our result improves the field-theoretical approach \cite{Ev} by obtaining an
exact expression for the parameters that appear in the flow of $\lambda$ and
$\xi$. This is a crucial feature used to obtain heuristically the asymptotic
behavior of the end-to-end distance for a weakly SARW that penalizes the
(self-)intersection of two random walks. Furthermore, the method here
presented is full of physical intuition and suitable of being applied to
discrete kinetically growing measure models.\
Following standard thinking we shrink all space and time magnitudes each step
the map is applied. We shrink time taking into account $O(\lambda)$
contributions to renormalized local times, generated by applying the
renormalization-group map to the weakly SARW that penalizes (self-)
intersections of walks. Length type magnitudes are shrunk as usual. We
present this, as a possible origin for the expression
$\left\langle w^{\alpha}(T)\right\rangle^{1/\alpha}\sim (DT)^{1/2}log^{1/8}T$
as T
tends to infinity, in $d=4$, for a weakly SARW that penalizes the
(self-)intersection of two random walks on a hierarchical lattice.\
\vspace*{10mm}
\section{Acknowledgments}
This work was partially supported by CONACYT REF 4336E9406, M\'{e}xico.
I wish to thank S.N. Evans, Department of Statistics, UC Berkeley, for
helpful discussion and ITD, UC Davis for its hospitality.
\baselineskip0.6cm
|
\section{Introduction}
\label{sec:Introduction}
The much heralded convergence of the Standard Model gauge couplings in
supersymmetric Grand Unified Theories (GUTs) \cite{EKN}, is continually being
challenged by ever more precise LEP measurements of the gauge couplings. It was
realized early on that the effect of the GUT particles responsible for the
onset of the unified theory, was not negligible \cite{GUTthresholds}. However,
because of the presumed great uncertainty in the GUT physics, such discussions
have been largely carried out in the context of the minimal $SU(5)$
supergravity model \cite{CAN}. Central to the study of these issues is the
technical point of how exactly these GUT (or lighter) particles decouple at
scales below their masses. Recent investigations \cite{smooth} reveal that a
``smooth" decoupling of particles leads to noticeable differences from the
step-function approximation. Moreover, these new effects coupled with the
latest LEP data on $\sin^2\theta_W$ and the determination of the top-quark
mass, lead to a greatly increased prediction for $\alpha_3(M_Z)$
\cite{Clavelli,Bagger,LP}, strongly suggesting that minimal $SU(5)$ GUT
thresholds are unable to bring the $\alpha_3(M_Z)$ prediction down to the
experimentally acceptable range \cite{Clavelli}. This impasse may be resolved
with a significant contribution from Planck-scale non-renormalizable operators
\cite{NROs,LP}, although such effects call into question the whole
field-theoretical approximation to the gauge coupling unification problem.
Even if Planck-scale physics can resolve the present $\alpha_3$ discrepancy
in minimal $SU(5)$, this GUT model suffers from a well known fine-tuning
\cite{DG} regarding the solution of the doublet-triplet splitting problem of
the Higgs pentaplets. At least three solutions to this problem (all involving
non-minimal GUT models) have been proposed: the missing-partner mechanism
\cite{MPM}, the sliding-singlet mechanism \cite{SSM}, and the
pseudo-goldstone-boson mechanism \cite{PGBM}. In the sliding-singlet mechanism
radiative corrections destroy the gauge hierarchy \cite{Lahanas}, whereas an
additional global or local $SU(6)$ symmetry is required in the
pseudo-goldstone-boson mechanism. It is very suggestive that the same
investigations that uncover the $\alpha_3$ discrepancy in minimal $SU(5)$, also
show that in the so-called Missing Doublet Model (MDM) \cite{MPM}, which has as
its central component the missing-partner mechanism, the $\alpha_3$ prediction
is decreased to acceptable values \cite{Bagger,Clavelli}. As we discuss below,
some variants of the missing-doublet model \cite{MNTY2,HMTY} also solve the
problematic situation with dimension-five proton decay operators in minimal
$SU(5)$, which require the Higgs triplet mass ($M_{H_3}$) to exceed the GUT
scale and the supersymmetric spectrum to be tuned in specific ways
\cite{AN,HMY}, especially when cosmological constraints are simultaneously
enforced \cite{LNP}. In fact, an updated analysis has recently shown
\cite{HMTYpd} that the upper bound on the Higgs triplet mass from unification
constraints ({\em i.e.},
$M_{H_3}\propto e^{-5\pi/3\alpha_3},\ \alpha_3<\alpha_3^{\rm max}$), and the
corresponding lower bound from proton decay constraints ({\em i.e.}, $\tau_p\propto
M^2_{H_3},\ \tau_p>\tau^{\rm min}_p$) are very close to each
other, leaving only a small window of allowed parameter space in minimal
$SU(5)$. Note also that, because of the rather large representations introduced
in the MDM (\r{75},\r{50},\rb{50}), it is necessary to assign some of these
Planck-size masses, in order to avoid the onset of a strongly-interacting GUT
below the Planck scale \cite{Clavelli,HMTY}. Thus, Planck-scale physics is
again unavoidable in this more realistic version of $SU(5)$ GUTs.
$SO(10)$ GUTs \cite{GFM,GN} have also received a great deal of attention lately
\cite{BB,Mohapatra,Raby,inspired}, with interesting successes in the area of
quark and lepton masses and mixings, although the $\tan\beta={\cal O}(50)$
prediction requires fine-tuning of the supersymmetric spectrum \cite{NR,Carena}
to reconcile it with radiative electroweak breaking. Assuming universal soft
supersymmetry breaking, the further constraints from $B(b\rightarrow s\gamma)$ and
cosmology strongly disfavor the model \cite{BOP}. However, most of these
shortcomings are overcome when one allows certain classes of non-universal
scalar masses \cite{OP,BOP}. More to the point, the successes of $SO(10)$ rely
on the existence of certain non-renormalizable operators (as originally
suggested in Ref.~\cite{EG+NS}) that are presumed to be obtained from a
string-derived model at the Planck scale. Despite initial claims \cite{Lykken},
no {\em consistent} $SO(10)$ GUT string model has been derived in the context
of free-fermionic strings \cite{Cleaver}. However, these failed attempts have
been enough to fuel a series of ``string-inspired" $SO(10)$ GUT models
\cite{inspired}, which are limited to certain type and number of
representations (those allowed by the level-two Kac-Moody
construction\footnote{At level two, the allowed unitary massless
representations are \r{1},\r{10},\r{16},\rb{16},\r{45},\r{54} \cite{ELN}. A
string model containing the \r{126},\rb{126} representations used in
traditional $SO(10)$ model building requires an unlikely level-five
construction \cite{ELN}.}), forcing model builders to rely heavily on
postulated effective non-renormalizable operators \cite{inspired}. Level-two
string $SO(10)$ GUT models have been consistently constructed in the context of
symmetric orbifolds \cite{Aldazabal}, but with limited phenomenological
success, especially in dealing with the doublet-triplet splitting problem.
In view of its field-theoretical successes, in this paper we revisit the
missing-doublet model as a well-motivated, realistic contender for a grand
unified model. We first review the original MDM and its features and
shortcomings (Sec.~\ref{sec:observable}). We then propose a simple extension of
the model to naturally suppress dimension-five proton decay operators
(Sec.~\ref{sec:observable}). Our most substantive contribution is to endow this
{\em supergravity} model with a hidden sector containing gauge and matter
degrees of freedom (Sec.~\ref{sec:hidden}). Hidden sector gaugino condensation
triggers supersymmetry breaking which, as we discuss, can be of the desired
magnitude for suitable choices of the hidden gauge group and hidden matter
spectrum. The matter condensates provide a new dynamical intermediate scale
which, via non-renormalizable interactions, generates a low-energy Higgs mixing
term $\mu$. With the introduction of right-handed neutrinos to the model, this
scale also becomes their mass scale, which provides a suitable see-saw spectrum
of neutrino masses (Secs.~\ref{sec:observable},\ref{sec:CBA}). We show that the
model is consistent with gauge coupling unification for experimentally
acceptable values of $\alpha_3(M_Z)$ and that dimension-five proton decay
operators are consistent with present limits even for large values of
$\tan\beta$ (Sec.~\ref{sec:unification}). Also, the out-of-equilibrium decays
of the right-handed neutrinos subsequent to inflation produce a lepton
asymmetry which is re-processed into a baryon asymmetry by strongly-interacting
Standard Model effects (sphalerons) at the electroweak scale
(Sec.~\ref{sec:CBA}). We finally compare the features of this traditional GUT
model with that of the readily string-derivable $SU(5)\times U(1)$ model, and
discuss the prospects of deriving the revamped MDM from string theory
(Sec.~\ref{sec:comparison}). We summarize our conclusions in
Sec.~\ref{sec:conclusions}.
\section{The observable sector}
\label{sec:observable}
The original MDM \cite{MPM} can be described by the following set of
observable sector fields: $\Sigma$ (\r{75}), $\theta$ (\r{50}), $\bar\theta$
(\rb{50}), $h$ (\r{5}), $\bar h$ (\rb{5}), $F_{1,2,3}$ (\r{10}'s),
$\bar f_{1,2,3}$ (\rb{5}'s), interacting via the superpotential
\begin{equation}
W=\coeff{1}{2}M_{75}\,\Tr{\Sigma^2}+\coeff{1}{3}\lambda_{75}\,\Tr{\Sigma^3}
+\lambda_4\,\bar\theta\Sigma h+\lambda_5\,\theta\Sigma\bar h
+M_{50}\,\theta\bar\theta+\lambda^{ij}_2 F_iF_jh
+\lambda^{ij}_1 F_i\bar f_j \bar h\ .
\label{eq:W}
\end{equation}
The expectation value of the \r{75} can be chosen such that the $SU(5)$ gauge
symmetry is broken down to the Standard Model one, in which case the scalar
potential that follows from Eq.~(\ref{eq:W}) gives
$\vev{\Sigma}\sim M_{75}/\lambda_{75}$. The $\bar\theta\Sigma h$,
$\theta\Sigma\bar h$, and $\theta\bar\theta$ terms in $W$ effect the
doublet-triplet mechanism via the mass matrix
\begin{equation}
\bordermatrix{
&\bar h_3&\bar\theta_3\cr
h_3&0&\lambda_4\vev{\Sigma}\cr
\theta_3&\lambda_5\vev{\Sigma}&M_{50}}\ ,
\label{eq:2/3}
\end{equation}
where the subscript `3' indicates the $SU(2)_L$ singlet, $SU(3)_C$ triplet
component of the corresponding $SU(5)$ representation. This matrix clearly
yields massive ($\sim\vev{\Sigma}\sim M_{\rm GUT}$) Higgs triplets ($h_3,\bar
h_3$), whereas the doublets ($h_2,\bar h_2$) remain massless. The $M_{50}$ term
is not required for a successful doublet-triplet splitting. However, it is
introduced in order to give large masses to the many leftover components of the
\r{50},\rb{50} representations. The last two terms in W (\ref{eq:W}) provide
the Yukawa matrices for the Standard Model fermions, implying the usual
relations ({\em e.g.}, $\lambda_b=\lambda_\tau$).
Despite the above natural solution to the doublet-triplet splitting problem,
the magnitude of the dimension-five ($d=5$) proton decay operators still needs
to be assessed. The crucial element in this calculation is the effective
$h_3\bar h_3$ mixing term. If in Eq.~(\ref{eq:W}) $M_{50}$ were allowed to
vanish, then there would be no mixing whatever, and the $d=5$ operators would
be negligible. In practice $M_{50}$ cannot vanish, and we are left with two
possibilities: (i) $M_{50}\sim \vev{\Sigma}$, and (ii)
$M_{50}\gg\vev{\Sigma}$. The first case implies an effective Higgs-triplet
mixing term of the same magnitude as in the minimal $SU(5)$ model, and
therefore similar difficulties in suppressing proton decay. However, this case
is not really viable since above the GUT scale the large \r{50},\rb{50}
representations increase the $SU(5)$ beta function so much that the gauge
coupling becomes non-perturbative before reaching the Planck scale
\cite{Clavelli,HMTY}. We are left with the second alternative with $M_{50}\sim
M$, where $M=M_{Pl}/\sqrt{8\pi}\approx10^{18}\,{\rm GeV}$ is the appropriate
gravitational scale. Unfortunately, this choice leads to a see-saw type mass
for the Higgs triplets: $m_{h_3,\bar h_3}\sim \vev{\Sigma}^2/M\sim10^{14}\,{\rm GeV}$,
and effective mixing of the same magnitude, which makes proton decay much too
fast.
Various variants of the MDM have been proposed to deal with the proton decay
problem in a more effective way \cite{MPM,MNTY2,HMTY}. Here we follow the
suggestion in Ref.~\cite{HMTY}, whereby the following additional fields are
introduced: $\theta'$ (\r{50}), $\bar\theta'$ (\rb{50}), $h'$ (\r{5}), $\bar
h'$ (\rb{5}). The superpotential for the model is that in Eq.~(\ref{eq:W}) with
$M_{50}\equiv0$, and supplemented by
\begin{equation}
W'=\lambda_4'\,\bar\theta'\Sigma h'+\lambda_5'\,\theta'\Sigma\bar h'
+M'_{50}\,\theta\bar\theta'+M'_{50}\,\theta'\bar\theta\ ,
\label{eq:W'}
\end{equation}
where we again take $M'_{50}\sim M$. These interactions lead to the following
generalized Higgs-triplet mass matrix
\begin{equation}
\bordermatrix{
&\bar h'_3&\bar\theta_3&\bar h_3&\bar\theta'_3\cr
h_3&0&\lambda_4\vev{\Sigma}&0&0\cr
\theta'_3&\lambda'_5\vev{\Sigma}&M'_{50}&0&0\cr
h'_3&0&0&0&\lambda'_4\vev{\Sigma}\cr
\theta_3&0&0&\lambda_5\vev{\Sigma}&M'_{50}}\ ,
\label{eq:2/3'}
\end{equation}
and effective interactions \cite{HMTY}
\begin{equation}
\left(\lambda_4\lambda'_5{\vev{\Sigma}^2\over M'_{50}}\right)h_3\bar h'_3+
\left(\lambda'_4\lambda_5{\vev{\Sigma}^2\over M'_{50}}\right)h'_3\bar h_3
\equiv M_{H_3}\,h_3\bar h'_3+M_{\bar H_3}\,h'_3\bar h_3\ .
\label{eq:mixings}
\end{equation}
Since there is no effective interaction between $h_3$ and $\bar h_3$ (the
only triplets that interact with the Standard Model fermions), the $d=5$
operator is negligible.
If the superpotential $W+W'$ were the complete model, we would have managed
to make all the non-minimal fields sufficiently heavy or non-interacting.
However, we would have left two pairs of Higgs doublets $h_2,\bar h_2$ and
$h'_2,\bar h_2'$ with no apparent use for the second pair, and if light, with
severe trouble with gauge coupling unification. Let us assume the existence of
a mass term $M_{h'}h'\bar h'$, with no specific origin for $M_{h'}$ for now.
Such a term contains $M_{h'}h'_3\bar h'_3$, which ``hooks up" the two
disconnected pieces in Eq.~(\ref{eq:mixings}) and allows $d=5$ proton decay to
occur, with an operator proportional to
\begin{equation}
{1\over M_{H_{\rm eff}}}\equiv{M_{h'}\over M_{H_3} M_{\bar H_3}}
\sim {M_{h'}\over [\vev{\Sigma}^2/M'_{50}]^2}\ ,
\label{eq:Heff}
\end{equation}
where $M_{H_{\rm eff}}$ is the effective Higgs triplet mass. Since in the
minimal $SU(5)$ model with $M_{H_{\rm eff}}=M_{H_3}\mathrel{\mathpalette\@versim>} 10^{17}\,{\rm GeV}$, the
present experimental bounds on proton decay are satisfied without strong
restrictions on the parameter space \cite{AN,HMY}, we effectively require
$M_{h'}\lsim10^{11}\,{\rm GeV}$.
But where does this intermediate scale come from? It has been suggested that
this scale could be generated dynamically via the breaking of a Peccei-Quinn
symmetry \cite{MSY,HMTY}. A more modern and economical approach to the
generation of intermediate scales, especially in the context of supergravity,
is to consider the condensation of a hidden sector gauge group that triggers
supersymmetry breaking. Non-renormalizable interactions coupling hidden sector
matter fields to observable fields may then naturally generate the
intermediate scale.\footnote{This mechanism is commonly available in string
model building \cite{decisive}.} Specifically, we add to our model
the following superpotential terms\footnote{The apparent asymmetry between
the $h\bar h$ and $h'\bar h'$ couplings may be understood on the basis
of additional local $U(1)$ quantum numbers, which are broken near the Planck
scale and are carried by both hidden and observable sector particles, as is
common in string model building \cite{decisive}. For further symmetry
arguments motivating these choices, see {\em e.g.}, Ref.~\cite{CKN}.}
\begin{equation}
W''= \lambda_7\, h\bar h\,{(T\bar T)^2\over M^3}
+\lambda'_7\, h'\bar h'\,{T\bar T\over M}\ ,
\label{eq:W''}
\end{equation}
where $T\bar T$ is a gauge-singlet hidden-sector composite ({\em e.g.}, \r{4}\rb{4} in
$SU(4)$). When the hidden sector condenses, we generate dynamically two mass
scales:
\begin{equation}
M_{h'}=\lambda'_7\,{\vev{T\bar T}\over M}\,,\qquad
\mu=\lambda_7 \,{\vev{T\bar T}^2\over M^3}\ .
\label{eq:scales}
\end{equation}
Note that for $\vev{T\bar T}/M\sim10^{10}\,{\rm GeV}$, we would obtain for the masses
of the extra pair of doublets $M_{h'}\sim10^{10}\,{\rm GeV}$, and an effective
Higgs-triplet mixing which satisfies proton decay constraints automatically.
We would also obtain dynamically\footnote{This dynamical generation of the
$\mu$ parameter via non-renormalizable interactions is also familiar from
string model-building \cite{decisive,Casasmu,CKN}.} a very desirable Higgs
mixing parameter $\mu\sim100\,{\rm GeV}$. In the next section we explore the hidden
sector of the model with these phenomenological constraints in mind.
One of the main model-building shortcomings of $SU(5)$ GUTs is the
not-so-obvious source of neutrino masses. In fact, neutrino masses can be
introduced by simply adding right-handed neutrino ($SU(5)$ singlet) fields
to the model. To implement the standard see-saw mechanism we introduce
three singlet fields $\nu^c_{1,2,3}$ with the following superpotential
\begin{equation}
W'''=\lambda^{ij}_3\, \bar f_i \nu^c_j h + \lambda^{ij}_6\, \nu^c_i\nu^c_j \,
{T\bar T\over M}\ .
\label{eq:W'''}
\end{equation}
After hidden sector condensation and electroweak symmetry breaking, we
obtain the following see-saw neutrino mass matrix
\begin{equation}
\bordermatrix{&\nu_j&\nu^c_j\cr
\nu_i&0&\lambda_3^{ij}v_2\cr
\nu^c_i&\lambda_3^{ji}v_2&\lambda^{ij}_6\vev{T\bar T}/M}\ ,
\label{eq:see-saw}
\end{equation}
and light neutrino see-saw masses
$m_{\nu}\sim\lambda^3_2 v^2_2/[\lambda_6\vev{T\bar T}/M]$. For simplicity,
in what follows we assume $\lambda^{ij}_6=\lambda^i_6\delta_{ij}$.
With our above desired value of $\vev{T\bar T}/M\sim M_{\nu^c}\sim10^{10}\,{\rm GeV}$,
and $\lambda_3 v_2\sim10\,{\rm GeV}$, typical see-saw light neutrino masses follow,
{\em i.e.}, $m_{\nu_\tau}\sim10\,{\rm eV}$. Further discussion of the consequences of this
see-saw matrix for the light neutrino masses and mixing angles is given in
Sec.~\ref{sec:CBA} below.
\section{The hidden sector}
\label{sec:hidden}
Our supergravity model is endowed with a hidden sector which communicates
with the observable sector via gravitational interactions (or via $U(1)$
gauge interactions broken near the Planck scale). The hidden sector consists
of a hidden gauge group and a set of matter representations, which for
convenience we take to be $SU(N_c)$ with $N_f$ flavors ($T_i,\bar T_i,\ i=1\rightarrow
N_f$) and $N_f<N_c$. This gauge group starts with a gauge coupling $g$ at the
Planck scale ($Q=M$), and becomes strongly interacting at the condensation
scale defined by
\begin{equation}
\Lambda=M e^{8\pi^2/\beta g^2}\ ,
\label{eq:Lambda}
\end{equation}
where the beta function is given by $\beta=-3N_c+N_f$. For simplicity we assume
that all the flavors are ``light", {\em i.e.}, they have masses\footnote{Massless
flavors lead to pathologies ({\em i.e.}, no vacuum), which can nonetheless be remedied
by invoking supersymmetry-breaking masses for the $T_i,\bar T_i$ fields
\cite{Peskin}.} $m\ll\Lambda$. At the condensation scale, the strongly
interacting theory is described in terms of composite ``meson" fields $T_i\bar
T_i$. The dynamics of this system can be obtained from an effective Lagrangian
with the following non-perturbative superpotential \cite{Seiberg}
\begin{equation}
W_{\rm non-pert}=(N_c-N_f)\,{\Lambda^{(3N_c-N_f)/(N_c-N_f)}\over
(\det T\bar T)^{1/(N_c-N_f)}}+\Tr(mT\bar T)\ .
\label{eq:Wnp}
\end{equation}
Minimization of the corresponding scalar potential results in the following
expectation values for the mesons fields $\vev{T\bar T}$ (we work in a diagonal
flavor basis)
\begin{equation}
\vev{T\bar T}=\Lambda^{(3N_c-N_f)/N_c}\,(\det m)^{1/N_c}\, m^{-1}
=\Lambda^3 \left({m\over\Lambda}\right)^{N_f/N_c}{1\over m}
=\Lambda^2\, x^{(N_f/N_c)-1}\ ,
\label{eq:TT}
\end{equation}
where in the last expression we have defined $x\equiv m/\Lambda$, with $x<1$.
Inserting this expectation value in $W_{\rm non-pert}$, we obtain
\begin{equation}
\vev{W}=N_c\, \Lambda^3\, x^{N_f/N_c}\ ,
\label{eq:Wvev}
\end{equation}
where $W$ includes all perturbative and non-perturbative contributions. In a
supergravity theory, the scale of supersymmetry breaking is determined by
the gravitino mass: $m_{3/2}=\vev{e^K\,W}$, where $K$ is the K\"ahler
potential. In simple models $K=\sum\phi_i\phi_i^\dagger$, and thus $\vev{K}=0$.
More complicated forms of $K$ are obtained in string models (where the dilaton
and moduli fields play an important role). For our present purposes, we simply
assume that $\vev{e^K}\sim1$. This implies that $\vev{W}$ is the sole source
of supersymmetry breaking, {\em i.e.},
\begin{equation}
m_{3/2}\sim \vev{W}\sim \left({\Lambda\over M}\right)^3\, x^{N_f/N_c}\, M\ ,
\label{eq:m3/2}
\end{equation}
where we have restored the units in the expression for $m_{3/2}$.
With the results in Eqs.~(\ref{eq:TT}) and (\ref{eq:m3/2}) for $\vev{T\bar T}$
and $m_{3/2}$, we can now investigate the conditions on $N_c$, $N_f$, and $x$
that would yield the desired results: $\vev{T\bar T}/M=10^p\,{\rm GeV}$ and
$m_{3/2}=10^q\,{\rm GeV}$, with $p\sim10$ and $q\sim3$. In terms of $p$ and $q$, we
can solve simultaneously Eqs.~(\ref{eq:TT}), (\ref{eq:m3/2}), and
(\ref{eq:Lambda}), to obtain
\begin{equation}
N_c={p-q\over18-q}\,N_f+{8\pi^2\over g^2}\,\,{\log_{10} e\over18-q}\ ,
\label{eq:Nc}
\end{equation}
and
\begin{equation}
x=10^{2N_c\,({3\over2}p-q-9)/\beta}\ .
\label{eq:x}
\end{equation}
Thus, for a given value of $g$ and $N_f$, we obtain $N_c$ (and thus $\beta$)
from Eq.~(\ref{eq:Nc}). With this value of $N_c$, $x$ is determined from
Eq.~(\ref{eq:x}), and $\Lambda$ from Eq.~(\ref{eq:Lambda}). For the desired
$p=10$ and $q=3$, and with the sensible inputs $N_f=1$ and $g=0.7$, we obtain
$N_c=5$, $x\approx0.01$, and $\Lambda\approx10^{13}\,{\rm GeV}$. That is, an $SU(5)$
hidden gauge group with one light flavor. The general constraints on $N_c$ and
$N_f$ for given values of $g$ are shown in Fig.~\ref{fig:NcNf}, for $q=2\to3$
({\em i.e.}, $m_{3/2}=100\,{\rm GeV}\to1\,{\rm TeV}$) and $p=10$.
We do not address here the calculation of the observable-sector
soft-supersymmetry-breaking scalar and gaugino masses, since these
depend on the specific choices for the K\"ahler function and the gauge
kinetic function, although their overall scale is already determined by
$m_{3/2}$. The ``flat" choice $K=\sum\phi_i\phi^\dagger_j$ leads to the usual
universal scalar masses, but this choice is not unique.
\section{Unification and proton decay}
\label{sec:unification}
The revamped MDM presented in the two previous sections contains several
departures from conventional gauge coupling unification: (i) there is a pair
of Higgs doublets with intermediate-scale masses ($M_{h'}\sim10^{10}\,{\rm GeV}$),
(ii) there is a richer structure of GUT particles, including two pairs of
Higgs triplets (from the \r{5},\rb{5} representations) with masses
$M_{H_3,\bar H_3}\sim10^{14}\,{\rm GeV}$, and (iii) there is a spectrum of masses for
the different components of the \r{75}, all near the unification scale. There
is also a
hidden gauge group, with an in-principle independent gauge coupling at the
Planck scale (denoted by $g$ in Sec.~\ref{sec:hidden}).\footnote{In the spirit
of string unified models one could assume that the observable and hidden gauge
couplings are related at the Planck scale or at the string scale ($M_{\rm
str}\approx4\times10^{17}\,{\rm GeV}$).} Fortunately, the issue of gauge coupling
unification in the observable sector has already been addressed in detail
in Ref.~\cite{HMTY}. Those calculations are directly applicable to our model
since the observable matter content and spectrum of the two models is the
same, even though the dynamics providing the intermediate scale are different.
Thus, here we limit ourselves to a brief summary of the relevant issues.
Writing down the one-loop RGEs for the gauge couplings, including a common
supersymmetric threshold at $M_{\rm SUSY}$, one can eliminate the unified
$SU(5)$ coupling and obtain the following two relations \cite{HMTY,Method}
\begin{eqnarray}
\left(3\alpha^{-1}_2-2\alpha^{-1}_3-\alpha^{-1}_1\right)(M_Z)&=&
{1\over2\pi}\left\{{12\over5}\ln{M_{H_3}M_{\bar H_3}\over M_{h'}M_Z}
-2\ln{M_{\rm SUSY}\over M_Z}-23.3\right\}
\label{eq:Relation1}\\
\left(5\alpha^{-1}_1-3\alpha^{-1}_2-2\alpha^{-1}_3\right)(M_Z)&=&
{1\over2\pi}\left\{36\ln{(M^2_V M_\Sigma)^{1/3}\over M_Z}
+8\ln{M_{\rm SUSY}\over M_Z}+12.1\right\}
\label{eq:Relation2}
\end{eqnarray}
In these relations, $M_V=3\sqrt{15}(g_5/\lambda_{75})M_{75}$ is the mass of the
GUT gauge bosons, and the explicit constants come from the splittings of the
\r{75} relative to the $M_\Sigma=5M_{75}$ mass of its (\r{8},\r{3}) component.
The above relations can be made more
accurate by the inclusion of realistic low-energy supersymmetric thresholds,
two-loop gauge coupling RGEs, and smooth decoupling of heavy particles. Once
this is done, and the latest values of the Standard Model gauge couplings
are input ({\em i.e.}, $\alpha^{-1}=127.9\pm0.2$, $\sin^2\theta_W=0.2314\pm0.0004$,
$\alpha_3=0.118\pm0.007$), one obtains the following 1$\sigma$ allowed
intervals \cite{HMTY}:
\begin{eqnarray}
1.4\times10^{17}\,{\rm GeV}\le&{M_{H_3}M_{\bar H_3}\over M_{h'}}&
\le 5.5\times10^{20}\,{\rm GeV}\ ,
\label{eq:H3range}\\
8.4\times10^{15}\,{\rm GeV}\le&\left(M^2_V M_\Sigma\right)^{1/3}&
\le2.6\times10^{16}\,{\rm GeV}\ .
\label{eq:Vrange}
\end{eqnarray}
It is evident that our choices above, {\em i.e.}, $M_{H_3}\sim M_{\bar H_3}\sim
10^{14}\,{\rm GeV}$ and $M_{h'}\sim10^{10}\,{\rm GeV}$, are perfectly consistent with the
constraint in Eq.~(\ref{eq:H3range}). The same is true for the
middle-of-the-road choice $M_V\sim M_\Sigma$ ({\em i.e.}, $\lambda_{75}\sim g_5$),
which yields a GUT scale close to $10^{16}\,{\rm GeV}$.
We recall that we have set the masses of the \r{50},\rb{50} representations
at the gravitational scale $M\approx10^{18}\,{\rm GeV}$ in order to prevent the
onset of a non-perturbative $SU(5)$ regime below the Planck scale. Nonetheless,
because of the needed GUT-scale \r{75} representation, the unified gauge
coupling grows above the unification scale. However, it has been demonstrated
that this coupling remains in the perturbative regime, {\em i.e.}, $\alpha\lsim0.1$
\cite{HMTY}. One could assume that the corresponding gauge coupling at the
gravitational scale ($g\approx1$) is related to the gauge coupling
from the hidden sector gauge group discussed in Sec.~\ref{sec:hidden}, as
would be the case in string models. This relation would help to further
constrain the viable hidden sector choices. For instance, assuming a
``super-unified" situation, where hidden and observable gauge couplings are
equal near the gravitational scale, the constraints on the hidden sector choice
can be read off Fig.~\ref{fig:NcNf} ($g=1$ curves).
Concerning proton decay, gauge-boson-mediated dimension-six operators depend
on $1/M^2_V$. From Eq.~(\ref{eq:Vrange}), $M_V$ is not expected to be much
below $10^{16}\,{\rm GeV}$, unless $\lambda_{75}\gg g_5$, but this case is unlikely
since $\lambda_{75}$ would be in the non-perturbative regime. Thus, we don't
expect a particular enhancement of dimension-six operators in this model. More
interesting is the situation with the dimension-five proton decay operators,
which depend on the effective Higgs triplet mass ($M_{H_{\rm eff}}$) defined
in Eq.~(\ref{eq:Heff}). The dominant proton partial lifetime is given by
\cite{HMY,HMTYpd}
\begin{equation}
\tau(p\rightarrow K^+\bar\nu_\mu)=2.0\times10^{31}\,{\rm y}
\left| {0.0056\,{\rm GeV}^3\over\beta}\,{0.67\over A_S}\,{\sin2\beta\over 1+y^{tK}}\,
{M_{H_{\rm eff}}\over 10^{17}\,{\rm GeV}}\,{\,{\rm TeV}^{-1}\over f}\right|^2\ ,
\label{eq:pdecay}
\end{equation}
where $\beta=(5.6\pm0.8)\times10^{-3}\,{\rm GeV}^3$ is the relevant hadronic matrix
element, $A_S$ is the short-distance renormalization factor, and $y^{tK}$
parametrizes the contribution of the third family relative to the first two
($|y^{tK}|\approx2$ for $m_t=175\,{\rm GeV}$) with an undetermined phase. The $f$
functions are the one-loop integrals which behave as $1/f\approx m^2_{\tilde
q}/m_{\widetilde W}$ for $m_{\tilde q}\gg m_{\widetilde W}$.
{}From unification constraints, Eq.~(\ref{eq:H3range}) indicates that
$M_{H_{\rm eff}}>1.4\times 10^{17}\,{\rm GeV}\approx10M_V$. In this case,
Eq.~(\ref{eq:pdecay})
and Ref.~\cite{AN} show that the present Kamiokande limit $\tau(p\rightarrow \bar\nu
K^+)>1\times10^{32}\,{\rm y}$ \cite{PDG}, is satisfied provided $\tan\beta$ is not too
large ($\tan\beta\lsim5$) and the universal scalar mass $m_0>300\,{\rm GeV}$. On the
other hand, in our model we obtain $M_{H_{\rm eff}}\gsim10^{18}\,{\rm GeV}
\approx100M_V$, and the experimental limit is satisfied rather comfortably,
even for large values of $\tan\beta$ and presently accessible supersymmetric
particle masses. For instance, for $m_{\tilde q}\approx 300\,(600)\,{\rm GeV}$ and
$m_{\widetilde W}\approx 80\,{\rm GeV}$, $\tan\beta\lsim10\,(40)$ is required. Thus,
$p\rightarrow \bar\nu K^+$ remains the dominant mode for proton decay, with good
prospects for observation at the upcoming SuperKamiokande experiment and the
proposed ICARUS facility. Note that the much-weakened proton-decay upper-bound
on $\tan\beta$ offers a new possibility in the study of Yukawa coupling
unification in $SU(5)$ GUTs ({\em i.e.}, $\lambda_b=\lambda_\tau$), which now also
allow the so-called ``large-$\tan\beta$" solution \cite{YU}.
\section{Cosmic baryon asymmetry}
\label{sec:CBA}
With the realization of significant electroweak baryon number violation at high
temperatures, which occurs through ($B$+$L$)-violating but
($B$--$L$)-conserving non-perturbative sphaleron interactions
\cite{sphalerons}, several new mechanisms for generating the cosmic baryon
asymmetry have been proposed \cite{Olive}. These mechanisms produce a
primordial lepton asymmetry (leptogenesis), which is then recycled by
sphaleron interactions into a baryon asymmetry at the electroweak scale. It is
important to note that primordial ($B$--$L$)-conserving asymmetries, such as
those produced in traditional $SU(5)$ GUT baryogenesis, are likely to be wiped
out by the sphaleron interactions \cite{erasure}. Therefore, in the context of
$SU(5)$ GUTs, the leptogenesis-based mechanisms may be unavoidable. Here we
consider the simplest of these mechanisms, based on the out-of-equilibrium
decay of right-handed neutrinos, as first suggested in
Ref.~\cite{FY},\footnote{Before the realization of the importance of the
sphaleron interactions, Ref.~\cite{MNTY2} pointed out the possibility of
generating a baryon asymmetry in the decay of right-handed neutrinos via baryon
number violating GUT interactions.} and extended to supersymmetry in
Refs.~\cite{CDO,MSYY}, and to $SU(5)\times U(1)$ GUTs in Refs.~\cite{ENO,ELNO}.
We note that the lepton-asymmetric decays of right-handed sneutrino condensates
\cite{AD,sneutrinos}, may provide an additional contribution to the lepton
asymmetry that we discuss below.
In order to satisfy the out-of-equilibrium condition in the decay of the
right-handed neutrinos, one could follow the standard procedure of demanding
that the $\nu^c_{1,2,3}$ decay rate be less than the expansion rate of the
Universe at the time of $\nu^c$ decay. This condition leads to constraints on
the $\lambda_3$ couplings of the right-handed neutrinos, that can be
undesirable when trying to use the same couplings to compute the corresponding
light see-saw neutrino masses. Even more problematic can be the need to obtain
the surviving lepton asymmetry solely from the decays of the lightest
right-handed neutrino ($\nu^c_1$), since the asymmetry produced in the decays
of $\nu^c_{2,3}$ is typically wiped out by the $\nu^c_1$ interactions. Such
potential difficulties have been exemplified in Ref.~\cite{ELNO}. Instead, here
we follow an alternative scenario \cite{NOS},\footnote{Below we show that in
our model, the traditional out-of-equilibrium scenario is also viable.} whereby
the right-handed neutrinos are produced in the decays of the inflaton
subsequent to inflation. The COBE data on the anisotropy of the cosmic
microwave brackground radiation, interpreted in the context of inflation,
allows one to deduce the inflaton mass to be $m_{\eta}\sim 10^{11}\,{\rm GeV}$ and the
reheating temperature $T_R\sim10^8\,{\rm GeV}$ \cite{CDO}. The $\nu^c$ then decay
immediately after inflation and out of equilibrium at the temperature $T_R\ll
M_{\nu^c}$, {\em as long as} $M_{\nu_c}<m_{\eta}\sim10^{11}\,{\rm GeV}$.
Interestingly, the constraint from proton decay (see Sec.~\ref{sec:observable})
ensures that this condition is satisfied {\em automatically}.
The primordial lepton asymmetry, when reprocessed by sphaleron interactions,
leads to a similar baryon asymmetry \cite{CDO}
\begin{equation}
{n_B\over n_\gamma}\sim{n_L\over n_\gamma}\sim
\left({m_{\eta}\over M_{Pl}}\right)^{1/2}\epsilon\sim 10^{-4}\ \epsilon\ ,
\label{eq:nB}
\end{equation}
where the asymmetry parameter ($\epsilon$) due to the decay of the
$i$th-generation neutrino and sneutrino is given by \cite{CDO}
\begin{equation}
\epsilon_i = {1\over 2\pi (\lambda^\dagger_3\lambda_3)_{ii}}
\sum_j \left({\rm Im}\,[(\lambda^\dagger_3\lambda_3)_{ij}]^2\right)\,
g(M^2_{\nu^c_j}/M^2_{\nu^c_i})\ ,
\label{eq:eps}
\end{equation}
with
\begin{equation}
g(x)=4\sqrt{x}\,\ln{1+x\over x}\ .
\label{eq:g}
\end{equation}
To proceed we need to manipulate the entries in $\lambda_3$, which has
remained as yet unspecified. We define the unitary rotation matrix $U$,
such that $\hat\lambda_3=U\lambda_3 U^\dagger$, where $\hat\lambda_3$ is
the diagonal matrix of eigenvalues of $\lambda_3$. Experience with the
quark mixing matrix leads us to assume that $U$ differs little from the
identity matrix: $U={\bf1}+R$, with \cite{ELNO}
\begin{equation}
R=\left(\begin{array}{ccc} 0&\theta_{12}&0\\ -\theta^*_{12}&0&\theta_{23}\\
0&-\theta^*_{23}&0\end{array}\right)\ .
\label{eq:R}
\end{equation}
With this ansatz we obtain to lowest non-vanishing order
\begin{eqnarray}
(\lambda^\dagger_3\lambda_3)_{ii}&=&|\hat\lambda^i_3|^2
+\sum_j|\hat\lambda^j_3|^2\, |\theta_{ij}|^2 \ ,
\label{eq:ii}\\
(\lambda^\dagger_3\lambda_3)_{ij}&=&|\theta_{ij}|e^{i\phi_{ij}}
\left[|\hat\lambda^i_3|^2-|\hat\lambda^j_3|^2\right]\quad (i\not=j)\ ,
\label{eq:ij}
\end{eqnarray}
where $\phi_{ij}={\rm Arg}\,[\theta_{ij}]$. Thus, Eq.~(\ref{eq:eps}) becomes
\begin{equation}
\epsilon_i={1\over2\pi}\,
{\sum_{j\not=i}|\theta_{ij}|^2\sin2\phi_{ij}\,
[|\hat\lambda^i_3|^2-|\hat\lambda^j_3|^2]^2\
g(M^2_{\nu^c_j}/M^2_{\nu^c_i})\over
|\hat\lambda^i_3|^2+\sum_j|\hat\lambda^j_3|^2\,|\theta_{ij}|^2}\ .
\label{eq:eps2}
\end{equation}
Because of the several unknown parameters in the above expressions, and the
inherent uncertainties in this type of calculations, we will be content with
presenting a plausible scenario leading to interesting lepton asymmetries
and see-saw neutrino masses. For simplicity let us assume that the $\lambda_6$
matrix is proportional to the unit matrix, {\em i.e.},
\begin{equation}
M_{\nu^c_1}=M_{\nu^c_2}=M_{\nu^c_3}=M_{\nu^c}
=\lambda_6\,{\vev{T\bar T}\over M}\ .
\label{eq:Mnu^c}
\end{equation}
The light neutrino mass matrix then becomes $M_\nu=\lambda_3\lambda^T_3
v^2_2/M_{\nu^c}$. If we neglect the CP violating phases (a not necessarily
justified approximation), the matrix $U$ which diagonalizes
$\lambda_3\lambda^\dagger_3$, also diagonalizes $\lambda_3\lambda^T_3$ and the
physical neutrino masses become (up renormalization group scaling
corrections \cite{chorus})
\begin{equation}
m_{\nu_i}\approx{(\hat\lambda^i_3 v_2)^2\over M_{\nu^c}}\ .
\label{eq:Mnu}
\end{equation}
Furthermore, in our ansatz the (small) neutrino mixing angles are given by
$\theta_{e\mu}=\theta_{12}$, $\theta_{e\tau}=0$, and
$\theta_{\mu\tau}=\theta_{23}$. As we will see shortly, these mixing angles are
unrestricted from lepton asymmetry considerations, and thus could accomodate
the MSW solution to the solar neutrino problem ($\nu_e\leftrightarrow\nu_\mu$)
and lead to interesting $\nu_\mu\leftrightarrow\nu_\tau$ oscillations at the
CHORUS and NOMAD, and P803 experiments at CERN and Fermilab respectively.
{}From Eq.~(\ref{eq:Mnu}) we see that
$m_{\nu_\tau}\approx (\hat\lambda_3^3 v_2)^2/M_{\nu^c}
=[\hat\lambda_3^3\, \sin\beta(174\,{\rm GeV})]^2/M_{\nu^c}$. With
$M_{\nu^c}\sim10^{10}\,{\rm GeV}$ and $\hat\lambda^3_3\approx0.1$, we get
$m_{\nu_\tau}\sim15\,(30)\,{\rm eV}$ for $\tan\beta\sim1\,(\tan\beta\gg1)$. This
range of tau neutrino masses provide an adequate and desirable hot dark matter
component of the Universe. Thus, in what follows we take $\hat\lambda^3_3=0.1$.
It is also natural to assume that the remaining eigenvalues of the $\lambda_3$
matrix are hierarchically smaller, {\em i.e.},
$\hat\lambda^1_3\ll\hat\lambda^2_3\ll\hat\lambda^3_3$. For instance,
$\hat\lambda^2_3\sim{1\over100} \hat\lambda^3_3$ yields $m_{\nu_\mu}\sim
10^{-3}\,{\rm eV}$, consistent with solutions to the solar neutrino problem via the
MSW mechanism. (These hierarchies are comparable to those in the up-quark
Yukawa matrix.)
Going back to the calculation of the lepton asymmetries, with our hierarchical
assumption for $\hat\lambda_3$, from Eq.~(\ref{eq:eps2}) we obtain
\begin{eqnarray}
\epsilon_1&\approx&{2\ln2\over\pi}\,\left(\hat\lambda^2_3\right)^2
\sin2\phi_{12}\sim 10^{-6}\,\phi_{12}\ ,
\label{eq:e1}\\
\epsilon_2&\approx&{2\ln2\over\pi}\,\left(\hat\lambda^3_3\right)^2
\sin2\phi_{23}\sim 10^{-2}\,\phi_{23}\ ,
\label{eq:e2}\\
\epsilon_3&\approx&{2\ln2\over\pi}\,\left(\hat\lambda^3_3\right)^2
|\theta_{23}|^2\sin2\phi_{23}\sim 10^{-2}\,|\theta_{23}|^2\phi_{23}\ .
\label{eq:e3}
\end{eqnarray}
With the expression for the estimated baryon asymmetry in Eq.~(\ref{eq:nB}), we
would get the desired result of ${\rm few}\times10^{-10}$ for $\phi_{12}\sim1$
and $\phi_{23}\ll1$. The natural choice would be maximal CP violation in
the $\theta_{12}$ entry in the rotation matrix $R$ (see Eq.~(\ref{eq:R}))
and no CP violation elsewhere in the matrix (unless new entropy diluting
sources are introduced to reduce $\epsilon_1+\epsilon_2+\epsilon_3$). These
results would be affected somewhat if one allows a non-trivial structure
to the matrix $\lambda_6$ ({\em i.e.}, relaxing the assumption in
Eq.~(\ref{eq:Mnu^c})).
We now remark that this model is also viable in the traditional
out-of-equilibrium scenario, where $\epsilon_1$ is the only surviving
asymmetry. The out-of-equilibrium condition at $T=M_{\nu^c_1}\sim10^{10}\,{\rm GeV}$,
\begin{equation}
\Gamma_{\nu^c_1}={(\lambda^\dagger_3\lambda_3)_{11}\over16\pi}\,
M_{\nu^c_1}<1.66 g^{1/2}_*\,{T^2\over M_{Pl}}=H\ ,
\label{eq:out}
\end{equation}
is satisfied for (using Eq.~(\ref{eq:ii}))
\begin{equation}
(\lambda^\dagger_3\lambda_3)_{11}=
|\hat\lambda^1_3|^2+|\hat\lambda^2_3|^2\,|\theta_{12}|^2\lsim10^{-6}\ ,
\label{eq:cond}
\end{equation}
which is consistent with our hierarchical assumption. However, in this case
the calculation of the leptonic asymmetry has a larger ($\sim10^{-2}$
\cite{CDO}) coefficient than in Eq.~(\ref{eq:nB}), requiring a non-maximal CP
violating phase $\phi_{12}\sim10^{-2}$.
Finally, let us comment on whether or not the sphaleron interactions may wash
away the leptonic asymmetry produced above. This could in principle happen if
the non-renormalizable operators obtained when integrating out the right-handed
neutrino fields, {\em i.e.}, $(\lambda_3/M_{\nu^c}) LLHH$, where $L$ is the lepton
doublet in $\bar f$ and $H$ the Higgs doublet in $h$, are in equilibrium with
the sphaleron interactions \cite{FY}. It has been shown \cite{CKO} that to
prevent the erasure of the asymmetry, one must demand
$M_{\nu^c}\mathrel{\mathpalette\@versim>}(\lambda_3)^2\, 3\times10^9\,{\rm GeV}$, which is always satisfied for
our choices of $\lambda_3$ and $M_{\nu^c}$.
\section{Comparison with $SU(5)\times U(1)$}
\label{sec:comparison}
The revamped MDM presented in the previous sections has several appealing
phenomenological features, constituting an interesting example of traditional
grand unified model building. Nonetheless, it is apparent that the model
is rather non-minimal or uneconomical. For instance, a \r{75} needs to be
used for GUT symmetry breaking, greatly increasing the size of the GUT particle
spectrum. Moreover, the \r{50},\rb{50} to effect the doublet-triplet splitting
problem make the unified gauge coupling so large above the GUT scale that they
need to be taken at the gravitational scale. The doublet-triplet splitting
is tamed, but proton decay can still be too fast because of the ``useless"
pieces of the \r{50},\rb{50} representations which need to be made heavy,
resulting in the otherwise-not-needed doubling of these representations and of
the Higgs pentaplets. Regarding the right-handed neutrinos, their (ad-hoc)
introduction has various desirable consequences. However, the Yukawa matrix
coupling them to the lepton doublets is arbitrary, with no particular
motivation for its desired hierarchical structure.
It is interesting to note that the above critique of the revamped MDM can be
circumvented altogether if one extends the gauge group from $SU(5)$ to
$SU(5)\times U(1)$ \cite{Barr,revitalized,Moscow}. Gauge symmetry breaking
down to the Standard Model gauge group occurs via vacuum expectation values
of the $H$ (\r{10}) and $\bar H$ (\rb{10}) Higgs representations. This is
possible because of the ``flipping" $u\leftrightarrow d$, $u^c\leftrightarrow d^c$, $e\leftrightarrow\nu$,
$e^c\leftrightarrow\nu^c$ involved in the assignment of the Standard Model particles
to the $\bar f=\{u^c,L\}$ (\rb{5}) and $F=\{Q,d^c,\nu^c\}$ (\r{10})
representations. Thus, $H$ and $\bar H$ contain one pair of neutral fields
$\nu^c_H,\nu^c_{\bar H}$, which get vevs along the flat direction
$\vev{\nu^c_H}=\vev{\nu^c_{\bar H}}$. There is no need for large GUT
representations for symmetry breaking. As is well known (and we review below),
this property takes on a much larger magnitude when one attempts to derive
these models in string model building.
The missing-partner mechanism, which above involved the couplings
$\bar\theta\Sigma h$ [(\rb{50})(\r{75})(\r{5})] and $\theta\Sigma\bar h$
[(\r{50})(\r{75})(\rb{5})], is now effected by the couplings $HHh$
[(\r{10})(\r{10})(\r{5})] and $\bar H\bar H\bar h$
[(\rb{10})(\rb{10})(\rb{5})]. First note that no additional representations
are needed besides the GUT-breaking Higgs ones. Moreover, the resulting
Higgs triplet matrix
\begin{equation}
\bordermatrix{
&\bar h_3&d^c_H\cr
h_3&0&\lambda_4\,\vev{\nu^c_H}\cr
d^c_{\bar H}&\lambda_5\,\vev{\nu^c_{\bar H}}&0}\ ,
\label{eq:f2/3}
\end{equation}
does not need a large non-zero (22) entry ({\em c.f.} Eq.~(\ref{eq:2/3}))
because the ``useless" components of the $H$ and $\bar H$ representations
are eaten by the GUT gauge bosons to become massive or become GUT Higgsinos.
This natural zero mass term for $d^c_H d^c_{\bar H}$ implies that the
dimension-five proton decay operators are negligible. We end up with a very
economical GUT Higgs spectrum and no threat of dimension-five operators.
Regarding neutrino masses, the right-handed neutrinos which had to be
introduced by hand in the revamped MDM, are now contained in the $F$ (\r{10})
representations. Indeed, the coupling $\lambda_3\bar f\nu^c h$ in
Eq.~(\ref{eq:W'''}) is here written as $\lambda_3\bar f e^c h$, with the
(unavoidable) right-handed electrons now introduced ``by hand". In $SU(5)\times
U(1)$ this coupling provides the charged lepton masses. On the other hand, the
coupling $\lambda_1 F\bar f\,\bar h$, which in Eq.~(\ref{eq:W}) provided the
down-quark masses, here provides the up-quark masses and Dirac neutrino masses.
(Also, the coupling $\lambda_2 FFh$, which in Eq.~(\ref{eq:W}) provided the
up-quark masses, here provides the down-quark masses.) Thus, the right-handed
neutrinos are unavoidable in $SU(5)\times U(1)$, and their Yukawa couplings
to the lepton doublets are equal to those of the up-quark Yukawa matrix,
providing (as discussed in Sec.~\ref{sec:CBA}) an automatic and desirable
hierarchy in the see-saw neutrino masses. An important distinction between
the see-saw mechanism in the revamped MDM and $SU(5)\times U(1)$ is the
manner in which the right-handed neutrinos get a mass. In the revamped MDM
this is through the superpotential term $\lambda_6\nu^c\nu^c\vev{T\bar T}/M$ in
Eq.~(\ref{eq:W'''}), whereas in $SU(5)\times U(1)$ there are two possible
sources: (i) through cubic couplings
$\lambda_6 F\bar H\phi\ni\lambda_6\vev{\nu^c_{\bar H}}\nu^c\phi$, where $\phi$
(with $\vev{\phi}=0$) are $SU(5)$ singlets \cite{revitalized}; and (ii)
through non-renormalizable couplings $\lambda_9 FF\bar H\bar H/M\ni
\lambda_9 (\vev{\nu^c_{\bar H}}^2/M)\nu^c\nu^c$ \cite{improved}. The second
form resembles that in the revamped MDM, although the mass scale is likely
higher ({\em i.e.}, $\vev{\nu^c_{\bar H}}^2/M\sim10^{14}\,{\rm GeV}$).
These two models also differ somewhat in the calculation of the cosmic baryon
asymmetry, besides the possible difference in the right-handed neutrino mass
spectrum. Indeed, because the $SU(5)\times U(1)$ gauge symmetry is broken along
a flat direction, there is a dilution factor ($\Delta$) in the computation of
the lepton asymmetry due to the entropy released by the late-decaying ``flaton"
field \cite{dilution,ENO}. However, these two effects ($\nu^c$ spectrum and
$\Delta$) tend to compensate each other and an acceptable baryon asymmetry is
typically obtained \cite{ENO,ELNO}.
There is another cosmological aspect of these models that sets them apart,
namely the breaking of the GUT symmetry down to the Standard Model one.
In the MDM, $SU(5)$ symmetry breaking via an arbitrary vev of the \r{75} leads
to several possible degenerate vacua \cite{HMPR}, at least in the context of
global supersymmetry. When supergravity effects are taken into account, if the
desired vacuum has zero cosmological constant, all the others will be lower in
energy, although essentially unreachable \cite{Weinberg}. Thus, if the vev
of the \r{75} can be arranged to be in the desired direction, the Universe
will remain in the desired broken phase. In contrast, in $SU(5)\times U(1)$
the breaking down to the Standard Model via the vevs of the \r{10},\rb{10}
along the F- and D-flat direction $\vev{\nu^c_H}=\vev{\nu^c_{\bar H}}$
is {\em unique} \cite{revitalized}.
Regarding the issue of unification, the revamped MDM requires non-minimal
representations to make this possible. In $SU(5)\times U(1)$ traditional
grand unification does not occur (although the non-abelian Standard Model gauge
groups do unify) and unification is not a test of the model. However, if string
unification is desired (at the scale $M_{\rm str}\approx4\times10^{17}\,{\rm GeV}$),
then non-minimal representations need to be added to the $SU(5)\times U(1)$
model \cite{gap}.
We have seen that the pair of \r{50},\rb{50} representations in the revamped
MDM need to be put at the gravitational scale. It is then natural to ask
whether this model can be obtained from the only known consistent theory of
quantum gravity, namely string theory. Because of some technical difficulties
which we review below, no attempts have been made to derive the MDM from
strings. It is of course well known that $SU(5)\times U(1)$ can be easily
derived from strings \cite{revamp,search}.
The prime constraint in string model-building is that of the massless
representations which are allowed when the corresponding gauge group $G$ is
represented by a ``level-$k$" Kac-Moody algebra on the world-sheet
\cite{ELN,GO}. The allowed representations must be unitary,
\begin{equation}
\sum_{i=1}^{\rm rank\, G} n_i m_i\le k\ ,
\label{eq:unitary}
\end{equation}
where $n_i$ are the Dynkin labels of the highest weight representation in
question, and $m_i$ are fixed positive integers for a given $G$. In the case of
$SU(n)$: $m_i=1\,,\,\forall i$. In our $SU(5)$ example then $\sum_{i=1}^4
n_i\le k$. Looking up the $n_i$ values, we see that for $k=1$, only
\r{1},\r{5},\rb{5},\r{10},\rb{10} are unitary. For $k=2$ we find in addition:
\r{15},\r{24},\r{40},\rb{40},\r{45},\rb{45},\r{50},\rb{50},\r{75}. Only
level-one constructions appear to be needed to derive $SU(5)\times U(1)$,
whereas at least level-two constructions are required in the MDM. However,
this is not the end of the story, since one can also ask whether the allowed
representations could possible be massless. This requires calculating the
so-called conformal dimension $h_r$ of the representation $r$,
\begin{equation}
h_r={C_r\over 2k+C_A}\ ,
\label{eq:h}
\end{equation}
where $C_r$ is the Casimir of $r$, and $C_A$ that of the adjoint
representation. If $h_r>1$, the representation is necessarily massive. For
$h_r\le1$ the representation could be massless, although this is not guaranteed
since other degrees of freedom may add their own contribution to the conformal
dimension making it exceed unity. It is not hard to see that in $SU(5)$ all
unitary representations at level one are also massless \cite{ELN}, and thus
$SU(5)\times U(1)$ models can be readily constructed at level one. The unitary
representations of interest for MDM model-building, which are allowed at level
two, have conformal dimensions
\begin{equation}
h_{50,\overline{50}}={42\over5(k+5)}\ ,\qquad h_{75}={8\over k+5}\ ,
\label{eq:h5075}
\end{equation}
and are not massless at level two. In fact, $k=4$ is required to make all these
representations massless. Such high-level Kac-Moody constructions have never
been attempted.
One intriguing possibility would be to construct level-two $SU(5)$ string
models (for recent attempts see Ref.~\cite{Aldazabal}), which should allow the
required large MDM representations, although with masses at the Planck scale.
Note that this is not necessarily a problem since we already require the
\r{50},\rb{50} to be at that mass scale. If the \r{75} is also raised to that
scale, the breaking of $SU(5)$ would occur at the string scale, and this may be
difficult to reconcile with gauge coupling unification. It is also worth
remarking that in a string model all gauge couplings are related at the string
scale, and with $SU(5)$ constructed at level two, the relation would be
$\sqrt{2}\,g_5=g_h$ \cite{Ginsparg}, where $g_h$ is the gauge coupling of the
hidden gauge group. Finally, the mass terms in Eqs.~(\ref{eq:W},\ref{eq:W'}),
which would not be allowed if the large MDM representations belonged to the
massless spectrum, are expected to arise when they belong to the string massive
spectrum. Of course, bridging the gap between the massless and massive spectrum
may create problems in obtaining the low-energy effective field theory, but
this question cannot be answered until an actual model is constructed along
these lines.
\section{Conclusions}
\label{sec:conclusions}
During the last few years, a great deal of attention has been paid to
supersymmetric grand unified theories in light of the precise LEP measurements
of the Standard Model gauge couplings. These analyses depend crucially on
the details of the low-energy supersymmetric spectrum and the heavy GUT
spectrum. Most of the effort to date has been focused on the minimal $SU(5)$
supergravity model, which appears to be running into difficulties regarding
unification and proton decay. In addition, there is the nagging doublet-triplet
splitting problem that receives no satisfactory explanation. Motivated by
these developments, we have reconsidered one of the alternatives to minimal
$SU(5)$, where the doublet-triplet splitting is dealt with in a reasonable
way via the missing-partner mechanism, and gauge coupling unification is not
in jeopardy. We have revamped this model to tame the dimension-five proton
decay operators, and to allow see-saw neutrino masses. In order to generate the
needed intermediate scale for the right-handed neutrino masses, we have endowed
the model with a ``modern" hidden sector which can generate dynamically the
desired intermediate scale, the scale of supersymmetry breaking, and the Higgs
mixing parameter $\mu$. The revamped MDM also provides for the cosmic baryon
asymmetry through the Fukugita-Yanagida mechanism via lepton-number-violating
decays of the right-handed neutrinos.
We have also contrasted the main features of the revamped MDM against the
``flipped" $SU(5)\times U(1)$ model, and basically shown that the former
can be considered as a ``poor man's" version of the latter. In the realm
of string model-building, $SU(5)\times U(1)$ fares rather well, whereas
the revamped MDM is very unlikely to be realized, except perhaps if one
allows $SU(5)$ symmetry breaking to occur at the string scale.
\section*{Acknowledgments}
We would like to thank John Ellis and Bruce Campbell for useful discussions.
This work has been supported in part by DOE grant DE-FG05-91-ER-40633.
|
\section{Introduction} \label{sec-intro}
Remarkable, exact results on the vacuum structure of four dimensional
$N=2$ supersymmetric quantum field theories with an $SU(2)$ gauge
symmetry and matter fields in the fundamental representation have
recently been obtained \cite{SW}, with generalizations to $SU(N_c)$
gauge groups \cite{Ncolor}. In these models, an $N=1$ supersymmetry
preserving mass perturbation for the adjoint superfield leads to
monopole condensation, confinement and chiral symmetry breaking.
There has been much interest in how such exact supersymmetric
results generalize to nonsupersymmetric gauge theories
\cite{othersusybreaking}.
In this paper we show how $N=2$, $N=1$ and softly broken
supersymmetric models can all be embedded in a single enlarged $N=1$
model in which coupling constants are treated as ``spurion'' fields.
The effective superpotential of the enlarged model can be exactly
determined from its symmetries and the $N=2$ limit. In the $N=2$
limit the $D$-terms are also exactly determined, but when
supersymmetry is broken the potential and superpotential each receive
contributions from additional $D$-terms of the enlarged model, which
vanish in the $N=2$ limit. Although constrained by the symmetries of
the model, these cannot be fully determined. If the $N=2$ symmetry
is broken solely by soft breaking masses, these unknown functions
typically are necessary to determine details of the vacuum structure.
The exact superpotential does however include all the lowest-dimension
gauge interaction terms, and continues to have singularity
structure\footnote{
Strictly speaking there is no longer a moduli space after
breaking $N=2$ supersymmetry. However, we can still discuss
monodromies in the configuration space of the theory, by means
of an external source used to traverse a closed loop.}
consistent with points at which monopoles become massless.
We can thus confirm that when the soft breaking masses are
perturbations to an $N=1$ preserving mass, the vacuum remains
essentially that of the $N=1$ model.
An important step in obtaining the exact results of \cite{SW} is to
treat the couplings in the $N=2$ models as spurion chiral
superfields in an enlarged $N=1$ model, which fail to propagate if
their kinetic term coefficients are taken to infinity \cite{natiral}.
Their scalar component vacuum expectation values (vevs) can thus be
frozen at any chosen values. The $F$-terms of the Wilsonian effective
Lagrangian of a supersymmetric model \cite{natiral,SV,CERN} must be
holomorphic in the fields as well as invariant under the gauge and
global symmetries of the model (assuming a supersymmetric
regularization scheme). The superpotential of the enlarged model is
therefore holomorphic in the couplings. A recent paper \cite{yale1}
showed that soft supersymmetry breaking interactions and mass terms
can be introduced in supersymmetric models without altering the
holomorphic constraint on the $F$-terms. The spurion fields can be
coupled to a sector that generates supersymmetry breaking expectation
values for the spurion field $f$-components\footnote{
Our superfield notation \cite{WB} is $\Phi = a_\Phi
+ \sqrt{2}\, \theta \psi_\Phi + \theta \theta f_\Phi$,
with $\Phi^\dagger\Phi\big|_D \equiv
\int d^2\theta\ d^2{\bar\theta}\ \Phi^\dagger\Phi = |f_\Phi|^2,\
\Phi\big|_F \equiv \int d^2\theta\ \Phi = f_\Phi,\
\Phi\big|_A \equiv a_\Phi$, etc.}.
Freezing out the spurions generates soft supersymmetry breaking masses
and interactions in the embedded model.
For large vevs of the adjoint scalar field $a_\Phi$ in the models we
study, the $SU(N_c)$ gauge symmetry is
broken to a subgroup of weakly interacting $U(1)$ gauge symmetries.
It has been argued \cite{SW} that the $N=2$ theory remains in the
Coulomb phase even for small scalar vev where the model is
strongly coupled. One might have expected the pure $SU(N_c)$
dynamics to operate unimpeded when the scalar vev was much smaller than
the strong interaction scale $\Lambda$, and the low energy theory
to exist in a confining phase with all the gauge bosons confined
within glueballs. However, the solutions' self-consistency
\cite{SW} suggests that for the special choice of couplings that
gives the $N=2$ model, the quantum effects that cause condensation
and confinement exhibit special cancellations, leading to the
surprising low energy dynamics of a strongly coupled Coulomb phase
with massless fermions, monopoles and dyons.
The solutions' consistency with our expectations for how these
theories ought to behave helps confirm that higher dimension terms
do not qualitatively change the solutions. Following the analysis
\cite{SW}, which does not determine operators with more than two
derivatives (${\cal O}(p^2)$) or four fermions, we will assume that
the gauge particle dynamics are essentially controlled by the
effective operators of lowest dimension, at sufficiently low energy
scales. The model's symmetries permit higher dimension operators
containing for example powers of $|WW/\Lambda^3|^2$, and (by $N=2$
supersymmetry) also containing superderivatives and powers of
$\Phi/\Lambda$. Since the scalar component of $\Phi$ is pinned at
$\langle a_\Phi \rangle \sim \Lambda$ when the supersymmetry is broken
to $N=1$, these terms are not at first sight negligible. However,
since they must be related by $N=2$ supersymmetry to the corresponding
higher derivative $W$ terms \cite{mans}, they should also be suppressed
in the infrared regime. Thus despite these caveats, a tractable
electric--magnetic duality is useful in determining the theory's
vacuum structure.
We expect that introducing $N=2$ breaking masses destroys
cancellations in the dynamics below the scale of the breaking mass,
and confines the $U(1)$ gauge interaction below that scale (via the
Higgs mechanism in the dual theory). Provided that breaking masses
are sufficiently small relative to the scale $\Lambda$ of the strong
interactions, the theory will continue to be described by a $U(1)$
gauge theory between the $SU(2)$ breaking scale parameterized by
$\langle a_\Phi \rangle$, pinned at $\sim \Lambda$, and the breaking
mass scale. Softly broken $N=1$ and $N=0$ models, when close in
parameter space to the $N=2$ model, are therefore anticipated to
occupy the same phase.
The paper is organized as follows. Section~\ref{sec-su2} explicitly
describes the class of models we study. We allow the models' coupling
constants to be chiral superfield spurions, whose values are
eventually frozen in all of spacetime \cite{natiral,yale1}. We then
review the analysis that leads to the Seiberg--Witten ansatz for the
vacuum structure for $SU(2)$ gauge symmetry, and monopole condensation
in $N=1$ models close in parameter space to the $N=2$ model. Allowing
soft supersymmetry breaking terms in our Lagrangian via nonvanishing
spurion field $f$-components, we show that to lowest order in the soft
breakings, gauge kinetic terms are unchanged by the introduction of
soft masses. However, the potential minimum in these models typically
depends upon insufficiently constrained contributions from $D$-terms.
Section~\ref{sec-matter} discusses obtaining similar results in $SU(N_c)$
gauge theories or with matter fields in the fundamental representation,
and reaching QCD by continuous interpolation from models of this type.
\newpage
\section{Supersymmetric $SU(2)$} \label{sec-su2}
\subsection{An enlarged model} \label{sec-2.1}
Consider an $N=1$ supersymmetric $SU(2)$ gauge theory, containing one
matter chiral superfield in the adjoint representation. Promoting
the coupling constants into chiral superfields (including a $D$-term
normalization, $K$, which we shall use to generate a squark mass)
\cite{yale1} yields the Lagrangian
\begin{equation} \label{smallLagrangian}
\begin{array}{c}
{\cal L} =
{1\over 4\pi}{\rm Im}\left( {1\over 2}\tau_0 W_\alpha W^\alpha
\Big|_F
\ +\ (\tau_0 + K^\dagger K) \Phi^\dagger e^V \Phi\Big|_D \right)
\ +\ m \Phi^2\Big|_F \ +\ {\rm h.c.}\\ \\
\ +\ \Lambda_m^2 \left(m^\dagger m\Big|_D + \beta_m m\Big|_F
+ {\rm h.c.}\right)
\ +\ \Lambda_\tau^2 \left( \tau_0^\dagger \tau_0 \Big|_D
+ \beta_\tau \tau_0 \Big|_F + {\rm h.c.}\right)\\ \\
\ +\ \Lambda_K^2 \left( K^\dagger K\Big|_D + \beta_K K\Big|_F
+ {\rm h.c.}\right)
\end{array}
\end{equation}
(note that $\Lambda_m$ and $K$ are dimensionless).
The spurion fields $i = \{m, \tau_0, K\}$ do not propagate in the
limit $\Lambda_i \rightarrow \infty$; we can fix their scalar
components $a_i$ at any expectation values we choose, and replace
$f_i = -\beta_i^\dagger + {\cal O}(1/\Lambda_i^2)$.
[An equivalent approach normalizes the spurion $D$-terms
conventionally, placing the $\Lambda_i$ suppression on
the spurion-to-physical-field coupling operators; for example
writing $m^\dagger m\Big|_D + (m/\Lambda_m)\Phi^2\Big|_F$. The
spurion vevs are then taken proportional to $\Lambda_i$, so their
classical and quantum evolutions are still relatively suppressed.]
In any case, if we choose $\langle a_{\tau_0} \rangle\neq 0$
and $\left(\langle a_m\rangle\ ,\ \langle a_{K} \rangle\ ,\
\langle f_i \rangle\ \right) \ =\ 0$,
the model reduces to $N=2$ supersymmetric $SU(2)$, studied in
\cite{SW}, whose bare Lagrangian is
\begin{equation} \label{neqone}
{\cal L}_{N=2}\ =\ {1\over 4\pi}{\rm Im} \langle a_{\tau_0}\rangle
\left( {1\over 2} W_\alpha W^\alpha \Big|_F
\ +\ \Phi^{\dagger} e^V \Phi\Big|_D \right)\ .
\end{equation}
In this notation, we will review the derivation \cite{SW} of the
$N=2$ model effective potential and its generalization to the
$N=1$ case of $\langle a_m \rangle \neq 0$. Nonzero values of
$\langle f_{\tau_0} \rangle$ and $\langle f_K\rangle$ generate
supersymmetry breaking mass terms for gauginos and squarks
respectively.
In the limit $\langle f_{\tau_0,K} \rangle \rightarrow\infty$
the gauginos and squarks decouple from the low energy theory,
leaving only $SU(2)$ gauge bosons and adjoint fermions.
The model has an anomaly-free $U(1)_R$ global symmetry, whose charge
assignments
\begin{equation} \label{U1Rcharges}
\matrix{
\theta && +1 \cr
W_\alpha && +1 \cr
\tau_0 && 0 \cr
\Phi && 0 \cr
m && +2 \cr
K && {\rm arbitrary} }
\end{equation}
significantly constrain the terms which can appear in the low energy
effective Lagrangian. Allowable $D$-term operators have net $R$-charge
of zero, whereas $F$-term operators must have net $R$-charge of $+2$.
The anomaly breaks a second $SU(2)_R$ symmetry of the embedded $N=2$
model to a discrete symmetry; in the $SU(2)$ gauge case this gives
the ${\bf Z_2}$ symmetry.
A further constraint applies to the spurion sources $\beta_i$, which
may be treated as decoupled chiral superfields in their own right.
Coupled only linearly to the $f$-components of $m,\tau$ and
$K$, they do not contribute perturbatively to 1PI diagrams.
Moreover, since scalar components $a_{\beta_i}$ do not couple at all
to other fields in the bare potential, they will only appear
suppressed by powers of $\Lambda_i$ in induced terms in the
effective Lagrangian.
The only dependence in the effective theory on $\beta_i$ vevs
thus occurs indirectly, through the $f$-component vevs $f_i$ of
coupling-field spurions, induced by $a_{\beta_i}$.
\subsection{The low energy effective theory in the $N=2$ limit}
\label{sec-2.2}
To fix our notation we briefly review the analysis \cite{SW} relevant
to the $N=2$ model in (\ref{neqone}). The form of the ${\cal O}(p^2)$
effective Lagrangian's superpotential can be deduced from its holomorphic
properties as described below. In the $N=2$ limit the effective
Lagrangian's $D$-terms are then determined by $N=2$ supersymmetry,
giving the effective Lagrangian as
\begin{equation} \label{N=2lowE}
{1\over 4\pi} {\rm Im}\left\{ {\partial{\cal F}(A) \over \partial A}
\bar A\Bigg|_D
\ +\ {1\over 2} {\partial^2{\cal F}(A) \over \partial A^2}
W_{\alpha} W^{\alpha}\Bigg|_F \right\} \ ,
\end{equation}
where $A$ is the $N=1$ chiral multiplet and ${\cal F}$ the
``prepotential'' \cite{SW,prepotential}. The exact solution for the
effective Lagrangian of the $N=2$ model determines the effective
Lagrangian of the enlarged model, and hence of the models with soft
supersymmetry breaking masses, up to terms compatible with the
symmetries of the enlarged model that vanish in the $N=2$ limit.
We shall discuss these extra terms in the following sections.
In the $N=2$ model the classical potential for $a_\Phi$,
$V_{cl} = {\rm Tr}([a_\Phi, a_\Phi^\dagger]^2)$,
is minimized (and vanishes) when $a_\Phi$ has any diagonal complex vev.
This breaks $SU(2)$ down to $U(1)$ and yields a classical moduli space,
parameterized by the gauge invariant superfield $U \equiv {\rm Tr}[\Phi^2]$.
We assume that at low energies the quantum theory also remains in the
Coulomb phase, so that the particle content is restricted to the
$U(1)$ gauge superfield $W_\alpha$ and the massless matter fields.
In the Coulomb phase at arbitrarily low scales, where we need consider
only the lowest dimension terms in the Lagrangian, the effective action
is simply quadratic in the gauge field, and a duality transformation is
conveniently derived \cite{SW}
by considering a functional integral over the gauge superfield $W$.
Imposing the condition Im$({\cal D} W) = 0$ via a Lagrange multiplier
$V$, incorporated by a new dual gauge field $W_D = i{\cal D}V$, we can
complete the square and perform the Gaussian integral, to obtain an
equivalent theory for the dual field with $\tau_D = -1/\tau$:
\begin{equation}
{\cal Z} \sim \int{\cal D}[W_D]\ \exp{\left({i\over 8\pi}
{\rm Im} \int \tau_D W_D^\alpha W_{D \alpha}\Big|_F \right)}\ .
\end{equation}
In order to maintain the explicit $N=2$ supersymmetry of (\ref{N=2lowE})
in the dual theory we must rewrite the matter $D$-terms
\begin{equation} \label{dualprepotential}
{\partial {\cal F}(A) \over \partial A} \bar{A} \ =\
{\partial {\cal F_D}(A_D) \over \partial A_D} \bar{A}_D \ ,
\end{equation}
with $\tau \ =\ \partial^2 {\cal F}(A)/ \partial A^2$ and
$\tau_D \ =\ \partial^2 {\cal F}_D(A_D)/ \partial A_D^2$. Thus
\begin{equation} \label{taudfda}
A_D \ =\ {\partial {\cal F(A)} \over \partial A}\ , \qquad
\tau \ =\ {\partial A_D \over \partial A}\ .
\end{equation}
This manipulation is independent of whether $\tau$ is a coupling
constant or, as in our enlarged model, a chiral superfield itself.
Duality therefore continues to hold at low energies even after
supersymmetry is broken by for example $\langle f_\tau \rangle \neq 0$,
induced by $f_{\tau_0} \neq 0$; in the dual theory, the trivially related
$\langle f_{\tau_D} \rangle$ drives supersymmetry breaking.
The theory is also invariant under shifts by integer $n$ in the real part
of $\tau$, since these correspond to unobservable shifts in the $\theta$
angle. Writing $\tau$ as the ratio of the two components of a vector
allows us to recognize the duality and $\theta$ angle transformations
\begin{equation} \label{tautransform}
\left( \begin{array}{cc} a & b\\ c & d \end{array} \right)
\left( \begin{array}{c} \tau \\ 1 \end{array} \right)
\end{equation}
as generating the group $SL(2,Z)$.
Now consider the effective theory for scales much less than $U$.
Its $F$-terms must be holomorphic and invariant under the global
$U(1)_R$ symmetry of (\ref{U1Rcharges}), and therefore of the form
\begin{equation}
{\cal L}_{\rm gauge} \ =\ {1\over 8\pi}{\rm Im} \left[
\tau(\tau_0,U)\, W_\alpha W^\alpha \Big|_F \right] \ .
\end{equation}
If $\tau$ were everywhere analytic in $U$, Im$(\tau)$ would be a harmonic
function, which would be unbounded below, resulting in an imaginary gauge
coupling. To avoid this $\tau$ must have singularities at finite values
of $U$, $U_i$, presumably due to composite states driven massless by
strong interactions. Thus the dual theory is weakly coupled near the
singularities. The composite (monopole or dyon) fields are light near
the singular points and must be included in the effective theory.
We therefore include ``hypermultiplet'' terms compatible with the
symmetries in the effective Lagrangian:
\begin{equation} \label{Lmonopole}
{\cal L}_{\rm dyon} \ =\ M^\dagger e^V M \Big|_D
\ +\ {\tilde M}^\dagger e^{-V} {\tilde M} \Big|_D
\ +\ \sqrt{2} A_D M {\tilde M} \Big|_F \ \ +\ {\rm h.c.}\ ,
\end{equation}
where $M$ and ${\tilde M}$ are the two $N=1$ chiral multiplets of an $N=2$
hypermultiplet, and $A_D \sim (U-U_i)$ determines the bound states' mass.
The simplest possibility for $\tau$, consistent with its weak-coupling limit
at large $U$ and with ${\bf Z_2}$ symmetry $U \leftrightarrow -U$, is a pair
of logarithmic singular points at each of which a single composite state
becomes massless. An $SL(2,Z)$ invariant function with two such singularities
is determined up to scalings by its behavior there and at infinity, and can be
interpreted as the modular parameter of a torus
(with $\tau$ the ratio of its two periods).
The torus described by $\tau$ corresponds to a cubic elliptic curve with roots
$e_1, e_2, e_3$,
\begin{equation} \label{ellcurve}
y^2 \ =\ 4 (x - e_1) (x - e_2) (x - e_3) \ .
\end{equation}
Up to $SL(2,Z)$ transformations, $\tau$ can be written
\begin{equation} \label{tau}
\tau \ =\ {\partial A_D \over \partial A} \ =\
{\int_{e_1}^{e_2} {dx / y} \over \int_{e_2}^{e_3} {dx / y}}\ .
\end{equation}
Seiberg and Witten showed that if the dyons are respectively a $(0,1)$
magnetic monopole and a $(1,-1)$ dyon, then one obtains a function $\tau$
consistent
with the one loop beta function at $U \rightarrow \infty$:
$\tau(U\rightarrow\infty) \sim -(i/\pi) \ln U$.
The charges are redefinable by circuiting infinity, which induces
$(n_m,n_e)\rightarrow (-n_m,-n_e - 2 n_m)$.
Furthermore, changing the $\theta$ angle of $\tau$ by $2\pi$ simply
switches the labelling of the monopole and dyon, leaving the two singular
points $U_i$ physically equivalent, as the ${\bf Z_2}$ symmetry requires.
For the pure gauge $N=2$ $SU(2)$ model, these boundary conditions
correspond to an elliptic curve
\begin{equation} \label{ellipticSU2}
y^2 \ =\ (x - U_i)(x + U_i)(x - U) \ ,
\end{equation}
where $U_2 = -U_1$ by the ${\bf Z_2}$ symmetry. The period integrals are
\begin{eqnarray} \label{aadintegrals}
A &\ =\ & {\sqrt{2}\over\pi} \int^{U_i}_{-U_i}
{dx \sqrt{x - U} \over \sqrt{x^2-U_i^2}} \\
&& \nonumber \\ A_D &\ =\ & {\sqrt{2}\over\pi} \int_{U_i}^U
{dx \sqrt{x-U} \over \sqrt{x^2-U_i^2}}\ .
\end{eqnarray}
These integrals can be expanded in powers of $z \equiv (U - U_i)/U_i$
with appropriate hypergeometric function expansions \cite{abramowitz},
yielding near $U = U_i$
\begin{eqnarray} \label{zexpansion}
A &\ \sim\ &\sqrt{U_i} \ \left( \ {4\over\pi}\ +\
{z\over 2 \pi} ( 1 - \ln{z\over 32} ) \ +\ {z^2\over 32 \pi}
( {3\over 2}\ +\ \ln{z\over 32} ) \ +\ {\cal O}(z^3) \right)
\nonumber \\
A_D &\ \sim\ & i\sqrt{U_i} \ \left( {z\over 2} \ -\ {z^2\over 32}
\ +\ {3 z^3\over 2^9}\ + \ {\cal O}(z^4) \right)
\nonumber \\
\tau_D &\ \sim\ & {i\over\pi} \left( -\ln{z\over 32}\ +\ {z\over 4}
\ -\ {13\over 256}z^2 \ +\ {\cal O}(z^3)\right)\ ,
\end{eqnarray}
The singular point $U_i$ can be obtained by matching to the one-loop
beta function result, in the perturbative large-$U$ regime, and up to
threshold corrections represented by a constant $c$ of order one, is
\begin{equation} \label{Lambdadefn}
U_i \approx \Lambda^2 \equiv c\Lambda_{UV}^2 \exp{(i\pi\tau_0)}\ ,
\end{equation}
where $\Lambda_{UV}^2$ is the Wilsonian scale associated with the ``bare''
Lagrangian. The scalar potential in the dual theory, in the vicinity
of either singular point where the dual theory is weakly coupled, is
thus (to order $|a_D|^2 / \Lambda^2$)
\begin{equation} \label{scalarpotential}
V\ \approx\ 2 |a_D|^2 \left(|a_M|^2 + |a_{\tilde M}|^2\right)
\ -\ \pi^2 \ { \left(|a_M|^2 - |a_{\tilde M}|^2\right)^2
\ +\ 16 |a_M a_{\tilde M}|^2 \over 2\ln{|a_D/16\Lambda|} }\ .
\end{equation}
This is minimized by $a_M = a_{\tilde{M}} = 0$, leaving $a_D$ as
a quantum moduli space. At $a_D = 0$ the monopoles are massless.
Before introducing $N=2$ supersymmetry breaking mass terms into the
model, it is worth discussing the region of validity of the effective
theory discussed above. The Wilsonian scale $\mu^2$ must
stay below the scale $U$ where the $SU(2)$ gauge group is broken to
the $U(1)$ subgroup. Furthermore, the monopole fields must have masses
in excess of $\mu$, since we have cut off the monopoles' contribution
to the $\tau$ function at their mass $\sim (U-U_i)/\sqrt{U_i}$.
(If instead the monopole mass were less than $\mu$, then some of the
gauge coupling's running would be due to loops computed in the low
energy effective theory with internal momenta below $\mu$,
and the $\tau$ function would not depend on $(U-U_i)/\sqrt{U_i}$.)
The vacuum structure of the theories can be obtained from the limit
$\mu \rightarrow 0$, so the effective theory is useful even close to
the point $a_D = 0$ where the monopoles are light.
If we are to extend the $N=2$ model's effective theory to models with
$N=2$ breaking masses these models must continue to exhibit an energy
range with a $U(1)$ gauge symmetry. As the $N=2$ breaking masses grow
we expect the theories' mass gap to increase (since the cancellations
preventing confinement of the $SU(2)$ gluons will break down). For
an energy range with a $U(1)$ gauge symmetry to exist, the scale at
which the $U(1)$ gauge boson is confined by the $SU(2)$ gauge dynamics,
and also therefore the $N=2$ breaking masses, must be less than the
$SU(2)$ breaking scale $U$. The
resulting tractable models occupy a ``ball'' in parameter space around
the $N=2$ model. At tree level these models possess an $SU(2)$ gauge
symmetry; a light adjoint of fermions; and a light adjoint of scalars,
not necessarily degenerate with the fermions.
We cannot rule out the possibility that the models' qualitative
properties change dramatically as the soft breakings are taken larger
than $\Lambda$ (necessary to recover the QCD limit), although for
small soft breakings we find that the behavior is smooth.
Even though introducing $N=2$ breaking terms lifts the degeneracy of
the moduli space, we can still discuss monodromies of the effective
$\tau$ function in what was the moduli space by coupling a chiral
superfield source $J(x)$ to the adjoint field $\Phi(x)$.
The additional term
\begin{equation} \label{source}
{\cal L}_J\ =\ \int d^2 \theta\ J \Phi \ +\ {\rm h.c.}
\end{equation}
shifts $\langle a_{\Phi} \rangle$ away from its ($J=0$) vacuum value
when $\langle f_J \rangle \neq 0$.
The additional interaction in (\ref{source}) can of course induce new
$J$-dependent terms in the effective Lagrangian ${\cal L}_{eff}^{J}$.
However, for small $N=2$ breaking masses a background source with
magnitude of order the small breaking mass is sufficient to explore
the monodromies around the original $N=2$ singular points. In the
case of the gauge kinetic term, the lowest order term induced by $J$,
\begin{equation} \label{Jinduced}
(\ J^\dagger J \ (D W / \Lambda^3)^2 \ ) \Big\vert_D \ \sim\
\vert f_J / \Lambda^3 \vert^2 \ F_{\mu\nu}^2 \ ,
\end{equation}
is clearly a subleading contribution near the $N=2$ limit, leaving the
monodromies and singularities of the broken model still controlled by
the $N=2$ original $\tau$ function. We therefore expect to find light
monopoles or dyons in the effective theory for $\langle a_\Phi \rangle$
near the $N=2$ singular points.
\subsection{Breaking to $N=1$ supersymmetry} \label{sec-2.3}
The $N=2$ supersymmetric model discussed in section~\ref{sec-2.2}
can be broken to an $N=1$ model by introducing a mass term for the
adjoint matter fields \cite{SW,IS}, corresponding in the enlarged model
to allowing $\langle a_m \rangle \neq 0$.
The effective Lagrangian is thus that of the $N=2$ model plus terms,
invariant under the $N=1$ supersymmetry and gauged and global $U(1)$
symmetries of the model, that vanish as the adjoint mass is turned off.
The effective superpotential can therefore receive corrections of the
form
\begin{equation} \label{N=1F}
\Delta W \ =\ m f(U/\Lambda, \tau_0)\Big|_F
\end{equation}
where $f$ is an unknown function. This unknown mass renormalization
arises from interactions of $A$ with $SU(2)$ gauge bosons at scales
above $U$.
Introducing this mass term can also give rise to additional
$D$-terms of the form
\begin{equation} \label{N=1D}
\Delta {\cal L}_D \ =\ m^\dagger m\ G \left( {A_D \over \Lambda} ,
{A_D^\dagger \over \Lambda}, \tau_0,
\tau_0^\dagger, {MM^\dagger \over \Lambda^2},
{\tilde{M} M^\dagger \over \Lambda^2},
{\tilde{M} \tilde{M}^\dagger \over \Lambda^2} \right)\Bigg|_D
\end{equation}
plus higher dimension terms in the gauge fields.
In general $G$ is some complicated unknown function\footnote{
Functions like $G$ induced by supersymmetry breaking appear in
$D$-terms of the effective Lagrangian, here and below. While there
are some constraints on these functions, for example from weak
coupling behavior and nonsingularity of the Kahler metric, these
are usually not sufficient to globally determine the behavior.}.
Note that the corrections in (\ref{N=1D}) to the kinetic energy terms
are suppressed by ${\cal O} (m^2 / \Lambda^2)$,and for small $m$ we
can consider them ${\cal O} (p^4)$. In this sense these corrections
do not destroy the ${\cal O}(p^2)$ exactness of the $N=2$ solution.
Upon eliminating $f_{A_D}$ from such terms and the canonical $D$-term
for $A_D$, the potential only receives contributions of the form
\begin{equation} \label{VfromWG}
{1\over 1 + G^\prime\big|_A}\left| {dW\over df_{A_D}}\right|^2
\end{equation}
with $G^\prime$ a function related to $G$. Now, supersymmetric
vacua satisfy $V = 0$, and $G$ is nonsingular unless the effective
theory completely breaks down. The extrema of $W$ thus coincide
with the minimum of the potential, and we have
\begin{equation} \label{neq1eqmot}
\begin{array}{c}
\sqrt{2}\, a_M a_{\tilde M}\ +\ m \left( d a_U / d a_D \right)\ =\ 0
\ \ \\ \\ a_D a_M\ =\ a_D a_{\tilde M}\ =\ 0 \ .
\end{array}
\end{equation}
As Seiberg and Witten found, the potential is minimized for
$\langle a_D \rangle\ =\ 0$ and the monopoles condense with
\begin{equation} \label{neq1minimum}
a_M\ =\ a_{\tilde M}\ =\ \left( -a_m {a_U}'(0) / \sqrt{2}
\right)^{1/2}\ .
\end{equation}
The theory possesses a mass gap, since the magnetic $U(1)$ gauge
boson acquires a mass by the Higgs mechanism, and electric charges
are confined.
\subsection{Breaking to $N=0$ with nonzero $f_m$} \label{sec-2.4}
As a first example of soft supersymmetry breaking consider giving a
nonzero vacuum expectation value to the $f$-component of $m$ in the
$N=1$ model. At tree level this generates an extra $N=1$ breaking
mass term for $a_\Phi$
\begin{equation} \label{fmterm}
2 {\rm Re}( f_m {\rm Tr}(a_\Phi^2) ) \ ,
\end{equation}
which is renormalized by gauge interactions, through the function $f$
in (\ref{N=1F}), but remains in the effective theory. In addition
the $D$-terms in (\ref{N=1D}) may generate additional terms
\begin{equation} \label{evalfmterm}
|f_m|^2 \ G\left({A_D \over \Lambda},
{A_D^\dagger \over \Lambda}, \tau_0, \tau_0^{\dagger},
{MM^\dagger \over \Lambda^2},
{\tilde{M} M^{\dagger} \over \Lambda^2},
{\tilde{M} \tilde{M}^\dagger \over \Lambda^2} \right)\Bigg|_A
\ +\ a_m f_m G \Big|_F \ +\ {\rm h.c.}
\end{equation}
Firstly, we note that the unconstrained terms are small
(${\cal O}(p^2)$ or higher) when
$m \ll \Lambda$ and $f_m \ll \Lambda^2$. Secondly, introducing the
soft breaking mass leaves the lowest dimension gauge kinetic term
unchanged from the $N=2$ model (ignoring not only the higher
dimension terms mentioned above, but also suppressed terms such as
$m^\dagger m(DW)^2/\Lambda^4|_D \sim |f_m^2/\Lambda^4|(F_{\mu\nu})^2$).
We conclude from the singularity structure of $\tau$ that even in the
softly broken model there remain two points at which monopole bound
states become massless. The soft breaking may generate $a_D$
interactions of unknown sign and similarly contribute to the masses
of the scalar components of the monopole fields.
When $f_m$ is the only $N=2$ breaking parameter, these terms are
suppressed by $f_m/\Lambda$ relative to the bare scalar mass term,
but unfortunately the tree level potential is then unbounded below
as $\langle a_U \rangle \rightarrow -\infty$. Higher dimension
$D$-terms, contributing for example $|a_U|$ terms to the potential,
determine whether it is truly unbounded or is instead minimized at
some finite $a_U$. Such unknown terms may however be small compared
to an additional $N=2$ breaking mass generated by $\langle a_m \rangle
\neq 0$. In the limit $f_m/\Lambda^2 \ll a_m/\Lambda \ll 1$, the bare
masses dominate the corrections to the pure $N=2$ model and we obtain
the potential
\begin{eqnarray} \label{fmscalarpotential}
V\ \approx\ {-8 \pi^2\over \ln{|a_D/16\Lambda|} } \left|
a_M a_{\tilde{M}} - {i \sqrt{2} \Lambda a_m }\right|^2
\ +\ { 2 |a_D|^2 } \left(|a_M|^2 + |a_{\tilde{M}}|^2\right)
\nonumber\\
\ -\ {\pi^2\over 2 \ln{|a_D/16\Lambda|}}
\left(|a_M|^2 - |a_{\tilde{M}}|^2\right)^2
\ +\ 4 {\rm Im} (f_m \Lambda a_D)
\end{eqnarray}
which is minimized (via (\ref{neq1minimum}), up to terms of order
$f_m^2$ and $a_m^2$) by $a_M = a_{\tilde M} = (i \sqrt{2} \Lambda
a_m )^{1/2}$ and $a_D = i 2 \sqrt{2}f_m^*/a_m$.
We thus obtain an ${\cal O} (p^0)$ solution of a model
with $N=0$ supersymmetry.
We conclude that up to small corrections the vacuum structure of a
model with both a small $N=1$ preserving mass and a small soft
supersymmetry breaking mass is equivalent to that of the pure $N=1$
model. There thus exist $N=0$ models which are close in parameter
space to Seiberg and Witten's $N=1$ model that exhibit monopole
condensation and confinement. Unfortunately the terms induced in the
potential by soft breaking from $D$-terms are not determined by the
super-, gauge or global symmetries of the model and thus we draw no
conclusions as to the vacuum structure when
$\langle a_m \rangle \rightarrow 0$.
This theme will recur in the discussion below.
\subsection{Breaking to $N=0$ with a gaugino mass} \label{sec-2.5}
The $N=2$ model can be perturbed directly to an $N=0$, softly broken
supersymmetric model by including a gaugino mass term. In the
enlarged model such a mass can arise from setting $\beta_\tau \neq 0$
and hence $\langle f_{\tau_0} \rangle \neq 0$. The effective
Lagrangian is again that of the $N=2$ model plus, potentially, all
additional terms consistent with the symmetries of the model that
vanish as the gaugino mass is switched off. The lowest dimension
gauge kinetic term is given by the $\tau$ function of $N=2$ and hence
the singularities indicating the presence of massless monopole fields
are unchanged by the soft supersymmetry breaking. There are however
extra terms generated in the dual theory close to $U_i$, arising from
the terms in the $N=2$ effective Lagrangian when
$\langle f_{\tau_0} \rangle \neq 0$. In particular, $f_{\tau_0}$
enters through $U_i$'s dependence (\ref{Lambdadefn}) on $\tau_0$;
to second order in $a_z$, with $z \equiv \left(U - U_i \right)/U_i$,
we have
\begin{eqnarray} \label{dualz2potential}
V &\approx &
\left|{\Lambda f_{\tau_0} \over 16}\right|^2
\left( -\ln{|a_z|\over 32}
+ {a_z + a_z^*\over 8}\left(\ln{|a_z|\over 32} + 1\right)
\right)^{-1}
\nonumber\\
&& \left|
8 - {32\sqrt{2}\pi \over \Lambda f_{\tau_0}} a_M^* a_{\tilde M}^*
- 2 a_z\left(\ln{|a_z|\over 32} - 1/2\right)
- a_z^*\left( 1 -
{4\sqrt{2}\pi\over f_{\tau_0} \Lambda} a_M^* a_{\tilde M}^*\right)
\right|^2
\nonumber\\
&& - \left|{\Lambda f_{\tau_0} \over 4}\right|^2 \left(a_
- {a_z^2
\over 16}
- {4\sqrt{2}\pi \over \Lambda^* f_{\tau_0}^*} a_M a_{\tilde M} a_z
\right) + {\rm h.c.} \nonumber\\
&& + {\left|\Lambda\right|^2 \over 2}|a_z|^2
\left(|a_M|^2 + |a_{\tilde M}|^2\right)
\nonumber\\
&& + {\pi^2 \over 2} \left(-\ln{|a_z|\over 32} + {a_z + a_z^*\over 8}
\right)^{-1}
\left(|a_M|^2 - |a_{\tilde M}|^2 \right)^2
\ .
\end{eqnarray}
There may in addition be $D$-terms that vanish in the $N=2$ limit
and give contributions to the superpotential of the form
\begin{equation} \label{ftaudirectD}
\Lambda^2 H(\tau_0^\dagger, \tau_0, z, z^\dagger)\Big|_D \ .
\end{equation}
Such terms may induce a scalar monopole mass contribution,
for example
\begin{equation} \label{mpolemassexample}
U_i^\dagger z \Big|_D \sim\ \Lambda^2 f_{\tau_0}^* f_z \ ,
\end{equation}
which on eliminating $f_z$ provide a shift proportional to
$f_{\tau_0}$ within $|f_z|^2$ in (\ref{dualz2potential}).
The resulting cross terms with the monopole fields induce a scalar
monopole mass. Its sign and magnitude, which would signify whether
they condense, is therefore undetermined by the symmetries.
These terms perturb, in an unknown fashion, the potential
(\ref{dualz2potential}).
One can however read off the gaugino mass in the effective theory,
arising from the $f$-component of the $\tau$ function, as
\begin{equation} \label{gluinomass}
m_{\tilde{\gamma}_D} \ =\ {m_{\tilde\gamma} \over a_{\tau}}
\ \sim\ \tau_D^\prime (a_{\tau_0}) f_{\tau_0}
\ \sim\ -f_{\tau_0} {\Lambda^2 \over a_D}\ .
\end{equation}
The gaugino and dual gaugino masses diverge as $a_D \rightarrow 0$,
which we may interpret as the decoupling of the massive gaugino
fields from the effective theory, as the decreasing monopole mass
drives its region of validity to zero. If the theory behaves
without singularity as $\langle f_{\tau_0} \rangle \rightarrow 0$,
the contribution of (\ref{ftaudirectD}) to the potential must cause
the vacuum expectation $\langle a_D \rangle$ to become nonzero and
proportional to $\langle f_{\tau_0}\rangle ^l$ with $l<1$, so that
$m_{\tilde{\gamma}_D}$ smoothly approaches zero in the $N=2$ limit.
As before if the gaugino mass is introduced as a perturbation to the
$N=1$ model ($\langle f_{\tau_0} \rangle \ll \langle a_m \rangle$),
then the corrections to the potential from the gaugino mass will be
small and the minima will be that of the $N=1$ model with monopole
condensation.
\subsection{Breaking to $N=0$ with a squark mass} \label{sec-2.6}
Allowing $\langle f_K \rangle \neq 0$ in (\ref{smallLagrangian})
generates a soft supersymmetry breaking squark mass, induced at tree
level by the term $|f_K|^2 |a_{\Phi}|^2$. Since $K$ has arbitrary
$U(1)$ charge it could not appear in the $F$-terms of the effective
theory, but it does appear in $D$-terms of the form
\begin{eqnarray} \label{effectiveKterm}
{\cal L}_K &\ =\ &\left[ K^\dagger K\ \Phi^\dagger \Phi\ G_1
\left(\tau_0, \tau_0^\dagger,
{\Phi^\dagger \Phi \over \Lambda^2},
{\Phi^2 \over \Lambda^2},
{M^\dagger M \over \Lambda^2},
{{\tilde M}^\dagger {\tilde M} \over \Lambda^2} \right)
\right. \nonumber \\ && \left.
\ +\ \ K^\dagger K\ \Phi^2 \ G_2 \left(\tau_0, \tau_0^\dagger,
{\Phi^\dagger \Phi \over \Lambda^2}, {\Phi^2 \over \Lambda^2},
{M^\dagger M \over \Lambda^2},
{{\tilde M}^\dagger {\tilde M} \over \Lambda^2} \right)
\right]_D \ ,
\end{eqnarray}
with $G_1$ and $G_2$ unknown functions. The bare squark mass is
preserved up to gauge renormalization, but the induced monopole mass
is undetermined and hence so is the vacuum structure. The $\tau$
function of the $N=2$ model is again preserved, indicating the
presence of massless monopole states at $a_D = 0$. The potential
minima when the $N=1$ model is perturbed will still lie close
(for small $\langle f_K \rangle$) to $a_D = 0$ and $a_M$ given by
(\ref{neq1minimum}), not altering the vacuum structure from the pure
$N=1$ case.
\section{$SU(N_c)$ Gauge Symmetry And Matter Fields} \label{sec-matter}
The exact results for $N=2$ supersymmetric QCD with an $SU(2)$ gauge
group have been generalized to the case of an $SU(N_c)$ gauge group
\cite{Ncolor}. The potential for the adjoint scalars is minimized by
a vev that classically breaks $SU(N_c)$ to $U(1)^{N_c}$. For large
scalar vev the $U(1)$ couplings can again be calculated in perturbation
theory and shown to be
consistent with the existence of $N$ points on the quantum moduli space
at each of which $N-1$ dyons become massless. The function $\tau$ is
again exactly specified (up to scaling) by the singularity structure.
Seiberg and Witten have also demonstrated how $N=2$ matter multiplets
can be included in the $SU(2)$ model. For heavy matter fields, the
$SU(2)$ dynamics are invariant at low energies, except for the
addition of an extra singularity at large $U = m/\sqrt{2}$
corresponding to where a matter field mass vanishes due to its Yukawa
coupling to the scalar vev. As the mass decreases, the singularity
moves towards the origin of the moduli space and merges with the pure
glue singularities as dictated by the requirement that the dyons fill
out multiplets of the relevant flavor symmetries. Again the $\tau$
function is exactly determined by the monodromy structure.
These results can also be generalized to $SU(N_c)$ gauge models
\cite{matterN}.
In each of these cases the exact form of the $\tau$ function in the
effective theory is known (up to higher dimensional contributions).
We can break the models to $N=0$ models with squark and gaugino
masses by promoting the bare couplings $\tau_0$ and $K$ to the
status of chiral superfields and allowing them to acquire nonzero
$f$-component vevs. As for the $SU(2)$ theory, the $D$-terms will
generate unknown contributions to the potential, but the gauge
multiplet's kinetic term is given exactly by $\tau$.
If the soft breaking masses are small relative to an $N=1$ preserving
adjoint matter mass then the models are pinned near the singularities
and the monopoles condense.
Among these models is the particularly interesting case of an
$SU(3)$ gauge symmetry and $N=1$ matter multiplets in the fundamental
representation. We may choose to introduce an $N=1$ supersymmetry
preserving mass for the adjoint matter field and soft supersymmetry
breaking masses for the gauginos and fundamental representation
scalars. At tree level these masses can be raised until the fields
are integrated from the theory, leaving QCD with quarks. Although we
cannot explicitly take this limit in the effective theory, since the
light degrees of freedom there are not those appropriate to QCD, we may
begin the interpolation from the $N=2$ result towards QCD. The behavior
of the theory is that monopoles condense and the quarks are confined.
This picture's consistency with the t'Hooft--Mandlestam picture
\cite{dualqcd} of quark confinement in QCD suggests that the models
are smoothly connected.
\section{Discussion} \label{sec-discussion}
We have embedded the $N=2$ and $N=1$ supersymmetric models possessing
exactly calculable results within a larger $N=1$ model, which in
certain limits induces soft supersymmetry-breaking masses in the
original theories. This procedure generates new contributions to
the effective Lagrangian, which unfortunately holomorphy and $U(1)$
symmetries do not completely determine. Thus the information
we are able to extract about the behavior of corresponding $N=0$
models is limited. However, at leading order in the breaking masses
the gauge kinetic terms are still completely determined, leaving
unchanged the $\tau$ function of the $N=2$ and $N=1$ models.
Since the singular behavior of $\tau$ is consistent with the original
ansatz that a monopole becomes massless at each of the singular
points, we expect that these are still the only singularities in the
low-energy effective Lagrangian of the nonsupersymmetric models.
If the soft supersymmetry breaking masses are induced as sufficiently
small perturbations to the $N=1$ preserving model, then monopole
condensation and confinement are preserved and seen not to depend on
the presence of exactly massless gauginos or squarks, nor on exact
supersymmetry degeneracies.
While we have encountered some limitations in applying Seiberg and
Witten's techniques to nonsupersymmetric models, the analysis
nevertheless reveals some details of condensation and confinement.
Our results are consistent with the supersymmetric models being
smoothly connected to nonsupersymmetric QCD-like models.
\begin{flushleft} {\Large\bf Acknowledgments} \end{flushleft}
The authors are grateful to Luis Alvarez-Gaum\'e, Jacques Distler,
Mans Henningson, Greg Moore, Nati Seiberg and Matt Strassler for
useful discussions and comments. This work was supported in part
under United States Department of Energy contract DE-AC02-ERU3075.
SDH acknowledges the hospitality of the Aspen and the Benasque
Centers for Physics, where some of this work was performed.
\newpage
\baselineskip=1.6pt
|
\section{Introduction}
The COBE experiment (Bennet et al. 1992, 1994; Smoot et al. 1992) has
stimulated a considerable amount of work on cosmic structure. Current tests
usually exploit the angular correlation function and several harmonic
amplitudes of the sky temperature field (see e.g, Adams et al. 1992;
Kashlinsky 1992; Efstathiou et al. 1992; Kofmann et al. 1993; Gorski et al.
1994). However, several more tests have been suggested over the years for a
thorough investigation of the properties of the anisotropy field of the
cosmic background radiation. These often involve the distribution and
features of hot and cold spots, which can provide useful checks of the
Gaussian nature of the fluctuations (Sazhin 1985; Bond \& Efstathiou 1987;
Coles \& Barrow 1987; Coles 1988; Mart\'\i nez-Gonz\'alez \& Sanz 1989; Gott
et al. 1990). Measurable quantities include the number of spots $N_{{\rm iso}%
}$ defined by isotemperature contours, the spot boundary curvature or genus $%
G$, the spot excursion area, and so on. A first analysis of COBE-DMR maps
along these lines has been performed by Torres (1994). A definition of
spots independent of isotemperature contours considers local maxima and
minima of temperature (Bond \& Efstathiou 1987, Vittorio \& Juszkiewicz
1987), and is thereby not connected to the topological features of spots.
Differing from $N_{{\rm iso}},$ the dependence of the number of (positive
and/or negative) peaks $N_{{\rm peak}}$ on threshold is not universal for
Gaussian fields. This latter approach was adopted by Fabbri \& Natale (1993,
1995) in studies of the 2-dimensional distribution of extragalactic IRAS
sources, but has not yet been applied to the cosmic background radiation.
In this work we analyze the statistics of local maxima and minima in
COBE-DMR 2-year maps. We found that in this kind of analysis the detector
noise must be taken into account very carefully since $N_{{\rm peak}}$ is
sensitive also to high order harmonics where noise dominates (cf. Fabbri
1992). However, due to a highly nonlinear dependence of $N_{{\rm peak}}$ on
the harmonic strengths, the presence of structured signals in COBE maps
reduces its value below the level measured in pure noise maps. (The
identification of genuine peaks in the radiation temperature is not required
at all in our analysis.) We find that the distributions of positive and
negative peaks are mutually consistent, and the results from this statistics
agree with those of earlier tests. Therefore, we find no evidence for
non-Gaussian features in the fluctuations. More precisely, fitting Gaussian
power-law models of cosmic structure to the peak distribution we recover a
clear anticorrelation between the spectral index $n$ and the predicted rms
quadrupole $Q_{{\rm rms-PS}}$ (Seljak \& Bertschinger 1993; Smoot et al.
1994): We get $Q_{{\rm rms-PS}}=17\pm 3$ $\mu $K for $n=1$ and $14\pm 3$ $%
\mu $K for $n=1.5$, where the error bars include uncertainties deriving from
the treament of noise as well as cosmic variance. These numbers, altough
they agree with previous evaluations of the quadrupole from higher order
harmonics, are not consistent with its direct determination providing $Q_{%
{\rm rms}}=6\pm 3$ $\mu $K (Bennet et al. 1994). So the recently discovered
discrepancy is confirmed by the properties of the peak distribution, which
depend on the harmonic content of the angular distribution up to $\ell \sim
50.$
\section{The peak number test}
We analyzed the 2-year 53(A+B) DMR maps processed with a 2.9$^{\circ }$
smoothing (Wright et al. 1993) and dipole subtraction. Considering the
Northern and Southern hemispheres separately, we constructed two
pole-centered maps, each containing 12892 pixels, using the coordinate
transformation $\theta _1=2\sin \left[ \frac 12\left( \frac \pi 2-\left|
b\right| \right) \right] ,$ $\phi _1=l,$ with $b$ and $l$ the Galactic
coordinates. After masking low Galactic latitudes, $\left| b\right|
<20^{\circ },$ we were left with 8412 pixels per map. We then looked for
temperature peaks using the algorithms of Fabbri \& Natale (1993, 1995).
Table 1 gives the no-threshold numbers of peaks, both actually detected and
extrapolated to the entire sky and to the North and South hemispheres (2nd
and 3rd column, respectively). Figure 1 reports the extrapolated numbers vs.
a threshold factor $\nu $. This is the peak height normalized to the sky rms
fluctuation $C^{\frac 12}(0)$; for cold spots, $N_{{\rm peak}}$ gives the
number of minima below $-C^{\frac 12}(0)\nu $. For distributions of only
positive {\it or} negative peaks the statistical errors at 1-sigma
confidence levels are evaluated as $\left( N_{{\rm peak}}/f_{{\rm U}}\right)
^{\frac 12},$ with\/ $f_{{\rm U}}=0.652$ the unmasked fraction of the sky.
Within such error limits, we find no significant difference between the
distributions of positive and negative peaks; this result provides support
for the Gaussian nature of cosmic perturbations. In the figure we also
report the full-sky average number of positive and negative peaks. This
average will be compared with theoretical models below, because of the
smaller (by a factor $\sqrt{2})$ relative error.
\begin{table*}
\caption[ ]{No-threshold peak numbers \label{peaks}}
\begin{flushleft}
\begin{tabular}{llll}
\hline \noalign {\smallskip}
Peak set & $(A+B)$ Maps & $(A+B)$ Maps & $(A-B)$ Maps\\
& $|b| > 20^{\circ}$ & Extrapolated$^{\dagger }$ & No
mask \\
\noalign {\smallskip}
\hline \noalign {\smallskip}
Hot, North & 50 & $76.6 \pm 10.8 $ & 95 \\
Cold, North & 47 & $72.0\pm 10.5 $ & 99 \\
Hot, South & 48 & $73.6 \pm 10.6 $ & 105 \\
Cold, South & 53 & $81.2 \pm 11.2 $ & 97 \\
Hot, total & 98 & $ 150.2 \pm 15.2 $ & 200 \\
Cold, total & 100 & $ 153.3 \pm 15.3 $ & 196 \\
Hot/Cold Average & 99 & $ 151.7 \pm 10.8 $ & 198 \\
\noalign {\smallskip}
\hline
\end{tabular}
$^{\dagger }$To full sky or hemispheres, including masked regions
\end {flushleft}
\end{table*}
\begin{figure}
\picplace{8.0cm}
\caption[ ] {The peak number $ N_{\rm peak}$ vs. threshold in
COBE-DMR maps. Open
squares (triangles) denote positive (negative) peaks. The four lowest curves
give the peak numbers extrapolated to
the Northern (dotted lines) and Southern (dash-dotted) halves of the sky.
The remaining curves refer to the full sky.
Filled circles give the average numbers
of positive and negative peaks. The full line describes a power-law model with
$n = 1$ and $Q_{%
{\rm rms-PS}} = 18.5$ $ \mu $K, corresponding to the fit
procedure S31 in Table 2}
\end{figure}
\begin{figure}
\picplace{8.0cm}
\caption[ ] {The noise harmonic coefficients
$A_{{\rm N}\ell }^2$ (containing the form factors $W_\ell ^2$) and the beam
shape coefficients $G_\ell $ vs. $\ell.$
Filled circles give $A_{{\rm N}\ell }^2$ as obtained from the noise map,
and the dotted line represents
the best fit curve corresponding to $\sigma _{\rm N}=0.936^{\circ }$
and $C_{\rm N}=5.74\times 10^{-5}$ mK$^2.$
Crosses give the experimental $G_\ell $, compared to Gaussian
approximations with
dispersion 2.9$^{\circ }$ (full line) and 3.1$^{\circ }$ (dash-dotted) }
\label{figure1}
\end{figure}
For the interpretation of the above data, we must consider that both cosmic
signal and noise contribute to the temperature field $T(\theta ,\phi )=\sum
a_{\ell m}Y_{\ell m}(\theta ,\phi )$. The expectation values of the numbers
of local maxima and minima are determined by the variances $a_\ell
^2=\sum_m\left\langle \mid a_{\ell m}\mid ^2\right\rangle $. For a Gaussian
field the full-sky $N_{{\rm peak}}$ is given by
\begin{eqnarray}
\label{Npeak} & &\left\langle N_{{\rm peak}}(\nu )\right\rangle =
\sqrt{\frac 2\pi }
\theta^{*}{}^{-2}\left\{ \gamma ^{*2}\nu {\rm exp}\left( -\frac 12\nu ^2\right)
\right. \nonumber \\
& &+ \frac{\gamma ^{*}\left( 1-\gamma ^{*2}\right) ^{\frac 32}}{\sqrt{2\pi }}
{\rm %
exp}\left( -u_2^2\nu ^2\right) +\sqrt{\frac \pi 6}{\rm erfc}\left( \sqrt{%
\frac 32}u_1\nu \right) \nonumber \\
& & - \frac 12\int_\nu ^\infty \left[ \gamma ^{*2}(y^2-1){\rm %
exp}\left( -\frac 12y^2\right) {\rm erfc}\left( \gamma ^{*}u_2y\right) \right.
\nonumber \\
& &+ \left. \left. u_1%
{\rm exp}\left( -\frac 32u_1^2y^2\right) {\rm erfc}\left( \gamma
^{*}u_1u_2y\right) \right] {\rm d}y \right\} ,
\end{eqnarray}
where
\begin{equation}
\label{u1u2}u_1=\left( 3-2\gamma ^{*2}\right) ^{-\frac 12},\;\;\;u_2=\left[
2\left( 1-\gamma ^{*2}\right) \right] ^{-\frac 12},
\end{equation}
and the properties of the anisotropy field are summarized by the parameters $%
\theta ^{*}$ and $\gamma ^{*}$ (Bond \& Efstathiou 1987; Fabbri 1992) given
by
\begin{eqnarray}
\label{tetastar} \theta ^{*2} & = & \left[ 2\sum_\ell \ell \left( \ell
+1\right)
W_\ell ^2a_\ell ^2\right] \nonumber \\
& \times & \left[ \sum_\ell \left( \ell -1\right) \ell \left(
\ell +1\right) \left( \ell +2\right) W_\ell ^2a_\ell ^2\right] ^{-1},
\end{eqnarray}
and
\begin{eqnarray}
\label{gammastar}\gamma ^{*2} & = & \left[ \sum_\ell \ell \left( \ell +1\right)
W_\ell ^2a_\ell ^2\right] ^2\left[ \sum_\ell W_\ell ^2a_\ell ^2\right]^{-1}
\nonumber \\
& \times & \left[ \sum_\ell \left( \ell -1\right) \ell \left( \ell +1\right)
\left( \ell +2\right) W_\ell ^2a_\ell ^2\right] ^{-1}.
\end{eqnarray}
Here $W_\ell $ are form factors taking into account beam shape and any
additional smearing effect.
Assuming that signal and noise are uncorrelated, their contributions to $%
a_\ell ^2$ add up in quadrature in the (A+B) maps, $a_\ell ^2=a_{{\rm S}\ell
}^2+a_{{\rm N}\ell }^2$, and we need to determine $a_{{\rm N}\ell }^2$ in an
independent way. This can be achieved by means of the (A$-$B) maps, which
can however be used to directly derive the coefficients $A_{{\rm N}\ell
}^2=W_\ell ^2a_{{\rm N}\ell }^2$ rather than the $a_{{\rm N}\ell }^2$.
Figure 2 reports the results of a harmonic best fit up to $\ell =30$
executed on the entire celestial sphere$.$ Increasing the number of
harmonics, best fits (as well as direct integration by means of $a_{\ell
m}=\int T(\theta ,\phi )Y_{\ell m}^{*}(\theta ,\phi ){\rm d}\Omega $) tend
to overrate large-$\ell $ amplitudes. We checked for this effect by
considering the peak statistics for noise maps.%
\begin{figure}
\picplace{8.0cm}
\caption[ ] {The peak number $ N_{\rm peak}$ vs. threshold in
COBE-DMR noise maps. Open
squares (triangles) denote positive (negative) peaks. The four lowest curves
give the peak numbers extrapolated to
the Northern (dotted lines) and Southern (dash-dotted) halves of the sky.
The remaining curves refer to the full sky. Filled circles give the
average statistics
of positive and negative peaks. The full line represents the best
fitting function}
\label{figure3}
\end{figure}
\begin{figure}
\picplace{6.5cm}
\caption[ ] {The acceptance regions in the $(n,Q_{{\rm rms-PS}})$ plane at 1-
and 2-sigma
confidence levels. Contours are computed with procedures S31 (full lines) and
F31 (dotted). Both procedures assume
$\sigma_{\rm S} = 3.1^{\circ}$,
$\sigma_{\rm N} = 0.936^{\circ},$ and $C_{\rm N}
=5.74 \cdot 10^{-5} $ mK$^2$. The former uses the measured
$A_{\rm N\ell }^2$ up to $ \ell = 30$.}
\label{figure4}
\end{figure}
Figure 3 gives the numbers of maxima and minima detected in the full-sky
noise maps vs. threshold $\nu .$ If using Eq.s (\ref{Npeak}-\ref{gammastar})
we generate the peak statistics from a set of 50 harmonics with best-fitted
amplitudes, we get a low-threshold excess of about 30 peaks with respect to
the data in Fig. 2; the discrepancy increases with the number of harmonics.
Since accurate theoretical calculations of peak statistics require at least $%
\sim $50 harmonics, we tried to fit an analytic form to a more limited set
of $A_{{\rm N}\ell }^2$. A satisfactory choice is
\begin{equation}
\label{anoise}A_{{\rm N}\ell }^2=C_{{\rm N}}\left( \ell +\frac 12\right)
\exp \left[ -\left( \ell +\frac 12\right) ^2\sigma _{{\rm N}}^2\right] .
\end{equation}
However the large error bars on individual $A_{{\rm N}\ell }^2$ make a
2-parameter fit for the function (\ref{anoise}) not very useful. A more
convenient solution to this problem is to fit the parameter $\sigma _{{\rm N}%
}$ directly on the noise-generated peak distribution; using the average
distribution of maxima and minima (represented by the filled circles in
Figure 3) and 99 harmonics the result is $\sigma _{{\rm N}}=0.936^{\circ
}\pm 0.071^{\circ }.$ (We checked however that a set of $\sim 50\div 60$
harmonics would provide a sufficient accuracy.) The fit turns out to be even
``too'' good, providing $\chi _{\min }^2=1.3$ against 9 degrees of freedom:
This means that data points at different thresholds are not uncorrelated.
The value found for $\sigma _{{\rm N}}$ is quite independent of $C_{{\rm N}}$%
, since peak statistics does not depend on the overall amplitude of
anisotropies as shown by Eq.s (\ref{tetastar}, \ref{gammastar}).We then
fitted the amplitude parameter of Eq. (\ref{anoise}) on the reported set of $%
A_{{\rm N}\ell }^2$ with $\ell =2\div 30,$ getting $C_{{\rm N}}=(5.74\pm
0.66)\times 10^{-5}$ mK$^2.$ The distribution described bu the full line in
Fig. 3 was generated using the function (\ref{anoise}) in the entire range $%
\ell =2\div 99.$ We checked that the peak distribution of the (A-B) maps is
also recovered to a good accuracy building up the harmonic spectrum with the
``measured'' $A_{{\rm N}\ell }^2$ up to $\ell =30$ and the function (\ref
{anoise}) at higher $\ell .$
We tested power-law spectra against the peak statistics in Fig. 1 with the
following procedure. For each spectrum, labelled by the spectral index $n$
and the predicted quadrupole $Q_{{\rm rms-PS}}=a_{\rm S2}/\sqrt{4\pi },$ we
generated the theoretical expectation values $a_{{\rm S}\ell }^2$ and then
\begin{equation}
\label{totalAl}W_\ell ^2a_\ell ^2=G_\ell ^2\exp \left[ -\left( \ell +\frac
12\right) ^2\sigma _{{\rm S}}^2\right] a_{{\rm S}\ell }^2+A_{{\rm N}\ell }^2.
\end{equation}
Here $G_\ell $ denote the measured beam-shape coefficients of COBE-DMR.
These have reported by Wright et al. (1994) up to $\ell =50.$ A more
extensive set of $G_\ell $ up to $\ell =99$ has been provided to us by Kogut
(private communication). The exponential factor in Eq. (\ref{totalAl}) takes
into account the 2.9$^{\circ }$ smearing on the map as well as the
additional smearing due to orbital motion, so that our best estimate is $%
\sigma _{{\rm S}}=3.1^{\circ }$. Note that $\sigma _{{\rm S}}$ should not be
confused with the approximate Gaussian beamwidth which was used in many
computations, but not in the present work. Also, it is quite different from
the phenomenological parameter $\sigma _{{\rm N}}$ of Eq. (\ref{anoise}).
The peak distributions calculated from Eq.s (\ref{Npeak}-\ref{gammastar})
are tested against the averaged distribution of positive and negative peaks.
We should notice that such theoretical distribution are affected by several
sources of uncertainty. Cosmic variance (White et al. 1993) affects the
cosmic-structure contribution to the harmonic coefficients $a_{\ell m}$, and
a quite similar variance pertains to noise; in fact, these two effects are
described by identical equations assuming that both are Gaussian processes.
Performing a limited number of simulations for superpositions of Gaussian
signal and noise (with fixed $a_\ell ^2$, the {\em expectation} values of
harmonic strengths), we found that the probability distribution of $N_{{\rm %
peak}}$ at a given threshold is roughly consistent with a Poisson
distribution having a width $\left\langle N_{{\rm peak}}\right\rangle
^{\frac 12}.$ However, in our case the $A_{{\rm N}\ell }^2$ themselves are
not fixed. When we use Eq. (\ref{anoise}) their errors at different $\ell $
are correlated due to the uncertainties on $C_{{\rm N}}$ and $\sigma _{{\rm N%
}};$ it can be shown that the corresponding contribution to the uncertainty
on the predicted $N_{{\rm peak}}$ is given by
\begin{eqnarray}
\label{vars} & &{\rm var}(N_{{\rm peak}}) = \left( \sum_\ell A_{{\rm N}\ell
}^2%
\frac{\partial \left\langle N_{{\rm peak}}\right\rangle }{\partial A_{{\rm N}%
\ell }^2}\right) ^2\frac{{\rm var}(C_{{\rm N}})}{C_{{\rm N}}^2}
\nonumber \\
& & +4 \left[
\sum_\ell \left( \ell +\frac 12\right) ^2A_{{\rm N}\ell }^2\frac{\partial
\left\langle N_{{\rm peak}}\right\rangle }{\partial A_{{\rm N}\ell }^2}%
\right] ^2\sigma _{{\rm N}}^2{\rm var}(\sigma _{{\rm N}}).
\end{eqnarray}
When we use the measured $A_{{\rm N}\ell }^2,$ a more familiar equation $%
{\rm var}(N_{{\rm peak}})=\sum_\ell \left[ \left( \partial \left\langle N_{%
{\rm peak}}\right\rangle /\partial A_{{\rm N}\ell }^2\right) ^2{\rm var}(A_{%
{\rm N}\ell }^2)\right] $ applies. For each computed model we calculated the
corresponding $\chi ^2$ combining quadratically the above errors with with
the experimental error bars reported in Fig.1. This procedure allowed us to
avoid a more extensive use of simulations.
Figure 4 gives the allowed regions at 1 and 2 sigmas in the $(n,Q_{{\rm %
rms-PS}})$ plane. Full lines give the contours calculated for $\sigma _{{\rm %
S}}=3.1^{\circ },$ using the $A_{{\rm N}\ell }^2$ of Fig. 2 up to $\ell =30$%
, and the analytic form (\ref{anoise}) with the optimal values of $\sigma _{%
{\rm N}}$ and $C_{{\rm N}}$ at higher $\ell .$ (This procedure is referred
to as S31 in the Figure and in Table 2.) Contours calculated for $\sigma _{%
{\rm S}}=2.9^{\circ }$ (case S29, not reported in the Figure) would be
hardly distinguishable from them. Using the analytic form (\ref{anoise}) in
the entire range $\ell =1\div 99$ (case F31) the contours are slightly
displaced to smaller values of $Q_{{\rm rms-PS}}$ (by $\sim 2\ \mu $K for $%
n=1)$.
\begin{table}
\caption[ ]{Coefficients of the $n-Q_{\rm rms-PS}$ regression defined by Eq.
(\ref{ab}) \label{regress}}
\begin{flushleft}
\begin{tabular}{lll}
\hline \noalign {\smallskip}
Procedure & $a$ ($\mu$K) & $b$ ($ \mu$K) \\
\noalign {\smallskip} \hline \noalign {\smallskip}
S31$^\dagger$ & $24.71\pm 0.65$ & $ -6.23\pm 0.45$ \\
S29 $^\ddagger$ & $24.54\pm 0.66$ & $ -6.16\pm 0.46$ \\
F31$^{\star }$ & $21.47\pm 0.60$ & $ -5.45\pm 0.42$ \\
\noalign {\smallskip} \hline
\end{tabular}
$^\dagger \sigma_{\rm N} =3.1^\circ$, using noise harmonic
amplitudes from Fig. (2) for $\ell \leq 30$.
$^{\ddagger }\sigma_{\rm N} =2.9^\circ$
$^{\star }$Noise described by Eq. (\ref{anoise}) for any $\ell $
\end {flushleft}
\end{table}
The parameters $n$ and $Q_{{\rm rms-PS}}$ are clearly anticorrelated, as
already found from analyses of harmonic amplitudes and the angular
correlation function (Seljak \& Bertschinger 1993; Torres et al. 1994; Smoot
et a. 1994). Minimizing $\chi ^2$ for fixed $n$ we identify a straight line
in the $(n,Q_{{\rm rms-PS}})$ plane,
\begin{equation}
\label{ab}Q_{{\rm rms-PS}}=a+bn,
\end{equation}
with the coefficient values listed in Table 2. From the above results,
taking into account differences arising from the S31 and F31 procedures, we
can conclude that $Q_{{\rm rms-PS}}=17\pm 3$ $\mu $K for $n=1$, and $Q_{{\rm %
rms-PS}}=14\pm 3$ $\mu $K for $n=1.5$. These numbers agree very well with
the most likely quadrupole $Q_{{\rm rms-PS}}$ derived from higher order
multipoles, but not with the quadrupole rms fluctuation of 6$\pm 3$ $\mu $K
directly fitted on two-year data; see Bennet et al. (1994) for a discussion
of this discrepancy, and Jing \& Fang (1994) for a possible explanation in
terms of an infrared cutoff in the spectrum.
\begin{acknowledgements}
We wish to thank A. Kogut for providing unpublished data on COBE-DMR
beam shape. This work is partially supported by Agenzia Spaziale
Italiana under Contract \# 94-RS-155, by the Italian Ministry for
University and Scientific and Technological Research (Progetti Nazionali e
di Rilevante Interesse per la Scienza), and by the European Union
under Contract \# CI1*-CT92-0013.
\end{acknowledgements}
|
\section{Introduction}
One of the basic assumptions of the standard Friedmann cosmology
is the Cosmological Principle which states
that the Universe is homogeneous and isotropic\cite{Weinberg}
so that
every point of the Universe is equivalent.
While isotropy has been reasonably well established
from, among others, the observation of the cosmic microwave
background radiation by COBE\cite{COBE},
homogeneity has been challenged by various observations
of the Large Scale Structures (LSS).
Recent galaxy redshift surveys such as CfA 1\cite{CfA1},
CfA 2\cite{CfA2}, SSRS1\cite{S1}, SSRS2\cite{CfA1} and the
pencil beam surveys\cite{pencil} have provided evidence for
the LSS such as filaments, sheets,
superclusters and voids, up to $200-300 h^{-1}$Mpc.
The current interpretation based on the observation of the LSS
is that the Cosmological Principle may not be applicable,
at least, to the local Universe,
although it may be applicable to the Universe on a very large scale
with a characteristic distance $\lambda$.
However,
one of the most remarkable consequences of the above
galaxy surveys is that the scale of the largest structures
in each survey is comparable with the extent of the survey itself.
Recently, it has been suggested that
from pencil beam surveys\cite{pencil}
as well as from new deep redshift survey (ESP survey) \cite{ESP},
$\lambda$ should be much larger than the
survey limits $\sim 500-600 h^{-1}$Mpc,
implying the absence of any tendency towards homogeneity up
to the present observational limits.
For example, Pietronero\cite{fractal} has attempted
to explain these phenomena
by suggesting that the LSS shows fractal properties of the Universe.
Even though this non--analytical distribution of matter means that
the Universe is not homogeneous,
the local isotropy preserves the fundamental assumption
that every point of the Universe is equivalent.
Unfortunately, however, a rigorous mathematical description
of such a Universe is extremely
difficult and in practice it is almost impossible.
Therefore, it is desirable to simplify the description of
this $inhomogeneous$ Universe to the extent that its
analytical study becomes possible in order to see,
at least, qualitative features of the matter distribution
and cosmological consequences.
History of cosmological models for an inhomogeneous Universe
dates back to as early as 1930's.
Lemaitre, Tolman and Dingle \cite{Lemaitre}
attempted to describe the evolution of the
$fluctuation$ in the mass distribution. Later in 1947, Bondi
\cite{Bondi} elaborated the model and discussed
observational consequences.
In their model, which we shall call Tolman-Bondi(TB) model,
the global Universe is that of the standard Friedmann cosmology,
implying homogeneity over the region of order $\lambda$.
Recently, in the frame work of the TB model,
Moffat and Tatarski\cite{Moffat} studied,
in order to describe the local inhomogeneity,
cosmology of a local void in the globally Friedmann Universe and
its effect on the measurement of the Hubble constant
and the redshift--luminosity distance relation.
Since the Universe modeled in \cite{Moffat} consists of
many expanding voids
and we happen to be located at the center of one of them,
the shell--crossing singularity occurs,
implying that different shells collide and the comoving
coordinate become meaningless.
Recently, based on the fact that there is no observational evidence of
approaching towards the homogeneity within the survey limit,
another cosmological model\cite{Kim} was proposed,
whose global feature is not asymptotic to the Friedmann cosmology.
In \cite{Kim}, the observable Universe is modeled
as being inside an expanding bubble
with the underdense center and
matter inside the bubble is
isotropically (but inhomogeneously) distributed when viewed
by an observer located at the center.
They proposed that such a Universe may be described by the following
inhomogeneous metric
\begin{equation}
\di \tau^2 = \di t^2 - R^2(t,r)[\di r^2+r^2 \di \Omega^2],
\end{equation}
where $R(t,r)$ is the scale-factor,
dependent on $r$ as well as on $t$.
Therefore, homogeneity in the Cosmological Principle is explicitly
violated, whereas isotropy remains intact.
Based on the high degree of isotropy of the cosmic microwave
background radiation measured by COBE\cite{COBE},
it was assumed that the observer
is located at the center of the bubble (or near it),
albeit the return of the pre--Copernican notion.
Whether this picture is correct or not
can only be decided when the results of the model are confronted
with the observation.
In \cite{Kim}, its cosmological consequences were qualitatively
discussed.
For example, the Hubble constant, the density parameter
and the age of the Universe all became scale--dependent,
whereas the analysis in \cite{Moffat} was simplified
to avoid the possible position--dependent age.
Moreover, because of the lack of the
light propagation solution on which every observation is based,
no explicit and testable cosmological results were derived,
which can be compared with the observation. Therefore,
it is interesting to
examine observational consequences of such a model and compare with
those of the TB model.
In this paper, we first present a general
redshift--luminosity distance relation for a certain class of
inhomogeneous cosmological models. Then we apply the result to
the model discussed in \cite{Kim} as an example.
The plan of this paper is as follows.
In Section II, we present
the redshift--luminosity distance relation
for the case of one $(t,r)$--dependent scale factor in the metric.
In order to proceed further to derive some specific observable
consequences, we have chosen the model proposed in \cite{Kim} as an
example and briefly summarize the model
in Section III.
In Section IV, we derive
modified results of the redshift--luminosity distance
relation and show that they reduce to the well-known
relations in the standard Friedmann cosmology for small $z$, i.e.,
for nearby objects.
Section V deals with some numerical results of this model which are
applicable for large $z$.
Also discussed in this Section is the
observed increase of density parameter
with the redshift in the framework of this model.
A brief summary and conclusions are given in Section VI.
\section{Redshift in a General Metric
with one Scale Factor }
With the exception of astronomical neutrinos and possible
future gravitational waves, most of
the cosmological measurements are based on the electromagnetic
waves, which travel along the null geodesic, i.e., $\di \tau=0$.
Considering only the radial propagation ($\di \Omega =0$),
we have, from Eq.(1),
\begin{equation}
\di r = - \frac{\di t}{R(t,r)}~,
\end{equation}
where the minus sign is chosen since $r$ decreases as $t$ increases.
With a given $R(t,r)$,
Eq.(2) appears as a simple first-order differential equation,
yielding $r$ as a function of $t$. However, complexity of
solving this differential equation becomes immediately apparent
because we will be dealing with the case in which $R(t,r)$
is not factorized into a separable form of $a(t)f(r)$.
The boundary condition to be imposed is as follows.
Since we measure a signal at $r=0$,
the boundary condition is $r(t=t_{received})=0$.
More specifically, we treat the solution of Eq.(2) as $r = r(t,t_0)$
which is a function of $t$
with the boundary condition $r(t=t_0, t_0)=0$.
In order to define the redshift,
we consider two successive wave crests, both of which leave $r$ and
reach us ($r=0$) at different times.
Suppose that two wave crests were emitted at time $t$ and $t+\Dt$
and received by us at time $t_0$ and $t_0+\Dtz$, respectively.
Then, from Eq.(2) and the definition of $r(t,t_0)$, we have
\beqarr
r & =& \int_t^{t_0} \frac{\di t'}{R(t', r(t',t_0))}
\\ \no
& =& \int_{t+\Dt}^{t_0+\Dtz}
\frac{\di t'}{R(t',r(t',t_0+\Dtz))}
{}~.
\eeqarr
It is to be noted that for each wave crest a proper
boundary condition has to be applied,
which is explicitly expressed in the form of $r(t,t_{received})$.
Since $\Dt$ and $\Dtz$ are extremely small
compared with the cosmological time scale,
it is sufficient to consider only up
to the first order in $\Dt$ or $\Dtz$.
Then, we have, from Eq.(3),
\beq
\frac{ \Dt}{R(t,r(t,t_0))}
= \frac{\Dtz}{R(t_0, r(t_0,t_0))}
\left[ 1+R(t_0,0)\int_t^{t_0} \di t' \frac{\rd}{\rd t_0}
\left( \frac{1}{R(t',r(t',t_0))} \right)
\right]~~,
\eeq
yielding the defining relation of the redshift, $z$, as
\beq
1+z \equiv \frac{\Dtz}{\Dt}
=\frac{R(t_0,0)}{R(t,r(t,t_0))}
\left[ 1+R(t_0,0) \int_t^{t_0}\di t'
\frac{\rd}{\rd t_0}\frac{1}{R(t',r(t',t_0))}
\right]^{-1}~~,
\eeq
where the boundary condition, $r(t_0,t_0)=0$ has been used.
Here and hereafter ${\rd \over \rd t_0}$ explicitly means
$\frac{\rd r}{\rd t_0} \frac{\rd}{\rd r}$.
In the case of the Robertson--Walker metric where
$R(t,r)\equiv S(t)$,
the scale factor of the standard Friedmann cosmology,
it is easy to see that Eq.(5) reduces
to the well-known relation, $(1+z)=S(t_0)/S(t)$.
In the real observation,
the most important definition of distance is
the {\it luminosity distance\/}.
As is well known, if a source at comoving distance $r$ emits
light at time $t$ and a detector at $r=0$ receives the light
at time $t_0$,
the luminosity distance, $d_L$, of a source in the standard Friedmann
cosmology is
\beq
d_L = r S(t_0) (1+z) = r S(t) (1+z)^2~~,
\eeq
where the second equality is due to the relation $S(t_0)=S(t)(1+z)$
in the standard Friedmann cosmology.
The luminosity distance in a spherically symmetric but
inhomogeneous Universe
was first examined by Bondi \cite{Bondi} in 1947.
In order to avoid the non-zero pressure,
however, two different scale factors
were introduced in the metric, as was originally
done by Lemaitre, Tolman and Dingle \cite{Lemaitre},
\beq
\di \tau^2 = \di t^2 - X^2(t,r) \di r^2 - Y^2(t,r) \di \Omega^2~~.
\eeq
One of the special features of the TB model is that the pressure
is always zero, as originally designed
so as to be applicable in the matter dominated era only.
The situation considered in \cite{Bondi} is such that a standard
source is at the $center$ and an observer at $(t,r,\theta,\varphi)$.
As was shown in \cite{Bondi}, the ratio of the absolute luminosity
to the apparent luminosity is simply given by $Y^2(t,r)(1+z)^2$.
Two comments are in order here.
First, we note that the light source is at the center,
implying that the light propagates out spherically
with constant surface energy density at any given time,
which is, in general, not the case in an inhomogeneous
Universe.
Secondly, $(t,r,\theta,\varphi)$ is
a coordinate of the observer, not of the source.
That is, in the standard cosmology notation,
$Y(t,r)$ physically corresponds to $rS(t_0)$.
Since we consider the situation in which
the position of the observer is
located at the center, the
light from its source off the center
does not even propagate outward in a
spherically symmetrical manner.
Following the picture of the Universe in \cite{Kim}
where the Universe is inside a bubble with the underdense center
(where the observer is located)
and with the highly dense shell,
the light would feel attraction toward the shell,
implying that the path of the light propagation is,
in general, not a straight line.
Moreover, its energy is not uniformly distributed
over the non-spherical shell at any given time.
Nevertheless, since the position of the observer
is fixed at the center (i.e., $r=0$) and
he/she receives the light that
propagates on a straight line,
we shall use the following definition of luminosity distance
\beq
d_L \equiv r(t,t_0)R(t_0, r(t=t_0,t_0))(1+z)~~,
\eeq
where the $(1+z)$ factor comes from the correction factor,
$(1+z)^2$, that appears in the relationship
between the absolute luminosity and
the apparent luminosity (hence only
one factor out of $(1+z)^2$ in the distance).
One factor of $(1+z)$ in the luminosity relation is due to
the decrease of energy
because of the redshift,
the other factor coming from the increase of the time
interval from $\Dt$ to $\Dtz$, which is also just $(1+z)$
by definition.
Of course, the above definition has to be justified
by performing the coordinate transformation
from the observer to the source in the inhomogeneous Universe.
This, however, is beyond the scope of this paper and thus,
based on its plausibility,
we shall assume its validity in this paper.
We caution the reader that the luminosity distance
should not be simply written as $r(t,t_0)R(t,r(t,t_0))(1+z)^2$,
for $R(t,r(t,t_0))$ is not simply given by $R(t_0,0)$
times a factor, $(1+z)$, as can be seen in Eq.(5).
\section{Scale--dependent cosmology}
In order to obtain some specific results of cosmological
consequences of the proposed inhomogeneous metric, Eq.(1),
we shall consider, as an example, the model of \cite{Kim}.
In this Section, we shall briefly summarize the model.
First, given the metric in Eq.(1), in order to accommodate the
$r$-dependence on the Ricci tensors,
the Einstein equation was also generalized as
\begin{equation}
R^{\mu \nu}-\frac{1}{2}g^{\mu \nu}R=-8\pi [G T^{\mu \nu}](t,r)~,
\end{equation}
where the $(t,r)$ dependence of the combination, $[GT^{\mu \nu}]$,
was explicitly noted.
When the non-vanishing elements of Ricci tensor calculated
from Eq.(1) are substituted into the
generalized Einstein equation in Eq.(9),
we obtain the following non-vanishing components.
\begin{eqnarray}
3 \frac{\dot{R}^2(t,r)}{R^2(t,r)}-2\frac{R''(t,r)}{R^3(t,r)}
+\frac{R'^2(t,r)}{R^4(t,r)}-4\frac{R'(t,r)}{rR^3(t,r)}
&=&8 \pi [G\rho] \\
2 \frac{\ddot{R}(t,r)}{R(t,r)}+\frac{\dot{R}^2(t,r)}{R^2(t,r)}
-\frac{R'^2(t,r)}{R^4(t,r)}-2\frac{R'(t,r)}{rR^3(t,r)}
&=& -8 \pi [Gp_r] \\
2 \frac{\ddot{R}(t,r)}{R(t,r)}+\frac{\dot{R}^2(t,r)}{R^2(t,r)}
-\frac{R''(t,r)}{R^3(t,r)}
+\frac{R'^2(t,r)}{R^4(t,r)}-\frac{R'(t,r)}{rR^3(t,r)}
&=&- 8 \pi [Gp_{\theta}] \\
2 \frac{\ddot{R}(t,r)}{R(t,r)}+\frac{\dot{R}^2(t,r)}{R^2(t,r)}
-\frac{R''(t,r)}{R^3(t,r)}
+\frac{R'^2(t,r)}{R^4(t,r)}-\frac{R'(t,r)}{rR^3(t,r)}
&=&- 8 \pi [Gp_{\varphi}]
\end{eqnarray}
where dots and primes denote, respectively, derivatives with
respect to $t$ and $r$.
Another non-vanishing Ricci tensor $R_{01}$ yields
\begin{equation}
R_{01} = 2 \left( \frac{\dot{R}'(t,r)}{R(t,r)}
-\frac{\dot{R}(t,r)R'(t,r)}{R^2(t,r)} \right)
= - 8\pi [GT_{01}]~~.
\end{equation}
As was discussed in \cite{Kim},
in order to maintain an inhomogeneous matter distribution,
it is essential to keep pressures and $T_{01}$ to be finite
so that $R(t,r)$ is kept from being factored out as $a(t)f(r)$,
in which case the Robertson--Walker metric is recovered.
This feature distinguishes this model from the TB model,
in which pressure was set to be zero to begin with.
Moreover, to avoid the sheer force, it was assumed in \cite{Kim}
that $p_r=p_{\theta}(=p_{\varphi})$. Then
a constraint on $R(t,r)$ is uniquely determined
from Eqs.(11) and (12) as
\beq
R(t,r) = \frac{ a(t) }{1-B(t)r^2}~~,
\eeq
where $a(t)$ and $B(t)$ are positive, arbitrary functions
of $t$ alone.
The negative sign on the right hand side
of Eq.(15) is chosen to avoid a locally closed Universe
(see below Eq.(16)).
Inserting Eq.(15) into Eq.(10) gives
\begin{equation}
\left[ \frac{\dot{R}(t,r)}{R(t,r)} \right]^2
= \frac{8\pi }{3}[G\rho](t,r) + \frac{4B(t)}
{a^2(t)} ~~.
\end{equation}
The term, $4B(t)/a^2(t)$, was interpreted
as a time-varying vacuum energy density in \cite{Kim}.
It should be noted here that the appearance of this
term, admittedly very surprising, is a consequence of
the metric in Eq.(1).
Let us briefly discuss physical implications of Eq.(16).
In our neighborhood (i.e., $r \ll 1$),
the left-hand side of Eq.(16) is reduced to $(\dot{a}/a)^2$,
implying that $a(t)$ represents more or less the scale factor
for our local Universe.
Moreover, since the standard Friedmann cosmology has been
successful in describing our local neighborhood,
any modifications of
the standard cosmology in this model
must be small in the local Universe.
The next question is whether our local neighborhood is flat or open.
If it is $flat$,
$B(t)$ should be treated, as can be seen on the right--hand side
of Eq.(16), as being small in accord with the assumption
of small modifications on the local neighborhood.
If the local neighborhood is $open$, however,
$B(t)$ itself is not small.
Upon writing $B(t)$ as $[1+b(t)]/4$,
Eq.(16) becomes
\begin{equation}
\left[ \frac{\dot{R}(t,r)}{R(t,r)} \right]^2
= \frac{8\pi }{3}[G\rho](t,r) + \frac{1}
{a^2(t)} + \frac{ b(t)}{a^2(t)}~~,
\end{equation}
implying that in the local neighborhood
(i.e., for $r \ll 1$, or equivalently
for $(\dot{R}/R)^2 \simeq (\dot{a}/a)^2$),
small is $b(t)$, but not $B(t)$.
It is interesting to mention here that the metric given by
Eq.(1) with the Einstein equation dictates the behavior of
the energy density,
pressure and momentum density of the Universe.
We first note
that the constraint on $R(t,r)$, Eq.(15), severely restricts
the behavior of $G\rho$, $Gp$ and $GT_{01}$
as functions of $t$ and $r$.
Substituting Eq.(15) into Eqs.(10), (11) and (14),
we obtain the following explicit expressions :
\begin{eqnarray}
\frac{8\pi}{3}G\rho &=&
\left(\frac{\dot{a}}{a} \right)^2
-\frac{1}{a^2}-\frac{ b}{a^2}
+ 2\left(\frac{\dot{a}}{a}\right)\left( \frac{\dot{b}r^2/4}
{1- \left[ {1+b \over 4} \right]r^2}\right)
+\left(\frac{\dot{b}r^2/4}
{1- \left[ {1+b \over 4} \right]r^2} \right)^2
\\
8\pi Gp&=&
\frac{1}{a^2}-
2\frac{\ddot{a}}{a} - \left( \frac{\dot{a}}{a}\right)^2
+\frac{b}{a^2}
- \left( 6 \frac{\dot{a}}{a}+
2\frac{\ddot{b}}{\dot{b}} \right)
\left( \frac{\dot{b}r^2/4}
{1- \left[ {1+b \over 4} \right]r^2}\right)
- 5 \left(
\frac{\dot{b}r^2/4}
{1- \left[ {1+b \over 4} \right]r^2} \right)^2
\\
8\pi GT^{01} &=& \frac{\dot{b}r}{a^2} ~~.
\end{eqnarray}
The following comments are in order here.
{}From Eqs.(18) and (19),
it is easy to see that as $ r$ approaches $\sqrt{\frac{4}{1+b}}$,
we have $p/\rho = -\frac{5}{3}$,
implying a bizarre equation of state.
In the standard inflationary scenario
in which a constant vacuum energy
is responsible for generating an inflationary period,
one has $ p/\rho = -1$.
Therefore, in the model under consideration,
an unusual scalar sector with
time-dependent vacuum energy,
which is as yet to be understood,
may be responsible for the inflation
which is much more rapid than the usual inflation.
By using the arbitrariness of $b(t)$, however,
this singularity with the unusual $p/\rho$ ratio
can be pushed far away from the particle
horizon, so that the local value of $p/\rho$
at the present matter dominated era can be made
to be positively small, as generally expected.
For this reason, in spite of this bizarre behavior,
we shall proceed to use this model as an example
for our following discussions.
\section{Perturbative Approach}
The linear relationship between the redshift and luminosity distance
with a constant coefficient (the Hubble constant)
for nearby objects (for $z \ll 1$) has been well established by
various observations.
In this Section,
we investigate whether or not the redshift--luminosity distance relation
for small $z$ in an inhomogeneous cosmological model discussed
in the previous Section still remains the same as in the standard
cosmology.
\subsection{Locally flat Universe}
We start with the light propagation equation
\beq
\frac{\di r}{\di t}=-\frac{1}{a(t)}+\frac{B(t)}{a(t)}\,r^2,
\eeq
which is determined by two arbitrary functions of $t$,
$a(t)$ and $B(t)$.
But unfortunately Eq.(21) cannot be solved
analytically because it is non-linear in $r$.
For a locally $flat$ Universe with small perturbations
to the standard cosmology,
$B(t)$ should be treated as being very small.
Therefore, we have, from Eq.(21),
\beq
r(t,t_0)=
\int_t^{t_0}\frac{\di t'}{a(t')}
-\int_t^{t_0} \di t' \frac{B(t')}{a(t')}
\left[
\int_{t'}^{t_0}\frac{ \di t''}{a(t'')} \right]^2
+{\cal O\/}(B^2)~~,
\eeq
yielding
\beq
\frac{\rd}{\rd t_0}\left[ \frac{1}{R(t,r(t,r_0))} \right]=
-\frac{2 B(t)}{a(t)a(t_0)} \int_t^{t_0}\frac{\di t'}{a(t')}
+ {\cal O\/}(B^2)~~,
\eeq
where we have used Eq.(15).
Now, the redshift is given by
\beqarr
1+z &\equiv& \frac{\Dtz}{\Dt}
\\ \no
&=& \frac{a(t_0)}{a(t)} \left[ 1-B(t)\left(
\int_t^{t_0}\frac{\di t'}{a(t')} \right)^2
+2\int_t^{t_0}\di t' \frac{B(t')}{a(t')}
\int_{t'}^{t_0}\frac{ \di t''}{a(t'')}
+ {\cal O\/}(B^2) \right]~~.
\eeqarr
In a special case where $B(t)$ is a constant,
which corresponds to the standard Friedmann cosmology,
the redshift simply reduces to the standard relation
\beq
1+z = \frac{a(t_0)}{a(t)}~~,
\eeq
where we have used the relation
\beq
2 \int_t^{t_0}\frac{\di t'}{a(t')}
\int_{t'}^{t_0}\frac{ \di t''}{a(t'')} = \left[
\int_t^{t_0}\frac{\di t'}{a(t')} \right]^2 ~~ .
\eeq
That is, the well-known redshift expression
in the standard cosmology is reproduced, as expected.
Therefore, this result strongly suggests that $a(t)$ plays a role of,
more or less, the scale factor of the standard Friedmann cosmology.
In the case of a locally flat Universe, therefore,
the behavior of $a(t)$ cannot be much different from that of the
standard Friedmann cosmology with $k=0$,
which is proportional to $t^{2 \over 3}$ in the matter dominated era.
For mathematical simplicity and illustrative purposes,
we assume that $a(t)$ behaves the same as that in the standard cosmology
and $B(t)$ can be expressed by a simple power law.
That is, we assume that $a(t)$ and $B(t)$ are, in the matter
dominated era, of the form
\beq
a(t)=\alpha t_0 \,
\left( \frac{t}{t_0} \right)
^{ {2 \over 3} }
{}~,~~
B(t)=\beta
\left( \frac{t}{t_0} \right)^n
{}~( n \ge 0)~~,
\eeq
where $\alpha$ and $\beta$ are dimensionless parameters
to be determined and $n$ is set to be non-negative because of
the observed increase of the matter density
as a function of $r$.\cite{Kim}
Substituting Eq.(27) into Eq.(22) yields
\beqarr
r(t,t_0) &=& \frac{3}{\alpha} \left[ 1-
\left( \frac{t}{t_0} \right)^ {1 \over 3} \right]
-\beta\frac{9}{\alpha^3}
\left[ \frac{2}{9 ( n+ {1 \over 3})(n+{2 \over 3})(n+1)}\right.
\\ \no
& &\left.- \frac{1}{(n+{1 \over 3})}
\left( \frac{t}{t_0} \right)^{ n+{1\over 3}}
+\frac{2} { (n+ {2 \over 3}) }
\left( \frac{t}{t_0} \right)^{n+{2\over 3}}
-\frac{1}{(n+1)}
\left( \frac{t}{t_0} \right)^{n+1} \right]
+{\cal O\/}(\beta^2)~~.
\eeqarr
In practice, what is measured is the redshift.
Therefore, we must express
$( t/t_0)$ in terms of the red shift.
By substituting Eqs.(27) into Eq.(24), we have
\beqarr
1+z &=&
\left( \frac{t}{t_0} \right)^{- {2 \over 3}}\left[
1-\beta \frac{9}{\alpha^2}\left( \frac{t}{t_0} \right)^n
\left\{1- \left( \frac{t}{t_0} \right)^{1\over 3} \right\}^2
\right.
\\ \no
& & \left. + \beta
\frac{2}{ \alpha^2(n+ {1 \over 3})(n+ {2 \over 3}) } \left\{ 1+3
(n+{1 \over 3})
\left( \frac{t}{t_0} \right)^{n+{2 \over 3}}-3 (n+ {2 \over 3})
\left( \frac{t}{t_0} \right)^{n+ {1 \over 3}}\right\} \right]
+ {\cal O\/}(\beta^2)~~,
\eeqarr
yielding
\beqarr
\left( \frac{t}{t_0} \right)
&\equiv& T_{(0)}+\beta T_{(1)}+{\cal O\/}(\beta^2)~~,
\eeqarr
where, for notational simplicity, we define $T_{(0)}(z)$
and $T_{(1)}(z)$ as
\beqarr
T_{(0)}&=& (1+z)^{- {3 \over 2}}
\\ \no
T_{(1)} &=&{3 \over 2\alpha^2}
\left[ \frac{ 2 }{(n+ {1 \over 3})(n+ {2 \over 3})}
(1+z)^{- {3 \over 2}}+
\frac{18n}{(n+ {1 \over 3})}(1+z)^{- {3 \over 2}(n+{4\over 3})}
\right.
\\ \no
& & \left.
-\frac{9n}{(n+ {2 \over 3})}(1+z)^{- {3 \over 2}(n+{5 \over 3})}
-9(1+z)^{-{3 \over 2}(n+1)}\right]~~.
\eeqarr
Inserting Eqs.(30) and (31) into the right--hand side of Eq.(28) gives
\beqarr
r(t,t_0)
& = &
r_{(0)}+\beta r_{(1)}
+{\cal O\/}(\beta^2)~~,
\eeqarr
where $r_{(0)}$ and $r_{(1)}$ are defined as
\beqarr
r_{(0)} &\equiv&
{3 \over \alpha} [1- T_{(0)}^{1 \over 3}]
\\ \no
r_{(1)} &\equiv&
-{1 \over \alpha^3}
\left[ T_{(0)}^{-{2 \over 3}} T_{(1)} \alpha^2
+ \frac{2}{ (n+ {1\over3})(n+{2 \over 3})(n+1)}
\right.
\\ \no
& &\left.
-\frac{9}{(n+ {1 \over 3})} T_{(0)}^{ n+{1 \over3}}
+\frac{18}{(n+{2 \over 3})} T_{(0)}^{n+{2 \over 3}}
-\frac{9}{(n+1)} T_{(0)}^{n+1} \right].
~~.
\eeqarr
It is easy to see that the $r_{(0)}$
term reproduces the result
in the standard Friedmann cosmology whereas
$r_{(1)}$ represents a correction term.
It is interesting to note that
for small $z$,
$r_{(1)}$ is zero up to the second order in $z$,
as can easily be seen by substituting Eq.(31) into Eq.(33).
That is, there is no modification of small $r(t)$ (i.e., for $z\ll 1$)
due to small $B$,
up to the second order.
Since the luminosity distance is $r(t,t_0)R(t_0,0)(1+z)$,
the redshift--luminosity distance relation for small $z$
remains intact, at least, up to the first order
in $B(t)$. This is a consequence of the plausible assumption that
$a(t)$ has very similar behavior of the scale factor
of the standard Friedmann cosmology in the matter dominated era.
\subsection{Locally Open Universe.}
Before we proceed, we make a brief comment on
the status of the density parameter $\Omega$.
One of the most challenging tasks in the observational
astrophysics is the measurement of the mass density of the Universe,
which is supposed to be a constant at any given time
in the standard Friedmann cosmology.
Various observations, however, indicate that
the mass density indeed appears to increase as we probe farther out
\cite{Schramm}\cite{Kolb}.
{}From direct observations,
the fraction of critical density associated
with luminous galaxies is $\Omega_{LUM}\leq 0.01$.
When extending the observation to distances beyond
the luminous part of galaxies,
we found that there exist galactic halos which have a mass
corresponding to $\Omega_{HALO} \simeq 0.1$.
On a larger scale such as the Virgo cluster,
modeling the local distortion of the Hubble flow
around the cluster yields $\Omega_{CLUSTER}=$0.1 to 0.2.
Recently, using the redshift measurements for the catalogue of galaxies
by the Infrared Astronomy Satellite (IRAS)\cite{IRAS},
it became apparent that galaxies out to about 100 Mpc
flow towards the Great Attractor with high peculiar velocity.
It was concluded that the observed dynamics on this scale requires
$\Omega_{IRAS} \sim 1 \pm 0.6$.
Based on the above observations,
we have a picture of the Universe in which
our local neighborhood is underdense and
the mass density increases with scale.
Therefore, we present, in this Subsection,
the redshift--luminosity distance
relation for the locally $open$ Universe.
As was discussed before,
$a(t)$ is more or less the scale-factor
of the locally open Universe in the standard Friedmann cosmology.
Therefore, it is more transparent to rewrite $B(t)$
as $[1+b(t)]/4$.
Hereafter, a small perturbation to the locally $open$ Universe
is represented by $b(t)$ rather than by $B(t)$.
It is convenient to
introduce a new coordinate $\Phi$ as defined by
$r(t,t_0) = 2 \tanh \Phi(t,t_0)$.
The light propagation equation is then reduced to
\beq
\frac{\di \Phi}{\di t}= -\frac{1}{2a(t)}
+\frac{b(t)}{2 a(t)} \sinh^2 \Phi~~.
\eeq
We consider two successive wave crests
that leave a comoving coordinate $\Phi$ at time $t$ and $t+\Dt$
and arrive at $\Phi=0$ at times $t_0$ and $t_0+\Dtz$, respectively,
which yields the following equality :
\beqarr
\Phi &=&\int_{t}^{t_0}\di t'
\left[ \frac{1}{2a(t')}-\frac{b(t')}{2a(t')}
\sinh^2 \Phi (t',t_0) \right]
\\ \no
&=&
\int_{t+\Dt}^{t_0+\Dtz}\di t'
\left[ \frac{1}{2a(t')}-\frac{b(t')}{2a(t')}
\sinh^2\Phi(t',t_0+\Dtz) \right]~~,
\eeqarr
from which the redshift relation is given by
\beq
1+z \equiv \frac{\Dtz}{\Dt}
= \frac{a(t_0)}{a(t)} \left[ \frac
{1-b(t)\sinh ^2 \Phi(t,t_0)}
{1-a(t_0) \int_{t}^{t_0}\di t' \frac{b(t')}{a(t')}
\frac{ \rd \Phi(t',t_0)}{\rd t_0}
\sinh 2\Phi (t',t_0) } \right].
\eeq
For an open Friedmann Universe where $b(t)=0$,
$(1+z)$ simply becomes $a(t_0)/a(t)$, which is
the standard result.
Since the light propagation equation is non-linear,
we will again use the perturbation method
by treating $b(t)$ as being small.
Then, the redshift in this picture becomes
\beq
1+z \simeq
\frac{a(t_0)}{a(t)}\left[
1-b(t)\sinh^2 \int_{t}^{t_0}\frac{\di t'}{2a(t')}
+ \int_{t}^{t_0}\di t'
\frac{b(t')}{2a(t')}\sinh \int_{t'}^{t_0}
\frac{\di t''}{a(t'')}
+{\cal O\/}(b^2)\right]~~,
\eeq
where we have used the relation,
$\rd \Phi(t,t_0)/\rd t_0 = 1/2 a(t_0)$.
Again to proceed further we need
specific functional forms of $a(t)$ and $b(t)$.
Since $a(t)$ cannot be too
different from the scale factor in the Friedmann
cosmology with $k=-1$,
we assume that,
in the matter-dominated era,
$a(t)$ satisfies the following differential
equation as in the standard Friedmann cosmology
with $k=-1$ :
\beq
2\, \frac {\ddot{a}(t)}{a(t)} +
\left( \frac{\dot{a}(t)}{a(t)}\right)^2
-\frac{1}{a^2(t)} \simeq 0 ~~.
\eeq
Here, only one of the two initial conditions
can be fixed as $a(t=0)=0$.
The solution of Eq.(38) may be parameterized by an angle, $\Psi$,
as
\beqarr
a(\Psi) &=& \alpha t_0 [\cosh \Psi -1]
\\ \no
t(\Psi) &=& \alpha t_0 [\sinh \Psi -\Psi]
{}~,
\eeqarr
where $t_0$ is the age of the Universe and $\alpha$
is a dimensionless parameter to be determined.
In the following, $\Psi_0$ is defined as $t(\Psi_0)=t_0$,
which would correspond to $(1-q_0)/q_0$ in the Friedmann cosmology
with $k=-1$, where $q_0$ is the deceleration parameter.
For illustrative purposes,
we consider, in this Section,
a simple case where the arbitrary function $b(t)$ behaves as
\beq
b(\Psi) = \beta \Psi~~,
\eeq
where $\beta$ is a dimensionless parameter to be treated
as a perturbation.
Then, from Eqs.(35) and (37), we have
\beq
\frac{a(t)}{a(t_0)} = \frac{1}{1+z} \left[
1+\frac{\beta}{2} \{ \sinh(\Psi_0-\Psi)-(\Psi_0-\Psi)\}
\right]
+ {\cal O\/}(\beta^2)
\eeq
and
\beqarr
\Phi(\Psi,\Psi_0) &=&
\frac{1}{2} (\Psi_0-\Psi)
\\ \no
& & -\frac{\beta}{4} \left[ \cosh
(\Psi_0-\Psi)
+\Psi\sinh (\Psi_0-\Psi)-\frac{\Psi_0^2-\Psi^2}{2}-1\right]
+ {\cal O\/}(\beta^2)~~.
\eeqarr
Now, we are ready to obtain the redshift--luminosity relation
based on
$d_L=r R(t_0,0)(1+z)$.
Recalling $r(t,t_0)=2\tanh\Phi(t,t_0)$,
$(\Psi_0-\Psi)$ should be expressed in terms of the redshift.
{}From the parameterized solution Eq.(39), $\Psi$ is
\beq
\Psi = \cosh^{-1}
\left[ 1+(
\cosh \Psi_0-1
)\frac{a(t)}{a(t_0)}
\right]~~.
\eeq
Substituting Eq.(41) into Eq.(43), we find
\beqarr
\Psi_0-\Psi
&=& \Psi_0-\cosh^{-1} \left[1+\frac{ \cosh \Psi_0-1}{1+z}
\right]
\\ \no
& &
-\frac{\beta}{2}
\sqrt{\frac{ \cosh \Psi_0-1} {1+\cosh\Psi_0 + 2 z} }
\left\{ \sinh (\Psi_0-\Psi)- (\Psi_0-\Psi)
\right\}
+ {\cal O\/}(\beta^2)~~.
\eeqarr
For simplicity, we define the zeroth order term of $\Psi$,
$\Psi^{(0)}$, as
\beq
\Psi^{(0)} =
\cosh ^{-1} \left[ 1+\frac{ \cosh{\Psi_0}-1}{1+z}
\right]~~.
\eeq
Then, from Eqs.(42), (44) and (45), we have
\beqarr
\Phi(\Psi,\Psi_0) &=&
\frac{1}{2} (\Psi_0 - \Psi^{(0)})
\\ \no
& & + \frac{\beta}{4}
\sqrt{\frac{ \cosh \Psi_0-1} {1+\cosh\Psi_0 + 2 z} }
\left[ (\Psi_0 - \Psi^{(0)})
- \sinh (\Psi_0 - \Psi^{(0)})
\right]
\\ \no
& & - \frac{\beta}{4} \left[
\cosh (\Psi_0 - \Psi^{(0)})
+\Psi^{(0)}\sinh (\Psi_0 - \Psi^{(0)})
-\frac{ \Psi_0^2-(\Psi^{(0)})^2}{2}-1
\right]
\\ \no
& & +{\cal O\/} (\beta^2)
\\ \no
&\equiv& \Phi^{(0)}(z)+\frac{\beta}{4} \Phi^{(1)}
+ {\cal O\/} (\beta^2)~~.
\eeqarr
In our neighborhood ($z\ll 1$),
$\Phi^{(0)}(z)$ becomes
\beq
\Phi^{(0)} = \frac{1}{2}
\sqrt{ \frac{\cosh \Psi_0-1}{\cosh \Psi_0 +1}}
\left[ z -
\frac{ \cosh \Psi_0+2}{2(\cosh\Psi_0+1)}
\,z^2 \right]
+ {\cal O\/}(z^3)~~.
\eeq
Using Eqs.(44), (45) and (46),
it can easily be seen that
$\Phi^{(1)}$ is zero up to the second order of $z$.
Finally, $r(t,t_0)$ is expressed in terms of the redshift by
\beqarr
r(t,t_0) &=& 2 \tanh \Phi(t,t_0)
\\ \no
& = & 2 \tanh \Phi^{(0)}(z)
+\frac{\beta}{2}\Phi^{(1)}(z)
(1-\tanh^2 \Phi^{(0)}(z)) +{\cal O\/}(\beta^2)~~
\\ \no
&=& \sqrt{ \frac{ \cosh \Psi_0 -1}{\cosh \Psi_0+1} }
\left[ z-\frac{ \cosh \Psi_0+2}{2(\cosh \Psi_0+1)}
\, z^2 \right] + {\cal O\/}(z^3, \beta^2)~~.
\eeqarr
If $\cosh \Psi_0=(1-q_0)/q_0$ as in the Friedmann cosmology,
the standard result,
$r = \sqrt{1-2q_0}[z-(1+q_0)z^2/2]+{\cal O\/}(z^3)$,
is recovered,
implying that the luminosity distance, $r R(t_0)(1+z)$,
has the same dependence on $z$ for the small redshift
as in the standard model.
To summarize, the standard results of the redshift--luminosity
distance relation in the Friedmann cosmology for $z \ll 1$
are preserved in this model,
regardless of whether the local Universe is assumed to
be open or flat.
\section{Numerical Results}
As was demonstrated in the previous section,
the most fundamental cosmological relation predicted by
this model, i.e.,
the redshift--luminosity relation,
is practically the same as in the standard cosmology
for small $z$. This is due to our plausible assumption that $a(t)$
behaves more or less the same as the scale factor in the standard
cosmology.
We now proceed to explore the relationship for large $z$.
Unfortunately,
it is impossible to do so analytically,
due partly to our
inability to analytically solve the light propagation equation.
In this Section, therefore, we resort to
numerical method and present some numerical answers in the form
of figures depicting
possible modifications
of the redshift--luminosity distance relation for large $z$.
Since our local Universe appears to be open as discussed before,
we only consider the second case of the previous section.
Thus, $a(t)$ is parameterized by an angle $\Psi$ as given in Eq.(39).
Since $b(t)$ is completely arbitrary except for being small,
we consider the following three representative
cases : (A) $b(\Psi)=\beta \Psi$,
(B) $b(\Psi)=\beta \Psi^4$ and
(C) $b(\Psi)=\beta [\cosh \Psi - 1]^2$,
where $\beta$ is a dimensionless parameter to be determined.
Note that the case (C) corresponds to the picture in which
there is a $constant$ vacuum energy density in the Universe,
as can be seen from Eq.(17).
Recalling that the Hubble expansion rate of the local Universe
at the present epoch, $\overline{H}_0\equiv \dot{R}/R|_{t=t_0,r\sim0}$,
is $[ \dot{a}(t)/a(t)]_{t=t_0}$,
the density parameter in our neighborhood is, from Eq.(18),
\beq
\overline{\Omega}_0
\equiv \frac{8 \pi G \rho(t=t_0,r\simeq 0)}
{3 \overline{H}_0^2}
= 1- \dot{a}_0^2 - b(t_0)\dot{a}_0^2~~,
\eeq
where the bar denotes the $local$ value and the subscript
zero represents the $present$ value.
With the definitions of
$ \overline{\Omega}_{0,a} \equiv 1-\dot{a}_0^2$ and
$\overline{\Omega}_{0,b} \equiv b_0 \dot{a}_0^2 $,
the density parameter and the pressure of the local Universe
can be expressed by
\beqarr
\overline{\Omega}_{0}
&=& \overline{\Omega}_{0,a}
-\overline{\Omega}_{0,b}~~,
\\
8 \pi\overline{ Gp_0} &=&\overline{\Omega}_{0,b}
\overline{H}_{0}^2~~,
\eeqarr
where Eq.(19) is used.
{}From the observation of small peculiar velocities of nearby galaxies,
we assume that the pressure of the local Universe is
relatively small compared with the energy density.
That is, $\overline{\Omega}_{0,b} $ is very small.
In the numerical calculations to be presented in this Section,
we assume $\overline{\Omega}_{0} =0.1$ and, for the sake of
definiteness, choose values
$\overline{\Omega}_{0,a} =0.1001$ and
$\overline{\Omega}_{0,b} =0.0001$, specifying the numerical values
of $\alpha$ and $\beta$.
The light propagation equation can then be numerically
integrated with the boundary condition $r(t=t_0, t_0)=0$.
To this end, we divide the numerical solution of
$r(t,t_0)$ into $N$ intervals and fit each interval using
\beq
r_i(t,t_0)= \delta_i \left[ t_0^{\gamma_i}
-t^{\gamma_i} \right]~~\mbox{ for\/}~~
\frac{i-1}{N} < \frac{t}{t_0} < \frac{i}{N}~~
(i=1,2,3...)~,
\eeq
which then yields the numerical values of $\delta_i$ and $\gamma_i$
for each interval. We have used $ N = 200$ in our numerical calculations.
Using the definition of the comoving distance, i.e.,
$r_i(t,t_0)=r_i(t+\Dt,t_0+\Dtz)$, we have the redshift,
for each interval, as
\beq
(1+z) \equiv \left( \frac{\Dtz}{\Dt} \right)_i
= \left( \frac{t}{t_0} \right)^{\gamma_i-1}
{}~.
\eeq
Then, the luminosity distance, $d_L=r R(t_0)(1+z)$,
for each interval, can easily be calculated from Eqs.(52) and (53).
In Fig.1, the results of the three cases discussed above
are presented by solid lines,
while the standard results with $\Omega=0.1$ and $\Omega=1.0$
by the dashed lines and dotted lines, respectively.
First, it is to be noted that on small scale, the
linear {\em Hubble diagrams\/} are preserved for all of the
three cases, as was shown by perturbative calculations.
In the cases of (A) and (C),
the redshift--luminosity distance relations in this model
are almost indistinguishable
and furthermore correspond to
the standard results with the density parameter
in the range $\Omega \sim$ $0.8-0.9$,
even though the local mass density in the model is set
to be $\overline{\Omega}_0=0.1$.
In the case (B), however, it appears to be the standard result
with $\Omega > 1$ although the difference is not very
significant. Nevertheless,
this qualitative feature was totally unexpected.
In this model, therefore, a
precise measurement of the redshift--luminosity
distance relation alone cannot provide information on $q_0$,
which is related to $\Omega_0$ as $2q_0=\Omega_0$,
contrary to the case of the standard cosmology.
Now it would be appropriate to discuss the meaning of the
$observed$ increase of $\Omega_0$ as we look farther out.
Every light signal we are receiving right now
contains information about the past in time.
That is, what we measure are,
for example, the redshift and $G\rho(t)$, not $G\rho(t_0)$.
Thus, we deduce physical quantities at the present time $t_0$,
using the standard cosmological evolution equations.
Recalling that $G\rho(t)$ in the matter dominated era in the standard
Friedmann cosmology is proportional to $1/S^3(t)$,
and $S(t)$ is just $S(t_0)/(1+z)$,
$G\rho(t_0)$ is $obtained$ from $G\rho(t)$ by multiplying $1/(1+z)^3$.
Dividing $G\rho(t_0)$ with the constant critical density
in the standard cosmology, $G\rho^{(s)}_c$,
we can deduce the density parameter at present time,
which we shall call $\Omega^{obs}_0$, as
\beq
\Omega^{obs}_0 \equiv \frac{G\rho^{obs}(t_0)}{G\rho^{(s)}_c}
\equiv \frac{G\rho(t,r(t))}{G\rho^{(s)}_c (1+z)^3}~~.
\eeq
Using Eqs.(52) and (53) for each interval in Eq.(18),
we have calculated $\Omega_0^{obs}$ versus the redshift.
The results are shown in Fig.2. As before, we have taken
the local value $\Omega_{0}^{obs} =0.1$. In all three cases,
the calculated values of $\Omega^{obs}_0$ are increasing
functions of $z$. ( A naive value in the standard cosmology
is supposed to be a constant.)
We can see that the cases of (A) and (C) are
practically identical, whereas
the case (B) shows a faster increase of $\Omega_0^{obs}$.
As can be seen in Fig.2, however, even in the case (B) the increase
is too slow to explain the IRAS data\cite{IRAS}
where $\Omega^{obs}_0$ is compatible to unity
(but with large errors) at the distance of about
several 100 Mpc($z \sim 0.0166$).
To fit the increase of $\Omega^{obs}_0$ up to unity at $\sim 100$ Mpc
requires drastic ( perhaps unrealistic ) changes
in the form of and parameters in $a(t)$ and $b(t)$,
which then would modify our predictions on the Hubble law.
\section{summary and conclusions}
We have studied how cosmological observables are modified
in an isotropic but inhomogeneous Universe
compared with those of the standard model.
In particular, the luminosity distance and the density
parameter as functions of the redshift have been examined
in the generalized Robertson--Walker spacetime with only one
$(t,r)$-dependent scale factor and
they were compared with the standard results.
When $R(t,r)$ is not factorized into the form of
$a(t)f(r)$, the simple
redshift--scale factor relation such as $(1+z)=a(t_0)/a(t)$
remains no longer valid.
First by solving light propagation equation, Eq.(2),
for radially propagating light with
the boundary condition $r(t=t_{received})=0$
and then considering two wave crests emitted at time $t$ and $t+\Dt$
which are received at $t_0$ and $t_0+\Dtz$, respectively,
we have obtained the general redshift--scale factor relation given by
Eq.(5).
The result is valid in an inhomogeneous Universe and is shown to be
reduced to the simple form $(1+z)=a(t_0)/a(t)$ in the case of
the homogeneous spacetime, i.e., the standard
Robertson--Walker spacetime.
Our general relations agree with the
results obtained in \cite{Bondi} and \cite{Moffat}.
We have applied the general redshift--scale factor relation
to the cosmological model in \cite{Kim}
where the Universe is pictured as being
inside a highly dense and rapidly expanding shell
with the underdense center.
First, for the nearby objects ($z\ll 1$),
the luminosity distances as functions of the redshift
are obtained analytically, using a perturbative method
for two cases where the underdense center is either flat or open
according to the definition of the standard Friedmann cosmology.
One of the most interesting features in \cite{Kim}
is that the scale factor $R(t,r)(=a(t)/(1-B(t)r^2) )$
is specified by two arbitrary functions, $a(t)$ and $B(t)$
(or $b(t)$), and $a(t)$ is very similar to the scale factor
of the standard Friedmann cosmology and $B(t)$ (or $b(t)$)
is the perturbation to the locally flat (or open) Universe.
Under the assumption that $a(t)$ behaves the same as that in the
standard cosmology,
it is shown analytically that the standard redshift--luminosity
distance relations in the Friedmann cosmology for $z \ll 1$ remains intact
for both cases.
Specifically,
since the corrections of order $\cal{O \/}(\beta)$
of these expressions can
be expanded as a power series of $\Sigma_{i=3}^{\infty} c_i z^i$
with some coefficients $c_i$ (that is,
zero up to the second order of $z$),
it has been shown that for nearby objects, in spite of its
different metric,
the cosmological model of \cite{Kim}
is not much different from the standard cosmology as far as
the Hubble law is concerned.
It is also interesting to note that
in spite of the special location of the observer, i.e., the return of
the pre--Copernican notion in the model \cite{Kim},
the results are almost the same as those of the standard model.
As for large $z$, the redshift--luminosity distance relations
given in Eqs. (32) and (48) are different
from those of the standard
cosmology and moreover it is expected
that the corrections would be larger when
${\cal O\/}(\beta^2)$ terms are included. In this case, as we
mentioned repeatedly, we cannot
use the perturbative method and thus
we have solved them numerically and obtained the results as shown
in Fig.1. and Fig.2.
Figure 1 shows the redshift--luminosity distance relation
in the cosmological model of \cite{Kim}.
Comparing them with the standard cosmology with $k=0$ (dotted curve)
and $k=-1$ (dashed curve),
we can easily see that for small $z$,
the redshift--luminosity distance relation of model \cite{Kim}
denoted by the solid lines
is almost the same as the standard one,
as was also shown in the explicit perturbative calculation.
But for $1<z<3$, the Hubble law of the model is very
similar to that of the standard cosmology with $k=0$,
not with that with $k=-1$,
in spite of the fact that the mass density of the
local Universe is set to be
$\overline{\Omega}_0=0.1$.
It is to be noted that although there is no substantial
deviation from the standard model
in our redshift--luminosity distance relation in Fig.1,
the deviation in the case of the TB model is more
prominent and is different from ours, both qualitatively and
quantitatively as was shown in \cite{Moffat}.
Figure 2 shows that the calculated density parameter
at the present time,
$\Omega^{obs}_0$, is an increasing function of the redshift.
Although, admittedly, our perturbative calculations may not be
rigorous in the sense that the functions $a(t)$
and $B(t)$ that appear in the scale factor
$R(t,r)$ could not be determined by
the equation of state, and furthermore
our numerical results can only explain the IRAS data qualitatively,
we feel that the qualitative
nature of our results are robust
because we have considered the general case with
$B(t)=\beta (t/t_0)^n ~(n >0)$,
$b(t)=\beta \Psi(t)$, $b(t)=\beta \Psi^4(t)$ and
$b(t)=\beta [ \cosh \Psi -1]^2$
with a variety of numerical values of the parameters involved,
all of which have led to similar results.
In summary, although the scale-dependent cosmology for
the inhomogeneous Universe as modeled in
\cite{Kim} implies the explicit running of $H_0$, $\Omega_0$ and $t_0$
as functions of $r$ because of the non-Robertson-Walker metric,
as far as the observables such as redshift--luminosity
distance relations are concerned, the results are
hardly different from those of the
standard model in our neighborhood, i.e. for small $z$.
Even for large $z$, the difference between the model
considered and the standard model with $k=0$
still remains small but
the model can be tested when the data from
galaxy redshift survey at long distance become available
and are compared with the predictions of the model on the
matter distribution and on the age of the Universe.\cite{Kim}
\acknowledgments
The authors would like to thank G. Feldman, M. Im
and Y.C. Pei for helpful discussions.
One of the authors (THL) wishes to thank
the Department of Physics and Astronomy,
the Johns Hopkins University for the hospitality extended
to him during the
completion of this work.
This work was supported in part by the Basic
Science Research Institute Program,
Ministry of Education, Korea,
Project No. BSRI-94-2418(THL) and by the National Science Foundation.
|
\subsection{\bf Introduction}
\hspace*{1em} It has been known that the Flavor Changing process in
down quark sector is sensitive to the mass difference among
up-type quark in the standard model (SM). GIM mechanism tells us that
the FCNC in one-loop level of SM vanishes when up type
quarks are degenerate. In the real world , there is large mass gap
among up quark sector ($M_u \ll M_c \ll M_t$).
Then the Flavor Changing Process in down quark sector like $b \rightarrow
s \gamma$ and $b \rightarrow s l^+ l^-$ are enhanced. If we extend the
fermion sector beyond the standard model, there are several different
possibilities. If we just add the chiral fermions
as fourth generation in sequential way, GIM mechanism still
works and the new contribution
comes from up-type quark ($t^{\prime}$) in the fourth generation. Therefore
we can get constraints
for the mass of $t^\prime$ and its Kobayashi Maskawa mixing to
light down-type quarks ($V_{CKM}^{t^\prime i}$, $i = d,s,b$)
However if we extend the fermion
sector in non-sequential way, the different aspect
arises (refs. \cite{branco}, \cite{nir}, \cite{morozumi},
\cite{gautam} and \cite{handoko}).
The tree level FCNC arises both in $Z$ and neutral Higgs sector.
Furthermore, the size of the FCNC depends on the structure of
the down quark mass matrix rather than up type quark masses.
Therefore we may study the structure of down type quark mass
matrix by studying the flavor changing process in down quark sector.
\subsection{\bf Non-sequential extention of fermion sector}
\hspace*{1em} We study the Standard Model (SM) with extended quark sector.
In addition to the three standard generations of quarks,
$N_d - 3$ down-type and $N_u - 3$ up-type vector like singlet quarks are
introduced (refs. \cite{branco}, \cite{nir} and \cite{morozumi}).
\begin{equation}
{I_W} = {1\over 2}\: : \:
\left(
\begin{array}{c}
u \\
d
\end{array} \right)_L
\left(
\begin{array}{c}
c \\
s
\end{array} \right)_L
\left(
\begin{array}{c}
t \\
b
\end{array} \right)_L ,
I_W = 0 \: : \:
\left(
\begin{array}{ccccc}
u_R & c_R & t_R & t^{\prime}_{L+R} & \cdots \\
d_R & s_R & b_R & b^{\prime}_{L+R} & \cdots
\end{array}\right) \: .
\end{equation}
In this model, there is tree level FCNC in $Z$ and Higgs sector.
In order to illustrate this point and explain the relation between the
size of FCNC and down-type quark mass matrix, let us introduce
a toy model. This is the so called "top-prime less" model
\begin{equation}
{I_W} = {1\over2} \: : \:
\left(
\begin{array}{c}
{t}^0 \\
{b}^0
\end{array}
\right)_L \: , \:
{I_W} = 0 \: : \:
\begin{array}{c}
{t_R}^0 \\
{b_R}^0
\end{array}
{b_{L+R}}^{0\prime} \: .
\end{equation}
The most general mass matrix for down quark sector $M_d$ is given by,
\begin{equation}
L_{mass} = -(\bar{{b_L}^0} \: \bar{{{b_L}^0}^{\prime}})
\left[
\begin{array}{cc}
m & 0 \\
J & m_4
\end{array} \right]
\left(
\begin{array}{c}
{b_R}^0 \\
{{b_R}^0}^{\prime}
\end{array}\right) \: .
\end{equation}
where $J$ is complex and $m$ and $m_4$ are real numbers. $m$ comes from the
vacuum expectation value of Higgs doublet.
$J$ leads to the mixing between left handed singlet quarks and right handed
ordinary quarks. In the limit of $J = 0$, the vector like quark decouples
from the ordinary quark.
$ M_d {M_d}^\dagger$ can be diagonalized by the following unitary
transformation :
\begin{equation}
\left( \begin{array}{c}
{b_L}^0 \\
{b_L}^{\prime 0}
\end{array} \right) =
\left(
\begin{array}{cc}
\cos \theta & - \sin \theta e^{-i \phi} \\
\sin \theta e^{i \phi} & \cos \theta
\end{array} \right)
\left(\begin{array}{c}
b_L \\
b_L^{\prime}
\end{array} \right) \: ,
\end{equation}
where $b^0$ and $b^{0\prime}$ indicate the weak basis.
When $m$ is much smaller than $M = \sqrt{|J|^2 + {m_4}^2}$,
the elements of the
unitary matrix are given by the following formulae
approximately :
\begin{eqnarray}
\cos \theta & \simeq & 1 \: , \\
\sin \theta e^{-i \phi} & \simeq & {m J \over M^2} \: .
\end{eqnarray}
Correspondingly, the mass eigenvalues
for heavy quark $(M_H)$ and light quark $(M_L)$ are given
by :
\begin{eqnarray}
M_L & = & m {m_4 \over M} \\
M_H & = & M \: .
\end{eqnarray}
There are two interesting limits :
\begin{enumerate}
\item $J \ll m_4$
\item $J \gg m_4$
\end{enumerate}
In the case of (1), the mixing between ordinary quark and
vector like quark is suppressed. The physical masses are given by
the diagonal elements of the mass matrix:
\begin{equation}
M_L = m \: , \: M_H = m_4 \: .
\end{equation}
In the case of (2),
the diagonal elements no longer reflect the physical masses :
\begin{equation}
M_L = m {m_4 \over |J|} \: , \: M_H = |J| \: .
\end{equation}
Note that the light quark mass vanishes in the limit
$m_4 = 0$. On the other hand,
the neutral current is written in terms of physical basis,
\begin{eqnarray}
L_Z & = & -{g \over 2 \cos \theta_W} \bar{{b^0_L}}
\gamma_{\mu} {b^0}_L Z^{\mu} \nonumber \\
& = & -{g \over 2 \cos \theta_W} \left[
\cos^2\theta \bar{b_L} \gamma_{\mu} b_L
-\cos \theta \sin \theta e^{-i \phi}
\bar{b_L} \gamma_{\mu} b_L^{\prime} + h.c.
\right]Z^{\mu} \: .
\end{eqnarray}
Thus the flavor diagonal coupling for neutral current ($Z^{bb}$)
and the FCNC coupling ($Z^{bb^\prime}$) are given by :
\begin{eqnarray}
Z^{bb} & = &\cos^2\theta = 1 - \sin^2\theta \simeq
1 - O\left( \left(m J \over M^2 \right)^2
\right) \: , \\
Z^{bb^\prime} & = & -\cos \theta \sin \theta e^{-i \phi}
\simeq -{m J\over M^2} \: .
\end{eqnarray}
Depending on the two cases mentioned above, the enhancement
and suppression of the FCNC occur respectively.
\begin{enumerate}
\item $J \ll m_4$ (suppression)
\begin{eqnarray}
M_L & = & m \: , \\
M_H & = & m_4 \: , \\
\left| Z^{bb^{\prime}}\right| & = & {m |J| \over {m_4}^2}
\ll {M_L \over M_H} \: .
\end{eqnarray}
\item $J \gg m_4$ (enhancement)
\begin{eqnarray}
M_L & = & m {m_4 \over |J|} \: , \\
M_H & = & \left| J \right| \: , \\
\left| Z^{bb^{\prime}} \right| & = & {m \over |J|}
\gg {M_L \over M_H} \: .
\end{eqnarray}
\end{enumerate}
For instance, if the physical mass of the vector like quark mass
is $M_H = 500$(GeV), the following mass matrices realize
the same mass eigenvalues for heavy and light quark masses
($M_H = 500$(GeV), $M_L = M_b = 5$(GeV)) while giving rise
to the different size of FCNC.
\begin{enumerate}
\item \( M_d = \left[
\begin{array}{cc}
5 & 0 \\
50 & 500
\end{array} \right] (GeV) \: , \:
Z^{bb^\prime} = 10^{-3} \: , \)
\item \( M_d = \left[
\begin{array}{cc}
50 & 0 \\
500 & 50
\end{array} \right] (GeV) \: , \:
Z^{bb^\prime} = 10^{-1} \: . \)
\end{enumerate}
This exercise tells us that the FCNC between singlet quark
and ordinary quark is enhanced when the off-diagonal element
in the mass matrix is larger than the diagonal element.
This kind of analysis can be extended to the general case with arbitrary
numbers of isosinglet up-type quarks and down-type quarks.
Here we just record the full lagrangian for the model (ref.\cite{handoko}).
\begin{eqnarray}
{\cal L} & = & {\cal L}_{W^{\pm}} + {\cal L}_{\chi^{\pm}} +
{\cal L}_A + {\cal L}_Z + {\cal L}_H + {\cal L}_{\chi^0} ,
\label{eqn:lagrangian}
\end{eqnarray}
where,
\begin{eqnarray}
{\cal L}_{W^{\pm}} & = & \frac{g}{\sqrt{2}}
V_{CKM}^{\alpha\beta} \bar{u}^{\alpha} \gamma^{\mu} \, L \,
d^{\beta} \, W_{\mu}^+ + h.c. , \\
{\cal L}_{\chi^{\pm}} & = & \frac{g}{\sqrt{2} M_W}
V_{CKM}^{\alpha\beta}
\bar{u}^{\alpha} \left( m_{u^{\alpha}} L -
m_{d^{\beta}} R \right) d^{\beta} \chi^+ + h.c. , \\
{\cal L}_A & = & \frac{e}{3} \left(
2 \bar{u}^{\alpha} \gamma^{\mu} u^{\alpha}
- \bar{d}^{\alpha} \gamma^{\mu} d^{\alpha}
\right) \, A_{\mu} , \\
{\cal L}_Z & = & \frac{g}{2\cos\theta_W}
\left\{ \bar{u}^{\alpha} \gamma^{\mu} \left[\left(
{Z_u}^{\alpha\beta} - \frac{4}{3} \sin^2\theta_W
\delta^{\alpha\beta}
\right) L - \frac{4}{3} \sin^2\theta_W \delta^{\alpha\beta} R
\right] u^{\beta} \right. \nonumber \\
& & + \left. \bar{d}^{\alpha} \gamma^{\mu}
\left[\left( \frac{2}{3} \sin^2\theta_W \delta^{\alpha\beta} -
{Z_d}^{\alpha\beta}\right) L + \frac{2}{3}\sin^2\theta_W
\delta^{\alpha\beta}
R\right] d^{\beta}\right\} \, Z_{\mu} , \\
{\cal L}_H & = & \frac{-g}{2 M_W} \left[
{Z_u}^{\alpha\beta} \bar{u}^{\alpha}
\left(m_{u^{\alpha}} L + m_{u^{\beta}} R\right) u^{\beta}
\right. \nonumber \\
& & \left. \: \: \: \: \: \: \: \:
+ {Z_d}^{\alpha\beta} \bar{d}^{\alpha}
\left(m_{d^{\alpha}} L + m_{d^{\beta}} R\right) d^{\beta}
\right] \, H , \\
{\cal L}_{\chi^0} & = & \frac{-ig}{2 M_W} \left[
{Z_u}^{\alpha\beta} \bar{u}^{\alpha}
\left(m_{u^{\alpha}} L - m_{u^{\beta}} R\right) u^{\beta}
\right. \nonumber \\
& & \left. \: \: \: \: \: \: \: \:
- {Z_d}^{\alpha\beta} \bar{d}^{\alpha}
\left(m_{d^{\alpha}} L - m_{d^{\beta}} R\right) d^{\beta}
\right] \, \chi^0 .
\end{eqnarray}
For $N_u = 3$ and $N_d = 4$, the diagonal elements and off-diagonal elements
of FCNC are given by the following equations :
\begin{eqnarray}
Z^{bb} & = & 1 - \left| {V_L}^{4b} \right|^2 \: , \\
Z^{bs} & = & -{{V_L}^{4b}}^\ast V_L^{4s} \: , \\
Z^{bb^\prime} & = & -{{V_L}^{4b}}^\ast V_L^{44} \: ,
\end{eqnarray}
where $V_L$ is a unitary matrix which diagonalizes
the down type quark mass matrix
\begin{equation}
{d_L}^{0a} = {V_L}^{ab} {d_L}^b \: .
\end{equation}
In the basis in which up-type quark is diagonalized, $3$ by $4$ part of
$V_L$ is just CKM matrix,
\begin{equation}
{V_L}^{i a} = V_{CKM}^{ia} \: \:
(i = 1,2,3 \: , \: a = 1,2,3,4) \: .
\end{equation}
This leads to the relation between the FCNC and the deviation of unitarity
of CKM matrix
\begin{equation}
Z^{bs} = -{{V_L}^{4b}}^\ast V_L^{4s}
= \sum_{i=1}^{3} {V_{CKM}^{ib}}^\ast V_{CKM}^{is} \: .
\end{equation}
Therefore the unitarity of the CKM matrix no longer holds.
This non-unitarity ``quadrangle'' relation is used to
constrain the FCNC coupling, e.g. $Z^{bs}$, in later section.
However,
the deviation is suppressed when the diagonal element
of singlet quark mass is infinite and keeping off-diagonal
element finite.
\subsection{Rare decays and Effect of FCNC}
\subsubsection{\it New Physics v.s. Standard Model
in $b \rightarrow s Z \rightarrow s l^+ l^-$}
\hspace*{1em} In the present beyond standard model, there
is FCNC which leads to
tree level coupling $Z^{bs}$, $Z^{bd}$ and $Z^{sd}$.
In the standard model, the same vertex comes from top quark one loop
diagrams. Therefore the condition that the $Z$ FCNC dominates over the
standard model contribution in $Z$ sector is :
\begin{equation}
\left| {Z^{bs} \over {V_{CKM}^{tb}}^\ast V_{CKM}^{ts}}
\right| > O(\alpha) \simeq 0.012 \: .
\end{equation}
Then, even tiny coupling for $Z$ FCNC, it can easily
dominates over the standard model contribution.
Experimental constraints and quadrangle constraints for the
FCNC in $bs$, $bd$ and $sd$ sectors are given in Table (\ref{tab:bound}).
{}From Table (\ref{tab:bound}) (refs. \cite{nir},
\cite{data} and \cite{handoko}),
$\left|{Z^{bs}}/{{V_{CKM}^{tb}}^\ast
V_{CKM}^{ts}} \right| \le 0.05 $ and $ \left| {Z^{bd}}/
{{V_{CKM}^{tb}}^\ast V_{CKM}^{td}} \right| \le 0.79$.
Therefore, there are allowed regions for $Z^{bs}$ and
$Z^{bd}$
couplings where the tree level $Z$ FCNC can dominate over the
1-loop top quark diagrams in the standard model.
We can expect the drastic change of the differential
decay rates
in the present beyond standard model in $b \rightarrow s l^+
l^-$ and $b \rightarrow d l^+ l^-$ process.
\begin{table}[t]
\begin{center}
\begin{tabular}{|l|l|l|} \hline
Decay process & Experiment constraint & Quadrangle constraint \\
\hline \hline
$B \rightarrow X_s \mu^+ \mu^-$
& $\left| \frac{Z^{bs}}{{V_{CKM}^{cb}}^\ast
V_{CKM}^{cs}}\right|
\le 0.047$
& $\left| \frac{{V_{CKM}^{tb}}^\ast V_{CKM}^{ts}}
{{V_{CKM}^{cb}}^\ast V_{CKM}^{cs}}\right| \ge 0.94$
\\ \hline
$B \rightarrow X_d \mu^+ \mu^-$
& $\left| \frac{Z^{bd}}{{V_{CKM}^{cb}}^\ast
V_{CKM}^{cd}}\right|
\le 0.23$
& $\left| \frac{{V_{CKM}^{tb}}^\ast V_{CKM}^{td}}
{{V_{CKM}^{cb}}^\ast V_{CKM}^{cd}}\right| \ge 0.29$
\\ \hline
$K^+ \rightarrow \pi^+ \nu \bar{\nu}$
& $\left| \frac{Z^{sd}}{{V_{CKM}^{cd}}^\ast
V_{CKM}^{cs}}\right|
\le 2.9 \times 10^{-4}$
& $\left| \frac{{V_{CKM}^{td}}^\ast V_{CKM}^{ts}}
{{V_{CKM}^{cd}}^\ast V_{CKM}^{cs}}\right| \ge 0$ \\
\hline
\end{tabular}
\caption{Upper bound for $\left| Z^{\alpha\beta} \right| $ from
experiments, with assuming $Z$ exchange tree diagrams are dominant.}
\label{tab:bound}
\end{center}
\end{table}
\subsubsection{\it New Physics v.s. Standard Model in
$b \rightarrow s \gamma$ and $b \rightarrow s g$ process}
\hspace*{1em} There is no flavor changing neutral currents
for $b \rightarrow s \gamma $ and $ b\rightarrow s g $ ($g$ :
gluon) processes in the beyond
standard model and standard model. Therefore, in order that
the new physics contribution dominates over the standard model contribution,
the FCNC coupling constant must be the same order as that of
the CKM matrix elements.
The condition for the new physics contribution dominates over the
standard model contribution is :
\begin{equation}
\left| {Z^{bs} \over {V_{CKM}^{tb}}^\ast V_{CKM}^{ts}}
\right| > O(1) \: .
\end{equation}
{}From Table (\ref{tab:bound}) (refs. \cite{nir},
\cite{handoko}) we can see that there will be no
significant contribution to $b \rightarrow s \gamma (g)$
process because $\left| {Z^{bs}}/{{V_{CKM}^{tb}}^\ast V_{CKM}^{ts}}
\right| < 0.05$.
For $b \rightarrow d \gamma (g)$, new physics contribution
can be significant because $\left| {Z^{bd}}/{{V_{CKM}^{tb}}^\ast V_{CKM}^{td}}
\right| < 0.79$.
Let us summarize the computation of $b \rightarrow d \gamma$
briefly. The amplitude is proportional to the magnetic
moment interaction (refs. \cite{inami}, \cite{gautam} and
\cite{handoko}),
\begin{equation}
T = \frac{G_F e}{4 \sqrt{2} \pi^2} \; \bar{d_L} \sigma_{\mu \nu}
b_R \epsilon^{\mu} q^{\nu} F \: ,
\end{equation}
where,
\begin{eqnarray}
F & = & Q_u {V_{CKM}^{tb}}^\ast V_{CKM}^{td} F^{cc}
\left( {m_t \over M_W} \right) \nonumber \\
& + & Q_d Z^{db} \left\{ {2 \over 3} \sin^2\theta_W
{F_1}^{NC}(0) + 2 {F_3}^{NC}(0) \right\} \\
& + & Q_d Z^{db^\prime} Z^{b^\prime b} \left\{
{F_3}^{NC}\left({m_{b^\prime} \over M_Z}\right)
+ {F_2}^{NC}\left({m_{b^\prime} \over m_H},
{m_{b^\prime} \over M_Z}\right) \right\} \nonumber
\end{eqnarray}
where $Q_u = 2/3$ and $Q_d=-1/3$, and
$m_H$ is the neutral Higgs mass.
The first term comes from the top quark loop.
The second and the third terms come from down type quark
loop. The second term represents the light down type quarks
$(d,b)$ loop and the third term represents the heavy down type quark
$(b^\prime)$. The details of the functions and their behavior
are given in ref.
\cite{handoko}.
\subsection{\bf Summary}
\hspace*{1em} We study the rare $B$ decays in the
non-sequential extention of
fermion sector. The effect of vector like down type singlet quark
is studied.
It has been shown that the tree level FCNC can contribute to
$ b \rightarrow s(d) l^+ l^-$ processes, while $b \rightarrow s
\gamma (g)$ does not receive the significant contribution due to
the New Physics. In $b \rightarrow d \gamma (g)$, the effect of
FCNC ($Z^{bd}$, $Z^{db^\prime}$, $Z^{bb^\prime}$) may be seen.
\bigskip
\noindent
{\large \bf Acknowledgments}
T. M. would like to thank I. Bigi,
J. Kodaira and T. Kouno for
comments. This work is supported by the
Grant-in-Aid for Scientific Research ($\sharp 06740220$) from the Ministry of
Education, Science and Culture, Japan. The work of L.T.H. was supported by
a grant from the Overseas Fellowship Program (OFP-BPPT), Indonesia.
|
\section{Introduction}
The relationship between the entropy of a physical black hole and its internal
degrees of freedom remains a subject of active research. A natural question to
ask in this regard is: can these degrees of freedom be effectively included in
a functional integral description of black hole entropy?
In an attempt to give an affirmative answer to this question, we investigate in
this paper a microcanonical functional integral when applied to a
quantum self-gravitating statistical system that includes spacetimes whose
topology and boundary conditions coincide with the ones of (either distorted or
Kerr-Newman) eternal black holes.
A proposal for the density of states of a gravitational
system defined in terms of a microcanonical functional integral has been
suggested recently in Ref. \cite{BrYo2}.
This integral is defined as a formal sum
over Lorentzian geometries. The black hole density of states is obtained
from this functional
integral when the latter is approximated semiclassically by using a
complex metric whose boundary data at its single boundary surface coincide with
the boundary data of a Lorentzian, stationary black hole.
The density of states defined accordingly equals the exponential of
one fourth of the area of the black hole horizon.
This proposal opens the possibility of determining the
thermodynamical properties of black hole systems starting from a sum
over real Lorentzian geometries.
However, several problems remain in this approach. First, a spacelike
hypersurface $\Sigma$ that describes the initial data of a Lorentzian black
hole has to cross necessarily the event horizon and eventually intersect the
interior singularity.
Second, a Lorentzian, stationary black hole is not a extremum of the
microcanonical action for a spacetime region with a single timelike
boundary surface.
This implies that the black hole density of states
whose boundary data correspond to the ones of a Lorentzian,
stationary black hole cannot be approximated semiclassically by
using the same Lorentzian metric that motivates its boundary conditions.
It is therefore necessary to use complex metrics to evaluate the
Lorentzian functional integral in a
steepest descents approximation. This procedure yields the correct result
for the entropy but conceals its origin: the interior of the Lorentzian
black hole literally disappears by virtue of this procedure, leaving
effectively only a periodically identified
Euclidean version of the ``right" wedge region of a
Kruskal diagram. The properties of the black hole interior become encoded in a
set of conditions at the so-called ``bolt" of the complex geometry.
In this approach the statistical origin of entropy and
its relationship to the
internal degrees of freedom of a black hole remain obscure.
The problems mentioned above and the role of internal
degrees of freedom in functional integral descriptions of black hole
thermodynamics can be addressed by explicitly
considering the eternal version of a black hole \cite{Ma}.
The excitations of the physical black
hole can be associated with the deformations of an initial global Cauchy
surface $\Sigma$ of the eternal black hole.
In general, the spatial slices $\Sigma$ that foliate an
eternal black hole are (deformed) Einstein-Rosen bridges with wormhole topology
$R^1 \times S^2$. The spacetime is composed of two wedges $M_+$ and $M_-$
located in the right and left sectors of a Kruskal diagram.
Internal and external degrees of freedom of the black hole can be
easily identified in this approach since the hypersurfaces $\Sigma$ are
naturally divided in two parts $\Sigma_+$ and $\Sigma_-$ by a bifurcation
two-surface $S_0$ where
the lapse function $N$ vanishes.
While the ``external" degrees of freedom of the original black hole are
naturally given by the initial data at $\Sigma_+$,
its ``internal" degrees of freedom can be identified with initial data defined
at $\Sigma_-$ \cite{FrMa2}.
\section{Microcanonical action and functional integral}
Consider a spacetime region whose three-dimensional
timelike boundary surface $B$ consists of
two disconnected parts $B_+$ and $B_-$.
The microcanonical action $S_m$ is the action appropriate to a variational
principle in which the fixed boundary conditions at the timelike boundaries
$B_\pm$ are not the spacetime three-geometry
but the surface energy density
$\varepsilon$, surface momentum density $j_a$, and boundary two-metric
$\sigma_{ab}$ \cite{BrYo2,BrMaYo}.
The covariant form of the microcanonical action for a general
spacetime whose timelike surfaces $B_+$ and $B_-$ are located in the
regions $M_+$ and $M_-$ respectively has been presented in Ref. \cite{Ma}. Its
Hamiltonian form is easily obtained under a
$3+1$ spacetime split by recognizing that there exists a direction of time at
the boundaries $B_\pm$ \cite{Ma}.
By introducing the momentum $P^{ij}$ conjugate to the three-metric
$h_{ij}$ of $\Sigma$
for the so-called ``tilted" foliation \cite{FrMa2}
and integrating the kinetic part of the volume integral, the action becomes
$S_m = \int_M d^4x [ P^{ij} \dot h_{ij} - N{\cal H} - V^i {{\cal H}_i}]$,
where the dot denotes
differentiation with respect to time, $V^i$ denotes the shift vector, and
the gravitational contributions to the Hamiltonian and momentum constraints are
given by the usual expressions.
Observe that the action vanishes identically for stationary solutions of
the vacuum Einstein equations describing stationary eternal black holes.
In this case $\dot h_{ij}=0$, the
constraint equations are satisfied, and no boundary terms remain in the
action.
Consider now the microcanonical functional integral for a
gravitational system whose timelike boundary
surfaces $B_\pm$ are located in $M_\pm$. The path integral
\begin{equation}
{\bar {\nu}}[\varepsilon_+, j_+, \sigma_+ ;
\varepsilon_-, j_-, \sigma_-] = \sum_{\sl M} \int {\cal D}H \exp(i S_m)
\label{ournu}
\end{equation}
is a functional of the energy density $\varepsilon_\pm$,
momentum density $j_\pm$, and two-metric $\sigma_{\pm}$ at the
boundaries $B_+$ and $B_-$.
The sum over $M$ refers to a sum over manifolds of different
topologies whose boundaries have topologies
$B_+ \equiv S_+ \times S^1 = S^2 \times S^1$ and
$B_- \equiv S_- \times S^1 = S^2 \times S^1$. The element $S^1$ is due to the
periodic identification in the global time direction at the boundaries when the
initial and final hypersurfaces are identified. The integral is a sum over
periodic Lorentzian metrics that satisfy the boundary conditions at $B_+$ and
$B_-$.
The eternal black hole functional integral ${\bar {\nu}}_{*}$ is obtained
when the boundary data $(\varepsilon_+, j_+, \sigma_+)$
and $(\varepsilon_-, j_-,
\sigma_-)$ of the geometries summed over correspond to the boundary data of a
general Lorentzian, stationary eternal black hole. The boundary
data of this solution can be determined at $S_+$ and $S_-$ for each slice
$\Sigma$. Observe that by virtue of the gravitational constraint equations,
these data determine
uniquely the size of the black hole horizon \cite{ensembles} and are such that
the two-metric is continuous at this horizon.
We evaluate now the functional integral in the semiclassical
approximation. This requires finding a four-metric that extremizes the
action $S_m$ and satisfies the boundary conditions $(\varepsilon_+, j_+,
\sigma_+)$ at $S_+$ and $(\varepsilon_-, j_-, \sigma_-)$
at $S_-$. Since the classical Lorentzian eternal black hole metric can be
periodically identified and placed on a
manifold whose two spatial boundaries have the desired topologies,
the resulting metric can be used to approximate the path integral.
The periodic identification alters neither the constraint
equations nor the boundary data.
In the semiclassical approximation the functional integral ${\bar{\nu}}_{*}$
becomes \cite{Ma}
\begin{equation}
{\bar{\nu}}_{*}[\varepsilon_+,j_+, \sigma_+; \varepsilon_- ,j_-,\sigma_-]
\approx \exp \big( i S_m[{\tilde N}, {\tilde V}, {\tilde h}] \big)
\approx \exp \big( 0 \big)\ ,
\label{seminuL} \end{equation}
since the microcanonical action
$S_m[{\tilde N}, {\tilde V}, {\tilde h}]$
evaluated at the periodically identified geometry vanishes identically
if the stationarity condition and the constraints are satisfied.
It is illustrative to consider now a
complex four-metric which also extremizes the microcanonical action for
eternal black hole boundary conditions and which can be used to reevaluate the
path integral (\ref{ournu}) in a steepest descent approximation.
This alternative approximation of the quantity ${\bar{\nu}}_{*}$ is useful in
understanding the relationship of the result (\ref{seminuL}) with the density
of states for an ordinary (that is, non-eternal) black hole \cite{BrYo2}.
The complex metric can be obtained from the
Lorentzian eternal black hole metric by a complexification map $\Psi$
defined
by $\Psi(N)= -iN$,
$\Psi(V^i) = -iV^i$. This map preserves the reflection symmetry and the
canonical variables $h_{ij}$ and $P^{ij}$ of the Lorentzian
solution. The complex geometry consists of two complex sectors ${\bar M}_+$
and
${\bar M}_-$ which join at the locus of points at which the lapse vanishes.
This geometry is also a solution of Einstein equations if one requires that
conical singularities do not exist at that locus for every $\Sigma$.
To do this, it is necessary to puncture each complex sector and to close
smoothly the geometry at the inner boundaries
$^3\! H_\pm = ^2\! H_\pm \times S^1$ of ${\bar M}_\pm$,
where $^2\! H_{\pm}$ denotes the
intersection of the slices $\Sigma_{\pm}$ with the black hole horizon
for the Lorentzian metric \cite{Ma}. After imposing these regularity
conditions,
the topology of each sector ${\bar M}_\pm$ becomes $R^2 \times S^2$.
However, each element $^3\! H_+$ and $^3\! H_-$ does contribute a term to the
microcanonical action for the complex geometry.
The regular complex metric is not included in the sum over Lorentzian
geometries ${\bar{\nu}}_{*}$ in (\ref{ournu}) but can be used to
approximate it by distorting the contours of
integration for both lapse and shift
into the complex plane \cite{BrYo2}. In this approximation the path
integral becomes
${\bar{\nu}}_{*}
\approx \exp ( i S_m[-i{\tilde N}, -i{\tilde V}, {\tilde h}])$,
where $S_m[-i{\tilde N}, -i{\tilde V}, {\tilde h}]$ is the
action of the complex metic when the smoothness of the
geometries at $^3 \! H_+$ and $^3 \! H_-$ is inforced.
This action turns out to be
$ S_m[-i{\tilde N}, -i{\tilde V}, {\tilde h}] = -{i} A_+/4
+{i} A_- /4$,
where $A_+$ and $A_-$ denote the surface area of the elements
${^2\!H_{\scriptscriptstyle +}}$ and ${^2\! H_{\scriptscriptstyle -}}$
\cite{Ma}.
Since the periodic identification and the complexification $\Psi$ do not alter
the boundary data nor the gravitational constraint equations, the area $A_+$ of
${^2\!H_{\scriptscriptstyle +}}$ coincides with the area $A_-$ of
${^2\!H_{\scriptscriptstyle -}}$:
$A_+ (\varepsilon_+, j_+, \sigma_+) = A_- (\varepsilon_-, j_-, \sigma_-) \equiv
A_H$.
This implies that, in agreement with (\ref{seminuL}), the eternal black hole
functional integral is
${\bar{\nu}}_{*}[\varepsilon_+ ,j_+, \sigma_+; \varepsilon_-,
j_-, \sigma_-] \approx \exp{( A_H/4 - A_H/4)} = \exp (0)$
in the ``zero-loop" approximation.
If the microcanonical functional integral (\ref{ournu}) is
interpreted as the density of states of the statistical system, it is possible
to express ${\bar{\nu}}_{*}$ approximately as
${\bar{\nu}}_{*}[\varepsilon_+, j_+, \sigma_+; \varepsilon_-, j_-, \sigma_-]
\approx \exp({\cal S} [\varepsilon_+, j_+, \sigma_+; \varepsilon_-, j_-,
\sigma_-]) $,
where ${\cal S}$ represents the total entropy of the
system. The above result
implies that the entropy for the system
in the semiclassical approximation is
\begin{equation}
{\cal S} \approx {1\over 4}{A_H} - {1\over 4} {A_H} = 0 \ ,
\label{entropy}
\end{equation}
where $A_H$ is the area of the horizon of the
physical eternal black hole solution that classically
approximates the system \cite{Ma}.
The total entropy is given formally by the subtraction
${\cal S} = {{\cal S}_+}[\varepsilon_+ ,j_+, \sigma_+] - {{\cal
S}_-}[\varepsilon_- ,j_-, \sigma_-]$, where ${{\cal S}_+}$ and
${{\cal S}_-}$ can be interpreted as the semiclassical entropies
associated with
the external ($M_+$) and internal ($M_-$) regions respectively of the eternal
black hole system.
\section{Conclusions}
The functional integral (\ref{seminuL}) refers to a
quantum-statistical system which is classically approximated by a general
stationary, eternal black hole solution of Einstein equations
within a region bounded by two timelike surfaces $B_+$ and $B_-$.
Its semiclassical value is a consequence of the choice of boundary data, the
gravitational
constraint equations, and the vanishing of the microcanonical action for the
four-geometries that satisfy the boundary conditions and approximate the path
integral. The calculation presented above applies to any distorted black hole
in the strong gravity regime.
It indicates that a pure state (of zero
entropy) can be defined not only for
matter fields perturbations propagating in the spacetime of an eternal black
hole but also for the gravitational field itself. This is
physically appealing: the initial data for the eternal black hole specified at
the spacelike hypersurface $\Sigma$ contain all the information required for
the evolution of both the exterior and interior parts of a physical black hole.
The entropy associated with $\Sigma$ must therefore equal zero.
Since in a microcanonical description it seems natural to
relate the external and internal degrees of freedom of a black hole with the
boundary data at the surfaces $B_+$ and $B_-$ respectively \cite{FrMa2}, we
believe that the microcanonical functional integral for eternal black hole
systems opens the possibility of extending path integral formulation of
black hole thermodynamics to situations when internal degrees of freedom
are present and allows the study of gravitational
statistical properties in terms of a single pure state \cite{Ma}.
These conclusions are in complete agreement with thermofield dynamics
descriptions of quantum processes and, in particular, with the
application of this approach to black hole thermodynamics developed originally
by Israel \cite{Is} for small perturbations. They strongly suggest that
the thermofield dynamics description of quantum field processes in a curved
background can be extended beyond perturbations to the gravitational field
itself of distorted eternal black holes.
I am grateful to Valeri Frolov and Werner Israel for helpful comments.
This work was supported by the Natural Sciences and Engineering Research
Council of Canada.
|
\section{Introduction}
In the Minimal Supersymmetric Standard Model (MSSM) \cite{1,2} two Higgs
doublets are necessary,
leading to five physical Higgs bosons $h^0, H^0, A^0$, and $H^\pm$
\cite{3,4}. If all supersymmetric
(SUSY) particles are very heavy, the charged Higgs boson $H^+$ decays
dominantly into $t \bar b$; the
decays $H^+ \rightarrow \tau^+ \nu$ and/or $H^+ \rightarrow W^+ h^0$ are
dominant below the $t \bar b$
threshold \cite{3,4b}. In ref.~\cite{5} all decay modes of $H^+$ including
the SUSY-particle modes were
studied in detail; it was shown that the SUSY decay modes $H^+ \rightarrow
\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i \bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j (i,j = 1,2)$
can be dominant in a large region of the MSSM parameter space due to large
$t$ and $b$ quark Yukawa
couplings and large $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi$- and $\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi$-mixings, and that this could have a
decisive impact on $H^+$
searches at future colliders.
Here $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i \, (\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j)$ are the scalar top (scalar bottom) mass eigenstates
which are mixtures of
$\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_L$ and $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_R$ ($\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_L$ and $\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_R$).
The standard QCD corrections are very large for the width of $H^+
\rightarrow c \bar s$ and can be
large ($+10$\% to $-50$\%) for that of $H^+\rightarrow t \bar b$ \cite{6a}.
The QCD corrections from the SUSY-particle loops are calculated within the
MSSM for $H^+\rightarrow
t \bar b$ in \cite{6b} and turn out to be non-negligible ($\sim$ 10\%)
for certain values of the MSSM parameters. This suggests that the QCD
corrections to
$H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ could
also be large. Therefore it should be examined whether the result in
\cite{5} remains valid after
including the QCD corrections.
In this paper we calculate the ${\cal O}(\alpha_s)$ QCD corrections to the
width of $H^+ \rightarrow
\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i \bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j$ within the MSSM. To the best of our knowledge they are
not known in the literature.
We obtain the complete ${\cal O}(\alpha_s)$ corrected width in the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi
renormalization scheme (i.e.
the \ifmmode{\overline{\rm MS}} \else{$\overline{\rm MS}$} \fi scheme with dimensional reduction \cite{7}) including all quark
mass terms and $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_L-\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_R$ mixings.
The main complication here is that the $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_L-\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_R$ mixing angles are
renormalized by the SUSY QCD
corrections. We find that the corrections to the $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ width are
significant but that the
$\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ mode is still dominant in a wide parameter range.
\section{Tree level result}
We first review the tree level results \cite{5}. The squark mass matrix in
the basis ($\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_L$, $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_R$),
with $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi=\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi$ or $\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi$, is given by \cite{3,4} \begin{equation}
\left( \begin{array}{cc}m_{LL}^2 & m_{LR}^2 \\ m_{RL}^2 & m_{RR}^2
\end{array} \right)=
(R^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi})^{\dagger}\left( \begin{array}{cc}m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_1}^2 & 0 \\ 0 &
m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_2}^2 \end{array}
\right)R^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}, \end{equation}
where
\begin{eqnarray}
m_{LL}^2 &=& M_{\tilde{Q}}^2+m_q^2+m_Z^2\cos 2\beta (I_q-Q_q\sin^2\theta_W), \\
m_{RR}^2 &=& M_{\{\tilde{U},
\tilde{D}\}}^2+m_q^2+m_Z^2\cos 2\beta Q_q\sin^2
\theta_W, \\
m_{LR}^2=m_{RL}^2 &=& \left\{ \begin{array}{ll}
m_t(A_t-\mu\cot\beta) & (\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi=\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi) \\
m_b(A_b-\mu\tan\beta) & (\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi=\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi) \end{array} \right. , \end{eqnarray}
and
\begin{equation} \label{5}
R^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}_{i\alpha}=\left(
\begin{array}{cc}\cos\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi & \sin\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi \\
-\sin\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi & \cos\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi \end{array}\right) . \end{equation}
Here the mass eigenstates $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i(i=1,2)$ (with $m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_1}<m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_2}$) are
related to the
SU(2)$_L$ eigenstates $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_{\alpha}(\alpha=L,R)$ as
$\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i=R^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}_{i\alpha}\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_{\alpha}$. Note that
in the sign convention used here the parameters $A_{t,b}$ correspond to
$(-A_{t,b})$ of ref.\cite{5}.
The tree-level decay width of $H^+ \rightarrow \ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i \bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j$ is then
given by (see Fig.~1a)
\begin{equation} \label{6}
\Gamma^{(0)}(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j) =\frac{N_C\kappa}{16\pi
m_H^3}|G_{ij}|^2\,, \end{equation}
where $m_H$ is the $H^+$ mass,
$\kappa=\kappa(m_H^2, m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2, m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2 )$, $\kappa(x,y,z)\equiv
((x-y-z)^2-4yz)^{1/2}$, $N_C=3$,
and \begin{eqnarray} \label{7}
G_{ij}&=& \frac{g}{\sqrt{2}m_W}
R^{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi}\left( \begin{array}{cc}
m_b^2\tan\beta+m_t^2\cot\beta-m_W^2\sin 2\beta & m_b(A_b\tan\beta+\mu) \\
m_t(A_t\cot\beta+\mu) &
2m_tm_b/\sin 2\beta \end{array} \right)(R^{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi})^{\dagger} \nonumber\\ \end{eqnarray}
are the $H^+\bar{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi}_i\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j$ couplings \cite{3,4}, with $g$ being the
SU(2) coupling.
\section{QCD virtual corrections}
The ${\cal O}(\alpha_s)$ QCD virtual corrections to
$H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j$ stem from the
diagrams of Fig.~1b (vertex corrections) and 1c (wave function
corrections). For simplicity we use
in this paper the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi renormalization scheme\footnote{Strictly speaking,
our renormalization
scheme is the $\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi'$ scheme \cite{7b} where the ``$\epsilon$-scalar
mass'' is absorbed into
$M_{\tilde{Q},\tilde{U},\tilde{D}}^2$.} for all parameters which receive
the QCD corrections, i.e.
$m_{t,b}$, $A_{t,b}$, and $M_{\tilde{Q},\tilde{U},\tilde{D}}$. The
renormalized squark
mixing angle $\theta_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}$ is then defined by the relations (1--5) in
terms of the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi parameters
$m_{t,b}$, $A_{t,b}$, and $M_{\tilde{Q},\tilde{U},\tilde{D}}$.
The one-loop corrected decay amplitudes $G_{ij}^{\rm corr}$ are expressed
as \begin{equation}
G_{ij}^{\rm corr}=G_{ij}+\delta G_{ij}^{(v)}+\delta G_{ij}^{(w)}, \end{equation}
where $G_{ij}$ are defined by (7) in terms of the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi parameters, and
$\delta G_{ij}^{(v)}$ and
$\delta G_{ij}^{(w)}$ are the vertex and squark wave function corrections,
respectively.
The vertex corrections $\delta G_{ij}^{(v)}$ are calculated from the graphs
of Fig.~1b as
\begin{eqnarray}
\delta G_{ij}^{(v)}&=& \frac{\alpha_s C_F}{4\pi} \left[ \{B_0(m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2,
0, m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2)+B_0
(m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2, 0, m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2) -B_0(m_H^2, m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2, m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2)
\right. \nonumber \\
&& -2(m_H^2-m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2-m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2)
C_0(m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2, \lambda^2, m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2)\}G_{ij} \nonumber \\ &&
-B_0(m_H^2, m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_k}^2,
m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_l}^2)G_{kl}S^{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi}_{ik}S^{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_{lj} \nonumber \\
&& +2\{ (\alpha_{LL})_{ij}(m_ty_2+m_by_1)+(\alpha_{RR})_{ij}(m_ty_1+m_by_2)\}
B_0(m_H^2, m_b^2, m_t^2) \nonumber\\
&& +2\{ (\alpha_{LL})_{ij}m_ty_2+(\alpha_{LR})_{ij}m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}y_1
+(\alpha_{RL})_{ij}m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}y_2+
(\alpha_{RR})_{ij}m_ty_1\} B_0(m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_t^2) \nonumber\\ && +
2\{ (\alpha_{LL})_{ij}m_by_1+(\alpha_{LR})_{ij}m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}y_1
+(\alpha_{RL})_{ij}m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}y_2+
(\alpha_{RR})_{ij}m_by_2\} B_0(m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_b^2) \nonumber\\ &&
+2\{ (m_t^2+m_b^2-m_H^2)m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}
((\alpha_{LR})_{ij}y_1+(\alpha_{RL})_{ij}y_2) \nonumber\\ &&
+(m_b^2+m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2-m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2)m_t
((\alpha_{LL})_{ij}y_2+(\alpha_{RR})_{ij}y_1) \nonumber\\ &&
+(m_t^2+m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2-m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2)m_b
((\alpha_{LL})_{ij}y_1+(\alpha_{RR})_{ij}y_2) \nonumber\\ && \left. +
2m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}m_tm_b((\alpha_{LR})_{ij}y_2+(\alpha_{RL})_{ij}y_1) \} C_0(m_b^2,
m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_t^2) \right] ,
\end{eqnarray}
where $C_F=4/3$,
\begin{equation}
S^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}=\left( \begin{array}{cc} \cos 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi & -\sin 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi \\
-\sin 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi & -\cos 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi
\end{array} \right) ,
\end{equation}
\[
\alpha_{LL}=\left( \begin{array}{rr}
\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi & -\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \\
-\sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi &
\sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \end{array} \right) , \;\;\;
\alpha_{LR}=\left( \begin{array}{rr}
-\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi & -\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \\
\sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi &
\sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \end{array} \right) , \]
\begin{equation}
\alpha_{RL}=\left( \begin{array}{rr}
-\sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi & \sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \\
-\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi &
\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \end{array} \right) , \;\;\;
\alpha_{RR}=\left( \begin{array}{rr}
\sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi & \sin\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \\
\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\sin\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi &
\cos\theta_\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\cos\theta_\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi \end{array} \right) ,
\end{equation}
\begin{equation}
y_1=\frac{g}{\sqrt{2}m_W}m_b\tan\beta=h_b\sin\beta, \;\;\;
y_2=\frac{g}{\sqrt{2}m_W}
m_t\cot\beta=h_t\cos\beta, \end{equation}
and $m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}$ is the gluino mass. A gluon mass $\lambda$ is introduced to
regularize
the infrared divergences.
Here we define the functions $A$, $B_0$, $B_1$ and $C_0$ as in \cite{pv}
($\Delta = 2/(4-D) - \gamma_E +\log 4\pi$): \begin{eqnarray}
A(m^2)&=&
\int\frac{d^Dq}{i\pi^2}\frac{1}{q^2-m^2}=m^2(\Delta +\log(Q^2/m^2)+1), \\
B_0(k^2, m_1^2, m_2^2)&=&
\int\frac{d^Dq}{i\pi^2}\frac{1}{(q^2-m_1^2)((q+k)^2-m_2^2)} \nonumber\\
&=&\Delta-\int_0^1dz\log\frac{(1-z)m_1^2+zm_2^2-z(1-z)k^2-i\delta}{Q^2}, \\
B_1(k^2, m_1^2, m_2^2)&=&
\frac{1}{k_{\mu}}\int\frac{d^Dq}{i\pi^2}
\frac{q_{\mu}}{(q^2-m_1^2)((q+k)^2-m_2^2)} \nonumber\\
&=&-\frac{\Delta}{2}+\int_0^1dz\,z\log
\frac{(1-z)m_1^2+zm_2^2-z(1-z)k^2-i\delta}{Q^2}, \\ C_0(m_1^2, m_2^2, m_3^2)&=&
\int\frac{d^Dq}{i\pi^2}\frac{1}{(q^2-m_1^2)((q+k_2)^2-m_2^2)
((q+p)^2-m_3^2)} \nonumber\\
&=&
-\int_0^1dx\int_0^1dy\int_0^1dz\delta(1-x-y-z)\,
\times \nonumber\\
&&(xm_1^2+ym_2^2+zm_3^2-xym_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2
-yzm_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2-xzm_H^2-i\delta)^{-1}.
\end{eqnarray}
Here $p$ and $k_2$ are respectively the external momenta of $H^+$ and
$\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j$, and
$Q$ is the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi renormalization scale. Note that $\Delta$ is omitted in
the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi scheme.
The squark wave function corrections $\delta G_{ij}^{(w)}$ are expressed as
\begin{equation} \label{17}
\delta G_{ij}^{(w)}=-\frac{1}{2}\left[ \dot{\Pi}_{ii}^{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi}(m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2)
+\dot{\Pi}_{jj}^{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}(m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2)\right] G_{ij}
-\frac{\Pi_{ii'}^{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi}(m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2)}{m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2-m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_{i'}}^2} G_{i'j}
-\frac{\Pi_{j'j}^{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}(m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2)}{m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2-m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_{j'}}^2} G_{ij'},
\end{equation}
where $i\neq i'$ and $j\neq j'$.
$\Pi_{ij}^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}(k^2)$ are the one-loop corrections to the two-point
functions of
$\bar{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}_i\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_j$, which are obtained from the graphs of Fig.~1c.
$\dot{\Pi}(k^2)$
denotes the derivative with respect to $k^2$. The last two terms in (\ref{17})
represent the corrections due to the renormalization of the $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi$-mixings.
The explicit forms are
\begin{eqnarray}
\dot{\Pi}_{ii}^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2)&=&
\frac{\alpha_sC_F}{4\pi}\left[ -3B_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, 0, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2)
-2B_1(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, 0, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2)
-4m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2\dot{B}_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, \lambda^2, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2) \right. \nonumber\\
&&-2m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2\dot{B}_1(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, 0, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2)
-4m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2\dot{B}_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_q^2) -4B_1(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2,
m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_q^2) \nonumber\\
&&\left. -4m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2\dot{B}_1(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_q^2)
+(-)^{i-1}4\sin 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi m_qm_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}\dot{B}_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2,
m_q^2) \right] ,
\end{eqnarray}
and
\begin{equation}
\Pi_{i'i}^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2)=\frac{\alpha_sC_F}{4\pi}
\left[ \frac{1}{2}\sin 4\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi (A(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_2}^2)-A(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_1}^2))
+4\cos 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi m_qm_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}B_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_q^2)\right] . \end{equation}
The one-loop corrected decay width in the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi scheme is then given by \begin{equation}
\Gamma(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j)
=\frac{N_C\kappa_{\rm pole}}{16\pi m_H^3}[|G_{ij}|^2 +2G_{ij}{\rm
Re}(\delta G_{ij}^{(v)}
+\delta G_{ij}^{(w)})]. \label{20} \end{equation}
Here $\kappa_{\rm pole}$ refers to $\kappa$ in (6) evaluated with pole
squark masses.
The width of (20) is infrared divergent.
In the numerical analysis we take the pole quark masses as inputs. The
\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi quark
masses are obtained from the pole quark masses by using \begin{eqnarray}
m_q(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}&=&m_q({\rm pole})-\frac{\alpha_sC_F}{4\pi}
\left[ 2m_q(B_0(m_q^2, 0, m_q^2)-B_1(m_q^2, 0, m_q^2)) \right. \nonumber\\
&&+\sin 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}(B_0(m_q^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_1}^2)
-B_0(m_q^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_2}^2))\nonumber\\ &&
\left. +m_q(B_1(m_q^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_1}^2) +B_1(m_q^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2,
m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_2}^2))\right] , \end{eqnarray}
which is derived from the graphs of Fig.~1d. Furthermore, in the phase
space term
$\kappa_{\rm pole}$ in (20) we have to take the pole squark masses given by
\begin{eqnarray}
m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2({\rm
pole})&=&m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}-\Pi_{ii}^{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi}(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2) \nonumber \\
&=&m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}+\frac{\alpha_sC_F}{4\pi}
\left[ -3A(0)+4m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2B_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, 0, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2)
+2m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2B_1(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, 0, m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2) \right. \nonumber \\
&&-\cos^22\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi A(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2)-\sin^22\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi
A(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_{i'}}^2)+4A(m_q^2)
\nonumber\\
&&+4m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2B_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_q^2)
+4m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2B_1(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2, m_q^2) \nonumber\\
&&\left. -(-)^{i-1}4\sin 2\theta_\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi m_qm_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi} B_0(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}^2, m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}^2,
m_q^2) \right] .
\label{22} \end{eqnarray}
\section{Gluon emission}
The infrared divergences in (\ref{20}) are cancelled by including the
${\cal O}(\alpha_s)$
contribution from real gluon emission from $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi$ and $\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ (Fig.~1e).
The decay width of $H^+(p)\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i(k_1)+\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j(k_2)+g(k_3)$ is
given
in terms of the \ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi parameters as \begin{equation}
\Gamma(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_jg)
=\frac{\alpha_sC_FN_C|G_{ij}|^2}{4\pi^2 m_H}
[(m_H^2-m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2-m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2)I_{12}
-m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}^2I_{11}-m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}^2I_{22}-I_1-I_2]. \label{23} \end{equation}
The functions $I_n$, and $I_{nm}$ are defined as \cite{8} \begin{equation}
I_{i_1\ldots i_n}=\frac{1}{\pi^2}
\int\frac{d^3k_1}{2E_1}\frac{d^3k_2}{2E_2}\frac{d^3k_3}{2E_3}
\delta^4(p-k_1-k_2-k_3)\frac{1}
{(2k_3k_{i_1}+\lambda^2)\ldots(2k_3k_{i_n}+\lambda^2)}. \end{equation}
The explicit forms of $I_{i_1\ldots i_n}$ are given in \cite{8}. In (\ref{23}),
$I_{11,22,12}$ are infrared divergent. We have checked that the infrared
divergences in
(\ref{23}) cancel those in (\ref{20}). In the numerical analysis we define
the corrected
decay width as
$\Gamma^{\rm corr}(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j)\equiv
\Gamma(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j)+
\Gamma(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_jg)$.
\section{Numerical results and conclusions} As in ref.\cite{5}, we choose
\{ $m_H$, $m_{t,b}$(pole), $M$, $\mu$, $\tan\beta$, $M_{\tilde{Q}}$, $A$ \}
as the basic input parameters of the MSSM, taking
$M=(\alpha_2/\alpha_s)m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}=(3/5\tan^2\theta_W)M'$,
$M_{\tilde{Q}}\equiv M_{\tilde{Q}}(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}=
M_{\tilde{U}}(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}=M_{\tilde{D}}(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}=M_{\tilde{L}}$ and
$A\equiv A_t(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}=A_b(Q)_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi}=A_{\tau}$. Here $M$ ($M'$) is
the SU(2) (U(1)) gaugino mass, $\alpha_2=g^2/4\pi$, and ($M_{\tilde{L}}$,
$A_{\tau}$)
are the mass matrix parameters of the slepton sector \cite{5}.
The parameters $M$, $M'$, $M_{\tilde{L}}$, and $A_{\tau}$ do not receive
${\cal O}(\alpha_s)$ QCD corrections.
The theoretical and experimental constraints for the basic input
parameters are described in ref.\cite{5}.
We take $m_Z=91.2$GeV, $m_W=80$GeV,
$m_t({\rm pole})=180$GeV \cite{mtop}, $m_b({\rm pole})=5$GeV,
$\sin^2\theta_W=0.23$
and $\alpha_2=\alpha_2(m_Z)=\alpha/\sin^2\theta_W =(1/129)/0.23=0.0337$.
For the running QCD coupling at the renormalization scale $Q$,
$\alpha_s=\alpha_s(Q)$,
we always take the one-loop expression $\alpha_s(Q)=
12\pi/\{(33-2n_f)\ln(Q^2/\Lambda_{n_f}^2)\}$,
with $\alpha_s(m_Z)=0.12$, and the number of quark flavors $n_f=5(6)$ for
$m_b<Q\le m_t$ (for $Q>m_t$).
We define the QCD corrections as the difference between the ${\cal
O}(\alpha_s)$ corrected
width $\Gamma^{\rm corr}_{ij}\equiv \Gamma^{\rm
corr}(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j)$
of eqs.(20) plus (23) and the
tree-level width $\Gamma_{ij}^{\rm tree}\equiv
\Gamma^{(0)}(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j)$
of eq.(6) where ($m_q(\rm pole)$, $M_{\tilde{Q}}$, $A$) are substituted for
($m_q$, $M_{\tilde{Q},\tilde{U},\tilde{D}}$, $A_q$). The QCD corrections depend
on the renormalization scale $Q$. We choose the optimum value of $Q$
($Q_{\rm opt}$)
such that $\displaystyle\Delta(Q)\equiv\sum_{q=t,b}\sum_{i=1,2}
(m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}({\rm tree})-m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}({\rm pole}))^2$ is minimized, where
$m_{\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i}(\rm tree)$ refers to
the $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_i$ mass defined by (1--4) calculated with $m_q(\rm pole)$,
$M_{\tilde{Q}}$, and $A$.
In order not to vary too many parameters, in the following we fix $\mu=300$GeV,
and take the values of $M$ and $\tan\beta$ such that
$m_{\tilde{\chi}_1^0}\simeq 50$GeV
as in \cite{5} where $\tilde{\chi}_1^0$ is the lightest neutralino.
In Fig.2 we show the $m_H$ dependence of the tree-level and corrected widths
$\displaystyle\Gamma^{\rm tree}(\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi})\equiv
\sum_{i,j=1,2}\Gamma^{\rm tree}_{ij}$ and
$\displaystyle\Gamma^{\rm corr}(\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi})\equiv
\sum_{i,j=1,2}\Gamma^{\rm corr}_{ij}$, and
the tree-level branching ratio $\displaystyle B^{\rm tree}(\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi})\equiv
\sum_{i,j=1,2}B^{\rm tree}(H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j)$ \cite{5} for (a)
$M_{\tilde{Q}}=250$GeV, $A=650$GeV, $\tan\beta=2$, $M=120$GeV, and (b)
$M_{\tilde{Q}}=136$GeV, $A=260$GeV, $\tan\beta=12$, $M=110$GeV.
In these two cases we have (in GeV units): (a) $Q_{\rm opt}=216.5$,
$m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}=380$, $(m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_1},m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_2}, m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_1},m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_2})\{(\rm
tree),(Q_{\rm opt})_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi},
(\rm pole)\}=$ $\{(60,429,251,254)$, $(70,420,251,254)$, $(58,428,258,262)\}$,
$m_{\tilde{\chi}_1^+}=94$, and
(b) $Q_{\rm opt}=213.5$, $m_{\ifmmode{\tilde{g}} \else{$\tilde{g}$} \fi}=349$, $(m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_1},m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_2},
m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_1},m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_2})\{(\rm tree), (Q_{\rm opt})_{\ifmmode{\overline{\rm DR}} \else{$\overline{\rm DR}$} \fi},(\rm pole)\}=$
$\{(81,302,62,193)$, $(74,292,120,164)$, $(66,302,126,171)\}$,
$m_{\tilde{\chi}_1^+}=98$.
Here $\tilde{\chi}_1^+$ is the lighter chargino. In both cases we see that the
$\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ mode dominates the $H^+$ decay in a wide $m_{H^+}$ range at the
tree level, and that the QCD corrections to the $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ mode are
significant,
but that as a whole they do not invalidate the $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ mode dominance.
In Table 1 we show the values of the $B^{\rm tree}(\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi})$, the QCD
corrections
$C\equiv(\Gamma^{\rm corr}(\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi})- \Gamma^{\rm tree}(\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}))
/\Gamma^{\rm tree}(\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi})$, and $C_{ij}\equiv(\Gamma^{\rm corr}_{ij}
-\Gamma^{\rm tree}_{ij})/ \Gamma^{\rm tree}_{ij}$ for
typical values of $M_{\tilde{Q}}$ and $A$, for (a)
$m_{H^+}=400$GeV, $\tan\beta=2$, $M=120$GeV, and
(b) $m_{H^+}=400$GeV, $\tan\beta=12$, $M=110$GeV.
We see again that the $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$ mode dominates in a wide region also
when the QCD corrections are included. The QCD corrections can be very large
at some points of $(M_{\tilde{Q}},A)$; e.g. $C=-0.692$ at (175GeV, 0GeV) and
$C_{12}=0.734$ at (225GeV, $-350$GeV) in Table 1b. This occurs when
$m_H\sim m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}+m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}$.
This enhancement is just a kinematical effect due to the QCD corrections to
$m_{\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i}$ and $m_{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi_j}$.
In conclusion, we have calculated the ${\cal O}(\alpha_s)$ QCD corrections
to the
decay width of $H^+\rightarrow\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi_i\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}_j$, including all quark mass
terms and $\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_L-\ifmmode{\tilde{q}} \else{$\tilde{q}$} \fi_R$ mixing.
We find that the QCD corrections are significant but that they do not
invalidate our
previous conclusion at tree-level about the dominance of the $\ifmmode{\tilde{t}} \else{$\tilde{t}$} \fi\bar{\ifmmode{\tilde{b}} \else{$\tilde{b}$} \fi}$
mode in
a wide MSSM parameter region.
\section*{Acknowledgements}
The work of Y.Y. was supported in part by the Fellowships of the Japan
Society for the Promotion of
Scienceand the Grant-in-Aid for Scientific
Research from the Ministry of Education, Science and Culture of Japan, No.
06-1923 and 07-1923.
The work of A.B., H.E., and W.M. was supported by the ``Fonds zur
F\"orderung der
wissenschaftlichen Forschung'' of Austria, project no. P10843-PHY.
The authors are grateful to Y. Kizukuri for the collaboration at the early
stage of this work.
\clearpage
|
\section{Theoretical Treatment of Charm and
\section{Introduction}
\subsection{Goals and Obstacles}
As explained in detail elsewhere in this
review, a precise measurement of the lifetimes of the
various weakly decaying charm
and beauty hadrons possesses great experimental
value per se as well as in searches for
$B^0 - \bar B^0$ and $D^0 - \bar D^0$ oscillations. Yet there exists
also a strong theoretical interest in determining and
interpreting those lifetimes;
I want to sketch that
first in rather qualitative terms and
give specifics later.
Weak decays of hadrons depend on fundamental parameters
of the Standard Model, in particular on the KM parameters and
quark masses. It is eminently important to reliably determine
their values from data.
Alas -- this is easier said than done theoretically (and
experimentally)! For in such an endeavour we have to face the
"Dichotomy of the Two Worlds". On the one hand there is the
"Theorists' World" where quarks and gluons are the relevant
strongly interacting entities; it is in this short-distance or
Femto World where theorists like to formulate their
fundamental
theories. On the other hand there is the "Real World" where
hadrons constitute the relevant degrees of freedom; it is in
that world where everyone (including theorists) lives and
measurements are performed. To formulate predictions from
the Theorists' World in the language of the Real World and to
translate findings from the Real World back into the idioms of
the Theorists' World represents the theoretical challenge one
faces.
One quantitative measure for the difference between the two
worlds is provided by the lifetimes of the weakly decaying
hadrons carrying the same flavour. On the quark level
there is obviously only a single lifetime for a given flavour.
Yet in the `Real
World' hadrons carrying the same flavour
quantum number possess different
(and even vastly different) lifetimes; e.g.,
$\tau (K^+)/\tau (K_S)\simeq 140$,
$\tau (D^+)/\tau (D^0)\simeq 2.5$ and
$0.9 \leq \tau (B^-)/\tau (B_d)\leq 1.2$. On the other hand
deviations of the
lifetime ratios from unity evidently decrease for
an
increasing heavy-flavour mass. This is as expected
in a simple `two-component' picture:
in the limit of
$m_Q$ -- the mass of the heavy-flavour quark $Q$ --
going to infinity
the `Spectator Ansatz' should hold
where the lifetimes of
all hadrons $H_Q$ containing $Q$ coincide; for finite
values of $m_Q$ there are {\em pre-asymptotic} corrections.
Two lines of reasoning support this qualitative picture:
\noindent ($\alpha$) The decay width of a quark $Q$ increases very
quickly
with
its mass $m_Q$:
$$\Gamma _Q \propto G_F^2 m_Q^5 \eqno(1)$$
for $m_Q \ll M_W$ changing into
$\Gamma _Q \propto \alpha m_Q^3/M_W^2$ for
$m_Q \gg M_W$ \footnote{The asymptotic scaling behaviour
$\Gamma _Q \propto m_Q^3$ (rather than
$\Gamma _Q\propto m_Q$) reflects
the coupling of the {\em longitudinal} $W$ boson -- the original
Higgs field -- to the top quark.}. Therefore for $m_Q$ sufficiently
large, its decay width is bound to exceed $\Lambda _{QCD}$,
i.e. the quark $Q$ decays before it can hadronize
\footnote{For top decays this happens for
$m_t\geq 130$ GeV \cite{RAPALLO}.}. Then there is
obviously a universal lifetime for such a heavy flavour and
the spectator ansatz applies trivially.
\noindent ($\beta$) Analysing all
{\em non}-spectator reactions explicitely, one finds their widths to
increase
with a smaller power of $m_Q$ than the spectator process;
thus one arrives
at the spectator ansatz as the asymptotic case.
The most relevant phenomenological question then is how quickly
the limit of universal lifetimes is approached.
\subsection{Phenomenology: Legends with Truths}
Some mechanisms had
been identified very early on that generate differences
in the lifetimes of hadrons $H_Q$ with the same heavy
flavour $Q$ : "Weak Annihilation" (=WA) of $Q$ with the light
valence antiquark for mesons or
"W Scattering" (=WS) with the valence diquark system for
baryons \footnote{A distinction is often made between W exchange
in the
s and in the t channel with the former case referred to
as `weak annihilation' and the latter as `W exchange'. This
classification
is however artificial since the two operators mix already under
one-loop renormalization in QCD, as discussed later on.
Both cases will summarily be
referred to as WA.}. Such an analysis had first been undertaken
for charm decays.
Since WA contributes to Cabibbo allowed decays of $D^0$, but not of
$D^+$
mesons (in the valence quark description), it creates a difference in
$\tau (D^0)$ vs. $\tau (D^+)$. However the WA rate is doubly
suppressed relative to the spectator rate, namely by the helicity
factor $(m_q/m_c)^2$ with
$m_q$ denoting the largest mass in the final state and by the
`wavefunction overlap' factor $(f_D/m_c)^2$
reflecting the practically zero
range of the low-energy weak interactions:
$$\Gamma ^{(0)}_{W-X}(D^0)
\propto G_F^2f^2_Dm_q^2m_c \, .\eqno(2)$$
Therefore it had
originally been suggested that already charm hadrons should
possess approximately equal lifetimes. It then came
as quite a
surprise when observations showed it to be otherwise -- in
particular since the first data suggested a considerably larger
value for $\tau (D^+)/\tau (D^0)$ than measured today. This
near-shock caused a re-appraisal of the theoretical situation; its
results at that time can be summarized in
four main points:
\noindent (i) One source for a lifetime difference had
too quickly been discarded as insignificant. Cabibbo-allowed
nonleptonic decays of $D^+$
-- but not of $D^0$ -- mesons
produce two antiquarks in the final state that
carry the same flavour:
$$D^+ = [c\bar d] \rightarrow (s\bar d u) \bar d $$
Thus one has to allow for the {\em interference}
between different
quark diagrams in $D^+$ , yet not in $D^0$ decays; the
$\bar d$ valence antiquark in $D^+$ mesons thus ceases to
play the role of an uninvolved bystander and a difference
in the $D^+$ vs. $D^0$ lifetimes will arise.
While this interference had been included in descriptions of
the $D\rightarrow K \pi$ two-body modes it was
ignored for total widths. For it was
thought (often without stating it
explicitely) that the required coherence
would not be maintained between the two amplitudes
when applied to inclusive transitions.
This assumption was challenged
in ref.\cite{PI} where it was argued that even the two
inclusive amplitudes remain
sufficiently coherent. The interference
turns out to be destructive, i.e. it prolongs $\tau (D^+)$
over $\tau (D^0)$, but only once the QCD radiative corrections
have been included. This effect is usually referred to as
`Pauli Interference' (=PI) although such a name would be
misleading if it is interpreted as suggesting that the interference
is automatically destructive.
\noindent (ii) It was argued \cite{SONI} that the helicity suppression
of the WA contribution to $D$ decays can be vitiated. The diagram in
Fig.1 contains gluon bremsstrahlung off the
initial antiquark line in the $W$-exchange reaction;
evaluating this particular diagram
explicitely one finds:
$$\Gamma ^{(1)}_{W-X}(D^0)\propto
(\alpha_s /\pi ) G_F^2(f_D/\aver{E_{\bar q}})^2m_c^5
\eqno(3)$$
with
$\aver {E_{\bar q}}$ denoting the average energy of the
initial antiquark $\bar q$
\footnote{The $1/\aver {E_{\bar q}}$ term in the amplitude
derives from the
propagator in the diagram.}. Using a non-relativistic
wavefunction for the decaying meson one has
$\aver{E_{\bar q}}\simeq m_q$. This
contribution, although of higher order in $\alpha_s$, would
dominate over the lowest order term
$\Gamma ^{(0)}_{W-X}$ since helicity suppression has
apparently been vitiated and the decay constant $f_D$
is now
calibrated by $\aver{E_{\bar q}}$ with
$f_D/\aver{E_{\bar q}} \sim {\cal O}(1)$ rather than
$f_D/m_c \ll 1$.
The spectator picture would still apply at asymptotic quark
masses, since $\Gamma ^{(1)}_{W-X}/
\Gamma _c\propto (f_D/\aver{E_{\bar q}})^2 \rightarrow 0$
as $m_c\rightarrow \infty$ due to $f_D \propto 1/\sqrt{m_c}$.
Yet
if eq.(3) were indeed to hold, it would have a
dramatic impact on the theoretical description of weak
heavy-flavour decays: the impact of this
particular pre-asymptotic
correction, namely WA, would be enhanced considerably and
actually be quite significant even in beauty decays.
Alternatively
it had been suggested \cite{MINKOWSKI}
that the wavefunction of the $D$ meson
contains a $c \bar q g$ component where the $c \bar q$ pair forms
a spin-one configuration with the gluon $g$ balancing the spin
of the $c\bar q$ pair.
\noindent Both effects, namely PI and WA, work in the same
direction, i.e. both enhance $\tau (D^+)$ over $\tau (D^0)$.
\noindent (iii) A rich structure emerges in the decays of
charm baryons \cite{BARYONS1,BARYONS2,BARYONS3}:
$\bullet$ On the one hand WS contributes to the Cabibbo
allowed $\Lambda _c$ and $\Xi _c^0$ decays; one should
also keep in
mind that WS is {\em not} helicity suppressed in baryon
decays already to lowest order in the strong coupling. It is still
reduced in size by the corresponding wavefunction overlap;
yet that is at least partially off-set by WS being described by
two-body phasespace versus the three-body phase space of the
spectator transition; this relative enhancement in
phase space can be
estimated to be roughly of order $16\pi ^2$.
$\bullet$ PI affects the
$\Lambda _c$, $\Xi _c^{0,+}$ and $\Omega _c$
widths in various
ways, generating destructive as well as constructive contributions!
This also strongly suggests that it is very hard to make
reliable numerical predictions for these baryonic lifetimes; yet
the overall qualitative pattern has been predicted:
$$\tau (\Xi _c ^0) < \tau (\Lambda _c) < \tau (\Xi _c ^+) \eqno(4a)$$
together with
$$\tau (\Lambda _c)< \tau (D^0) < \tau (D^+)
\; . \eqno(4b)$$
\noindent (iv) Pre-asymptotic corrections might be sizeable in the
lifetime ratios of beauty hadrons.
\noindent Reviews of these phenomenological descriptions
can be found in \cite{RUCKL,BRADLEE}.
It turned out, as discussed in more detail later on, that some
of the phenomenological descriptions anticipated the
correct results: it is PI that provides the main
engine behind
the $D^+$-$D^0$ lifetime ratio; $\Lambda _c$ is considerably
shorter-lived than $D^0$; the observed charm baryon
lifetimes do obey the hierarchy stated in eq.(4).
Nevertheless the phenomenological treatments had significant
shortcomings, both of a theoretical and of a phenomenological
nature: (i) No agreement had emerged in the literature about how
corrections in particular due to WA and WS scale with the heavy
quark mass $m_Q$. (ii) Accordingly no clear predictions could be
made on the lifetime ratios among beauty hadrons, namely whether
$\tau (B^+)$ and $\tau (B_d)$ differ by a few to several percent
only, or by 20 - 30 \%, or by even more! (iii) No unequivocal
prediction on
$\tau (D_s)$ or $\tau (B_s)$ had appeared. (iv) In the absence of a
systematic
treatment it is easy to overlook relevant contributions, and that is
actually what happened; or the absence of certain corrections had to
be postulated in an ad-hoc fashion. Thus there existed an
intellectual as well as practical need for a description based
on a systematic theoretical framework rather than a set of
phenomenological prescriptions.
\subsection{From Phenomenology to Theory}
In the last few years we have succeeded
in showing that the non-perturbative corrections to heavy-flavour
decays can be expressed
through a {\em systematic} expansion in {\em inverse} powers of
$m_Q$. A simple analogy with nuclear $\beta$ decay can
illustrate this point. There are two effects distinguishing the decays
of neutrons bound in a nucleus from the decay of free neutrons:
\noindent (a) nuclear binding effects;
\noindent (b) the impact of Pauli statistics correlating the electrons
surrounding
the nucleus with those emerging from $\beta$ decay.
\noindent The typical energies of the bound electrons
-- $\epsilon _{el}$ -- are
certainly small compared to $E_{release}$, the energy released in the
decay; let us assume -- although this is not true in reality -- that also
the nuclear binding energies $\epsilon _{nucl}$ were small compared
to $E_{release}$. In that case nuclear $\beta$ decays would proceed,
to a good approximation, like the decays of {\em free} neutrons;
corrections to this simple `spectator' picture could be computed via
an expansion in powers of $\epsilon _{nucl}/E_{release}$ and
$\epsilon _{el}/E_{release}$. In practice, however, the corrections for
nuclear $\beta$ decay are incorporated by explicitely using the wave
functions of the bound nucleons and electrons (obtained with the
help of some fairly massive computer codes).
There arise analogous corrections to the decay rate for a quark $Q$
inside a hadron $H_Q$:
\noindent (a) interactions of the decaying quark with other partons
in the
hadron; this includes WA of
$Q$ with the light
valence antiquark for mesons or WS
with the valence
diquark system for baryons; they correspond to K
capture of bound electrons by a heavy nucleus in the preceding
example.
\noindent (b) PI effects of the decay products with other
partons in the hadron; e.g.: $b\bar u \rightarrow c\bar u d \bar u$
or $c\bar d \rightarrow u \bar d s \bar d$.
They prolong the lifetimes of $D^+$ and $B^-$ mesons.
\noindent The difference to the example of nuclear $\beta$ decay
is quite
obvious: even in the limit $m_Q\rightarrow \infty$ a non-relativistic
bound-state treatment is inapplicable since the dynamical
degrees of
freedom of the heavy-flavour hadron $H_Q$ cannot fully be
described by a hadronic wavefunction. The most reliable approach is
then to evaluate weak decay rates of heavy-flavour hadrons through
an expansion in powers of $\mu _{had}/m_Q$ where
$\mu _{had}$ represents a hadronic scale $\leq$ 1 GeV.
The first few terms in this series should yield a good approximation
for beauty decays;
the situation for charm decays is a priori unclear
(and at present remains so a
posteriori as well); this will be discussed in detail later on.
\footnote{As already mentioned, top quarks decay weakly
before they can hadronize \cite{RAPALLO}.}
The vice of hadronization is then transformed into (almost) a virtue:
the weak decays of heavy-flavour hadrons constitute an intriguing
and novel laboratory for
studying strong dynamics through their
interplay with the weak forces -- and this is the secondary
motivation for studying them.
To be more specific: the heavy-flavour mass $m_Q$ provides an
expansion parameter that allows to deal with the
non-perturbative dynamics of QCD in a novel way.
The formalism to be employed
combines the $1/m_Q$ expansion with other elements derived from
QCD proper without having to invoke a `deus ex machina' --
in contrast to phenomenological descriptions.
There is one concept underlying, in one form or another, all efforts to
deal with hadronization, namely the notion of quark-hadron duality
(henceforth referred to as {\em duality} for short). In its broadest
formulation it can be stated as follows: sufficiently inclusive
transition rates between hadronic systems can be calculated in
terms of quarks and gluons. While the general validity of
this concept has
not been established in a rigorous fashion, new light has been shed
on its nature, validity and applicability by $1/m_Q$ expansions.
Lifetimes obviously represent the most inclusive quantity where
duality should be applicable.
Nonleptonic transitions are certainly more complex than semileptonic
ones; yet I will argue later that while there probably exists a
quantitative difference in the degree to which duality holds in
semileptonic and in nonleptonic decays, there is {\em no qualitative}
one. Deatiled measurements of
lifetimes are then theoretically important not only to translate data
on the semileptonic branching ratio into a determination of the
semileptonic
width, but also in their own right.
Dedicated and comprehensive studies of both charm and beauty
decays are called for. The KM parameters relevant for charm decays
-- $V(cs)$
and $V(cd)$ -- are well known through unitarity constraints of the
3x3 KM
matrix, in contrast to the situation in beauty decays, and
I consider it unlikely that charm decay studies can improve on that.
On the other hand those can be used to calibrate our theoretical
tools before applying them to beauty decays.
Before concluding this general introduction I want to point out a less
straightforward aspect of accurate lifetime measurements: decay rate
evolutions in proper time for neutral mesons will not follow a single
exponential function when particle-antiparticle oscillations occur.
For there exist two distinct mass eigenstates with
$\Delta m \equiv m_1-m_2\neq 0 \neq
\Delta \Gamma \equiv \Gamma _1-\Gamma _2 $. The quantity
$\Delta m$ generates a deviation of the form
$e^{-\Gamma t}\cos \Delta mt$
or $e^{-\Gamma t}\sin \Delta mt$;
$\Delta \Gamma \neq 0$
leads to the emergence of a second exponential.
The general expression reads as follows:
$$d\Gamma (B[D]\rightarrow f)/dt \; \propto
e^{-\Gamma t}\cdot G(t)$$
$$G(t) = a+be^{-\Delta \Gamma t}+
ce^{-1/2\Delta \Gamma t}\cos \Delta mt +
de^{-1/2\Delta \Gamma t}\sin \Delta mt\eqno(5a)$$
where
$$a=|A(f)|^2\left[ \frac{1}{2}
(1+|\frac{q}{p}\bar \rho (f)|^2)
+ Re[ \frac{q}{p}\bar \rho (f)]
\right] $$
$$b=|A(f)|^2\left[ \frac{1}{2}
[ 1+|\frac{q}{p}\bar \rho (f)|^2]
- Re[ \frac{q}{p}\bar \rho (f)]
\right] $$
$$c=|A(f)|^2\{
1-|\frac{q}{p}\bar \rho (f)|^2\}\; , \;
d= 2|A(f)|^2 Im[ \frac{q}{p}\bar \rho (f)] \; , \;
\bar \rho (f)= \frac{\bar A(f)}{A(f)}
\eqno(5b) $$
with $\bar A(f)$ and $A(f)$ denoting the amplitude for
$\bar B [\bar D] \rightarrow f$ and $B [D] \rightarrow f$, respectively.
\section{Preview of the Predictions on the
Lifetime Ratios for Beauty and Charm Hadrons}
In this section I summarize the numerical results and
sketch the main elements of the underlying theoretical
treatment in a way that can satisfy the casual reader. A more
in-depth discussion of the theoretical concepts and tools
will be presented in subsequent sections.
Expanding the width for the decay of a heavy-flavour hadron
$H_Q$ containing $Q$
into an inclusive final state $f$ through order $1/m_Q^3$
one obtains
\cite{BUV,BS,SV}
$$\Gamma (H_Q\rightarrow f)=\frac{G_F^2m_Q^5}{192\pi ^3}|KM|^2
\left[ c_3^f\matel{H_Q}{\bar QQ}{H_Q}_{norm}+
c_5^f\frac{
\matel{H_Q}{\bar Qi \sigma \cdot G Q}{H_Q}_{norm}}{m_Q^2}+ \right.
$$
$$\left. +\sum _i c_{6,i}^f\frac{\matel{H_Q}
{(\bar Q\Gamma _iq)(\bar q\Gamma _iQ)}{H_Q}_{norm}}
{m_Q^3} + {\cal O}(1/m_Q^4)\right] , \eqno(6)$$
where the dimensionless coefficients $c_i^f$ depend on the
parton level
characteristics of $f$ (such as the ratios of the final-state quark
masses
to $m_Q$); $KM$ denotes the appropriate combination of KM
parameters,
and $\sigma \cdot G = \sigma _{\mu \nu}G_{\mu \nu}$
with $G_{\mu \nu}$ being the gluonic field strength tensor. The last
term
in eq.(6)
implies also the summation over the four-fermion operators with
different light flavours $q$. The expectation values of the local
operators appearing on the right-hand side of eq.(6)
contain the
relativistic normalization of the state $|H_Q\rangle$:
$$\matel{H_Q}{O_i}{H_Q}_{norm} \equiv
\matel{H_Q}{O_i}{H_Q}/2M_{H_Q} \, .\eqno(7) $$
It is through the quantities
$\matel{H_Q}{O_i}{H_Q}$ that the dependence on the {\em decaying
hadron} $H_Q$, and
on
non-perturbative forces in general, enters, instead of through
wavefunctions as in nuclear $\beta$ decay.
Since these are matrix
elements
for on-shell hadrons $H_Q$, one sees that $\Gamma (H_Q\rightarrow f)$ is
indeed
expanded into a power series in $\mu _{had}/m_Q < 1$. For
$m_Q\rightarrow \infty$ the contribution from the lowest dimensional
operator obviously dominates; here it is the dimension three
operator $\bar QQ$. A heavy quark expansion yields:
$$\matel{H_Q}{\bar QQ}{H_Q}_{norm}=1+
{\cal O}(1/m_Q^2)\, , \eqno(8)$$
i.e. $\matel{H_Q}{\bar QQ}{H_Q}_{norm}=1$ for $m_Q\rightarrow \infty$,
reflecting the unit of heavy-flavour common to all
hadrons $H_Q$.
Eqs.(6,8) show that the leading nonperturbative corrections are
of order $1/m_Q^2$ rather than $1/m_Q$; therefore they can be
expected to be rather small in beauty decays. The expectation
values appearing in the $1/m_Q^2$ and
$1/m_Q^3$ contributions can be related to other observables
and their size thus be extracted in a reliable
manner. Applying the
master formula eq.(6) to lifetimes one arrives at the following
general results:
\noindent $\bullet$ The leading contribution to the total decay
width is given by the first term on the
right-hand-side of eq.(8) that is
{\em common} to all hadrons
of a given heavy-flavour quantum number.
For $m_Q\rightarrow \infty$ one has thus derived -- from QCD proper -- the
spectator picture attributing equal lifetimes to all hadrons of a
given heavy-flavour! This is not a surprising result -- after all
without hadronization there is of course only a unique lifetime --,
but it is gratifying nevertheless.
\noindent $\bullet$ Lifetime differences first arise at order
$1/m_Q^2$ and are controlled by the expectation
values of two dimension five operators, to be discussed later. These
terms, which had
been overlooked in the original phenomenological analyses, generate
a lifetime difference between heavy-flavour {\em baryons} on one
side and {\em mesons} on the other. Yet apart from
small isospin or $SU(3)_{fl}$ breaking they shift the
meson widths by the same amount and thus do {\em not}
lead to differences among the meson lifetimes.
\noindent $\bullet$ Those emerge at
order $1/m_Q^3$ and are expressed through the expectation values
of four-fermion operators which are proportional to
$f_M^2$ with $f_M$ denoting the decay constant for the meson
$M$. Contributions from what is referred to as
WA and PI in the
original phenomenological descriptions
(see the discussion in Sect.1) are systematically and
consistently included. Further
contributions to the baryon-meson lifetime difference also
arise at this level due to WS.
\noindent $\bullet$ Since the transitions $b\rightarrow c l \nu$ or
$c\rightarrow s l \nu$ are described by an isosinglet operator one can
invoke the isospin invariance of the strong interactions to
deduce for the semileptonic widths
$$ \Gamma _{SL}(B^-)= \Gamma _{SL}(B_d) \eqno(9a)$$
$$ \Gamma _{SL}(D^+)= \Gamma _{SL}(D^0) \eqno(9b)$$
and therefore
$$\frac{\tau (B^-)}{\tau (B_d)}=
\frac{BR_{SL}(B^-)}{BR_{SL}(B_d)} \eqno(10a)$$
$$\frac{\tau (D^+)}{\tau (D^0)}=
\frac{BR_{SL}(D^+)}{BR_{SL}(D^0)} \eqno(10b)$$
up to small corrections due to the KM [Cabibbo]
suppressed transition $b\rightarrow u l \nu$
[$c\rightarrow d l \nu$] which changes isospin by half a unit.
The spectator picture goes well beyond eqs.(9): it
assigns the same semileptonic width to all hadrons of
a given heavy flavour. Yet such a property cannot be
deduced on {\em general} grounds: for one had to rely
on $SU(3)_{Fl}$ symmetry to relate
$\Gamma _{SL}(D_s)$ to $\Gamma _{SL}(D^0)$ or
$\Gamma _{SL}(B_s)$ to $\Gamma _{SL}(B_d)$ and
no symmetry can be invoked to relate the semileptonic
widths of mesons and baryons. There is actually a WA
process that generates semileptonic decays on the
Cabibbo-allowed level for $D_s$ [and also for
$B_c$], but not for $D^0$ and $D^+$ nor for
$B_d$, $B^-$ and $B_s$ mesons: the hadrons are produced
by gluon emission off the $\bar s$ [or the $\bar c$] line.
Yet since the relative weight of WA is significantly reduced
in meson decays, one does not expect this mechanism to
change $\Gamma _{SL}(D_s)$ significantly relative to
$\Gamma _{SL}(D^0)$. Actually there are corrections
to the semileptonic
widths arising already in order $1/m_Q^2$. On rather
general grounds one predicts the expectation values
$\matel{P_Q}{\bar QQ}{P_Q}$ and
$\matel{P_Q}{\bar Qi\sigma \cdot G Q}{P_Q}$ to be largely
independant of the flavour of the light antiquark in the
meson and therefore
$$\Gamma _{SL}(D_s)\simeq \Gamma _{SL}(D^0) \eqno(11a)$$
$$\Gamma _{SL}(B_s)\simeq \Gamma _{SL}(B_d) \eqno(11b)$$
like in the naive spectator picture, but for
non-trivial reasons. On the other hand,
as explained later, the values of the expectation values
for these operators are different when taken between
baryon states and one expects
$$\Gamma _{SL}(\Lambda _Q) > \Gamma _{SL}(P_Q)
\eqno(11c)$$
The remarks above can be summarized as follows:
$$\Gamma (\Lambda _Q) = \Gamma ([Q\bar q]^0) =
\Gamma ([Q\bar q]^{\pm}) + {\cal O}(1/m_Q^2) \eqno(12a)$$
$$\Gamma (\Lambda _Q) > \Gamma ([Q\bar q]^0) =
\Gamma ([Q\bar q]^{\pm}) + {\cal O}(1/m_Q^3) \eqno(12b)$$
$$\Gamma (\Lambda _Q) > \Gamma ([Q\bar q]^0) >
\Gamma ([Q\bar q]^{\pm}) + {\cal O}(1/m_Q^4) \eqno(12c)$$
Quantitative predictions for the lifetime ratios of beauty
hadrons through order $1/m_b^3$ are given in Table \ref{TABLE1}
together with present data.
\begin{table}
\begin{tabular} {|l|l|l|}
\hline
Observable &QCD ($1/m_b$ expansion) &Data \\
\hline
\hline
$\tau (B^-)/\tau (B_d)$ & $1+
0.05(f_B/200\, \,\mbox{MeV} )^2
[1\pm {\cal O}(10\%)]>1$ &$1.04 \pm 0.05$ \\
&(mainly due to {\em destructive} interference) & \\
\hline
$\bar \tau (B_s)/\tau (B_d)$ &$1\pm {\cal O}(0.01)$
& $ 0.98\pm 0.08$ \\
\hline
$\tau (\Lambda _b)/\tau (B_d)$&$\sim 0.9\; ^*$ & $0.76\pm 0.06$
\\
\hline
\end{tabular}
\centering
\caption{QCD Predictions for Beauty Lifetimes}
\label{TABLE1}
\end{table}
The expectations \cite{MIRAGE,MARBELLA,DS}
for the lifetimes of charm hadrons are juxtaposed to the data in
Table \ref{TABLE2}.
The overall agreement is rather good given the theoretical and
experimental uncertainties. It has to be kept in mind that the
{\em numerical} predictions on {\em baryon} lifetimes involve
quark model results for the size of the relevant expectation
values. This is indicated in Tables 1 and 2 by the asterisk.
There is however one obvious discrepancy: the observed
$\Lambda _b$ lifetimes is significantly shorter than predicted.
While -- as indicated just above -- predictions on {\em baryon}
lifetimes have to be taken with a grain of salt, one cannot change
the prediction on
$\tau (\Lambda _b)/\tau (B_d)$ with
complete theoretical impunity; this will be explained later on.
\begin{table}
\begin{tabular} {|l|l|l|}
\hline
Observable &QCD ($1/m_c$ expansion) &Data \\
\hline
\hline
$\tau (D^+)/\tau (D^0)$ & $\sim 2 \; \; \; $
[for $f_D \simeq 200$ MeV] &$2.547 \pm 0.043$ \\
&(mainly due to {\em destructive} interference) & \\
\hline
$\tau (D_s)/\tau (D^0)$ &$1\pm$ few $\times 0.01$
& $ 1.125\pm 0.042$ \\
\hline
$\tau (\Lambda _c)/\tau (D^0)$&$\sim 0.5\; ^*$ &
$0.51\pm 0.05$\\
\hline
$\tau (\Xi ^+ _c)/\tau (\Lambda _c)$&$\sim 1.3\; ^*$ &
$1.75\pm 0.36$\\
\hline
$\tau (\Xi ^+ _c)/\tau (\Xi ^0 _c)$&$\sim 2.8\; ^*$ &
$3.57\pm 0.91$\\
\hline
$\tau (\Xi ^+ _c)/\tau (\Omega _c)$&$\sim 4\; ^* $&
$3.9 \pm 1.7$\\
\hline
\end{tabular}
\centering
\caption{QCD Predictions for Charm Lifetimes}
\label{TABLE2}
\end{table}
These numerical predictions bear out the general expectations
expressed above. This
will be discussed below in more detail.
For proper perspective on the role played by the heavy quark
expansion some comments should be made at this point:
\noindent (i) As stated before, the early phenomenological
treatments had already identified the relevant mechanisms by
which lifetime differences arise, namely PI, WA and WS.
Furthermore the authors of ref.\cite{PI} had argued
-- correctly -- that PI constitutes the dominant effect. Yet it
is the heavy quark expansion that provides a firm
theoretical underpinning
to these expectations. In particular it clarifies -- both
conceptually and quantitatively -- the role of WA and
how its weight scales with $1/m_Q$.
\noindent (ii) Contributions of order $1/m_Q$ would
dominate all other effects -- if they were present! The heavy
quark expansion shows unequivocally that they are absent in
total rates; this has important ramifications, to be pointed out
later.
\noindent (iii) Corrections of order $1/m_Q^2$ differentiate
between the decays of mesons and of baryons. They had been
overlooked before. The heavy quark expansion has identified them
and basically determined their size.
\section{Methodology of the Heavy Quark Expansion for Fully
Integrated Rates}
The weak decay of the heavy quark $Q$ inside the heavy-flavour
hadron $H_Q$
proceeds within a cloud of light degrees of freedom (quarks,
antiquarks
and gluons) with which $Q$ and its decay products can interact
strongly. It
is the challenge for theorists to treat these initial and final state
hadronization effects. Among the existing four Post-Voodoo
theoretical technologies -- QCD Sum Rules,
Lattice QCD, Heavy Quark Effective Theory (HQET) and Heavy Quark
Expansions -- only the last one deals with inclusive decays. On
the other hand it benefits, as
we will see, from the results of the other three
technologies, since those can determine the size of some of
the relevant expectation values.\footnote{It should be noted that
contrary
to frequent claims in the literature HQET -- as it is usually
and properly defined -- per se does {\em not} allow to treat
lifetimes or even the total semileptonic width:
for those observables strongly depend on $m_Q$ whereas it
is the special feature of HQET that $m_Q$ is removed from
its Lagrangian.}
In analogy to the treatment of
$e^+e^-\rightarrow hadrons$ one describes the transition rate into an
inclusive final state $f$ through the imaginary part of a
forward scattering operator evaluated to second order in the weak
interactions \cite{BUV,BS,SV}:
$$\hat T(Q\rightarrow f\rightarrow Q)=
i \, Im\, \int d^4x\{ {\cal L}_W(x){\cal L}_W^{\dagger}(0)\}
_T\eqno(13)$$
where $\{ .\} _T$ denotes the time ordered product and
${\cal L}_W$ the relevant effective weak Lagrangian expressed on
the
parton level. If the energy released in the decay is sufficiently large
one can express the {\em non-local} operator product in eq.(13) as an
infinite sum of {\em local} operators $O_i$ of increasing dimension
with
coefficients $\tilde c_i$
containing higher and higher powers of
$1/m_Q$.\footnote{It should be kept in mind,
though, that
it is primarily the {\em energy release} rather than $m_Q$ that controls the
expansion.} The width for $H_Q\rightarrow f$ is then obtained by
taking the
expectation value of $\hat T$ between the state $H_Q$:
$$\matel{H_Q}{\hat T (Q\rightarrow f\rightarrow Q)}{H_Q} \propto
\Gamma (H_Q\rightarrow f) = G_F^2 |KM|^2
\sum _i \tilde c_i^{(f)} \matel{H_Q}{O_i}{H_Q}
\eqno(14)$$
This master formula holds for a host of different inclusive
heavy-flavour decays: semileptonic, nonleptonic and radiative
transitions,
KM favoured or suppressed etc. For semileptonic and nonleptonic
decays treated through order $1/m_Q^3$ it takes the form already
given in eq. (6):
$$\Gamma (H_Q\rightarrow f)=\frac{G_F^2m_Q^5}{192\pi ^3}|KM|^2
\left[ c_3^f\matel{H_Q}{\bar QQ}{H_Q}_{norm}+
c_5^f\frac{
\matel{H_Q}{\bar Qi\sigma \cdot G Q}{H_Q}_{norm}}{m_Q^2}+ \right.
$$
$$\left. +\sum _i c_{6,i}^f\frac{\matel{H_Q}
{(\bar Q\Gamma _iq)(\bar q\Gamma _iQ)}{H_Q}_{norm}}
{m_Q^3} + {\cal O}(1/m_Q^4)\right] $$
Integrating out the two internal loops in Figs.2 and 3
yields the operators $\bar QQ$ and $\bar Qi\sigma \cdot G Q$,
respectively; the
black boxes represent ${\cal L_W}$.
Likewise the diagrams in Figs.4 generate four-quark
operators; Fig.4a and 4b differ in how the light quark flavours
are connected inside the hadron $H_Q$:
cutting a quark line $q$ in Fig.2 and connecting it to the $\bar q$
constituent of the $H_Q$ meson, as shown in Fig.4a, one has a WA
transition operator; cutting instead the
$\bar q$ line and connecting it to the $H_Q$ constituents, see
Fig.4b, leads to the four-fermion operator describing PI.
Inspection of these diagrams suggests a rule-of-thumb
that is borne out by explicit computations: the dimensionless
coefficients $c^f_d$ are smaller for the two-loop diagrams of
Figs.2 and 3 than for the one-loop diagrams of Figs.4.
Integrating out the internal loops of Fig.3 actually generates
not only the dimension-five chromomagnetic operator, but also
dimension-six quark-gluon operators,
namely $\bar Q (iD_{\mu}G_{\mu \nu})\gamma _{\nu}Q$ and
$\bar Q\sigma _{\mu \nu}G_{\mu \rho}\gamma _{\rho}
iD_{\nu}Q$. Using the equations of motion those can be
reduced to four-quark operators. Yet their coefficients are
suppressed by $\alpha _S/\pi$ relative to the coefficient
coming from Figs.4; therefore one can discard their contributions
at the present level of accuracy.
One important general observation should be made at this
point. The two operators $\bar QQ$ and $\bar Qi\sigma \cdot G Q$ obviously
do not contain any light quark or antiquark fields; those enter
first on the dimension-six level. Their expectation values are,
as discussed below, practically insensitive to the light degrees
of freedom up to terms of order $1/m_Q^3$. The
inclusive semileptonic and
non-leptonic widths through order $1/m_Q^2$ are thus controled
by the {\em decay} of a single quark $Q$,
yet occurring in a non-trivial background field; the
{\em spectator} contribution universal to the widths for all
hadrons $H_Q$ of a given heavy flavour constitutes its
asymptotic part:
$$\Gamma (H_Q) \simeq \Gamma _{decay}(H_Q) +
{\cal O}(1/m_Q^3)\; , \eqno(15a)$$
$$\Gamma _{decay}(H_Q) \simeq \Gamma _{spect}(Q) +
{\cal O}(1/m_Q^2)\; . \eqno(15b)$$
However $\Gamma _{decay}$ possesses a different value
for mesons and baryons, as discussed below.
The expansion introduced above will be of practical use only
if its first few terms provide a good approximation and if
one can determine the corresponding expectation values
in a reliable manner. The latter can be achieved in two
complementary ways: One relates the matrix element in
question to other observables
through a heavy quark expansion; or one harnesses
other second-generation theoretical
technologies, namely QCD sum rules, QCD
simulations on the lattice and Heavy Quark Effective Theory,
to calculate these quantities. The results obtained so far
are listed below:
(a) Using the equations of motion one finds for the leading operator
$\bar QQ$:
$$\bar QQ= \bar Q\gamma _0Q -
\frac{\bar Q[(i\vec D)^2-(i/2)\sigma \cdot G]Q}{2m_Q^2}+
g_S^2\frac{\bar Q\gamma _0t^iQ
\sum _q \bar q\gamma _0 t^iq}{4m_Q^3} +
{\cal O}(1/m_Q^4)\eqno(16)$$
with the sum in the last term running over the light quarks $q$; the
$t^i$ denote the
colour $SU(3)$ generators. I have ignored total derivatives in this
expansion since they do not contribute to the $H_Q$ expectation
values. There emerge two dimension-five operators, namely
$\bar Q (i\vec D)^2 Q$ and $\bar Q \sigma \cdot G Q$ with $\vec D$
denoting the covariant derivative. The first one represents the
square of the spatial momentum of the heavy quark $Q$ moving in
the soft gluon background and thus describes its kinetic
energy\footnote{Since it is not a Lorentz scalar, it cannot
appear in
eq.(6).}. The second one constitutes the chromomagnetic operator
that already appeared in eq.(6). The
relevant dimension-six operators can be expressed as four-fermion
operators. Since $\bar Q\gamma _0 Q$ constitutes the Noether
current for the heavy-flavour quantum number one has
$\matel{H_Q}{\bar Q\gamma _0Q}{H_Q}_{norm}=1$ leading to eq.(8).
(b) The first non-perturbative corrections arise at
order $1/m_Q^2$, and they can enter in two ways, namely
through the chromomagnetic operator in the OPE of $\hat T$ or
through the $1/m_Q$ expansion of
the $H_Q$
expectation values of the local operator $\bar QQ$.
(c) For the chromomagnetic operator one finds
$$\matel{P_Q}{\bar Qi\sigma \cdot G Q}{P_Q}_{norm} \simeq
\frac{3}{2} (M_{V_Q}^2-M_{P_Q}^2) \eqno(17a)$$
where $P_Q$ and $V_Q$ denote the pseudoscalar and vector mesons,
respectively. For $\Lambda _Q$ and $\Xi _Q$ baryons one has
instead
$$\matel{\Lambda _Q}{\bar Qi\sigma \cdot G Q}{\Lambda _Q}
\simeq 0 \simeq \matel{\Xi _Q}{\bar Qi\sigma \cdot G Q}{\Xi _Q}\eqno(17b)$$
since the light diquark system inside $\Lambda _Q$ and $\Xi _Q$
carries no spin. These baryons thus represent (though only through
order $1/m_Q^2$) a simpler system than the mesons where the light
antiquark of course carries spin. For $\Omega _Q$ baryons on the
other hand the situation is different: for the light di-quarks
$ss$ form a spin-one configuration. The $\Omega _Q$
expectation value is then given by the spin-splitting in the
baryon masses:
$$\matel{\Omega _Q}{\bar Qi\sigma \cdot G Q}{\Omega _Q}_{norm}
\simeq \frac {4}{3} [M^2(\Omega _Q^{(3/2)}) -
M^2(\Omega _Q)]\eqno(17c)$$
(d) We thus have
$$\matel{\Lambda _Q}{\bar QQ}{\Lambda _Q}_{norm}=1 -
\frac{\langle (\vec p_Q)^2\rangle _{\Lambda _Q}}{2m_Q^2}+
{\cal O}(1/m_Q^3)\eqno(18a)$$
$$\matel{P_Q}{\bar QQ}{P_Q}_{norm}=1 -
\frac{\langle (\vec p_Q)^2\rangle _{P_Q}}{2m_Q^2}+
\frac{3}{8} \frac{M_{V_Q}^2-M_{P_Q}^2}{m_Q^2}+
{\cal O}(1/m_Q^3)\eqno(18b)$$
with the notation $\langle (\vec p_Q)^2\rangle _{H_Q}
\equiv \matel {H_Q}{\bar Q (i\vec D)^2 Q}{H_Q}_{norm}$.
The reason for the appearance of the kinetic energy term
in eqs. (18) is quite transparent: The first two terms on the r.h.s. of
eqs.(18) represent the mean value of the factor
$\sqrt{1-\vec v^2}$ reflecting the time dilation slowing down the
decay of the quark $Q$ moving inside $H_Q$.
(e) The value of $\langle (\vec p_Q)^2\rangle _{H_Q}$
has not been determined yet in an accurate way
(details will be given later); we know, though, it
has to obey the inequality \cite{VOLOSHIN,OPTICAL}
$$\langle (\vec p_Q)^2\rangle _{H_Q}\geq \frac{1}{2}
\matel{H_Q}{\bar Qi\sigma \cdot G Q}{H_Q}_{norm}
\equiv \aver{\mu _G^2}_{P_Q}\eqno(19)$$
which -- because of eq.(17) -- is useful only for $H_Q=P_Q$.
\footnote{One should note that this inequality has now been derived
in a fully field-theoretical treatment of QCD, rather than a merely
quantum-mechanical one. Its proper interpretation
will be discussed later.}
On the other hand, the difference in the kinetic energy of Q inside
baryons and mesons can be related to the masses of charm and
beauty hadrons \cite{BUVPREPRINT}:
$$ \langle (\vec p_Q)^2\rangle _{\Lambda _Q}-
\langle (\vec p_Q)^2\rangle _{P_Q} \simeq
\frac{2m_bm_c}{m_b-m_c}\cdot
\{ [\langle M_B\rangle -M_{\Lambda _b}]-
[\langle M_D\rangle -M_{\Lambda _c}] \} \eqno(20)$$
with $\langle M_{B,D}\rangle$ denoting the `spin averaged' meson
masses:
$$ \langle M_B\rangle \equiv \frac{1}{4}(M_B+3M_{B^*})
\eqno(21)$$
and likewise for $\langle M_D\rangle$. In deriving eq.(20) it was
implicitely assumed that
the $c$ quark can be treated also as heavy; in that case
$\langle (\vec p_c)^2\rangle _{H_c} \simeq
\langle (\vec p_b)^2\rangle _{H_b}$ holds.
(f) Eqs.(17-18,20) show that the two
{\em dimension-five} operators do
produce differences in $P_Q$ vs. $\Lambda _Q/\Xi _Q$ vs.
$\Omega _Q$ lifetimes
(and semileptonic widths) of order
$1/m_Q^2$; i.e., $\Gamma _{decay}(H_Q)$, as defined in
eqs.(15), depends on whether $H_Q$ represents a (pseudoscalar)
meson or a baryon. In the latter case it is
sensitive to whether the light diquark system carries spin
zero ($\Lambda _Q$, $\Xi _Q$) or
one ($\Omega _Q$). The origin of such a difference is quite transparent.
(i) The heavy quark $Q$ can be expected to
possess a somewhat different kinetic
energy inside a meson or a baryon; it can also be affected by spin-spin
interactions with the light di-quarks.
This difference in the heavy quark motion means that
Lorentz time dilation prolongs the lifetime of the quark $Q$
to different degrees in baryons than in mesons.
(ii) While there exists a spin interaction
between $Q$ and the antiquark $\bar q$ in the meson or the
$ss$ system in the $\Omega _Q$ baryon,
there is no such coupling inside $\Lambda _Q$
or $\Xi _Q$.
\noindent Yet these effects do not generate a significant difference
among
{\em meson} lifetimes since their expectation values satisfy isospin
and $SU(3)_{fl}$ symmetry to a good degree of accuracy.
(g) Differences in meson lifetimes are generated at order
$1/m_Q^3$ by {\em dimension-six} four-quark operators
describing PI as well as WA; the weight of the latter is, as
explained in more detail in the next section, greatly reduced.
The expectation values of these operators
look very similar to the one controlling $B^0-\bar B^0$ oscillations:
$$\matel{H_Q(p)}{(\bar Q_L\gamma _{\mu}q_L)
(\bar q_L\gamma _{\nu}Q_L}{H_Q(p)}_{norm}\simeq
\frac{1}{4m_{H_Q}}f^2_{H_Q}p_{\mu}p_{\nu}\eqno(22)$$
with $f_{H_Q}$ denoting the decay constant for the meson $H_Q$;
the so-called bag factor has been set to unity, i.e.
{\em factorization}
has been assumed.
(h) The situation becomes much more complex for $\Lambda _Q$ and
baryon decays in general. To order $1/m_Q^3$ baryons lose the
relative simplicity mentioned above: there are several different
ways in which the valence quarks of the baryon can be contracted
with the quark fields in the four-quark operators; furthermore
WS is {\em not} helicity suppressed and thus can make a sizeable
contribution to lifetime differences; also the PI effects can
now be constructive
as well as destructive. Finally one cannot take
recourse to factorisation as a limiting case.
Thus there emerge three types of
numerically significant mechanisms at this order in baryon
decays -- in contrast to meson decays where there is a
{\em single} dominant
source for lifetime differences -- and their strength cannot be
expressed in terms of
a single observable like $f_{H_Q}$. At present we do not know
how to determine the relevant matrix elements in a
model-independant way. The best available guidance and
inspiration is to be gleaned from quark model calculations with
their inherent uncertainties. This analysis had already been
undertaken
in the framework of phenomenological models
\cite{BARYONS1,BARYONS2,BARYONS3} with the following
qualitative results:
\noindent -- WS contributes to $\Lambda _Q$ and $\Xi _Q^0$
decays;
\noindent -- a {\em destructive} interference affects
$\Lambda _Q$, $\Xi _Q$ and $\Omega _b$ transitions.
\noindent -- a {\em constructive} interference enhances
$\Xi _c$ and $\Omega _c$ decays.
\noindent One thing should be obvious already at this point:
with terms of different signs and somewhat uncertain size
contributing to differences among baryon lifetimes one has to
take even semi-quantitative predictions with
a grain of salt!
\noindent The probability amplitude for WS or interference to
occur is expressed in quark models by the wavefunction for
$Q$ and one of the light quarks at zero spatial separation. This
quantity is then related to the meson wavefunction through the
hyperfine splitting in the baryon and meson masses
\cite{BARYONS3,MARBELLA}:
$$\frac{|\psi _{\Lambda _Q}^{(Qq)}(0)|}{|\psi _{P_Q}(0)|}\simeq
\frac{2m_q^*(M_{\Sigma _Q}-M_{\Lambda _Q})}
{(M_{P_Q}-m_q^*)(M_{V_Q}-M_{P_Q})}\eqno(23)$$
with $m_q^*$ denoting the {\em phenomenological constituent}
mass
of the light quark $q$ to be employed in these models. For the
baryonic expectation values one then obtains
$$\matel{\Lambda _Q}{\bar Q\Gamma _{\mu}Q
\bar q \Gamma _{\mu} q}{\Lambda _Q}_{norm} \sim
\frac{1}{4\aver{\mu ^2_G}_{P_Q}}(M_{\Sigma _Q}-M_{\Lambda _Q})
m^*_q F^2_{P_Q}M_{P_Q}\eqno(24)$$
and likewise for $\Xi _Q$.
For the $\Omega _Q$ one has instead
$$\matel{\Omega _Q}{\bar Q\Gamma _{\mu}Q
\bar q \Gamma _{\mu} q}{\Omega _Q}_{norm} \sim
\frac{5}{6\aver{\mu ^2_G}_{P_Q}}(M_{\Sigma _Q}-M_{\Lambda _Q})
m^*_q F^2_{P_Q}M_{P_Q} \; . \eqno(25)$$
reflecting the different spin substructure which already
had caused the difference in eqs.(17a) vs. (17b).
For simplicity one has assumed in these expressions that
the hyperfine splitting in the baryon masses are universal.
To summarize this discussion:
\noindent $\bullet$ The non-perturbative corrections
to total widths through order $1/m_Q^3$ are expressed in
terms of three non-trivial (types of) matrix elements:
$\aver{\mu _G^2}_{H_Q}$, $\aver{(\vec p_Q)^2}_{H_Q}$ and
$\matel{H_Q}{(\bar Q\Gamma _i q)(\bar q\Gamma _i Q)}{H_Q}$.
\noindent $\bullet$ The size of $\aver{\mu _G^2}_{H_Q}$ is
well-known for the mesons and $\Lambda _Q$ and $\Xi _Q$
baryons; reasonable estimates can be obtained for
$\Omega _Q$ baryons.
\noindent $\bullet$ A lower bound exists for the mesonic
expectation value $\aver{(\vec p_Q)^2}_{P_Q}$, but not
for the baryonic one. Various arguments strongly suggest
that $\aver{(\vec p_Q)^2}_{H_Q}$ is close to its lower
bound. Furthermore the quantity
$\aver{(\vec p_Q)^2}_{\Lambda _Q}-\aver{(\vec p_Q)^2}_{P_Q}$,
which is highly relevant for the difference in baryon
vs. meson lifetimes (and semileptonic widths) can be extracted
from mass measurements.
\noindent $\bullet$ The expectation value of the
four-quark operators between {\em mesons} can be expressed
in terms of a single quantity, namely the decay
constant $f_{P_Q}$.
\noindent $\bullet$ However the {\em baryonic} expectation
values of the four-quark operators constitute a veritable
Pandora's box, which at present is beyond theoretical
control. On the other hand measurements of inclusive
weak decay rates and the resulting
understanding of the role played by WS and PI present us
with a novel probe of the internal structure of the
heavy-flavour baryons.
\section{Comments on the Underlying Concepts}
Before explaining these
predictions in more detail and commenting on the comparison with
the data in the next subsection, I want to add four general remarks
and elaborate on them. While I hope they will elucidate the
underlying concepts, they are not truly essential for following the
subsequent discussion and the more casual reader can ignore them.
(A) {\em On the validity of the $1/m_Q$ expansion in the presence of
gluon emission:} The WA contribution to the decay width of
pseudoscalar mesons to lowest order in the strong coupling has
been sketched in eq.(2).
The diagram in Fig.1 contains gluon bremsstrahlung off the
initial antiquark line in the $W$-exchange reaction;
evaluating this diagram
explicitely one finds, as stated before:
$$\Gamma ^{(1)}_{W-X}\propto
(\alpha_s /\pi ) G_F^2(f_{H_Q}/\aver{E_{\bar q}})^2m_Q^5
\eqno(26)$$
with
$\aver {E_{\bar q}}$ denoting the average energy of the
initial antiquark $\bar q$.
Since $\Gamma ^{(1)}_{W-X} \propto 1/\aver{E_{\bar q}}^2$,
it depends very sensitively on the
low-energy quantity $ \aver{E_{\bar q}}$, and
its magnitude is thus quite
uncertain. The presence of a $1/\aver{E_{\bar q}}^2$
(or $1/m_q^2$) term would also mean that the nonperturbative
corrections to even inclusive decay widths of heavy-flavour
hadrons are not `infrared safe' and cannot be treated consistently through an
expansion in inverse powers of the heavy quark mass
$m_Q$ only.
\noindent Fortunately the contribution stated in eq.(26)
turns out to be
spurious for {\em inclusive} transitions.
This can best be seen by
studying the imaginary part of the forward scattering
amplitude $Q\bar q \rightarrow f \rightarrow Q\bar q$ as shown in Fig.5.
There are actually three poles in this amplitude indicated
by the broken lines
which
represent different final states: one with an on-shell
gluon and the other two with an off-shell gluon
materializing as a $q\bar q$ pair and involving interference
with the spectator decay amplitude. These processes are
all of the same order in the strong coupling and therefore
have to be summed over for an {\em inclusive} decay.
Analysing carefully the analyticity properties of the sum of these
forward
scattering amplitudes one finds \cite{MIRAGE,WA}
that it remains finite in the limit
of $\aver{E_{\bar q}} \rightarrow 0$, i.e. the amplitude for the
{\em inclusive} width does {\em not} contain terms of order
$1/\aver{E_{\bar q}}^2$ or even $1/\aver{E_{\bar q}}$! The
contribution of WA can then be described in terms of
the expectation value of a {\em local} four-fermion
operator although the final state is dominated by low-mass
hadronic systems.
\noindent To summarize: the quantity $f_{H_Q}$ is calibrated by the
large mass $m_Q$ rather
than by $\aver{E_{\bar q}}$;
WA is then only
moderately significant even in charm decays.
The $1/m_Q$ scaling thus persists to hold even in the presence of
radiative corrections -- but only for {\em inclusive} transitions!
This caveat is not of purely academic interest. For
it provides a nice toy model illustrating the {\em qualitative}
difference between inclusive and exclusive transitions. The latter are
represented here by the three separate cuts in Fig.5.
For charm decays they correspond to the reactions
$[c\bar u] \rightarrow s \bar d g$ and $[c\bar u] \rightarrow s \bar d u\bar u$.
The two individual transition rates quite sensitively depend on
a low energy scale - $\aver{E_{\bar q}}$ -
which considerably enhances the rate for the first
transition while reducing it for the second one.
Yet the dependance on $1/\aver{E_{\bar q}}$ disappears
from their sum!
(B) {\em On the fate of the corrections of order
$1/m_Q$:}
The most important element of eq.(2) is -- the one that
is
{\em missing}! Namely there is no term of order $1/m_Q$ in the
total decay width whereas such a correction definitely exists for the
mass formula: $M_{H_Q}=m_Q(1+\bar \Lambda /m_Q+{\cal
O}(1/m_Q^2))$ (and likewise for differential decay distributions).
Hadronization in the initial state does generate corrections
of order $1/m_Q$ to the total width, as does hadronization in the
final state. Yet
local colour symmetry enforces that they
cancel against each other!
\footnote{This can be nicely illustrated in quantum mechanical toy
models: The total rate for the transition $Q\rightarrow q$ depends on the
{\em local} properties of the potential, i.e. on the potential around
$Q$ in a neighourhood of size
$1/m_Q$ only; the nature of the resulting spectrum for $q$
-- whether it is discrete or continuous, etc. -- depends
of course on the long range properties of the potential,
namely whether it is confining or not.}
This can be understood in
another more compact (though less intuitive) way as well: with the
leading operator $\bar QQ$ carrying dimension three only
dimension-four operators can generate
$1/m_Q$ corrections; yet there is no independent dimension-four
operator \cite{CHAY,BUV}
{\em once the equation of motion is imposed} -- unless one abandons
local
colour symmetry thus making the operators $\bar Qi\gamma \cdot
\partial Q$
and $\bar Qi\gamma \cdot B Q$ independent of each other
($B_{\mu}$ denotes the gluon field)! The leading
non-perturbative corrections to fully integrated decay widths are
then
of order $1/m_Q^2$ and their size is controlled by two
dimension-five operators, namely the chromomagnetic and the
kinetic
energy operators, as discussed above. Their contributions
amount to no more than 10 percent
for $B$ mesons -- $(\mu _{had}/m_b)^2$ $\simeq
{\cal O}\left( (1\, GeV/m_b)^2\right)
\sim {\cal O}(\% )$ -- as borne out by the detailed analysis
previewed in Table 1.
(C) {\em Which mass is it?}
For a $1/m_Q$ expansion it is of course important
to understand which kind of quark mass is to be employed there,
in particular since
for confined quarks there exists no a priori natural choice.
It had been claimed that the pole mass can and therefore
should conveniently be used. Yet such claims
turn out to be fallacious \cite{POLEMASS,BRAUN}: QCD, like QED, is
not
Borel summable; in the high order terms
of the perturbative series there arise instabilities
which are customarily referred to as
(infrared) renormalons representing poles in the Borel plane; they
lead to an {\em additive} mass renormalization generating an
{\em irreducible uncertainty}
of order $\bar \Lambda$ in the size of the pole mass:
$m_Q^{pole}=m_Q^{(0)}(1+c_1\alpha_s + c_2\alpha_s ^2 +...+ c_N\alpha_s ^N) +
{\cal O}(\bar \Lambda )=m_Q^{(N)}(1+{\cal O}(\bar \Lambda/
m_Q^{(N)}))$. While this effect
can safely be ignored in a purely perturbative treatment, it negates
the
inclusion of non-perturbative corrections $\sim {\cal O}(1/m_Q^2)$,
since those are
then parametrically smaller than the uncertainty
$\sim {\cal O}(1/m_Q)$ in the definition of
the
pole mass. This problem can be taken care of through
Wilson's prescriptions for the operator product expansion:
$$\Gamma (H_Q\rightarrow f)=\sum _i c_i^{(f)}(\mu )
\matel{H_Q}{O_i}{H_Q}_{(\mu )}\eqno(27)$$
where a momentum scale $\mu$ has been introduced to allow a
consistent
separation of contributions from Long Distance and Short Distance
dynamics -- $LD > \mu ^{-1} > SD$ -- with the latter contained in
the coefficients $c_i^{(f)}$ and the former lumped into the matrix
elements. The quantity $\mu$ obviously represents an auxiliary
variable which drops out from the observable, in this case the decay
width. In the limit
$\mu \rightarrow 0$ infrared renormalons emerge in the coefficients; they
cancel
against ultraviolet renormalons in the matrix elements.
Yet that does
{\em not} mean that these infrared renormalons are irrelevant and
that
one can conveniently set $\mu =0$! For to incorporate both
perturbative
as well as non-perturbative corrections one has to steer a careful
course
between `Scylla' and `Charybdis': while one wants to pick
$\mu \ll m_Q$ so as to make a heavy quark expansion applicable,
one
also has to choose $\mu _{had}\ll \mu$ s.t. $\alpha _S(\mu ) \ll 1$;
for otherwise the {\em perturbative} corrections become
uncontrollable. Wilson's
OPE allows to incorporate both perturbative and non-perturbative
corrections, and {\em this underlies also a consistent formulation of
HQET};
the scale $\mu$ provides an infrared cut-off that automatically
freezes out infrared renormalons. For the asymptotic difference
between the hadron and the quark mass one then has to write
$\bar \Lambda (\mu ) \equiv
(M_{H_Q}-m_Q(\mu ))_{m_Q\rightarrow \infty}$.
This nice feature does not come for free,
of course: for one has to use a `running' mass $m_Q(\mu )$
evaluated at an
intermediate scale $\mu$ which presents a technical
complication. On the other hand this quantity can reliably
be extracted from data \cite{OPTICAL}; furthermore it drops out
from
lifetime {\em ratios}, the main subject of this discussion.
(D) {\em On Duality:}
Quark-hadron duality equates transition rates calculated on
the quark-gluon level with the observable ones involving
the corresponding hadrons -- provided one sums over a
sufficient number of final states.
Since the early days of the quark model this concept has
been invoked in many different formulations; among other
things it has never been clearly defined what
constitutes a sufficient number of final states. This lack of
a precise formulation emerged since duality had never been
derived from QCD in a rigorous fashion; furthermore a certain
flexibility in an unproven, yet appealing intuitive concept can
be of considerable heuristic benefit.
\noindent Heavy quark expansions assume the validity of duality.
Nevertheless they provide new and fruitful insights into its
workings. As already stated, heavy quark expansions are based
on an OPE, see eq.(26), and as such are properly defined in Euclidean
space. There are two possible limitations to the procedure by which
duality is implemented in heavy quark expansions. (i) The
size of the coefficients $c_i^{(f)}$ is controlled by short-distance
dynamics. In concrete applications they are actually obtained
within perturbation theory. Those computations involve
integrals over all momenta. However for momenta below
the scale $\mu$ a perturbative treatment is unreliable.
The contributions from this regime might turn out to be
numerically insignificant for the problem at hand; yet in
any case one can undertake to incorporate them through the
expectation values of higher-dimensional operators -- the
so-called condensates. This procedure is the basis of what is
somestimes referred to as the `practical' version of the OPE.
Yet it is conceivable that there are significant short-distance
contributions from non-perturbative dynamics as well;
instantons provide an illustration for such a complication,
although their relevance is quite unclear at present \cite{MISHA}.
(ii) Once the OPE has been constructed and its coefficients
$c_i^f$ determined there, one employs a dispersion relation to
analytically continue them into Minkowski space. This means,
strictly speaking, that in Minskowski space only `smeared'
transition rates can be predicted, i.e. transition rates
averaged over some finite energy range. This situation was
first analyzed in evaluating the cross section for
$e^+e^- \rightarrow had $ \cite{POGGIO}.
Through an OPE QCD allows to compute
$$\aver{\sigma (e^+e^- \rightarrow had; E_{c.m.})}
\equiv \frac{1}{\Delta E_{sm}}
\int ^{E_{c.m.}+\Delta E_{sm}}_{E_{c.m.}-\Delta E_{sm}}
dE'_{c.m.} \sigma (e^+e^- \rightarrow had; E'_{c.m.})$$
with
$0< \Delta E_{sm} \ll E_{c.m.}$. If the cross section happens to be
a smooth function of $E'_{c.m.}$ -- as it is the case well above
production thresholds --, then one can effectively take the limit
$\Delta E_{c.m.} \rightarrow 0$ to predict
$\sigma (e^+e^- \rightarrow had; E_{c.m.})$ for a
{\em fixed} c.m. energy $E_{c.m.}$. This scenario can be referred to
as {\em local} duality. Yet close to a threshold, like for charm or
beauty production, with its resonance structure one has to retain
$\Delta E_{sm} \sim \mu _{had}\sim 0.5 - 1$ GeV. The same
considerations are applied to heavy-flavour decays. Based on
{\em global} duality one can predict `smeared' decay rates, i.e.
decay rates averaged over a finite energy interval. If the energy
release is large enough the decay rate will be a smooth function of it,
smearing will no longer be required and {\em local} duality
emerges. How large the energy release has to be for this
simplification to occur cannot be predicted (yet). For the onset of
local duality is determined by terms that cannot be seen in any
finite order of the $1/m_Q$ expansion. However rather general
considerations lead to the following expectations: (i) This onset
should be quite abrupt around some energy scale exceeding usual
hadronic scales. (ii) Local duality should hold to a good degree of
accuracy for beauty decays, even for nonleptonic ones. (iii) On the
other hand there is reason for concern that this onset might not have
occurred for charm decays. This is sometimes referred to by saying
that inclusive charm decays might receive significant contributions
from (not-so-)`distant cuts'.
\section{Size of the Matrix Elements}
As discussed before, see eq.(27), the size of matrix elements depends
in general on the scale $\mu$ at which they are evaluated. This will
be addressed separately for the different cases.
The scalar dimension-three operator $\bar QQ$ can be expanded in
terms of $\bar Q\gamma _0Q$ -- with
$\matel{H_Q}{\bar Q\gamma _0Q}{H_Q}_{norm} = 1$ --
and operators of dimension five and higher, see eq.(15).
\noindent (i) Dimension-five operators
\noindent Employing eq.(16) for $b$ and $c$ quarks one finds:
$$\aver{\mu _G^2}_{B}\equiv
\matel{B}{\bar b\frac{i}{2}\sigma \cdot Gb}{B}_{norm} \simeq
\frac{3}{4} (M^2_{B^*}-M^2_B)\simeq 0.37\, (GeV)^2, \eqno(28a)$$
$$\aver{\mu _G^2}_D\equiv
\matel{D}{\bar c\frac{i}{2}\sigma \cdot Gc}{D}_{norm} \simeq
\frac{3}{4} (M^2_{D^*}-M^2_D)\simeq 0.41\, (GeV)^2,\eqno(28b)$$
i.e., these two matrix elements that have to coincide in the
infinite mass limit are already very close to each other.
They are already evaluated at the scale that is
appropriate for beauty and
charm decays.
For the description of $B_c$ decays one needs the expectation
value of the beauty as well as charm chromomagnetic operator
taken between the $B_c$ state. While the $B_c^*$ and $B_c$ masses
have not been measured yet, one expects the theoretical
predictions \cite{QUIGG} on them to be fairly reliable. With
$M(B^*_c)\simeq 6.33$ GeV and $M(B_c)\simeq 6.25$ GeV one obtains
$$\matel{B_c}{\bar b\frac{i}{2}\sigma \cdot Gb}{B_c}_{norm} \simeq
\matel{\bar B_c}{\bar c\frac{i}{2}\sigma \cdot Gc}{\bar B_c}_{norm}
\simeq 0.75\; (GeV)^2 \; ; \eqno(29)$$
i.e., double the value as for the mesons with light antiquarks.
{}From eq.(20) one obtains for the {\em difference}
in the kinetic energy
of the heavy quark inside baryons and mesons
$$\langle (\vec p_b)^2\rangle _{\Lambda _b}-
\langle (\vec p_b)^2\rangle _B
\simeq -(0.07 \pm 0.20)\, (GeV)^2 \eqno(30)$$
using available data; for charm the same value holds
to this order. We see that at present one cannot tell
whether the kinetic energy of the heavy quark
inside baryons exceeds that inside mesons or not. The
uncertainty in the mass of $\Lambda _b$ constitutes the bottleneck;
it would be quite desirable to decrease
its uncertainty down to below 10 MeV or even less. For
the overall size of $\langle (\vec p_Q)^2\rangle _{H_Q}$ we have
the following lower bounds from eq.(19):
$$\aver{(\vec p_b)^2}_B\geq 0.37 (GeV)^2\eqno(31)$$
$$\aver{(\vec p_c)^2}_D\geq 0.40 (GeV)^2\eqno(32)$$
To be conservative one can add a term of at most
$\pm 0.15\; (GeV)^2$ to the right-hand side of eqs.(31,32)
reflecting the uncertainty in the scale at which the
expectation values are to be evaluated
\cite{OPTICAL}. There are various
arguments as to why the size of $\aver{(\vec p_b)^2}_B$ will
not exceed this lower bound significantly. One should
note that such values for $\aver{(\vec p_b)^2}_B$ are
surprisingly large: for they state that the momenta of the
heavy quark inside the hadron is typically around 600 MeV.
An analysis based on QCD sum rules yields a value that is
consistent with the preceding discussion
\cite{QCDSR}:
$$\aver{(\vec p_b)^2}_B = (0.5\pm 0.1)\; (GeV)^2\eqno(33)$$
It can be expected that practically useful
values for $\aver{(\vec p_{c,b})^2}_{D,B}$ will be derived from
lattice QCD in the forseeable future
\cite{GUIDO}.
The leading non-perturbative corrections are thus controlled
by the following parameters:
$$\frac{\langle \mu ^2_G\rangle _B}{m_b^2}\simeq 0.016 \; \; ,
\frac {\langle (\vec p_b)^2\rangle _B}{m_b^2} \sim 0.016
\eqno(34a)$$
$$\frac{\langle \mu ^2_G\rangle _D}{m_c^2}\simeq 0.21 \; \; ,
\frac {\langle (\vec p_c)^2\rangle _D}{m_c^2} \sim 0.21
\eqno(34b)$$
$$\frac{\aver{\mu ^2_G}_{\Omega _c}}{m_c^2}
\simeq 0.12 \eqno(35)$$
The $\Omega _c$ mass hyperfine splitting has not been
measured yet. A quark model estimate of 70 MeV has
been used in evaluating eq.(17b).
\noindent (ii) Dimension-six Operators
\noindent The size of the decay
constants $f_B$, $f_{B_s}$, $f_D$ and
$f_{D_s}$ has not been determined yet with good accuracy.
Various theoretical technologies yield
$$f_D \sim 200\, \pm \, 30 \, MeV \eqno(36)$$
$$f_B \sim 180\, \pm \, 40 \, MeV \eqno(37)$$
$$\frac{f_{B_s}}{f_B} \simeq 1.1 - 1.2 \eqno(38)$$
$$\frac{f_{D_s}}{f_D} \simeq 1.1 - 1.2 \eqno(39)$$
where a higher reliability is attached to the predictions for the
ratios of decay constants, i.e. eqs.(38,39), than for their
absolute size, eqs.(36,37). Recent studies by WA75 and
by CLEO on $D_s^+\rightarrow \mu ^+ \nu $ yielded values for $f_{D_s}$
that are somewhat larger, but still consistent with these predictions.
It can be expected that lattice QCD will produce
more precise results on these decay constants in the foreseeable
future and that those will be checked by future measurements
in a significant way. Yet there are two subtleties to be kept in mind
here:
\noindent ($\alpha$) The matrix element of the four-quark operator
is related to the decay constant through the assumption of
`vacuum saturation'. Such an ansatz cannot hold as an identity; it
represents an approximation the validity of which has to depend on
the scale at which the matrix element is evaluated. The quantities
$f_B$ and $f_D$ are measured in $B\rightarrow \tau \nu ,\, \mu \nu$ and
$D\rightarrow \tau \nu ,\, \mu \nu$, respectively and thus probed at the
heavy-flavour mass, $m_B$ and $m_D$. Yet for the strong
interactions controling the size of the expectation value of
the four-quark operator the heavy-flavour mass is a completely
foreign parameter. If vacuum saturation makes (approximate)
sense anywhere, it has to be at ordinary hadronic scales
$\mu \simeq 0.5 - 1$ GeV. That means the decay constant that
is observed at the heavy-flavour scale has to be evaluated at the
hadronic scale $\mu$; this is achieved by the so-called hybrid
renormalization to be described later.
\noindent ($\beta$) The difference in decay widths generated by
the four-quark operators is stated as being proportional to
$f^2_{H_Q}/m_Q^2$ with
$f_{H_Q}= F_{H_Q}[1 - |\bar \mu |/m_Q +{\cal O}(1/m_Q^2)]$;
$F_{H_Q}$ represents the leading or `static' term which
behaves like $1/\sqrt{m_Q}$ for $m_Q\rightarrow \infty$
and therefore $F_D > F_B$. Thus
asymptotically one has $F^2_{H_Q}/m_Q^2$ which indeed vanishes
like
$1/m_Q^3$. Yet the decay constant $f_{H_Q}$ -- partly as a
consequence of its
usual definition via an axialvector rather than a pseudoscalar
current -- contains large pre-asymptotic corrections which lead to
$f_D \sim f_B$. Then it is not clear which value to use
for $f_{H_Q}$ when calculating width differences, the asymptotic
one -- $F_{H_Q}$ -- or the one including the pre-asymptotic
corrections -- $f_{H_Q}$ where numerically
$F_{H_Q} > f_{H_Q}$ holds. .
It can be argued that the main impact of some of the
dimension-seven operators in the OPE for the meson decay width is
to renormalise the $F^2_{H_Q}/m_Q^2$ term
to $f^2_{H_Q}/m_Q^2$; yet this
issue can be decided only through computing the contributions from
dimension-seven operators. I will return to this issue when
discussing beauty and charm decays specifically.
Both these considerations also apply when determining the {\em
baryonic} expectation values of four-quark
operators \cite{MARBELLA}. Yet -- as already pointed out -- that is a
considerably more murky affair, as will become quite apparent
in the subsequent discussions of the beauty and charm lifetimes.
While there are significant uncertainties and ambiguities in the
values of the masses of beauty and charm quarks their difference
which is free of renormalon contributions is
tightly constrained:
$$m_b-m_c\simeq \langle M\rangle _B - \langle M\rangle _D +
\langle (\vec p)^2\rangle \cdot
\left( \frac{1}{2m_c}- \frac{1}{2m_b}\right)
\simeq 3.46 \pm 0.04\, GeV\, .\eqno(40)$$
This value agrees very well with the one extracted from
an analysis of energy spectra in semileptonic $B$ decays
\cite{VOLOSHIN2}.
Lifetime ratios are hardly sensitive to this difference; we list it here
mainly for completeness.
\section{Predictions on Beauty Lifetimes}
Within the Standard Model there is no real uncertainty
about the weak forces driving beauty
decays; at the scale $M_W$ they are given by the Lagrangian
$${\cal L}_W^{\Delta B=1}(\mu = M_W)=
\frac{4G_F}{\sqrt{2}}
[V_{cb}\bar c_L\gamma _{\mu}b_L+
V_{ub}\bar u_L\gamma _{\mu}b_L]\cdot
[V_{ud}^*\bar d_L\gamma _{\mu}u_L +
V_{cs}^*\bar s_L\gamma _{\mu}c_L] \eqno(41)$$
for non-leptonic transitions with an analogous expression for
semileptonic ones. In eq.(41) I have ignored Cabibbo-suppressed
transitions and also the $b\rightarrow t$ coupling since here we are not
interested in Penguin contributions. Radiative QCD corrections lead
to a well-known renormalization at scale $m_b$, which is often
referred to as {\em ultra-violet} renormalization:
$${\cal L}_W^{\Delta B=1}(\mu = m_b)=
\frac{4G_F}{\sqrt{2}} V_{cb}V_{ud}^*
\{ c_1(\bar c_L\gamma _{\mu}b_L)(\bar d_L\gamma _{\mu}u_L)
+c_2(\bar d_L\gamma _{\mu}b_L)(\bar c_L\gamma _{\mu}u_L)
\} \eqno(42)$$
for $b\rightarrow c\bar ud$ transitions and an analogous expression
for $b\rightarrow c \bar cs$;
the QCD corrections are lumped together into the coefficients $c_1$
and $c_2$ with
$$c_1=\frac{1}{2}(c_+ + c_-), \; \; c_2=\frac{1}{2}(c_+ - c_-)
\eqno(43a)$$
$$c_{\pm}=\left[ \frac{\alpha _S(M_W^2)}{\alpha _S(m_b^2)}
\right] ^{\gamma _{\pm}}, \; \;
\gamma _+=\frac{6}{33-2N_f}=-\frac{1}{2}\gamma _-
\eqno(43b)$$
in the leading-log approximation; $N_f$ denotes the number
of active flavours. Numerically this amounts to:
$$c_1(LL) \simeq 1.1, \; \; c_2(LL) \simeq -0.23 \, .
\eqno(44a)$$
Next-to-leading-log corrections modify mainly $c_2$
\cite{CHERNYAK}:
$$c_1(LL+NLL) \simeq 1.13, \; \; c_2(LL+NLL) \simeq -0.29 \, .
\eqno(44b)$$
Radiative QCD corrections thus lead to a mild enhancement of the
original coupling -- $c_1>1$ -- together with the appearance of an
induced operator with a different colour flow: $c_2\neq 0$. Later on
we will also include the so-called `hybrid' renormalization reflecting
radiative corrections in the domain from $m_b$ down to
$\mu _{had}$, the scale at which the hadronic expectation values
are to be evaluated.
Such perturbative corrections affect the overall scale of
the semileptonic and non-leptonic width (and
thus of the semileptonic branching ratio); but
by themselves they cannot generate {\em differences} among
the lifetimes of beauty hadrons. Those have to be
initiated by non-perturbative contributions although their
numerical size depends also on perturbative corrections.
The semileptonic and non-leptonic widths of
beauty {\em hadrons} $H_b$ through
order $1/m_b^2$ are given by
$$\frac{\Gamma _{SL,decay}(H_b)}{\Gamma _0} =
\matel{H_b}{\bar bb}{H_b}_{norm}\cdot
\left[ \eta _{SL}I_0(x,0,0) +
\frac{\aver{\mu _G^2}_{H_b}}{m_b^2}
(x\frac{d}{dx}-2)I_0(x,0,0)\right] \; , \eqno(45a)$$
$$\frac{\Gamma _{NL,decay}(H_b)}{\Gamma _0} = N_C\cdot
\matel{H_b}{\bar bb}{H_b}_{norm}\cdot
\left[ \frac{A_0}{3}
[I_0(x,0,0)+ \rho _{c\bar c} I_0(x,x,0) +\right.
$$
$$\left. \frac{\aver{\mu _G^2}_{H_b}}{m_b^2}
(x\frac{d}{dx}-2)(I_0(x,0,0)+I_0(x,x,0))] -
\frac{4}{3} A_2 \frac{\aver{\mu _G^2}_{H_b}}{m_b^2} \cdot
[I_2(x,0,0,) + I_2(x,x,0)] \right] \; , \eqno(45b)$$
$$\Gamma _0 \equiv
\frac{G_F^2m_b^5}{192 \pi ^3}|V(cb)|^2 \; . \eqno(45c)$$
The following notation has been used here:
$I_0$ and $I_2$ are phase-space factors, namely
$$I_0(x,0,0)= (1-x^2)(1-8x+x^2)-12x^2\log x \eqno(46a)$$
$$I_2(x,0,0)=(1-x)^3\; , \; \; x=(m_c/m_b)^2 \eqno(46b)$$
for $b \rightarrow c \bar ud/c l\bar \nu$ and
$$I_0(x,x,0)=v(1-14x-2x^2-12x^3) + 24x^2(1-x^2)
\log \frac{1+v}{1-v}\, , \, v=\sqrt{1-4x}\eqno(46c)$$
$$I_2(x,x,0)=v(1+\frac{x}{2}+3x^2)-3x(1-2x^2)
\log \frac{1+v}{1-v} \, , \eqno(46d)$$
for $b\rightarrow c \bar cs$
transitions. The quantities $\eta _{SL}$, $\rho _{c\bar c}$,
$A_0$ and $A_2$ represent the
QCD radiative corrections. More specifically one has
$$A_2=(c_+^2-c_-^2)\; \; \; , \; \; \;
A_0= (c_-^2+2c_+^2)\cdot J \eqno(46e)$$
with $J$ reflecting the effect
of the subleading logarithms \cite{PETRARCA} and
$$\eta _{SL}\simeq 1+
\frac{2}{3}\frac{\alpha_s}{\pi}g(m_c/m_b,m_l/m_b,0)
\; . \eqno(46f)$$
The function $g$ can be computed numerically for arbitrary
arguments \cite{PHAM} and analytically for the most
interesting case $m_l=0$ \cite{NIR}.
The allowed values
for $\rho _{c\bar c}$,which reflects the fact that QCD
radiative corrections are quite sensitive to the final state quark
masses, can be found in ref.\cite{BAGAN}.
With $x\simeq 0.08$ one obtains
$$I_0(x,0,0)|_{x=0.08}\simeq 0.56 \; \; , \; \;
I_2(x,0,0)|_{x=0.08}\simeq 0.78 \; \; for \; \;
b\rightarrow c \bar ud$$
$$I_0(x,x,0)|_{x=0.08}\simeq 0.24 \; \; , \; \;
I_2(x,x,0)|_{x=0.08}\simeq 0.32 \; \; for \; \;
b\rightarrow c \bar cs$$
Since these functions are normalized to unity for $x=0$, one
notes that the final-state quark masses reduce the
available phase space quite considerably.
Some qualitative statements can illuminate the dynamical situation:
\noindent (i) As indicated in eqs.(15),
$\Gamma _{SL/NL,decay}$ differ from the naive
spectator result in order $1/m_b^2$.
\noindent (ii) Since $\, \bar bi\sigma \cdot G b\, $ and
$\, \bar b (i\vec D)^2b\, $
are $SU(3)_{Fl}$ singlet operators and
their expectation values are practically isospin and even
$SU(3)_{Fl}$ invariant, one obtains, as stated before,
$\tau (B_d)\simeq \tau (B^-)\simeq \bar \tau (B_s)$
through order $1/m_b^2$ and likewise
$\tau (\Lambda _b)\simeq \tau (\Xi _b)$. Yet the
meson and baryon lifetimes get differentiated on this level
since $\aver{\mu _G^2}_B > 0 =
\aver{\mu _G^2}_{\Lambda _b,\, \Xi _b}$ and
$\aver{(\vec p_b)^2}_B \neq
\aver{(\vec p_b)^2}_{\Lambda _b}$, see eqs.(20,30).
\noindent (iii) The semileptonic branching ratios of
$\Lambda _b$ and $\Xi _b$ baryons remain unaffected
in order $1/m_b^2$
due to $\aver{\mu _G^2}_{\Lambda _b,\, \Xi _b}\simeq 0$,
whereas for $B$ mesons it is (slightly) reduced \cite{BUV};
i.e., one finds
$$BR_{SL}(B) < BR_{SL}(\Lambda _b) +
{\cal O}(1/m_b^3)\; . $$
Beyond order $1/m_b^2$, however, this relation is changed, as
discussed later on.
\subsection{$B^-$ vs. $B_d$ Lifetimes}
Differences in
$\tau (B^-)$ vs. $\tau (B_d)$ are generated by the local
dimension-six four-quark
operators $(\bar b_L\gamma _{\mu}q_L)
(\bar q_L\gamma _{\nu}b_L)$
which explicitely depend on the
light quark flavours. Such corrections are of order
$1/m_b^3$. Based on this scaling law one can
already infer from the observed $D$ meson lifetime ratios that the
various $B$ lifetimes will differ by no more than 10\% or so.
Assuming factorization for the expectation values of these
four-fermion operators, i.e. applying eq.(22), one obtains
$$\frac{\Gamma (B_d)-\Gamma (B^-)}{\Gamma (B)}
\sim \frac{\Gamma _{nonspect}(B)}{\Gamma _{spect}(B)}
\propto \frac{f_B^2}{m_b^2}\; , \eqno(47)$$
as discussed above.
It turns out that WA can change the $B$ lifetimes by
no more than1\% or so; due to interference with the spectator
reaction, they could conceivably
{\em prolong} the $B_d$ lifetime
relative to the $B^-$ lifetime, rather than reduce it. The dominant
effect is clearly provided by PI, which produces a negative
contribution to the $B^-$ width:
$$\Gamma (B^-)\simeq \Gamma _{spect}(B)+
\Delta \Gamma _{PI}(B^-)\eqno(48a)$$
$$ \Delta \Gamma _{PI}(B^-)\simeq \Gamma _0\cdot
24\pi ^2\frac{f_B^2}{m_b^2}
\left[ c_+^2-c_-^2+\frac{1}{N_C}(c_+^2+c_-^2)\right]
\; .\eqno(48b)$$
Eq.(48b) exhibits an intriguing result:
$\Delta \Gamma _{PI}(B^-)$ is positive for $c_+=1=c_-$, i.e.
PI acts {\em constructively} and thus would shorten the
$B^-$ lifetime in the {\em absence} of radiative QCD corrections
\footnote{This shows that the term of `Pauli Interference' should
{\em not} be construed as implying that the interference
is a priori destructive.}.
Including those as evaluated on the leading-log and
next-to-leading-log level -- $c_+\simeq 0.84$, $c_-\simeq 1.42$ --
turns PI into a {\em destructive} interference prolonging
the $B^-$ lifetime, albeit by a tiny amount only at this point.
In eq.(48b) only ultraviolet renormalization has been incorporated.
Hybrid renormalization \cite{HYBRID}
down to the hadronic scale
$\mu _{had}$ amplifies this effect considerably; one
obtains
$$ \Delta \Gamma _{PI}(B^-)\simeq \Gamma _0\cdot
24\pi ^2\frac{f_B^2}{M_B^2}\kappa ^{-4}
\left[ (c_+^2-c_-^2)\kappa ^{9/4}+\frac{c_+^2+c_-^2}{3}
-\frac{1}{9}(\kappa ^{9/2}-1)(c_+^2-c_-^2)\right] \, , $$
$$\kappa \equiv \left[ \frac{\alpha_s (\mu _{had}^2)}
{\alpha_s (m_b^2)}\right] ^{1/b}, \;
b=11-\frac{2}{3} n_F \eqno(49)$$
Putting everything together one finds:
$$\frac{\tau (B^-)}{\tau (B_d)}\simeq
1+0.05 \cdot \frac{f_B^2}{(200\; MeV)^2}\; , \eqno(50)$$
i.e. the lifetime of a charged $B$ meson is predicted to
definitely {\em exceed} that of a neutral $B$ meson by
typically several percent.
Three comments are in order for properly evaluating
eq.(49):
\noindent $\bullet$ Although WA could conceivably prolong the
$B_d$ lifetime as stated before, its numerical significance pales
by comparison to that of PI. PI acts destructive in
$B^-$ decays once radiative QCD corrections are included to the best
of our knowledge. The lifetime of $B^-$ mesons therefore has to
exceed that of $B_d$ mesons.
\noindent $\bullet$ While the sign of the effect is predicted in an
unequivocal manner, its magnitude is not. The main uncertainty in
the prediction on $\tau (B^-)/\tau (B_d)$ is given by our present
ignorence concerning the size of $f_B$.
\noindent $\bullet$ Even a precise measurement of the lifetime ratio
$\tau (B^-)/\tau (B_d)$ would not automatically result in an
exact determination of $f_B$ by applying eq.(50). For corrections
of order $1/m_b^4$ have not (or only partially -- see the discussion
in the
preceding subsection) been included. Unless those have been
determined, one cannot extract the size of $f_B$ even from
an ideal measurement with better than a roughly 15 \%
uncertainty.
\subsection{$B_s$ Lifetimes}
Very little $SU(3)_{Fl}$ breaking is anticipated between the $B_d$
and $B_s$ expectation values of the two dimension-five operators.
\footnote{This issue will be addressed in some detail in the later
discussion of $D^0$ vs. $D_s$ lifetimes.} Among the contributions
from dimension-six operators
WA affects $B_d$ and $B_s$ lifetimes somewhat differently due to
different colour factors for these two decays. Yet it is quite irrelevant
in either case. Thus one predicts the $B_d$ and the average $B_s$
lifetimes to practically coincide:
$$\tau (B_d) \simeq \bar \tau (B_s) \pm {\cal O}(1\%)
\eqno(51)$$
The term `average $B_s$ lifetime' is used for a reason:
$B_s - \bar B_s$ oscillations generate two neutral beauty mesons
carrying strangeness that differ in their mass as well as in their
lifetime. Due basically to $m_t^2 \gg m_c^2$ one finds
$\Delta m(B_s) \gg \Delta \Gamma (B_s) \neq 0$. While
$\Delta m(B_s)$ can be calculated in terms of the expectation value
of the {\em local} four-fermion operator
$(\bar b_L\gamma _{\mu }q_L)(\bar b_L\gamma _{\mu }q_L)$
(since $m_b \ll m_t$), the situation is more complex for
$\Delta \Gamma (B_s)$, since the underlying operator is nonlocal.
One can however apply a heavy quark expansion; to lowest
nontrivial
order one obtains \cite{BSBS}
$$\frac{\Delta \Gamma (B_s)}{\bar \Gamma (B_s)}
\equiv \frac{\Gamma (B_{s,short})-\Gamma (B_{s,long})}
{\bar \Gamma (B_s)}\simeq 0.18\cdot
\frac{(f_{B_s})^2}{(200 \, MeV)^2} \eqno(52)$$
for $f_{B_s}$ not much larger than 200 MeV. Comparing
eqs.(50-52) leads to the intriguing observation that the largest
lifetime difference among $B^-$, $B_d$ and $B_s$ mesons is
generated by a very subtle source:
$B_s - \bar B_s$ oscillations! There are also two different
lifetimes for neutral $B$ mesons without strangeness; yet $\Delta
\Gamma (B_d)$
is suppressed by $\sim \sin \theta ^2_C$ relative to
$\Delta \Gamma (B_s)$.
One can search for the existence of two different $B_s$ lifetimes by
comparing $\tau (B_s)$ as measured in
$B_s \rightarrow \psi \eta /\psi \phi$ on one hand and in
$B_s\rightarrow l \nu X$ on the other. The
former decay predominantly leads to a CP even final state
and thus would,
to good accuracy, reveal $\tau (B_{s,short})$; the latter
exhibits the
average lifetime $\bar \tau (B_s)=
[\tau (B_{s,long} - \tau (B_{s,long})]/2$; for semileptonic $B_s$
decays involve the CP even and odd components in a nearly
equal mixture. Thus
$$\tau (B_s \rightarrow l \nu D^{(*)}) -
\tau (B_s\rightarrow \psi \eta /\psi \phi ) \simeq
\frac{1}{2} [\tau (B_{s,long})-\tau (B_{s,short})]\simeq $$
$$\simeq 0.09\cdot
\frac{(f_{B_s})^2}{(200 \, MeV)^2}. \eqno(53)$$
Whether an effect of this size is large enough to be ever observed
in a real experiment, is doubtful. Nevertheless one should search
for it even if one has sensitivity only for a 50\% lifetime difference
or so. For while eq.(52) represents the best presently available
estimate, it is not a `gold-plated' prediction. It is conceivable that the
underlying computation {\em underestimates} the actual lifetime
difference!
\subsection{$B_c$ Lifetime}
$B_c$ mesons with their two heavy constituents
-- $B_c=(b\bar c)$ -- represent a highly intriguing system that
merits special efforts to observe it. One expects a
rich spectroscopy probing the inter-quark potential at distances
intermediate to those that determine quarkonia spectroscopy in the
charm and in the beauty system. The Isgur-Wise function for the
striking channel $B_c\rightarrow l \nu \psi$ can be computed. What is most
relevant for our discussion here is that its overall decays, and thus
its lifetime, reflect the interplay of three classes of transitions,
namely
the decay of the $b$ quark, that of the $\bar c$ (anti)quark and WA
of $b$ and $\bar c$:
$$\Gamma (B_c)\simeq \Gamma _{b\rightarrow c} (B_c) +
\Gamma _{\bar c \rightarrow \bar s}(B_c) + \Gamma _{WA}(B_c) \eqno(54)$$
While $b\rightarrow c$ and $\bar c \rightarrow \bar s$ transitions do not
interfere with each other in any practical way and one can thus
cleanly separate their widths, the situation is much more
delicate concerning $\Gamma _{b \rightarrow c}$ and
$\Gamma _{WA}$, as briefly explained later. Yet for the
moment I ignore the latter although
its helicity and wavefunction suppressions represented by the
factors
$m_c^2/m_b^2$ and $f^2_{B_c}/m_b^2$ are relatively mild
($f_{B_c} \simeq 400-600$ MeV \cite{QUIGG})
and partially offset by the
numerical factor $16\pi ^2$ reflecting the enhancement of two-body
phase space -- relevant for WA -- over three-body phase space
appropriate for the spectator decay.
Concerning the other two transitions one would naively expect
the $\bar c$ decay to dominate:
$\Gamma _{\bar c\rightarrow \bar s}(B_c) \sim \Gamma (D^0)
\simeq (4\times 10^{-13}\, sec)^{-1} >
\Gamma _{b \rightarrow c}(B_c) \sim \Gamma (B) \simeq
(1.5\times 10^{-12}\, sec)^{-1}$. It had been suggested that for
a tightly bound system like $B_c$ the decay width should be
expressed not in terms of the usual quark masses (whatever
they are), but instead in terms of effective quark masses
reduced by something like a binding energy
$\mu _{BE} \sim 500$ MeV.
{\em If} so, then the $B_c$ width would be reduced considerably
since due to $\Gamma _Q \propto m_Q^5$ one would find
a relative reduction
$(\Gamma _Q + \Delta \Gamma _Q)/\Gamma _Q \sim
(m_Q - \mu _{BE})^5/m_Q^5 \sim 1 -5\mu _{BE}/m_Q$.
Even more significantly, beauty decays would become more
abundant than charm decays in the $B_c$ transitions since the
binding energy constitutes a higher fraction of $m_c$ than
of $m_b$: $\mu _{BE}/m_c > \mu _{BE}/m_b$.
However this conjecture that might look quite
plausible at first sight, turns out to be fallacious!
For it is manifestly
based on the existence of nonperturbative corrections of
order $1/m_Q$;
yet as discussed in Sect.4 those are absent in fully integrated
decay rates like lifetimes due to a non-trivial cancellation between
$1/m_Q$ contributions from initial-state and final-state radiation!
The leading corrections arise in order $1/m_Q^2$ and
they enter through the first two terms on the
right-hand-side of eq.(54):
$$\Gamma _{b \rightarrow c}(B_c) = \Gamma _{b \rightarrow c,decay}(B_c)
+ {\cal O}(1/m_b^3) \eqno(55a)$$
$$\Gamma _{\bar c \rightarrow \bar s}(B_c) =
\Gamma _{\bar c \rightarrow \bar s,decay}(B_c)
+ {\cal O}(1/m_c^3) \eqno(55b)$$
The quantities $\Gamma _{b \rightarrow c,decay}(B_c)$ and
$\Gamma _{\bar c \rightarrow \bar s}(B_c)$ are defined as in
eqs.(45) and (68); i.e., the differences between
$\Gamma _{decay}(B)$ and $\Gamma _{b \rightarrow c,decay}(B_c)$
enter through the different expectation values of the
same operators $\bar QQ$ and $\bar Q\frac{i}{2}\sigma \cdot G Q$,
and likewise for
$\Gamma _{decay}(D)$ (to be discussed in detail
in Sect.7) vs. $\Gamma _{\bar c \rightarrow \bar s,decay}(B_c)$.
For the $B_c$ meson provides a different environment for
quark decay than either $B$ or $D$ mesons. The relevance of
that difference is illustrated by eqs.(28,29).
Similarly one expects $\aver{(\vec p_b)^2}_{B_c}$,
$\aver{(\vec p_c)^2}_{B_c}$ to differ from
$\aver{(\vec p_b)^2}_B$, $\aver{(\vec p_c)^2}_D$.
While there are large corrections of order $1/m_c^2$ in
$\Gamma _{\bar c\rightarrow \bar s,decay}(B_c)$ that reduce the
corresponding semileptonic branching ratio considerably,
they largely cancel against each other in the total width.
Therefore
$$ \Gamma _{\bar c\rightarrow \bar s,decay}(B_c) \sim
\Gamma _{decay}(D) \simeq \Gamma (D^0) \eqno(56)$$
Writing the $\Delta B=1$ width of the $B_c$ meson as a
simple incoherent sum $\Gamma _{b\rightarrow c}+\Gamma _{WA}$
actually represents an oversimplification. For there
arises considerable interference between higher order
WA and spectator processes. Yet for the purposes of
our discussion here such effects can be ignored;
they will be discussed in detail in a forthcoming
publication. The $1/m_b^2$ contributions to
$\Gamma _{b\rightarrow c,decay}(B_c)$ are small, namely around
a few percent. The numerical impact of
$\Gamma _{WA}(B_c \rightarrow X_{\bar c s})$, which is formally
of order $1/m_b^3$, is nevertheless sizeable due to the
large decay constant, the merely mild helicity suppression,
given by $m_c^2/m_b^2$, and the enhancement
of WA by a factor $\sim 16\pi ^2$ due to its two-body
kinematics. One finds:
$$\Gamma _{b \rightarrow c}(B_c) \geq \Gamma _{b\rightarrow c}(B)\; .
\eqno(57)$$
Comparing eqs.(56,57) one concludes that
the naive expectation turns out to be basically
correct, i.e. $\tau (B_c)$ is well below $10^{-12}$ sec and charm
decays
dominate over beauty decays with
an ensuing reduction in the `interesting' branching ratios
like $B_c\rightarrow l\nu \psi$ or $B_c\rightarrow \psi \pi$.
\subsection{Beauty Baryon Lifetimes}
It is quite natural to assume that the kinetic energy of the $b$ quark
is practically the same inside the $\Lambda _b$
and $\Xi _b$ baryons:
$$\aver{(\vec p _b)^2}_{\Lambda _b}\simeq
\aver{(\vec p _b)^2}_{\Xi _b} \eqno(58a)$$
Together with eqs.(15,17a) this yields
$$\matel{\Lambda _b}{\bar bb}{\Lambda _b}=
\matel{\Xi _b}{\bar bb}{\Xi _b} + {\cal O}(1/m_b^3)
\eqno(58b)$$
The leading {\em differences} among the $\Lambda _b$ and
$\Xi _b$ lifetimes
then arise in
order $1/m_b^3$: they are generated by four-quark
operators analogous to those
that had already been identified in the phenomenological
studies of charm baryons\cite{BARYONS1,BARYONS2,BARYONS3}.
Some complexities arise, though, due to the
presence of the two transitions $b\rightarrow c \bar ud$
and $b\rightarrow c \bar cs$; one finds
$$\Gamma (\Lambda _b) = \Gamma _{decay}(\Lambda _b) +
\Gamma _{WS}(\Lambda _b)
-|\Delta \Gamma _{PI,-}(\Lambda _b, b\rightarrow c\bar ud)| \eqno(59a)$$
$$\Gamma (\Xi ^0 _b) = \Gamma _{decay}(\Xi _b) +
\Gamma _{WS}(\Xi _b)
- |\Delta \Gamma _{PI,-}(\Xi _b, b\rightarrow c\bar cs)|
\eqno(59b)$$
$$\Gamma (\Xi ^- _b) = \Gamma _{decay}(\Xi _b)
-|\Delta \Gamma _{PI,-}(\Xi _b, b\rightarrow c \bar ud)|
-|\Delta \Gamma _{PI,-}(\Xi _b, b\rightarrow c\bar cs)| \eqno(59c)$$
where \footnote{The channel $b \rightarrow \tau \nu q$ has been ignored
here for simplicity.}
$$\Gamma _{decay}(\Lambda _b/\Xi _b)
\equiv 2 \Gamma _{SL,decay}(\Lambda _b/\Xi _b)+
\Gamma _{NL,decay}(\Lambda _b/\Xi _b) \eqno(60a)$$
$$\frac{\Gamma _{SL,decay}(\Lambda _b/\Xi _b)}{\Gamma _0} =
\matel{\Lambda _b/\Xi _b}{\bar bb}{\Lambda _b/\Xi _b}_{norm}\cdot
\eta _{SL}I_0(x,0,0) \eqno(60b)$$
$$\frac{\Gamma _{NL,decay}(\Lambda _b/\Xi _b)}{\Gamma _0} =
\matel{\Lambda _b/\Xi _b}{\bar bb}{\Lambda _b/\Xi _b}_{norm}\cdot
A_0\cdot [I_0(x,0,0)+ \rho _{c\bar c} I_0(x,x,0)] \eqno(60c)$$
using the same notation as in eqs.(45);
$$\Gamma _{WS}(\Lambda _b, \Xi _b) \equiv 2\tilde \Gamma _0
c_-^2\matel{\Lambda _b/\Xi _b}{\bar b_L\gamma _{\mu}b_L
\bar u_L \gamma _{\mu}u_L}{\Lambda _b/\Xi _b}_{norm}
\eqno(60d)$$
$$\frac{\Delta \Gamma _{PI,-}(\Lambda _b, b\rightarrow c\bar ud)}
{\tilde \Gamma _0}
\equiv - c_+(2c_- - c_+)
\matel{\Lambda _b}
{\bar b_L\gamma _{\mu}b\bar d_L\gamma _{\mu}d_L +
\frac{2}{3}\bar b\gamma _{\mu}\gamma _5b
\bar d_L\gamma _{\mu}d_L}
{\Lambda _b}_{norm} \eqno(60e)$$
$$\frac{\Delta \Gamma _{PI,-}(\Xi _b, b\rightarrow c\bar cs)}
{\tilde \Gamma _0}\equiv
- c_+(2c_- - c_+)
\matel{\Xi _b}
{\bar b_L\gamma _{\mu}b\bar s_L\gamma _{\mu}s_L +
\frac{2}{3}\bar b\gamma _{\mu}\gamma _5b
\bar s_L\gamma _{\mu}s_L}
{\Xi _b}_{norm} \eqno(60f)$$
$$\tilde \Gamma _0 \equiv 48\pi ^2 \Gamma _0 \eqno(60g)$$
$\Delta \Gamma _{PI,-}(\Xi _b, b\rightarrow c\bar ud)$ is obtained
from eq.(60e) by taking the expectation value between
$\Xi _b$ rather than $\Lambda _b$ states. Those expectation
values vanish for the wrong charge-flavour combination; i.e.,
$\matel{\Xi _b^0}{\bar b_L\gamma _{\mu}b\bar d_L\gamma _{\mu}d_L}
{\Xi _b^0}=0=
\matel{\Lambda _b}{\bar b_L\gamma _{\mu}b\bar s_L\gamma _{\mu}s_L}
{\Lambda _b}$.
$\Gamma _{decay} $ includes the naive
spectator term:
$\Gamma _{decay} = \Gamma _{spect} + 1/m_b^2$ contributions; the
latter are practically identical for $\Lambda _b$ and
$\Xi _b$, but differ for $B$ mesons. $\Gamma _{WS}$ denotes
the contribution due to WS
and $\Delta \Gamma _{PI,-}$ the reduction due to
destructive
interference in the channels $b\rightarrow c\bar ud$ and
$b\rightarrow c \bar cs$, respectively.
\footnote{When $|\Delta \Gamma _{PI,-}(b\rightarrow c \bar ud)|$
is used in eq.(59c), one has,
strictly speaking, to evaluate the expectation value for the
state $\Xi _b$.}
On general grounds one thus obtains an inequality:
$$\tau (\Xi _b^-) > \tau (\Lambda _b)\, , \, \tau (\Xi _b^0)
\eqno(61)$$
Naively one might expect $\Gamma (\Xi _b^0) >
\Gamma (\Lambda _b)$ to hold, since the $b\rightarrow c \bar cs$ part
of the width which is reduced by PI in $\Xi _b^0$ decays
is smaller than the $b\rightarrow c \bar ud$ component which suffers
from PI reduction in $\Lambda _b$ decays. On the other hand since
phase space is more restricted for $b\rightarrow c \bar cs$ than for
$b\rightarrow c \bar ud$, one would likewise expect the degree of
(destructive) interference to be higher for the former than
the latter; it is then quite conceivable that actually
$\Gamma (\Xi _b^0) < \Gamma (\Lambda _b)$ holds. The
{\em sign} of the difference in the $\Lambda _b$ and
$\Xi _b^0$ lifetimes therefore provides us with valuable
information on the strong dynamics.
There are two complementary ways to transform these qualitative
predictions into quantitative ones.
\noindent (i) One evaluates the required expectation values
explicitely within a quark model, as expressed in eq.(17b, 23-25).
Since the model also predicts the baryon masses in terms of its
parameters, one can cross check it with the observed spectroscopy.
This will become increasingly relevant in the future, yet at present
provides little guidance.
The expressions given in eqs.(23-25) have to be augmented
by the radiative QCD corrections:
$$\matel{\Lambda _b}{\bar b\Gamma _{\mu}b
\bar q \Gamma _{\mu} q}{\Lambda _b}_{norm} \sim
\frac{1}{4\aver{\mu ^2_G}_B}(M_{\Sigma _b}-
M_{\Lambda _b})
m^*_q F^2_BM_B\cdot \kappa ^{-4} \eqno(62a)$$
$$\matel{\Omega _b}{\bar b\Gamma _{\mu}b
\bar q \Gamma _{\mu} q}{\Omega _b}_{norm} \sim
\frac{1}{4\aver{\mu ^2_G}_B}(M_{\Sigma _b}-
M_{\Omega _b})
m^*_q F^2_BM_B\cdot \kappa ^{-4} \eqno(62b)$$
with $\kappa$ as defined in eq.(49).
It should be noted that here -- unlike in the case of meson
decays -- the sign of the PI contribution is quite robust
under radiative corrections: it is proportional to
$- c_+(2c_- - c_+)$ which is negative already for
$c_+=1=c_-$. Furthermore $\Gamma _{WS}$ is proportional
to $c_-^2$ and thus colour-enhanced, since the baryon
wavefunction is purely antisymmetric in colour space.
This prescription yields lifetime differences
of not more than a few percent in eq.(59) and
$$ \frac{\tau (\Lambda _b)}{\tau (B_d)} \simeq 0.9 - 0.95 \; .
\eqno(63a)$$
(ii) As will be discussed in the next subsection, the
pattern predicted for charm baryon lifetimes agrees
with the observations within the experimental and
theoretical uncertainties. Taking this as prima facie
evidence that the heavy quark expansion through order
$1/m_Q^3$ applies -- at least in a semi-quantitative fashion --
already to charm lifetime ratios, one can extrapolate to the
weight of these pre-asymptotic corrections in beauty
decays using scaling like $1/m_Q^2$ and $1/m_Q^3$
(or $f_M^2/m_Q^2$).
That way one
again finds that the differences in eq.(56) amount to not more
than a few to several percent \cite{STONE2}:
$$ \frac{\tau (\Lambda _b)}{\tau (B_d)} \simeq 0.9 \; .
\eqno(63b)$$
Similarly one estimates $\tau (\Xi _b)$ through a
simple $1/m_Q^3$ scaling behaviour from the charm baryon
lifetimes:
$$\frac{\tau (\Xi _b^-)}{\tau (\Lambda _b)} \sim 1.1 \eqno(64a)$$
$$\frac{|\tau (\Xi _b^0)- \tau (\Lambda _b)|}
{\tau (\Lambda _b)} < 0.1 \eqno(64b)$$
Obviously and not surprisingly there is some numerical fuzziness
in these predictions; yet they seem to be unequivocal in stating
that the $B_d$ lifetime exceeds the $\Lambda _b$ lifetime
and the average beauty baryon lifetime by about 10 percent.
However this prediction appears to be in serious (though not yet
quite conclusive) disagreement with the data. If the predictions
were based exclusively on adopting the quark model results for the
baryonic expectation values, one could abandon them in a
relatively light-hearted way: for it should not come as a shocking
surprise that quark model results for baryonic matrix elements can
be off-target by a substantial amount. Yet we have encountered
a more serious problem here: data seem to contradict also the prediction
based on an extrapolation from the observed lifetime pattern in the
charm family; furthermore -- as discussed next -- the observed
lifetime ratios of charm hadrons can be reproduced, within the
expected uncertainties. This allows only one conclusion: if
$\tau (B_d)$ indeed exceeds $\tau (\Lambda _b)$ by 25 - 30 \%,
then a `theoretical price' has to be paid, namely that
\noindent $\bullet$ the charm mass
represents too low of a scale for allowing to go beyond merely
qualitative
predictions on charm baryon (or even meson) lifetimes,
since it appears that corrections of order $1/m_c^4$ and higher
are still important;
\noindent $\bullet$ that the present agreement between theoretical
expectations and data on charm baryon lifetimes is largely
accidental and most likely would not survive in the face of more
precise measurements!
\noindent At the same time an intriguing puzzle arises:
\noindent $\bullet$ Why are the quark model results for the
relevant expectation values so much off the mark for
beauty baryons? It is the deviation from unity in the lifetime
ratios that is controlled by these matrix elements; finding
a 30 \% difference rather than the expected 10 \% then represents
a 300 \% error!
Some new features emerge in $\Omega _b$ decays: since the
$ss$-diquark system forms a spin-triplet, there are
$1/m_b^2$ contributions to the semileptonic and
nonleptonic $\Omega _b$ widths from the chromomagnetic
operator. In order $1/m_b^3$ there arises a destructive
interference in the $b\rightarrow c \bar cs$ channels. However a
detailed discussion of $\Omega _b$ decays seems
academic at the moment.
\subsection{Semileptonic Branching Ratios of Beauty
Hadrons}
As briefly discussed before, one expects the semileptonic branching
ratios for $B_d$ and $B_s$ mesons to practically coincide,
in particular since semileptonic decays probe $\bar \tau (B_s)$,
the average $B_s$ lifetime.
It has already been pointed out that through order
$1/m_b^2$ the expected value for $BR_{SL}(\Lambda _b)$
exceeds that for $BR_{SL}(B)$ by a few percent.
In order $1/m_b^3$ the nonleptonic widths of both $B_d$ and
$\Lambda _b$ states receive new contributions, with
$\Gamma _{NL}(\Lambda _b)$ getting further enhanced relative
to $\Gamma _{NL}(B)$ as expressed in eq.(63b). Thus one
predicts
$$BR_{SL}(\Lambda _b) < BR_{SL}(B_d) < BR_{SL}(B^-)$$
with the inequalities indicating differences of a few to
several percent. If the present trend in the data persists, i.e.
if the total $\Lambda _b$ width exceeds $\Gamma (B)$ by,
say, 20 -25 \%, one would interprete this discrepancy as
meaning that the nonleptonic -- but not the semileptonic --
width has received a unforeseen enhancement. In that case
one expects $BR_{SL}(\Lambda _b)$ to fall below
$BR_{SL}(B_d)$ by $\sim$ 20 \%.
Putting all these observations together one concludes
that the beauty lifetime averaged over $B_d$, $B^-$,
$B_s$ and $\Lambda _b$ states should yield a value that
is {\em smaller} -- by a few percent -- than
$\frac{1}{2} \left[ BR_ {SL}(B_d) + BR_ {SL}(B^-)\right]
\equiv \aver{BR_{SL}(B)}$:
$$\aver{BR_{SL}(b)}< \aver{BR_{SL}(B)}$$
\section{Predictions on Charm Lifetimes}
Considering the wealth of rather precise experimental information
available on the lifetimes of charm hadrons one feels the urge to
apply heavy quark expansions to charm decays as well. Yet in doing
so one has to keep in mind that such a treatment might fail here
for two basic reasons, one of which has just been stated:
\noindent (i) The charm quark mass does not provide a
sufficiently large scale to make the $1/m_c$ expansion
converge quickly. To obtain an estimate of the size of the
expansion parameter $\mu _{had}/m_c$ one can take the
square root of the expression in eq.(35)
representing
the $1/m_c^2$ corrections:
$$\frac{\mu _{had}}{m_c}\sim \sqrt{\frac{\aver {\mu _G^2}_D}
{m_c^2}}\simeq 0.45 \, .\eqno(65)$$
This is not a small number although it is at least smaller
than unity.
\noindent (ii) As
discussed in Sect.4 one has to go beyond
{\em global} duality and invoke {\em local} duality to predict the
decay widths of real hadrons from heavy quark expansions properly
defined in Euclidean space. Yet a priori it is not clear at all
whether contributions from `distant cuts' can be ignored since the
charm quark mass does not exceed ordinary hadronic scales by a
large amount. This concern is a posteriori strengthened by the
following observation: Equating the observed semileptonic width of
$D$ mesons with its theoretical expression through order $1/m_c^2$
(and assuming $|V(cs)| \simeq 1$) leads to the requirement
$"m_c" \simeq 1.6$ GeV. This is however a high value relative
to what is derived from charmonium spectroscopy, namely
$m_c \leq 1.4$ GeV. A difference of $0.2$ GeV in $m_c$ might
appear quite innocuous -- till one realizes that the corresponding
semileptonic width depending on $m_c^5$ changes by a factor of
two when $m_c$ is shifted by those $0.2$ GeV! At this point one
might suspect that corrections of higher order in $1/m_c$ contribute
{\em constructively} boosting the theoretical value. The analysis of
ref.\cite{DIKEMAN} finds however that terms of order $1/m_c^3$
show a
tendency to {\em decrease} $\Gamma _{SL}(D)$, and their inclusion
does
not help at all in reproducing the measured value of
$\Gamma _{SL}(D)$ with $m_c\simeq 1.4$ GeV. This
suggests that contributions from `distant cuts' which cannot be seen
in any finite order of the $1/m_c$ expansion create this problem.
For a different opinion on this point, see ref.\cite{CHERNYAK}.
I draw the following somewhat tentative conclusions from these
observations:
\noindent $\bullet$ Predictions for the absolute size of charm hadron
lifetimes cannot be trusted.
\noindent $\bullet$ However it is quite conceivable that lifetime
{\em ratios} do not suffer from such a fundamental uncertainty due
to not-so-distant cuts. I will then explore the working hypothesis
that the
ratio of lifetimes can be trusted in principle -- though in
practise only with a quite considerable grain of salt!
The quark level Lagrangian is again well known. On the
Cabibbo-allowed level there is now a single
non-leptonic transition described by
$${\cal L}_W^{\Delta C=1}(\mu = m_c)=
\frac{4G_F}{\sqrt{2}} V_{cs}V_{ud}^*
\{ c_1(\bar s_L\gamma _{\mu}c_L)(\bar u_L\gamma _{\mu}d_L)
+c_2(\bar u_L\gamma _{\mu}c_L)(\bar s_L\gamma _{\mu}d_L)
\} \eqno(66)$$
where the short-distance coefficients $c_{1,2}$ are now evaluated
at a lowerJscale than in beauty decays yielding
$$ c_1(LL+NLL)\simeq 1.32 \, , \; \;
c_2(LL+NLL)\simeq - 0.58 \eqno(67)$$.
The expressions for the semileptonic and non-leptonic
decay widths of charm hadrons $H_c$
are quite analogous to the ones for beauty hadrons
(though simpler since there is only one non-leptonic
decay class rather than two); through order $1/m_c^2$
they are given by
$$\frac{\Gamma _{SL,decay}(H_c)}{\Gamma _0}=
\matel{H_c}{\bar cc}{H_c}_{norm}\cdot
\left[ I_0(x,0,0) +
\frac{\aver{\mu _G^2}_{H_c}}{m_c^2}
(x\frac{d}{dx}-2)I_0(x,0,0)\right] \; , \eqno(68a)$$
$$\frac{\Gamma _{NL,decay}(H_c)}{\Gamma _0} = N_C\cdot
\matel{H_c}{\bar cc}{H_c}_{norm}\cdot
\left[A_0[I_0(x,0,0) +
\frac{\aver{\mu _G^2}_{H_c}}{m_c^2}
(x\frac{d}{dx}-2)I_0(x,0,0)] - \right.$$
$$\left. 8A_2 \frac{\aver{\mu _G^2}_{H_c}}{m_c^2}
\cdot
I_2(x,0,0,)\right]\; . \eqno(68b)$$
where now
$$\Gamma _0 \equiv
\frac{G_F^2m_c^5}{192 \pi ^3}|V(cs)|^2 \; , \; \;
x=\frac{m_s^2}{m_c^2}\eqno(68c)$$
and the radiative corrections lumped into $A_0$
and $A_2$ are given by the appropriate values for
$c_+$ and $c_-$, see eq. (67). With $x\sim 0.012$ one finds:
$$I_0(x,0,0)|_{x=0.012}\simeq 0.91 \; \; , \; \;
I_2(x,0,0)|_{x=0.012}\simeq 0.96 \; , $$
i.e. there is much less phase space suppression than for
$b\rightarrow c$ transitions.
\subsection{$D^+$ vs. $D^0$ Lifetimes}
Analogously to the case of $\tau (B^-)$ vs. $\tau (B_d)$ the
$D^+$ and $D^0$ lifetimes get first differentiated in
order $1/m_c^3$
when PI and WA intervene. Both affect the lifetime
ratio in the same direction, namely they enhance
$\tau (D^+)$ over $\tau (D^0)$ with the
{\em destructive interference} due
to PI being the {\em dominant} effect. Quantitatively one finds through
order $1/m_c^3$ by employing eqs.(48,49) with the appropriate
substitutions and using $f_D \sim 200$ MeV:
$$\frac{\tau (D^+)}{\tau (D^0)} \sim 2 \eqno(69)$$
For proper perspective one has to keep four observations in mind:
\noindent (i) While the expectation expressed in eq.(69) does not
coincide numerically with the measured value --
$\tau (D^+) /\tau (D^0) = 2.547 \pm 0.043$ -- it agrees with it to
within $\sim 25$ \%. Such a deviation could be ascribed to
$1/m_c^4$ contributions ignored in eq.(69).
\noindent (ii) On the other hand, as discussed before, there is reason
to doubt the reliability and thus validity of heavy quark expansions
for treating nonperturbative corrections in
{\em charm} decays. Yet I
adopt, as already stated, the working hypothesis that a heavy quark
expansion can be employed for treating the {\em ratios} of lifetimes,
though not the lifetimes themselves. Such a conjecture is tested
a posteriori by analysing the whole pattern of the expected and the
observed lifetimes of the various charm hadrons.
\noindent (iii) Even the perturbative corrections
contain sizeable numerical uncertainties. To cite but one glaring
example: the size and the nature of the socalled `hybrid'
renormalization reflecting dynamics between the scales
$m_c$ and $\mu _{had} \sim 0.5 - 1$ GeV is quite important
quantitatively (they considerably enhance the destructive
interference in $D^+$ decays.) Yet a leading-log treatment of those
corrections seems woefully inadequate for dealing with such a small
slice in momentum space.
\noindent (iv) One should note also the following: while WA plays
only a relatively
minor role in inclusive rates (generating only a 10 \% or so
difference in $D^+$ vs. $D^0$ lifetimes), it is likely to play a
considerably more significant role in {\em exclusive} modes.
\subsection{$D_s^+$ vs. $D^0$ Lifetimes}
Since the impact of WA is reduced relative to that of PI in meson
decays, it is natural to compare $\tau (D_s)$ to
$\tau (D^0)$ rather than to $\tau (D^+)$. Such a comparison
-- and of their
semileptonic branching ratios -- touches upon several intriguing
dynamical issues. A priori $\tau (D_s)$ and $\tau (D^0)$ could
differ substantially from each other due to $SU(3)_{Fl}$
breaking and in particular due to a different weight of WA
in the two cases. Yet the heavy quark expansion
strongly suggests
those two lifetimes to be close to each other.
The analysis proceeds in several steps \cite{DS}.
The two operators
contributing in order $1/m_c^2$ are singlets under the
light quark flavours; yet $SU(3)_{Fl}$ breaking could enter
through their expectation values. For the chromomagnetic
operator one has:
$$\frac{1}{M_D}\matel{D^0}{\bar c i\sigma \cdot G c}{D^0}
\simeq \frac{3}{2} (M^2_{D^{*0}}-M^2_{D^0})\eqno(70)$$
$$\frac{1}{M_{D_s}}\matel{D_s}{\bar c i\sigma \cdot G c}{D_s}
\simeq \frac{3}{2} (M^2_{D_s^*}-M^2_{D_s})\eqno(71)$$
Since the measured values for the $D-D^*$ and the $D_s-D_s^*$
mass splittings are practically identical
( $M_{D^{*0}} - M_{D^0}=142.12 \pm 0.07$ MeV and
$M_{D_s^*} - M_{D_s}=141.6 \pm 1.8$ MeV ), the chromomagnetic
operator cannot induce an appreciable difference between
$\tau (D^0)$ and $\tau (D_s)$.
\footnote{The observation that the hyperfine splitting is largely
independant of the flavour of the spectator (anti)quark can be
understood intuitively in quark models.}
Through order $1/m_c^3$ there are four distinct
sources for a difference in $\Gamma (D_s)$ vs.
$\Gamma (D^0)$ exceeding the 1\% level:
(a) The decay $D_s\rightarrow \tau \nu$; (b) PI in those Cabibbo suppressed
$D_s$ decays that are driven by the quark level transition
$c\rightarrow s\bar su$; (c) the effects of $SU(3)_{Fl}$ breaking on the
expectation values of the kinetic energy operator;
(d) WA in nonleptonic $D^0$ and in nonleptonic as well as
semileptonic $D_s$ decays. Corrections listed under
(a), (b) and (d) are generated by dim-6 operators
whereas the much less familiar correction (c) is derived from a
dim-5 operator.
\noindent (a) The width for the decay $D_s\rightarrow \tau \nu$ is
completely determined in terms of the axial decay constant for the
$D_s$ meson:
$$\Gamma (D_s) = \frac{G_F^2m^2_{\tau}f^2_{D_s}M_{D_s}}
{8\pi}|V(cs)|^2(1-m^2_{\tau}/M^2_{D_s})^2 . \eqno(72)$$
For $f_{D_s}\simeq 210$ MeV one gets
$$\Gamma (D_s\rightarrow \tau \nu ) \simeq 0.03 \Gamma (D^0) .
\eqno(73)$$
This effect necessarily {\em reduces} $\tau (D_s)$ relative to
$\tau (D^0)$.
\noindent (b) In $D_s$ decays PI appears in the
Cabibbo-suppressed
$c\rightarrow s \bar su$ channel. Its weight is expressed in terms of the
matrix elements of the two four-fermion operators
$$\matel{D_s}{(\bar c_L\gamma _{\mu}s_L)
(\bar s_L\gamma _{\mu}c_L)}{D_s}, \; \;
\matel{D_s}{(\bar c_L\gamma _{\mu}\lambda ^as_L)
(\bar s_L\gamma _{\mu}\lambda ^ac_L)}{D_s} \eqno(74)$$
with known coefficients that are computed
perturbatively (including the `hybrid' renormalization of these
operators
down from the scale $m_c$). The effect of PI is
most reliably estimated from the observed difference in the
$D^+$ and $D^0$ widths. It is easy to see that the structure of
the operators responsible for PI in $D_s$ decays is exactly the
same as in $D^+$ decays if one replaces the $d$ quark with the
$s$ quark and adds the extra factor $\tan ^2\theta _C$; it is then
{\em destructive} as well. Thus one arrives at:
$$\delta \Gamma _{PI}(D_s) \simeq S\cdot \tan ^2\theta _C
(\Gamma (D^+)-\Gamma (D^0)) \simeq -S\cdot 0.03
\Gamma (D^0)\eqno(75)$$
where the factor $S$ has been introduced to allow for $SU(3)_{Fl}$
breaking in the expectation values of the four-fermion operators.
The factor $S$ is expected to exceed unity somewhat; in an
factorization ansatz it is given by the ratio
$f^2_{D_s}/f^2_D \simeq 1.4$, see eq.(39). Thus
$$ \delta \Gamma _{PI}(D_s) \simeq - 0.04 \Gamma (D^0) \, .
\eqno(76)$$
\noindent (c) The impact of the
{\em chromomagnetic}
operator on the $D_s-D^0$ lifetime ratio can be derived
from the hyperfine splitting. Inserting the observed meson
masses one obtains
$$ \matel{D^0}{\bar c \frac{i}{2} \sigma \cdot G c}{D^0}_{norm}
\simeq 0.413 \pm 0.002 \eqno(77a)$$
$$ \matel{D_s}{\bar c \frac{i}{2} \sigma \cdot G c}{D_s}_{norm}
\simeq 0.433 \pm 0.006 \eqno(77b)$$
from which one can conclude: there is -- not surprisingly --
very little $SU(3)_{Fl}$ breaking in these matrix elements and
a sizeable difference in $\tau (D_s)$ vs. $\tau (D^0)$
{\em cannot} arise from this source.
\noindent The second dim-5 operator generating
$1/m_c^2$ corrections is the kinetic energy operator
$\bar c (i\vec D)^2c$ where $\vec D$ denotes the covariant
derivative. Its expectation value describes the largely
non-relativistic
motion of the charm quark in the gluon background field
inside the charm
hadron. It reflects Lorentz
time dilation and thus prolongs the {\em hadron} lifetime. On
general grounds one expects it to extend $\tau (D_s)$ over
$\tau (D^0)$, as seen as follows. The spatial wavefunction should
be more concentrated around the origin for $D_s$ than for
$D^0$ mesons. This implies, via the uncertainty principle, the
mean value of $(\vec p_c)^2$ to be larger for $D_s$ than for
$D^0$; in other words the charm quark undergoes more Fermi
motion as constituent of $D_s$ than of $D$ mesons. The lifetime of
the charm quark is then prolonged by time dilation to a higher
degree inside $D_s$ than inside $D^0$ mesons.
\noindent While the trend of this effect is quite transparent, its
size is not yet clear. The relevant matrix elements
$\matel{D_s}{\bar c (i\vec D)^2 c}{D_s}$ and
$\matel{D}{\bar c (i\vec D)^2 c}{D}$ will be determined from
QCD sum rules and lattice simulations in the future. Yet in the
meantime one can estimate their size from the meson
masses in the charm and beauty family according to the following
prescription:
$$ \matel{D_s}{\bar c (i\vec D)^2 c}{D_s} -
\matel{D}{\bar c (i\vec D)^2 c}{D} \simeq
\frac{2m_bm_c}{m_b-m_c}
\{ [\aver{M_{D_s}}-\aver{M_D}]-
[\aver{M_{B_s}}-\aver{M_B}]\}
\eqno(78)$$
where as before $\aver{M_{D,D_s,B,B_s}}$ denote the
spin averaged masses. Accordingly one finds
$$\frac{\Delta \Gamma (D_s/D)}{\bar \Gamma}\equiv
\frac{\Gamma _{Fermi}(D_s)- \Gamma _{Fermi}(D)}
{\bar \Gamma } \simeq $$
$$ \simeq -\frac{m_b}{m_c(m_b-m_c)}
\times \{ [\aver{M_{D_s}}-\aver{M_D}]-
[\aver{M_{B_s}}-\aver{M_B}]\} \, .\eqno(79)$$
Since the $B_s^*$ mass has not been measured yet, one
cannot give a specific numerical prediction and has
to content oneself with semi-quantitative statements.
A 10 MeV shift in any of the terms $\aver{M}$
corresponds numerically to the kinetic energy operator
generating approximately a 1\% change in the ratio
$\tau (D_s)/\tau (D^+)$. Invoking our present understanding
of the heavy quark kinetic energy and its relationship to
the hyperfine splitting one arrives at the following
{\em conjecture}:
$$ \frac{\Delta \Gamma (D_s/D)|_{(a)+(b)+(c)}}{\bar \Gamma}
\sim {\cal O}(+\, few\, \% )\eqno(80)$$
\noindent (d) The mechanisms listed above in
(a) - (c) taken together can be expected to extend the
$D_s$ over the $D^0$ lifetime by at most a few percent.
Comparing that estimate with the data allows us to
conclude that WA cannot change $\tau (D_s)/\tau (D^0)$
by more than 10 \% . It is however quite unclear
how to refine this estimate at present. For the quantitative
impact of WA on charm
meson lifetimes in general and on the ratio
$\tau (D_s)/\tau (D^0)$ in particular
is the most obscure theoretical element in the
analysis. The uncertainty centers
mainly on the question of how much the WA amplitude suffers
from helicity suppression.
\noindent In the valence quark approximation
the answer is easily given to {\em lowest} order
in the strong coupling: the WA
amplitude is (helicity) suppressed by the ratio
$m_q/m_c$, where $m_q$ denotes the largest quark mass
in the final state. For a proper QCD treatment one has to use
the {\em current} rather than the larger
{\em constituent} mass, at least for
the $1/m_c^3$ corrections. The WA amplitude is then small in
$D^0$ decays where the helicity factor reads
$m_s/m_c \sim 0.1$ and a fortiori in $D_s$ decays where
only non-strange quarks are present in the final state.
The emission of semi-hard gluons that is included by
summing the leading log terms in the perturbative expansion
cannot circumvent this suppression \cite{MIRAGE}. For such gluon
corrections -- when properly accounted for -- drive the hybrid
renormalization of the corresponding four-fermion operators;
however they preserve the Lorentz structure of the lowest order
term and therefore do not eliminate the helicity suppression.
A helicity allowed amplitude arises in perturbation theory only at
the subleading level of order $\alpha_s (m_c^2)/\pi$ and is thus expected
to be numerically insignificant.
\noindent On the other hand {\em nonperturbative} dynamics
can quite naturally vitiate helicity suppression and thus provide
the dominant source of WA. These nonperturbative effects enter
through {\em non-factorizable} contributions to the hadronic
matrix elements, as analyzed in considerable detail in
refs.\cite{WA,DS}. As such we do not know (yet), how to predict
their weight from first principles. However, as shown in
ref.\cite{WA}, a detailed experimental study of the width of
semileptonic decays and their lepton spectra -- in particular
in the endpoint region -- in $D^0$ vs. $D_s$ and/or in $B^0$
vs. $B^+$ decays would allow us to extract the size of the
matrix elements that control the weight of WA in all
inclusive $B$ and $D$ decays. Before such data become
available, we can draw only qualitative, or at best
semi-quantitative conclusions: WA is not expected to
affect the total lifetimes of $D_s$ and $D^0$ mesons by more
than 10 - 20 \%, and their ratio by less.
Furthermore WA does not
necessarily decrease $\tau (D_s)/\tau (D^0)$; due to its
reduced amplitude its leading impact on the lifetime might
be due to its interference with the spectator amplitude and thus
it might even {\em enhance} $\tau (D_s)/\tau (D^0)$!
To summarize our findings on the $D_s-D^0$ lifetime ratio:
(i) $SU(3)_{Fl}$ breaking in the expectation values of the
dim-5 operators generating the leading nonperturbative
corrections of order $1/m_c^2$ can -- due to
`time dilatation' -- increase $\tau (D_s)$ by 3-5 \% over
$\tau (D^0)$. (ii) On the $1/m_c^3$ level there arise
three additional effects. Destructive interference in
Cabibbo suppressed $D_s$ decays increases $\tau (D_s)$
again by 3-5 \%, whereas the single mode
$D_s\rightarrow \tau \nu$ decreases it by 3 \%. The three
phenomena listed so far combine to yield
$\tau (D_s)/\tau (D^0) \simeq 1.0 - 1.07$.
(iii) Any difference over and above that has to be
attributed to WA. Therefore one has to interprete the
measured $D_s-D^0$ lifetime ratio as more or less direct
evidence for WA to contribute no more than 10-20 \% of
the lifetime ratio between charm mesons. This is consistent
with the indirect evidence discussed above that WA does not
constitute the major effect there. (iv) These predictions can
be refined in the future by two classes of more accurate data:
analyzing the {\em difference} in the
semileptonic spectra of charged and neutral mesons in the charm
and beauty family allows to extract the size of the matrix elements
that control the weight of WA; measuring the masses of
$\Lambda _b$ and $B_s^*$ to better than 10 MeV allows to
determine the expectation values of the kinetic energy
operator.
\subsection{Charm Baryon Lifetimes}
The lifetimes of the weakly decaying charm baryons reflect a
complex interplay of different dynamical mechanisms
increasing or decreasing transition rates
\cite{MARBELLA,BARYONS3}:
$$\Gamma (\Lambda _c^+)= \Gamma _{decay}(\Lambda _c^+) +
\Gamma _{WS}(\Lambda _c^+) -
|\Delta \Gamma _{PI,-}(\Lambda _c)| \eqno(81a)$$
$$\Gamma (\Xi _c^0)= \Gamma _{decay}(\Xi _c^0) +
\Gamma _{WS}(\Xi _c^0) +
|\Delta \Gamma _{PI,+}(\Xi _c^0)| \eqno(81b)$$
$$\Gamma (\Xi _c^+)= \Gamma _{decay}(\Xi _c^+) +
|\Delta \Gamma _{PI,+}(\Xi _c^+)| -
|\Delta \Gamma _{PI,-}(\Xi _c^+)| \eqno(81c)$$
The explicit expressions for $\Gamma _{decay}$, $\Gamma _{WS}$
and $\Delta \Gamma _{PI,-}$ are obtained from eqs.(60) by the
obvious substitutions; for $\Delta \Gamma _{PI,+}$ one finds
$$\Delta \Gamma _{PI,+}(\Xi _c)\equiv
\frac{G_F^2|V(cs)|^2m_c^2}{4\pi}
c_+(2c_- + c_+)
\matel{\Xi _c}
{\bar c_L\gamma _{\mu}c\bar s_L\gamma _{\mu}s_L +
\frac{2}{3}\bar c\gamma _{\mu}\gamma _5 c
\bar s_L\gamma _{\mu}s_L}
{\Xi _c}_{norm} \eqno(82)$$
On rather general grounds one then concludes:
$$\tau (\Xi _c^0) < \tau (\Xi _c^+)\; , \; \; \;
\tau (\Xi _c^0) < \tau (\Lambda _c^+) \eqno(83)$$
To go further requires calculating the relative weight of
matrix elements of the four-quark operators.
At present this can be done only by
using quark wavefunctions as obtained from a potential ansatz
(or from the quark model). In doing so one also has to
pay proper attention to the normalization point appropriate
for such an evaluation, i.e., one has to include ultraviolet as
well as hybrid renormalization. One then arrives at the
numbers quoted in Table 2.
It has been argued \cite{MARBELLA}
that in these expressions one should not
use the `real' value for the decay constant, as expressed by
$f_D$, and the
observed mass splitting, but solely the leading contribution
in a $1/m_Q^3$ expansion: $F_D \sim 400$ MeV and
$M_{\Sigma ^+_c} - M_{\Lambda _c}\sim 400$ MeV.
One reason for that is self-consistency since the widths have
been calculated through order $1/m_Q^3$ only; the other --
and maybe the more telling one -- is based on the needs of
phenomenology: for otherwise one cannot reproduce the
observed {\em magnitude} of the lifetime differences.
The predictions
given in Table 2 agree remarkably well with the data
within the experimental and theoretical uncertainties.
\footnote{One should point out that those predictions
were made before these data became available; data at that
time suggested considerably larger lifetime ratios.}
In the comparison of the $\Lambda _c$ and $D$ lifetimes
one should note the following: WS and PI {\em counteract}
each other in $\Gamma (\Lambda _c)$, see eq.(81a);
nevertheless $\Gamma (\Lambda _c)\sim 2\cdot \Gamma (D)$
can be obtained. This results from three effects:
(i) WA is very much reduced in $D$ decays. (ii) The
{\em baryonic} expectation values of the four-quark
operators are evaluated, as stated above, with the
`static' decay constant, which is considerably larger
than $f_D$ used for the {\em mesonic} expectation values.
(iii) The $1/m_c^2$ contributions enhance
$\Gamma (\Lambda _c)$ over $\Gamma (D)$, as briefly discussed
below.
Unfortunately there appears now a rather unpleasant
`fly in the ointment': as discussed above, extrapolating from
these prima facie successful predictions to the lifetimes of
beauty baryons leads to a less-than-successful prediction on
$\tau (\Lambda _b)/\tau (B_d)$.
The decays of $\Omega _c$ baryons require -- and deserve --
a separate treatment: the $ss$ diquarks carry spin one
and the resulting spin-spin interactions of $c$ with $ss$
lead to new effects. To be more specific: through order
$1/m_c^3$ one has
$$\Gamma (\Omega _c) = \Gamma _{decay}(\Omega _c) +
|\Delta \Gamma _{PI,+}(\Omega _c)| \eqno(84)$$
with both quantities on the right-hand-side of eq.(84)
differing from the corresponding ones for $\Lambda _c$
or $\Xi _c$ decays.
(i) Firstly,
$\matel{\Omega _c}{\bar c i\sigma \cdot G c}{\Omega _c}\neq 0
= \matel{\Lambda _c}{\bar c i\sigma \cdot G c}{\Lambda _c}$, see
eqs.(17b,17c,35). Secondly, it is quite conceivable that
these spin-spin interactions can create an
appreciable difference in the
kinetic energy of the $c$ quark inside $\Omega _c$ and
$\Lambda _c$ baryons. Thus there arise differences of order
$1/m_c^2$ in the total as well as in the
semileptonic widths of $\Omega _c$ and $\Lambda _c$ baryons:
$$\Gamma _{decay}(\Omega _c) \neq \Gamma _{decay}(\Lambda _c)
\; \; , \; \;
\Gamma _{SL,decay}(\Omega _c) \neq \Gamma _{SL,decay}(\Lambda _c)
\eqno(85)$$
(ii) These spin-spin interactions also
affect the expectation
values of the dimension-six four-fermion operators that control
the strength of WS and the interference effects, see eqs.(24,25),
in order $1/m_c^3$.
\noindent Taking everything together one estimates the
$\Omega _c$ to be the shortest lived charm baryon, as
shown in Table 2. This is quite remarkable since it means that
the constructive interference in $\Omega _c$ decays outweighs
the {\em combined} effect of constructive interference and
WS in $\Xi _c^0$ decays. It is actually not unexpected on
intuitive grounds: in $\Omega _c$ transitions interference
can happen with {\em both} constituent $s$ quarks whereas in
$\Xi _c$ baryons there is only one such $s$ quark.
This shows that inclusive weak decay rates
can be harnessed to probe the internal structure
of charm baryons in a novel way.
\subsection{Semileptonic Branching Ratios of Charm
Hadrons}
As expected, the ratio of the semileptonic $D^+$ and $D^0$
branching ratios agrees, within the errors, with the ratio
of their lifetimes:
$$2.23 \pm 0.42=
\frac{BR_{SL}(D^+)}{BR_{SL}(D^0)}\simeq
\frac{\tau (D^+)}{\tau (D^0)}=\, 2.547\, \pm \, 0.043
\eqno(86)$$
Also the {\em absolute} size of, say, $BR_{SL}(D^0)$ is of
interest. If the $D^+ - D^0$ lifetime difference is
generated mainly by a destructive interference
affecting $D^+$ transitions, then $D^0$ decays should
proceed in a largely normal way and the
semileptonic branching ratio of $D^0$ mesons should be at its
`normal' value. In a parton model description one finds
$BR(D\rightarrow l\nu X_s) \equiv BR(c\rightarrow l\nu s) \sim
16\%$. If that number represented the proper
yardstick for normal decays, one would actually conclude
that $D^+$ decays more or less normally, whereas the
nonleptonic decays of $D^0$ are strongly enhanced; i.e.,
it would imply that it is actually WA that provides the
dominant mechanism for the lifetime difference. This
long-standing `fly in the ointment' can be removed now.
For at order $1/m_Q^2$ the chromomagnetic operator
$\bar Qi\sigma \cdot G Q$ generates an isoscalar enhancement in the
nonleptonic widths of the charged and neutral mesons
\cite{BUV}.
It turns out that the corresponding reduction in the
semileptonic branching ratios of $B$ mesons is rather
insignificant, namely $\sim$ 2 \% . Yet for charm mesons
this reduction is quite large, namely around
40 \%, since it is
scaled up by $(m_b/m_c)^2$ and less reduced
by colour factors. The semileptonic branching
ratio for $D$ mesons is thus around 9 \% through
order $1/m_c^2$ -- before the dimension-six
four-fermion operators reduce
$\Gamma _{NL}(D^+)$ by a factor of almost two and
enhance $\Gamma _{NL}(D^0)$ by a moderate amount.
As discussed above, we have to take these numbers for
charm transitions with a grain of salt; on the other hand
the emerging pattern in the ratio of semileptonic branching
ratios as well as their absolute size is self-consistent!
It should be apparent from the discussion on $\tau (D_s)$
vs. $\tau (D^0)$ that the semileptonic widths and branching
ratios of $D_s$ and $D^0$ mesons practically coincide
through order $1/m_c^2$. In order $1/m_c^3$ WA
can contribute to $\Gamma _{SL}(D_s)$ (as well as to
$\Gamma _{NL}(D_s)$ and $\Gamma _{NL}(D^0)$).
Yet from $\tau (D_s)\simeq \tau (D^0)$ one infers that the
relative weight of WA is quite reduced; therefore one expects
$BR_{SL}(D_s)$ to differ from $BR_{SL}(D^0)$ by not more
than 10\% or so in {\em either} direction:
$$1\, \pm \sim {\rm few}\% = \frac{\tau (D_s)}{\tau (D^0)}
\simeq \frac{BR_{SL}(D_s)}{BR_{SL}(D^0)}=1\, \pm \sim 10\%
\eqno(87)$$
The semileptonic width is not universal for all
charm hadrons due to ${\cal O}(1/m_c^2)$
contributions even when one ignores Cabibbo suppressed
channels. One actually estimates
$$ \Gamma _{SL}(D) / \Gamma _{SL}(\Lambda _c) /
\Gamma _{SL}(\Omega _c) \, \sim \,
1/1.5/1.2 \, .\eqno(88)$$
Therefore the ratio of
semileptonic branching ratios of mesons and
baryons does not faithfully reproduce the ratio of their
lifetimes. Instead one expects
$$BR_{SL}(\Lambda _c) > BR_{SL}(D^0)\times
\tau (\Lambda _c)/\tau (D^0)
\simeq 0.5\, BR_{SL}(D^0) \, . \eqno(89)$$
\section{What Have We Learned and What Will We Learn?}
Second generation theoretical technologies have been developed for
treating heavy-flavour decays that are directly related to QCD
without the need for invoking a `deus ex machina'. As far as
{\em inclusive} heavy-flavour decays are concerned, the relevant
technology is based on an operator expansion in inverse powers
of the heavy quark mass \footnote{To repeat it one last time: This
is {\em not} HQET!}. It allows to express the leading nonperturbative
corrections through the expectation values of a small number of
dimension-five and -six operators. Basically all such matrix
elements relevant for {\em meson} decays can reliably be related to
other
observables; this allows to extract their size in a
model-independant way.
Baryons, however, possess a more complex internal
structure, which becomes relevant for their decays
in order $1/m_Q^3$. At present one has
to rely on quark model calculations to determine
the expectation values of the dimension-six operators relevant for
lifetime differences. The numerical results of such computations
are of dubious reliability; accordingly predictions for lifetime ratios
involving heavy-flavour {\em baryons} suffer from larger
uncertainties than those involving only mesons. There is
also a positive side to this, though: measuring the lifetimes (and if
at all possible also the semileptonic branching ratios) of the
various heavy-flavour baryons sheds new light onto their
internal strong dynamics. Comprehensive studies of inclusive
weak decays in the charm baryon family are particularly
promising from a practical point
for such studies on hadronic structure; for the
pre-asymptotic corrections are much larger here than
for beauty baryons.
At present there exists one rather glaring phenomenological
problem for the heavy quark expansion and -- not surprisingly, as
just indicated -- it concerns baryon decays: the observed
$\Lambda _b$ lifetime is clearly
shorter than predicted relative to the $B_d$ lifetime.
Unless future measurements move it up significantly, one
has to pay a theoretical price for that failure. As stated above,
for the time being one
has to employ a quark model to determine the size of the
baryonic expectation values of the relevelant dimenion-six
operators. It is not surprising per se that quark model computations
can yield numerically incorrect results, in particular when sizeable
cancellations occur between different contributions. The more
serious problem is provided by the following aspects: the
analoguous treatment of charm baryons, using the same quark
models, had yielded predictions
that give not only the correct pattern, but also numerical values for
the lifetime ratios which are surprisingly close to the data
considering the uncertainties due to higher order terms in the
$1/m_c$ expansion. Furthermore scaling these charm lifetime
ratios up to the beauty scale, using a $1/m_Q^2$ and $1/m_Q^3$
behaviour for the pre-asymptotic corrections from the
dimension-five and -six operators, respectively, also tells us
that the $\Lambda _b$ lifetime should be shorter than the
$B_d$ lifetime by no more than 10 \%.
To the degree that the observed value for
$\tau (\Lambda _b)/\tau (B_d)$ falls below 0.9 one has to draw the
following conclusion: one cannot trust
the numerical results of quark model
calculations for {\em baryonic} matrix elements -- not very
surprising by itself; yet furthermore and more seriously
it would mean that $1/m_Q^4$ or even higher order
contributions are still relevant in charm baryon decays before
fading away for beauty decays.
Then one had
to view the apparently successful predictions on the lifetimes
of charm baryons as largely coincidental -- a quite sobering
result!
A {\em future} discrepancy between the predictions on
$\tau (B^+)/\tau (B_d)$ or
$\tau (B_d)/\bar \tau (B_s)$ and the data \footnote{It has to be
stressed again that $\bar \tau (B_s)$ refers to the algebraic
average of the two $B_s$ lifetimes.} -- in particular an
observation that the lifetime for $B^+$ mesons is definitely
{\em shorter}
than for $B_d$ mesons -- would have quite fundamental
consequences. For the leading deviation of these ratios from
unity arises at order $1/m_b^3$ and should provide a good
approximation since the expansion parameter is small:
$\mu _{had}/m_b \sim 0.1$. The size of this term is given by
the expectation value of a four-fermion operator expressed in terms
of $f_B$. A failure in this simple situation would
raise very serious doubts about the validity or at least
the practical relevance
of the $1/m_Q$ expansion for treating even fully
inclusive nonleptonic transitions; this would leave only
semileptonic transitions in the domain of their applicability.
Such a breakdown of quark-hadron duality would
{\em a priori} appear
as a quite conceivable and merely disappointing
scenario. However such an
outcome would have to be seen as quite puzzling
{\em a posteriori}; for in our
analysis we have not discerned any sign indicating the existence
of such a fundamental problem or a qualitative distinction between
nonleptonic and semileptonic decays \cite{BLOKMANNEL,OPTICAL}.
Thus even a failure would
teach us a valuable, albeit sad lesson about the intricacies of the
strong interactions; for the heavy quark expansion is
directly and unequivocally based on QCD with the only additional
assumption concerning the workings of
quark-hadron duality!
The theoretical analysis of the lifetimes of heavy-flavour hadrons
can be improved, refined and extended:
\noindent $\bullet$ {\em improved} by a better
understanding of
quark-hadron duality, preferably by deriving it from QCD or
at least by analysing how it operates in the lepton and photon
spectra of semileptonic and radiative $B$ decays, respectively;
\noindent $\bullet$ {\em refined} by a reliable determination of
in particular, but not only, the baryonic expectation values of
the relevant dimension-six operators;
\noindent $\bullet$ {\em extended} by treating $\Xi _b$ (and
$\Omega _b$) decays.
The extension will certainly have been made very soon;
the refinement should be obtained in the
foreseeable future, probably
from a lattice simulation of QCD; when finally
quark-hadron duality will be derived from QCD cannot be
predicted, of course.
There is a corresponding list of future measurements that are
most likely to probe and advance our understanding:
\noindent $\bullet$ measure $\tau (\Lambda _b)$,
$\tau (\Xi ^- _b)$ and $\tau (\Xi ^0 _b)$
{\em separately} with good accuracy;
\noindent $\bullet$ confirm $\tau (B_d)\simeq
\bar \tau (B_s)$ within an accuracy of very few percent;
\noindent $\bullet$ verify that $\tau (B^+)$ {\em exceeds}
$\tau (B_d)$ by a few to several percent;
\noindent $\bullet$ in charm decays determine the lifetimes
of $\Xi _c^{0,+}$ and $\Omega _c$ baryons with a few percent
accuracy and measure $\tau (D_s)/\tau (D^0)$ with 1\%
precision.
\noindent {\bf Acknowledgements:} I am deeply indebted to
my collaborators B. Blok, M. Shifman, N. Uraltsev
and A. Vainshtein for sharing their insights with me.
Prodding from many experimental colleagues, in
particular S. Paul and V. Sharma, has been instrumental in
making
me face reality (I think). I have been fortified by many excellent
chinese dinners prepared late at night by Bai-Ju's in South Bend.
|
\section{Introduction}
After the celebrated discovery of the top quark \cite{abe}, the Higgs boson is
the last missing link in the Standard Model (SM).
The detection of this particle and the study of its characteristics are among
the prime objectives of present and future high-energy colliding-beam
experiments.
Following Bjorken's proposal \cite{bjo}, the Higgs boson is currently being
searched for with the CERN Large Electron-Positron Collider (LEP1) and the
SLAC Linear Collider (SLC) via $e^+e^-\to Z\to f\bar fH$.
At the present time, the failure of this search allows one to rule out the
mass range $M_H\le64.3$~GeV at the 95\% confidence level \cite{jan}.
The quest for the Higgs boson will be continued with LEP2 by exploiting the
Higgs-strahlung mechanism \cite{ell,iof}, $e^+e^-\to ZH\to f\bar fH$.
In next-generation $e^+e^-$ linear supercolliders (NLC), also
$e^+e^-\to\bar\nu_e\nu_eH$ via $W^+W^-$ fusion and, to a lesser extent,
$e^+e^-\to e^+e^-H$ via $ZZ$ fusion will provide copious sources of
Higgs bosons.
Once a novel scalar particle is discovered, it will be crucial to decide if it
is the very Higgs boson of the SM or if it lives in some more extended Higgs
sector.
To that end, precise knowledge of the SM predictions will be mandatory,
{\it i.e.}, quantum corrections must be taken into account.
The status of the radiative corrections to the production and decay processes
of the SM Higgs boson has recently been summarized \cite{pr}.
Since the top quark is by far the heaviest established elementary particle,
with a pole mass of $M_t=(180\pm12)$~GeV \cite{abe}, the leading high-$M_t$
terms, of ${\cal O}(G_FM_t^2)$, are particularly important, and it is desirable
to acquire information on their quantumchromodynamical (QCD) corrections.
During the last year, a number of papers have appeared in which the two-loop
${\cal O}(\alpha_sG_FM_t^2)$ corrections to various Higgs-boson production and
decay processes are presented.
These processes include
$H\to f\bar f$, with $f\ne b$ \cite{hll} and $f=b$ \cite{ks1,kwi},
$Z\to f\bar fH$ and $e^+e^-\to ZH$ \cite{ks2},
$e^+e^-\to\bar\nu_e\nu_eH$ via $W^+W^-$ fusion \cite{ks3},
$gg\to H$ \cite{ks3,gam},
and more \cite{ks3}.
In this paper, we shall take the next step and tackle with three-loop
${\cal O}(\alpha_s^2G_FM_t^2)$ corrections.
To keep matters as simple as possible, we shall restrict our considerations to
light Higgs bosons, with $M_H\ll M_t$, and to reactions with colourless
particles in the initial and final states.
Such reactions typically involve the $\ell^+\ell^-H$, $W^+W^-H$, and $ZZH$
couplings together with gauge couplings of the $W$ and $Z$ bosons to leptons.
We are thus led to incorporate the next-to-leading QCD corrections in the
low-$M_H$ effective $\ell^+\ell^-H$, $W^+W^-H$, and $ZZH$ interaction
Lagrangians.
This will be achieved in Section~2.
Recently, the ${\cal O}(\alpha_s^2G_FM_t^2)$ correction to $\Delta\rho$ has
been calculated and found to be sizeable \cite{avd}.
This is relevant for present and future precision tests of the standard
electroweak theory.
It is of great theoretical interest to find out whether the occurrence of
significant ${\cal O}(\alpha_s^2G_FM_t^2)$ corrections is specific to
$\Delta\rho$ or whether this is a common feature in the class of electroweak
observables with a quadratic $M_t$ dependence at one loop.
In the latter case, there must be some underlying principle which is able to
explain this phenomenon.
Our analysis will put us into a position where we can investigate this matter
for four independent quantities.
We shall return to this issue in Section~5.
The complete evaluation of the one-loop electroweak correction to a process
which involves more than four external particles is enormously intricate.
To our knowledge, the literature does not contain a single example of such a
calculation.
However, the so-called improved Born approximation (IBA) \cite{iba} allows us
to conveniently extract at least the dominant fermionic loop corrections.
As a by-product of our analysis, we shall illustrate the usefulness of the IBA
for Higgs-boson production and decay in high-energy $e^+e^-$ collisions.
The appropriate formalism will be developed in Section~3.
This paper is organized as follows.
In Section~2, we shall extend the low-$M_H$ effective $\ell^+\ell^-H$,
$W^+W^-H$, and $ZZH$ interaction Lagrangians to ${\cal O}(\alpha_s^2G_FM_t^2)$.
In the $G_F$ formulation of the electroweak on-shell scheme, knowledge of the
QCD-corrected $W^+W^-H$ coupling is sufficient to control the related four-
and five-point Higgs-boson production and decay processes which emerge by
connecting one or both of the $W$ bosons with lepton lines, respectively.
Contrariwise, the corresponding processes involving a $ZZH$ coupling receive
additional QCD corrections from the gauge sector, which we shall evaluate by
invoking the IBA in Section~3.
In Section~4, we shall quantitatively analyze the phenomenological consequences
of our results.
Section~5 contains our conclusions.
\section{Effective Lagrangians}
Throughout this paper, we shall work in the electroweak on-shell
renormalization scheme \cite{aok}, with $G_F$ as a basic parameter, and define
$c_w^2=1-s_w^2=M_W^2/M_Z^2$ \cite{sir}.
In particular, this implies that the lowest-order formulae are expressed in
terms of $G_F$, $c_w$, $s_w$, and the physical particle masses.
The self-energies of the $W$, $Z$, and Higgs bosons to
${\cal O}(\alpha_s^2G_FM_t^2)$ for zero external four-momentum squared
will be the basic ingredients of our analysis.
While the results for the $W$ and $Z$ bosons are now well established
\cite{avd}, the Higgs-boson self-energy requires a separate analysis,
which will be performed here.
Our calculation will proceed along the lines of Ref.~\cite{avd}.
We shall employ dimensional regularization in $n=4-2\epsilon$ space-time
dimensions and introduce a 't~Hooft mass, $\mu$, to keep the coupling constants
dimensionless.
We shall suppress terms containing $\gamma_E-\ln(4\pi)$, where $\gamma_E$ is
Euler's constant.
These terms may be retrieved by substituting
$\mu^2\to4\pi e^{-\gamma_E}\mu^2$.
In the modified minimal-subtraction ($\overline{\mbox{MS}}$) scheme \cite{msb},
these terms are subtracted along with the poles in $\epsilon$.
This is also true for the relation between the $\overline{\mbox{MS}}$ and pole
masses of the quarks, so that these terms are also absent when the quark masses
are renormalized according to the on-shell scheme.
Since we wish to extract the leading high-$M_t$ terms, we may neglect the
masses of all virtual particles, except for the top quark.
As usual, we shall take $\gamma_5$ to be anticommuting for $n$ arbitrary.
We shall choose a covariant gauge with an arbitrary gauge parameter
for the gluon propagator.
This will allow us to explicitly check that our final results are gauge
independent.
The requirement that the expressions for physical observables be
renormalization-group (RG) invariant will serve as a further check for our
calculation.
Large intermediate expressions will be treated with the help of FORM 2.0
\cite{ver}.
The tadpole integrals which enter the one- and two-loop calculations may be
solved straightforwardly, even for arbitrary powers of propagators.
The three-loop case is more involved.
After evaluating the traces, the scalar integrals may be reduced by
decomposing the scalar products in the numerator into appropriate combinations
of the factors in the denominator.
Subsequently, recurrence relations derived using the integration-by-parts
method \cite{che} may be applied to reduce any scalar Feynman integral to a
small number of so-called master diagrams, which remain to be calculated by
hand.
More technical details may be found in Ref.~\cite{avd}.
Prior to listing our results, we shall introduce our notation.
We take the colour gauge group to be SU($N_c$);
$C_F=(N_c^2-1)/(2N_c)$ and $C_A=N_c$ are the Casimir operators of its
fundamental and adjoint representations, respectively.
As is usually done for SU($N_c$), we fix the trace normalization of the
fundamental representation to be $T_F=1/2$.
In our numerical analysis, we set $N_c=3$.
We explicitly include five massless quark flavours plus the massive top quark
in our calculation, so that we have $n_f=6$ active quark flavours altogether,
{\it i.e.}, we must not consider $n_f$ as a free parameter.
We denote the QCD renormalization scale by $\mu$.
We evaluate the strong coupling constant, $\alpha_s(\mu)$, at next-to-leading
order (two loops) in the $\overline{\mbox{MS}}$ scheme, from
\begin{equation}
\label{as}
{\alpha_s(\mu)\over\pi}={1\over\beta_0\ln(\mu^2/\Lambda_{\overline{MS}}^2)}
\left[1-{\beta_1\over\beta_0^2}\,{\ln\ln(\mu^2/\Lambda_{\overline{MS}}^2)
\over\ln(\mu^2/\Lambda_{\overline{MS}}^2)}\right],
\end{equation}
where $\Lambda_{\overline{MS}}$ is the asymptotic scale parameter appropriate
for $n_f=6$ and \cite{wil}
\begin{eqnarray}
\label{beta}
\beta_0&\hspace*{-2.5mm}=\hspace*{-2.5mm}&{1\over4}\left({11\over3}C_A-{2\over3}n_f\right)={7\over4},
\nonumber\\
\beta_1&\hspace*{-2.5mm}=\hspace*{-2.5mm}&{1\over16}\left({34\over3}C_A^2-2C_Fn_f-{10\over3}C_An_f\right)
={13\over8}
\end{eqnarray}
are the first two coefficients of the Callan-Symanzik beta function of QCD.
We define $a=4h=\alpha_s(\mu)/\pi$,
$x_t=[G_Fm_t^2(\mu)/8\pi^2\sqrt2]$,
$X_t=(G_FM_t^2/8\pi^2\sqrt2)$,
$l=\ln[\mu^2/m_t^2(\mu)]$, and
$L=\ln(\mu^2/M_t^2)$,
where $m_t(\mu)$ and $M_t$ are the $\overline{\mbox{MS}}$ and pole masses of
the top quark, respectively, and $G_F$ is Fermi's constant.
Using the two-loop relation between $m_t(M_t)$ and $M_t$ \cite{gra} along with
the RG equation for $m_t(\mu)$, we find
\begin{eqnarray}
\label{mass}
{m_t(\mu)\over M_t}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&1+hC_F(-3L-4)
+h^2C_F\left\{L^2\left({9\over2}C_F-{11\over2}C_A+n_f\right)
+L\left({21\over2}C_F-{185\over6}C_A
\right.\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.
{13\over3}n_f\right)-12\zeta(2)+6
+C_F\left[-12\zeta(3)+6\zeta(2)(8\ln2-5)+{7\over8}\right]
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.
C_A\left[6\zeta(3)+8\zeta(2)(-3\ln2+1)-{1111\over24}\right]
+n_f\left[4\zeta(2)+{71\over12}\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&1-a\left(L+{4\over3}\right)
-a^2\left({3\over8}L^2+{35\over8}L+9.125\,451\right).
\end{eqnarray}
Riemann's zeta function takes on the values $\zeta(2)=\pi^2/6$,
$\zeta(3)\approx1.202\,057$, and $\zeta(4)=\pi^4/90$.
The numerical constants \cite{avd}
\begin{eqnarray}
S_2&\hspace*{-2.5mm}=\hspace*{-2.5mm}&{4\over9\sqrt3}\mathop{{\mbox{Cl}}_2}\nolimits\left({\pi\over3}\right)
\approx0.260\,434,\nonumber\\
D_3&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&-3.027\,009,\nonumber\\
B_4&\hspace*{-2.5mm}=\hspace*{-2.5mm}&16\mathop{{\mbox{Li}}_4}\nolimits\left({1\over2}\right)-{13\over2}\zeta(4)-4\zeta(2)\ln^22
+{2\over3}\ln^42
\approx-1.762\,800,
\end{eqnarray}
where $\mathop{{\mbox{Cl}}_2}\nolimits{}$ is Clausen's function and $\mathop{{\mbox{Li}}_4}\nolimits{}$ is the quadrilogarithm,
occur in the evaluation of the three-loop master diagrams.
In the following , we shall frequently make use of the QCD expansion of
$\Delta\rho$ through ${\cal O}(\alpha_s^2G_FM_t^2)$.
{}For the reader's convenience, we shall list it here for $N_c=3$ and $n_f=6$.
The $\overline{\mbox{MS}}$ and on-shell results read \cite{avd}
\begin{eqnarray}
\Delta\bar\rho&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&
N_cx_t\left[1+a(2\,l-0.193\,245)
+a^2\left({15\over4}l^2+2.025\,330\,l-3.969\,560\right)\right],
\\
\label{drhoos}
\Delta\rho&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&
N_cX_t[1-2.859\,912\,a-a^2(5.004\,846\,L+14.594\,028)],
\end{eqnarray}
respectively.
To start with, we shall construct the low-$M_H$ effective $\ell^+\ell^-H$
interaction Lagrangian through ${\cal O}(\alpha_s^2G_FM_t^2)$.
In the following, bare quantities will be labelled with the superscript 0.
The bare $\ell^+\ell^-H$ Lagrangian reads
\begin{equation}
\label{llh}
{\cal L}_{\ell\ell H}=-m_\ell^0\bar\ell^0\ell^0{H^0\over v^0},
\end{equation}
where $v$ denotes the Higgs vacuum expectation value.
The renormalizations of the lepton mass and wave function do not receive
corrections in ${\cal O}(\alpha_s^nG_FM_t^2)$, where $n=0,1,2$,
so that we may replace $m_\ell^0$ and $\ell^0$ with their renormalized
counterparts.
In the $G_F$ formulation of the on-shell scheme, we have \cite{hff}
\begin{equation}
\label{du}
{H^0\over v^0}=2^{1/4}G_F^{1/2}H(1+\delta_u),
\end{equation}
with
\begin{equation}
\delta_u=-{1\over2}\left[{\Pi_{WW}(0)\over M_W^2}+\Pi_{HH}^\prime(0)\right].
\end{equation}
Here, $\Pi_{WW}(q^2)$ and $\Pi_{HH}(q^2)$ are the $W$- and Higgs-boson
self-energies for external momentum $q$, respectively, and the subscript $u$
is to remind us that this term appears as a universal building block in the
radiative corrections to all production and decay processes of the Higgs boson.
Consequently, the renormalized version of Eq.~(\ref{llh}) reads
\begin{equation}
\label{lllh}
{\cal L}_{\ell\ell H}=-2^{1/4}G_F^{1/2}m_\ell\bar\ell\ell H(1+\delta_u).
\end{equation}
The one-loop expressions for $\Pi_{WW}(q^2)$ and $\Pi_{HH}(q^2)$ have been
presented in Ref.~\cite{hzz}.
The leading-order QCD corrections to $\Pi_{WW}(q^2)$ and $\Pi_{HH}(q^2)$ for
arbitrary quark masses have been found in Refs.~\cite{djo,hll}, respectively.
The ${\cal O}(\alpha_sG_FM_t^2)$ term of $\delta_u$ has independently been
obtained in Ref.~\cite{kwi} by using the computational technique outlined
above at the two-loop level.
Here, we shall extend this analysis to ${\cal O}(\alpha_s^2G_FM_t^2)$.
The ${\cal O}(\alpha_s^2G_FM_t^2)$ term of $\Pi_{WW}(0)$ may be found in
Ref.~\cite{avd}.
The Feynman diagrams pertinent to $\Pi_{HH}(q^2)$ in
${\cal O}(\alpha_s^2G_FM_t^2)$ come in twenty different topologies.
Typical examples are depicted in Fig.~\ref{one}.
We shall renormalize the strong coupling constant and the top-quark mass
according to the $\overline{\mbox{MS}}$ scheme.
The appropriate counterterms are listed in Ref.~\cite{gra}.
In this way, we obtain
\begin{eqnarray}
\label{pihh}
\Pi_{HH}^\prime(0)&\hspace*{-2.5mm}=\hspace*{-2.5mm}&N_cx_t\left\{{2\over\epsilon}+2l-{4\over3}
+hC_F\left(-{6\over\epsilon^2}+{5\over\epsilon}+6l^2-10l-{37\over6}\right)
+h^2C_F\left[27\zeta(3)+6
\right.\right.
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_F\left({12\over\epsilon^3}-{12\over\epsilon^2}
+{1\over\epsilon}\left(24\zeta(3)-{119\over6}\right)
+l\left(72\zeta(3)-{93\over2}\right)+24B_4-108\zeta(4)
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.
106\zeta(3)+{331\over12}\right)
+C_A\left({22\over3\epsilon^3}-{83\over3\epsilon^2}
+{1\over\epsilon}\left(-12\zeta(3)+{77\over3}\right)
+{22\over3}l^3
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.
14l^2+l\left(-36\zeta(3)-{961\over18}\right)-12B_4+54\zeta(4)
-{55\over3}\zeta(3)-7\right)
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.\left.
n_f\left(-{4\over3\epsilon^3}+{10\over3\epsilon^2}-{8\over3\epsilon}
-{4\over3}l^3+{65\over9}l-{32\over3}\zeta(3)-3\right)\right]\right\}.
\end{eqnarray}
When we combine Eq.~(\ref{pihh}) with the corresponding expression for
$\Pi_{WW}(0)$ \cite{avd}, the ultraviolet divergences cancel, and we obtain
\begin{eqnarray}
\bar\delta_u&\hspace*{-2.5mm}=\hspace*{-2.5mm}&N_cx_t\left\{{7\over6}
+hC_F\left(7l-2\zeta(2)+{19\over3}\right)
+h^2C_F\left[243S_2-{449\over6}\zeta(3)-{14\over3}\zeta(2)+{79\over3}
\right.\right.
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_F\left(21l^2+l\left(-12\zeta(2)-{1\over2}\right)+4B_4+2D_3
-{1053\over2}S_2+2\zeta(4)+{599\over3}\zeta(3)-{259\over9}\zeta(2)
\right.\nonumber\\
&\hspace*{-2.5mm}-\hspace*{-2.5mm}&\left.
{3043\over72}\right)
+C_A\left({77\over6}l^2+l\left(-{22\over3}\zeta(2)+{1097\over18}\right)
-2B_4-D_3+{1053\over4}S_2+15\zeta(4)
\right.\nonumber\\
&\hspace*{-2.5mm}-\hspace*{-2.5mm}&\left.\left.\left.\!\!
{509\over6}\zeta(3)
-{73\over3}\zeta(2)
+{953\over24}\right)
+n_f\left(\!-{7\over3}l^2+l\left({4\over3}\zeta(2)-{73\over9}\right)
\!-{8\over3}\zeta(3)+{14\over3}\zeta(2)-{55\over12}\right)\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&
{7\over6}N_cx_t\left[1+a(2\,l+0.869\,561)
+a^2\left({15\over4}l^2+6.010\,856\,l-2.742\,226\right)\right].
\end{eqnarray}
With the help of Eq.~(\ref{mass}), we may eliminate $m_t(\mu)$ in favour of
$M_t$, which leads to
\begin{eqnarray}
\label{duos}
\delta_u&\hspace*{-2.5mm}=\hspace*{-2.5mm}&N_cX_t\left\{{7\over6}
+hC_F\left(-2\zeta(2)-3\right)
+h^2C_F\left[243S_2-{449\over6}\zeta(3)-{98\over3}\zeta(2)+{121\over3}
\right.\right.
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_F\left(4B_4+2D_3-{1053\over2}S_2+2\zeta(4)+{515\over3}\zeta(3)
+\zeta(2)\left(112\ln2-{745\over9}\right)-{146\over9}\right)
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_A\left(L\left(-{22\over3}\zeta(2)-11\right)
-2B_4-D_3+{1053\over4}S_2+15\zeta(4)-{425\over6}\zeta(3)
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.\left.\left.
\zeta(2)\left(-56\ln2-{17\over3}\right)-{2459\over36}\right)
+n_f\left(L\left({4\over3}\zeta(2)+2\right)
-{8\over3}\zeta(3)+14\zeta(2)+{83\over9}\right)\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&{7\over6}N_cX_t[1-1.797\,105\,a-a^2(3.144\,934\,L+16.200\,847)].
\end{eqnarray}
Equation~(\ref{duos}) reproduces the ${\cal O}(G_FM_t^2)$ and
${\cal O}(\alpha_sG_FM_t^2)$ terms found in Refs.~\cite{hff,hll}, respectively.
We observe that the new ${\cal O}(\alpha_s^2G_FM_t^2)$ term in Eq.~(\ref{duos})
enhances the QCD correction and thus supports the screening of the
leading-order $M_t$ dependence.
The choice $\mu=M_t$ is singled out, since it eliminates the terms containing
$L$ in Eq.~(\ref{duos}).
The nonlogarithmic coefficient of $(\alpha_s/\pi)^2$ in Eq.~(\ref{duos}) is
relatively large;
it exceeds the corresponding coefficient of $\Delta\rho$ in Eq.~(\ref{drhoos})
by approximately 11\%.
If we consider the ratio of the coefficient of $(\alpha_s/\pi)^2$ to the one of
$\alpha_s/\pi$, the difference is even more pronounced;
the corresponding numbers for Eqs.~(\ref{duos}) and (\ref{drhoos}) are
roughly 9 versus 5.
A phenomenologically interesting application of Eq.~(\ref{lllh}) is to study
the effect of QCD corrections on $\Gamma(H\to\ell^+\ell^-)$.
The corrections through ${\cal O}(\alpha_s^2G_FM_t^2)$ to this observable may
be accommodated by multiplying the Born formula \cite{hff} with
\begin{eqnarray}
\label{kllh}
K_{\ell\ell H}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&(1+\delta_u)^2\nonumber\\
&\hspace*{-2.5mm}=\hspace*{-2.5mm}&1+2\delta_u,
\end{eqnarray}
where we have suppressed terms of ${\cal O}(G_F^2M_t^4)$ in the second line.
This implies that $\delta_u$ is gauge independent and RG invariant in these
orders.
In order to avoid double counting, the ${\cal O}(G_FM_t^2)$ term must once be
subtracted when the full one-loop correction \cite{hff} is included.
A detailed numerical analysis will be presented in Section~4.
Next, we shall derive the ${\cal O}(\alpha_s^2G_FM_t^2)$ correction to the
low-$M_H$ effective $W^+W^-H$ interaction Lagrangian.
In contrast to the $\ell^+\ell^-H$ case, we are now faced with the task of
computing genuine three-point amplitudes at three loops, which, at first sight,
appears to be enormously hard.
Fortunately, in the limit that we are interested in, this problem may be
reduced to one involving just three-loop two-point diagrams by means of a
low-energy theorem \cite{ell,vai}.
Generally speaking, this theorem relates the amplitudes of two processes which
differ by the insertion of an external Higgs-boson line carrying zero
four-momentum.
It allows us to compute a loop amplitude, ${\cal M}(A\to B+H)$, with an
external Higgs boson which is light compared to the virtual particles by
differentiating the respective amplitude without that Higgs boson,
${\cal M}(A\to B)$, with respect to the virtual-particle masses.
More precisely \cite{ell,vai},
\begin{equation}
\label{let}
\lim_{p_H\to0}{\cal M}(A\to B+H)={1\over v}\sum_i
{m_i\partial\over\partial m_i}{\cal M}(A\to B),
\end{equation}
where $i$ runs over all massive virtual particles which are involved in the
transition $A\to B$.
Here, it is understood that the differential operator does not act on factors
of $m_i$ appearing in coupling constants, since this would generate tree-level
interactions involving the Higgs boson that do not exist in the SM.
This theorem has variously been applied at leading order \cite{ell,vai} and has
even made its way into standard text books \cite{oku}.
Special care must be exercised if this theorem is to be applied beyond leading
order.
Then, it must be formulated for the bare quantities of the theory, and the
renormalization must be performed after the left-hand side of Eq.~(\ref{let})
has been constructed \cite{ks1}.
The beyond-leading-order version of this theorem \cite {ks1} has recently been
employed to find the ${\cal O}(\alpha_sG_FM_t^2)$ corrections to
$\Gamma\left(H\to b\bar b\,\right)$ \cite{ks1},
$\Gamma\left(Z\to f\bar fH\right)$, and $\sigma(e^+e^-\to ZH)$ \cite{ks2}.
A comprehensive review of higher-order applications of this and related
low-energy theorems may be found in Ref.~\cite{ks3}.
An axiomatic formulation of these soft-Higgs theorems has recently been
introduced in Ref.~\cite{kil}.
Proceeding along the lines of Refs.~\cite{ks2,ks3}, we find the bare $W^+W^-H$
interaction Lagrangian including its genuine vertex corrections to be
\begin{equation}
{\cal L}_{W^+W^-H}=2(M_W^0)^2(W_\mu^+)^0(W^{-\mu})^0{H^0\over v^0}
\left[1-{(m_t^0)^2\partial\over\partial(m_t^0)^2}\,
{\Pi_{WW}(0)\over(M_W^0)^2}\right],
\end{equation}
where it is understood that $\Pi_{WW}(0)$ is expressed in terms of the bare
top-quark mass, $m_t^0$, while all other quark masses are put to zero.
We renormalize the $W$-boson mass and wave function
according to the electroweak on-shell scheme by substituting
\begin{eqnarray}
(M_W^0)^2&\hspace*{-2.5mm}=\hspace*{-2.5mm}&M_W^2+\delta M_W^2,\nonumber\\
(W_\mu^\pm)^0&\hspace*{-2.5mm}=\hspace*{-2.5mm}&W_\mu^\pm(1+\delta Z_W)^{1/2},
\end{eqnarray}
with the counterterms
\begin{eqnarray}
\delta M_W^2&\hspace*{-2.5mm}=\hspace*{-2.5mm}&\Pi_{WW}(0),\nonumber\\
\delta Z_W&\hspace*{-2.5mm}=\hspace*{-2.5mm}&-\Pi_{WW}^\prime(0).
\end{eqnarray}
{}For dimensional reasons, $\delta Z_W$ does not receive corrections in the
orders that we are interested in.
Using Eq.~(\ref{du}), we thus obtain
\begin{equation}
\label{lwwh}
{\cal L}_{W^+W^-H}=2^{5/4}G_F^{1/2}M_W^2W_\mu^+W^{-\mu}H(1+\delta_{WWH}),
\end{equation}
where
\begin{equation}
\delta_{WWH}=\delta_u+\delta_{nu}^{WWH}
\end{equation}
and the non-universal part herein may be calculated from
\begin{equation}
\label{dwwhnu}
\delta_{nu}^{WWH}=\left[1-{(m_t^0)^2\partial\over\partial(m_t^0)^2}\right]
{\Pi_{WW}(0)\over(M_W^0)^2}.
\end{equation}
In Ref.~\cite{avd}, $\Pi_{WW}(0)$ is expressed in terms of renormalized
parameters.
Thus, we have to undo the top-quark mass renormalization \cite{gra} before we
can apply Eq.~(\ref{dwwhnu}).
Then, after evaluating the right-hand side of Eq.~(\ref{dwwhnu}), we
reintroduce the renormalized top-quark mass and so obtain a finite result for
$\delta_{nu}^{WWH}$, which we combine with $\delta_u$ to get $\delta_{WWH}$.
If we define the top-quark mass according to the $\overline{\mbox{MS}}$
scheme, then the result is
\begin{eqnarray}
\bar\delta_{WWH}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&N_cx_t\left\{-{5\over6}
+hC_F\left(-5l-2\zeta(2)+{7\over3}\right)
+h^2C_F\left[243S_2-{449\over6}\zeta(3)-{14\over3}\zeta(2)+{79\over3}
\right.\right.
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_F\left(-15l^2+l\left(-12\zeta(2)+{83\over2}\right)+4B_4+2D_3
-{1053\over2}S_2+2\zeta(4)+{383\over3}\zeta(3)
\right.\nonumber\\
&\hspace*{-2.5mm}-\hspace*{-2.5mm}&\left.
{43\over9}\zeta(2)
+{377\over72}\right)
+C_A\left(-{55\over6}l^2+l\left(-{22\over3}\zeta(2)-{331\over18}\right)
-2B_4-D_3+{1053\over4}S_2
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.
15\zeta(4)-{293\over6}\zeta(3)-{29\over3}\zeta(2)
-{691\over24}\right)
+n_f\left({5\over3}l^2+l\left({4\over3}\zeta(2)+{11\over9}\right)
-{8\over3}\zeta(3)
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.\left.\left.
2\zeta(2)+{53\over12}\right)\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&
-{5\over6}N_cx_t\left[1+a(2\,l+0.382\,614)
+a^2\left({15\over4}l^2+4.184\,802\,l+1.343\,710\right)\right].
\end{eqnarray}
The corresponding formula written in terms of $M_t$ reads
\begin{eqnarray}
\label{dwwhos}
\delta_{WWH}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&N_cX_t\left\{-{5\over6}
+hC_F\left(-2\zeta(2)+9\right)
+h^2C_F\left[243S_2-{449\over6}\zeta(3)+{46\over3}\zeta(2)+{49\over3}
\right.\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_F\left(4B_4+2D_3-{1053\over2}S_2+2\zeta(4)+{443\over3}\zeta(3)
+\zeta(2)\left(-80\ln2+{551\over9}\right)-{614\over9}\right)
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_A\left(L\left(-{22\over3}\zeta(2)+33\right)
-2B_4-D_3+{1053\over4}S_2+15\zeta(4)-{353\over6}\zeta(3)
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.\left.\left.
\zeta(2)(40\ln2-23)+{1741\over36}\right)
+n_f\left(L\left({4\over3}\zeta(2)-6\right)
-{8\over3}\zeta(3)-{14\over3}\zeta(2)-{49\over9}\right)\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&-{5\over6}N_cX_t[1-2.284\,053\,a-a^2(3.997\,092\,L+10.816\,384)].
\end{eqnarray}
We recover the ${\cal O}(G_FM_t^2)$ and ${\cal O}(\alpha_sG_FM_t^2)$ terms of
Refs.~\cite{hww,ks3}, respectively.
Similarly to $\Delta\rho$ and $\delta_u$, the ${\cal O}(\alpha_s^2G_FM_t^2)$
term of Eq.~(\ref{dwwhos}) supports the screening of the one-loop $M_t$
dependence by the leading-order QCD correction.
Here, the coefficient of $(\alpha_s/\pi)^2$ is by 26\% smaller than in the
case of $\Delta\rho$, but it, too, is about five times bigger than the
coefficient of $\alpha_s/\pi$.
{}From Eq.~(\ref{lwwh}) it follows on that $\Gamma(H\to W^+W^-)$ receives the
correction factor
\begin{equation}
\label{kwwh}
K_{WWH}=1+2\delta_{WWH}.
\end{equation}
Thus, both $\delta_{nu}^{WWH}$ and $\delta_{WWH}$ are gauge independent and RG
invariant to the orders that we are working in.
The tree-level formula for $\Gamma(H\to W^+W^-)$ and its full one-loop
correction may be found in Ref.~\cite{hww}.
In order for the Higgs boson to decay into a $W^+W^-$ pair, it must satisfy
$M_H>2M_W$.
On the other hand, the high-$M_t$ approximation is based on $M_H\ll M_t$.
Since these two conditions conflict with each other \cite{abe}, the
application of Eq.~(\ref{lwwh}) to $\Gamma(H\to W^+W^-)$ is somewhat academic.
However, the first condition is relaxed to $M_H>M_W$ or removed altogether
if one or both of the $W$ bosons are allowed to leave their mass shells,
respectively.
In order to avoid gluon exchange between the $W^+W^-H$ vertex and the external
fermions, we restrict our considerations to leptonic currents.
The resulting class of processes includes
$H\to(W^+)^*W^-\to\ell^+\nu_\ell W^-$,
$H\to W^+(W^-)^*\to W^+\ell^-\bar\nu_\ell$,
$H\to(W^+)^*(W^-)^*\to\ell^+\nu_\ell\ell^{\prime-}\bar\nu_{\ell^\prime}$,
as well as $e^+e^-\to\bar\nu_e\nu_e(W^+)^*(W^-)^*\to\bar\nu_e\nu_eH$ via
$W^+W^-$ fusion.
The Born formulae for these $1\to3$, $1\to4$, and $2\to3$ processes may be
found in Refs.~\cite{riz,hzgg,eezh}, respectively.
Since $G_F$ is defined through the radiative correction to the muon decay,
which is a charged-current process, the $W$-boson propagator does not
receive radiative corrections in the orders of interest here.
Therefore, the correction factors of all these processes coincide with the one
for $\Gamma(H\to W^+W^-)$.
Finally, we shall treat the $ZZH$ interaction.
The procedure is very similar to the $W^+W^-H$ case.
Application of the low-energy theorem (\ref{let}) to the bare $Z$-boson vacuum
polarization induced by the top quark yields
\begin{equation}
{\cal L}_{ZZH}=(M_Z^0)^2Z_\mu^0Z^{\mu0}{H^0\over v^0}
\left[1-{(m_t^0)^2\partial\over\partial(m_t^0)^2}\,
{\Pi_{ZZ}(0)\over(M_Z^0)^2}\right].
\end{equation}
Again, $(M_Z^0)^2=M_Z^2+\delta M_Z^2$, with $\delta M_Z^2=\Pi_{ZZ}(0)$, and
$Z_\mu^0=Z_\mu$.
Together with Eq.~(\ref{du}), we then have
\begin{equation}
\label{lzzh}
{\cal L}_{ZZH}=2^{1/4}G_F^{1/2}M_Z^2Z_\mu Z^\mu H(1+\delta_{ZZH}),
\end{equation}
where $\delta_{ZZH}=\delta_u+\delta_{nu}^{ZZH}$,
with the non-universal part,
\begin{equation}
\delta_{nu}^{ZZH}=\left[1-{(m_t^0)^2\partial\over\partial(m_t^0)^2}\right]
{\Pi_{ZZ}(0)\over(M_Z^0)^2},
\end{equation}
being separately finite, gauge independent, and RG invariant.
Starting from the expression for $\Pi_{ZZ}(0)$ listed in Ref.~\cite{avd}
and repeating the steps of the $W^+W^-H$ analysis, we obtain
\begin{eqnarray}
\bar\delta_{ZZH}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&N_cx_t\left\{-{5\over6}
+hC_F\left(-5l-2\zeta(2)+{25\over3}\right)
+h^2C_F\left[243S_2-{449\over6}\zeta(3)-{14\over3}\zeta(2)+{79\over3}
\right.\right.
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_F\left(-15l^2+l\left(-12\zeta(2)+{155\over2}\right)+4B_4+2D_3
-{1053\over2}S_2+2\zeta(4)+{383\over3}\zeta(3)
\right.\nonumber\\
&\hspace*{-2.5mm}-\hspace*{-2.5mm}&\left.
{259\over9}\zeta(2)+{593\over72}\right)
+C_A\left(-{55\over6}l^2+l\left(-{22\over3}\zeta(2)+{65\over18}\right)
-2B_4-D_3+{1053\over4}S_2
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.
15\zeta(4)-{293\over6}\zeta(3)-{73\over3}\zeta(2)
+{613\over24}\right)
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.\left.
n_f\left({5\over3}l^2+l\left({4\over3}\zeta(2)-{25\over9}\right)
-{8\over3}\zeta(3)+{14\over3}\zeta(2)-{35\over12}\right)\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&
-{5\over6}N_cx_t\left[1+a(2\,l-2.017\,386)
+a^2\left({15\over4}l^2-4.815\,198\,l-1.086\,685\right)\right]
\end{eqnarray}
in the $\overline{\mbox{MS}}$ scheme and
\begin{eqnarray}
\label{dzzhos}
\delta_{ZZH}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&N_cX_t\left\{-{5\over6}
+hC_F\left(-2\zeta(2)+15\right)
+h^2C_F\left[243S_2-{449\over6}\zeta(3)+{46\over3}\zeta(2)+{49\over3}
\right.\right.
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_F\left(4B_4+2D_3-{1053\over2}S_2+2\zeta(4)+{443\over3}\zeta(3)
+\zeta(2)\left(-80\ln2+{335\over9}\right)-{1019\over9}\right)
\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&C_A\left(L\left(-{22\over3}\zeta(2)+55\right)
-2B_4-D_3+{1053\over4}S_2+15\zeta(4)-{353\over6}\zeta(3)
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.\left.\left.\!
\zeta(2)\left(40\ln2-{113\over3}\right)+{3697\over36}\right)
+n_f\left(L\left({4\over3}\zeta(2)-10\right)
-{8\over3}\zeta(3)-2\zeta(2)-{115\over9}\right)\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&-{5\over6}N_cX_t[1-4.684\,053\,a-a^2(8.197\,092\,L+6.846\,779)]
\end{eqnarray}
in the on-shell scheme.
The ${\cal O}(G_FM_t^2)$ and ${\cal O}(\alpha_sG_FM_t^2)$ terms of
Eq.~(\ref{dzzhos}) agree with those found in Refs.~\cite{hzz,ks2},
respectively.
Again, the ${\cal O}(\alpha_s^2G_FM_t^2)$ term of Eq.~(\ref{dzzhos}) reinforces
the potential of the QCD corrections to reduce the leading-order $M_t$
dependence.
Comparing $\delta_{ZZH}$ with $\Delta\rho$, $\delta_u$, and $\delta_{WWH}$,
we observe that it has the largest $\alpha_s/\pi$ coefficient but the smallest
$(\alpha_s/\pi)^2$ coefficient, the ratio of the latter to the former only
being about 1.5.
The $(\alpha_s/\pi)^2$ coefficient of $\delta_{ZZH}$ is by 53\% smaller than
the one of $\Delta\rho$ in Eq.~(\ref{drhoos}).
{}From Eq.~(\ref{lzzh}), we infer that $\Gamma(H\to ZZ)$ receives the
correction factor
\begin{equation}
\label{kzzh}
K_{ZZH}=1+2\delta_{ZZH}.
\end{equation}
The Born formula for $\Gamma(H\to ZZ)$ and its full one-loop
correction may be found in Ref.~\cite{hzz}.
Since the condition $2M_Z<M_H\ll M_t$ is likely to be unrealistic \cite{abe},
the high-$M_t$ approximation underlying Eq.~(\ref{lzzh}) is of limited
usefulness for $H\to ZZ$.
We can evade this problem by allowing for one or both of the $Z$ bosons to go
off-shell.
In addition to the information contained in Eq.~(\ref{lzzh}), we then need to
account for the corresponding corrections arising from the gauge sector.
However, in order not to invoke unknown QCD corrections, we have to restrict
ourselves to the inclusion of lepton lines.
The form of the additional corrections depends on the considered reaction.
It is useful to divide the phenomenologically relevant processes into three
classes:
\begin{itemize}
\item[(1)] $H\to Z^*Z\to f\bar fZ$, $Z\to Z^*H\to f\bar fH$,
and $e^+e^-\to ZH$;
\item[(2)] $H\to Z^*Z^*\to f\bar ff^\prime\bar f^\prime$ and
$e^+e^-\to Z^*\to Z^*H\to f\bar fH$ (via Higgs-strahlung);
\item[(3)] $e^+e^-\to e^+e^-Z^*Z^*\to e^+e^-H$ (via $ZZ$ fusion).
\end{itemize}
Here, $f$ and $f^\prime$ stand for neutrinos and charged leptons.
The results
for $H\to f\bar fZ$ at tree level \cite{riz} and at one loop \cite{pr},
for $Z\to f\bar fH$ at tree level \cite{zhgg} and at one loop \cite{zffh},
for $e^+e^-\to ZH$ at tree level and at one loop \cite{eezh},
for $H\to f\bar ff^\prime\bar f^\prime$ at tree level \cite{hzgg},
and for $e^+e^-\to f\bar fH$ at tree level \cite{eezh}
are in the literature.
In the next section, we shall discuss the corrections to these processes in
${\cal O}(\alpha_s^nG_FM_t^2)$, with $n=0,1,2$.
\section{Corrections from the gauge sector}
The IBA \cite{iba} provides a
systematic and convenient method to incorporate the dominant corrections of
fermionic origin to processes within the gauge sector of the SM.
These are contained in $\Delta\rho$ and
$\Delta\alpha=1-\alpha/\overline\alpha$, which parameterizes the running of the
fine-structure constant from its value, $\alpha$, defined in Thomson scattering
to its value, $\overline\alpha$, measured at the $Z$-boson scale.
The recipe is as follows.
Starting from the Born formula expressed in terms of $\alpha$, $c_w$, $s_w$,
and the physical particle masses, one substitutes
\begin{eqnarray}
\alpha&\hspace*{-2.5mm}\to\hspace*{-2.5mm}&\overline\alpha={\alpha\over1-\Delta\alpha},\nonumber\\
c_w^2&\hspace*{-2.5mm}\to\hspace*{-2.5mm}&\overline c_w^2=1-\overline s_w^2=c_w^2(1-\Delta\rho).
\end{eqnarray}
To eliminate $\overline\alpha$ in favour of $G_F$, one exploits the relation
\begin{equation}
{\sqrt2\over\pi}G_F={\overline\alpha\over\overline s_w^2M_W^2}
={\overline\alpha\over\overline c_w^2\overline s_w^2M_Z^2}(1-\Delta\rho),
\end{equation}
which correctly accounts for the leading fermionic corrections.
We shall now employ the IBA to find the additional corrections through
${\cal O}(\alpha_s^2G_FM_t^2)$ to the four- and five-point processes with a
$ZZH$ coupling, which we have classified in Section~2.
We shall always assume that the Born formulae are written in terms of
$G_F$, $c_w$, $s_w$, and the physical particle masses.
The generic correction factor for class~(1) reads \cite{ks2,zffh}
\begin{eqnarray}
\label{k1}
K_1^{(f)}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&{(1+\delta^{ZZH})^2\over1-\Delta\rho}\,
{\overline v_f^2+a_f^2\over v_f^2+a_f^2}\nonumber\\
&\hspace*{-2.5mm}=\hspace*{-2.5mm}&1+2\delta^{ZZH}+
\left(1-8c_w^2{Q_fv_f\over v_f^2+a_f^2}\right)\Delta\rho,
\end{eqnarray}
where $v_f=2I_f-4s_w^2Q_f$, $\overline v_f=2I_f-4\overline s_w^2Q_f$,
$a_f=2I_f$, $Q_f$ is the electric charge of $f$ in units of the positron
charge, $I_f$ is the third component of weak isospin of the left-handed
component of $f$, and we have omitted terms of ${\cal O}(G_F^2M_t^4)$ in the
second line.
Similarly, the correction factor for class~(2) is given by \cite{gro}
\begin{eqnarray}
\label{k2}
K_2^{(ff^\prime)}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&
{(1+\delta^{ZZH})^2\over(1-\Delta\rho)^2}\,
{\overline v_f^2+a_f^2\over v_f^2+a_f^2}\,
{\overline v_{f^\prime}^2+a_{f^\prime}^2\over v_{f^\prime}^2+a_{f^\prime}^2}
\nonumber\\
&\hspace*{-2.5mm}=\hspace*{-2.5mm}&1+2\delta^{ZZH}+
2\left[1-4c_w^2\left({Q_fv_f\over v_f^2+a_f^2}
+{Q_{f^\prime}v_{f^\prime}\over v_{f^\prime}^2+a_{f^\prime}^2}\right)\right]
\Delta\rho.
\end{eqnarray}
Here and in the following, we neglect interference terms of five-point
amplitudes with a single fermion trace, since, in the kinematic regime of
interest here, these are strongly suppressed, by $\Gamma_V/M_V$, with $V=W,Z$.
Such terms have recently been included in a tree-level calculation of
$\Gamma(H\to2V\to4f)$ for $M_H\ll M_W$ \cite{asa}.
The correction factor for case~(3) is slightly more complicated because the
electron and positron lines run from the initial state to the final state.
Allowing for generic fermion flavours, $f$ and $f^\prime$,
the Born cross section may be evaluated from
\begin{equation}
\label{zfus}
\sigma(ff^\prime\to ff^\prime H)={G_F^3M_Z^4\over64\pi^3\sqrt2}
\left[(v_f^2+a_f^2)(v_{f^\prime}^2+a_{f^\prime}^2)A
\pm4v_fa_fv_{f^\prime}a_{f^\prime}B\right],
\end{equation}
where
\begin{equation}
\label{int}
A=\int_{M_H^2/s}^1dx\int_x^1dy{a(x,y)\over\left[1+s(y-x)/M_Z^2\right]^2},
\end{equation}
and similarly for $B$, $\sqrt s$ is the centre-of-mass energy, and the
plus/minus sign refers to an odd/even number of antifermions in the initial
state.
The process under case~(3), with an $e^+e^-$ initial state, requires the plus
sign.
The integrands read
\begin{eqnarray}
a(x,y)&\hspace*{-2.5mm}=\hspace*{-2.5mm}&
\left({2x\over y^3}-{1+2x\over y^2}+{2+x\over2y}-{1\over2}\right)
\left[{z\over1+z}-\ln(1+z)\right]
+{x\over y^2}\left({1\over y}-1\right){z^2\over1+z},\nonumber\\
b(x,y)&\hspace*{-2.5mm}=\hspace*{-2.5mm}&
\left(-{x\over y^2}+{2+x\over2y}-{1\over2}\right)
\left[{z\over1+z}-\ln(1+z)\right],
\end{eqnarray}
where $z=(y/M_Z^2)(s-M_H^2/x)$.
The inner integration in Eq.~(\ref{int}) has been carried out analytically
in Appendix~A of Ref.~\cite{eezh}.
By means of the IBA, we obtain the correction factor for Eq.~(\ref{zfus}) as
\begin{eqnarray}
\label{k3}
K_3^{(ff^\prime)}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&
{(1+\delta^{ZZH})^2\over(1-\Delta\rho)^2}\,
{(\overline v_f^2+a_f^2)(\overline v_{f^\prime}^2+a_{f^\prime}^2)A
\pm4\overline v_fa_f\overline v_{f^\prime}a_{f^\prime}B\over
(v_f^2+a_f^2)(v_{f^\prime}^2+a_{f^\prime}^2)A
\pm4v_fa_fv_{f^\prime}a_{f^\prime}B}\nonumber\\
&\hspace*{-2.5mm}=\hspace*{-2.5mm}&1+2\delta^{ZZH}+
2\left[1-{4c_w^2\over1+r}\left({Q_fv_f\over v_f^2+a_f^2}
+{Q_{f^\prime}v_{f^\prime}\over v_{f^\prime}^2+a_{f^\prime}^2}\right)
\right.\nonumber\\&\hspace*{-2.5mm}-\hspace*{-2.5mm}&\left.
{2c_w^2\over1+1/r}\left({Q_f\over v_f}+{Q_{f^\prime}\over v_{f^\prime}}\right)
\right]\Delta\rho,
\end{eqnarray}
where
\begin{equation}
r={\pm4v_fa_fv_{f^\prime}a_{f^\prime}B\over
(v_f^2+a_f^2)(v_{f^\prime}^2+a_{f^\prime}^2)A}.
\end{equation}
We wish to point out that, in the limit $r\to0$, $K_3$ coincides with $K_2$.
Detailed analysis reveals that $r$ is quite small in magnitude whenever
$e^+e^-\to e^+e^-H$ via $ZZ$ fusion is phenomenologically relevant.
In fact, if we consider energies $\sqrt s>150$~GeV and demand that the total
cross section of this process be in excess of $10^{-2}$~fb$^{-1}$, then we
find $|r|<1\%$.
This concludes our discussion of the additional QCD corrections to the
processes under items~(1)--(3) originating in the gauge sector.
\section{Numerical results}
We are now in a position to explore the phenomenological implications of our
results.
We shall take the values of our input parameters to be
$M_W=80.26$~GeV,
$M_Z=91.1887$~GeV \cite{lep},
$M_t=180$~GeV \cite{abe},
and $\alpha_s^{(5)}(M_Z)=0.118$ \cite{bet}.\footnote{Note that this value does
not include results from lattice computations.}
The latter corresponds to $\alpha_s^{(6)}(M_t)=0.1071$, which entails that
$\Lambda_{\overline{MS}}^{(6)}=91$~MeV in Eq.~(\ref{as}).
If we use the one-loop formula for $\alpha_s^{(6)}(\mu)$, {\it i.e.},
Eq.~(\ref{as}) with the second term within the square brackets discarded,
$\Lambda_{\overline{MS}}^{(6)}$ comes down to 41~MeV.
Any perturbative calculation to finite order depends on the choice of
renormalization scheme and, in general, also on one or more renormalization
scales.
It is generally believed that the scheme and scale dependences of a
calculation up to a given order indicate the size of the unknown higher-order
contributions, {\it i.e.}, they provide us with an estimate of the theoretical
uncertainty.
Of course, the central values and variations of the scales must be judiciously
chosen in order for this estimate to be meaningful.
If the perturbation series converges, then the scheme and scale dependences
are expected to decrease as the respective next order is taken into account.
This principle has recently been confirmed for $\Delta\rho$ \cite{avd,yr}.
Here, we have the opportunity to carry out similar studies for the three
additional observables $\delta_u$, $\delta_{WWH}$, and $\delta_{ZZH}$.
Similarly to Ref.~\cite{avd}, we have presented our results in the on-shell
and $\overline{\mbox{MS}}$ schemes as functions of a single renormalization
scale, $\mu$.
In the $\overline{\mbox{MS}}$ scheme, one could, in principle, introduce
individual renormalization scales for the coupling and the mass.
{}For simplicity, we have chosen not to do so.
It is natural to define the central value of $\mu$ in such a way that, at
this point, the radiative correction is devoid of logarithmic terms.
This leads us to set $\mu=\xi M_t$ in the on-shell scheme and $\mu=\xi\mu_t$,
where $\mu_t=m_t(\mu_t)$, in the $\overline{\mbox{MS}}$ scheme.
We may obtain $\mu_t$ as a closed function of $M_t$ by iterating
Eq.~(\ref{mass}), with the result that
\begin{eqnarray}
\label{mut}
{\mu_t\over M_t}&\hspace*{-2.5mm}=\hspace*{-2.5mm}&1-4HC_F
+H^2C_F\left\{-12\zeta(2)+6
+C_F\left[-12\zeta(3)+6\zeta(2)(8\ln2-5)+{199\over8}\right]
\right.\nonumber\\
&\hspace*{-2.5mm}+\hspace*{-2.5mm}&\left.
C_A\left[6\zeta(3)+8\zeta(2)(-3\ln2+1)-{1111\over24}\right]
+n_f\left[4\zeta(2)+{71\over12}\right]\right\}
\nonumber\\
&\hspace*{-2.5mm}\approx\hspace*{-2.5mm}&1-{4\over3}A-6.458\,784\,A^2,
\end{eqnarray}
where $A=4H=\alpha_s(M_t)/\pi$.
{}For $M_t=180$~GeV, Eq.~(\ref{mut}) yields $\mu_t=170.5$~GeV, in good
agreement
with the exact fix point of Eq.~(\ref{mass}), which is $\mu_t=170.6$~GeV.
\begin{table}[ht]
\caption{Relative deviations (in \%) of $\Delta\rho$, $\delta_u$,
$\delta_{WWH}$, and $\delta_{ZZH}$ from the respective one-loop results due to
their corrections up to ${\cal O}(\alpha_s)$ and ${\cal O}(\alpha_s^2)$.
The renormalization scale dependence is investigated by choosing $\mu=\xi M_t$,
with $\xi$ variable.
}\label{tab:os}
\medskip
\begin{tabular}{|c|c|c|c|c|c|c|c|c|} \hline\hline
\rule{0mm}{5mm}$\xi$ &
\multicolumn{2}{c|}{$\Delta\rho/\Delta\rho^{(1)}-1$ [\%]} &
\multicolumn{2}{c|}{$\delta_u/\delta_u^{(1)}-1$ [\%]} &
\multicolumn{2}{c|}{$\delta_{WWH}/\delta_{WWH}^{(1)}-1$ [\%]} &
\multicolumn{2}{c|}{$\delta_{ZZH}/\delta_{ZZH}^{(1)}-1$ [\%]} \\ \cline{2-9}
\rule{0mm}{5mm} &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ \\ \hline
1/4 & $-11.68$ & $-11.88$ & $ -7.34$ & $ -8.65$ & $ -9.33$ & $ -9.35$ &
$-19.13$ & $-16.58$ \\
1/2 & $-10.63$ & $-11.72$ & $ -6.68$ & $ -8.34$ & $ -8.49$ & $ -9.24$ &
$-17.40$ & $-16.83$ \\
1 & $ -9.75$ & $-11.44$ & $ -6.12$ & $ -8.01$ & $ -7.78$ & $ -9.04$ &
$-15.96$ & $-16.76$ \\
2 & $ -9.00$ & $-11.11$ & $ -5.66$ & $ -7.67$ & $ -7.19$ & $ -8.79$ &
$-14.74$ & $-16.51$ \\
4 & $ -8.36$ & $-10.74$ & $ -5.26$ & $ -7.35$ & $ -6.68$ & $ -8.51$ &
$-13.70$ & $-16.15$ \\
\hline\hline
\end{tabular}
\end{table}
\begin{table}[ht]
\caption{Relative deviations (in \%) of $\Delta\bar\rho$, $\bar\delta_u$,
$\bar\delta_{WWH}$, and $\bar\delta_{ZZH}$ from the respective one-loop results
due to their corrections up to ${\cal O}(\alpha_s)$ and ${\cal O}(\alpha_s^2)$.
The renormalization scale dependence is investigated by choosing
$\mu=\xi\mu_t$, with $\xi$ variable.
}\label{tab:ms}
\medskip
\begin{tabular}{|c|c|c|c|c|c|c|c|c|} \hline\hline
\rule{0mm}{5mm}$\xi$ &
\multicolumn{2}{c|}{$\Delta\bar\rho/\Delta\rho^{(1)}-1$ [\%]} &
\multicolumn{2}{c|}{$\bar\delta_u/\delta_u^{(1)}-1$ [\%]} &
\multicolumn{2}{c|}{$\bar\delta_{WWH}/\delta_{WWH}^{(1)}-1$ [\%]} &
\multicolumn{2}{c|}{$\bar\delta_{ZZH}/\delta_{ZZH}^{(1)}-1$ [\%]} \\
\cline{2-9}
\rule{0mm}{5mm} &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ &
${\cal O}(\alpha_s)$ & ${\cal O}(\alpha_s^2)$ \\ \hline
1/4 & $-15.63$ & $-11.15$ & $-10.70$ & $ -8.25$ & $-12.96$ & $ -8.67$ &
$-24.09$ & $-15.15$ \\
1/2 & $-10.89$ & $-11.55$ & $ -6.87$ & $ -8.22$ & $ -8.71$ & $ -9.09$ &
$-17.77$ & $-16.56$ \\
1 & $ -8.96$ & $-11.19$ & $ -5.63$ & $ -7.79$ & $ -7.15$ & $ -8.86$ &
$-14.69$ & $-16.51$ \\
2 & $ -8.64$ & $-10.88$ & $ -5.82$ & $ -7.56$ & $ -7.11$ & $ -8.71$ &
$-13.49$ & $-16.16$ \\
4 & $ -9.24$ & $-10.85$ & $ -6.81$ & $ -7.70$ & $ -7.92$ & $ -8.84$ &
$-13.41$ & $-15.93$ \\
\hline\hline
\end{tabular}
\end{table}
In Tables~\ref{tab:os} and \ref{tab:ms}, we investigate the $\xi$ dependence of
$\delta_u$, $\delta_{WWH}$, and $\delta_{ZZH}$ and their $\overline{\mbox{MS}}$
counterparts, respectively.
{}For comparison, we also include the results for $\Delta\rho$ and
$\Delta\bar\rho$.
To be specific, we consistently evaluate these quantities to leading and
next-to-leading order in QCD and study their relative deviations from their
respective one-loop values, which we denote by the superscript (1),
{\it e.g.}, $\Delta\rho^{(1)}=N_cX_t$, {\it etc.}
Notice that the on-shell and $\overline{\mbox{MS}}$ results coincide at one
loop.
In our ${\cal O}(\alpha_s)$ analysis, we use the one-loop formula for
$\alpha_s(\mu)$ with $\Lambda_{\overline{MS}}^{(6)}=41$~MeV and omit the
${\cal O}(\alpha_s^2)$ terms in Eqs.~(\ref{mass}) and (\ref{mut}).
{}For the time being, let us concentrate on the entries for $\xi=1$ and
assess the effect of the QCD corrections as well as their scheme dependence.
We observe that, in both schemes, the QCD corrections are throughout negative,
even for $\bar\delta_u$ and $\bar\delta_{WWH}$, where the
${\cal O}(\alpha_s)$ and ${\cal O}(\alpha_s^2)$ terms are in part positive.
This is due to the fact that we consistently compute all QCD parameters,
{\it i.e.}, $\alpha_s(\mu)$, $m_t(\mu)$, and $\mu_t$, to the orders under
consideration.
The reduction in $x_t$, which occurs as an overall factor in the
$\overline{\mbox{MS}}$ formulae, happens to overcompensate the positive
effect of these particular coefficients.
Inclusion of the ${\cal O}(\alpha_s^2)$ terms in
$(\Delta\rho,\delta_u,\delta_{WWH},\delta_{ZZH})$ increases the size of the
QCD corrections by $(17,31,16,5)\%$, respectively.
In the $\overline{\mbox{MS}}$ case, the increments amount to
$(25,38,24,12)\%$ of the respective ${\cal O}(\alpha_s)$ corrections.
As might be expected, the scheme dependence of the QCD corrections to this
quadruplet of quantities is dramatically reduced, by $(68,55,71,80)\%$, as we
pass from ${\cal O}(\alpha_s)$ to ${\cal O}(\alpha_s^2)$.
Let us now also include the other $\xi$ values in our consideration.
Within each scheme, we determine the scale dependence of the QCD correction to
a given quantity by comparing its largest and smallest values in the interval
$1/4\le\xi\le4$.
As expected, the scale dependence is drastically decreased when we take the
${\cal O}(\alpha_s^2)$ terms into account, namely by $(66,37,68,87)\%$ and
$(89,87,93,86)\%$ in the on-shell and $\overline{\mbox{MS}}$ schemes,
respectively.
The exceptionally small reduction of the scale dependence in the case of
$\delta_u$ is due to the fact that $\delta_u$ has the smallest
${\cal O}(\alpha_s)$ term and the largest ${\cal O}(\alpha_s^2)$ term of all
four on-shell quantities.
\begin{table}[ht]
\caption{Coefficients of the correction factors in the form of Eq.~(\ref{kfac})
for the various Higgs-boson decay rates and production cross sections discussed
in the text.
In the last line, $x=B/A$, where $A$ and $B$ are given by Eq.~(\ref{int}),
and terms of ${\cal O}(x^2)$ have been neglected.
}\label{tab:k}
\medskip
\begin{tabular}{|c|c|c|c|} \hline\hline
$K$ & $C_1$ & $C_2$ & $C_3$ \\ \hline
$K_{\ell\ell H}$ & $7/3$ & $-1.797$ & $-16.201$ \\
$K_{WWH}$ & $-5/3$ & $-2.284$ & $-10.816$ \\
$K_{ZZH}$ & $-5/3$ & $-4.684$ & $-6.847$ \\
$K_1^{(\nu)}$ & $-2/3$ & $-7.420$ & $4.774$ \\
$K_1^{(\ell)}$ & $-1.272$ & $-5.249$ & $-4.445$ \\
$K_2^{(\nu\nu)}$ & $1/3$ & $6.261$ & $-53.330$ \\
$K_2^{(\nu\ell)}$ & $-0.272$ & $-14.025$ & $32.824$ \\
$K_2^{(\ell\ell)}$ & $-0.878$ & $-6.323$ & $0.113$ \\
$K_3^{(\ell\ell)}$ & $-0.878-2.353\,x$ & $-6.323+9.281\,x$ &
$0.113-39.416\,x$ \\
\hline\hline
\end{tabular}
\end{table}
In the remainder of this section, we shall stick to the on-shell scheme.
In Eqs.~(\ref{kllh}), (\ref{kwwh}), (\ref{kzzh}), (\ref{k1}), (\ref{k2}), and
(\ref{k3}), we have presented correction factors for various Higgs-boson
production cross sections and decay rates in terms of $\Delta\rho$, $\delta_u$,
$\delta_{WWH}$, and $\delta_{ZZH}$.
It is instructive to cast these correction factors into the generic form
\begin{equation}
\label{kfac}
K=1+C_1\Delta\rho^{(1)}\left[1+C_2a\left(1+{7\over4}aL\right)+C_3a^2\right],
\end{equation}
where $C_i$ $(i=1,2,3)$ are numerical coefficients.
Notice that we have kept the full $\mu$ dependence in Eq.~(\ref{kfac}).
We could have written Eqs.~(\ref{drhoos}), (\ref{duos}), (\ref{dwwhos}), and
(\ref{dzzhos}) in the same way.
The fact that the coefficient of $aL$ is universal may be understood by
observing that $K$ represents a physical observable, which must be RG invariant
through the order of our calculation, and that, to leading order of QCD, $K$
only implicitly depends on $\mu$, via $a$.
In fact, the coefficient of $aL$ is nothing but $\beta_0$ of Eq.~(\ref{beta}).
The outcome of this decomposition is displayed in Table~\ref{tab:k}.
In the case of $K_3^{(\ell\ell)}$, we have treated $x=B/A$, where $A$ and $B$
are defined in Eq.~(\ref{int}), as an additional expansion parameter and
discarded terms of ${\cal O}(x^2)$.
This is justified because, in practice, $|x|\ll1$,
{\it e.g.}, for $\sqrt s=300$~GeV and $M_H=100$~GeV, we find
$x\approx-5.233\cdot10^{-2}$.
While in the case of the three basic corrections, $K_{\ell\ell H}$,
$K_{WWH}$, and $K_{ZZH}$, $C_2$ and $C_3$ are both negative, this is not
in general so.
In fact, in all composite corrections, except for $K_1^{(\ell)}$, the
${\cal O}(\alpha_s^2)$ terms partially compensate the ${\cal O}(\alpha_s)$
ones.
In $K_2^{(\nu\nu)}$, we even find a counterexample to the heuristic rule
\cite{ks2} that, in the $G_F$ formulation of the on-shell scheme, the
${\cal O}(G_FM_t^2)$ terms get screened by their QCD corrections.
In the latter case, we also encounter a gigantic value of $C_3$.
Both features may be ascribed to the fact, that, in ${\cal O}(G_FM_t^2)$, the
$\delta_{ZZH}$ and $\Delta\rho$ terms of $K_2^{(\nu\nu)}$ largely cancel.
The extraordinarily large value of $C_3$ in $K_2^{(\nu\ell)}$ is also
accompanied by a suppression of $C_1$.
The $C_1$ values of $K_2^{(\ell\ell)}$ and $K_3^{(\ell\ell)}$ are relatively
small, too.
We are thus in the fortunate position that the leading high-$M_t$ corrections
to the $2\to3$ and $1\to4$ processes of Higgs-boson production and decay with a
$ZZH$ coupling, for which full one-loop calculations have not yet been
performed, are throughout quite small.
Thus, there is hope that the subleading fermionic corrections to these
processes will not drastically impair the situation.
However, the IBA does not provide us with any information on the bosonic
corrections.
\begin{table}[ht]
\caption{Full one-loop weak corrections (in \%) to various Higgs-boson decay
rates and production cross sections and their ${\cal O}(G_FM_t^2)$ terms.
In the last line, we have used $\protect\sqrt s=175$~GeV.
}\label{tab:full}
\medskip
\begin{tabular}{|c|c|c|c|} \hline\hline
Observable & $M_H$ [GeV] & ${\cal O}(\alpha)$ weak [\%] &
${\cal O}(G_FM_t^2)$ [\%] \\ \hline
$\Gamma(H\to\tau^+\tau^-)$ & 75 & 1.792 & 2.369 \\
$\Gamma(H\to\nu\bar\nu Z)$ & 105 & 1.275 & $-0.677$ \\
$\Gamma(H\to\ell^+\ell^-Z)$ & 105 & $-1.220$ & $-1.292$ \\
$\Gamma(Z\to\nu\bar\nu H)$ & 65 & 0.024 & $-0.677$ \\
$\Gamma(Z\to\ell^+\ell^-H)$ & 65 & 0.296 & $-1.292$ \\
$\sigma(e^+e^-\to ZH)$ & 75 & $-2.293$ & $-1.292$ \\
\hline\hline
\end{tabular}
\end{table}
In this context, it is interesting to revisit processes for which the full
one-loop weak corrections are known and to investigate in how far the
${\cal O}(G_FM_t^2)$ terms play a dominant r\^ole there.
Here, we are only interested in reactions which already proceed at tree level.
Specifically, we shall consider
$Z\to f\bar fH$ \cite{zffh} for $M_H=65$~GeV,
$H\to\tau^+\tau^-$ \cite{hff} and $e^+e^-\to ZH$ \cite{eezh} for $M_H=75$~GeV,
and $H\to f\bar fZ$ \cite{pr} for $M_H=105$~GeV, where $f=\nu,\ell$.
Our analysis of $\sigma(e^+e^-\to ZH)$ will refer to LEP2 energy,
$\sqrt s=175$~GeV.
In all these cases, the quantumelectrodynamical (QED) and weak corrections are
separately finite and gauge independent at one loop.
In Table~\ref{tab:full}, we compare the full one-loop weak corrections to these
processes with their ${\cal O}(G_FM_t^2)$ terms.
In the case of $H\to\tau^+\tau^-$, $H\to\ell^+\ell^-Z$, and $e^+e^-\to ZH$,
the ${\cal O}(G_FM_t^2)$ terms give a reasonably good account of the full
corrections, while they come out with the wrong sign in the other cases.
However, the full calculations for $M_H=65$~GeV give very small results
anyway.
On the other hand, the ${\cal O}(G_FM_t^2)$ term for $H\to\nu\bar\nu Z$ is
suppressed due to a partial cancellation between $\delta_{ZZH}$ and
$\Delta\rho$ in $K_1^{(\nu)}$ and cannot be expected to dominate the full
correction.
Whenever the full correction is known, it should be included on the right-hand
side of Eq.~(\ref{kfac}) with the ${\cal O}(G_FM_t^2)$ term subtracted.
In conclusion, the radiative corrections considered in Table~\ref{tab:full} all
appear to be well under control.
\section{Conclusions}
In this paper, we have presented the three-loop ${\cal O}(\alpha_s^2G_FM_t^2)$
corrections to the effective Lagrangians for the interactions of light Higgs
bosons with pairs of charged leptons, $W$ bosons, and $Z$ bosons in the SM.
While the demand for corrections in this order is certainly more urgent in the
gauge sector \cite{avd}, where precision test are presently being carried out,
our analysis is also interesting from a theoretical point of view, since it
allows us to recognize a universal pattern.
In addition to $\Delta\rho$, we have now three more independent observables
with quadratic $M_t$ dependence at our disposal for which the QCD expansion is
known up to next-to-leading order, namely $\delta_u$, $\delta_{WWH}$, and
$\delta_{ZZH}$.
In the on-shell scheme of electroweak and QCD renormalization, these four
electroweak parameters exhibit striking common properties.
In fact, the leading- and next-to-leading-order QCD corrections act in the
same direction and screen the ${\cal O}(G_FM_t^2)$ terms.
Even the sets of $\alpha_s/\pi$ and $(\alpha_s/\pi)^2$ coefficients each
lie in the same ball park.
{}For the choice $\mu=M_t$, the coefficients of $\alpha_s/\pi$ range between
$-1.797$ and $-4.684$, and those of $(\alpha_s/\pi)^2$ between $-6.847$ and
$-16.201$.
If we compare this with the corresponding coefficients of the ratio
$\mu_t^2/M_t^2$, which are $-2.667$ and $-11.140$,
then it becomes apparent that the use of the top-quark pole mass is the origin
of these similarities.
Here, $\mu_t=m_t(\mu_t)$, for which we have presented a closed two-loop
formula.
If we express the QCD expansions in terms of $\mu_t$ rather than $M_t$ and
choose $\mu=\mu_t$,
then the coefficients of $\alpha_s/\pi$ and $(\alpha_s/\pi)^2$ nicely group
themselves around zero;
they range from $-2.017$ to 0.870 and from $-3.970$ to $1.344$, respectively.
This indicates that the perturbation expansions converge more rapidly if we
renormalize the top-quark mass according to the $\overline{\mbox{MS}}$ scheme.
Without going into details, we would like to mention that the study of
renormalons \cite{ren} offers a possible theoretical explanation of this
observation.
Since the on-shell and $\overline{\mbox{MS}}$ results coincide in lowest
order, this does, of course, not imply that the QCD corrections are any
smaller in the $\overline{\mbox{MS}}$ scheme.
It just means that, as a rule, the ${\cal O}(G_FM_t^2)$ terms with $M_t$
replaced by the two-loop expression for $\mu_t$ are likely to provide fair
approximations for the full three-loop results.
Furthermore, we have demonstrated that, similarly to $\Delta\rho$, the scheme
and scale dependences of $\delta_u$, $\delta_{WWH}$, and $\delta_{ZZH}$ are
considerably reduced when the next-to-leading-order QCD corrections are taken
into account.
Armed with this information, we have made rather precise predictions for a
variety of production and decay processes of low-mass Higgs bosons at present
and future $e^+e^-$ colliders.
In all the cases considered here, the radiative corrections appear to be well
under control now.
\bigskip
\centerline{\bf ACKNOWLEDGEMENTS}
\smallskip\noindent
We would like to thank Bill Bardeen, Kostja Chetyrkin, and Michael Spira for
very useful discussions.
One of us (BAK) is indebted to the FNAL Theory Group for inviting him as a
Guest Scientist.
He is also grateful to the Phenomenology Department of the University of
Wisconsin at Madison for the great hospitality extended to him during a
visit when a major part of his work on this project was carried out.
|
\section{Introduction}
Several aspects of the collective transport associated with the phason
in charge- or spin-density waves (CDW or SDW) are still not well
understood. One of the intriguing phenomena is the electromechanical
effect observed in CDW\cite{Brill,Moz,Xi1,Xi2,Jac} and more recently in
SDW\cite{Brown}. First of all, most of the elastic moduli increase upon
entrance into the CDW or SDW state, often with a sharp dip at $T_c$, the
transition temperature. Second, some of the elastic moduli in CDW or
SDW soften when the density wave is depinned by an external electric
field in excess of the depinning threshold field $E_T$. Third, the change
in the elastic moduli due to depinning of the density wave depends on
the frequency $\omega$ of the flexural vibration\cite{Xi3}, and
decreases like $\omega^{-p}$ with $p\approx 1$. This behavior is similar
to the frequency dependence of the change in the dielectric constant
upon depinning in CDW and SDW\cite{Cava}.
We have shown earlier\cite{MV1} in the collisionless limit
that the hardening of the elastic
constants can be understood in terms of the reduction in the
quasiparticle screening of the ion potential due to the formation of
the density wave state. The electromechanical effect was interpreted
as an additional screening contribution from the collective mode of
the density wave condensate (phason) liberated by depinning. A later
extension of that theory to the experimentally more relevant
hydrodynamic limit\cite{VM1} did not modify the above picture
qualitatively. However in these papers it was assumed that the phonon
simply couples to the electronic density, and the effect of the
long-range Coulomb interaction was neglected.
The purpose of the present paper is twofold. First, we shall develop
the theory of the electron-phonon coupling for a strongly anisotropic
system, which will enable us to distinguish between the behavior of
transverse and longitudinal sound waves propagating in various
directions. Second, we shall include the effect of the long-range Coulomb
interaction following Kadanoff and Falko\cite{KF}. In Section II we
construct the electronic stress tensor (which couples to the deformation
tensor of the sound wave) for a quasi-one dimensional system following
the method of comoving frame\cite{Tsu}. In Section III we concentrate
on the quasiparticle contribution to the stress tensor correlation
functions corresponding to the pinned case. Section IV is devoted to
the examination of the coupling of the stress tensor to the phason,
which is relevant to the electromechanical effect. Our conclusions are
summarized in Section V.
A preliminary report
on this work has already been published elsewhere\cite{VM2}.
\section{Electron-phonon coupling}
Following Tsuneto\cite{Tsu} let us assume that both the ionic potential
and the electronic wavefunction are deformed in the
slowly varying sound field
${\bf u}({\bf r},t)={\bf u}\cos({\bf qr}-\omega t)$
imposed externally (extreme tight-binding limit):
\begin{eqnarray}
V({\bf r})\rightarrow&V[{\bf r}-{\bf u}({\bf r},t)]\nonumber\\
\psi({\bf r})\rightarrow&\psi[{\bf r}-{\bf u}({\bf r},t)]
(1+\nabla{\bf u})^{-1/2}.
\end{eqnarray}
This displacement is generated by the unitary operator
$U=\exp[-{\bf u}({\bf r},t)\nabla]$. Therefore the deformed
wavefunction is an eigenfunction of the transformed Hamiltonian
$h=Uh_0U^+$, where $h_0=\varepsilon(-i\nabla)$ with $\varepsilon
({\bf p})$ being the zone-periodic electronic energy spectrum.
The transformed Hamiltonian $h$ is then expanded in terms of the
deformation tensor $\nabla_iu_j$, which is much smaller than one, even
though the displacement ${\bf u}$ itself may be many times of the
lattice constant for sound propagation. We obtain $h=h_0+h_{el-ph}$,
where the Hamiltonian for the electron-phonon coupling is given by
\begin{equation}
h_{el-ph}=\sum_{i,j}(\nabla_iu_j)i\nabla_jv_i(-i\nabla),
\end{equation}
with ${\bf v}({\bf p})=\partial\varepsilon({\bf p})/\partial {\bf p}$,
the velocity of the Bloch electron.
The matrix element of $h_{el-ph}$ between Bloch states is evaluated as
\begin{equation}
\langle {\bf p}+{\bf q}|h_{el-ph}|{\bf p}\rangle =
-i\sum_{i,j}q_ju_i\tau_{ij}({\bf p}),
\end{equation}
where the stress tensor
\begin{equation}
\tau_{ij}({\bf p})=mv_i({\bf p})v_j({\bf p}),
\end{equation}
and $m$ is the bare electron mass. This expression generalizes
the stress
tensor used for an isotropic metal\cite{KF}.
In orthorombic symmetry the sound wave polarized in the ${\bf i}$
direction and propagating in the ${\bf j}$ direction couples to the
$\tau_{ij}$ component of the stress tensor, and in order to determine
the effect of that coupling on the frequency (or velocity) of the
sound, we have to evaluate the appropriate stress tensor correlation
function $\langle [\tau_{ij},\tau_{ij}]\rangle$. Once this is known,
the renormalized sound velocity can be calculated in the weak
coupling limit as
\begin{equation}
c=c_0\{1-\langle [\tau,\tau ]\rangle /2Mc_0^2\},
\end{equation}
where $c_0$ is the sound velocity without electron-phonon coupling,
$M$ is the ion mass, and for clarity we have suppressed the indeces
both for the stress tensor component and for the sound velocity.
For a highly anisotropic ($t_a\gg t_b\gg t_c$) tight-binding
dispersion
\begin{equation}
\varepsilon({\bf p})=-2t_a\cos(ap_x)-2t_b\cos(bp_y)
-2t_c\cos(cp_z)-\mu,
\end{equation}
widely used for CDW and SDW materials\cite{Yam}, the velocity and
stress tensor components are easily obtained, and their values on
the open Fermi surface can conveniently be expressed by the
component of ${\bf p}$ perpendicular to the chains (${\bf x}$
direction). However, since the Green`s functions for CDW and SDW
are usually written in the left-right spinor representation\cite{VM1},
involving measuring momenta from $\pm{\bf Q}/2$ with
the density wave wavevector ${\bf Q}=(2p_F,\pi/b,\pi/c)$,
we should express the stress tensor elements in a compatible
manner. It turns out, that for each stress tensor component the
term proportional to the unit matrix dominates, therefore we have
\begin{eqnarray}
\tau_{xx}=&mv_F^2\{1-[t_b/t_a\sin(ap_F)]^2
\sin^2(bp_y)\}\nonumber\\
\tau_{yy}=&mv_y^2[1+\cos(2bp_y)]\nonumber\\
\tau_{xy}=&mv_Fv_y\sqrt{2}\cos(bp_y).
\end{eqnarray}
Here $p_F$ is the Fermi momentum,
$v_F=2at_a\sin(ap_F)$ is the Fermi velocity in the chain
direction, while $v_y=\sqrt{2}bt_b\ll v_F$ is a typical velocity
in the perpendicular direction. In the followings we restrict
our study to the $x-y$ plane, since behavior involving the $z$
direction should be similar to that of the $y$ direction. We note,
that all components of the stress tensor depend on momentum through
the combination $\varphi=bp_y$ only.
\section{Pinned density waves}
In this section we consider sound propagation with no applied
electric field. The density wave is pinned, therefore the condensate
is unable to contribute to correlation functions, including that
for the stress tensor. Mathematically this situation can be
simulated by setting the coupling of the stress tensor
(and of the density) to the phason
to zero. Then the stress tensor couples only to the density
fluctuations, resulting in the well known Coulomb screening\cite{KF}:
\begin{equation}
\langle [\tau,\tau]\rangle=\langle[\tau,\tau]\rangle_0-
{\langle[\tau,n]\rangle_0\langle[n,\tau]\rangle_0\over
q^2/4\pi e^2+\langle[n,n]\rangle_0}.
\end{equation}
Here $n$ stands for the electronic particle density, and
$\langle[A,B]\rangle_0$ denotes a correlation function, in which
only the effect of impurity scattering is taken into account.
The density correlator $\langle[n,n]\rangle_0$ in the presence of
impurity scattering was evaluated in \cite{VM1}. A straightforward
extension of that calculation confirms that under the circumstances
of the sound experiment ($lq\ll 1$, where $l$ is the mean free path)
the stress tensor correlator has two distinct contributions:
\begin{equation}
\langle[\tau,\tau]\rangle_0=\langle\tau(\varphi)\rangle_{\varphi}^2
\langle[n,n]\rangle_0+\langle[\delta\tau(\varphi)]^2\rangle_{\varphi}
\langle[n,n]\rangle_0^{no vertex}.
\end{equation}
The first contribution features
only the average
of the stress tensor $\langle\tau(\varphi)\rangle_{\varphi}=
(2\pi)^{-1}\int_{-\pi}^\pi d\varphi\tau(\varphi)$, and is
proportional to the diffusive (vertex corrected) density correlator.
The second contribution containes only the fluctuating part
of the stress tensor component
$\delta\tau(\varphi)=\tau(\varphi)-\langle\tau(\varphi)
\rangle_{\varphi}$, and therefore it is proportional to the density
correlator $\langle[n,n]\rangle_0^{no vertex}$
calculated without vertex corrections. According to the same argument,
only the average of the stress tensor couples to the density,
therefore
\begin{equation}
\langle[\tau,n]\rangle_0=
\langle[n,\tau]\rangle_0=\langle\tau(\varphi)\rangle_{\varphi}
\langle[n,n]\rangle_0.
\end{equation}
Combining Eqs.(8)-(10) we see that in the long wavelength limit
appropriate for the sound experiment ($q\approx 1/L$, where $L$
is the sample size), the average part of the stress tensor
($s$-wave component, proportional to density) is
completely screened out by the Coulomb interaction, and only the
fluctuating part contributes to the correlation function:
\begin{equation}
\langle[\tau,\tau]\rangle=\langle[\delta\tau(\varphi)]^2
\rangle_{\varphi}\langle[n,n]\rangle_0^{no vertex}.
\end{equation}
The situation here is the same as in the electronic Raman scattering,
where the long-range Coulomb interaction suppresses the density
(charge) fluctuations, and only non $s$-wave channels
survive\cite{AG}.
The evaluation of $\langle[n,n]\rangle_0^{no vertex}$ can be done
starting with the results of \cite{VM1}. The calculation is
rather technical, therefore we delegate it to the Appendix, and
we give here the results only. Without vertex corrections there is
no diffusion pole, and both the wavenumber ${\bf q}$ and the
frequency $\omega$ could be set to zero. However we keep a finite
(but small) frequency for finite imaginary part of the correlator:
\begin{equation}
\langle[n,n]\rangle_0^{no vertex}=N_F{i\tilde\Gamma_{qp}(1-\tilde f)
\over \omega+i\tilde\Gamma_{qp}},
\end{equation}
where $N_F$ is the density of states at the Fermi surface. The
corresponding "unrenormalized" condensate density $\tilde f$
(for the general formula see Eq.(32) in the Appendix) and
quasiparticle damping $\tilde\Gamma_{qp}$ are evaluated in two
limiting cases, close to $T_c$ and close to zero temperature as
\begin{eqnarray}
\tilde f(T\rightarrow T_c)=&-2({\Delta\over 4\pi T})^2
\psi^{\prime\prime}({1\over 2}+{\Gamma\over 2\pi T})\approx
7\zeta(3)({\Delta\over 2\pi T})^2\nonumber\\
\tilde f(T\rightarrow 0)=&1-3\pi\alpha/16,
\end{eqnarray}
and
\begin{eqnarray}
\tilde\Gamma_{qp}(T\rightarrow T_c)=&2\Gamma\nonumber\\
\tilde\Gamma_{qp}(T\rightarrow 0)=&{9\pi\over 32}{\Delta
\alpha^{8/3}\over u_0^2(5-4\alpha^{2/3})}{G\over T}e^{G/T}.
\end{eqnarray}
Here $\Delta$ is the density wave order parameter, $\Gamma=
\Gamma_F+\Gamma_B/2$ is a combination of the impurity forward
and backscattering rate, $\alpha=\Gamma/\Delta$, $u_0^2=1-
\alpha^{2/3}$ and $G=\Delta u_0^3$ is the density wave gap.
As it is seen from Eq.(13), $\tilde f$ increases linearly in
$(T_c-T)$ below $T_c$, but it is slightly less than one at $T=0$
($\Gamma$ is usually an order of magnitude smaller than $T_c$).
Nevertheless, the temperature dependence of the sound velocity
in the pinned case
will still be qualitatively the same as in the collisionless
limit\cite{MV1}. The relative change of the sound velocity
compared to the normal state ($c_n$) is easily obtained from Eqs.(5),
(11) and (12) as
\begin{equation}
(c-c_n)/c_0=\lambda\tilde f,
\end{equation}
where the effective coupling
\begin{equation}
\lambda={N_F\over 2Mc_0^2}\langle[\delta\tau(\varphi)]^2
\rangle_{\varphi}.
\end{equation}
Using Eq.(7), these effective couplings for the various
sound waves are
\begin{eqnarray}
\lambda_{xx}=&{N_F(mv_F^2)^2\over 16Mc_0^2}[{t_b\over t_a\sin(
ap_F)}]^4\nonumber\\
\lambda_{yy}=&{N_F(mv_y^2)^2\over 4Mc_0^2}\nonumber\\
\lambda_{xy}=&{N_F(mv_Fv_y)^2\over 2Mc_0^2}.
\end{eqnarray}
Since $v_y/v_F\approx t_b/t_a\approx 1/10$ in many quasi one
dimensional materials, we expect that the relative increase of
the sound velocity below $T_c$ will be a factor $10^2$ smaller
for longitudinal sound than for transverse sound.
\section{Electromechanical effect}
If an external electric field in excess of the threshold field
$E_T$ of the nonlinear conductivity is applied in the chain direction,
then the condensate is depinned and is able to contribute to
various correlation functions\cite{MV1}. The best known example
is of course
the conductivity itself, but the situation is the same for the
stress tensor correlator as well. The collective contribution to
$\langle[\tau,\tau]\rangle$ can be obtained if we allow both the
stress tensor and the density to couple to the phason (in the
previous section this coupling was blocked due to pinning).In this
case the stress tensor correlator has another contribution
$\langle[\tau,\tau]\rangle^{coll}$ in addition to the one
calculated in the previous section, namely:
\begin{equation}
\langle[\tau,\tau]\rangle^{coll}={U\langle[\tau,\delta\Delta]
\rangle^{Coul}\langle[\delta\Delta,\tau]\rangle^{Coul}\over
1-U\langle[\delta\Delta,\delta\Delta]\rangle^{Coul}}.
\end{equation}
Here $U$ is the on-site Coulomb repulsion responsible for the
formation of the SDW state (for CDW it should be replaced by the
phonon propagator, but that does not affect our conclusions),
$\delta\Delta$ is the phase fluctuation of the order
parameter, and
$\langle[A,B]\rangle^{Coul}$ is the correlation function of
quantities $A$ and $B$ including the effect of the long-range
Coulomb interaction (like in Eq.(8)). As we have seen earlier, in
the long wavelength limit this yields:
\begin{equation}
\langle[A,B]\rangle^{Coul}=\langle[A,B]\rangle_0-\langle[A,n]
\rangle_0\langle[n,B]\rangle_0/\langle[n,n]\rangle_0.
\end{equation}
First we consider if the allowed coupling to the phason actually
takes place for various sound waves. According to Eq.(18) we
need to examine $\langle[\tau,\delta\Delta]\rangle^{Coul}$,
which is given by Eqs.(19) and (10) as
\begin{equation}
\langle[\tau,\delta\Delta]\rangle^{Coul}=\langle[\tau,\delta
\Delta]\rangle_0-\langle\tau(\varphi)\rangle_{\varphi}\langle
[n,\delta\Delta]\rangle_0.
\end{equation}
The density-phason correlator $\langle[n,\delta\Delta]\rangle_0$
was evaluated in \cite{VM1}. Here we only reiterate that result
in the limit of experimental interest $lq\ll c_0/v_F$ (dynamic
limit):
\begin{equation}
\langle[n,\delta\Delta]\rangle_0=iN_F{\langle\zeta(\varphi)
\rangle_{\varphi}\over 2\Delta}f_d,
\end{equation}
where in our two dimensional geometry the wavenumber ${\bf q}$
appears in $\zeta(\varphi)=v_Fq_x+\sqrt{2}v_yq_y\cos\varphi$,
and the condensate density in the dynamic limit is given by:
\begin{eqnarray}
f_d(T\rightarrow T_c)=&{\Delta^2\over 2\pi T\Gamma_B}
\psi^\prime({1\over 2}+{\Gamma\over 2\pi T})\approx
{\pi\Delta^2\over 4T\Gamma_B}\nonumber\\
f_d(T\rightarrow 0)=&1.
\end{eqnarray}
We recall that the same $f_d$ appears in the current-phason
correlator $\langle[j,\delta\Delta]\rangle_0$, as well as in
the phason propagator\cite{MV2}.
Note that $f_d$ increases from zero much faster below $T_c$
than $\tilde f$ does, and that at zero temperature it saturates
exactly to $1$. The stress tensor-phason correlator can be
calculated similarly:
\begin{equation}
\langle[\tau,\delta\Delta]\rangle_0=iN_F{\langle\tau(\varphi)
\zeta(\varphi)\rangle_{\varphi}\over 2\Delta}f_d.
\end{equation}
Now we shall examine Eq.(20) for different sound waves in order
to determine if there is a collective contribution to the
corresponding stress tensor correlator. We consider longitudinal
and transverse sound waves propagating in the ${\bf x}$ and
${\bf y}$ directions. Clearly, the second (screening) term in
Eq.(20) is nonzero only for the longitudinal sound
propagating in the chain direction (${\bf q}\parallel {\bf u}
\parallel {\bf x}$),
in which case it completely cancels the first term, leading to
no collective contribution. The other longitudinal sound
propagating perpendicular to the chains (${\bf q}\parallel
{\bf u}\parallel {\bf y}$) does not couple to the phason either,
because $\langle\tau_{yy}(\varphi)\cos\varphi\rangle_{\varphi}
=0$ (see Eq.(7)). The coupling of the transverse wave
propagating in the chain direction (${\bf q}\parallel {\bf x}$
and ${\bf u}\parallel {\bf y}$) is controlled by $\langle
\tau_{xy}(\varphi)\rangle_{\varphi}=0$, yielding again no
collective contribution. This means that in all of the above
three cases there will be no electromechanical effect.
For the rest of this section we will concentrate on the only
interesting case, when the transverse sound propagates
perpendicular to the chains (${\bf q}\parallel {\bf y}$ and
${\bf u}\parallel {\bf x}$). In this case there will be coupling
to the phason, since
\begin{equation}
\langle[\tau_{xy},\delta\Delta]\rangle^{Coul}=
\langle[\tau_{xy},\delta\Delta]\rangle_0=
iN_F{f_d\over 2\Delta}mv_Fv_y^2q_y
\end{equation}
is nonzero. Now we have to consider the denominator in Eq.(18).
Since $\langle[n,\delta\Delta]\rangle_0=0$ for $q_x=0$,
therefore $\langle[\delta\Delta,\delta\Delta]\rangle^{Coul}=
\langle[\delta\Delta,\delta\Delta]\rangle_0$, and we can use
the result for the phason propagator calculated in \cite{VM1},
which in our case reduces to
\begin{equation}
1-U\langle[\delta\Delta,\delta\Delta]\rangle_0=
{UN_Ff_d\over (2\Delta)^2}[(v_yq_y)^2-i\omega\Gamma_{ph}].
\end{equation}
Here $\Gamma_{ph}$ is the phason damping rate and is given by
\begin{eqnarray}
\Gamma_{ph}(T\rightarrow T_c)=&2\Gamma_B\nonumber\\
\Gamma_{ph}(T\rightarrow 0)=&{8u_0^2T\Gamma_B\over
3\alpha^{4/3}G}e^{-G/T}.
\end{eqnarray}
The phason damping freezes out for low temperature, and
approaches $2\Gamma_B$ at $T_c$. Note the discontinuity in
$\Gamma_{ph}$ at $T_c$ (approaching from above $\Gamma_{ph}
\approx 2\pi^3T/7\zeta(3)$), which is the consequence of the
finite order parameter $\Delta$ below $T_c$ exceeding almost
immediately the energy scale set by $\omega$ and $vq$.
We are now able to write down the total correlation function
for this sound wave in the unpinned case. Using Eqs.(11),(18),
(24) and (25) we obtain
\begin{eqnarray}
\langle[\tau_{xy},\tau_{xy}]\rangle=&N_F(mv_Fv_y)^2
\times\nonumber\\
\times&\left [
{i\tilde\Gamma_{qp}(1-\tilde f)\over \omega+i\tilde
\Gamma_{qp}}+{(v_yq_y)^2f_d\over(v_yq_y)^2-i\omega\Gamma_{ph}}
\right ].
\end{eqnarray}
The above equation means that the collective contribution
(the second term on the right hand side) due
to the moving condensate recovers some of the screening of the
ion motion lost because of the decrease in the number of quasi
particles. In fact, at zero temperature it overcompensates
somewhat, since $f_d>\tilde f$. Therefore we expect the
electromechanical effect for the transverse sound polarized
in the chain direction only. According to Eq.(27), close to
$T_c$ the electromechanical effect on the sound velocity should
be much smaller than the temperature effect, while at low
temperatures the softening is somewhat bigger than the hardening
was upon cooling. Although the collective contribution in Eq.(27)
does have a frequency dependence, it does not appear to describe
the suppresion of the electromechanical effect when the
frequency is increased\cite{Xi3}. This is rather puzzling, although
the above results for sound propagation may not translate
literally to the flexural experiment.
\section{Conclusions}
We have derived for the first time the appropriate stress tensor for
quasi-one dimensional electron systems.
Under conditions of a sound velocity measurement ($lq\ll 1$) the
long-range Coulomb interaction has a simple role to suppress the
$s$-wave channel as in the theory of electronic Raman scattering.
In highly anisotropic systems like Bechgaard salts the increase of the
sound velocity in the density wave state is two orders of magnitude
smaller for longitudinal sound waves than for transverse ones. We also
find, that a sound wave with polarization perpendicular to the chain
direction can not couple to the phason, because the density wave
condensate can only move parallel to the chains. The coupling of the
longitudinal sound propagating in the chain direction is screened away
by the Coulomb interaction, which leaves us only the transverse
sound wave propagating perpendicular to the chains as the one which
does couple to the phason, and shows the electromechanical effect.
Recently Britel {\it et.al.} measured the elastic constant
$c_{44}$
of (Ta$_{1-x}$Nb$_x$Se$_4$)$_2$I at 15MHz in the geometry
${\bf u}\parallel {\bf x}$\cite{Mon}, and found a relative
reduction of order $10^{-4}$ in the presence of an electric field
approximately $10E_T$. This seems to be consistent with our
analysis, although the observed effect appears to be a little
too small.
\section*{Acknowledgments}
This publication is sponsored by the U.S.-Hungarian Science and Technology
Joint Fund in cooperation with the National Science Foundation and the
Hungarian Academy of Sciences under Project No. 264/92a, which enabled one
of us (A.V.) to enjoy the hospitality of USC. This work is also supported
in part by the National Science Foundation under Grant No. DMR92-18317 and
by the Hungarian National Research Fund under Grants No. OTKA15552 and
T4473.
\section*{Appendix}
We evaluate here the density correlation function $\langle[n,n]
\rangle_0^{no vertex}$ (Eqs.(12)-(14) in the text). We start with the
corresponding thermal product (See \cite{VM1}):
\begin{equation}
\langle[n,n]\rangle_0^{no vertex}=N_F[1-\pi T\sum_n F(iu_n,iu_n^\prime)],
\end{equation}
where $N_F$ is the density of states, and $u_n$ and $u_n^\prime$ are
related to the Matsubara frequencies $\omega_n$ and $\omega_n^\prime=
\omega_{n-\nu}$ by
\begin{equation}
\omega_n/\Delta=u_n[1-\alpha(u_n^2+1)^{-1/2}],
\end{equation}
with $\Delta$ the order parameter, $\alpha=\Gamma/\Delta$ and
$\Gamma=\Gamma_F+\Gamma_B/2$ a combination of the impurity forward
and backscattering rates. Neglecting vertex corrections in the
relevant formulas in \cite{VM1} leads to:
\begin{equation}
F(u,u^\prime)={1+{1+uu^\prime\over (1-u^2)^{1/2}(1-u^{\prime 2})
^{1/2}}\over \Delta[(1-u^2)^{1/2}+(1-u^{\prime 2})^{1/2}]},
\end{equation}
where $u$ and $u^\prime$ are analytic continuations of $iu_n$ and
$iu_n^\prime$, and in the absence of the diffusion pole the
wavenumber ${\bf q}$ was already set to zero.
While evaluating the correlation function we follow the standard
method\cite{Zit}. Expanding up to linear order in $\omega$ we obtain
\begin{equation}
\langle[n,n]\rangle_0^{no vertex}=N_F(1-\tilde f+i\omega I),
\end{equation}
where
\begin{equation}
\tilde f={\pi T\over\Delta}\sum_n(u_n^2+1)^{-3/2},
\end{equation}
and
\begin{eqnarray}
I=&{1\over 2\Delta}\int_G^\infty{dE\over 2T}\cosh^{-2}\left ({E\over
2T}\right )\times\nonumber\\
\times&\left [{h^\prime\over {\rm Re}(1-u^2)^{1/2}}-{\rm Re}
(1-u^2)^{-3/2}\right ].
\end{eqnarray}
Here $h^\prime=(1/2)[1+(|u|^2+1)/|u^2-1|]$, and $G=\Delta u_0^3$ is
the gap with $u_0^2=1-\alpha^{2/3}$. The above equations can be
evaluated in two limiting cases, with temperature close to $T_c$ and
close to zero, and can be brought to the form of Eqs.(12)-(14) in
the text.
|
\section{Introduction}
For an agent\footnote{Unless otherwise indicated, we will use the term
``agent'' to refer to both individual agents and sets of agents.} to
be able to perform an action, it must satisfy both the physical and
knowledge preconditions of that action \cite{Moore:85,Morgenstern:87}.
For example, for an agent to pick up a particular tower of blocks, it
must (1)~know how to pick up towers in general, (2)~be able to
identify the tower in question, and (3)~ have satisfied the (physical)
preconditions or constraints associated with picking up towers (e.g.,
it must have a free hand). These conditions must hold whether the
agent is planning an action on its own or is involved in a
collaborative planning effort with other agents.
In this paper, we provide an axiomatization of knowledge preconditions
for the SharedPlan model of collaborative activity
\cite{Grosz-Sidner:90,LGS:90,Grosz-Kraus:93}. This model draws upon
past work \cite{Moore:85,Morgenstern:87}, but adapts it to the
collaborative situation. We briefly describe the SharedPlan framework
in Section~\ref{sp}, and then, in Section~\ref{know-prec}, present our
axiomatization of knowledge preconditions. In Section~\ref{role}, we
demonstrate the use of knowledge preconditions in accounting for
information-seeking subdialogues, such as those in Figure~\ref{ac}. We
then compare our approach to the alternative accounts
\cite{Litman-Allen:87,Lambert-Carberry:91,Ramshaw:91}.
\begin{small}
\begin{figure}[hbtp]
\hspace*{0.2in}
\psfig{figure=knowp-d.eps,width=3.1in}
\caption{\small Information-Seeking Subdialogues [Grosz, 1974]\label{ac}}
\end{figure}
\end{small}
\vspace*{-3ex}
\section{SharedPlans}
\label{sp}
The SharedPlan formalism is a mental-state model of collaborative
plans with roots in Pollack's~\shortcite{Pollack:90} work on
single-agent plans. For a group of agents $GR$ to have a {\em full
SharedPlan} (FSP) for an act $\alpha$, they must satisfy the
requirements given in Figure~\ref{sp-reqts}. When the agents have
satisfied only a subset of these requirements, they are said to have a
{\it partial SharedPlan} (PSP).\footnote{This description of a PSP is
only a rough, though useful, approximation to the formal definition
given by Grosz and Kraus~\shortcite{Grosz-Kraus:93}.} The bracketed
terms in Figure~\ref{sp-reqts} indicate the operators used by Grosz
and Kraus~\shortcite{Grosz-Kraus:93} to formalize each requirement.
Requirement~(1) in Figure~\ref{sp-reqts} refers to the agents' {\em
recipe} \cite{Pollack:90} for $\alpha$. Recipes are modeled in Grosz
and Kraus's definitions as sets of constituent acts and constraints.
To perform an act $\alpha$, an agent must perform each constituent act
in $\alpha$'s recipe according to the constraints of that recipe.
Actions themselves may be further decomposed into act-types and
parameters. We will represent an action $\alpha$ as a term of the
form $\bar{\alpha}(p_1,\ldots,p_n)$ where $\bar{\alpha}$ represents
the act-type of the action and the $p_i$ its parameters.
\begin{figure}[t]
\begin{small}
For a group of agents $GR$ to have an FSP for $\alpha$
\begin{enumerate}
\item $GR$ must have mutual belief of a {\em recipe} for $\alpha$
\item For each single-agent constituent act of the recipe, there must
be an agent $G_{\beta_i} \in GR$, such that
\begin{enumerate}
\vspace*{-1ex}
\item $G_{\beta_i}$ intends to perform $\beta_i$ {\bf [Int.To]}
\item $G_{\beta_i}$ believes that it can perform $\beta_i$ {\bf [BCBA]}
\item $G_{\beta_i}$ has a full individual plan for $\beta_i$ {\bf [FIP]}
\item The group $GR$ has mutual belief of (2a)-(2c)
\item Each member of $GR$ intends that $G_{\beta_i}$ succeed {\bf [Int.Th]}
\end{enumerate}
\item For each multi-agent constituent act of the recipe, there must
be a subgroup of agents $GR_{\beta_i} \subseteq GR$ such that
\begin{enumerate}
\item $GR_{\beta_i}$ mutually believe that
they can perform $\beta_i$ {\bf [MBCBAG]}
\item $GR_{\beta_i}$ has a full SharedPlan for $\beta_i$ {\bf [FSP]}
\item The group $GR$ has mutual belief of (3a)-(3b)
\item Each member of $GR$ intends that $GR_{\beta_i}$ succeed {\bf [Int.Th]}
\end{enumerate}
\end{enumerate}
\vspace*{-2ex}
\caption{\small FSP Requirements \label{sp-reqts}}
\end{small}
\end{figure}
\section{Knowledge Preconditions}
\label{know-prec}
Grosz and Kraus~\shortcite{Grosz-Kraus:93} use the operators BCBA
(read ``believes can bring about'') and MBCBAG (read ``mutually
believe can bring about group'') to formalize respectively
requirements~(2b) and~(3a) in Figure~\ref{sp-reqts}. Although these
operators are intended to specify the conditions under which an agent
is able to perform an action, their definitions explicitly require
only that an agent satisfy the physical preconditions or constraints
associated with an action to be able to perform it. Because an agent
is not truly capable of performing an act unless it possesses the
appropriate knowledge, the definitions of BCBA and MBCBAG must be
augmented with an axiomatization of knowledge preconditions. The
following observations made by Morgenstern~\shortcite{Morgenstern:87},
but recast in our terminology, must be represented in such an
axiomatization:
\begin{enumerate}
\vspace*{-1ex}
\item Agents need to know recipes for the acts they
perform.
\vspace*{-1.5ex}
\item All agents have some primitive acts in their
repertoire.
\vspace*{-1.5ex}
\item Agents must be able to identify the parameters of the acts they
perform.
\vspace*{-1.5ex}
\item Agents may know only some descriptions of an act.
\vspace*{-3.5ex}
\item Agents know that the knowledge necessary for complex acts
derives from that necessary for their component acts.
\vspace*{-1.5ex}
\end{enumerate}
Our axiomatization of knowledge preconditions is based on
Morgenstern's observations, but adapted to the requirements of
individual and shared mental-state plans.\footnote{A comparison of our
formalization with those of Morgenstern~\shortcite{Morgenstern:87} and
Moore~\shortcite{Moore:85} can be found elsewhere \cite{KEL:94}.} We
use the predicates {\em has.recipe\/} and {\em id.params\/} to
represent explicitly observations~(1) and (3) above. The remaining
observations are implicitly represented by the way in which these two
knowledge precondition relations are defined. Observation~(2) is
modeled as the base case of {\em has.recipe}, and observation~(5) is
modeled by the use of {\em has.recipe\/} within the recursive plan
definitions.
Observation~(4) requires that the knowledge precondition relations be
intensional, rather than extensional; within their scope it should not
be possible to freely substitute one representation of an action for
another. We thus define {\em has.recipe\/} and {\em id.params\/} to
hold of action {\em descriptions}, rather than actions. Action
descriptions are intensional objects; one action description can be
substituted for another only if the descriptions are the same. For
example, although $555$-$1234$ and $phone$-$number(speech$-$lab)$ may
be extensionally equivalent, the descriptions $\rule[1.2ex]{.05em}{.24em}^{\hspace*{-0.05em}\rule[.6ex]{.25em}{.05em}}\!\,{555}$-${1234}\:\rule[1.2ex]{.05em}{.24em}^{\hspace*{-.26em}\rule[.6ex]{.25em}{.05em}}$
and $\rule[1.2ex]{.05em}{.24em}^{\hspace*{-0.05em}\rule[.6ex]{.25em}{.05em}}\!{phone}$-${number(speech}$-${lab)}\:\rule[1.2ex]{.05em}{.24em}^{\hspace*{-.26em}\rule[.6ex]{.25em}{.05em}}$ are not. By
convention, we will omit the corner quote notation in what follows and
simply take the appropriate arguments of the predicates to represent
action descriptions rather than actions.
Although Morgenstern's observations are most naturally expressed
informally in terms of knowledge, we formalize them using belief to
allow for the possibility of an agent's being incorrect. Although it
is true that an agent cannot successfully act unless its beliefs about
recipes and parameters are correct, having to {\it know} the recipes
and parameters is too strong a requirement for collaborating agents
\cite{KEL:94}.
\subsection{Determining Recipes: {\em has.recipe}}
\label{recipe}
For an agent to be able to perform an act $\alpha$, it must know {\em
how\/} to perform $\alpha$; i.e., it must have a recipe for the act.
The relation $has.recipe(G,\alpha,R,T)$ is used to represent that
agent $G$ has a recipe $R\,$ for an act $\alpha$ at time $T$. It is
formalized as follows:
\begin{footnotesize}
\begin{tabbing}
{}~~~~~~~~$has.recipe(G,\alpha,R,T) \Leftrightarrow$\\
{}~~~~~~~~(1)~~~~$[$\=$basic.level(\alpha)~~\wedge$\\
\>$BEL(G,basic.level(\alpha),T) \wedge R=R_{Empty}]~~\vee$ \\
{}~~~~~~~~(2)~~~~$[\neg basic.level(\alpha)~~\wedge$\\
{}~~~~~~~~(2a) \>$R=\{\beta_i,\rho_j\}~~\wedge$\\
{}~~~~~~~~(2a1) \>~~~~$\{$\=$[|G| = 1 \wedge BEL(G,R \in
Recipes(\alpha),T)]~~\vee$\\
{}~~~~~~~~(2a2) \>\>$[|G| > 1 \wedge MB(G,R \in Recipes(\alpha),T)]\}]$
\end{tabbing}
\end{footnotesize}
Clause~(1) of the definition models Morgenstern's second observation,
namely that agents do not need a recipe to perform a {\em basic-level\/}
action, i.e., one executable at will
\cite{Pollack:90}.\footnote{Basic-level actions are by their
nature single-agent actions.} For non-basic-level actions
(Clause~(2)), the agent of $\alpha$ (either a single agent (2a1) or a
group of agents (2a2)) must believe that some set of acts, $\beta_i$,
and constraints, $\rho_j$, constitute a recipe for
$\alpha$.
\subsection{Identifying Parameters: {\em id.params}}
\label{params}
An agent must also be able to identify the parameters of an act
$\alpha$ to be able to perform it. For example, if an agent is told
{\it ``remove the flywheel,''} as in the dialogue of Figure~\ref{ac},
the agent must be able to identify the flywheel in question. The
relation $id.params(G,\alpha,T)$ is used to represent that agent $G$
can identify the parameters of act $\alpha$ at time $T$. If $\alpha$
is of the form $\bar{\alpha}(p_1,...,p_n)$, then
$id.params(G,\alpha,T)$ is true if $G$ can identify each of the $p_i$.
To do so, $G$ must have a description of each $p_i$ that is suitable
for $\bar{\alpha}$. The relation {\em id.params\/} is defined as
follows:
\begin{footnotesize}
\begin{tabbing}
{}~~~~$id.params(G,\bar{\alpha}(p_1,\ldots,p_n),T) \Leftrightarrow$\\
{}~~~~~~~$(\forall i, 1 \leq i \leq n)~has.sat.descr(G,p_i,{\cal
F}(\bar{\alpha},p_i),T)$
\end{tabbing}
\end{footnotesize}
The ability to identify an object is highly context dependent. For
example, as Appelt points out~\shortcite[pg. 200]{Appelt:85}, ``the
description that one must know to carry out a plan requiring the
identification of `John's residence' may be quite different depending
on whether one is going to visit him, or mail him a letter.'' The
function ${\cal F}$ in the above definition is an oracle function
intended to model the context-dependent nature of parameter
identification. This function returns a suitable {\em identification
constraint\/} \cite{Appelt-Kronfeld:87} for a parameter $p_i$ in the
context of an act-type $\bar{\alpha}$. For example, in the case of
sending a letter to John's residence, the constraint produced by the
oracle function would be that John's residence be described by a
postal address.
The relation $has.sat.descr(G,P,C,T)$ holds of an agent $G$, a
parameter description $P$, an identification constraint $C$, and a
time $T$, if $G$ has a suitable description, as determined by $C$, of
the object described as $P$ at time $T$. To formalize this relation,
we utilize Kronfeld's~\shortcite{Kronfeld:86} notion of an
individuating set. An agent's individuating set for an object is a
maximal set of terms such that each term is believed by the agent to
denote that object. For example, an agent's individuating set for
John's residence might include its postal address as well as an
identifying physical description such as ``the only yellow house on
Cherry Street.'' To model individuating sets, we introduce a function
$IS(G,P,T)$; the function returns an agent $G$'s individuating set at
time $T$ for the object that $G$ believes can be described as $P$.
This function is based on similar elements of the formal language that
Appelt and Kronfeld~\shortcite{Appelt-Kronfeld:87} introduce as part
of their theory of referring. The function returns a set that
contains $P$ as well as the other descriptions that $G$ has for the
object that it believes $P$ denotes.
For an agent to suitably identify a parameter described as $P$, the
agent must have a description, $P^\prime$, of the parameter such that
$P^\prime$ is of the appropriate sort. For example, for an agent to
visit John's residence, it is not sufficient for the agent to believe
that the description ``John's residence'' refers to the place where
John lives. Rather, the agent needs another description of John's
residence, one such as ``the only yellow house on Cherry Street,''
that is appropriate for the purpose of visiting him. To model an
agent's ability to identify a parameter (described as $P$) for some
purpose, we thus require that the agent have an individuating set for
the parameter that contains a description $P^\prime$ such that
$P^\prime$ satisfies the identification constraint that derives from
the purpose. The definition of {\em has.sat.descr\/} is thus as
follows:\footnote{A more precise account of what it means to be able
to identify an object is beyond the scope of this paper; for further
details, see the discussions by Hobbs~\shortcite{Hobbs:85}, Appelt and
Kronfeld~\shortcite{Appelt:85,Kronfeld:86,Appelt-Kronfeld:87}, and
Morgenstern~\shortcite{Morgenstern-thesis}.}
\begin{footnotesize}
\begin{tabbing}
{}~~~~$has.sat.descr(G,P,C,T) \Leftrightarrow$\\
{}~~~~~~$\{$\=$[$\=$|G| = 1~\wedge$\\
\>\>~~$(\exists P^\prime) BEL(G,[$\=$P^\prime \in
IS(G,P,T)~~\wedge$\\
\>\>\>${\it suff\!.for.id}(C,P^\prime)],T)]~~\vee$\\[0.5ex]
\>$[|G| > 1~\wedge$\\
\>\>~~$(\exists P^\prime) MB(G,(\forall G_j \in G)
[$\=$P^\prime \in IS(G_j,P,T)~~\wedge$\\
\>\>\>${\it suff\!.for.id}(C,P^\prime)],T)]\}$
\end{tabbing}
\end{footnotesize}
The predicate ${\it suff\!.for.id}(C,P^\prime)$ is true if the
constraint $C$ applies to the parameter description $P^\prime$. The
oracle function ${\cal F}(\bar{\alpha},p_i)$ in {\em id.params\/}
produces the appropriate identification constraint on $p_i$ given
$\bar{\alpha}$.
\section{The Role of Knowledge Preconditions in Language Processing}
\label{role}
We now show how the requirements of knowledge preconditions can be
used in discourse processing. Our model of discourse processing is
based on the theory of discourse structure proposed by Grosz and
Sidner~\shortcite{Grosz-Sidner:86}. According to their theory,
discourse structure consists of three interrelated components: a
linguistic structure, an attentional state, and an intentional
structure. The linguistic structure consists of discourse
segments\footnote{The term discourse segment is a generalization of
the term subdialogue. Whereas the term discourse segment applies to
all types of discourse, the term subdialogue is reserved for segments
that occur within dialogues.} and an embedding relationship among
them; the bold rule in Figure~\ref{ac} indicates the linguistic
structure of that discourse. Attentional state is an abstraction of
the discourse participants' focus of attention; it serves as a record
of those entities that are salient at any point in the discourse.
Intentional structure is comprised of discourse segment purposes and
their interrelationships, particularly that of {\em dominance}. A
discourse segment purpose, or DSP, is a
Gricean-like~\shortcite{Grice:69} intention that leads to the
initiation of a discourse segment. One DSP is dominated by another if
the satisfaction of the first provides part of the satisfaction of the
second.
Intentional structure plays a central role in discourse processing; an
agent's comprehension of the utterances in a discourse relies on the
recognition of this structure \cite{Grosz-Sidner:86}. Grosz and
Sidner~\shortcite{Grosz-Sidner:90} proposed SharedPlans to provide a
basis for recognizing intentional structure. They argued that
discourses are fundamentally collaborative, and hence that a model of
{\it shared\/} plans provides a more appropriate basis for discourse
processing than a model of single-agent plans. However, the
connection between SharedPlans and intentional structure was never
specified.
\subsection{SharedPlans as Intentional Structure}
We have developed a model of discourse processing that provides that
connection \cite{KEL:94}. Figure~3 illustrates the role of
SharedPlans in modeling intentional structure. Each segment of a
discourse has an associated SharedPlan. The purpose of the segment is
taken to be intention that (Int.Th \cite{Grosz-Kraus:93}) the
discourse participants form that plan. This intention is held by the
agent who initiates the segment. In what follows, we will refer to
that participant as the initiating conversational participant or ICP;
the other participant is the OCP \cite{Grosz-Sidner:86}. Dominance
relationships between DSPs are modeled using {\it subsidiary}
relationships between SharedPlans. One plan is subsidiary to another
if the completion of the first plan contributes to the completion of
the second. Subsidiary relationships are discussed in more detail in
Section~\ref{recog}
The utterances of a discourse are understood in terms of their
contribution to the SharedPlans associated with the segments of the
discourse. Those segments that have been completed at the time of
processing an utterance have a full SharedPlan (FSP) associated with
them (e.g., segment~(2) in Figure~3), while those that have not have a
partial SharedPlan (PSP) (e.g., segments~(1) and~(3) in Figure~3).
\psfig{figure=struct-ijcai.eps,width=3.37in}
\begin{center}
Figure 3: Modeling Intentional Structure
\end{center}
For each utterance of a discourse, an agent must determine whether the
utterance begins a new segment of the discourse, contributes to the
current segment, or completes it \cite{Grosz-Sidner:86}. For an
utterance to begin a new segment, it must indicate the initiation of a
subsidiary plan. This case is described in further detail below. For
an utterance to contribute to the current segment, it must advance the
partial SharedPlan associated with the segment towards completion.
That is, it must establish one of the beliefs or intentions required
for the discourse participants to have a full SharedPlan, but missing
from their current partial SharedPlan. For an utterance to complete
the current segment, it must indicate that the purpose of that segment
has been satisfied. For that to be the case, the SharedPlan
associated with the segment must be an FSP rather than a PSP. That
is, all of the beliefs and intentions required of an FSP, as indicated
in Figure~\ref{sp-reqts}, must have been established over the course
of the segment.
A detailed description of the implemented algorithms used in modeling
each of these cases can be found elsewhere \cite{KEL:94}. Here, we
focus on the use of knowledge preconditions in accounting for the
initiation of information-seeking subdialogues. We use the dialogue
in Figure~\ref{ac} as an example and assume the role of the Expert
(participant ``E'') in analyzing the discourse.\footnote{For
simplicity of exposition, we will take participant ``E'' to be female
and participant ``A'' to be male.} The dialogue in Figure~\ref{ac}
was extracted from a larger discourse in which the Expert and
Apprentice (participant ``A'') are collaborating on removing the pump
of an air compressor. We thus take the purpose of the larger
discourse to be
\begin{footnotesize}
\begin{tabbing}
{}~~~DSP$_1$=$Int.Th(e,FSP(\{a,e\},remove(pump(ac1),\{a\})))$\footnotemark\\
{}~~~~~~where $ac1$ represents the air compressor the agents\\
{}~~~~~~~~are working on.
\end{tabbing}
\end{footnotesize}
\footnotetext{We have omitted the time parameters for simplicity of
exposition.}
\subsection{Accounting for the Initiation of New Discourse Segments}
\label{recog}
To make sense of an {\em utterance}, an agent must provide an
explanation for it in the form of an answer to the question, ``Why did
the speaker say that to me?'' \cite{Sidner-Israel:81}. An OCP must
provide a similar explanation for an ICP's initiation of a new
discourse segment. This explanation takes the form of an answer to
the question ``Why does the ICP want to engage in a segment with
purpose DSP$_j$ at this point in our discourse?''; i.e., ``How is
DSP$_j$ related to what we were talking about before?'' Subsidiary
relationships between SharedPlans provide the basis for modeling the
OCP's reasoning.
One plan is subsidiary to another if the completion of the first plan
contributes to the completion of the second. The most basic example
of this relationship occurs within the FSP definition itself. As
indicated in Figure~\ref{sp-reqts}, a full plan for an act $\alpha$
includes full plans for each subact in $\alpha$'s recipe as components
(requirements~(2c) and (3b)). The plans for the subacts thus
contribute to the plan for $\alpha$ and are therefore subsidiary to
it.
Subsidiary relationships may also arise in response to the other
requirements of the FSP definition. For example, as discussed in
Section~\ref{know-prec}, the BCBA operator used to model
requirement~(2b) specifies that to be able to perform an act $\alpha$,
an agent must (1)~have a recipe for $\alpha$ ($has.recipe$), (2)~be
able to identify the parameters of $\alpha$ ($has.sat.descr$), and
(3)~have satisfied the constraints associated with performing
$\alpha$. The first of these requirements provides an explanation for
the first subdialogue in Figure~\ref{ac}.
The purpose of this subdialogue is represented as\footnote{We describe
a method for recognizing DSPs of this form elsewhere \cite{KEL:94}.}
\begin{footnotesize}
\begin{tabbing}
{}~~~~~DSP$_2$=\=$Int.Th(a,FSP(\{a,e\},$\\
\>~~$Achieve(has.$\=$recipe(a,$\\[-0.5ex]
\>\>$remove(flywheel(ac1),\{a\}),R))))$
\end{tabbing}
\end{footnotesize}
and can be glossed as ``the Apprentice
intends that the agents collaborate on his obtaining a recipe for the
act of removing the flywheel of the air compressor.'' To account for
the Apprentice's initiation of this subdialogue, the Expert must
determine the relationship of DSP$_2$ to the purpose of the agents'
preceding discourse, namely DSP$_1$. In this case, the Expert can
reason that the Apprentice wants to engage in the subdialogue to
obtain a recipe for the act of removing the flywheel so that he will
be able to perform that act as part of the agents' SharedPlan to
remove the pump. The plan in DSP$_2$ is thus subsidiary to that in
DSP$_1$ by virtue of a knowledge precondition requirement of the
latter plan.
Figure~\ref{kp1-graphic} illustrates this analysis. Each box in the
figure corresponds to a discourse segment and contains the SharedPlan
used to model the segment's purpose. The SharedPlans are labeled so
as to be co-indexed with the DSPs discussed above. The arrows
indicate subsidiary relationships between SharedPlans, as explained by
the text that adjoins them. When plan Pj is subsidiary to plan Pi,
DSP$_j$ is dominated by DSP$_i$.
\begin{small}
\begin{figure*}[bt]
\hspace*{1in}
\psfig{figure=knowp-ijcai2.eps,width=5.7in}
\caption{Analysis of the First Subdialogue in
Figure~\protect\ref{ac}\label{kp1-graphic}}
\end{figure*}
\end{small}
The information represented within each SharedPlan in
Figure~\ref{kp1-graphic} is separated into two parts. Those beliefs
and intentions that have been established at the time of the analysis
are shown above the dotted line, while those that remain to be
established, but that are used in determining subsidiary
relationships, are shown below the line. The index in square brackets
to the right of each constituent indicates the FSP requirement from
which the constituent arose.
As indicated in Figure~\ref{kp2-graphic}, the initiation of the second
subdialogue in Figure~\ref{ac} is explained similarly.
This time, however, it is the need to identify parameters of acts
(requirement~(2) above) that leads to the initiation of the
subdialogue. In addition, the parameter in question is a parameter of
an act in a subsidiary individual plan of the Apprentice's.
\begin{small}
\begin{figure*}[tb]
\hspace*{1in}
\psfig{figure=knowp-2-ijcai2.eps,width=4.9in}
\caption{Analysis of the Second Subdialogue in
Figure~\protect\ref{ac}\label{kp2-graphic}}
\end{figure*}
\end{small}
\subsection{Discussion}
In this paper, we have shown that information-seeking subdialogues may
be explained on the basis of knowledge precondition requirements. Our
account of such subdialogues fits within a general framework for
discourse processing in which the purpose of a subdialogue is modeled
using a SharedPlan and is related to the purposes of other
subdialogues based on the requirements of the FSP definition.
Elsewhere~\cite{KEL:94}, we show that correction and subtask
subdialogues, among others, may also be accounted for in this manner.
In contrast, alternative plan-based accounts of dialogue understanding
introduce multiple types of plans to account for the utterances in a
discourse. For example, Litman and Allen~\shortcite{Litman-Allen:87},
propose the use of two types of plans to model clarification and
correction subdialogues: {\em discourse plans\/} and {\em domain
plans}. Domain plans represent knowledge about a task, while
discourse plans represent conversational relationships between
utterances and plans. Litman and Allen provide operators for the
following discourse plans:
\begin{itemize}
\vspace*{-1ex}
\item INTRODUCE-PLAN: introduce a new plan for discussion
\vspace*{-1ex}
\item CONTINUE-PLAN: execute the next action in a plan
\vspace*{-1ex}
\item TRACK-PLAN: talk about the execution of an action
\vspace*{-1ex}
\item MODIFY-PLAN: introduce a new plan by modifying a previous one
\vspace*{-1ex}
\item CORRECT-PLAN: correct a plan
\vspace*{-1ex}
\item IDENTIFY-PARAMETER: identify a parameter of an action in a plan
\end{itemize}
Under our approach, the recognition of discourse plans is unnecessary.
The {\em fact} that a speaker is using an utterance to, for example,
introduce a plan, or track a plan, or identify a parameter, need not
be explicitly recognized for the purposes of utterance interpretation.
Furthermore, we would argue that such facts are not intended to be
recognized (cf. Grice~\shortcite{Grice:69}). Rather, they simply fall
out of recognizing the relationship of an utterance to the current
discourse structure, i.e., the currently active SharedPlans. For
example, INTRODUCE-PLAN corresponds to initiating a new discourse
segment, CONTINUE- or TRACK-PLAN to contributing to the current
segment, and IDENTIFY-PARAMETER to initiating a new segment to satisfy
a $has.sat.descr$ knowledge precondition requirement. Although the
initiation of a new SharedPlan corresponds to the initiation of a new
discourse segment under our approach, it is the SharedPlan that must
be recognized and not a discourse plan that refers to that SharedPlan.
Lambert and Carberry~\shortcite{Lambert-Carberry:91} have extended
Litman and Allen's approach by introducing a third type of plan. {\em
Problem-solving plans}, such as BUILD-PLAN and INSTANTIATE-VARS, are
used to model the process by which agents construct domain plans.
Under our approach, the need to explicitly recognize problem-solving
plans is also avoided. The fact that an agent is building a plan or
instantiating a variable is a byproduct of understanding an utterance
by relating it to the current discourse structure. BUILD-PLAN
corresponds to initiating a new discourse segment to satisfy a
$has.recipe$ knowledge precondition requirement, while
INSTANTIATE-VARS corresponds to initiating one to satisfy a
$has.sat.descr$ requirement. Unlike Lambert and Carberry's approach,
however, and Litman and Allen's as well, our approach actually
recognizes this structure. The other approaches are essentially
utterance-to-utterance based and thus do not recognize discourse
segments as separate units.
Ramshaw~\shortcite{Ramshaw:91} has added a different third type of
plan, {\em exploration plans}, to Litman and Allen's two types.
Exploration plans are intended to model the process by which agents
explore courses of actions. Although we have not yet incorporated
such reasoning into our model, we hypothesize that the exploration of
plans can be modeled, without the introduction of a new plan type, by
reasoning about an agent's {\it potential intentions} and the process
by which they become full-fledged intentions
\cite{Grosz-Kraus:93,BIP:88}.
These alternative approaches share an important property that
distinguishes them from our approach; they take a {\em data-structure}
view of plans, rather than a {\em mental phenomenon} view
\cite{Pollack:90}. Whereas data-structure plans are essentially
``recipes-for-action,'' mental phenomenon plans are a ``structured
collection of beliefs and intentions''
\cite[pg. 77]{Pollack:90}.\footnote{Although Lambert and
Carberry~\shortcite{Lambert-Carberry:91} adopt
Pollack's~\shortcite{Pollack:90} terminology in presenting their
theory, their ``plans'' are not mental state plans in Pollack's
sense.} Data-structure plans thus describe {\em what} an agent is
doing with an utterance, but not {\em why} the agent is doing it. For
example, although the constraints of Litman and Allen's
IDENTIFY-PARAMETER discourse plan force the plan to be related to
another plan that involves the parameter to be identified,
IDENTIFY-PARAMETER does not explain why this information is desired;
it does not capture that agents need to know parameters {\em to be
able to} perform acts involving them.\footnote{Although Litman and
Allen {\em describe} IDENTIFY-PARAMETER as providing ``a suitable
description for a term instantiating a parameter of an action such
that the hearer is then able to execute the action''
\cite[pg. 173]{Litman-Allen:87}, the IDENTIFY-PARAMETER operator
itself does not include a {\em formalization} of the last condition,
i.e., that the parameter description should enable the execution of
the action.} It thus fails to model the essential knowledge
precondition nature of identifying a parameter. Although it is
possible to impose a mental phenomenon interpretation on top of a
data-structure plan, doing so does not result in a mental phenomenon
plan \cite{Pollack:90}. Saying that G1 {\em intends} to IDENTIFY a
PARAMETER fails to address why G1 intends to do so.
The need to explain an utterance is not unique to interpretation.
Moore and Paris~\shortcite{Moore-Paris:93} have shown that a similar
need exists in generation. In particular, they have argued that
RST-based text plans must be augmented with intentional structure.
Otherwise, a system has no record of why it said what it did and is
thus unable to respond effectively if a hearer does not understand or
accept its utterances.
\section{Conclusion}
In this paper, we have presented an axiomatization of knowledge
preconditions for the SharedPlan model of collaborative activity
\cite{Grosz-Kraus:93}. We have also shown how the requirements of
knowledge preconditions can be used to account for information-seeking
subdialogues in discourse. Our account of this phenomenon fits within
a general framework for discourse processing in which SharedPlans and
relationships among them are used to model the intentional component
of Grosz and Sidner's~\shortcite{Grosz-Sidner:86} theory of discourse
structure. Unlike the alternative approaches, our approach recognizes
and makes use of discourse structure. In addition, it does not require
the introduction of new plan types.
\begin{small}
|
\section{Introduction}
Since the wormhole-based time machine was proposed \cite{Tho}
much efforts have been directed towards finding a mechanism that could
"protect causality" and destroy such a time machine. One of the most
popular ideas is that the creation of
the time machine might be prevented by quantum effects since as it is
claimed in \cite{KimT} "at any event in spacetime, which can be joined to
itself by a closed null geodesic, the vacuum fluctuations of a massless
scalar field should produce a divergent renormalized stress-energy
tensor". The considerations leading to such a claim I shall call hereafter
"FKT approach". \par
In essence, the FKT approach amounts to the following \cite{Fro,KimT} (see
also \cite{Kli}). The vacuum
expectation value of the stress-energy tensor
$\langle T_{\mu\nu}\rangle$ of the field $\phi$ in the (multiply
connected) spacetime $M$ containing a time machine is found by
applying some
differential operator $D_{\mu\nu}$ to the Hadamard function
\begin{equation}
\label{DefG}
G^{(1)}(X,X')\equiv\bigl\langle\{\phi(X),\phi(X')\}\bigr\rangle.
\end{equation}
To find $G^{(1)}$ it is proposed to use the formula
\begin{equation}
\label{Cov}
G^{(1)}(X,X^{\prime})=G^\Sigma\equiv
\sum_n{\widetilde G}^{(1)}(X, \gamma^n X^{\prime}) .
\end{equation}
Here $\widetilde G^{(1)}$ is the Hadamard function of $\phi$ in the
spacetime $\widetilde M$, which is the universal covering space for $M$,
and $\gamma^n X\in \widetilde M$ is the $n$-th inverse image of $X\in M$
($\gamma^0X$ is identified in (\ref{Cov}) with $X$). The advantage of the
use of (\ref{Cov}) is that $\widetilde G^{(1)}$ is supposed to have the
Hadamard form:
\begin{equation}
\label{Had}
{\widetilde G}^{(1)}(X,X')=\tilde u\sigma^{-1}+\tilde v\ln{|\sigma|}
+\mbox{nonsingular terms}
\end{equation}
where $\sigma$ is half the square of the geodesic distance between $X$ and
$ X'$, and $\tilde u,\ \tilde v$ are some smooth functions. We might think
thus that
\begin{equation}
\begin{array}{lcl}\displaystyle{
\langle T_{\mu\nu}\rangle^{{\rm ren}}_{M\vphantom{\widetilde M}}
}&=
\displaystyle{\langle T_{\mu\nu}\rangle^{{\rm ren}}_{\widetilde M} + \sum
_{n\not=0}\lim_
{ X'\to X\atop{}} D_{\mu\nu}{\widetilde G}^{(1)}
(X, \gamma^n X^{\prime})}\\
\strut&&\\&\to&
\displaystyle{
\langle T_{\mu\nu}\rangle^{{\rm ren}}_{\widetilde M} + \sum
_{n\not=0}\;\lim_
{X'\to X\atop X\to\, \mbox{\scriptsize horizon}} D_{\mu\nu}
(\tilde u\sigma_n^{-1}+\tilde v\ln{|\sigma_n|}).
}\end{array}
\label{Tren}
\end{equation}
Here $\sigma_n\equiv \sigma(X,\gamma^nX')$ and the subscript "ren"
(renormalized) has appeared because renormalization of
$\langle {\bf T}\rangle_{\widetilde M}$ and of
$\langle {\bf T}\rangle_{M\vphantom{\widetilde M}}$
requires substraction of the same terms. The last series in (\ref{Tren})
diverges (since $\sigma_n\to0$, when $X$ approaches the horizon), so the
conclusion is made that the appearance of a closed
timelike curve must be prevented (unless some effects of quantum gravity
remedy the situation) by the infinite increase of the energy density.
\smallskip \par
The goal of this paper is to state that there is actually {\em no\/} reasons
to expect that the energy density diverges at the Cauchy horizon in the
general case. In particular, we cite a few examples in which the
stress-energy tensor of the massless scalar field (as found in the usual way
--- through quantization directly in $M$) remains bounded as the horizon
is approached. These examples are not of course anywhere near an adequate
model of the time machine (for example, the wave equation on the
two-dimensional cylinder has no solutions, except for constant, continuous
at the Cauchy horizon). However, they can well serve as counterexamples to
statements like that cited above. So it seems important to find out
whether they indicate only some loop-holes in the FKT approach, which can
then be used in the general case, or whether this approach must be fully
revised. Our analysis in Section~2 shows that the latter is true --- by
several independent reasons one cannot extract any information from expression
(\ref{Tren}). In Section~3 some special cases are considered necessary to
support statements from Section~2.
\section{Analysis}
\subsection{Going to the universal covering}
Formula (\ref{Cov}), combined with some implicit assumptions serves as a
basis for
the overall FKT approach since one cannot use (\ref{Had}) in multiply
connected spacetimes, where $\sigma$ is not defined. To discuss its validity
and to reveal these assumptions
let us first state the simple fact that most properties of the Hadamard
function, {\em including\/} the validity of (\ref{Had}), depends on the
choice of the vacuum appearing in definition (\ref{DefG}). So,
formulas like (\ref{Cov}) are meaningless until we specify the vacuua
$|0\rangle $ and $|\tilde 0\rangle $ in $M$ and ${\widetilde M}$
respectively. We come thus to the problem of great importance in our
consideration --- how, given $|0\rangle $, one could determine corresponding
$|\tilde 0\rangle $?
The above-mentioned assumptions concern just this problem. They must
be something like the following
\begin{enumerate}
\item For any vacuum on $M$ there exists a vacuum $|\tilde 0\rangle $ on $
\widetilde{M}$ such that (\ref{Cov}) holds.
\item The function $\widetilde{G}^{(1)}$ corresponding to $|\tilde 0\rangle $
has the "Hadamard form" (\ref{Had}).
\item ${G}^{(1)}$ determines $\widetilde{G}^{(1)}$ uniquely.
\end{enumerate}\par
The validity of Assumption 1 is almost obvious in the simplest cases (see
below), but it was not proven in the general
case. (One can meet the references to \cite{Ban} in this connection. Note,
however, that the functions $K_C$ which stand there in the analog of our
formula (\ref{Cov}), are actually not defined%
\footnote{I am
grateful to Dr. Parfyonov, who explained to me this issue.} in our case,
i.~e.\ when $|\Gamma|=\infty$.)
Assumption 2 seems still more arbitrary. The validity of (\ref{Had}) was
proven not for {\em any\/} state, but only for some specific class of states
(see \cite[Sect.~2c]{Ful}) and there is no reason to believe that our $%
|\tilde0\rangle$ belongs just to this class.
Assumption 3 is definitely untrue. In the following section we construct
as an example a
class of vacuua $|\tilde0\rangle_f$ such that (\ref{Cov}) is satisfied for
any $f$ while $\widetilde{G}_f$ differ for different $f$. This
nonuniqueness is far from harmless. As we argue below it makes, in fact,
expression (\ref{Tren}) meaningless.
\subsection{The expression for the stress-energy tensor}
Expression (\ref{Tren}) is the main result of the FKT approach
and (\ref{Cov}) is needed only to justify it. So let us state first that
\begin{enumerate}
\item (\ref{Tren}) does not follow (or, at least, does not follow
immediately) from (\ref{Cov}), since
\begin{enumerate}
\item To write $\lim D_{\mu \nu }\sum \widetilde G^{(1)}=\sum \lim D_{\mu
\nu } \widetilde G^{(1)}$ without a special proof one must be sure that
the series
$\sum \widetilde G^{(1)}$ and $\sum D_{\mu \nu }\widetilde G^{(1)}$ converge
uniformly, while it is clear that they do not (at least as long as (\ref{Had}%
) holds). This nonuniformity manifests itself, in particular, in the fact
that, in general, one cannot drop the nonsingular terms in
$\widetilde G^{(1)}$. In Subsection~$3.2$ we shall show that the last
series in (\ref{Tren}) can diverge {\em off\/} the Cauchy horizon though
(\ref{Had},\ref{Cov}) hold and $G^{(1){\rm ren}}$ (and $\langle {\bf
T}\rangle^{{\rm ren}}$) are smooth there.
\item Even when $|\tilde 0\rangle $ belongs to the above-mentioned class, (%
\ref{Had}) is proven not for {\em any\/} $X,\ X^{\prime }$, but only for $%
X^{\prime }$ lying in the ''sufficiently small'' neighborhood of $X$. It is
necessary, in particular, that $\sigma (X,X^{\prime })$ would be defined
uniquely. To provide this in Ref. \cite{Ful}, for example, $X$ and $%
X^{\prime }$ are required not to lie respectively near points
$\underline{\vphantom{y}x},
\mbox{ and }\underline{y}$ connected by a null geodesic with a
point conjugate to $\underline{\vphantom{y}x}$ before $\underline{y}$. To
violate this condition for the points $X'$ and $\gamma X'$ it suffices to
separate the mouths of the wormhole widely enough and to fill the space
between them with the conventional matter \cite[Prop. 4.4.5]{HawEl}.
\end{enumerate}
\vbox{Thus, we see that (\ref{Tren}) must be regarded as an
independent assumption. We can, however, neither use nor check it in view
of the aforementioned ambiguity:}
\item Like the Hadamard function,
$\langle T_{\mu\nu}\rangle^{{\rm ren}}_{\widetilde M}$
depends on which vacuum we choose, while from the FKT standpoint all
vacuua $|\tilde0\rangle$ satisfying (\ref{Cov}) are equivalent. This
equivalence is of a fundamental nature --- the only physical object is the
spacetime $M$, while $\widetilde M$ and $|\tilde0\rangle$ are some
auxiliary matters and as long as (\ref{Cov}) holds we cannot apply any
extraneous criteria to distinguish among them. So, we have no way
of determining what to substitute in (\ref{Tren}) as
$\langle T_{\mu\nu}\rangle^{{\rm ren}}_{\widetilde M}$.
In Subsection~$3.2$ we shall see that choosing different $|\tilde0\rangle$
(even when $\widetilde M$ is a part of the Minkowski plane)
one can make $\langle T_{\mu\nu}\rangle^{{\rm ren}}_{\widetilde M}$ finite
or infinite at the horizon at will.\par
\end{enumerate}
Let me note in passing that there is no point in using (\ref{Tren}) unless
we decide that $|\tilde0\rangle$ is among the very "good" and convenient
vacuua. For an arbitrary $|\tilde0\rangle$,
it is not a bit easier to find $\langle {\bf T}\rangle^{{\rm
ren}}_{\widetilde M}$ than $\langle {\bf T}\rangle^{{\rm
ren}}_{M\vphantom{\widetilde M}}$.
\subsection{Interpretation}
Suppose that $\langle\Psi|{\bf T}|\Psi\rangle^{{\rm ren}}_M$ for some
$|\Psi\rangle $ does diverge
at the Cauchy horizon. Suppose further that it is $\langle {\bf
T}\rangle^{{\rm ren}}_M$ that stands in the right side of the Einstein
equations (though it is not obvious, see \cite{Bir} for the literature and
discussion). Does this really mean that owing to the quantum effects the
time machine $M$ cannot be created? I think that the answer is negative.
It well may be that
$\langle\Phi|{\bf
T}|\Phi\rangle^{{\rm ren}}_M$ does not diverge for some other state $
|\Phi\rangle$ (example see in
Subsection~$3.3$).
Why must we restrict ourselves to the state $|\Psi\rangle$? To prove that
the Einstein equations and QFT are incompatible in $M$ one must have
proven that the expected stress-energy tensor tends to
infinity for {\em any\/} quantum state, or at least for any state satisfying
some reasonable physical conditions (say, stability).
\section{Examples}
Let us find the expectation value of the stress-energy tensor in a few
specific cases. We restrict our consideration to the two-dimensional
cylinder $M$ obtained from the plane $(\tau ,\chi )$ by identifying
$\chi\leadsto \chi +H $ and endowed with the metric
\begin{equation}
\label{Metr}
ds^2=C(-d\tau^2+d\chi^2)=C\,dudv .
\end{equation}
Here $u\equiv\chi-\tau,\ v\equiv\chi+\tau$; $C$ is a smooth function on $M$.
To find in the ordinary way $\langle {\bf T}\rangle $ for the free real
scalar field $\phi $
\begin{equation}
\label{Weq}\Box \phi =0,\qquad \phi (\chi +H,\tau )
\cases{\phi(\chi,\tau)&for the non-twisted field\cr
-\phi(\chi,\tau)&for the twisted field\cr}
\end{equation}
we must first of all specify the vacuum we consider. That is we must choose
a linear space of solutions of (\ref{Weq}) and an "orthonormal" basis \cite
{Bir} $U=\{u_n\}$ in it. In particular, this will define the Hadamard
function:
$$
G^{(1)}(X,X')=\sum_n u_n(X)u^*_n(X')+\mbox{complex conjugate}.
$$
A possible choice of $U$ for the non-twisted field is
\begin{equation}
u_n=|4\pi n|^{-1/2}e^{2\pi i H^{-1}(n\chi -|n|\tau )} \quad
n=\pm1,\pm2\dots \label{ConVac}
\end{equation}
The vacuum $|0\rangle_C$ defined by (\ref{ConVac}) (the "conformal" vacuum) is
especially attractive as the expressions for the Hadamard function $G_C^{(1)}$
and for the stress-energy tensor $\langle {\bf T}\rangle _C $ are already
obtained (see \cite[the neighborhood of formula (6.211)]{Bir}):
\begin{eqnarray}
\nonumber\langle T_{ww}\rangle _C^{\rm ren} &=&
-\frac{\pi\epsilon}{12H^2}+\frac{1}{24\pi}
\left[
\frac{C,_{ww}}{C}-\frac{3}{2}\frac{{C,_w}^2}{C^2}\label{Tcon
\right]
,\quad w=u,v\\
&&\strut\\
\nonumber\langle T_{uv}\rangle _C^{\rm ren}&=&
\langle T_{vu}\rangle _C^{\rm ren}=-RC/(96\pi).
\end{eqnarray}
Here $\epsilon=-1/2$ or 1 depending on whether $\phi$ is twisted
or untwisted and $R$ is the curvature of $M$.
Though the absence of a solution corresponding to $n=0$ in (\ref{ConVac})
may seem artificial, it is, in
fact, an inherent feature of $|0\rangle_C$, which is to describe the vacuum
of $\phi$ as a massless limit of the "natural" vacuum of a massive field (cf
\cite[below (4.220)]{Bir}). One could start, however, from another vacuum
for the massive field and arrive at another theory (see below) with the
basis $U^{\prime}$
$$
U^{\prime}=
U\cup u_0\equiv
(2H)^{-1/2}(F\tau+i/F) .
$$
Where the real constant $F$ is a free parameter. Choosing different
$F\neq0$ we
obtain different vacuua $|0\rangle_F$ and Hadamard functions $G^{(1)}_F$.
It is easy to see that
\begin{equation}
\label{GF}G_F^{(1)}=G_C^{(1)}+\frac{F^2}{H}\tau\tau^{\prime}+const.
\end{equation}
\subsection{Two-dimensional time machines in the conformal vacuum state}
As a first example let us consider the Misner spacetime, which is the
quadrant $\alpha <0,$ $\beta >0$ of the Minkowski plane $%
ds^2=d\alpha d\beta $ with points identified by the rule $(\alpha _{0,}\beta
_0)\mapsto (A\alpha _{0,}\beta _0/A)$. The coordinate transformation%
$$
u=-W^{-1}\ln |W\alpha |,\quad v=W^{-1}\ln (W\beta )
$$
delivers the isometry between Misner space and $M$ with%
$$
C=e^{2W\tau },\qquad H=W^{-1}\ln A .
$$
$W$ here is an arbitrary parameter with dimension of mass. Substituting this
in (\ref{Tcon}) we immediately find%
$$
\langle T_{ww'}\rangle_C^{\rm ren}=
-W^2\left(\frac{\epsilon\pi}{12\ln^2A}+\frac{1}{48\pi}\right)\delta_{ww'}.
$$
\noindent The metric in coordinates $\alpha,\beta$ is "good" (smooth,
nondegenerate) near the Cauchy horizons $\alpha=0$ or $\beta=0$. So, the
proper basis of an observer approaching to one of them with a finite
acceleration
is related to the basis $D\equiv\{\partial_\alpha,\partial_\beta\}$ by a
finite Lorentz transformation. Thus the quantities we are to examine are,
in fact, the components of $\langle {\bf T}\rangle_C^{\rm ren}$ in the
basis $D$, which are
$$
\begin{array}{c}
\langle T_{\alpha\alpha}\rangle_C^{\rm ren}=T\alpha^{-2},\quad
\langle T_{\beta\beta}\rangle_C^{\rm ren}=T\beta^{-2},\quad
\langle T_{\alpha\beta}\rangle_C^{\rm ren}=
\langle T_{\beta\alpha}\rangle_C^{\rm ren}=0,
\\ \strut\\
\displaystyle T\equiv -\left(\frac{\pi\epsilon}{12\ln^2A}+
\frac{1}{48\pi}\right).
\end{array}
$$\par
Now let us use the above simple method to find $\langle {\bf
T}\rangle^{\rm ren}$ for two time machines more (see also \cite{Yur}).
Consider first the cylinder $S$ obtained from the strip
\begin{equation}
\begin{array}{c}
ds^2=W^{-2}\xi^{-2}(-d\eta^2+d\xi^2)=\xi^{-2}d\alpha d\beta, \\
\strut \\ \mbox{where }\alpha\equiv(\xi-\eta)/W,\
\beta\equiv(\xi+\eta)/W;\quad
\eta\in(-\infty,\infty),\;\xi\in[1,A]. \label{StMod
\end{array}
\end{equation}
by gluing points $\eta=\eta_0,\ \xi=1$ with the points $\eta=A\eta_0,\ \xi=A$.
This spacetime was considered in detail in \cite{Fro} where it was called
the "standard model". A simple investigation shows that the Cauchy
horizons $\alpha=0$ and $\beta=0$ divide $S$ into three regions. Causality
holds in the "inner" region $\widetilde S:\ \alpha, \beta>0$ and violates
in $I^\pm(\widetilde S)$. Introducing new coordinates $u,\;v$:
$$
u\equiv W^{-1}\ln \alpha,\quad v\equiv W^{-1}\ln \beta
$$
we find that $\widetilde S$ like the Misner space%
\footnote{In spite of their apparent similarity these spaces are
significantly distinct. For example, the Misner spacetime is geodesically
incomplete \cite{HawEl}, and the standard model is not \cite{Me}. This
may be of importance if one would like to separate $X$ and $X'$ "widely
enough" (see item 1b in the previous section).}
is isometric to $M$.
This time
$$
C=\cosh^{-2}W\tau,\quad H=W^{-1}\ln A,
$$
which yields
\begin{equation}
\begin{array}{c}
\langle T_{\alpha\alpha}\rangle_C^{\rm ren}=T\alpha^{-2},\quad
\langle T_{\beta\beta}\rangle_C^{\rm ren}=T\beta^{-2},\\
\strut\\
\langle T_{\alpha\beta}\rangle_C^{\rm ren}=
\langle T_{\beta\alpha}\rangle_C^{\rm ren}=(1/12\pi)(\alpha+\beta)^{-2}.
\label{TFro}
\end{array}
\end{equation}\par
Consider lastly the spacetime obtained by changing $\xi^{-2}\to\eta^{-2}$
in (\ref{StMod}). This spacetime is similar to the standard model, but has
a somewhat more curios causal structure --- there are two causally
nonconnected regions separated by the time machine. $\langle {\bf
T}\rangle_C^{\rm ren}$ differs from that in (\ref{TFro}) by the
off-diagonal (bounded) terms
$$
\langle T_{\alpha\beta}\rangle_C^{\rm ren}=
\langle T_{\beta\alpha}\rangle_C^{\rm ren}=-(1/12\pi)(\alpha-\beta)^{-2}.
$$\par
So, we see that in all three cases the vacuum energy density (associated
with {\em some} vacuum states)
does grow infinitely as one approaches to the Cauchy horizon. A few
comments are necessary, however:
\begin{enumerate}
\item
The divergence in discussion is not at all something peculiar to the time
machine: the passage to the limit $A\to\infty$ shows that
precisely
the same divergence (with $T=-1/(48\pi)$) takes place in $\widetilde M$ though
(in the case of Misner space) $\widetilde M$ is merely
a part of the Minkowski plane. This suggests that for the time machine
too, the divergence of the stress-energy tensor is a consequence not of
its causal or topological structure but rather of the unfortunate choice
of the quantum state.
\item The twisted field at $A=e^{\sqrt{2}\pi}$ has the
bounded $\langle {\bf T}\rangle_C^{\rm ren}$ (cf. \cite{Sush}).
\item Let us consider nonvacuum states now (see Subsection~$2.3$). The
first example is a two-particle state $|1_n1_{-n}\rangle$ with the particles
corresponding to the $n$-th and $-n$-th modes of (\ref{ConVac}).
$\langle1_{-n}1_n| {\bf T}|1_n1_{-n}\rangle^{\rm ren}$ is readily found using
\cite[eq.~(2.44)]{Bir}:
$$
\langle1_{-n}1_n| T_{\gamma\gamma}|1_n1_{-n}\rangle^{\rm
ren}=T'\gamma^{-2},\quad
\langle1_{-n}1_n| T_{\alpha\beta}|1_n1_{-n}\rangle^{\rm ren}
=\langle T_{\alpha\beta}\rangle^{\rm ren}_C
$$
with $T'\equiv T+2\pi nH^{-2},$ and $ \gamma\equiv\alpha,\beta$. Thus we
see that there {\em are\/} states
with the bounded energy density of the untwisted field.\par
Another yet example is the equilibrium state at a nonzero temperature
$|t\rangle$. Expression (4.27) of \cite{Bir} gives
$$
\langle t| T_{ww}| t\rangle^{\rm ren}
=\langle T_{ww}\rangle^{\rm ren}_C+\frac{\pi}{2H^2}\sum_{m=1}^\infty
\sinh^{-2}\frac{\pi m}{k_BtH}.
$$
So, for any $H$ there exists such temperature $t$ that
$\langle t| T_{\gamma\gamma}| t\rangle^{\rm ren}$ does not diverge at the
horizon.
\end{enumerate}
\subsection{Another vacuum}
The conformal vanuum is not suited for verifying or exemplifying most of
statements made in Section 2., since the Hadamard function does not exist
in this state. So consider now the new vacuum $|0\rangle_f$ on the plane
$(\tau,\chi)$
defined by the modes
\begin{equation}
u'_p\equiv\cases{
\displaystyle\frac{1}{ 2\sqrt{\pi\omega} } e^{ip\chi-i\omega\tau},&
$\omega\geq\delta$\cr
\strut&\cr
\displaystyle\frac{1}{ 2\sqrt{\pi} } e^{ip\chi}
(f^{-1}\cos\omega\tau-i\omega^{-1}f\sin\omega\tau),&
$\omega<\delta$.\cr}
\label{AnVac}
\end{equation}
where $\omega\equiv|p|$, $\delta$ is an arbitrary positive constant:
$\delta<1$ and $f$ is an arbitrary smooth positive function:
$f(\omega\geq \delta)=\sqrt\omega$. The modes (\ref{AnVac}) are obtained
>from that
defining the conformal vacuum on the plane by a Bogolubov transformation of
the low-frequency modes so as to avoid the infrared divergence without
affecting the ultraviolet behavior of $\langle {\bf T}\rangle$.
The asymptotic form of ${\widetilde G}^{(1)}_f$ does not depend on $f$:
\begin{equation}
\forall f\qquad \widetilde G^{(1)}_f=-1/(2\pi)\ln|\Delta u\Delta
v|+\mbox{smooth, bounded function} .
\label{GAs}
\end{equation}
If we retain only the first term, we obtain (in the flat case)
$$
\lim_{X^{\prime}\to X} D_{\alpha\alpha}{\widetilde G}^{(1)}(X,\gamma^nX')=
\ln^2(A)\, \alpha^{-2}A^{-n}n^{-2}.
$$
So, the last series in (\ref{Tren}) diverges not only at the horizon, but
everywhere on $M$ (cf. Subsection $2.2$).\par
$\langle{\bf T}\rangle_f$ can be found from (\ref{GF}) (see
\cite[Sect.~$6.4$]{Bir}). For any $C$ we have:
$$
\langle T_{ww}\rangle_f^{\rm ren}=
\frac{1}{8\pi}\left[
-\frac{\;W^2}{6}+ \int^\delta_0
(f^{-2}\omega^2 +f^2-\omega)
\,d\omega\right],\quad
\langle T_{uv}\rangle_f^{\rm ren}=
\langle T_{uv}\rangle_C^{\rm ren}.
$$
Having taken an appropriate $f(\omega)$ one can make $\langle
T_{\alpha\alpha}\rangle^{\rm ren}_f$ infinite or zero at the horizon, as
we have stated in Subsection $2.2$.\par
To illustrate some more statements from Section~2.\ let
us first find $G^\Sigma$. To this end note that it has the form
\begin{equation}
G^\Sigma=\sum_n\int_{-\infty}^{\infty}h(p)e^{inHp}\,dp + c.\ c.
\label{GSum
\end{equation}
with
$$
h\equiv\cases{
\displaystyle\frac{1}{ 4\pi\omega } e^{ip\Delta\chi-i\omega\Delta\tau},&
$\omega\geq1/2$\cr
\strut&\cr
\displaystyle\frac{1}{ 4\pi } e^{ip\Delta\chi}
(f^{-2}\cos\omega\tau\cos\omega\tau'+
\omega^{-2}f^2\sin\omega\tau\sin\omega\tau')
,&
$\omega<1/2$.\cr}
$$
The function $h(p)$ can be written as a sum: $h=(h-h_0)+h_0$, where
$$
h_0\equiv\frac{1}{4\pi\sqrt{1+p^2}}\;e^{ip\Delta\chi}
\; e^{-i\sqrt{1+p^2}\Delta\tau}.
$$
The first summand is a smooth function falling off at infinity like
$p^{-2}$ and
the second summand ($h_0$) is a holomorphic (but for $p=\pm i$) function
admitting the following estimate:
$$
|h_0|\leq C|x|^{-1/2}\;e^{|(\Delta\chi-\Delta\tau)y|}.
$$
Hence \cite{Fed} we can apply the Poisson formula to (\ref{GSum})
and obtain:
$$
G^\Sigma=2\pi H^{-1}\sum_n h(2\pi H^{-1}n) +c.\ c.
$$
We see thus that $G^\Sigma$ is indeed the Hadamard function and it
corresponds to the vacuum
$|0\rangle_F$ with $F=f(0)$.\par
{\em Remark 1.\/} This does not mean, however, that $G^\Sigma$ will be a
Hadamard function of some reasonable state for {\em any\/} $\widetilde
G^{(1)}$.
One can easily construct, for example, such a vacuum that $\widetilde
G^{(1)}(\chi,\chi')$ will not be invariant under translations
$\ \chi,\chi'\mapsto\chi+H,\chi'+H$ and
$G^\Sigma(\chi,\chi')$, as a consequence, will not even be symmetric.\par
{\em Remark 2.\/} For all $G^{(1)}_f$ with the same $f(0)$ the Hadamard
functions
$G^\Sigma$ are the same. This proves our statement from Subsection~$2.1$.
To find $\langle {\bf T}\rangle_F$ note that it differs from
$\langle {\bf T}\rangle_C$ only by the term arising from the second
summand in (\ref{GF})
(cf.~\cite[ eqs. (4.20), (6.136)]{Bir}):
$$
\Delta\langle T_{ww'}\rangle^{\rm ren}=\frac{F^2}{2H}
(\tau,_w\tau,_{w'}-
1/2\,\eta_{ww'}\eta^{\lambda\delta}\tau,_\lambda\tau,_\delta).
$$
That is
\begin{eqnarray*}
\langle T_{ww}\rangle_F^{\rm ren}&=&
\displaystyle{\frac{F^2}{8H\vphantom{\ln^2A}}-
W^2\left(\frac{\epsilon\pi}{12\ln^2A}+
\frac{1}{48\pi}\right)},\\
&&\strut\\
\langle T_{uv}\rangle_F^{\rm ren}&= &
\langle T_{vu}\rangle_F^{\rm ren}=
\langle T_{uv}\rangle_C^{\rm ren} .
\end{eqnarray*}
So, for all three time machines considered here there exists a vacuum,
such that the expectation value of the stress-energy tensor is bounded in
the causal region.
\section{Conclusion}
Thus, we have seen that one cannot obtain any information about the energy
density near the Cauchy horizon employing the FKT approach. In the absence
of any other general approach this means that all we have is a few simple
examples. In some of them the energy
density diverges there and in some do not. So, the time machine perhaps is
stable and perhaps is not. This seems to be the most strong assertion we
can make.
|
\subsection{The case of zero mixing}
In this case $A_t=0$, and the expression
for $G$ in (\ref{G}) simplifies a lot.
Using (\ref{redefinicion}) and
(\ref{betasns}) one can, to one loop, expand $\lambda$ as
\begin{eqnarray}
\label{exp}
\lambda(\overline{m}_{t})&=&
\lambda(M)-\beta_{\lambda}(M)\log\frac{M^2}{\overline{m}_{t}^2} \nonumber \\
&=& \frac{1}{4}(g^2+g'^2)+\frac{3}{8\pi^2}h_t^2(h_t^2-\lambda)
\log\frac{M^2}{\overline{m}_{t}^2}
\end{eqnarray}
while ${\rm d}^2G/{\rm d}x^2$ is found to give
\begin{equation}
\label{Gseg}
\frac{{\rm d}^2G}{{\rm d}x^2} = \frac{3}{8\pi^2}
h_t^2(h_t^2-\lambda)\log\frac{M^2+\overline{m}_{t}^2}{M^2}
-\frac{3}{16\pi^2}h_t^4\frac{M_Z^2}{M^2+\overline{m}_{t}^2}
\end{equation}
where all couplings are considered at the scale $M$
(remember that the $G$-term was
frozen at that scale) and we have neglected ${\cal O}(g^4)$ terms.
Putting (\ref{exp}) and (\ref{Gseg}) together, the effective coupling
(\ref{lambdaeffec}) can be written as
\begin{equation}
\label{efflambda}
\lambda_{\rm eff}(\overline{m}_{t})=\frac{1}{4}(g^2+g'^2)
+\frac{3}{8\pi^2}h_t^2(h_t^2-\lambda_{\rm eff})
\log\frac{M^2+\overline{m}_{t}^2}{\overline{m}_{t}^2}
-\frac{3}{16\pi^2}h_t^4\frac{M_Z^2}{M^2+\overline{m}_{t}^2}
\end{equation}
where all couplings on the right-hand side are
considered at the scale $M$.
Comparison of (\ref{exp}) with (\ref{efflambda}) shows that the effect of
including the $G$-term in
(\ref{potencialsub}) is, apart from the last (mixed
Yukawa--gauge) term in (\ref{efflambda}) which is numerically unimportant,
to resum the series
$$ \sum_{n=1}^{\infty}\frac{(-1)^{n+1}}{n}
\left(\frac{\overline{m}_{t}^{2n}}{M^{2n}}\right)$$
from the expansion parameter $\log(M^2/\overline{m}_{t}^2)$ to the physical one
$\log((M^2+\overline{m}_{t}^2)/\overline{m}_{t}^2)$,
which goes to zero in the supersymmetric limit
$M\rightarrow 0$, and makes the radiative
corrections vanish in that limit.
In other words the resummation
from the $G$-function allows us to replace the
expansion (\ref{exp}) by the expansion
\begin{equation}
\label{expMs}
\lambda_{\rm eff}(\overline{m}_{t})=
\lambda_{\rm eff}(M_S)-\beta_{\lambda_{eff}}(M_S)t
\end{equation}
where the expansion parameter is
\begin{equation}
\label{t}
t=\log\frac{M_S^2}{\overline{m}_{t}^2}
\end{equation}
and
\begin{equation}
\label{MS}
M_S^2=M^2+\overline{m}_{t}^2 .
\end{equation}
\subsection{The case $A_t\neq 0$}
Using again (\ref{redefinicion}) and (\ref{betasns}) one can write
\begin{eqnarray}
\label{expmix}
\lambda(\overline{m}_{t}) & = & \lambda(M)-
\beta_{\lambda}(M)\log\frac{M^2}{\overline{m}_{t}^2}\nonumber \\
&=& \frac{1}{4}(g^2+g'^2)+\frac{3}{8\pi^2}h_t^2(h_t^2-\lambda)
\log\frac{M^2}{\overline{m}_{t}^2} \\
&+ &
\frac{3}{8\pi^2} h_t^2
\left[\left(h_t^2-\frac{1}{8}(g^2+g'^2)\right)\frac{A_t^2}{M^2}
-\frac{1}{12}h_t^2\frac{A_t^4}{M^4}\right] \nonumber
\end{eqnarray}
In the absence of the $G$-contribution in (\ref{lambdaeffec}),
Eq.~(\ref{expmix})
gives the usual one-loop leading-log
contribution to the Higgs mass. Introducing
now the $G$-function, one can write (\ref{lambdaeffec}) as:
\begin{equation}
\label{efflamix}
\lambda_{\rm eff}(\overline{m}_{t})
= \frac{1}{4}(g^2+g'^2)+\frac{3}{8\pi^2}h_t^2
(h_t^2-\lambda_{\rm eff})
\log\frac{M_S^2}{\overline{m}_{t}^2} + \Delta_{\rm th}\lambda_{\rm eff} ,
\end{equation}
where the `threshold' contribution to the coupling is given by
\begin{eqnarray}
\label{umbral}
\Delta_{\rm th}\lambda_{\rm eff} & = &
\frac{3}{64 v^4\pi^2}\left\{\left(2m_t^2-\frac{1}{2}M_Z^2+m_t A_t\right)^2
\log\left(1+\frac{m_tA_t-\frac{1}{4}M_Z^2}{M_S^2}\right)
\right. \nonumber \\
&+ & \left(2m_t^2-\frac{1}{2}M_Z^2-m_t A_t\right)^2
\log\left(1-\frac{m_tA_t+\frac{1}{4}M_Z^2}{M_S^2}\right)
\\
&- & m_t A_t \left\{
\left(M_S^2-\frac{1}{4}M_Z^2+m_t A_t \right)
\left[\log\left(1+
\frac{m_t A_t-\frac{1}{4}M_Z^2}{M_S^2}\right)-1\right]\right.
\nonumber \\
& - & \left.\left. \left(M_S^2-\frac{1}{4}M_Z^2-m_t A_t \right)
\left[\log\left(1-
\frac{m_t A_t+\frac{1}{4}M_Z^2}{M_S^2}\right)-1\right]\right\}
\right\} \nonumber
\end{eqnarray}
and it can be easily checked that
\begin{equation}
\label{expumbral}
\Delta_{\rm th}\lambda_{\rm eff}=
\frac{3}{8\pi^2} h_t^2\left[\left(h_t^2-\frac{1}{8}(g^2+g'^2)\right)
\frac{A_t^2}{M_S^2}
-\frac{1}{12}h_t^2\left(\frac{A_t^2}{M_S^2}\right)^2\right]
+\cdots
\end{equation}
where the ellipsis denotes the contribution from higher-dimensional
terms in $m_t^2/M_S^2$ and $A_t^2/M_S^2$.
We can see that the effect of taking into account the $G$-function in
(\ref{potencialsub}) is twofold:
\begin{itemize}
\item
It makes a resummation from $\log(M^2/\overline{m}_{t}^2)$
to the physical expansion parameter
$\log(M_S^2/\overline{m}_{t}^2)$ in the one-loop radiative correction,
as well as resummation from $A_t^2/M^2$ to $A_t^2/M_S^2$ in the
threshold term of (\ref{expmix}).
\item
It generalizes the threshold correction of Eq.~(\ref{expmix})
to include higher order effects
in powers of $A_t^2/M_S^2$ in Eq.~(\ref{umbral}).
\end{itemize}
Therefore the net effect of the $G$-function is to change the expansion
(\ref{expmix}) into
\begin{equation}
\label{expMsmix}
\lambda_{\rm eff}(\overline{m}_{t})=\lambda_{\rm eff}(M_S)-
\beta_{\lambda_{\rm eff}}(M_S)t
\end{equation}
where
\begin{equation}
\label{inicial}
\lambda_{\rm eff}(M_S)=
\frac{1}{4}(g^2+g'^2)+\Delta_{\rm th}\lambda_{\rm eff} .
\end{equation}
\subsection{Two-loop leading-log expansion}
As we have seen in Ref.~\cite{CEQW} one can obtain the two-loop leading-log
correction by expanding the parameter $\lambda$
to order $(\log(M^2/\overline{m}_{t}^2))^2$, as
\begin{equation}
\label{expmix2}
\lambda(\overline{m}_{t})=\lambda(M)-\beta_{\lambda}(M)\log\frac{M^2}{\overline{m}_{t}^2}
+\frac{1}{2}\beta'_{\lambda}\left(\log\frac{M^2}{\overline{m}_{t}^2}\right)^2,
\end{equation}
where, since we are now considering two-loop corrections, we have
evaluated the quartic coupling at the on-shell top quark mass scale
and the prime denotes derivative with respect to $\log(Q^2)$.
In the previous subsection we have seen that
the effect of including the $G$-function
in the effective potential is to resum $M^2$
to $M_S^2$ in the first two terms of
(\ref{expmix2}) and replace the threshold correction in the first one by
$\Delta_{\rm th}\lambda_{\rm eff}$ in (\ref{umbral}) (in Ref.~\cite{CEQW},
this resummation effect was assumed to be true to assure a proper
behaviour of the radiative corrections for $M = 0$).
To prove resummation in the last term of
(\ref{expmix2}) one would need to use the whole
two-loop effective potential in the
MSSM. In the absence of such a calculation
we shall assume that this happens, since we
have already proved that $t$ in (\ref{t}) is the
physical expansion parameter in the
one-loop calculation, and in addition the
numerical relevance of the resummation in the
two-loop corrections is expected to be tiny. We shall therefore
consider as a starting point
\begin{eqnarray}
\label{efflamix2}
\lambda_{\rm eff}(\overline{m}_{t})& = & \lambda_{\rm eff}(M_S)-\beta_{\lambda_{\rm eff}}(M_S)t
+\frac{1}{2}\beta'_{\lambda_{\rm eff}}(\overline{m}_{t})t^2 +\cdots \nonumber \\
& = & \lambda_{\rm eff}(M_S)-\beta_{\lambda_{\rm eff}}(\overline{m}_{t})t
-\frac{1}{2}\beta'_{\lambda_{\rm eff}}(\overline{m}_{t})t^2+\cdots
\end{eqnarray}
where $\lambda_{\rm eff}(M_S)$ is given in (\ref{inicial})
and $t$ and $M_S$ are obtained from Eqs.~(\ref{t}) and
(\ref{MS}).
Defining $\beta_{\lambda_{\rm eff}} = a_{\lambda_{\rm eff}} \lambda_{\rm eff} + b_{\lambda_{\rm eff}}$,
it follows that
\begin{equation}
\lambda_{\rm eff}(\overline{m}_{t}) = \lambda_{\rm eff}(M_S) \left( 1
- a_{\lambda_{\rm eff}}(\overline{m}_{t}) \; t \right) - b_{\lambda_{\rm eff}}(\overline{m}_{t}) \; t \;
\left( 1
- a_{\lambda_{\rm eff}}(\overline{m}_{t}) \; t \right) -\frac{1}{2}\beta'_{\lambda_{\rm eff}}(\overline{m}_{t})t^2.
\end{equation}
{}From the renormalization group equations of the quartic coupling,
it follows that the
term $1 - a_{\lambda_{\rm eff}}(\overline{m}_{t}) \;t = \xi^{-4}(\overline{m}_{t})$, where
$\xi$, the Higgs field anomalous dimension, is:
\begin{equation}
\label{anomdim}
\xi(\overline{m}_{t})=1+\frac{3}{32\pi^2}h_t^2(\overline{m}_{t})\;t.
\end{equation}
Recalling that $v^2(M_S) = v^2(\overline{m}_{t}) \; \xi^{-2}(\overline{m}_{t})$, from
Eq.~(\ref{masahiggs}) we get
\begin{equation}
\label{decmasa}
m_h^2(\overline{m}_{t}) = m_h^2(M_S)\;\xi^{-2}(\overline{m}_{t})+
\Delta_{\rm rad\; }m_h^2(\overline{m}_{t}).
\end{equation}
In the above,
\begin{equation}
\label{masafroz}
m_h^2(M_S) = 2 \lambda_{\rm eff}(M_S)\;v^2(M_S),
\end{equation}
$\lambda_{\rm eff}(M_S)$ is given in Eq.~(\ref{inicial})
with all couplings and masses
evaluated at the scale $M_S$,
and $\Delta_{\rm rad\; }m_h^2(\overline{m}_{t})$ is given by
\begin{eqnarray}
\label{radmas}
\Delta_{\rm rad\; }m_h^2(\overline{m}_{t})& = &
2v^2(\overline{m}_{t})\left[
- b_{\lambda_{\rm eff}}(\overline{m}_{t}) \; t \;
\left( 1
- a_{\lambda_{\rm eff}}(\overline{m}_{t}) \; t \right) -\frac{1}{2}\beta'_{\lambda_{\rm eff}}(\overline{m}_{t})t^2
\right] \nonumber \\
& = & \frac{3}{4\pi^2}\frac{\overline{m}_{t}^4}{v^2(\overline{m}_{t})}\;t
\left[1+\frac{1}{16\pi^2}\left(\frac{3}{2}h_t^2-32\pi\alpha_3\right)t
\right]
\end{eqnarray}
where all couplings in (\ref{radmas}) are evaluated at
the scale $Q^2=\overline{m}_{t}^2$,
$\alpha_3$ is the strong gauge coupling, and we have used
$\beta_{h_t}$ in the evaluation of (\ref{radmas})~\cite{CEQW}.
Observe that, if we replace $\Delta_{\rm th}\lambda_{\rm eff}$
by its expansion in powers of $A_t/M_S$, Eq.~(\ref{expumbral}),
we neglect the small terms depending on the weak gauge couplings,
and we re-express the values of the couplings at $M_S$ by their
expressions at $\overline{m}_{t}$ using the appropriate $\gamma$- and
$\beta$-functions, we obtain
\begin{eqnarray}
m_h^2& = & M_Z^2\left( 1-\frac{3}{8\pi^2}\frac{\overline{m}_{t}^2}
{v^2}\ t\right) \nonumber \\
\label{mhsm}
& + & \frac{3}{4\pi^2}\frac{\overline{m}_{t}^4}{v^2}\left[ \frac{1}{2}\tilde{X}_t + t
+\frac{1}{16\pi^2}\left(\frac{3}{2}\frac{\overline{m}_{t}^2}{v^2}-32\pi\alpha_3
\right)\left(\tilde{X}_t t+t^2\right) \right]
\end{eqnarray}
where
\begin{eqnarray}
\label{stopmix}
\tilde{X}_{t} & =& \frac{2 A_t^2}{M_{S}^2}
\left(1 - \frac{A_t^2}{12 M_{S}^2} \right).
\end{eqnarray}
Equation (\ref{mhsm}) is equivalent to
Eq.~(9) of Ref.~\cite{CEQW} in the large $\tan\beta$ regime.
Although we have used, as simplifying hypothesis,
the case where $\tan\beta\gg 1$ and $\mu = 0$, all
the results are also valid for
the case $m_A\sim M_S$ and any value of $\tan\beta$
and $\mu$. The only
change in the final results is that
\begin{eqnarray}
\label{relacion}
(g^2+g'^2) & \longrightarrow & (g^2+g'^2)\cos^2 2\beta \nonumber \\
A_t & \longrightarrow & A_t-\mu\cot\beta
\end{eqnarray}
in $\lambda(M)$ and $\lambda_{\rm eff}(M_S)$.
Finally, we want to conclude this section with a comment about
the physical interpretation of the decomposition (\ref{decmasa}).
The term $m_h^2(M_S)$ comes from the
scale independent part of the MSSM effective
potential frozen at the scale $M_S$, where we have already subtracted the
contribution
evolving with $\log(M_S^2/\overline{m}_{t}^2)$.
Since this term is scale independent, it is
evolved to the scale
$\overline{m}_{t}$ with the corresponding power of the anomalous dimension
of the Higgs field.
On the contrary, the term $\Delta_{\rm rad\; }m_h^2$, which arises
from renormalizable terms after the stops
are decoupled at the high scale, is computed at
the low scale $\overline{m}_{t}$. Had we considered the whole MSSM effective potential
at the scale $\overline{m}_{t}$~\cite{ERZ}, we would have been neglecting the stop
decoupling at the scale $M_S$.
This is shown in Fig.~1 (dotted lines) where we
plot $m_h$ as a function of $M_S$ for
a pole top-quark mass
$M_t=175$ GeV,
vanishing mixing, $A_t=\mu=0$, and large ($\tan\beta=15$) and small
[infrared (IR) fixed point solution:
$\sin\beta\sim (200\; {\rm GeV}/M_t)$] values of $\tan\beta$.
Had we evolved the Higgs mass obtained from
the whole MSSM effective potential
(including the logarithmic terms) at $M_S$, to the scale
$\overline{m}_{t}$ with the anomalous dimension factor $\xi^{-2}(\overline{m}_{t})$,
we would have made an error associated with the
non-exact scale invariance of the effective
potential in the one-loop approximation,
as was observed in~\cite{CEQR}. This is shown in the
dashed lines of Fig.~1
and was also noticed in Ref. \cite{CEQW}. In fact the latter procedure
would lead to an expression of $\Delta_{\rm rad\; }m_h^2$,
where the second term inside the
square brackets has an extra factor of 2
[see Eq.~(\ref{mhsm})]. The solid
lines in Fig.~1 show the corresponding
value for the Higgs mass, while considering
Eq.~(\ref{decmasa}) with the whole expression for
$\Delta_{\rm th} \lambda_{\lambda_{\rm eff}}$
from Eq.~(\ref{umbral}).
The dependence of (\ref{decmasa}) on the mixing is shown in Fig.~2
(solid lines) where we plot $m_h$ as a function of $A_t$, for $\mu=0$,
the same values of $\tan\beta$ as in Fig.~1 and
$M_S=1$~TeV. For comparison we also plot
(dashed lines) the
corresponding mass, using the approximate expression for the
threshold contribution to the Higgs quartic couplings,
Eq.~(\ref{mhsm}). We see
that this approximation, which was used in~\cite{CEQW}, is very
good up to the maximum of the
curve, which means that the absolute upper bound on
the mass of the lightest Higgs boson
previously obtained remains unchanged for any
value of the mixing. Of course for very large
values of the mixing there is a departure between the two curves.
We have deliberately omitted from Figs.~1 and 2 the small D-term
contributions in Eq.~(\ref{umbral}) to compare with
previous results which did not make use of them \cite{ERZ,CEQW}.
They will be included in the general analysis of section 3.
\setcounter{equation}{0{The general case}
Having understood the simplified case explained in section 2, we
can proceed with the general case. Let us first focus on the
behaviour of the renormalized Higgs quartic couplings for low
values of the CP-odd Higgs mass and values of the
mass parameter $m_Q$ different from $m_U$.
In this case, the effective potential at scales above
${\rm Max}(m_Q,m_U,m_D)$ is
obtained through the general expression, Eq.~(\ref{potgeneral}).
Expanding
the effective potential in powers of the Higgs fields, it is
easy to identify the form of the effective quartic couplings
at the scale $m_Q$, $\lambda_i(m_Q)$, in a way completely analogous
to what we have done for the case $m_Q = m_U$ in section 2.
Assuming for definiteness that $m_Q > m_U > m_D$,
\begin{equation}
\lambda_i(m_Q) = \lambda_i(m_U) +
\beta_{i}^{QU}(m_U) \; \widetilde{t}_{QU}
+ \left(\beta_{i}^{QU}\right)' \; \frac{\;\widetilde{t}_{QU}^{\; 2}}{2}
\end{equation}
[see e.g. the second equality in Eq.~(\ref{efflamix2})]
where $\widetilde{t}_{QU} = \log(m_Q^2/m_U^2)$, $\beta_i$
denotes the $\beta$-function of the Higgs quartic couplings,
$\beta_i = d \lambda_i/d\log(Q^2)$ and the superscript $QU$
denotes its expression at
energy scales between $m_Q$ and $m_U$. In general,
we can write,
\begin{equation}
\beta_i = a_i \lambda_i + b_i, \;\;\;\;\;\;\;\;\;
\beta_i' \simeq a_i b_i + b'_i ,
\end{equation}
where we have only kept the dominant, Yukawa-coupling-dependent
contributions to $\beta'_i$. We omit the scale dependence of
$\beta_i'$ since it is a higher-order effect.
Hence,
\begin{equation}
\lambda_i(m_U) = \lambda_i(m_Q) \left( 1 - a_i \;
\widetilde{t}_{QU} \right) - b^{QU}_i(m_U)
\; \widetilde{t}_{QU} \left( 1 -
a_i \; \widetilde{t}_{QU} \right) -
\left(\beta_{i}^{QU}\right)' \; \frac{\;\widetilde{t}_{QU}^{\; 2}}{2}.
\end{equation}
The coefficients $a_i$ are linear combinations of the
anomalous dimensions of the Higgs fields $H_i$, which are independent
of the squark fields.
The same procedure as used above may be used to connect the value of
the quartic couplings at $m_D$ with their values at $m_U$,
and finally the values at $m_D$ with their values at $\overline{m}_{t}$.
Keeping only the dominant terms, we obtain
\begin{eqnarray}
\lambda_i(\overline{m}_{t}) & = &\lambda_i(m_Q) \left( 1
- a_i(\overline{m}_{t}) \; \widetilde{t}_Q \right)
- \left(\beta_{i}^{QU}\right)' \frac{\;\widetilde{t}_{QU}}{2}
\left(\widetilde{t}_{Q} + \widetilde{t}_{U}\right)
\nonumber\\
&-& b^{QU}_i(\overline{m}_{t}) \; \widetilde{t}_{QU} \; \left[ 1 -
a_i(\overline{m}_{t}) \left(\widetilde{t}_{Q} + \; \widetilde{t}_{U}
\right) \right]
\nonumber\\
& - & b^{UD}_i(\overline{m}_{t}) \; \widetilde{t}_{UD} \left[ 1 -
a_i(\overline{m}_{t}) \; \left(\widetilde{t}_{U} +
\widetilde{t}_D \right) \right]
- \left(\beta_{i}^{UD}\right)' \; \frac{\;\widetilde{t}_{UD}}{2}
\left(\widetilde{t}_{U} + \widetilde{t}_D \right)
\nonumber\\
& - & b_i(\overline{m}_{t}) \; \widetilde{t}_{D} \left( 1 -
a_i (\overline{m}_{t})\; \widetilde{t}_{D} \right)
-\beta_{i}' \; \frac{\;\widetilde{t}_{D}^{\; 2}}{2},
\label{lambimt}
\end{eqnarray}
where $\beta_i$ without any superscript
denotes the $\beta$-function values once
the two stops and the two
sbottoms are decoupled, the superscript $XY$, with $X,Y = Q,U,D$,
denotes the functional
form of the $\beta$-functions at scales between $m_{X}$ and
$m_{Y}$, $\widetilde{t}_{XY} = \widetilde{t}_X - \widetilde{t}_Y$
and $\widetilde{t}_X = \log(m_X^2/\overline{m}_{t}^2)$.
Similar expressions are obtained for a
different hierarchy of the squark mass parameters, with the
only difference that, in the case $m_U > m_Q$ and/or
$m_D > m_Q$, stops and sbottoms should be decoupled at
different scales. For simplicity of presentation
we shall first discuss the
result for the Higgs mass matrix elements in the case under
study and we shall present below
the result in the most general case.
The contribution of
the quartic couplings to the Higgs mass matrix elements is
then given by
\begin{eqnarray}
M^2_{12} &=& 2 v^2 [\sin \beta \cos \beta (\lambda_3 + \lambda_4) +
\lambda_6 \cos^2 \beta + \lambda_7 \sin^2 \beta ] - m_A^2
\sin\beta \cos\beta
\nonumber\\
M^2_{11} &=& 2 v^2 [\lambda_1 \cos^2 \beta + 2
\lambda_6 \cos \beta \sin \beta
+ \lambda_5 \sin^2 \beta] + m_A^2 \sin^2\beta
\label{mijl}\\
M^2_{22} &=& 2 v^2 [\lambda_2 \sin^2 \beta +2 \lambda_7 \cos \beta
\sin \beta + \lambda_5 \cos^2 \beta] + m_A^2 \cos^2\beta,
\nonumber
\end{eqnarray}
where all terms should be evaluated at the same scale and
we have also included the dependence on the
CP-odd Higgs mass~\cite{HH}. In the above, $v_1 = v \cos\beta$
and $v_2 = v \sin\beta$ are the $H_1$ and $H_2$ vacuum expectation
values, respectively.
Equation (\ref{lambimt}) has a clear interpretation:
The factor $\lambda_i(m_Q)$ in the first term
contains the tree level terms
and all finite contributions to the quartic
couplings, arising from
the existence of squark mixing and the fact that
$m_D \neq m_Q \neq m_U$.
The factor involving $a_i$ in the first term contains exactly
the terms necessary to rescale the Higgs mass
matrix elements by the appropriate anomalous dimension factors
from the scale $m_Q$ to the scale $\overline{m}_{t}$, together with the ones
necessary to re-express the
vacuum expectation values of the Higgs fields
appearing in Eq.~(\ref{mijl}) at the scale $\overline{m}_{t}$,
in terms of their
values at the scale $m_Q$ (for the complete expression of the
$\beta$-functions of the Higgs quartic couplings,
see for example Ref.~\cite{HH}). For instance, singling out this
contribution
of the $\lambda_2$ coupling to the matrix element $M_{22}^2$, which
we shall denote by $K_{22}$, we obtain
\begin{eqnarray}
K_{22}(\overline{m}_{t}) &=&2 \; \lambda_2(m_Q)
\left[ 1 - a_2 \log\left( \frac{m_Q^2}{\overline{m}_{t}^2}
\right) \right] \; v_2^2(\overline{m}_{t})
\nonumber\\
&=& 2 \; \lambda_2(m_Q) \; \xi_2^{-4}(\overline{m}_{t}) \; v_2^2(\overline{m}_{t})
\\
& & \nonumber \\
&=& 2 \; \lambda_2(m_Q) \; v_2^2(m_Q) \; \xi_2^{-2}(\overline{m}_{t})\nonumber,
\end{eqnarray}
where $\xi_2$ is the anomalous dimension factor of the Higgs
fields $H_2$.
The contribution of the other couplings to the Higgs mass matrix
elements present similar properties.
Thus, both the tree level term and all finite terms leading
to the non-trivial matching of the quartic couplings at the
scale $m_Q$ may be treated in the same way as the terms proceeding
from the higher-dimensional operator contributions to the
Higgs mass matrix elements. That is, they may be frozen
at the scale $m_Q$ and rescaled with the appropriate anomalous
dimension factors to obtain their expressions at low energies.
This generalizes the result obtained in section 2 for the case
of degenerate squark masses, $m_Q = m_U$.
The following terms in Eq.~(\ref{lambimt}) are the ones which
would be obtained even in the presence of trivial matching
conditions for the quartic Higgs couplings and, as has been
clearly explained in the previous section, are associated
with the scale-dependent contributions to the effective potential.
The dominant leading-log
contribution to the quartic couplings proceeds from the terms
in the $\beta$-function proportional to the fourth power of
the top and bottom quark Yukawa couplings, which are given
by
\begin{equation}
\lambda_i(\overline{m}_{t}) - \lambda_i(m_Q) = - \frac{b_i^Y}{2}
\left( \log\left(\frac{m_Q^2}{\overline{m}_{t}^2}\right)
+ \log\left(\frac{m_X^2}{\overline{m}_{t}^2}\right) \right),
\label{llog}
\end{equation}
where $b_1^Y = -3 h_b^4/8 \pi^2$ and $b_2^Y = -3 h_t^4/8 \pi^2$,
$h_b$ and $h_t$ are the bottom and top quark Yukawa
couplings, and $X = D, U$ for $i = 1,2$,
respectively. Although $b_3$ and
$b_4$ also present a quartic dependence on the Yukawa couplings,
such dependence is absent from $b_3 + b_4$.
As may be easily proved using Eq.~(\ref{mijl}), and
following the same procedure
as in section 2, in the case of no mixing
the contribution of the
higher-order operator to the \lq 22' (\lq 11')
Higgs mass matrix elements
allows us to replace the factors $m_Q^2$ and $m_U^2$
($m_Q^2$ and $m_D^2$)
in the leading order expressions by $m_{Q}^2
+ \overline{m}_{t}^2$ and $m_U^2 + \overline{m}_{t}^2$
($m_Q^2 + m_b^2$ and $m_D^2 + m_b^2$).
The same occurs with
the D-term contributions proportional to the square of
products of the weak couplings and the Yukawa couplings.
It is hence convenient to define
\begin{eqnarray}
t_Q^U & = & \log\left(\frac{m_Q^2 + \overline{m}_{t}^2}{\overline{m}_{t}^2}\right) \;\;\;\;\;
{\rm and} \;\;\;\;
t_U = \log\left(\frac{m_U^2 + \overline{m}_{t}^2}{\overline{m}_{t}^2}\right),
\nonumber\\
t_Q^D & = & \log\left(\frac{m_Q^2 + m_b^2}{m_b^2}\right) \;\;\;\;\;
{\rm and} \;\;\;\;
t_D = \log\left(\frac{m_D^2 + m_b^2}{m_b^2}\right) \;\;\;\;\;
\label{tqtu}
\end{eqnarray}
while $t_{QU} = t_Q^U - t_U$ and $t_{QD} = t_Q^D - t_D$.
Observe that in the denominators of the expressions for
$t_D$ and $t_Q^D$ we have written $m_b$ instead of $\overline{m}_{t}$
since we know that, upon consideration of the whole
supersymmetric contributions to the physical masses,
radiative corrections should
vanish in the supersymmetric limit; we should thus use
expansion parameters $t_Q^D$ and $t_D$, which vanish
in the limit $m_Q,\; m_D\rightarrow 0$. This
change has negligible effects on the Higgs mass computation.
In the general case,
the way to proceed to obtain the Higgs mass matrix elements
at the scale $\overline{m}_{t}$ is the following: the
CP-even Higgs mass matrix elements may be decomposed
in three terms, namely:
\begin{eqnarray}
{\cal{M}}^2_{ij}(\overline{m}_{t}) & = &
\overline{M}^2_{ij}(\overline{m}_{t})
+ \left(\widetilde{{\cal{M}}}^2_{ij}
(M_{st}) \right)_{\widetilde{t}}
\left(\xi_{i}^{\widetilde{t}}(\overline{m}_{t})
\xi_j^{\widetilde{t}}(\overline{m}_{t})\right)^{-1}
\nonumber\\
& + &
\left(\widetilde{{\cal{M}}}^2_{ij}(M_{sb}) \right)_{\widetilde{b}}
\left(\xi_{i}^{\widetilde{b}}(\overline{m}_{t})
\xi_j^{\widetilde{b}}(\overline{m}_{t})\right)^{-1},
\label{m2ijmt}
\end{eqnarray}
where $\xi_{i}(\overline{m}_{t})$ denote
the anomalous dimension factors,
\begin{eqnarray}
\xi_1^{\widetilde{b}}(\overline{m}_{t}) =
1 + \frac{3 h_b^2}{32 \pi^2} \; t^{\widetilde{b}}_1 \;\;\;\; &&
\;\;\;\;\;\;\;
\xi_2^{\widetilde{b}} (\overline{m}_{t})= 1 + \frac{3 h_t^2}{32 \pi^2} \;
t^{\widetilde{b}}_1
\nonumber\\
\xi_1^{\widetilde{t}}(\overline{m}_{t}) =
1 + \frac{3 h_b^2}{32 \pi^2} \; t^{\widetilde{t}}_1 \;\;\;\; &&
\;\;\;\;\;\;\;
\xi_2^{\widetilde{t}} (\overline{m}_{t})= 1 + \frac{3 h_t^2}{32 \pi^2} \;
t^{\widetilde{t}}_1
\label{anomdij}
\end{eqnarray}
where for convenience we define,
\begin{equation}
t^{\widetilde{b}}_1 = {\rm Max}(t_Q^D,t_D) \;\;\;\;\;\;\;\;\;
t^{\widetilde{t}}_1 = {\rm Max}(t_Q^U,t_U),
\end{equation}
and also, for later use,
\begin{equation}
t^{\widetilde{b}}_2 = {\rm Min}(t_Q^D,t_D)
\;\;\;\;\;\;\;\;\;
t^{\widetilde{t}}_2 = {\rm Min}(t_Q^U,t_U).
\label{tbtt}
\end{equation}
In the above,
$\left(\widetilde{{\cal{M}}}^2_{ij}(M_{st})\right)_{\widetilde{t}}$
$\left(\left(\widetilde{{\cal{M}}}^2_{ij}(M_{sb})
\right)_{\widetilde{b}}\right)$
is the contribution to the
mass matrix elements coming from the terms
frozen at the scale $M_{st}$ ($M_{sb}$),
where the stops (sbottoms) are decoupled,
with $M_{st}^2 = {\rm Max}(m_Q^2 +\overline{m}_{t}^2,m_U^2+\overline{m}_{t}^2)$
$\left(M_{sb}^2 = {\rm Max}(m_Q^2 + m_b^2,m_D^2+ m_b^2)\right)$,
and $\overline{M}^2_{ij}$ are obtained from the mass matrix
elements $M^2_{ij}$, Eq.~(\ref{mijl}), by considering the
one- and two-loop leading logarithm
contributions in the
renormalizable Higgs quartic coupling expressions.
For example, in the case $m_Q > m_U > m_D$, these
contributions to the quartic couplings $\overline{\lambda}_i$
are given by
\begin{equation}
\overline{\lambda}_i = \lambda_i -
\left(\lambda_i(m_Q) - \lambda_i^{{\rm tree}}(m_Q) \right)
\left( 1 - a_i(\overline{m}_{t}) \; \widetilde{t}_{Q} \right) ,
\end{equation}
where the $\lambda_i$ are given in Eq.~(\ref{lambimt}),
$\lambda_1^{{\rm tree}} = \lambda_2^{{\rm tree}} =
-\left(\lambda_3^{{\rm tree}}+\lambda_4^{{\rm tree}}\right) =
\left(g^2 + g'^2\right)/4$,
and,
in order to simplify the presentation, we include all
D-term contributions in the definition of the quartic couplings
$\overline{\lambda}_i$.
Observe that, since $\overline{M}_{ij}^2(\overline{m}_{t})$
contains only the one- and two-loop
leading logarithm expressions independent of the mixing mass
terms, the quartic couplings $\lambda_5$, $\lambda_6$ and
$\lambda_7$ give no contribution to
$\overline{M}_{ij}^2(\overline{m}_{t})$. Replacing the quartic coupling
$\beta$-functions by their dominant Yukawa coupling dependence,
we obtain, in the general case
\begin{eqnarray}
\overline{\lambda}_1 &=&
\frac{g^2 + g'^2}{4} \left[ 1 - \frac{3}{8 \pi^2} h_b^2
\;t^{\widetilde{b}}_1 \right]
\nonumber\\
& + & \frac{3}{16 \pi^2}\; h_b^4\;
\left(t^{\widetilde{b}}_1 - t^{\widetilde{b}}_2 \right)\; \left[
1
+ \frac{1}{16 \pi^2}
\left( \frac{3}{2} \;h_b^2 + \frac{1}{2}\;h_t^2
- 8\; g_3^2 \right) \left(t^{\widetilde{b}}_1 +
t^{\widetilde{b}}_2 \right) \right]
\nonumber\\
& + &
\frac{3}{8 \pi^2}\; h_b^4\;
t^{\widetilde{b}}_2 \left[
1 + \frac{1}{16 \pi^2}
\left( \frac{3}{2} \;h_b^2 + \frac{1}{2}\;h_t^2
- 8\; g_3^2 \right)
t^{\widetilde{b}}_2 \right] + \Delta_1^D,
\label{lambda1}
\end{eqnarray}
\begin{eqnarray}
\overline{\lambda}_2 &=&
\frac{g^2 + g'^2}{4} \left[ 1 - \frac{3}{8 \pi^2} h_t^2
\; t^{\widetilde{t}}_1 \right]
\nonumber\\
& + & \frac{3}{16 \pi^2}\; h_t^4\;
\left(t^{\widetilde{t}}_1 - t^{\widetilde{t}}_2 \right) \;
\left[ 1
+ \frac{1}{16 \pi^2}
\left( \frac{3 \;h_t^2}{2} + \frac{h_b^2}{2}
- 8\; g_3^2 \right)\left(
t^{\widetilde{t}}_1 + t^{\widetilde{t}}_2 \right) \right]
\nonumber\\
& + &
\frac{3}{8 \pi^2}\; h_t^4\;t^{\widetilde{t}}_2 \; \left[
1 + \frac{1}{16 \pi^2}
\left( \frac{3 \;h_t^2}{2} + \frac{h_b^2}{2}
- 8\; g_3^2 \right) t^{\widetilde{t}}_2 \right] + \Delta_2^D,
\label{lambda2}
\end{eqnarray}
\begin{eqnarray}
\overline{\lambda}_3 + \overline{\lambda}_4 & = & \Delta_3^D + \Delta_4^D
\nonumber\\
& - & \frac{g^2 + g'^2}{4} \left[ 1 -
\frac{3}{16 \pi^2} h_b^2
\; t^{\widetilde{b}}_1 -
\frac{3}{16 \pi^2} h_t^2
\; t^{\widetilde{t}}_1 \right],
\label{lambda4}
\end{eqnarray}
where we have already performed the top and bottom mass resummations,
leading to the change $\widetilde{t}_X \rightarrow t_X$, with
$X = Q,U,D$. In the above, $h_t$ and $h_b$ denote the top and bottom
quark Yukawa couplings at the scale $\overline{m}_{t}$ and $g_3$ is
the strong gauge coupling
at the same scale. The $\Delta_i^D$ terms are
the additional
leading-log D-term contributions appearing through
the $\beta$-functions of the quartic couplings, which contain
additional terms proportional to the square of the product of
weak gauge couplings and Yukawa couplings.
Their two-loop leading-log contributions are
very small and can be ignored. Interestingly enough,
once these terms are considered together with the terms
written explicitly in Eqs.
(\ref{lambda1})--(\ref{lambda4}),
one recovers the expressions for the D-terms
first found in Ref.~\cite{B}.
Defining $\Delta_3^D =
\left( \Delta_3^D \right)_U + \left( \Delta_3^D \right)_D$,
for $m_Q \geq m_D$, we obtain
\begin{eqnarray}
\Delta_1^D & =& \frac{1}{16 \pi^2} g'^2 h_b^2 \; t_{QD}
\nonumber\\
\left(\Delta_3^D\right)_D
& =& - \frac{1}{32 \pi^2} g'^2 h_b^2 \; t_{QD}
\end{eqnarray}
while for $m_D \geq m_Q$ we get
\begin{eqnarray}
\Delta_1^D & =& -\frac{3}{32 \pi^2} \left(\frac{g'^2}{3} +
g^2\right)
h_b^2 \; t_{QD}
\nonumber\\
\left(\Delta_3^D\right)_D
& =& - \frac{3}{64 \pi^2} \left( g^2 - \frac{g'^2}{3} \right)
h_b^2 \; t_{QD}
\\
\Delta_4^D & = & \frac{3}{32 \pi^2} g^2 h_b^2 \; t_{QD} .
\nonumber
\end{eqnarray}
Analogously, for $m_Q \geq m_U$ we obtain
\begin{eqnarray}
\Delta_2^D & =& \frac{1}{8 \pi^2} g'^2 h_t^2 \; t_{QU}
\nonumber\\
\left(\Delta_3^D \right)_U
& =& - \frac{1}{16 \pi^2} g'^2 h_t^2 \; t_{QU}
\end{eqnarray}
while for $m_U \geq m_Q$ we get
\begin{eqnarray}
\Delta_2^D & =&- \frac{3}{32 \pi^2} \left(-\frac{g'^2}{3} +
g^2\right)
h_t^2 \; t_{QU}
\nonumber\\
\left(\Delta_3^D\right)_U
& =& -\frac{3}{64 \pi^2} \left( g^2 + \frac{g'^2}{3} \right)
h_t^2 \; t_{QU}
\\
\Delta_4^D & = & \frac{3}{32 \pi^2} g^2 h_t^2 \; t_{QU} .
\nonumber
\end{eqnarray}
Finally notice that $\tan\beta$ is fixed at the scale $m_A$, for
$m_A\leq \overline{m}_{t}$, while for $m_A\geq\overline{m}_{t}$, $\tan\beta$
is given by
\begin{equation}
\label{tanbeta}
\tan\beta(\overline{m}_{t})=\tan\beta(m_A)\left[
1+\frac{3}{32\pi^2}(h_t^2-h_b^2)\log
\frac{m_A^2}{\overline{m}_{t}^2} \right] .
\end{equation}
For the case in which the CP-odd Higgs mass $m_A$ is
lower than $M_S = {\rm Max} \left(M_{st},M_{sb}\right)$,
we should have decoupled (strictly speaking)
the heavy Higgs doublet at the scale $m_A$,
and defined an effective quartic coupling
for the light Higgs as $\lambda(m_A)=m_h(m_A)/2v^2$, the low
energy value of it being obtained by the running of the Standard
Model renormalization group equations from the scale $m_A$ to
$\overline{m}_t$ \cite{CEQW}.
For simplicity we have ignored, in our analytical
approximation, the effect of the heavy Higgs doublet decoupling at the
intermediate scale. We partially compensate this effect by relating the
value of $\tan\beta$ at the scale $\overline{m}_t$ with its corresponding
value at the scale $m_A$ through its renormalization-group running,
Eq.~(\ref{tanbeta}).
The Higgs mass matrix elements at the scales $M_{st},M_{sb}$
may be inferred
from the second derivative
of the one-loop effective potential, Eq.~(\ref{potgeneral}), at the
minimum values for the Higgs fields. Consequently, they
can be obtained from the expressions
given in Refs.~\cite{ERZ,B}, where all parameters should be
assumed to be given at the scale at which the matrix element
is evaluated, i.e. $M_{st}$ or $M_{sb}$.
As we have shown above,
the dominant D-term contributions to the Higgs masses are already
taken into account in the one-loop leading logarithmic
expressions, much as it happens in the case
of $m_Q = m_U$, where
the dominant D-term contribution comes from the first two terms in
Eq.~(\ref{efflamix}). The D-term contribution of
$\Delta_{\rm th} \lambda_{\rm eff}$ becomes relevant only for
very large mixing and gives a very small correction to the
top quark Yukawa coupling effect. For completeness, we shall
include all ${\cal{O}}(g^2h_q^2,g'^2h_q^2)$ D-term contributions,
where $q = t,b$.
In order to obtain the expressions
for $\left(\widetilde{{\cal{M}}}^2_{ij}\right)_{\widetilde{t}} \;,
\left(\widetilde{{\cal{M}}}^2_{ij}\right)_{\widetilde{b}}$
one must subtract from
the expressions of the matrix elements at the
scale $M_{st}$, $M_{sb}$, respectively,
the contributions coming from
the term dependent on the CP-odd Higgs mass and from the
leading order logarithmic expressions.
This procedure leaves in the matrix elements all the terms
that should be frozen at the scales $M_{st}$, $M_{sb}$, including
all terms depending on the squark mixing
mass parameters. Using the expressions given
in Refs.~\cite{ERZ,B}, and Eqs.~(\ref{lambimt})--(\ref{lambda4}),
we obtain,
\begin{eqnarray}
\label{cpevenst}
&&
\left(\widetilde{{\cal{M}}}_{ij}^2(M_{st})
\right)_{\widetilde{t}}
=
\frac{3}{8 \pi^2 v^2}
\left[
\begin{array}{cc}
\widetilde{\Delta}_{11}^{\widetilde{t}} +
\left(\widetilde{\Delta}'_{11}\right)^{\widetilde{t}}
& \widetilde{\Delta}_{12}^{\widetilde{t}} +
\left(\widetilde{\Delta}'_{12}\right)^{\widetilde{t}} \\
\widetilde{\Delta}_{12}^{\widetilde{t}} +
\left(\widetilde{\Delta}'_{12}\right)^{\widetilde{t}}
& \widetilde{\Delta}_{22}^{\widetilde{t}} +
\left(\widetilde{\Delta}'_{22}\right)^{\widetilde{t}}
\end{array}
\right],
\end{eqnarray}
\begin{eqnarray}
\label{cpevensb}
&&
\left(\widetilde{{\cal{M}}}_{ij}^2 (M_{sb})
\right)_{\widetilde{b}}
=
\frac{3}{8 \pi^2 v^2}
\left[
\begin{array}{cc}
\widetilde{\Delta}_{11}^{\widetilde{b}} +
\left(\widetilde{\Delta}'_{11}\right)^{\widetilde{b}}
& \widetilde{\Delta}_{12}^{\widetilde{b}} +
\left(\widetilde{\Delta}'_{12}\right)^{\widetilde{b}} \\
\widetilde{\Delta}_{12}^{\widetilde{b}} +
\left(\widetilde{\Delta}'_{12}\right)^{\widetilde{b}}
& \widetilde{\Delta}_{22}^{\widetilde{b}} +
\left(\widetilde{\Delta}'_{22}\right)^{\widetilde{b}}
\end{array}
\right],
\end{eqnarray}
where
\begin{eqnarray}
\label{delta11b}
\widetilde{\Delta}_{11}^{\widetilde{b}}& = &
\frac{m_b^4}{\cos^2\beta}
\left[\log \left(\frac{m_{\;\widetilde{b}_1}^2 m_{\;\widetilde{b}_2}^2}
{\left(m_Q^2 + m_b^2\right)\left(m_D^2 + m_b^2\right)}\right)
+\frac{2 A_b(A_b-\mu\tan\beta)}
{m_{\;\widetilde{b}_1}^2-m_{\;\widetilde{b}_2}^2}
\log\frac{m_{\;\widetilde{b}_1}^2}{m_{\;\widetilde{b}_2}^2}\right]
\nonumber \\
& + & \frac{m_b^4}{\cos^2\beta}
\left[\frac{A_b(A_b-\mu\tan\beta)}
{m_{\;\widetilde{b}_1}^2-m_{\;\widetilde{b}_2}^2}
\right]^2
g(m_{\;\widetilde{b}_1}^2,m_{\;\widetilde{b}_2}^2),
\end{eqnarray}
\begin{eqnarray}
\label{delta11t}
\widetilde{\Delta}_{11}^{\widetilde{t}}& = &
\frac{m_t^4}{\sin^2\beta}
\left[\frac{\mu(-A_t+\mu\cot\beta)}
{m_{\;\widetilde{t}_1}^2-m_{\;\widetilde{t}_2}^2}
\right]^2
g(m_{\;\widetilde{t}_1}^2,m_{\;\widetilde{t}_2}^2),
\end{eqnarray}
\begin{eqnarray}
\label{delta22t}
\widetilde{\Delta}_{22}^{\widetilde{t}} & = &
\frac{m_t^4}{\sin^2\beta}
\left[\log \left(\frac{m_{\;\widetilde{t}_1}^2 m_{\;\widetilde{t}_2}^2}
{\left(m_Q^2 + m_t^2\right)\left(m_U^2 + m_t^2\right)} \right)
+\frac{2 A_t(A_t-\mu\cot\beta)}
{m_{\;\widetilde{t}_1}^2-m_{\;\widetilde{t}_2}^2}
\log\frac{m_{\;\widetilde{t}_1}^2}
{m_{\;\widetilde{t}_2}^2}\right] \nonumber \\
& + & \frac{m_t^4}{\sin^2\beta}
\left[\frac{A_t(A_t-\mu\cot\beta)}{m_{\;\widetilde{t}_1}^2-
m_{\;\widetilde{t}_2}^2}
\right]^2
g(m_{\;\widetilde{t}_1}^2,m_{\;\widetilde{t}_2}^2),
\end{eqnarray}
\begin{eqnarray}
\label{delta22b}
\widetilde{\Delta}_{22}^{\widetilde{b}} & = &
\frac{m_b^4}{\cos^2\beta}
\left[\frac{\mu(-A_b+\mu\tan\beta)}{m_{\;\widetilde{b}_1}^2-
m_{\;\widetilde{b}_2}^2}
\right]^2
g(m_{\;\widetilde{b}_1}^2,m_{\;\widetilde{b}_2}^2),
\end{eqnarray}
\begin{eqnarray}
\label{delta12t}
\widetilde{\Delta}_{12}^{\widetilde{t}} & = &
\frac{m_t^4}{\sin^2\beta}
\frac{\mu(-A_t+\mu\cot\beta)}
{m_{\;\widetilde{t}_1}^2-m_{\;\widetilde{t}_2}^2}
\left[\log\frac{m_{\;\widetilde{t}_1}^2}{m_{\;\widetilde{t}_2}^2}
+\frac{A_t(A_t-\mu\cot\beta)}
{m_{\;\widetilde{t}_1}^2-m_{\;\widetilde{t}_2}^2}
g(m_{\;\widetilde{t}_1}^2,m_{\;\widetilde{t}_2}^2)\right],
\nonumber\\
\end{eqnarray}
\begin{eqnarray}
\label{delta12b}
\widetilde{\Delta}_{12}^{\widetilde{b}} & = &
\frac{m_b^4}{\cos^2\beta}
\frac{\mu(-A_b+\mu\tan\beta)}
{m_{\;\widetilde{b}_1}^2-m_{\;\widetilde{b}_2}^2}
\left[\log\frac{m_{\;\widetilde{b}_1}^2}{m_{\;\widetilde{b}_2}^2}
+\frac{A_b(A_b-\mu\tan\beta)}
{m_{\;\widetilde{b}_1}^2-m_{\;\widetilde{b}_2}^2}
g(m_{\;\widetilde{b}_1}^2,m_{\;\widetilde{b}_2}^2)\right],
\nonumber\\
\end{eqnarray}
and
\begin{eqnarray}
\left(\widetilde{\Delta}'_{11}\right)^{\widetilde{b}}
& = & M_Z^2 \left[ 2 m_b^2
f^{\;\widetilde{b}}_1
- m_b A_b f^{\;\widetilde{b}}_2 \right], \;\;\;\;\;\;\;\;\;\;
\left(\widetilde{\Delta}'_{11}\right)^{\widetilde{t}} = M_Z^2
m_t \mu \cot\beta f^{\;\widetilde{t}}_2\;
\nonumber\\
\left(\widetilde{\Delta}'_{22}\right)^{\widetilde{t}}
& = & M_Z^2 \left[ -2 m_t^2
f^{\;\widetilde{t}}_1
+ m_t A_t f^{\;\tilde{t}}_2 \right], \;\;\;\;\;\;\;\;\;
\left(\widetilde{\Delta}'_{22}\right)^{\widetilde{b}} = - M_Z^2
m_b \mu \tan\beta f^{\;\widetilde{b}}_2\;
\nonumber\\
\left(\widetilde{\Delta}'_{12}\right)^{\widetilde{t}}
& = & M_Z^2 \left[ m_t^2 \cot\beta
f^{\;\widetilde{t}}_1
- m_t \frac{ A_t \cot\beta + \mu}{2}
f^{\;\widetilde{t}}_2 \right],
\nonumber\\
\left(\widetilde{\Delta}'_{12}\right)^{\widetilde{b}} & = & M_Z^2 \left[
- m_b^2 \tan\beta
f^{\;\widetilde{b}}_1
+ m_b \frac{A_b \tan\beta + \mu}{2} f^{\;\widetilde{b}}_2 \right],
\end{eqnarray}
where
\begin{equation}
g(m_1^2,m_2^2)=2-\frac{m_1^2+m_2^2}{m_1^2-m_2^2}
\log\frac{m_1^2}{m_2^2},
\end{equation}
\begin{eqnarray}
f^{\;\widetilde{t}}_1 & = & \frac{m_Q^2 - m_U^2}
{m_{\;\widetilde{t}_1}^2 - m_{\;\widetilde{t}_2}^2}
\left(\frac{1}{2} - \frac{4}{3}
\sin^2\theta_W\right) \log \left(
\frac{m_{\;\widetilde{t}_1}}{m_{\;\widetilde{t}_2}}\right)
\nonumber\\
&+& \left(\frac{1}{2} - \frac{2}{3} \sin^2\theta_W \right)
\log\left(\frac{m_{\;\widetilde{t}_1} m_{\;\widetilde{t}_2}}{m_Q^2 + m_t^2}
\right)
\\
& + & \frac{2}{3} \sin^2\theta_W \log \left(
\frac{m_{\;\widetilde{t}_1}m_{\;\widetilde{t}_2}}
{m_U^2 + m_t^2} \right) \nonumber,
\end{eqnarray}
\begin{eqnarray}
f^{\;\widetilde{b}}_1 & = & \frac{m_Q^2 - m_D^2}
{m_{\;\widetilde{b}_1}^2 - m_{\;\widetilde{b}_2}^2}
\left(-\frac{1}{2} + \frac{2}{3}
\sin^2\theta_W\right) \log \left(
\frac{m_{\;\widetilde{b}_1}}{m_{\;\widetilde{b}_2}}\right)
\nonumber\\
&+& \left(-\frac{1}{2} + \frac{1}{3} \sin^2\theta_W \right)
\log\left(\frac{m_{\;\widetilde{b}_1} m_{\;\widetilde{b}_2}}{m_Q^2 + m_b^2}
\right)
\\
&-& \frac{1}{3} \sin^2\theta_W \log \left(
\frac{m_{\;\widetilde{b}_1}m_{\;\widetilde{b}_2}}
{m_D^2 + m_b^2} \right)\nonumber,
\end{eqnarray}
\begin{eqnarray}
f^{\;\widetilde{t}}_2 & = &
m_t\; \frac{A_t - \mu \cot\beta}{m_{\;\widetilde{t}_1}^2 -
m_{\;\widetilde{t}_2}^2} \left[-\frac{1}{2}
\log\left( \frac{m_{\;\widetilde{t}_1}^2}
{m_{\;\widetilde{t}_2}^2}\right)\right. \nonumber \\
& + &\left.
\left(\frac{4}{3} \sin^2\theta_W
- \frac{1}{2} \right)
\frac{m_Q^2 - m_U^2}
{m_{\;\widetilde{t}_1}^2 - m_{\;\widetilde{t}_2}^2}
g(m_{\;\widetilde{t}_1}^2, m_{\;\widetilde{t}_2}^2) \right],
\end{eqnarray}
\begin{eqnarray}
f^{\;\widetilde{b}}_2 & = &
m_b\; \frac{A_b - \mu \tan\beta}{m_{\;\widetilde{b}_1}^2 -
m_{\;\widetilde{b}_2}^2}
\left[ \frac{1}{2} \log\left( \frac{m_{\;\widetilde{b}_1}^2}
{m_{\;\widetilde{b}_2}^2}\right) \right. \nonumber \\
& + & \left.
\left( \frac{1}{2} -\frac{2}{3} \sin^2\theta_W
\right)
\frac{m_Q^2 - m_D^2}
{m_{\;\widetilde{b}_1}^2 - m_{\;\widetilde{b}_2}^2}
g(m_{\;\widetilde{b}_1}^2, m_{\;\widetilde{b}_2}^2) \right].
\end{eqnarray}
In the above, all terms should be computed at the scale
$M_{st}$ or $M_{sb}$ depending on whether they are associated to
stop or sbottom contributions
and all ${\cal{O}}(g^4,g^2g'^2,g'^4)$ terms are ignored.
It is easy to show that, apart from very small terms of
order $M_Z^2/(m_{\;\widetilde{t}_1}^2 + m_{\;\widetilde{t}_2}^2)$,
Eqs.~(\ref{cpevenst}), (\ref{cpevensb})
vanish in the case of zero squark mixing.
It is also straightforward to show that in the limit of large
$\tan\beta$, $m_Q = m_U$ and $m_b = \mu =0$,
$\widetilde{\Delta}_{12}^{\widetilde{t}} = 0$
and $3 \left(\widetilde{\Delta}_{22}^{\widetilde{t}} +
\left(\widetilde{\Delta}_{22}'\right)^{\widetilde{t}}
\right)/16\pi^2v^4$
reproduces the expression of $\Delta_{\rm th} \lambda_{\rm eff}$
given in Eq.~(\ref{umbral}).
All parameter values in the expressions
for the matrix elements $\widetilde{{\cal{M}}}_{ij}^2$
can be expressed in terms of
their values at the scale $\overline{m}_{t}$ by making use of the corresponding
$\beta$-and $\gamma$-functions.
Having computed the renormalization group improved Higgs mass
matrix elements at the scale $\overline{m}_{t}$, the neutral CP-even Higgs
mass eigenvalues can be easily derived. They read
\begin{eqnarray}
m^2_{h(H)} &=& \frac{{\rm Tr} {\cal M}^2 \mp \sqrt{({\rm Tr}
{\cal M}^2)^2 - 4 \det {\cal M}^2}}{2}
\label{mhH}
\end{eqnarray}
where
\begin{equation}
{\rm Tr}{\cal M}^2 = {\cal{M}}_{11}^2 + {\cal{M}}_{22}^2 \;\; ; \;\;\;\;\;
\det {\cal M}^2 = {\cal{M}}_{11}^2 {\cal{M}}_{22}^2 -
\left( {\cal{M}}_{12}^2 \right)^2 ,
\label{detm2}
\end{equation}
and the ${\cal{M}}_{ij}$ are the renormalized Higgs mass matrix elements
at the scale $\overline{m}_{t}$, Eq.~(\ref{m2ijmt}).
{}From the matrix elements,
the mixing angle $\alpha$ is also determined by~\cite{HH}:
\begin{eqnarray}
\sin 2\alpha = \frac{2{\cal{M}}_{12}^2}
{\sqrt{\left({\rm Tr} {\cal M}^2\right)^2-4\det {\cal M}^2}}
\end{eqnarray}
\begin{eqnarray}
\cos 2\alpha = \frac{{\cal{M}}_{11}^2-{\cal{M}}_{22}^2}
{\sqrt{\left({\rm Tr} {\cal M}^2\right)^2-4\det {\cal M}^2}}
\end{eqnarray}
Concerning the running of the squark mass parameters, the
dominant contribution comes from gluino-induced effects, which
are absent in the case of heavy gluino particles, as we are
considering within this work. The remaining contributions
are small and have a somewhat complicated dependence on the squark and
Higgs spectrum. We shall ignore them within
our approximation. We have further defined the light stop and sbottom
masses as the values obtained using Eqs. (\ref{masast}), (\ref{masasb}),
while taking the running mass parameters at the scale $\overline{m}_{t}$
and adding the QCD-dependent vacuum polarization effects. A more
precise definition of the squark masses may be obtained
by computing the squark effective potential and adding the
full vacuum polarization contributions to the squark masses,
much as we have done in the case of the Higgs bosons. We shall
concentrate on this subject elsewhere.
Figures~3 and 4 show the dependence of the Higgs mass $m_h$
for varying values of $m_Q = A_t = m_A$ and
for a fixed
value of the right-handed mass parameters $m_U = m_D =1$ TeV and
$m_U = m_D = 100$ GeV, respectively. The supersymmetric
mass parameter $\mu$ has been set to zero. The solid lines
represent the value of the Higgs mass by performing the
renormalization group improvement of the effective potential
method, as explained in this work. The dotted lines represent
the values obtained from the
one-loop effective potential, while ignoring the stop decoupling
and taking all values to be given at the scale $\overline{m}_{t}$
\cite{ERZ}. The dashed
lines are the values obtained for the Higgs mass, while considering
that the effective potential is scale-invariant, that is by
taking the second derivatives at the scale $M_S$ and rescaling
them to the scale $\overline{m}_{t}$ through the appropriate anomalous dimension
factors (see e.g. Ref.~\cite{CEQW}).
For low values of $m_U$ and $m_D$, as those shown in
Fig. 4, the last method becomes accurate for all values of
$m_Q < 600$ GeV, while in the second method the departure from the
proper renormalization group improved values is faster. Observe that
the behaviour shown in Fig. 4 is very similar to the one shown
in Fig. 1. Since the main purpose of Figs.~3 and 4 is to compare
the results in this paper with
other different approaches,
based on the one-loop MSSM effective potential, we have
plotted in them the running Higgs mass at the scale $\overline{m}_t$,
and turned the small D-terms threshold corrections off,
i.e. we have put $\widetilde{\Delta}'_{ij}=0$.
In the following we shall turn these D-terms on and consider
pole Higgs masses.
The Higgs masses $m_{h,H,A}$ defined in Eq.~(\ref{mhH}) are all running
masses obtained from the effective potential\footnote{In fact, as was
noticed in Ref.~\cite{ERZ}, the mass of the CP-odd Higgs, $m_A$,
turns out to be scale independent at one loop.}, and evaluated at
the top-quark mass scale. To compute the physical (propagator
pole) masses $M_{h,H,A}$ one has to correct for the fact that the
effective potential is defined at zero external momentum. In fact, the
pole and running Higgs masses are related by (see e.g. Ref.~\cite{CEQR})
\begin{equation}
\label{polemasas}
M^2_{\varphi}=m^2_{\varphi}+{\rm Re}\Delta\Pi_{\varphi}(M^2_{\varphi})
\end{equation}
where ${\varphi}=h,H,A$ and
\begin{equation}
\label{delpolariz}
\Delta\Pi_{\varphi}(M^2_{\varphi})=\Pi(M^2_{\varphi})-\Pi(0) ,
\end{equation}
$\Pi_{\varphi}(q^2)$ being the renormalized self-energy
of the corresponding Higgs boson.
We have computed the Higgs self-energies, at the one-loop
level, from the top/stop and bottom/sbottom sectors. The
corresponding expressions can be found in Appendix A.
Figures~5--7 show the variation of the pole Higgs mass $M_h$
as a function of $m_Q$ for
fixed $m_A$ and several fixed values of $m_U$ and of the stop
mixing parameter $A_t$. Although in general, for
a fixed moderate value of $m_U$ and a fixed value of $A_t$, the
Higgs mass increases together with $m_Q$,
for large values of $A_t$ and
moderate values of $m_U$ or
for small values of
$m_U$ and moderate values of $A_t$,
situations can be observed for which the Higgs mass
decreases with larger values of $m_Q$.
An understanding
of this effect may be obtained by making use of the approximation
of Ref.~\cite{CEQW}: although
$\log(M_{\rm SUSY}/\overline{m}_{t})$, with
$M_{\rm SUSY}^2 \equiv
\left(m_{\;\widetilde{t}_1}^2 + m_{\;\widetilde{t}_2}^2 \right)/2$,
increases together with the
squark-mass parameters, leading to larger values of the
Higgs mass, the squark mixing contributions are
maximized for values of $A_t^{\rm max} \simeq 2.4 M_{\rm SUSY}$. Hence,
for fixed values of $A_t$ a delicate balance between these
two effects should be present in order to maximize the Higgs
mass. Observe also that for small values of $m_U$, such as
the ones analysed in Fig.~7, the configurations which
maximize the lightest CP-even Higgs mass $M_h$
correspond to situations
for which one of the stops may become light. Indeed, the
curves are cut since we have introduced the (crude) experimental
constraint $m_{\;\widetilde{t}_{2}}\stackrel{>}{{}_\sim} 45$ GeV on them.
An important effect arising in the case of small values
of one of the mass parameters, for instance $m_U^2 \simeq M_Z^2$,
and $m_Q^2 \gg m_U^2$ as shown in Fig.~7, is that for the same
value of $M_{\rm SUSY}$, the maximal value for the Higgs mass is always
lower than in the case of $m_Q = m_U$. This is due to the fact that
the requirement of having stop masses above the present
experimental bound implies $A_t \leq m_Q \simeq \sqrt{2} M_{\rm SUSY}$.
Hence, $A_t$ is always significantly lower than $A_t^{\rm max}$,
which, as explained above, is the value that maximizes the Higgs mass
for that particular value of the average stop mass scale $M_{\rm SUSY}$.
One could enquire about the stop and sbottom
vacuum polarization contributions in the pole Higgs boson mass definitions.
We have checked that these contributions do not give a significant
effect in the determination of the neutral Higgs boson masses, unless one
of the squarks becomes light (i.e. $\;\widetilde{t}$ and/or $\;\widetilde{b}$) and its couplings
to the Higgs fields
(i.e. $A_t$, $A_b$, $\mu$) are large. This behaviour is illustrated in
Fig.~8 where we plot the neutral Higgs boson running
($m_h$, $m_H$, $m_A$) masses (dotted lines) and pole ($M_h$, $M_H$,
$M_A$) masses (solid lines) as functions of $A_t$, for fixed values of the
other supersymmetric parameters. We also plot the mass of the lightest
stop $m_{\;\widetilde{t}_{2}}$ (dashed line).
We see that the departure between the running
masses and the pole masses occurs in all cases
for large values of $A_t$ and
small values of $m_{\;\widetilde{t}_{2}}$.
The dependence of the Higgs mass on $m_D$ becomes relevant only
for large values of $\tan\beta$.
In Fig.~9 we show the variation of $M_h$
as a function of $m_D$ for $\tan\beta=60$, fixed values of $m_Q$, $m_U$
and $m_A$, and two values of $\mu$: $\mu=1$ TeV (solid curves) and
$\mu=2$ TeV (dashed curves). The radiative corrections
to $M_h$ induced by the
bottom/sbottom propagation are always of negative sign and become
only relevant for very large values of the $\mu$-parameter.
The behaviour with $\mu$ can be understood
from the fact that the larger $\mu$, the larger mixing in the sbottom sector
and hence lighter sbottom masses are obtained.
In this example, we have chosen $m_b(\overline{m}_{t}) = 3$ GeV, which corresponds
to a bottom quark pole mass $M_b \simeq 5$ GeV.
All the curves are cut by the experimental
constraint $m_{\;\widetilde{b}_{2}}\stackrel{>}{{}_\sim} 45$ GeV.
Finally Fig.~9 also exhibits the dependence of the Higgs mass on the
parameter $A_t$, its dependence with respect to $A_b$ being tiny.
\setcounter{equation}{0{Conclusions}
We have presented a renormalization group improvement of the
effective potential computation of the neutral Higgs masses in the
MSSM. The method provides the first calculation of two-loop
leading order corrections to the Higgs masses valid for
any value of the soft supersymmetry breaking squark mass parameters,
$m_Q$, $m_U$, $m_D$, $A_t$ and $A_b$, the CP-odd mass $m_A$,
the supersymmetric Higgs mass $\mu$
and $\tan\beta$. This generalization is essential for the
computation of the Higgs masses and mixing angles in the presence
of light squarks. Our method uses
explicit decoupling of stops and sbottoms
at their corresponding mass scales, leaving threshold effects in the
effective potential (and coupling constants)
frozen at the decoupling scales and evolving, in the
squared mass matrix, with the anomalous dimensions of the Higgs
fields.
The threshold effects achieve a complete matching of the
effective potential for scales above and below the decoupling
scales, and include all higher order (non-renormalizable)
terms arising from the whole MSSM effective
potential. The effect of considering non-renormalizable threshold
effects in the effective potential is twofold: on the one hand
it triggers resummations in
the renormalization group expansion of the parameters,
leading to `physical' expansion parameters; on the other hand,
it enables to consider the general case of
arbitrary left--right squark mixing,
as well as general left- and right-handed soft supersymmetry breaking
squark masses.
We have corrected the running neutral Higgs boson masses with
one-loop self-energy diagrams, where top- and bottom-quarks,
stops and sbottoms propagate, to define the corresponding
pole masses. The numerical effect of polarizations is relevant only
under special circumstances: light squarks and large mixing.
We have analysed
the general pattern of Higgs
masses for general values of the supersymmetric parameters.
We have found regions in the parameter space where the
radiative corrections become large and negative. They are
characterized by large values of the mixing-mass parameters,
where the stability of the electroweak minimum can be endangered
by the presence of charge and color breaking minima.
Our results also allow the evaluation of the relevant
radiatively corrected Higgs couplings through the corresponding
value of the Higgs angle $\alpha$ \cite{CEQW}.
We have neglected, throughout the whole calculation, the possible
contribution coming from light charginos/neutralinos. Their effect
can be easily included in the threshold terms, as well as in the
running of the $\beta$- and $\gamma$-functions, where they appear
as ${\cal O}(g^4,g^2g'^2,g'^4)$
terms and are thus numerically unimportant.
In Ref.~\cite{CEQW} we have shown that, in the case of a heavy
supersymmetric spectrum, our analytical expressions reproduced
the Higgs mass spectrum with an
error of less than 2--3 GeV.
It can be easily checked that light charginos and neutralinos
can increase the Higgs
mass in $\stackrel{<}{{}_\sim}$ 2--3 GeV \cite{topc,Marc}, with respect to heavy ones,
which is indeed
within the errors of our different approximations.
Nevertheless for completeness, we include in appendix B
the leading-log ${\cal{O}}(g^4,g'^2g^2,g'^4)$
chargino and neutralino contributions to the CP-even Higgs
masses and the mixing angle. Throughout this work,
we have also implicitly assumed that the gluino masses are of
order $M_S$. If the gluinos were, instead, much lighter than the
characteristic squark masses, the running of the
third generation Yukawa couplings
would be different, inducing a small indirect effect
on the two-loop Higgs mass computation. The third generation
Yukawa coupling running is also modified by the presence of
light charginos/neutralinos in the spectrum. These two loop
contributions to the CP-even Higgs masses are also presented
in appendix B.
Our present analysis reproduces, with a high level
of accuracy, the values of
the Higgs masses and mixing angles, for the
previously studied case of
degenerate left- and right-handed squark mass parameters,
and for values of the squark left--right mixing
mass parameters lower than the ones giving the maximal
values of the lightest CP-even Higgs mass. This comparison
holds up to a tiny difference coming from the inclusion in this
work of the small D-term threshold contributions and
vacuum-polarization effects. This confirms
previous results on the upper bound on the lightest Higgs boson
mass in the MSSM.
\section*{Acknowledgements}
We would like to thank
A. Brignole, J.R. Espinosa, H. Haber, S. Peris
and F. Zwirner for interesting discussions.
\newpage
|
\section*{Acknowledgments}
I thank S. Kuhlmann, J. Huston, J. Linnemann, T. Ferbel, and
S. Snyder for information about the experiments. This work was
supported in part by U.S. National Science Foundation grant
number PHY-9507683.
|
\section{Introduction} \label{sec 1}
Let $K$ be any compact, connected set in the plane.
The complement of $K$ has one unbounded component and its
topological boundary is called the {\em frontier} of $K$,
denoted $\mbox {\rm \small frontier}(K)$. The example we are most interested in is when $K$
is the range of a planar Brownian motion run for a finite time
(see Figure~\ref{outer}).
In this case, Mandelbrot (1982) conjectured that the Hausdorff
dimension $\dim (\mbox {\rm \small frontier}(K))$ is $4/3$. Rigorously, the best proven upper
bound on the dimension is $3/2 - 1/(4\pi^2) \approx 1.475$ by
Burdzy and Lawler (1990). Burdzy (1989) proved that
$\mbox {\rm \small frontier}(K)$ has infinite length; our main result improves this
to a strict dimension inequality:
\begin{figure}
\vbox{ \centerline{ \hbox{
\psfig{figure=path4.ps,height=2.5in}
$\hphantom{xxx}$
\psfig{figure=outline4.ps,height=2.5in}
}}}
\caption{ \label{outer} A Brownian path and its frontier}
\end{figure}
\begin{th} \label{th main}
Let $B[0,t]$ denote the range of a planar Brownian motion,
run until time $t>0$. There is an $\epsilon > 0$ such that with probability 1,
The Hausdorff dimension $ \, \dim (\mbox{\rm \small frontier}(B[0,1]))$ is at least $1 + \epsilon$.
Moreover, with probability 1,
$$
\inf_{t>0} \, \inf_V \, \dim \Big(\mbox{\rm \small frontier}(B[0,t]) \cap V \Big) \, \geq \, 1 + \epsilon \, ,
$$
where the inner infimum is over all open sets $V$ that intersect $\, \mbox{\rm \small frontier}(B[0,t])$.
\end{th}
\noindent{\em Remarks:} The uniformity in $t$ implies that
$ \, \dim (\mbox{\rm \small frontier}(B[0,\tau])) \geq 1 + \epsilon$ almost surely for any
positive random variable $\tau$ (which may depend on the Brownian motion).
We also note that
our proof shows that the frontier can be replaced in the statement of
the theorem by the boundary of any connected component of
the complement $B[0,t]^c$.
(One can also infer this from the statement of the theorem
by using conformal invariance of Brownian motion).
As explained at the end of Section \ref{sec 6},
The result also extends to the frontier of the planar Brownian
bridge (which is a closed Jordan curve by Burdzy and Lawler (1990)).
Bishop and Jones (1994) proved that if a compact, connected set
is ``uniformly wiggly at all scales'', then it has dimension
strictly greater than 1. Here we adapt this to a stochastic setting
in which the set is likely to be wiggly at each scale, given
the behavior at previous scales. The difficulty is in handling
statistical dependence.
\noindent{\bf Definitions:} Let $G$ be a compact set in the plane
with complement $ G^c$, and let
$\eta>0$. Denote by \bcorG the set
$\{ z \in G : \mbox{\rm dist} (z, G^c) > \eta \cdot \mbox{\rm diam}(G) \}$.
Say that the compact set $K \, $ \mbox{\bf $\, \eta$-surrounds} $G$ if $K$
topologically separates \mbox{{\rm core}$(G,\eta)$ } from $ G^c$,
i.e., if \mbox{{\rm core}$(G,\eta)$ } is disjoint from the unbounded component of $(K \cap G)^c$.
\begin{figure}
\centerline{ \psfig{figure=gen4.ps,height=2.5in} }
\caption{ \label{gen4} The Gosper island }
\end{figure}
\begin{th} \label{th any K}
Let $G_0$ be the {\bf Gosper Island}, defined in the next section
and illustrated in
Figure~\ref{gen4}. There exists an absolute
constant $\eta_0 > 0$ with the following property.
Suppose that $ c_0 >0$, and $K$ is a random compact connected
subset of the plane such that for
all homothetic
images $G=z+r G_0$ of $G_0$ with $r \in (0, 1)$ and $z$ in the plane:
\begin{equation} \label{eq gosper-wiggly}
{\bf{P}} \Big [ K \, \mbox{ \rm $\eta_0$-surrounds } G
\, \mbox { \rm \underbar{or} } \; \, K \cap \crG0 = \emptyset
\, \, \Big| \, \, \sigma(K \setminus G^{\circ} ) \Big ] > c_0,
\end{equation}
where the conditioning is on the $\sigma$-field generated by
the random set $K$ outside the interior $G^{\circ}$ of $ G$.
Then there is an $\, \epsilon > 0 $, depending only on $c_0$, such that
$$
\dim (\mbox {\rm \small frontier}(K)) \geq 1 + \epsilon
$$
with probability 1. More generally, $\dim (\mbox {\rm \small frontier}(K) \cap V) \geq 1 + \epsilon$
for any open $V$ intersecting $\mbox {\rm \small frontier}(K)$.
\end{th}
\begin{figure}
\centerline{ \psfig{figure=Surr+Miss.ps,height=2.5in} }
\caption{ \label{Surr+Miss} A Brownian motion which surrounds the core
and one which misses it. }
\end{figure}
\noindent{\sc Remarks: } { \bf 1.} In fact, the proof in Section $ 3 \,$
shows that
with probability 1, for any connected component $\Omega$
of $K^c$ and any open $V$ intersecting $\partial \Omega$,
there is a John domain $\Omega_{\rm J} \subset \Omega$ with
closure $\overline{\Omega_{\rm J}}$ contained in
$V$, such that
$\dim (\partial \Omega \cap \partial \Omega_{\rm J}) \geq 1 + \epsilon$.
\newline {\bf 2.} The constant $\eta_0$
will be chosen in the next section to ensure that no ``macroscopic''
line segment can be wholly contained within a $2 \eta_0$-neighborhood
of the Gosper Island's boundary $\partial G_0$.
The appearance of the Gosper Island might seem strange at this point,
but is explained as follows. The hypothesis on $K$ that
guarantees ``wiggliness'' should be local to handle dependence (thus
it must hold inside each $G$ conditioned on $K \cap G^c$).
If $ K \, $ $\, \eta_0$-surrounds $G$,
\underbar{or} $\; K \cap \crG0 = \emptyset$,
then $\mbox {\rm \small frontier}(K)$ cannot intersect \crG0.
Having thus controlled $\mbox {\rm \small frontier}(K)$ inside $G$, away from the
boundary of $G$, we must worry about how $\mbox {\rm \small frontier}(K)$ behaves
near boundaries of cells $G$, as these run over a
partition of the plane. If a small neighborhood of the union
of the boundaries of cells $G$ of a fixed size
contains no straight line segments of length comparable to
$\, \mbox{\rm diam} \, (G) \,$, then
no significant flatness can be introduced near cell boundaries.
To apply the argument with the same constants on every scale,
we need a self-similar tiling where tile boundaries have
no straight portions; the Gosper Island yields such a tiling.
Proving Theorem \ref{th any K} is the main effort of the paper and
is organized as follows. Section~\ref{sec 2} summarizes
notation and useful facts about the Gosper Island.
We also discuss the notion of a Whitney decomposition with respect
to these tiles. Section~\ref{sec 3} constructs a random tree
of Whitney tiles for $K$ and reduces Theorem \ref{th any K} to a
lower bound on the expected growth rate of the tree, via some general
propositions on random trees.
In Section~\ref{sec 4} we state a variant of
Jones's Traveling Salesman Theorem adapted to
the current setting. In Section~\ref{sec 5} this theorem is used
to derive the required lower bound on the expected growth rate
of the ``Whitney tree'' mentioned above, which
then finishes the proof of Theorem~\ref{th any K}. In
Section~\ref{sec 6} we verify that the range of planar Brownian motion,
killed at an independent exponential time,
satisfies the hypothesis of Theorem~\ref{th any K};
this easily yields Theorem \ref{th main}. Finally,
section 7 gives a hypothesis on the random set $K$
that is weaker than (\ref{eq gosper-wiggly}), but still
implies the conclusion of Theorem \ref{th any K}.
\section{Gosper Islands and Whitney tiles} \label{sec 2}
The standard hexagonal tiling of the plane is not self-similar,
but can be modified to obtain a self-similar tiling.
Replacing each hexagon by the union of seven smaller hexagons
(of area $1/7$ that of the original -- see Figure \ref{Sub}) yields a new tiling
of the plane by $18$-sided polygons; denote by $d_1$ the Hausdorff
distance between each of these polygons and the hexagon it approximates.
Applying the above operation to each of the seven smaller hexagons
yields a $54$-sided polygon with Hausdorff distance $7^{-1/2} \cdot d_1$ from
the $18$-sided polygon, which also has translates that tile
the plane. Repeating this operation (properly scaled) ad infinitum,
we get a sequence of polygonal tilings of the plane, that
converge in the Hausdorff metric to a tiling of the plane by translates
of a compact connected
set $G_0$ called the ``Gosper Island'' (see Gardner (1976)
and Mandelbrot (1982)).
\begin{figure}
\centerline{ \psfig{figure=Sub.ps,height=1.5in} }
\caption{ \label{Sub} Substitution defining Gosper island }
\end{figure}
\begin{figure}
\vbox{
\centerline{ \hbox{
\psfig{figure=gen0.ps,height=1.0in}
$\hphantom{xxx}$
\psfig{figure=gen1.ps,height=1.0in}
$\hphantom{xxx}$
\psfig{figure=gen2.ps,height=1.0in}
$\hphantom{xxx}$
\psfig{figure=gen3.ps,height=1.0in}
}}}
\caption{ \label{Gosper} First four generation of the construction}
\end{figure}
\noindent{\bf Notation:} We normalize $G_0$ to be centered at the origin
and have diameter 1.
Denote by ${\cal{D}}_0$ the set of translates of $G_0$ that form a tiling
of the plane (depicted in Figure \ref{TileUnion}).
This tiling is {\bf self-similar}, i.e., there is a complex number $\lambda$
with $|\lambda| >1$ such that for each tile $G \in {\cal{D}}_0$,
the homothetic image $\lambda \cdot G$ is the union of tiles in ${\cal{D}}_0$.
(For the tiling by Gosper Islands, $|\lambda| = 7^{1/2}$.)
For each integer $n$, we denote by ${\cal{D}}_n$ the scaled tiling
$\{ \lambda^{-n} \cdot G \; : \: G \in D_0 \}$,
and let $\, {\cal{D}} = \cup_{n=0}^{\infty} {\cal{D}}_n$.
If $G \in {\cal{D}}_n$ we say that $G$ is a tile of index $n$ and write
$||G||=n$. Every tile $ G \in {\cal{D}}_n$ is contained in a unique
tile of ${\cal{D}}_{n-1}$, denoted $\, \mbox{\rm parent}(G)$.
Each tile $G$ is centrally symmetric
about a ``center point'' $z$;
for any $\theta >0$, denote by $\theta \odot G = z+ \theta \cdot (G-z)$
the expansion of $G$ by a factor $\theta$ around $z$.
\begin{figure}
\centerline{ \psfig{figure=Tiling.ps,height=2.5in} }
\caption{ \label{TileUnion} A self-similar tiling of the plane }
\end{figure}
We record several simple properties of the tiling by Gosper Islands,
which will be useful later.
\begin{enumerate}
\item \label{min-dist}
There is some minimal distance $d_0$ between
any two nonadjacent tiles of ${\cal{D}}_0$.
\item \label{eta}
There is an $\eta_0 > 0$ such that any line segment
of length $d_0$ must intersect $\mbox{\rm core}(G, 2\eta_0)$ for some $G \in {\cal{D}}_0$.
(The existence of $\eta_0$ follows by a compactness argument from the fact that
$\partial G_0$ contains no straight line segments.)
\item \label{inrad}
The Gosper Island $G_0$ contains an open disk centered at the origin
which in turn contains $\lambda^{-1} G_0$.
\item \label{odot}
The blow-up $\lambda^3 \odot D$ contains $\lambda \odot
\mbox{\rm parent} (D)$ for any $D \in {\cal{D}}$ (see Figure \ref{WhitTile}).
\item \label{neighbor}
If $||G|| = ||G'|| - 1$ for neighboring tiles
$G$ and $G'$, then $\lambda \odot G$ contains $\lambda \odot G'$.
(See Figure~\ref{WhitTile}.)
\item \label{union}
The blow-up $\lambda \odot G$ is contained in $\bigcup (\lambda \odot G'')$
where the union is over all neighbors $G''$ of $G$
of index $||G||$.
\item \label{Jordan}
The boundary of $G_0$ is a Jordan curve.
To see this note that when we replace each segment or
length $r$ by the
three segments of the next generation, they remain
within distance $r \sqrt{3}/14$ of the segment. Thus
the limiting arc is within
$$ r \frac {\sqrt {3}}{14} \sum_{n=0}^\infty (\frac 1{\sqrt{7}})^n
= {r \sqrt{21}\over 14(\sqrt{7} -1)} \approx r(0.198892),$$
of the segment. If $I_1, I_2, I_3$ are consecutive segments
of length $r$ then $\mbox{\rm dist} (I_1, I_3) = r$, so this shows the
limiting arcs corresponding to them are at least distance
$r/2$ apart. Thus the boundary of the Gosper Island is a
Jordan curve, indeed, is the image of the unit circle under
a map $f$ satisfying
$$ \frac 1C \leq {|f(x) - f(y)| \over |x-y|^\alpha } \leq C,$$
where $\alpha = \frac 12 \log 7/\log 3$.
\item \label{annulus}
For any $\eta>0$, there is a topological annulus
with a rectifiable boundary, which separates $\mbox{\rm core}(G_0,\eta)$ from
the boundary $\partial G_0$ of the Gosper island. \newline
(By the previous property, the interior $G_0^\circ$ of $G_0$
is simply connected, so this annulus can be obtained, for instance,
by applying the Riemann mapping theorem.)
\end{enumerate}
\noindent{\bf Definitions: }
Let $K$ be a compact connected subset of the plane.
We say that
$\, G \in {\cal{D}} \, $ is a {\bf Whitney~tile} for $\, K \,$ if
$\, \lambda \odot G \,$ is disjoint from $\, K$,
but $\, \lambda \odot \mbox{\rm parent}(G) \, $ intersects $\, K$.
(See Figure~\ref{WhitTile}.)
Let $W_K$ denote the set of Whitney tiles for $K$.
This collection is called a Whitney decomposition of
$K^c$, since it decomposes $K^c$ into a countable union
of tiles (disjoint except for their boundaries) each with
diameter comparable to its distance from $K$. See Figure
\ref{TileChain}.
A chain of adjacent tiles
$\{ G_1 , G_2 , \ldots , G_j \} \, $ in $\, W_K \,$ such that
$\, G_i \subset
\lambda^{5} \odot G_1 \, $ and $\, ||G_{i}|| \geq ||G_{i-1}|| \,$
for all $\, i \in \{2, \ldots , j \} \, $ is called a {\bf Whitney~chain}
(see Figure~\ref{TileChain}).
Given $G \in W_K$, define $W_K^G \subset W_K$ to be the set of
tiles $G'$ such that
there is a Whitney chain
$\{ G_1 , G_2 , \ldots , G_j \}$ with $G_1 = G $ and $G_j = G'$.
\begin{figure}
\centerline{ \psfig{figure=WhitDefnSim.ps,height=2.5in} }
\caption{ \label{WhitTile} Boundary misses $\lambda \odot Q$, but
hits $\lambda \odot \mbox{\rm parent}(Q)$. }
\end{figure}
\begin{figure} \centerline{ \psfig{figure=Dec+Chain.ps,height=2.5in} }
\caption{ \label{TileChain} Whitney decomposition and a chain of tiles }
\end{figure}
Note the following property of the Whitney decomposition,
which holds for any connected component $\Omega$ of $K^c$:
\begin{equation} \label{Whit-property}
\mbox{ \rm
For any open $V$ intersecting $\, \partial \Omega\, $, there
is a tile } G_* \in W_K \mbox{ with } \lambda^{5} \odot G_*
\subset V.
\end{equation}
\begin{lem} \label{lem 1}
If $G_1 , G_2 \in W_K$ are adjacent then $||G_1|| - ||G_2||$
is 0 or $\pm 1$.
\end{lem}
\noindent{\sc Proof:} Suppose $\, ||G_1|| - ||G_2|| \geq 2$. Let
$G$ be the tile of index $\, ||G_2||+1 \,$ that contains $\, G_1$,
and observe that $G$ is adjacent to $G_2$. Then by Property \ref{neighbor}
of the Gosper tiling, \newline
$\lambda \odot \mbox{\rm parent} (G_1) \subset
\lambda \odot G \subset \lambda \odot G_2$
and maximality of $G_2$ is violated. $\fbox{\hphantom{} \vphantom{x}}$
\begin{lem} \label{lem 2}
Suppose ${\cal{C}} \subset W_K \cap {\cal{D}}_n$ is a collection of tiles whose
union topologically surrounds a smaller Whitney tile $G \in
W_K \cap {\cal{D}}_{n+k}$ where $k > 0$. Then ${\cal{C}}$ surrounds a
a point of $K$.
\end{lem}
\noindent{\sc Proof:} The $k$-fold parent of $G$ is a tile $D \in{\cal{D}}_n$
which is surrounded by ${\cal{C}}$. Applying Property \ref{union} inductively shows that
the union of $\lambda \odot G'$ for $G' \in {\cal{C}}$ surrounds whatever part
of $\lambda \odot D$ it does not contain. Thus maximality of $G$
implies that $\lambda \odot D$ intersects $K$, and any point of
intersection is surrounded by ${\cal{C}}$. $\fbox{\hphantom{} \vphantom{x}}$
\begin{lem} \label{lem 3}
Suppose that $ \, G \in W_K \, $ and that there is a Whitney chain from
some larger tile outside $\, \lambda^{5} \odot G \, $ to $\, G$. For any
$\, n > ||G|| \, $ define $\, \mbox{\rm Wall} (G,n) \,$ to be the set
$ \, W_K^G \cap {\cal{D}}_n$.
(See Figure~\ref{TileWall}.) Let
$$
E_n = \bigcup \{D \, : \, D \in \mbox{\rm Wall} (G,n) \}
\cup \partial (\lambda^{5} \odot G) \, .
$$
Then $ \, E_n \,$ is a connected set which topologically separates
$\, G \,$ from
$\, K \,$. Furthermore, If $\, \Gamma \, $ is a Jordan curve separating
$\, G \, $ from the
complement of $\, \lambda^{5} \odot G \, $, then every component of
$\, \bigcup \{D \, : \, D \in \mbox{\rm Wall} (G,n) \} \,$ intersecting the
domain bounded by $\, \Gamma \,$ also intersects $\Gamma$.
\end{lem}
\begin{figure}
\centerline{ \psfig{figure=Wall+Ugh.ps,height=2.5in} }
\caption{ \label{TileWall} The sets Wall(G,n) and U(G,h). }
\end{figure}
\noindent{\sc Proof:} By Lemma \ref{lem 1} any path which connects
$G$ to $K$ must hit Whitney tiles of every index
larger than $||G||$. Thus any such path either hits
$\mbox{\rm Wall}(G,n)$ or must leave $\lambda^{5} \odot G$, proving that
$E_n$ separates $G$ from $K$. Connectedness follows from the last
assertion of the lemma for $\Gamma = \partial (\lambda^{5} \odot G)$,
so it remains only to prove the last assertion.
Suppose to the contrary that there is a component $U$
of $\, \bigcup \{D \, : \, D \in \mbox{\rm Wall} (G,n) \} \,$ which intersects
the domain bounded by $\Gamma$ but is disjoint from $\Gamma$ itself.
The union of all
Whitney tiles which are in the unbounded component of
$U^c$ and are adjacent to $U$ is a connected set.
By Lemma \ref{lem 1} all of these tiles have index
$n-1$ or $n+1$. By connectedness and
Lemma \ref{lem 1}, they must all have a single
index. Suppose they all have index
$n+1$. Since tiles in $U$ can be connected to $G$ by
Whitney chains which don't cross any tile of index $n+1$, this means
$G$ is in a bounded component of the complement of $U$, which
contradicts our assumption that $G$ could be connected by a
Whitney chain to a larger tile outside $\lambda^{5} \odot G$.
Thus the adjacent tiles must all have index $n-1$.
But then by Lemma \ref{lem 2}
these adjacent tiles must also surround a point of
$K$, which implies that $K$ is not connected, another
contradiction. $\fbox{\hphantom{} \vphantom{x}}$
The next two lemmas are needed in order to
show that if ``major portions'' of a wall of Whitney tiles
can be covered by a thin
strip, then $K$ must intersect the core of an appropriate tile $G''$
without $\eta$-surrounding it;
the latter event is controlled by the hypothesis of Theorem~\ref{th any K}.
\begin{lem} \label{lem short segment}
Fix $G' \in {\cal{D}}$ and $\beta \in (0, \eta_0)$. Let
$\hat U$ be any connected set intersecting both \newline
$\partial (\lambda^{5} \odot G')$ and $\lambda^3 \odot G'$. Suppose that
$\hat U \cap (\lambda^{5} \odot G')$ is contained in an infinite open strip
of width $2 \beta \mbox{\rm diam} (G')$.
Then there is a tile $G''$ contained in $\lambda^{5} \odot G'$ and
of the same index as $G'$, such that
$\hat U$ intersects $\mbox{\rm core}(G'', 2\eta_0-2\beta)$.
\end{lem}
\noindent{\sc Proof:} Pick a point $x \in \hat U \cap (\lambda^3 \odot G')$
and choose $y \in \hat U \cap \partial (\lambda^4 \odot G)$ connected
to $x$ inside $\hat U \cap (\lambda^4 \odot G)$.
By Property \ref{min-dist} of the tiling, the segment
$\overline{xy}$ has length at least $\, d_0 \cdot \mbox{\rm diam} (G')$,
so by Property \ref{eta} of the tiling, there is a tile
$G''$ of the same index as $G'$, such that $\overline{xy}$
contains some point $z \in \, \mbox{\rm core}(G'', 2 \eta_0)$.
By Property \ref{inrad} and convexity of disks, $z \in \lambda^5 \odot G$,
and therefore $G'' \subset \lambda^5 \odot G$.
Observe that $\mbox{\rm dist} (z, \hat U) < {2 \beta \mbox{\rm diam} (G')}$,
for if not, removing the open disk centered at $z$ of radius
$2 \beta \mbox{\rm diam} (G')$ from the infinite strip would
contradict the connectedness of $\hat U$.
This observation implies the assertion of the lemma.
$\fbox{\hphantom{} \vphantom{x}}$
Let $G$ be any Whitney tile with $\lambda^{5} \odot G$ not
containing all of $K$. Let $\gamma$ be a circle
centered at $\mbox{\rm center} (G)$
which separates $\lambda^4 \odot G$ from
$\partial (\lambda^5 \odot G)$. (Such a circle exists
by Property \ref{inrad} of the tiling.) For any
positive integer $h$, let $U(G,h)$ be the union of
all tiles $D \in W_K^G$ of index $||G|| + h$ such that $D$ intersects the
disk bounded by $\gamma$ (see Figure~\ref{TileWall}).
\begin{lem} \label{lem wiggly -> beta}
Choose $a_2$ so that $\lambda^{3 - a_2} \leq \eta_0 / 2$.
With $G$ as above, let $G'$ be a tile in $W_K^G$ with
$||G|| < ||G'|| <
||G|| + h - a_2$ such that $\lambda^{5} \odot G'$ is contained in
the disk bounded by $\gamma$.
Suppose that $U (G,h) \cap (\lambda^{5} \odot G')$ is covered by
an open strip of width $2 \beta \mbox{\rm diam} G'$ with $\beta < \eta_0 /4$.
Then
there is a tile
$G'' \subset \lambda^{5} \odot G'$ of the same index as $G'$,
such that $K$ intersects $\mbox{\rm core} (G'', \eta_0) $ without
$\eta_0$-surrounding $G''$.
\end{lem}
\noindent{\sc Proof:} By Lemma~\ref{lem 3},
$U (G,h) \cup \gamma$ is connected
and therefore satisfies the hypotheses of Lemma~\ref{lem short segment};
let $G''$ be a tile as in the
conclusion of that lemma.
Since $2\eta_0-2\beta > 3 \eta_0/2$, we can
pick a point $u$ in $U(G,h) \cap \mbox{\rm core}(G'', 3 {\eta_0}/2)$.
This clearly prevents $K$ from $\eta_0$-surrounding $G''$.
For any Whitney tile $D$ of index $||G||+h$, the blow-up
$\lambda^3 \odot D$ intersects $K$ (by Property \ref{odot} of the tiling).
Since $U (G,h) \cap (\lambda^{5} \odot G')$ is a union of tiles
of index $||G||+h$, it follows that
$$
\mbox{\rm dist}(u,K) < |\lambda|^{3-||G||-h} \leq |\lambda|^{3-a_2-||G'||} \leq
{\eta_0 \over 2} \mbox{\rm diam} (G'') \, ,
$$
by the choice of $a_2$.
Therefore $K$ intersects $\mbox{\rm core}(G'',\eta_0)$.
$\fbox{\hphantom{} \vphantom{x}}$
\section{A tree of Whitney tiles} \label{sec 3}
Fix a compact, connected $K \subset \mathop{\raise .45ex\hbox{${\bf\scriptstyle{|}}$$, a tile $G_* \in W_K$,
and a positive integer, $h$. We construct a tree $T = T(K , G_* , h)$
of Whitney tiles. The root of $T$ is $G_*$ and the remaining
generations of $T$ are defined recursively as follows.
Assume $T$ has been defined up to generation $n$
and for each $G$ in $T_n$, the $n^{th}$ generation of $T$, define
$\widetilde T_{n+1} (G)$
to be the set of tiles $D$ with the following properties:
\begin{quote}
1. $||D|| = ||G_*|| + (n+1)h$;
2. $D \in W_K^G$;
3. $\lambda^{5} \odot D \subset \lambda^{5} \odot G$.
\end{quote}
Let $ T_{n+1} (G)$ be
a subcollection of $\widetilde T_{n+1} (G)$ which has maximal cardinality
among all subcollections ${\cal{C}}$ for which
the expanded tiles $\{ \lambda^{6} \odot D : D \in {\cal{C}} \}$ are disjoint.
By maximality, $\bigcup\{ \lambda^{7} \odot D \, : \, D \in T_{n+1} (G) \}$
contains all tiles in $\widetilde T_{n+1} (G)$ and therefore
\begin{equation} \label{eq:prune}
|\widetilde T_{n+1} (G)| \leq |\lambda|^{14} |T_{n+1} (G)| \, .
\end{equation}
The children of $G$ in $T$ are defined to be the collection $ T_{n+1} (G)$.
Some trivial inductive observations are that $T_n \subset {\cal{D}}_{nh + ||G_*||}$,
that each $G \in {\cal{D}}_n$ is connected to $G_*$ by a Whitney chain, and
that the sets $\lambda^{5} \odot G$ are disjoint as $G$ runs over any $T_n$.
\noindent{\bf Some tree terminology:} $\:$ Let $V$ be a countable set.
\begin{description}
\item{(i)}
A mapping {\bf T} from a probability space ${\cal S}$ to the set
of trees on the vertex set $V$ is {\bf measurable} with respect to a
$\sigma\mbox{-field} $ $ {\cal{F}}$ on ${\cal S}$, if for any pair $\{v,v'\} \subset V$,
the event $\left[\{v,v'\} \mbox{\rm is an edge of {\bf T} } \right]$
is in $ {\cal{F}}$.
\item{(ii)}
For any tree $T$ with vertex set contained in $V$,
and any element $v \in V$, define
$\mbox{\rm trunc}_v (T)$ to be null if $v$ is
not a vertex of $T$, and otherwise let $\mbox{\rm trunc}_v (T)$ be $T$ with the
part below $v$ removed; more precisely, the vertices of $\mbox{\rm trunc}_v (T)$ are
the vertices of $T$ not separated from the root by $v$, and the edges
are the edges of $T$ spanning pairs of vertices in this smaller vertex set.
\end{description}
For any $G \in {\cal{D}}$, let ${\cal{F}}_G$
denote the $\sigma\mbox{-field}$ generated by the events
$\{ D \cap K \neq \emptyset \}$ for all tiles $D$ for which either
$||D|| \leq G$ or the interior of $D$ is disjoint from
$\lambda^{5} \odot G$.
\begin{lem} \label{lem tree-dep}
On the event $||G|| = nh + ||G_*||$, the random variable $\mbox{\rm trunc}_G \circ T$
is measurable with respect to ${\cal{F}}_G$.
\end{lem}
\noindent{\sc Proof:} Suppose $||G|| = nh + ||G_*||$ and consider an
event of the form
$$\{ \{ D , D' \} \mbox{ is in the edge set of } \mbox{\rm trunc}_G (T) \} ,$$
where $D$ and $D'$ are tiles of index $mh + ||G_*||$ and $(m+1) h + ||G_*||$
respectively, \newline
with $\lambda^{5} \odot D' \subset \lambda^{5} \odot D$.
If $m \geq n$ and $\lambda^{5} \odot D$ is not disjoint from
$\lambda^{5} \odot G$, then the edge $\{ D , D' \}$ cannot be in
$\mbox{\rm trunc}_G (T)$. If $m < n$ or $\lambda^{5} \odot D$ is
disjoint from $\lambda^{5} \odot G$ then the event that
$\{ D , D' \}$ is an edge of $T$ is the union of events witnessed
by particular Whitney chains of tiles, all tiles being either
disjoint from $\lambda^{5} \odot G$ or of index at most
$||G||$, so the event is measurable with respect to ${\cal{F}}_G$. $\fbox{\hphantom{} \vphantom{x}}$
The next lemma requires the traveling salesman theorem
described in the next section, so its proof is delayed until
Section~\ref{sec 5}.
\begin{lem} \label{lem main}
Assume the random set $K$ satisfies the hypotheses of
Theorem~\ref{th any K}. Fix any tile $G_*$ and $h > 0$
and let $T$ be the random tree $T(K , G_* , h)$.
There are constants $c_1 , c_2 > 0$ such that for any tile
$G \in {\cal{D}}_{nh+||G_*||}$,
\begin{equation} \label{eq main}
{\bf{E}} [\# T_{n+1} (G) \| {\cal{F}}_G ] \geq c_1 h |\lambda|^h - c_2 |\lambda|^h
\end{equation}
on the event that $G_* \in W_K$, the tile $G$ is in $T_n$,
and $\lambda^{5} \odot G_*$ does not contain $K$. The constants
$c_1$ and $c_2$ depend only on $c_0$.
\end{lem}
To prove Theorem \ref{th any K}, we also need two
general lemmas concerning trees.
Define the
{\bf boundary} $\partial T$ of the infinite rooted tree $T$ to be the set of
infinite self-avoiding paths from the root.
The next lemma is implicit in Hawkes (1981) and can be found
in a stronger form in Lyons (1990). For convenience,
we include the short proof.
\begin{lem} \label{lem lyons}
Let $T$ be an infinite rooted tree. Given constants
$C>0$ and $ \theta>1$, put a
metric on $\partial T$ by
\begin{equation} \label{eq: metric}
\mbox{\rm dist} (\xi , \xi') = C \theta^{-n} \mbox{ if }
\xi \mbox{ \rm and } \xi' \mbox{ share exactly } n \mbox{ edges.}
\end{equation}
Suppose that independent percolation with parameter $p \in (0,1)$
is performed on $T$, i.e., each edge of $T$ is
erased with probability $1-p$ and retained with probability $p$,
independently of all other edges.
If
$$
\dim(\partial T) < \alpha = {\log (1/p) \over \log \theta }
$$
then
with probability 1, all the connected components of retained edges in
$T$ are finite.
\end{lem}
\noindent{\sc Proof:}
It suffices to show that the connected component of the root
is finite almost surely. For any vertex $v$ of $T$,
denote by $|v|$ the number of edges between $v$ and the root.
By the dimension hypothesis
and the definition of the metric on $\partial T$,
there must exist cut-sets $\Pi$ in $T$ for which the
$\alpha$-dimensional cut-set sum
$$
\sum_{v \in \Pi} \theta^{-|v| \alpha} = \sum_{v \in \Pi} p^{|v|}
$$
is arbitrarily small. But for any cutset $\Pi$, the right-hand side is
the expected number of vertices in $\Pi$ which are connected to the root
after percolation; this expectation bounds the probability
that the connected component of the root is infinite.
$\fbox{\hphantom{} \vphantom{x}}$
The next lemma formalizes the notion of a random tree
which ``stochastically dominates'' the family tree of a
branching process. We require the analogue of a filtration in our setting.
\noindent{\bf Definition:} $\:$
Let $V$ be a countable set and let $T$ be a random tree
with vertex set contained in $V$, i.e., $T$ is a
measurable mapping from some
probability space $\, \langle {\cal S}, {\cal A}, {\bf{P}} \rangle$
to the set of trees on the vertex set $V$.
Say that $\sigma\mbox{-field}$s $\{{\cal{F}}_v \, : \, v \in V\}$ on ${\cal S}$
form a {\bf tree-filtration}
if for any $v,w \in V$ and any $A \in {\cal{F}}_v$,
the event
$A \cap \{ w \mbox{ \rm is a descendant of } v \mbox{ in } T \} \, $
is ${\cal{F}}_w$-measurable.
\begin{lem} \label{lem tree}
Let $V$ be a countable set and let $T$ be a random tree
with vertex set contained in $V$. We assume that $T$ is rooted
at a fixed $v_* \in V$.
Assume that $b > 1$ and a tree-filtration $\{ {\cal{F}}_v \, : \, v \in V \}$
exists such that $\mbox{\rm trunc}_v (T)$ (defined before Lemma~\ref{lem tree-dep})
is ${\cal{F}}_v$-measurable for each $v \in V $, and the conditional expectation
$$
{\bf{E}} ( \mbox{number of children of } v \mbox{ in } T \| {\cal{F}}_v) \geq b .
$$
If every vertex of $T$ has at most $M$ children and at least
$m$ children, $m \geq 0$, then
\begin{enumerate}
\item The probability that $T$ is infinite is at least $1-q > 0$,
where $q$ is the unique fixed point in $[0,1)$ of the polynomial
$$
\psi (s) = s^m + {b-m \over M-m} (s^M - s^m) \, .
$$
(Observe that $q = 0$ when $m > 0$.)
\item ${\bf{P}} (T \mbox{ is infinite } \| \tilde {\cal{F}}_{v_*} ) \geq 1 - q$ for any
$\tilde {\cal{F}}_{v_*} \subset {\cal{F}}_{v_*}$.
\item If $\partial T$ is endowed with the metric (\ref {eq: metric}),
then $\dim (\partial T) \geq \log b / \log \theta$ with probability
at least $1 - q$.
\end{enumerate}
\end{lem}
\noindent{\sc Proof:}
\noindent{\bf 1.} Let $ |T_n|$ be the size of the $n^{th}$ generation
$T_n$ of $T$.
and let
$\psi_n (s)$ denote the $n$-fold iterate of $\psi$.
We claim that for $s \in [0,1]$,
\begin{equation} \label{eq claim 1}
{\bf{E}} s^{ |T_n|} \leq \psi_n (s) .
\end{equation}
When $n = 1$, convexity of $x \mapsto s^x$ implies that
$$
s^{|T_1|} \leq s^m +
{|T_1| - m \over M-m} (s^M - s^m)
$$
and the claim follows by
taking expectations:
$${\bf{E}} s^{|T_1|} \leq s^m + {{\bf{E}} |T_1| - m \over M-m} (s^M - s^m)
\leq \psi_1 (s)$$
since $s^M - s^m \leq 0$. For $n > 1$ proceed by induction. Let
$|T_{n+1} (v)|$ be the number of children of $v$ if $v \in T_n$ and zero
otherwise, and use the argument from the $n=1$ case to
see that ${\bf{E}} (s^{ |T_{n+1} (v)|} \| {\cal{F}}_v ) \leq \psi (s)$ on the event
$v \in T_n$ (which is an event in ${\cal{F}}_v$). Giving $V$ an arbitrary
linear order (denoted ``$<$''), we have in particular
$${\bf{E}} (s^{ |T_{n+1}(v)|} \| T_n \mbox{\rm and }
T_{n+1} (w) \mbox { for } w < v \mbox { in } T_n)
\leq \psi (s) \, .
$$
for $v \in T_n$. Since
$$
s^{ |T_{n+1}|} =\prod_{v \in T_n} s^{ |T_{n+1}(v)|} \, , \;
\mbox{ this yields }
\; \; {\bf{E}} (s^{ |T_{n+1}|} \| T_n) \leq \psi (s)^{ |T_n|} \, .
$$
Taking expectations and applying the induction hypothesis with
$\psi (s)$ in place of $s$ gives
$$
{\bf{E}} s^{ |T_{n+1}|} \leq {\bf{E}} \Big(\psi (s)^{ |T_n|}\Big)
\leq \psi_n (\psi (s)) = \psi_{n+1} (s) \, ,
$$
proving the claim (\ref{eq claim 1}).
From (\ref{eq claim 1}) we see that
${\bf{P}} ( |T_n| = 0) \leq E q^{ |T_n|} \leq \psi(q) =q $,
establishing the first conclusion of the lemma.
\noindent{\bf 2.}
By copying the derivation of~(\ref{eq claim 1}), inserting an extra
conditioning on $\tilde {\cal{F}}_{v_*}$, one easily verifies
that $ {\bf{E}} (s^{ |T_n|} \| \tilde {\cal{F}}_{v_*}) \leq \psi_n (s)$, and
the rest of the argument is the same as in the first part.
\noindent{\bf 3.}
Let $T' (v)$ be the connected component of the
subtree of $T$ below $v$ after removing each vertex of $T$ below $v$
independently with probability $1-p$. For $p > 1/b$, let $q_p
\in (0,1)$ solve $q_p = 1 + (bp/M) (q_p^M - 1)$.
We apply the second part of the lemma to $T' (v)$ conditioned
on ${\cal{F}}_v$ to see that
$${\bf{P}} (T' (v) \mbox{ is infinite } \| {\cal{F}}_v ) \geq 1 - q_p$$
for $v \in T$. By Lemma \ref{lem lyons}, the
event $\{ \dim (\partial T) < |\log p| / \log \theta \}$
is contained up to null sets in the event $\{ T' (v)
\mbox{ is finite for all } v \in T_n \}$. Thus
\begin{eqnarray*}
{\bf{P}} \Big( \dim (\partial T) < |\log p| / \log \theta \, \Big| \, T_n \Big)
& \leq &
{\bf{P}} \Big( \cap_{v \in T_n} T' (v) \mbox{ finite } \, \Big| \, T_n \Big) \\[2ex]
& = & \prod_{v \in T_n} {\bf{P}} \Big( T' (v) \mbox{ finite } \, \Big| \, T_n \, , \,
T' (w) \mbox{ finite for all } w < v \mbox{ in } T_n \Big) \\[2ex]
& \leq & q_p^{ |T_n|}
\end{eqnarray*}
since each event conditioned on is in the corresponding ${\cal{F}}_v$.
Taking expectations yields
$${\bf{P}} \Big( \dim (\partial T) < {|\log p| over \log \theta} \Big)
\leq \psi_n (q_p) .$$
Since $q_p < 1$ for each $p > 1/b$, this goes to $q$ as
$n \rightarrow \infty$, proving the last conclusion
of the lemma.
$\fbox{\hphantom{} \vphantom{x}}$
\noindent{\sc Proof of Theorem}~\ref{th any K}:
Put a metric on $\partial T$ by
$$ \mbox{\rm dist} (\xi , \xi') = |\lambda|^{-(n+1)h - ||G_*||} \mbox{ if }
\xi \mbox{\rm and } \xi' \mbox{ share exactly } n \mbox{ edges.}
$$
Each $\xi = (G_1 , G_2 , \ldots) \in \partial T$ defines a
unique limiting point $\phi (\xi) \in \mbox {\rm \small frontier}(K)$ which is the
decreasing limit of the set $\lambda^{5} \odot G_n$. If
$\xi = (G_1 , G_2 , \ldots)$ and $\xi' = (G_1' , G_2' , \ldots)$
share exactly $n$ edges, then by definition of $ T_{n+1}$, the expanded tiles
$\lambda^{6} \odot G_{n+1}$ and $\lambda^{6} \odot G_{n+1}'$
are disjoint. Since $\phi (\xi) \in \lambda^{5} \odot G_{n+1}$
and $\phi (\xi') \in \lambda^{5} \odot G_{n+1}'$, it
follows from Property \ref{min-dist} of the tiling that
$$|\phi (\xi) - \phi (\xi')| \geq d_0 |\lambda|^{-(n+1)h - ||G_*||} .$$
Thus
$$|\phi (\xi) - \phi (\xi')| \geq d_0 \cdot \mbox{\rm dist}(\xi , \xi')$$
and since the range of $\phi$ is included in $\mbox {\rm \small frontier}(K) \cap
\lambda^{5} \odot G_*$ it follows that
\begin{equation} \label{eq local dim}
\dim (\mbox {\rm \small frontier}(K) \cap \lambda^{5} \odot G_*) \geq \dim (\partial T) .
\end{equation}
From Lemma \ref{lem tree} and the conclusion of Lemma~\ref{lem main}, we
see that
\begin{equation} \label{eq dim bound}
\dim (\partial T (K , G_* , h)) \geq { \log ( c_1 h |\lambda|^h -
c_2 |\lambda|^h ) \over h \log |\lambda|}
\end{equation}
with probability 1, on the event that $G_* \in W_K$ and $\lambda^{5}
\odot G_*$ does not contain $K$. Choose $h$
to maximize the RHS of~(\ref{eq dim bound}). Since the maximum
is greater than 1, there is an $\epsilon > 0$ for which
$$\dim (\partial T (K , G_* , h)) \geq 1 + \epsilon$$
with probability 1 on this event. Finally, let $\Omega$ be any connected
component of $K^c$.
By property (\ref{Whit-property}
of the Whitney
decomposition,
for any open $V$ intersecting $\partial \Omega$,
there is a tile $G_* \in W_K$
with $\lambda^{5} \odot G_* \subset V$, and the theorem
follows from~(\ref{eq local dim}).
$\fbox{\hphantom{} \vphantom{x}}$
\noindent{\bf Remark:} A planar domain $\Omega$ is called
a {\bf John domain} if
there is a base point $z_0 \in \Omega$
and a constant $C >0$ so
that any point $x \in \Omega$ can be joined to $z_0$ by a
curve $\gamma_x \subset \Omega$ so that
$ \mbox{\rm dist}(z, \partial \Omega) \geq C|x-z| $ for any $z \in \gamma_x$.
John domain were introduced by Fritz John in 1961, and
some basic facts about them can be found
in N{\"a}kki and V{\"a}is{\"a}l{\"a} (1994).
With the notation of the above
proof, if $G_* \in W_K$ is contained in a component $\Omega$
of $K^c$, choose for every tile $G \neq G_*$ in the tree
$T (K , G_* , h)$, a Whitney chain leading to $G$ from its unique
ancestor in the previous generation of the tree.
For each tile $G'$ in this chain, there is an open disk containing
it which is contained in $\lambda \odot G'$ (by Property \ref{inrad}
of the tiling).
The union of all these open disks as $G'$ runs over the chosen Whitney
chain for $G$ and $G$ runs over $T (K , G_* , h)$,
is a John domain $\Omega_{\rm J}$ satisfying
$\dim (\partial \Omega \cap \partial \Omega_{\rm J}) \geq 1 + \epsilon$.
\section{The traveling salesman theorem} \label{sec 4}
Given a set $E$ in the plane and another bounded
plane set $S$, we define
$$ \beta_E (S) = (\mbox{\rm diam} (S))^{-1} \inf_{ L \in \cal L}
\sup_{z \in E \cap S} \mbox{\rm dist} (z, L),$$
where $\cal L$ is the set of all lines $L$ intersecting $S$.
\begin{th}[Jones 1990] \label{th jones}
If $E \subset \mathop{\raise .45ex\hbox{${\bf\scriptstyle{|}}$ $ then the length of the shortest
connected curve $\Gamma$ containing $E$ is bounded between
(universal) constant multiples of
$$\mbox{\rm diam}(E) + \sum_Q \beta_E (3 \odot Q)^2 \mbox{\rm diam} (Q) , $$
where the sum is over all dyadic squares in the plane and
$3 \odot Q$ is the union of a 3 by 3 grid of congruent squares
with $Q$ as the central square.
\end{th}
A simpler proof of this Theorem, and an extension to higher dimensions,
are given in Okikiolu (1992).
The theorem easily implies that the
length $|\Gamma|$ of any curve $\Gamma$ which passes
within $r$ of every point of $E$ satisfies
\begin{equation} \label{eq jones}
\mbox{\rm diam}(E) + \sum_{\mbox{\rm diam}(Q) \geq r} \beta_E
(3 \odot Q)^2 \mbox{\rm diam} (Q) \leq c_3 |\Gamma| ,
\end{equation}
where the sum is over all dyadic squares in the plane with
diameter at least $r$.
For every set $S$, there is a dyadic square $Q$ of side length at most
$2 \mbox{\rm diam} (S)$ for which $S \subset 3 \odot Q$.
Picking $S = \lambda^{5} \odot G$ for some tile $G$ and
$Q$ accordingly, we get
$$\beta_E (\lambda^{5} \odot G)^2 \mbox{\rm diam} (G) \leq
9 |\lambda|^{-10} \beta_E (3 \odot Q)^2 \mbox{\rm diam} (Q) $$
and since each expanded square $3 \odot Q$ contains a bounded number of
expanded tiles $\lambda^{5}
\odot G$ for tiles $G$ with $\sqrt{2} |\lambda|^{5} \mbox{\rm diam} (G) \geq
\mbox{\rm diam} (Q)$, it follows that the length of any curve passing
within $r$ of every point of $E$ satisfies
\begin{equation} \label{eq jones2}
|\Gamma| \geq c_4 \left ( \mbox{\rm diam}(E) + \sum_{\mbox{\rm diam}(G) \geq r}
\beta_E (\lambda^{5} \odot G)^2 \mbox{\rm diam} (G) \right ) .
\end{equation}
We require the following corollary, which uses an idea from
Bishop and Jones (1994).
\begin{cor} \label{cor BJ}
Let $\gamma$ be a Jordan curve with length denoted $|\gamma|$ and
let ${\cal{C}}$ be a collection of Whitney tiles of index $n$. Let $U$
denote $\bigcup_{D \in {\cal{C}}} D$ and suppose that $\gamma \cup
U$ is connected. Then there is a constant $c_5$
such that the cardinality of ${\cal{C}}$ is at least
$$c_5 |\lambda|^n \left ( - |\gamma| + \sum_{G' \in \Xi ({\cal{C}})} \beta_U
(\lambda^{5} \odot G')^2 \mbox{\rm diam} (G') \right ) \; , $$
where $\Xi ({\cal{C}})$ is the collection of tiles $G'$ of index at most $n$
for which $\lambda^{5} \odot G'$ intersects $U$.
\end{cor}
\noindent{\sc Proof:} Let ${\cal{C}}_\circ$ be the collection of circles of
radius $r := |\lambda|^{-n} $ centered at points $\mbox{\rm center} (D)$
for $D \in {\cal{C}}$. Since neighboring tiles in ${\cal{C}}$ give rise to
intersecting circles in ${\cal{C}}_{\circ}$, we see that $\Gamma := \gamma \cup
\bigcup_{\Theta \in {\cal{C}}_{\circ}} \Theta $ is connected and passes within
$r$ of every point of $\gamma \cup U$. Furthermore, any
connected finite union of closed curves is a closed
curve, and hence $\Gamma$ is a curve of length at most
$|\gamma| + 2 \pi \# ({\cal{C}}) |\lambda|^{-n} $.
Combining this with~(\ref{eq jones2}) shows that
$$\# ({\cal{C}}) \geq {|\lambda|^n \over 2 \pi } \, \Big (
c_4 \sum_{\mbox{\rm diam} (G') \geq r} \beta_U (\lambda^{5} \odot G')^2
\mbox{\rm diam} (G') - |\gamma| \Big ) .$$
Since all tiles in $\Xi ({\cal{C}})$ have diameter at least $r$, this proves
the lemma. $\fbox{\hphantom{} \vphantom{x}}$
\section{ Expected offspring in the Whitney tree} \label{sec 5}
\noindent{\sc Proof of Lemma}~\ref{lem main}:
Fix $G_*$ and $h$
as in the statement of the lemma and let $G$ be
any tile in $T_n$. Let $\gamma$ be the circle separating
$|\lambda|^4 \odot G$ from $\partial (|\lambda|^5 \odot G)$,
which was used in Lemma \ref{lem wiggly -> beta}.
Let ${\cal{C}}$ be the
collection of tiles $D \in W_K^G$ of index $||G|| + h$ intersecting the
disk bounded by $\gamma$. The union of all tiles in
${\cal{C}}$ is the set $U= U (G,h)$
defined before Lemma \ref{lem wiggly -> beta}.
We want to
show that the expected cardinality of $ T_{n+1} (G)$ is large.
Since the cardinality of $ T_{n+1}(G)$ is at least
$|\lambda|^{-14}$ times the cardinality of $\widetilde T_{n+1}(G)$
by (\ref{eq:prune}), and $\widetilde T_{n+1} (G) $ is a superset of ${\cal{C}}$,
it suffices to show that
$${\bf{E}} (\# {\cal{C}} \| {\cal{F}}_G) \geq c_1' h |\lambda|^h - c_2' |\lambda|^h \, .$$
To do this, we will apply Corollary~\ref{cor BJ} to ${\cal{C}}$,
so that the set $U$ defined in that corollary
is the same as $U(G,h)$ defined above.
We will be able to bound from below the
summands in Corollary~\ref{cor BJ} for most, but not all,
``intermediate-sized'' tiles $G'$.
Pick an integer $a_3>1$ so that $|\lambda|^{3-a_3} < d_0$,
where $d_0$ is the minimal distance between nonadjacent tiles in ${\cal{D}}_0$.
Let $\widetilde \gamma$ be a circle concentric with $\gamma$,
with a smaller radius:
$\mbox{\rm rad}(\widetilde \gamma)=\mbox{\rm rad}(\gamma)-|\lambda|^{5-a_3}$ (see Figure \ref{GammaCir}).
\begin{figure}
\centerline{ \psfig{figure=GammaCir.ps,height=2.5in} }
\caption{ \label{GammaCir} The circles $\gamma$ and $\tilde \gamma$ }
\end{figure}
For $a_3< j < h $, let $W_K^G (j)$ be the set of tiles
$G' \in W_K^G$
such that $ ||G'||=||G|| + j$ and $\lambda^5 \odot G'$ intersects
the disk bounded by $\widetilde \gamma$ . For any such tile $G'$
the blow-up $\lambda^5 \odot G'$ is contained in the disk bounded
by $\gamma$.
Fix any tile $G' \in W_K^G(j)$ with $a_3 < j \leq ||G|| + h - a_2$,
where $a_2$ was specified in Lemma~\ref{lem wiggly -> beta}.
That lemma implies
\begin{eqnarray}
&& {\bf{P}} \left ( \beta_U (\lambda^{5} \odot G') \geq
{\eta_0 \over 4 } \, \Big| \, {\cal{F}}_G \mbox {\rm and } W_K^G(j) \right )
\label{eq asdf} \\[2ex]
& \geq & {\bf{P}} \Big[ \bigcap_{G''}
\Big( K \, \mbox{ \rm $\eta_0$-surrounds } G''
\, \mbox { \rm \underbar{or} } \;
K \cap \! \mbox{\rm core}(G'',\eta_0) = \emptyset \Big)
\, \Big| \, {\cal{F}}_G \mbox {\rm and } W_K^G(j) \Big] \, \nonumber ,
\end{eqnarray}
where the intersection is over all tiles
$G'' \subset \lambda^{5} \odot G'$
such that $||G''|| = ||G'||$.
The set of such tiles $G''$ for a fixed $G'$ has cardinality
$|\lambda|^{10}$.
Enumerating these and multiplying
conditional probabilities using the hypotheses of
Lemma~\ref{lem main} (since the $\sigma$-fields ${\cal{F}}_G$ and
$\sigma(W_K^G(j))$ are contained in $\sigma(K \setminus G''^{\circ})$ )
gives a lower bound of
$c_0^{|\lambda|^{10}} $
for~(\ref{eq asdf}), and implies that
$$
{\bf{E}} \Big( \beta_U (\lambda^{5} \odot G')^2 \, \Big| \,
{\cal{F}}_G \mbox{\rm and } W_G^K (j) \Big)
\geq \left ( {\eta_0 \over 4 } \right )^2
c_0^{|\lambda|^{10}} = c_6>0 \, .
$$
(This is the definition of $c_6$).
Since $\gamma$ is outside $\lambda^4
\odot G$, the distance from $\gamma$ to $\lambda^3 \odot G$
is at least $|\lambda|^3 d_0 \cdot \mbox{\rm diam}(G)$,
by Property \ref{min-dist} of the tiling.
Therefore the distance from
$\widetilde \gamma$ to $ \lambda^3 \odot G$ is at least
$ (|\lambda|^3 d_0 - \lambda^{5-a_3} )\cdot \mbox{\rm diam}(G)$,
which is greater than $ d_0 \cdot \mbox{\rm diam}(G) \,$ by the choice of $a_3$.
By Lemma~\ref{lem 3},
the union of $\widetilde \gamma$ with all the tiles in $W_K^G (j)$ is a connected
set. Since it intersects both $\lambda^3 \odot G$ and $\widetilde \gamma$,
it follows that the cardinality of
$W_K^G (j)$ is at least $d_0 |\lambda|^j$.
Thus for each $j \in (a_3, h-a_2]$ we have
$${\bf{E}} \Big( \sum_{G' \in W_K^G (j)} \beta_U (\lambda^{5} \odot G')^2 \, \Big| \,
{\cal{F}}_G \mbox{\rm and } W_G^K (j) \Big)
\geq c_6 d_0 |\lambda|^j .
$$
By Corollary~\ref{cor BJ},
$${\bf{E}} (\# {\cal{C}} \| {\cal{F}}_G) \geq c_5 |\lambda|^{(n+1)h+||G_*||} \left (
-|\gamma| + \sum_{j = a_3+1}^{h - a_2}
|\lambda|^{-nh-||G_*|| - j} c_6 d_0 |\lambda|^j \right ) .
$$
Summing gives
$${\bf{E}} ( \# {\cal{C}} \| {\cal{F}}_G) \geq c_5 \left (
c_6 d_0 (h - a_2-a_3)
|\lambda|^h - |\gamma| \cdot |\lambda| ^{(n+1)h + ||G_*||} \right )$$
which proves the lemma since $|\gamma| = 2 \pi |\lambda|^{4 - nh - ||G_*||}$.
$\fbox{\hphantom{} \vphantom{x}}$
\section{The Brownian frontier: Proof of Theorem 1.1}
\label{sec 6}
Let ${\bf{P}}_x$ denote the law of a planar
Brownian motion $\{B(t) \}_{t \geq 0}$
started at $x$. We use ${\bf{P}}_0$ unless indicated explicitly otherwise.
Let $\tau_{\exp}$ be a positive random variable,
independent of the Brownian
motion, which is exponential of mean 1 (i.e., its density is $e^{-t}$).
We will verify~(\ref{eq gosper-wiggly}) for $K = B[0,\tau_{\exp}]$.
by Brownian scaling, this will imply the first assertion of
Theorem~\ref{th main}.
\noindent{\bf Notation: } for any compact planar set
$S$, denote by $\tau_S = \min \{ t \geq 0 \, : \, B(t) \in S \}$
the first hitting time of $S$,
which is almost surely finite if $S$ has positive logarithmic capacity.
Given $\eta \in (0,1/10)$,
let $J_0$ be a rectifiable closed Jordan curve,
which is the exterior boundary of a topological annulus separating
$\mbox{\rm core}(G_0, \eta)$ from $\partial G_0$. (Here the constant $1/10$ can be
replaced by any constant smaller than the inradius of $G_0$,
and the existence of $J_0$ is guaranteed by
Property \ref{annulus} of the tiling.)
For the rest of this section, consider
a homothetic image $G=z_{\rm cen}+r G_0$ of the Gosper Island $G_0$,
with $r \in (0, 1)$ and $z_{\rm cen}$ in the plane.
Also, denote by $J=z_{\rm cen}+r J_0$
the image of $J_0$ in $G$.
We must obtain estimates which are uniform in the location and scale of
$G$, as well as in the structure of the Brownian range outside $G$.
\begin{lem} \label{lem:fastloop}
For every $x \in \mbox{{\rm core}$(G,\eta)$ } $,
\begin{equation} \label{eq:loop}
{\bf{P}}_x \Big( B[0,\tau_J] \; \,
\mbox{\rm $\eta$-surrounds } G \Big) \geq c_7(\eta) >0 \, .
\end{equation}
Furthermore, there exists $c_8(\eta)>0$ such that
\begin{equation} \label{eq:fast}
{\bf{P}}_x \Big( B[0,\tau_J] \; \mbox{ \rm $\eta$-surrounds } G \: \mbox{ \rm and }
\tau_J < \tau_{\exp} < \tau_{_{ \partial G}} \Big) \geq
c_8(\eta) {\bf{P}}_x(\tau_{\exp} < \tau_{_{ \partial G}}) \, .
\end{equation}
\end{lem}
\noindent{\sc Proof:} The first estimate is immediate for $G_0$,
and the general case follows by scaling.
For the second, observe that by Brownian scaling,
$\; \inf_{y \in J} {\bf{P}}_y (\tau_{_{ \partial G}} > \mbox{\rm diam}(G)^2) \: $ is a
positive constant depending
only on $J_0$, hence only on $\eta$. Also, clearly
${\bf{P}}(\tau_J < 1) >1/2$ and therefore
$$
{\bf{P}}(\tau_J < \tau_{\exp} < \tau_J+ \mbox{\rm diam}(G)^2) >
{e^{-2} \over 2} \mbox{\rm diam}(G)^2 \, , \; \mbox{ since } \, \mbox{\rm diam}(G) <1.
$$
Applying (\ref{eq:loop}), lack of memory of exponential variables, and
the strong Markov property
of Brownian motion at the stopping time $\tau_J$,
then shows that the left-hand side of
(\ref{eq:fast}) is at least a constant multiple of $\mbox{\rm diam} (G)^2$.
On the other hand, for any $x \in G$ we have
${\bf{P}}_x(\tau_{\exp} < \tau_{_{ \partial G}}) \leq {\bf{E}}_x ( \tau_{_{ \partial G}}) \leq \mbox{\rm diam}(G)^2$.
This completes the proof.
$\fbox{\hphantom{} \vphantom{x}}$
\begin{lem} \label{lem 3 parts}
There exists $c_9 (\eta) > 0$ such
that for any tile $G$, for any $x \in \mbox{{\rm core}$(G,\eta)$ } $ and any $A \subset
\partial G$,
$$
{\bf{P}}_x \Big(B[0,\tau_{_{ \partial G}}] \; \,
\mbox{ \rm $\eta$-surrounds } G \; \mbox{\rm and } B(\tau_{_{ \partial G}}) \in A \Big)
\geq c_9(\eta) {\bf{P}}_x (B(\tau_{_{ \partial G}}) \in A) .
$$
\end{lem}
\noindent{\sc Proof:}
Recall that $J$ is a Jordan curve of finite length
separating $\mbox{{\rm core}$(G,\eta)$ } $ from $\partial G$. The Harnack principle (see,
e.g.,~Bass (1995, Theorem~1.20)) implies
that there is a constant \newline $c_{10} = c_{10} (J_0)$
such that for any $y,z \in J$
and for any $A \subset \partial G$,
\begin{equation} \label{eq:harnack}
{\bf{P}}_y (B(\tau_{_{ \partial G}}) \in A) \geq c_{10} {\bf{P}}_z (B(\tau_{_{ \partial G}}) \in A) .
\end{equation}
Therefore for any $x \in \mbox{{\rm core}$(G,\eta)$ } $,
\begin{eqnarray}
&{\bf{P}}_x ( B[0,\tau_J] \; \, \mbox{ \rm $\eta$-surrounds } G \; \mbox{\rm and }
B(\tau_{_{ \partial G}}) \in A ) & \nonumber \\[3ex]
&= {\bf{E}}_x \Big( \mbox{\bf 1}_{\{ B[0,\tau_J] \; \mbox{ \rm $\eta$-surrounds } G \}}
\cdot {\bf{P}}_{\! B(\tau_J)} [B(\tau_{_{ \partial G}}) \in A ] \Big) & \label{eq:one-ave}
\end{eqnarray}
Applying the Harnack inequality (\ref{eq:harnack}) with $y=B(\tau_J)$
and then invoking the estimate
(\ref{eq:loop}) from the previous lemma, we find that the expression
(\ref{eq:one-ave}) is at least $c_7(\eta) c_{10} {\bf{P}}_z (B(\tau_{_{ \partial G}}) \in A)$,
for any $z \in J$. Finally, taking $z=B(\tau_J)$ and
averaging with respect to
${\bf{P}}_x$ using the strong Markov property gives
$$
{\bf{P}}_x ( B[0,\tau_J] \; \, \mbox{ \rm $\eta$-surrounds } G \mbox{\rm and }
B(\tau_{_{ \partial G}}) \in A ) \geq c_7(\eta) c_{10} {\bf{P}}_x (B(\tau_{_{ \partial G}}) \in A) \, ,
$$
for any Borel set $A \subset \partial G$.
This proves the lemma.
$\fbox{\hphantom{} \vphantom{x}}$
\begin{figure}
\centerline{ \psfig{figure=B1B2B3.ps,height=2.5in} }
\caption{ \label{B1B2B3} The partition of the Brownian trajectory }
\end{figure}
Given $\eta>0$, we abbreviate $\tau_{_{\rm C}}= \tau_{\mbox{\rm core}(G, \eta)}$ and
partition the Brownian trajectory into three
pieces:
\begin{enumerate}
\item Until the first time $\tau_{_{\rm C}}$ that the path visits $\mbox{\rm core}(G, \eta)$.
\newline
Formally, define $B^{(1)}(t) = B(t \wedge \tau_{_{\rm C}})$ for $t \geq 0$,
where $t \wedge s$ is shorthand for $\min \{t,s\}$.
\item From time $\tau_{_{\rm C}}$ until the next visit to $\partial G$, denoted
$\tau_{_{{\rm C}, \partial G}}=
\min \{ t \geq \tau_c \, : \, B(t) \in \partial G \}$. \newline
Define $B^{(2)}(t) = B((t+\tau_{_{\rm C}}) \wedge \tau_{_{{\rm C}, \partial G}})$ for $t \geq 0$.
\item After time $\tau_{_{{\rm C}, \partial G}}$. \newline
Denote $B^{(3)}(t)=B(t+\tau_{_{{\rm C}, \partial G}})$ for $t \geq 0$.
\end{enumerate}
The idea now is that $B^{(1)}$ and $B^{(3)}$ determine the Brownian range
outside $G^\circ$, and $B^{(2)}$ has a substantial chance
of $\eta$-surrounding $G$, even when we condition on its endpoints.
However, we still have to take the exponential killing into account.
Define the random variable
$$
I = \left\{ \begin{array}{lcr}
1 & \mbox{\rm if} & \tau_{\exp} < \tau_{_{\rm C}} \\
2 & \mbox{\rm if} & \tau_{_{\rm C}} \leq \tau_{\exp} < \tau_{_{{\rm C}, \partial G}} \\
3 & \mbox{\rm if} & \tau_{_{{\rm C}, \partial G}} \leq \tau_{\exp}
\end{array} \right. \,
$$
that indicates in which part of the motion the exponential killing occurred.
Finally, define
$$
\widetilde \tau_{\exp} = \left\{ \begin{array}{lcr}
\tau_{\exp} & \mbox{\rm if} & I=1 \\
\tau_{_{\rm C}} & \mbox{\rm if} & I=2 \\
\tau_{\exp}-\tau_{_{{\rm C}, \partial G}} & \mbox{\rm if} & I=3
\end{array} \right.
$$
\begin{pr} \label{pr:brown}
For any $\eta \in (0, 1/10)$ there is a constant $c_0= c_0(\eta) > 0$
such that for all homothetic
images $G=z_{\rm cen} +r G_0$ of $G_0$ with $r \in (0, 1)$
and $z_{\rm cen}$ in the plane:
\begin{eqnarray} \label{eq gosper-pf}
{\bf{P}} \Big ( B[0, \tau_{\exp}] \, \mbox{ \rm $\eta$-surrounds } G
\, \mbox { \rm \underbar{or} } \; B[0, \tau_{\exp}] \cap \mbox{{\rm core}$(G,\eta)$ } = \emptyset
\, \, \Big| \, \, {\cal{A}}_G \Big ) > c_0
\; \; \:
& \; & \;
\end{eqnarray}
where the conditioning is on the $\sigma$-field ${\cal{A}}_G$
generated by $I, \, B^{(1)}, \, B^{(3)} \mbox{\bf 1}_{\{I=3 \} } \,$ and $\, \widetilde \tau_{\exp}$.
\end{pr}
\noindent{\sc Proof:}
\, On the event $\{I=1\}$, the set $B[0, \tau_{\exp}]$
is disjoint from $\mbox{{\rm core}$(G,\eta)$ } $. \newline
To handle the case $\{I=2\}$, we use the strong Markov property at
time $\tau_{_{\rm C}}$ and apply the estimate (\ref{eq:fast}) to $B^{(2)}$.
Denoting
$\tau_{_{{\rm C}, J}}= \min \{ t \geq \tau_c \, : \, B(t) \in \partial G \}$, this gives
\begin{eqnarray*}
& {\bf{P}}_0 \Big( B[\tau_{_{\rm C}},\tau_{_{{\rm C}, J}}] \;
\mbox{ \rm $\eta$-surrounds } G \: \mbox{ \rm and }
\tau_{_{{\rm C}, J}} < \tau_{\exp} < \tau_{_{{\rm C}, \partial G}} \, \Big| \, I \geq 2 ; \, B^{(1)} \Big) & \\[3ex]
& \geq
c_8(\eta) {\bf{P}}_0 \Big( \tau_{_{\rm C}} < \tau_{\exp} < \tau_{_{{\rm C}, \partial G}})
\, \Big| \, I \geq 2 ; \, B^{(1)} \Big) \, . &
\end{eqnarray*}
This proves (\ref{eq gosper-pf}) on the event $I=2$.
Only the case $I=3$ remains.
By using the strong Markov property at time $\tau_{_{\rm C}}$ and
applying Lemma \ref{lem 3 parts} to $B^{(2)}$, we see that
for any $A \subset \partial G$,
$$
{\bf{P}}_0 \Big(B[\tau_C,\tau_{_{{\rm C}, \partial G}}] \; \,
\mbox{ \rm $\eta$-surrounds } G \; \mbox{\rm and }
B(\tau_{_{{\rm C}, \partial G}}) \in A \, \Big| \, B^{(1)} \Big)
\geq c_9(\eta) {\bf{P}}_0 \Big(B(\tau_{_{{\rm C}, \partial G}}) \in A \, \Big| \, B^{(1)} \Big) \, .
$$
In other words,
$$
{\bf{P}}_x \Big( B[\tau_C,\tau_{_{{\rm C}, \partial G}}] \; \,
\mbox{ \rm $\eta$-surrounds } G \, \Big| \, B^{(1)}, \, B(\tau_{_{{\rm C}, \partial G}}) \Big)
\geq c_9(\eta) \, .
$$
An application of the strong Markov property
at time $\tau_{_{{\rm C}, \partial G}}$ shows that this lower bound
is still valid if we insert an additional conditioning on
$I \geq 2$ and on $B^{(3)}$. \newline
Finally, since ${\bf{P}}_0(I =3 \| I \geq 2) \geq e^{-1}/2$ and
$\widetilde \tau_{\exp}$ is conditionally independent of $B[0,\tau_{_{{\rm C}, \partial G}}]$
given $I=3$, this completes the proof of the proposition.
$\fbox{\hphantom{} \vphantom{x}}$
To obtain the uniformity in Theorem~\ref{th main}, we will need the
following general observation.
\begin{lem} \label{lem semicont}
Let $\Gamma : [0,\infty) \rightarrow \mathop{\raise .45ex\hbox{${\bf\scriptstyle{|}}$$ be any continuous path
and let $t > 0$. For any open disk $U$ intersecting
$\mbox{\rm \small frontier}(\Gamma[0,t])$ such that
$\Gamma (t) \notin \overline{U} $,
there is a $\delta > 0$ such that for any $s \in
[t , t + \delta ]$, we have
\begin{equation} \label{eq:contain}
U \cap \mbox{\rm \small frontier} ( \Gamma [0,t]) \, \supset \,
U \cap \mbox{\rm \small frontier} (\Gamma [0,s]) \, \neq \emptyset \, .
\end{equation}
\end{lem}
\noindent{\sc Proof:}
By hypothesis $U$ intersects the unbounded component, $\Omega$.
of $\Gamma [0,t]^c$, so there is a point $u\in U$ with and
an unbounded curve starting from $u$ and contained in $\Omega$.
Using the convexity of $U$, we can append to this curve a line-segment
connecting $u$ to a nearest point $x$ on $\Gamma [0,t]$,
and thus obtain an unbounded curve $\psi$
starting at $x$ and contained in $\Omega \cup \{x\}$.
Choose $\delta>0$ small enough so that
$\Gamma [t,t + \delta]$ is disjoint from the curve $\psi$.
This gives the right-hand side of (\ref{eq:contain})
for $s \in [t , t + \delta ]$.
If we also require that $\Gamma [t,t + \delta]$ is disjoint from
$U$, then the left-hand side of (\ref{eq:contain})
follows from the general fact that
$$
\mbox{\rm \small frontier} ( \Gamma [0,s]) \subset \mbox{\rm \small frontier} (\Gamma [0,t]) \cup \Gamma [t,s] \, .
$$
$\fbox{\hphantom{} \vphantom{x}}$
\noindent{\sc Proof of Theorem~\ref{th main}:} $\:$
The random set $B[0, \tau_{\exp}] \setminus G^\circ$ is completely
determined by the variables generating the $\sigma$-field
${\cal{A}}_G$ defined in Proposition \ref{pr:brown}, so the
proposition implies that $K=B[0, \tau_{\exp}]$ satisfies
the hypothesis (\ref{eq gosper-wiggly}) of Theorem \ref{th any K}.
Since $B[0,\tau_{\exp}]$ has the same distribution as
$\sqrt{\tau_{\exp}} \cdot B[0,1]$,
this establishes the first assertion of
Theorem~\ref{th main}.
For $t>0$, Let $A_t$ be the event that
$\dim (\mbox{\rm \small frontier} (B[0,s]) \cap V) \geq 1 + \epsilon$ simultaneously for
all open disks $V$ that intersect $\mbox{\rm \small frontier}(B[0,t])$ and have
rational centers and radii. Theorem \ref{th any K}
and Proposition \ref{pr:brown} give ${\bf{P}}_0(A_1)>0$, and we must
show that
${ \displaystyle {\bf{P}}_0 ( \cap_{t>0} A_t ) = 1}$.
Denote by $A_{\rm I\!\!\!Q}= \cap_{s \in {\rm I\!\!\!Q}_+} A_s$ the
intersection over all positive rational times.
Now
Brownian scaling and countable additivity
imply that ${\bf{P}}_0 (A_{\rm I\!\!\!Q})=1$, so it suffices to prove that
$A_{\rm I\!\!\!Q} \subset A_t$ for all $t>0$.
Fix $t>0$ and an open disk $V$ that intersects $\mbox{\rm \small frontier}(B[0,t])$.
Since $\mbox{\rm \small frontier} (B[0,t])$ is connected, it must intersect some
(random) open disk $U = U(V,t)$ with rational center
and radius such that $U \subset V$ and $B(t) \notin \overline{U}$. By the
previous lemma,
there is a rational $s$ such that
$$
\mbox{\rm \small frontier} (B[0,t]) \cap U \, \supset \, \mbox{\rm \small frontier} (B[0,s]) \cap U \neq \emptyset \, .
$$
This implies that $A_{\rm I\!\!\!Q} \subset A_t$,
and completes the proof of the theorem.
$\fbox{\hphantom{} \vphantom{x}}$
Finally, we consider the {\bf planar Brownian bridge} $B_{\rm br}$,
which may be defined either by conditioning the Brownian path
to return to the origin, or by $B_{\rm br}(t) = B(t)-tB(1)$ for $t \in [0,1]$.
For every $t<1$, the restrictions $B_{\rm br}|_{[0,t]}$
and $B|_{[0,t]}$ have mutually absolutely continuous laws
(these laws are measures on the space of
continuous maps from $[0,t]$ to the plane.)
Therefore by Theorem \ref{th main},
for every {\em fixed} $t \in (0,1)$,
\begin{equation} \label{eq:bridge1}
\dim \Big(\mbox{\rm \small frontier}(B_{\rm br}[0,t]) \Big) \, \geq 1 + \epsilon \; \, \mbox{ a.s.}
\end{equation}
Consider a sequence of annuli $\{A_n\}$ of modulus $2^{-n}$
around the origin. The probability that $B_{\rm br}$
surrounds the origin in $A_n$ is bounded away from 0,
so the Blumenthal 0--1 law implies that
with probability 1, there is some rational
$t<1$ such that $\mbox{\rm \small frontier}(B_{\rm br}[0,1]) = \mbox{\rm \small frontier} (B_{\rm br}[0,t]) $
(see Burdzy and Lawler (1990)).
Thus by (\ref{eq:bridge1}), with probability 1,
$$
\dim \Big(\mbox{\rm \small frontier}(B_{\rm br}[0,1]) \Big) \, \geq
\inf_{t \in {\rm I\!\!\!Q} \cap (0,1)} \, \dim \Big(\mbox{\rm \small frontier}(B_{\rm br}[0,t]) \Big)
\geq \, 1 + \epsilon \,.
$$
\section{Concluding remarks}
It can be shown (Krzysztof Burdzy, personal communication)
that $\dim (\mbox{\rm \small frontier}(B[0,1]))$ is almost surely constant;
this fact is not required for the
arguments in this paper.
The conjecture that the Brownian frontier has dimension $4/3$ is
related to well-known conjectures
concerning self-avoiding random walks, which in turn are a model for
long polymer chains. In that context, the exponent $4/3$
first appeared in the non-rigorous
considerations of Flory (1949); see also de Gennes (1991).
Theorem \ref{th any K} is stated for general random sets,
rather than just Brownian motion, in view of potential
applications to the ranges and level-sets of other
stochastic processes.
Besides the range of Brownian motion, another natural random set that satisfies
the hypothesis of Theorem \ref{th any K} is the support
of {\bf super-Brownian motion}, i.e. the intersection of all closed sets
that are assigned full measure by this measure-valued diffusion
throughout its lifetime. (For the definitions see, e.g.,
Dawson, Iscoe, and Perkins (1989).)
Equivalently, this random set may be characterized as
the set of points ever visited by the path-valued process
constructed by Le-Gall (1993). (This process is often referred to as
``The Brownian snake''.).
We are grateful to Steve Evans for enlightening discussions of
super-Brownian motion.
To allow for further applications,
we state below a variant of Theorem \ref{th any K}
which obtains the same conclusions under weaker
hypotheses on the random set $K$.
We omit the proof, which requires the estimates
obtained by Pemantle (1994) for the probability that a Wiener sausage
covers a straight line segment.
For any set $S \subset \mathop{\raise .45ex\hbox{${\bf\scriptstyle{|}}$$ and any $\epsilon > 0$,
let $S^\epsilon$ denote the set $\{ x : |x-y| \leq \epsilon \mbox{ for some }
y \in S \}$.
Say that $K$ {\em is $\eta , \delta$-flat} inside $S$
if there is some line segment $\ell$ of length $\eta \mbox{\rm diam} (S)$ covered by
$K^{\delta \mbox{\rm diam} (S)}$, having $\ell^{\eta \mbox{\rm diam} (S)}$ inside $G$
with $\ell$ not topologically surrounded by $K \cap G \cap
(\ell^{\delta \mbox{\rm diam} (S)})^c$.
\begin{th} \label{th any K2}
Let $G_0$ be the Gosper Island, and
let $K$ be a random compact connected
subset of the plane.
Suppose that for some $\delta_0 > 0$, the following hypothesis
on $K$ is satisfied, where the supremum is over $r \in (0,1)$
and ${\bf x}$ in the plane.
\begin{equation} \label{eq sup condition}
\sup_{G = {\bf x} + r G_0} \mbox {\rm ess sup}\;
{\bf{P}} \left [ K \mbox{ is } \eta_0 , \delta_0 \mbox{-flat inside }
G \| K \cap G \neq \emptyset , \sigma (K \cap G^c) \right ] < 1 .
\end{equation}
Then there is an $\epsilon > 0$ for which
$\; \dim (\mbox {\rm \small frontier}(K)) \geq 1 + \epsilon \, $
with probability 1.
\end{th}
\noindent{\sc Remark:} The intuition behind the two-part
definition of $\eta , \delta$-flatness is
that for
$\mbox {\rm \small frontier}(K)$ to be close to straight (thus for $K$ to be flat),
$K$ itself must nearly cover a line segment and this must
happen somewhere that is not completely encircled by $K$.
For the special case when $K$ is the range of planar Brownian motion,
it seems likely that methods directly adapted to this case will yield
better estimates for $\dim (\mbox {\rm \small frontier}(K))$
than those obtainable by our methods. Indeed, Gregory Lawler
has informed us that immediately
after he learned of our Theorem \ref{th main}
(but without seeing its proof), he proved
(using completely different methods)
that the dimension of the Brownian frontier
can be expressed in terms of the
``double disconnection exponent'' of Brownian motion.
This allowed Lawler to deduce that $\dim (\mbox{\rm \small frontier}(B[0,1])) > 1.01$ a.s.,
by invoking recent estimates of Werner (1994) on disconnection exponents.
We refer the reader to Lawler's forthcoming paper
for this and several other striking results on the Brownian frontier.
Finally, we note an application to simple random walk on the
square lattice $Z\!\!\!Z^2$.
Given a subset $S$ of $Z\!\!\!Z^2$,
say that a lattice point $x \in S$ is on the {\bf outer boundary}
of $S$ if $x$ is adjacent to some point in the unbounded component
of $Z\!\!\!Z^2 \setminus S.$
We remark that using the strong approximation
results of Auer (1990) and our construction of the Whitney
tree in Section 3, it is easy to derive the following.
\begin{cor} \label{cor:srw}
Let $\{S(k)\}$ denote simple random walk on $Z\!\!\!Z^2$,
and let $\epsilon>0$ be as in Theorem \ref{th main}.
Then for every $\epsilon_1 < \epsilon$ we have
$$
\lim_{n \to \infty}
{\bf{P}}\{ \mbox{There are more than } \: n^{(1+\epsilon_1)/2} \,
\mbox{ points on the outer boundary of } \: S[0,n]. \} = 1
$$
\end{cor}
\renewcommand{\baselinestretch}{1.0}
|
\section{Introduction}
The main motivation for the models of CP violation has been to
explain the $\epsilon'/\epsilon $ parameter for neutral kaon decay
simultaneously preserving
the vanishing smallness of the electric dipole moment of the neutron.
The idea that CP violation can be accomodated in extensions of the
Higgs sector
was put forward by Weinberg \cite{Wein1} in his three Higgs doublet
model, and T.D. Lee \cite{Lee1} in his two Higgs doublet model (2HDM)
with spontaneous CP violation. In these models the Yukawa couplings
have to be constrained to suppress flavor changing neutral currents.
A natural condition is to require that either $\phi_1$ couples to the
up quarks and $\phi_2$
to the down quarks or $\phi_1$ couples to all the quarks and $\phi_2$
does not couple at all. A discrete symmetry $\phi_1 \longrightarrow
-\phi_1$ is imposed on the Higgs potential to ensure suppression. In
majority opinion the minimal standard model suffers from inadequate
CP violation coming from the CKM matrix to be considered a viable
model for baryogenesis at the electroweak scale. So the 2HDM with
with CP violation in the Higgs sector has attracted attention for the
past several years in the context of electroweak baryogenesis.
McLerran-Shaposhnikov-Turok and Voloshin \cite{MacL} had shown that
in this model, the triangle diagram with one gauge boson leg and two
scalar legs and top quark in the loop contributes a correction to the
effective action $\sim {\cal O} N_{cs}$, where $\cal O$ is an operator
that can be found out by explicit evaluation of the diagram and
$N_{CS}$ is the Chern-Simons (CS) number. This correction biases the
CS number and since $\bigtriangleup N_B = n_{f} \bigtriangleup
N_{CS}$, nett baryon number results, with the sign determined by the
sign of $<\cal O >$.
Recently Yue-Liang-Wu \cite{Wu} has generalised the model of Glashow
and Weinberg \cite{Wein2} and has proposed that the value of
$\epsilon'/\epsilon $ and the limits on EDM can be well explained
with less stringent conditions on the Yukawa couplings. The Higgs
boson-fermion couplings are now much more general and no discrete
symmetries are imposed on the Higgs potential. In this class of
models fermions acquire mass from both $\phi_1$ and $\phi_2$.
Consequently in addition to the triangle diagrams with $\phi_1$
on the external legs, diagrams with $\phi_2$ also on the external legs
start contributing to the effective action.
The MSTV proposal has been criticised on the grounds that it is a higher
order adiabatic effect, of the order $(\langle \phi \rangle^T/T)^4$.
In their considerations the relative phase between the two Higgs vacuum
expectation values was treated constant throughout bubble evolution.
Here we provide the details of the argument \cite{sbdanduay2} that the
inclusion of the dynamics of the
relative phase removes the adiabatic suppression,
making it an effect
of the order $(\langle \phi \rangle^T/T)^2$. We have also previously shown
\cite{sbdanduay1} that the uncertainties
of physical parameters of bubble formation \cite{Dineetal} can be
circumvented if we consider a string induced phase transition
\cite{Yajnik}. This is the scenario we shall be considering here as
well. It may be noted that since according to \cite{sbdanduay1}, the
string induced bubbles provide adiabatic conditions, all B-genesis
mechanisms relying on such conditions are workable. In particular the
mechanisms of Cohen and Kaplan \cite{Cohen2} and Cohen-Kaplan-Nelson
\cite{Cohen3} can also proceed through string induced bubbles.
We recapitulate here the corresponding bubble
profile ansatz including the time evolution of the relative phase and
calculate the resultant contribution to the effective action.
Putting together the two sources of enhancement, viz., dynamics of the
relative phase and additional triangle diagrams,
we show that the mechanism has sufficient potential to give adequate
baryogenesis.
\section{The triangle diagram and correction to the effective action}
It was first pointed out by Turok and Zadrony \cite{Turok} and then
explicitly calculated by McLerran-Shaposhnikov-Turok and Voloshin
\cite{MacL}, that for 2HDM a term biasing the Chern-Simons number with
a CP odd chemical potential is contributed by
the triangle diagram of figure 1. This contribution to the effctive
action at finite temperature is
\begin{equation}
\Delta S={-7\over{4}}{\zeta(3)}{\left(m_{t}\over{\pi T}\right)^2}
{ g\over 16{{\pi }^2}}
{1\over{{v_1}^2}}
\times \int (
{\cal D}_{i}\phi_{1}^{\dagger}\sigma^{a}{\cal D}_{0}\phi_1 +
{\cal D}_{0}\phi_{1}^\dagger \sigma^{a}{\cal D}_{i}\phi_1)
\epsilon^{ijk} F_{jk}^{a}d^{4}x
\end{equation}
where $m_{t}$ is the mass of the top quark, $\zeta $ is the Riemann Zeta
function, and $\sigma^{a}$ are the Pauli matrices.
For homogeneous but time varying configurations of the Higgs fields
and in the ${A_0}^{a}=0$ gauge,
\begin{eqnarray}
\Delta S_1 &=& {-i7\over{4}}{\zeta(3)}{\left(m_{t}\over{\pi T}\right)^2}
{2\over{{v_1}^2}}
\int dt[{\phi_1}^\dagger{\cal D}_{0}\phi_{1} -
{\cal D}_{0}\phi_1^\dagger \phi_{1}] N_{CS}\\
& =& {\cal O}_{1} N_{CS}
\end{eqnarray}
Clearly, here the top quark acquires mass only from $\phi_1$ as
demanded by the Glashow-Weinberg natural flavor conservation (NFC)
criteria. In Wu's model, the FC criterion is satisfied
under relaxed conditions on the Higgs-fermions couplings. In this
case the general Yukawa interactions can be written as
\begin{equation}
L_Y = \bar{q_{iL}} {(\Gamma_{D}^{a})}_{ij} D_{jR} \phi_a
+\bar{q_{iL}} {(\Gamma_{U}^{a})}_{ij} U_{jR} \bar{\phi_a}
+ \bar{l_{iL}} {(\Gamma_{E}^{a})}_{ij} E_{jR} \phi_a + h.c
\end{equation}
where $q_{i}$, $l_i$ and$\phi_a$ are $SU(2)_L$ doublet quarks,
leptons and Higgs bosons, $U_i$, $D_i$, $E_i$ are $SU(2)_L$ singlets.
$i = 1,2,...n $ is a generation label and $a = 1,2$ is a Higgs
doublet label. ${\Gamma^a}_F$ $(F = U, D, E)$ are the Yukawa coupling
matrices.
According to Wu the Glashow-Weinberg criteria can be replaced by a
theorem which states that the flavor conservation for the neutral
currents is natural in the Higgs sector or equivalently, the matrices
${\Gamma^a}_F$ $(F = U, D, E)$ are diagonalizable simultaneously by a
biunitary or biorthogonal transformation, if and only if the square
$n \times n$ ${\Gamma^a}_F$ are represented in terms of the linear
combinations of a complete set of $n \times n$ matrices $
({\Omega^\alpha}_F, \alpha = 1,2,...n)$.
\begin{equation}
{\Gamma^a}_F = \sum_{\alpha} {g^F}_{a\alpha} {\Omega_F}^\alpha
\end{equation}
where, ${\Omega_F}^\alpha$ satisfy the following orthogonal condition
${\Omega_F}^\alpha {({\Omega_F}^\beta)}^\dagger = {L^\alpha}_F
\delta_{\alpha\beta}$,\\ ${({\Omega_F}^\alpha)}^\dagger
{\Omega_F}^\beta= {R^\alpha}_F \delta_{\alpha\beta}$ , with the
normalization $\sum_\alpha {L^\alpha}_F = \sum_\alpha {R^\alpha}_F =
1$.
This generalised Yukawa coupling prompts us to consider another set
of triangle diagrams as shown in fig-2. The contribution at
finite temperature from these diagrams to the effective action can
similarly be calculated to be
\begin{equation}
\Delta S_2 = {-i7\over{4}}{\zeta(3)}{\left(1\over{\pi T}\right)^2}
{{\Gamma}^{2}}^2 \int dt[{\phi_2}^\dagger({\cal D}_{0}\phi_{2}) -
({\cal D}_{0}\phi_2)^\dagger \phi_{2}] N_{CS}
\end{equation}
\begin{equation}
\Delta S_3 = {-i7\over{4}}{\zeta(3)}{\left(1\over{\pi T}\right)^2}
{\Gamma}^{1}{\Gamma}^{2}\int dt[{\phi_1}^\dagger({\cal
D}_{0}\phi_{2}) -
({\cal D}_{0}\phi_1)^\dagger \phi_{2}] N_{CS}
\end{equation}
where $\Gamma_{1(2)}$ is the Yukawa coupling when the fermions couple
to $\phi_{1(2)}$.
\subsection{The bubble profile}
Now to find the bubble profile we use the following ansatze for the
finite temperature vacuum expectation value of the Higgs fields
\begin{eqnarray}
{\phi_{1}}^0 &=& {\rho}_{1}(r,t) e^{-i\theta(t)} \\
{\phi_{2}}^0 &=& {\rho}_{2}(r,t) e^{i\omega(t)}
\end{eqnarray}
where, as pointed out by Cohen-Kaplan-Nelson \cite{Cohen1} we can
fix the unitary gauge ensuring $\theta$ is the physical pseudo-scalar
orthogonal to the Goldstone boson eaten by Z.
This gauge fixing gives the relation between $\omega$ and $\theta$
to be,
\begin{equation}
\partial_\mu \omega = {(\rho_1 / \rho_2)}^2 \partial_\mu \theta
\end{equation}
We assume that the phase transition takes place when a combination of
$\rho_{1}$ and
$\rho_{2}$ becomes massless in a particular direction $\gamma $ in
the ${\phi_{1}}^0 - {\phi_{2}}^0 $ plane. Hence, in the bubble
profile we may take
\begin{equation}
{\rho}_1(r,t)=\rho(r,t)\cos{\gamma} \hskip 1true in
\rho_2(r,t)= {\rho}(r,t)\sin{\gamma}
\end{equation}
with this parametrization and finite temperature corrections,
the effective potential for the $\rho$, $\theta$ and $\omega$ is
\begin{eqnarray}
V_{T}(\rho,\theta,\omega)&=&{M_1}^{2}(T){\rho}^2
- ET{\rho}^{3} + {K_{1} \over {4}} {\rho}^4 \nonumber \\
&&+ {M_2}^{2}(T)(\cos{\xi} \cos{(\omega + \theta)}
+ {C_1}\sin{\xi}\sin{(\theta + \omega)} ){\rho}^2 \nonumber \\
&&+ {K_{2}\over {4} }{\rho}^4 (\cos^2{(\theta + \omega)}
+{C_{1}} \sin^{2}{(\theta + \omega )})
\end{eqnarray}
where, ${M_{1}}^2$ and ${M_{2}}^2$ are temperature corrected mass
parameters and $K_{1}$ and $K_{2}$ are combinations of quartic coupling
constants with small temperature dependent corrections. The constant
$C_{1}$ is the ratio of ${\lambda _5}$ and $\lambda _{6}$ in the
standard parameterisation of the 2HDM \cite{hhguide}, and $\xi$ is
the phase of the neutral component of zero temperature vacuum
expectation value of $\phi_2$.
Rescaling ${\rho \longrightarrow ({2ET/{K_{1}}}){\rho}^{\prime}},
r \longrightarrow{{r^\prime}/{2ET\over {K_{1}}}}$ and $t$
$\longrightarrow{{t^\prime} /{2ET\over {K_{1}}}}$
omitting the primes the potential can be written as,
\begin{eqnarray}
V_{T}(\rho,\theta, \omega)&=&{\left( {2ET}\over {K_{1}}\right)}^{4}
[{K_{1}\over{2}}(1+E_{1}){\rho}^{2}
- {K_{1}\over {2}}{\rho}^{3}+ {K_{1}\over {4}}
{\rho}^{4} \nonumber \\
&&+ {K_{2}\over{2}}(1+E_{2})(\cos{\xi} \cos(\omega + \theta)+
{C_1}\sin\xi \sin(\theta + \omega) ){\rho}^{2} \nonumber \\
&&+ {K_{2}\over {4} }{\rho}^4 (\cos^2(\theta + \omega) +
{C_{1}} \sin^{2}(\theta + \omega ))]
\end{eqnarray}
where $E_{1} = \frac{M_{1}^{2}K_{1}}{2 E^{2} T^{2}}-1 $
and $E_{2} = \frac{M_{2}^{2}K_{1}}{2 E^{2} T^{2}}-1 $.
The geometry of string induced bubbles is cylindrical. With this in mind,
the time dependence of $\rho$ and $\theta$ can
be found by solving the following equations,
\begin{eqnarray}
\lefteqn{ {\frac{{\partial}^{2}\rho}{\partial t^{2}}} -
{\frac{{\partial}^{2}\rho}{\partial r^{2}}}
-\frac{1}{r}\frac{\partial\rho}{\partial r}
+ \rho{\left( {\partial \theta}\over
{\partial t} \right)}^2 - {3\over2}K_{1}{\rho}^2
+ K_1{\rho}^3+ {K_{1}\over 2}(1+E_{1})\rho } \nonumber\\
&& +K_{2}C_{1}
\sin^{2}(\theta + \xi){{\rho}^3}
+ {K_{1}\over 2}(1+E_{2}) (\cos{\theta} + C_{1}
\sin{\xi} \sin{(\xi + \theta}))\rho = 0
\end{eqnarray}
\begin{eqnarray}
\lefteqn{ \frac{\partial^2\theta}{\partial t^2} + \frac{K_2}{4}
C\sin{2(\theta+\xi)}{\rho^2} } \nonumber\\
&&+\frac{K_1}{2}(1+E_2)(C \sin\xi\cos{(\theta + \xi)}
-\sin(\theta+\omega-\xi)) = 0
\end{eqnarray}
The time independent solution of $\rho$ can be found by imposing the
boundary condition $\rho \longrightarrow 0$ as $r \longrightarrow
\infty $ and ${{\partial \rho}\over{\partial r}} \longrightarrow
0$ as $ r \longrightarrow 0$. Subsequently, as the nontrivial minimum
becomes favorable, the same solutions begin to evolve in time.
The parameters used in the equations are,
\noindent
\centerline{\vbox{\noindent $K_{1}$, $K_{2}$ = 0.1 -- $0.001$, \\
$C_{1} = 1.1$, $C = C_{1}-1$, $E_{1} = -0.074$, $E_{2} = -0.07$,\\
$\xi = 0.2$, $E\sim 0.01$}}
The time evolution of the bubble profile $\rho(r,t)$
has been reported earlier \cite{sbdanduay1}.
In fig-3, we show the time evolution of the relative phase in the
regions where $\rho$ has become nonzero. If the initial reference value
is zero, it oscillates to reach the stationary value dictated by the
2HDM effective potential at the relevant temperature. We assume this
value to be $O(1)$ since no natural reasons prevent it from being so.
\subsection{Evaluation of the operator}
Now we can use these solutions to evaluate the average value of the
operators as,
\begin{equation}
{\cal O}_1 + {\cal O}_2 = 28 {\zeta(3)}{\left(1\over{\pi }\right)^2}
{ E^{2} \over {K_{1}^2}}A_{1} {\int {{\rho}^2}{\frac{\partial
\theta}{\partial t}} dt}
\end{equation}
and,
\begin{equation}
{\cal O}_3 = 28 {\zeta(3)}{\left(1\over{\pi }\right)^2}
{ E^{2} \over {K_{1}^2}}A_{2} {\int {{\rho}^2}{\frac{\partial
\theta}{\partial t}}{e^{i(\theta + \omega)} dt}}
\end{equation}
where $A_1 = {({\Gamma_1}\cos\gamma)}^2 + {({\Gamma_2}\sin\gamma)}^2$
and $A_2 =\Gamma_1\Gamma_2 sin2\gamma$. These parameters
have to be determined from the phenomenology of the 2HDM which is yet
far from being tested by current experiments.
Knowing the solutions to eqn.s (14)-(15), we can estimate the integral
in eqn. (16) to be ${\rho_{\infty}}^2\Delta\theta$ where
$\Delta\theta$ is the nett
change in the relative phase $\theta$ at any given point as the bubble
wall sweeps past it, and $\rho_{\infty}=1$.
There is an additional contribution from the transient part of
$\theta$, which can also be calculated numerically, but is not significant.
As for the term ${\cal O}_3$, its CP odd part has the magnitude
$\rho_{\infty}^2\Delta(\sin\theta)$ from arguments already given.
Putting in other known factors, we see that
\begin{equation}
{\cal O}_1 + {\cal O}_2 \simeq\ A_1\left(E/K_1 \right)^2 \Delta\theta
\end{equation}
and a similar contribution from ${\cal O}_3$.
Recall that $E$ is the dimensionless cubic
self-coupling induced by thermal loops, and $K_1$ involves the quartic
self-couplings of the 2HDM. For naturalness we would like
$\Delta\theta$ to be $O(1)$ but the remaining factor
is numerically a small magnitude,
perhaps between $1$ and $10^{-4}$. Note however that the effect is
not suppressed by
the physics of the process viz., the bubble wall dynamics. We emphasise
again that this is the consequence of the dynamics of the relative
phase of the 2HDM.
\section{Estimation of the asymmetry}
To estimate the baryon asymmetry we assume the presence of high
temperature sphaleron processes inside
the bubble wall. The rate of such transitions per unit time per
unit volume is of the order $\sim \kappa{\alpha_{w}T}^4$, $\kappa
\sim 1$ \cite{Amb}. The number of fermions
created per unit time in the bubble wall is given by
\begin{equation}
B = \kappa {(\alpha _{w} T)}^{4} l S\times
{\frac {1}{T}}{\frac {{\cal O}} {l}}
\end{equation}
where we have made use of a well established master formula
\cite{BocShap} \cite{Cohen1}, and
where $l$ and $S$ are the thickness and the surface area of the bubble
wall respectively. From which we get the baryon to photon ratio to be
\begin{eqnarray}
\bigtriangleup \equiv {\frac {n_{B}}{s}}& \simeq& {\frac{1}{N_{eff}}}
{(\alpha_{w})}^4 {\cal O } \nonumber \\
& \simeq & 10^{-8}\times \left( \frac{E}{K_1} \right)^2 \Delta\theta
\end{eqnarray}
where we have used $\alpha_w$ $\sim$ $10^{-\frac{3}{2}}$ and $N_{eff} \sim
100$. This answer easily accomodates the observed value of this number.
It is worth emphasising the physics of this answer which is fairly
robust against changes in the specific particle physics models. The
thermal rate contributes $10^{-6}$ through $\alpha_w^4$, and another
$10^{-2}$ is contributed by $N_{eff}$. The remaining smallness of the
answer follows from smallness of the thermal vacuum value $\langle
\phi \rangle^T$ $\sim ET/K_1$, which is small on the scale set by
$T_c$. The appearance of this particular physical quantity has to do
with our picture of the process occuring in bubble walls as the phase
transition is yet in progress.
\section{Conclusion}
The MSTV proposal has been
criticised \cite{Cohen1} on the grounds that the operator ${\cal O}$
is of the order of $(\phi/T)^4$. Since $\phi$, i.e., the temperature
dependent vacuum value $\langle \phi \rangle^T$ is $\ll T$,
the effect was
thought to be unacceptably small. But with time variation of the
relative phase allowed, we find
the relevent operator is of the order of $(\phi/T)^2$.
The effect is therefore not intrinsically suppressed.
We also note that
the same analysis could be fruitfully applied to the proposals of
\cite{Cohen2} and \cite{Cohen3}.
Secondly, we have shown that in principle there are more diagrams
in the 2HDM contributing to the effective action. Although in making
numerical estimates we are hampered by large number of unknown
parameters in the 2HDM,
it is worth remembering the existence of these effects
as potential sources of enhancement in electroweak baryogenesis.
\section{Acknowledgment}
This work has been supported in part by the Department of Science and
Technology, Government of India.
|
\section{Introduction}
\label{intro}
There are various ideas of generalization of the manifold
notion. For example Aronszajn and Marshall
\cite{1af,27af}
have developed theory of so-called subcartesian spaces.
Mostov \cite{28af} and Spallek \cite{36af,59af}
have defined spaces called at present Mostov spaces.
Roman Sikorski
defined differential spaces (d-spaces for short)
\cite{3bo}.
There is a
strong requirement for notions more general than manifolds
in theoretical physics both classical and quantum .
Here are few examples.
\begin{itemize}
\raggedright
\item[ a)] Theory of singularities in General Relativity; space-time
with its singular boundary is not a manifold
\cite{33af,1bo},
\item[ b)] The space of solutions of Einstein equation (superspace) is
not a manifold \cite{12bo},
\item[ c)] A class of non-differential processes described in book by
Arnold \cite{13bo}.
\end{itemize}
The generalizations of the manifold concept have not found
applications in physics yet. Among reasons of such situation I
can mention excessive abstractness of some
theories and small amount of theorems useful in practice.
Differential spaces in the sense of Sikorski are
generalization of the manifold concept by omitting the condition
of existence of local diffeomorphisms to ${{\rm I\hspace{-0.5mm}R}}^n$. The
generalization
admit to considerations larger class of object than manifolds.
Simultaneously, mathematical tools (definitions and theorems) of
the d-spaces theory is similar to that one usually used in
ordinary differential geometry.
From the point of view of
physics, where differential geometry plays a very important role,
differential geometry on d-spaces is a natural tool for
generalization of a physical theory. Other important argument
which distinguishes the d-spaces theory is a great number of useful
theorems.
Among various structures existing in a space-time (e.g.
topological, chronological, metrical, spinorial etc. ) only
differential structure is not well elaborated. Since the
structure may play a fundamental role in quantization of gravity
\cite{14bo}
every result
concerning d-structure of space-time models seems to be valuable.
\pagebreak
\section{Differential spaces in the sense of Sikorski}
\label{ds}
A differential space in the sense of Sikorski is defined as follows.
Let $(M,{\rm top}(M))$ be a topological space and \mbox{${\cal C}_0$} \
a set of real functions on $M$.
A function \fun{f}{M}{{\rm I\hspace{-0.5mm}R}} is local \mbox{${\cal C}_0$} -function on $M$ if for every
$p_0\in M$ there is a neighbourhood $U\in{\rm top}(M)$ and
$\varphi\in\mbox{${\cal C}_0$}$ such that $f|_U=\varphi|_U$.
The set of all local \mbox{${\cal C}_0$} \ functions on $M$ is denoted by
$(\mbox{${\cal C}_0$})_{_M}$.
\begin{defin}
A set of real functions ${\cal C}$ on $M$ is said to be closed with
respect to localization if ${\cal C}=({\cal C})_{_M}$.
\end{defin}
\begin{defin}
A pair \dspa{M}{C} \ is said to be a differential space in the sense of
Sikorski if:
\begin{itemize}
\label{defds}
\raggedright
\item[1)] ${\cal C}$ is closed with respect to localization, i.e.
${\cal C}=({\cal C})_{_M}$,
\item[2)] ${\cal C}$ is closed with respect to superposition with smooth
functions on ${{\rm I\hspace{-0.5mm}R}}^n$, $n\in{\rm I\hspace{-0.4mm}l\hspace{-1.1mm}N}$, i.e. for any
$\ckow{\alpha}{1}{2}{n}\in {\cal C}$ and any
$\omega\in C^{\infty}({\rm I\hspace{-0.5mm}R}^n)$, $\omega\circ
(\ckow{\alpha}{1}{2}{n})\in {\cal C}$.
\item[3)] $M$ is equipped with the topology ${\tau}_{_{\cal C}} $ which
is the weakest topology on $M$ in which functions of ${\cal C}$ are
continuous.
\end{itemize}
\end{defin}
A picture of our world is formed by measures providing us with
information coded in form of real functions. One can say that reality is
given to us by a family of real functions existing on $M$.
If the family is sufficiently "complete" one can identify it with the
d-structure ${\cal C}$. Then the first axiom of definition \ref{defds}
guarantees consistency of a local physics with the global one. The second
axiom means that information obtained by superposition with smooth functions
on
${{\rm I\hspace{-0.5mm}R}}^n$ do not lead to a new information not contained in ${\cal C}$
(see \cite{13af}).
Let us test how the above
described idea works in practice on example of cosmic string's space-time
with a scalar field.
\pagebreak
\section{D-space of cosmic string's space-time}
\label{cosmic}
The space-time of cosmic string $(M,g)$ is a pseudoriemannian manifold
with conical type quasiregular singularity and equipped with the metric
\begin{equation}
g=-dt^2+k^{-2}d\rho^2+{\rho}^2 d\phi^2 +dz^2 , \label{metric}
\end{equation}
where $k=(1-\Delta/2\pi)$ , $t, z \in {\rm I\hspace{-0.5mm}R}$, $\rho\in (0, \infty)$,
$\phi\in \langle 0, 2\pi ) $ and $\Delta $ stands for
the deficit of angle. This manifold
is isometric to $(C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2, \iota^*\eta^{(5)})$, where
$C^\circ$ is two-dimensional cone without vertex,
$\iota : C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2$ $ \rightarrow{\rm I\hspace{-0.5mm}R}^5$ is an embedding and
${\eta}^{(5)}$ is the five-dimensional Minkowski metric.
The singular boundary in this approach is represented by the set
${\rm S}\times {\rm I\hspace{-0.5mm}R}^2$, where ${\rm S}$ denotes vertex of the cone
$C^{\bullet}$ and the dot under $C$ denotes that now the vertex is
taken into account.
The presented dipheomorphic picture of a space-time of cosmic string is
interesting for two reasons. First, it shows
the sense of the conical singularity in a demonstrative
manner \cite{16af,15af}. Second, it enables
us to construct immediately the differential structure
in the sense of Sikorski for
cosmic string's space-time both with and without singularity.
In this approach, d-space of a cosmic string constitutes d-subspace
$(C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2,$ $ ({\cal E}_5)_{C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2})$
of
the d-space $({\rm I\hspace{-0.5mm}R}^5, {\cal E}_5 )$,
where
${\cal E}_5 =C^{\infty} ({\rm I\hspace{-0.5mm}R}^5)$
and the symbol $(\cdot)_{C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2}$ denotes an operation
of taking closure with respect to localization (definition
\cite{3bo,12af,62af}).
The cosmic string's space-time with singularity is not a manifold but it is
still a d-space in the sense of Sikorski. The set
$C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2 \subset {\rm I\hspace{-0.5mm}R} ^5$ can be treated as a d-subspace
of
the d-space $({\rm I\hspace{-0.5mm}R}^5, {\cal E}_5 )$. Then the d-space
$(C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2, ({\cal E}_5)_{C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2})$
represents background d-space of cosmic string's space-time with
singularity.
The above mentioned d-spaces of cosmic string are not convenient from
technical and methodological point of view. Therefore, in the following,
two auxiliary d-spaces \tildedsp{P^\circ} \ and \tildedsp{P^\bullet} \ will be investigated rather
than
$(C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2, ({\cal E}_5)_{C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2})$
and
$(C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2, ({\cal E}_5)_{C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2})$,
respectively.
The auxiliary d-spaces are defined as follows.
\begin{defin}
Let the set
$\tilde{P}^\circ:={\rm I\hspace{-0.5mm}R}\times
(0,\infty)\times \langle 0,2\pi\rangle \times {\rm I\hspace{-0.5mm}R}$ be a "parameter
space".
One can define the following d-structure
$\mbox{$ \tilde{\cal P}^\circ$} :={\rm Gen}(\ckow{\tilde{{\alpha}}^{\circ}}{0}{1}{4})$,
where functions
\fun{{\tilde\alpha}^{\circ}_{i}}{\tilde{P}^\circ}{\rm I\hspace{-0.5mm}R}, $i=0,1,\dots ,4$
are given by means of the following formulas
\begin{quote}
\raggedright
${\tilde\alpha}^{\circ}_{0}(\tilde p):=t$ \linebreak
${\tilde\alpha}^{\circ}_{1}(\tilde p):=\rho\cos\phi$ \linebreak
${\tilde\alpha}^{\circ}_{2}(\tilde p):=\rho\sin\phi$\linebreak
${\tilde\alpha}^{\circ}_{3}(\tilde p):=z$\linebreak
${\tilde\alpha}^{\circ}_{4}(\tilde p):=\rho$,
$\tilde{p}\in \tilde{P}^\circ $.
\end{quote}
\end{defin}
The pair \tildedsp{P^\circ} \ is a d-space useful for description of interior of
cosmic string's space time.
\begin{defin}
Let
$\tilde{P}^\bullet:={\rm I\hspace{-0.5mm}R}\times
\langle 0,\infty)\times \langle 0,2\pi\rangle \times {\rm I\hspace{-0.5mm}R}$,
be the "prolonged parameter space".
One can define the following d-structure
$\mbox{$ \tilde{\cal P}^\bullet$} :={\rm Gen}(\ckow{{\tilde{\alpha}}^{\bullet}}{0}{1}{4})$.
The functions
\fun{\tilde{\alpha}^{\bullet}_{i}}{\tilde{P}^\bullet}{{\rm I\hspace{-0.5mm}R}} are
defined as follows
$$\tilde{\alpha}^{\bullet}_{i}(\tilde{p}_\bullet):=
\lim_{\tilde{p}\rightarrow\tilde{p}_\bullet}
{\tilde\alpha}^{\circ}_{i}(\tilde{p}), $$
where $\tilde{p}\in\tilde{P}^\circ,\tilde{p}_\bullet\in
\tilde{P}^\bullet$ and $i= 0,1,\dots ,4$.
\end{defin}
The \tildedsp{P^\bullet} \ is used for description of a cosmic string's space time with
singularity.
Both \tildedsp{P^\circ} \ and \tildedsp{P^\bullet} \ are not Hausdorff topological spaces. Therefore
they are not dipheomorphic to
$(C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2, ({\cal E}_5)_{C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2})$
and
$(C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2, ({\cal E}_5)_{C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2})$,
respectively.
However, after some identifications of points of \mbox{$ \tilde{P}^\circ$} \ and of \mbox{$ \tilde{P}^\bullet$} \
one can obtain d-spaces dipheomorphic to above mentioned background
d-spaces of cosmic string's space-time (see \cite{16af,15af,62af} for
details). In addition, every statement concerning smoothness and
differential dimension true for \tildedsp{P^\circ} \ and \tildedsp{P^\bullet} \ holds also for
$(C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2, ({\cal E}_5)_{C^{\circ}\times {\rm I\hspace{-0.5mm}R}^2})$
and
$(C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2, ({\cal E}_5)_{C^{\bullet}\times {\rm I\hspace{-0.5mm}R}^2})$
(see \cite{62af} for details). Thus, without loss of correctness one can
confine further considerations to \tildedsp{P^\circ} \ and \tildedsp{P^\bullet} .
\section{Physical fields and differential structure for cosmic string}
\label{scfield}
Let us consider normal modes of a Klein-Gordon field on the cosmic
string's space-time background. The modes have the following form:
$$\fun{\tilde{\Psi}^{\circ}_{\epsilon,l,\beta}}{\tilde{P}^{\circ}}{\mbox{\rm\hspace{1.2mm},$$
\begin{equation}
\label{psito}
\tilde{\Psi}^{\circ}_{\epsilon,l,\beta}(t,\rho,\phi,z)=
N_{\epsilon,l,\beta}e^{-i\epsilon t}e^{i\beta
z}e^{il\phi}\rho^{|l|/k}F(l,k; \rho),
\end{equation}
where
$\epsilon,\beta\in {\rm I\hspace{-0.5mm}R}$,
$l\in\calk$,
$N_{\epsilon,l,\beta}$
is the normalization constant and $k\in (0, 1) $ is defined in
formula (\ref{metric}). The $F$ is an analytical function. Its detailed
form is without meaning for further studies.
\begin{defin}
Let \dspa{M}{C} \ be a d-space. A complex function \fun{f}{M}{\mbox{\rm\hspace{1.2mm} \
is said to be smooth complex function if \ ${\rm Re}f$ and ${\rm Im}f$
are smooth ones.
\end{defin}
It is easy to check that
\begin{stw}
\label{smootho}
For every
$\epsilon, \beta\in{\rm I\hspace{-0.5mm}R}$, $l\in\calk$ and $k\in (0, 1)$
the normal modes
\mbox{$\tilde{\Psi}^{\circ}_{\epsilon,l,\beta}$} \
are smooth functions on \tildedsp{P^\circ}.
\end{stw}
\begin{stw}
\label{smoothb}
For every $\epsilon ,\beta \in {\rm I\hspace{-0.5mm}R}$ and $l\in \calk$ the
modes naturally prolonged to singularity
\begin{equation}
\label{psitb}
\mbox{$\tilde{\Psi}^{\bullet}_{\epsilon,l,\beta}$} (p):=\lim_{q\rightarrow p}\mbox{$\tilde{\Psi}^{\circ}_{\epsilon,l,\beta}$} (q), q\in\mbox{$ \tilde{P}^\circ$}, p\in\mbox{$ \tilde{P}^\bullet$},
\end{equation}
are
\begin{itemize}
\raggedright
\item[{\bf a)}]
smooth functions on \tildedsp{P^\bullet} \ for $k=1/n$, $n=1,2,\dots$,
\item[{\bf b)}] not smooth functions on \tildedsp{P^\bullet} \ for other $k\in (0,1)$.
\end{itemize}
\end{stw}
Following proposition \ref{smootho}
one can say that
the d-space of a cosmic string's gravitational field is
from the very beginning "prepared" for insertion of a Klein-Gordon field
on its background. The fact of insertion of the Klein-Gordon field is from
the point of view of cosmic string's space time interior the only an
indication which functions among already existing in \mbox{$ \tilde{\cal P}^\circ$} \ are the normal
modes.
The situation is much more interesting when one takes into account
the conical singularity (proposition \ref{smoothb}). Then, in general,
the insertion
of the scalar field enforces a supplementation of \mbox{$ \tilde{\cal P}^\bullet$} .
The situation clarifies the following proposition.
\begin{stw}
Let $\hat{\cal P}^{\bullet}$ be the smallest d-structure containing
\mbox{$\tilde{\Psi}^{\bullet}_{\epsilon,l,\beta}$} \ as smooth functions;
$\hat{\cal P}^{\bullet}:= {\rm
Gen}(\ckow{\tilde{\alpha}^\bullet}{0}{1}{4},\mbox{$\tilde{\Psi}^{\bullet}_{\epsilon,l,\beta}$} )$. Then the d-space
$(\mbox{$ \tilde{P}^\bullet$} ,\mbox{$ \hat{\cal P}^{\bullet}$} )$ has the following properties
\begin{itemize}
\item[{\bf a)}]
$(\mbox{$ \tilde{P}^\bullet$} ,\mbox{$ \hat{\cal P}^{\bullet}$} ) = \tildedsp{P^\bullet}$ for $k=1/n$, $n=1,2,\dots$,
\item[{\bf b)}]
$(\mbox{$ \tilde{P}^\bullet$} ,\mbox{$ \hat{\cal P}^{\bullet}$} )$ \ is not dipheomorphic to \tildedsp{P^\bullet} \ for $k\ne 1/n$,
$n=1,2,\dots$,
\item[{\bf c)}]
{\rm dim} $T_{p}(\mbox{$ \tilde{P}^\bullet$},\mbox{$ \hat{\cal P}^{\bullet}$}) = 6$ and {\rm dim} $T_{p}\tildedsp{P^\bullet} = 5$
for $p\in {\bf S}$, where {\bf S} denotes the set of singular points.
\end{itemize}
\end{stw}
Details of proofs one can find in \cite{62af} .
For $\Delta\ne 2\pi (1-1/n)$ the insertion of a Klein-Gordon field
changes global properties of cosmic string's background d-space. The new
background d-space $(\mbox{$ \tilde{P}^\bullet$},\mbox{$ \hat{\cal P}^{\bullet}$})$ is not dipheomorphic to \tildedsp{P^\bullet} . For
example the dimension of its tangent spaces at singular points is 6
while for \tildedsp{P^\bullet} \ is 5. That means that $(\mbox{$ \tilde{P}^\bullet$},\mbox{$ \hat{\cal P}^{\bullet}$})$ do not represent
background d-space of a cosmic string.
In the case of $\Delta = 2\pi (1-1/n)$ the supplemented d-structure \mbox{$ \hat{\cal P}^{\bullet}$}
\ and \mbox{$ \tilde{\cal P}^\bullet$} \ are the same so the background d-space of a cosmic string's
space-time does not change.
The similar result holds for an electromagnetic field and gravitational
radiation on cosmic string's space-time background \cite{62af}.
Thus, one can formulate the following theorem.
\begin{tw}
\label{nostringA}
The conical space-time of cosmic string equipped with the metric
{\rm (}\ref{metric}{\rm )} can be a background for
a smooth Klein-Gordon field, electromagnetic field and
a smooth field of gravitational radiation the only
in the case of the following discrete spectrum of the deficit of angle:
$$\Delta=2\pi (1-1/n), $$
where $n=1,2,\dots $\ .
\end{tw}
\section{Comparison with results obtained by means of
field theory methods}
\label{gradiation}
The issues of previous section were obtained by means of strictly
geometrical methods within the theory of differential spaces in the
sense of Sikorski. There appear the question: whether an echo of these
anticipations one can find among results obtained by the field theory
methods?
In the present section I will discuss some results of papers by Hacyan
and Sarmiento \cite{44af} and also by Aliev and Gal'tsov
\cite{43af,56af}.
{\bf 1) }
Sarmiento and Hacyan have studied
the energy density spectrum of vacuum surrounding cosmic
string for massless fields with spin 0, 1/2, and 1. Briefly speaking
they have shown that the energy density of vacuum
is given by
the following formula
\begin{equation}
\label{vacuum}
\frac{de}{d\omega}=\frac{\hbar {\omega}^{3}}{2{\pi}^{2}c^{3}}\left(1-
\frac{1}{2\pi\omega k\rho}\sin (\frac{\pi}{k}){\rm
I}_{_{\frac{1}{k}}}(2\omega\rho)\right),
\end{equation}
where $k$ has meaning as in (\ref{metric}).
When
$k=1/n$, $n=1,2,\ldots$ and so for deficits of angle $\Delta=2\pi (1-1/n)$
the energy density is like for the Minkowski space-time:
$$\frac{de}{d\omega}=\frac{\hbar {\omega}^{3}}{2{\pi}^{2}c^{3}} . $$
One can say that in this case the energy density is not distorted.
The result is independent of spin.
{\bf 2) }
The conical nature of space-time cosmic string geometry is a cause of
a class of effects absent in the Minkowski space-time (for example: the
lensing effect \cite{51af}). One of the most interesting effect
is a gravitational
radiation emitted by a freely moving particle (conical bremsstrahlung).
The total energy emitted by the particle in the form
of gravitational radiation is given
by the formula
\begin{equation}
\label{radiation}
E=\frac{1}{k}{\sin}^{2}(\frac{\pi}{k})f(v, k, d),
\end{equation}
where $f$ \ is a function of parameters of motion. Its explicit form one can
find in the article by Aliev \cite{43af}. It is easy to check that
the conical bremsstrahlung effect vanishes when
$k=1/n$, $n=1,2,\ldots $ like in the above section.
\section{Final remarks}
Results concerning cosmic string's space-time obtained in previous
sections could be interpreted as follows. The assumption of smoothness of
physical fields introduces a new kind of coupling. Usually the coupling
of a gravitational field and physical fields holds throughout the
energy-momentum tensor in Einstein equations. The new coupling has a
global nature and carries out through differential structure of cosmic
string leading to changes in global properties of this
space-time.
When one enforce vanishing of such "differential coupling" then
the gravitational field of a cosmic string becomes quantized and
simultaneously every effect mentioned in the section \ref{gradiation}
vanishes.
|
\section{Introduction}
Multilayer films formed from transition metals and semiconductors have
long been studied because of their unusual superconducting
properties\cite{ruggiero} and because of possible application as x-ray
optical elements.\cite{barbee85} Many unusual phenomena have been
produced, ranging from the observation of dimensional crossover in
weakly coupled superconducting Nb layers in Nb/Ge
multilayers\cite{ruggiero} to the occurrence of bcc Ge in short-period
Mo/Ge multilayers.\cite{wilson} Unusual magnetic properties have
recently been observed in Fe/Si multilayers by workers at
ETH\cite{toscano} and Argonne.\cite{fullerton} A large
antiferromagnetic (AF) interlayer coupling in these multilayers
manifests itself in hysteresis loops as a high saturation field and a
low remanent magnetization. Similar magnetization curves are
associated with large interlayer coupling in metal/metal multilayers
like Fe/Cr and Co/Cu.\cite{demokritov,mosca} Much consideration has
been given to whether the coupling in the Fe/Si system has the same
origin as in the metal/metal multilayers.\cite{bruno,singh} Therefore
the question of whether the spacer layer in the Fe/Si multilayers is a
metal or semiconductor is of particular interest.
Previous work on Nb/Si,\cite{fullerton2} Co/Si,\cite{miura}
Ni/Si,\cite{wang} and Mo/Si\cite{stearns} multilayers have shown that
there is a strong tendency towards compound formation at the
metal/silicon interface. In general these multilayers consist of
polycrystalline metal layers separated by an amorphous silicon layer
which is bounded on either side by a layer of intermixed material.
The intermixed silicide layers in these films were amorphous unless
they were annealed at several hundred
$^{\circ}$C.\cite{miura,holloway} These previously studied multilayers
were therefore likely in their as-grown state to have
metal/semiconductor character because of the presence of the amorphous
silicon layer.
In order to investigate the character of the spacer layer in the Fe/Si
multilayer system, we have grown a large number of films with
different substrate temperatures, substrate types, and layer
thicknesses. When the Si spacer layer thickness is greater than about
20\AA, we find that the metal layers are crystalline but that the
spacer layers are amorphous, similar to the situation in other
transition metal/silicon systems. When the Si spacer layer thickness
is less than about 20\AA thick, the iron silicide spacer layer forms a
crystalline silicide with either the B2 or DO$_3$ structure. The B2
structure consists of two interpenetrating simple cubic sublattices
and is identical to the CsCl structure for a 1:1 ratio of Fe and
Si,\cite{vonkanel} while the DO$_3$ structure is an fcc lattice with
two inequivalent Fe sites.\cite{kudrnovsky} Extensive growth
experiments, described below, suggest that crystallinity of the spacer
layer is crucial for occurrence of the antiferromagnetic interlayer
coupling, in keeping with previous suggestions.\cite{fullerton} Since
both the B2 and DO$_3$ phases are metallic,\cite{vonkanel,kudrnovsky}
the fact that crystallinity is required for antiferromagnetic coupling
suggests that the coupling in Fe/Si has a common origin with that
observed in metal/metal multilayers.
\section{Experimental Methods}
The Fe/Si multilayers are grown in the ion-beam sputtering
(IBS) chamber whose layout is shown schematically in
Figure~\ref{chamber}. The system base pressure is typically about
2$\times$10$^{-8}$ torr. The ion gun is a 3 cm Kauffman source with
focusing optics.\cite{commonwealth} The energy of the ions leaving the
gun can be modulated by raising and lowering the voltage on the
acceleration grids, creating in effect an electrical shutter. The Ar
ions are incident on the sputter target at 1000V at an angle of about
45$^{\circ}$. The Ar pressure is maintained in the
2-3$\times$10$^{-4}$ range by a flow-controller coupled to a
capacitance manometer.\cite{MKS} Four 3" diameter sputter targets are
mounted on a tray which can be rotated by a stepper motor.\cite{MDC}
Layer thickness is monitored by a quartz-crystal oscillator which is
placed in close proximity to the substrates. The substrates are about
25 cm above the targets, clamped to a copper tray. The temperature of
the tray is monitored by a thermocouple and can be varied between
-150$^{\circ}$C and +200$^{\circ}$C.\cite{rsipaper} Three films are
grown per chamber pumpdown.
The thickness monitor, the controller for the stepper motor and the
ion-beam power supply are all interfaced to a personal computer which
has been programmed using the ASYST instrument control
package.\cite{keithley} When the system is depositing a multilayer,
the computer sends the material parameters to the thickness monitor,
rotates the stepper motor to its new orientation, and turns the ion
gun on. When the desired thickness is reached, the thickness monitor
turns the ion-gun off and prompts the computer for the next layer.
The basic design of the system is similar to one previously described
by Kingon {\it et al.}\cite{kingon}
The substrates for multilayers growth include glass coverslips,
oxidized silicon wafers, MgO (001) and Al$_2$O$_3$(0\={2}11). The
first two substrates, which are used for growth of polycrystalline
films, were rinsed in solvents before loading into the vacuum chamber.
The second two, which are used for epitaxial growth, are cleaned
according to a recipe reported by Farrow and coworkers.\cite{farrow}
The typical deposition rate for Fe is 0.2\AA/s while that for Si is
about 0.3\AA/s. All films are capped with a 200\AA\ Ge oxidation
barrier. The magnetic and structural properties of the films are
stable for at least one year. Ge is used for capping instead of Si in
order to prevent interference with element-specific soft x-ray
fluorescence measurements, which will be reported
elsewhere.\cite{carlisle}
X-ray diffraction characterization has been performed using a 18kW
rotating anode system outfitted with a graphite monochromator. All
spectra are taken using the Cu K$_{\alpha}$ wavelength. Conventional
high-resolution electron microscopy and electron diffraction have been
performed in order to characterize the microstructure of the
as-deposited films in cross-section. Magnetization curves are
obtained using a vibrating sample magnetometer. All the data shown
here were taken at room temperature.
\section{Results}
Overall the magnetic properties of the Fe/Si multilayers made by IBS
are similar to those made previously by magnetron
sputtering.\cite{fullerton,foiles} Definitive confirmation of AF
interlayer coupling in our multilayers has been obtained by polarized
neutron reflectivity measurements.\cite{ankner} For some unknown
reason the magnetic properties of our multilayers are closer to those
of the magnetron-sputtered multilayers than those reported on in a
previous study using IBS, where much lower saturation fields were
observed.\cite{inomata} The differences between the previous IBS-grown
films and ours may be related to the lower ion-beam voltage used by
Inomata {et al.}\cite{inomata} Comparisons on the basis of layer
thickness are made here only between films grown during the same
deposition run in order to insure that the relative layer thicknesses
are meaningful. Films with similar layer thicknesses have been grown
many times to establish reproducibility of the observed trends.
\subsection{Layer-Thickness Dependence of Properties}
\label{thicksec}
Forty- and fifty-repeat multilayers have been grown with t$\rm_{Fe}$ =
14, 20, 30, 40, and 50\AA\ and t$\rm _{Si}$ = 14 and 20\AA.
Magnetization curves for 50-repeat (Fe30\AA/Si20\AA) and
(Fe30\AA/Si14\AA) multilayers grown on glass at nominal RT (about
+60$^{\circ}$C) are shown in Figure~\ref{sithick}. On the y-axis of
this plot is the magnetic moment of the multilayer normalized to the
moment of an equivalent volume of bulk Fe. The magnetization curve of
the (30/20) multilayer looks much like that of an Fe film, while the
magnetization curve of the (30/14) multilayer shows the high
saturation field and low remanence which characterize AF interlayer
coupling. At its saturation field the magnetization of the (30/14)
multilayer is about the same as for the (30/20) multilayer. Both of
these films have moments only about half as large as an equivalent
volume of bulk Fe. Our observation of AF coupling for Si thicknesses
between 10 and 20\AA\ and the disappearance of coupling for Si thicker
than 20\AA\ confirm previous observations on magnetron-sputtered
films.\cite{fullerton}
X-ray diffraction spectra for these multilayers are shown in
Figures~\ref{sithickxrdlow} and ~\ref{sithickxrdhi}.
Figure~\ref{sithickxrdlow} shows the small-angle x-ray scattering
(SAXS) data with peaks at angles
\begin{equation}
n^2 \lambda^2 \, = \, 4 \Lambda^2 sin^2 \theta \; + \; 2\delta
\label{lowbragg}
\end{equation}
\noindent where $\lambda$ is the x-ray wavelength, $\rm \Lambda \:
\equiv \: t_{Fe} \: + \: t_{Si}$ is the multilayer bilayer period and
$\delta$ is the index of refraction for x-rays.\cite{fullerton3} This
grazing incidence data gives information about the quality of the
multilayer interfaces. Figure~\ref{sithickxrdlow} shows four
low-angle peaks for both films, indicating a reasonably strong
composition modulation along the growth direction. (The higher
frequency oscillations between 1$^{\circ}$ and 4$^{\circ}$ are
finite-thickness fringes from the Ge cap layer.) The most notable
difference between the two spectra is that the multilayer peaks are
broader for the AF-coupled t$\rm _{Si}$ = 14\AA\ film, indicating more
fluctuations in bilayer period and probably more interface roughness.
Using the spacing between peak positions to eliminate the unknown
$\delta$ from Eqn.~\ref{lowbragg} gives values of the bilayer period
$\Lambda$ for the two films. For the multilayer with nominal layering
of (Fe30\AA/Si20\AA)x50, the derived value for $\Lambda$ is (41.82
$\pm$ 0.07)\AA, while for the (Fe30\AA/Si14\AA)x50 film $\Lambda$ =
(38.10 $\pm$ 0.04)\AA. $\Lambda$ is 8.2\AA\ shorter than the nominal
value for the t$\rm _{Si}$ = 20\AA\ film, and 5.9\AA\ shorter than
nominal for the t$\rm _{Si}$ = 14\AA\ film. Although some of the
discrepancy between the nominal and observed bilayer period may be due
to calibration inaccuracies, most is undoubtedly due to intermixing of
the Fe and Si layers, in keeping with observations in the other
metal/Si multilayers.\cite{fullerton2,holloway} Throughout this paper
we will continue for convenience to refer to the films in terms of
their nominal layer thicknesses.
Comparison of the magnetization data to the x-ray data can give some
further insight into the question of intermixing. Because of the
presumed interdiffusion of the Fe and Si layers, the magnetic moment
of the Fe layers is also reduced from the nominal value. The missing
magnetic moment can be expressed as an equivalent thickness of Fe.
Figure~\ref{missmom} shows a plot of missing moment in units of \AA
ngstroms of Fe versus missing bilayer period determined from
multilayer peak positions in SAXS for films grown at room temperature
(RT). The plot shows that while the diffusion-induced reduction in
bilayer period varies between 1 and 8\AA, the missing Fe moment per
bilayer (for both interfaces) is consistently between 10 and 12\AA.
The one outlier in Figure~\ref{missmom} is for a film which had t$\rm
_{Fe}$ = 20\AA, the thinnest Fe for which we have ever observed
interlayer coupling. Other groups have previously observed a moment
reduction of 12-14\AA\ per bilayer in polarized neutron reflectivity
measurements on uncoupled Fe/Si multilayers with thick Si
layers.\cite{dufour,ankner2}
The disparity between the magnetic moment reduction and the bilayer
period reduction numbers may at first appear to be puzzling. This
disparity occurs because the moment and bilayer period are affected by
different aspects of the structure. In calculating the moment
reduction in \AA\ the assumption has been made that the Fe layer has
the magnetization of bulk Fe. This is equivalent to assuming that
there is no Si in the Fe layer, which is undoubtedly false. In
calculating the missing bilayer period, the assumption has also been
made that the spacer layer is pure Si, also clearly false. The fact
that the missing magnetic moment is almost constant irrespective of
the reduction in bilayer period suggests that the spacer layer is
non-magnetic independent of Si thickness. The lack of variation of
the missing moment is then explained by the diffusion of a constant
number of iron atoms into the silicon layer, irrespective of its
thickness. The wide variation of the measured bilayer period is most
likely related to the varying orientation and crystallinity of the
spacer layer, neither of which affects the magnetic moment if the
spacer itself is non-magnetic.
Figure~\ref{sithickxrdhi} shows the high-angle x-ray spectra where
peak positions give information about the orientation and
crystallinity of the films. The intense peak near 70$^{\circ}$ in
this plot is due to the Si substrate. Included are data for an
(Fe40\AA/Si14\AA)x40 antiferromagnetically coupled multilayer and for
an (Fe30\AA/Si20\AA)x40 uncoupled multilayer, both grown on oxidized
Si(001) at RT. The peaks for the (40/14) film are narrower than for
the (30/20). The Scherrer formula gives 78\AA\ or about two bilayer
periods for the coherence length of the (40/14) film and 34\AA\ or
about one bilayer period for the coherence length of the (30/20) film.
Coherence lengths in IBS-sputtered antiferromagnetically coupled films
are often as long as 200\AA. Fullerton {\it et al.} have inferred
that the spacer layer in thin-Si Fe/Si multilayers must be crystalline
based on their observation of coherence lengths longer than a bilayer
period.\cite{fullerton} In keeping with its superior crystallinity,
the (40/14) multilayer has one superlattice satellite on the low-angle
side of the Fe(002) peak. Typically only one satellite on the
low-angle side of the Fe (011) or (002) x-ray peak is observed for
polycrystalline multilayers grown on glass, in agreement with
observations by Foiles {et al.}\cite{foiles2}
The thin-Si multilayers which have AF coupling usually show a mixed
[001] and [011] orientation when grown on glass substrates at RT.
Occasionally t$\rm _{Si}$ = 14\AA\ films with a pure (011) orientation
are obtained at RT. The variation in texture may be due to changes in
film stress under slightly different deposition conditions. Stress
induced during deposition has been postulated to explain the mixed Mo
texture found in Mo/Ge multilayers.\cite{wilson} In contrast to the
thin-Si case, the thicker-Si Fe/Si multilayers which do not show
interlayer coupling always have a pure (011) texture. Since the (011)
plane is close-packed for the bcc crystal structure, one would expect
the (011) orientation to be energetically favored for the Fe in a
multilayer with amorphous Si. Films grown at nominal RT on glass or
oxidized Si substrates typically had rocking curves about 10$^{\circ}$
wide indicating a moderate amount of orientation.
Transmission electron microscopy (TEM) has been used to further
investigate the morphology of the films. TEM cross-sectional images
of an (Fe30\AA/Si20\AA)x50 multilayer and an (Fe40\AA/Si14\AA)x50
multilayer grown during the same deposition run are shown in
Figures~\ref{TEMlowres}a and \ref{TEMlowres}b, respectively. The most
salient features of the (30/20) multilayer are the long lateral
continuity of the layers and the smoothness of the interfaces. Since
there is no interlayer coherence in the (30/20) film, the crystalline
grains have a high aspect ratio. The (40/14) multilayer also has
long, continuous layer planes but has rougher interfaces, consistent
with the SAXS data.
Transmission electron selected-area diffraction patterns for the
(30/20) and (40/14) films are shown in parts c and d of
Figure~\ref{TEMlowres}. The (30/20) films shows only a Fe(011) ring,
consistent with the high-angle x-ray diffraction scans. The (40/14)
film, on the other hand, displays spots corresponding to the (011) and
(002) reflections seen using x-rays. The presence of spots rather
than rings in the (40/14) image implies the presence of large,
oriented crystallites in the film. Most interestingly, the (40/14)
image includes a faint spot near what would be the Fe(001) position
were the Fe(001) peak not forbidden by symmetry in the bcc crystal
structure. The (001) peak is allowed in the B2 and DO$\rm _3$ crystal
structures. The B2 structure is found in the equilibrium phase
diagram only at 10-22\% Si range of composition,\cite{binary} but
workers at ETH have grown this crystal structure throughout the range
of composition on Si substrates using MBE.\cite{onda} The DO$_3$ phase
found in the equilibrium phase diagram is Fe$_3$Si, which is
ferromagnetic.\cite{binary} Clearly a ferromagnetic spacer phase is
not consistent with the observation of antiferromagnetic interlayer
coupling, although a non-stoichiometric DO$_3$-structure phase might
have different magnetic order. The B2 and $\epsilon$ iron silicide
phases have both been previously suggested as possible candidates for
the spacer layer in AF-coupled Fe/Si
multilayers.\cite{fullerton,foiles2,dekoster} The position of the
(001) TEM spots is not consistent with the $d$-spacings of the
$\epsilon$ phase.
According to the powder-diffraction files for the B2 and DO$_3$
structures, only the (111) peak of the fcc-family DO$_3$ does not
coincide with a B2 peak. The (111) peak would be expected to be very
weak in the diffraction patterns formed from cross-sectional specimens
of the film. The reason is that a small number of grains contributes
to the cross-sectional image, and the probability of sampling a grain
with its (111) planes in the observable direction is small because of
the random in-plane orientation. Future work will include electron
diffraction studies of a (40/14) specimen prepared in the plan-view
geometry, where the number of grains which are sampled is considerably
larger and the odds of observing the fcc (111) peak are improved.
High-resolution TEM images of the (30/20) and (40/14) multilayers are
displayed in Figure~\ref{TEMhires}. The (30/20) film is shown in
\ref{TEMhires}a to have a crystalline Fe layer and amorphous spacer
layer, similar to the morphology seen before in
Mo/Si\cite{stearns,holloway} and Co/Si multilayers.\cite{miura} The
(40/14) multilayer in Fig.~\ref{TEMhires}b on the other hand is made
up entirely of crystalline layers. The coherence between the Fe and
silicide spacer is clearly evidenced by the continuity of atomic layer
planes from the Fe layer into the spacer. Some crystallites in the
(40/14) film extend all the way from the substrate to the surface of
the film. The small coherence lengths observed in x-ray diffraction
data for the uncoupled thicker-Si films are explained by the presence
of the amorphous layers. The lack of crystallinity in the spacer
layer of t$\rm _{Si}$ = 20\AA\ films is presumably due to insufficient
time for full interdiffusion and ordering in the thicker layers. A
kinetic mechanism for the lack of crystallization is supported by
experiments which show that intentional placement of Fe in the Si
layer allows thicker spacer layers to
crystallize.\cite{foiles,mattson}
Another striking feature of the image in Figure~\ref{TEMhires}b is the
periodic modulation that occurs in the silicide spacer layer. The
modulation originates from scattering by inequivalent planes of atoms.
Simulation of this image using a multiple-scattering computer
calculation may be helpful in positively identifying the crystal
structure of the spacer layer phase.
Dark-field images of the (40/14) multilayer can help answer questions
about the texture of the film as well. Figure~\ref{TEMdark}a shows
the same bright-field image as in Figure~\ref{TEMlowres}b. Dark-field
images were formed using (001), (002) and (011) spots from the
diffraction pattern shown in Figure~\ref{TEMlowres}d. The resulting
micrographs are shown in Figures~\ref{TEMdark}b, c, and d
respectively. Panels a and b of this figure show the same region of
the (40/14) multilayer. The brightness of the spacer layers in this
dark-field image demonstrates that the (001) reflection does indeed
come from the spacer layer and is not the forbidden (001) spot of bcc
Fe. Figures~\ref{TEMdark}c and d also show the same region (although
a different region than panels a and b). The bright areas in these
two images are the complement of one another; where one is bright, the
other is dark and vice versa. The dark-field images in panels c and d
of Figure~\ref{TEMdark} demonstrate convincingly that the orientation
of the film evolves from predominantly (011) to predominantly (002) as
the thickness increases. The reason for the change in orientation
with film thickness is not obvious; it may be related to the
bilayer-period-number dependence discussed in
Section~\ref{bilayersec}.
The effect of varying the Fe thickness has also been studied.
Magnetic properties for films with 20\AA\ $\le$ t$\rm _{Fe}$ $\le$
50\AA\ are found to change only slightly in keeping with the expected
inverse proportionality of the saturation field with t$\rm
_{Fe}$.\cite{demokritov} SAXS peaks tend to broaden and even split
with increasing Fe thickness, indicating increased disorder in the
layering. The splitting of these peaks may indicate different bilayer
periods in areas of the film with the (011) and (001) textures. When
the Fe is made less than 20\AA\ thick, the Fe high-angle diffraction
peaks disappear and so does the AF coupling. The disappearance of
crystalline Fe peaks near t$\rm _{Fe}$ = 20\AA\ is consistent with
previous results on evaporated Fe/Si multilayers.\cite{dufour} Thus
poor crystallinity of the Fe layers appears to suppress the interlayer
coupling even when the Si thickness is favorable. The lack of AF
coupling in films with poorly crystalline Fe may be related to the
lack of a template for the crystalline iron silicide spacer to grow
on.
\subsection{Dependence of Properties on Growth Temperature and Post-Growth
Annealing}
\label{tempsec}
Depositing the multilayers at different substrate temperatures is an
obvious way of influencing the composition and crystallinity of the
spacer layer phase in the Fe/Si multilayers. Fullerton has suggested
that the interlayer of Fe/Si multilayers is improved by
high-temperature growth.\cite{fullerton5} We have grown films on glass
substrates at various temperatures between -150 and +200$^{\circ}C$.
The effect of substrate temperature on the interlayer coupling of the
films is illustrated in Figure~\ref{magtemp}, where magnetization
curves for three (Fe40\AA/Si14\AA)x40 multilayers grown at
-150$^{\circ}$C, +60$^{\circ}$C (nominal RT) and +200$^{\circ}$C are
shown. The data show that as the substrate temperature increases the
saturation field increases indicating larger AF coupling. The
saturation magnetization also decreases, suggesting a larger degree of
interdiffusion in the films grown at higher temperatures.
The suspicion that more interdiffusion occurs at higher substrate
temperatures is confirmed by examination of the SAXS spectra for the
three films, shown in Figure~\ref{lowxrdtemp}. The film grown at
reduced temperature has 7 peaks while the film grown at nominal RT has
5 and the film grown at +200$^{\circ}$C has only 4. Quantitative
modelling of low-angle x-ray data has shown that the suppression of
higher-order peaks may be due to either interdiffusion or cumulative
roughness.\cite{fullerton3,payne} Certainly larger cumulative
roughness could also occur at higher growth temperatures, but one
would expect very rough growth to suppress AF coupling due to an
increased number of pinholes and larger magnetostatic interlayer
coupling.\cite{altbir} Since higher growth temperatures seem to
enhance rather than suppress the coupling, it seems more likely that
high substrate temperatures are promoting interdiffusion rather than
roughness. Studies of Mo/Si multilayers showed that a growth
temperature of 150$^{\circ}$C gives maximum SAXS reflectivity, which
the authors attribute to greater interface smoothness than for RT
deposition.\cite{stock} Smaller bilayer periods in multilayers grown
at higher temperatures support the claim of increased interdiffusion.
Fitting Eqn.~\ref{lowbragg} to peak positions from
Figure~\ref{lowxrdtemp} gives $\Lambda$ = 52.7, 49.3, and 43.8 \AA\
respectively for the -150$^{\circ}$, +60$^{\circ}$, and +200$^{\circ}$
multilayers versus the nominal value of 54\AA.
Higher substrate temperatures may also promote ordering of the Fe and
Si atoms in the crystalline spacer layer. In the fully ordered B2
phase, the Fe and Si atoms sit on different simple cubic sublattices.
The sublattice order can occur irrespective of whether or not the Fe
to Si ratio is 1:1. It is interesting to speculate whether the AF
coupling is dependent on the degree of ordering in the spacer layer.
An ordering-dependent coupling seems plausible in light of the
Fermi-surface theories of coupling in metal/metal
multilayers.\cite{bruno2,stiles} A well-ordered B2 or DO$\rm _3$ phase
would have more well-defined Fermi surface features than a random
solid solution. Unfortunately the (001) silicide peak has only been
observed by TEM, making experimental attempts to address this issue
difficult. Further studies with x-ray diffraction and soft x-ray
fluorescence are underway.
The crystallinity of the films also varies with growth temperature.
Surprisingly, films grown at both low and high temperatures on glass
substrates always have only the (011) texture, while films grown at
nominal RT often have mixed (001) and (011) textures. The multilayers
deposited on heated and cooled substrates do differ greatly in that
those grown at low temperature have amorphous spacer layers, while
those grown at high temperatures have long crystalline coherence
lengths. The reasons for the strange temperature dependence of growth
texture are not understood, although one presumes that they have to do
with the kinetics of growth. It is not clear why the (001) texture
should appear at all, although it has also been seen in Mo/Ge
multilayers.\cite{wilson} An oscillatory dependence of film texture on
spacer layer thickness and deposition conditions has been reported for
NiFe/Cu multilayers grown by IBS.\cite{nakatani} The (001) texture has
not been reported in polycrystalline magnetron-sputtered Fe/Si
multilayers, and may be due to some peculiarity of IBS growth.
A logical extension to the growth temperature studies is to try
annealing the Fe/Si multilayers grown at lower substrate temperatures
to see if their properties evolve towards those of the multilayers
grown at higher temperatures. As far as the magnetic properties are
concerned, the answer is ``no.'' Annealing the uncoupled RT-grown
(Fe30\AA/Si20\AA)x40 and low-temperature-grown (Fe40\AA/Si14\AA)x40
multilayers at +200$^{\circ}$C for two hours had almost no effect on
their magnetic properties beyond a slight magnetic moment reduction.
A subsequent 300$^{\circ}$C anneal for two hours once more produced a
moment reduction and a decrease in coercive field in the uncoupled
multilayers. A very low coercive field for annealed Fe/Si films is
not surprising given the well-known softness of Fe-Si alloys. A
300$^{\circ}$C anneal even eliminated the interlayer exchange coupling
of a RT-grown (Fe40\AA/Si14\AA)x50 film used as a control. For this
(40/14) multilayer, the 300$^{\circ}$C anneal caused the SAXS peaks to
narrow and reduced their number from 5 to 4. At the same time the
bilayer period decreased from 49.4\AA\ to 46.0\AA. High-angle x-rays
spectra (not shown) indicated that the Fe lattice constant slightly
decreased, which is consistent with increased diffusion of Si in the
Fe layer.\cite{foiles2} These x-ray and magnetization results imply
that annealing primarily promotes interdiffusion of the Fe and
silicide layers. With sufficient interdiffusion the spacer layer may
become ferromagnetic, which would explain the suppression of
antiferromagnetic interlayer coupling. These Fe/Si multilayers show
less thermal stability than Mo/Si multilayers with comparable layer
thicknesses, which do not show changes in SAXS spectra until
400$^{\circ}$C.\cite{stock} There was no sign of the solid-state
amorphization previously observed in Fe/Si multilayers with thicker
layers.\cite{gupta}
Whatever process occurs during annealing, it does not enhance the
interlayer coupling the way that +200$^{\circ}$C growth does. This is
hardly surprising given that annealing will tend to drive the
multilayer towards its equilibrium state, presumably a mixture of
different iron silicide phases. There is no reason to think that the
crystalline Fe/Fe$_x$Si$_{1-x}$ multilayer should be an
intermediate phase during the annealing. In the future the kinetics
of Fe/Si multilayer growth at different substrate temperatures will be
investigated further by employing an ion-assist gun to improve atomic
surface mobility.
\subsection{Dependence of Properties on Number of Bilayers}
\label{bilayersec}
One puzzling aspect of the interlayer exchange coupling in the Fe/Si
system has been the dependence of its strength on the number of
bilayers in the multilayer. This trend is illustrated in
Figure~\ref{binumber}, where magnetization curves for
(Fe40\AA/Si14\AA)xN multilayers with 2, 12 and 25 repeats are
displayed. (The 2-repeat multilayer is just an Fe/Si/Fe trilayer.)
Although the trilayer has magnetic properties like bulk Fe, the
25-repeat multilayer data has a magnetization curve similar to the
40-repeat multilayer data shown above. The magnetization curve for
the 12-repeat multilayer falls in between that for the thicker and
thinner films. Evidence for AF coupling which is stronger near the
top of an Fe/Si multilayer than near the substrate has previously been
described by Fullerton {\it et al.}\cite{fullerton4} Presumably the
increase of coupling with bilayer-number is a manifestation of the
same phenomenon. The interlayer coupling in Co/Cu multilayers also
increases with the number of bilayer periods up to about 25
bilayers.\cite{rupp}
One would not expect interlayer coupling that is quantum-mechanical in
nature to be affected much by total film thickness. The unusual
thickness dependence therefore raises the question of whether there is
quantum-mechanical coupling at all, or whether some other mechanism
might determine the shape of the magnetization curves. Disordered
magnetic materials such as small amorphous Fe particles can have low
remanence and high saturation fields without any layering at all. The
magnetization curves of these Fe particles are in fact quite similar
to those of the Fe/Si multilayers.\cite{grinstaff} This resemblance
might lead to speculation that the topmost Fe layers in Fe/Si
multilayers are discontinuous and that the magnetic properties are
dominated by particle shape. However, the existence of half-order
peaks in polarized neutron reflectometry measurements in the IBS-grown
Fe/Si multilayers\cite{ankner} and the magnetron-sputtered
multilayers\cite{fullerton4} gives unambiguous evidence that the
magnetic properties are due to magnetic order rather than structural
disorder. In addition, TEM pictures such as Figure~\ref{TEMlowres}
show that the Fe layers are continuous in films with both high and low
saturation fields.
How then does the number of bilayer periods influence the AF coupling
strength? It has been suggested that the difference between thin and
thick multilayers grown at nominal RT is that the substrates of thick
multilayers have time to rise to a higher temperature (about
+60$^{\circ}$C for our system) during the longer
growth.\cite{fullerton5} This idea seems reasonable in light of the
larger coupling in samples grown on heated substrates as described
above. In order to investigate this idea, a
(Fe100\AA/Si14\AA/Fe100\AA) film was grown on glass at
+200$^{\circ}$C. The magnetization curve for this film is shown in
Figure~\ref{trilayers}. Also shown in this figure are data for a
(Fe100\AA/Si14\AA/Fe100\AA) trilayer deposited at nominal RT and for a
(Fe100\AA/Si14\AA/Fe100\AA) trilayer deposited at +200$^{\circ}$C,
both grown on a 500\AA-thick a-Si buffer. The trilayer deposited
directly on glass at elevated temperature has only slightly less
remanence and higher saturation field than the trilayer grown at RT
whose data are shown in Figure~\ref{binumber}. This result implies
that it is not substrate temperature alone which causes bilayer-number
effects. The magnetization curves of the trilayers grown on buffer
layers, on the other hand, look much more like typical t$\rm _{Fe}$ =
40\AA\ 40-repeat multilayer results. An epitaxial
(Fe100\AA/Si14\AA/Fe100\AA) trilayer grown directly on an MgO(001) at
+200$^{\circ}$C substrate also has strong AF coupling (data not
shown). Undoubtedly the strong AF coupling of the trilayer grown
directly on the MgO is due the superior surface quality of the
single-crystal substrate.
The take-away lesson from all of these results is that substrate
roughness is probably responsible for the reduced interlayer coupling
in (Fe40\AA/Si14\AA) multilayers with a low number of bilayers.
Conformal growth may propagate this roughness up from the substrate
into the multilayer. Parkin {et al.} have found that the interlayer
coupling in MBE-grown Co/Cu multilayers is very sensitive to the
substrate and the buffer layer type, perhaps due to pinholes through
the Cu layers.\cite{parkin} Presumably thin Fe layers grown directly
on glass are so wavy that pinhole and magnetostatic coupling dominate
the interlayer interactions for the first few bilayer periods. Recent
calculations show that magnetostatic effects associated with
propagating roughness can give interlayer ferromagnetic coupling of
the same order of magnitude as the coupling derived from quantum-well
effects.\cite{altbir} Ongoing polarized neutron reflectivity
experiments may give more information on the variation of the coupling
with position in the thicker multilayers.\cite{ankner}
\subsection{Growth on Single-Crystal Substrates}
That Fe films can be grown epitaxially on MgO and $\rm Al_2O_3$
substrates is well-known.\cite{metoki} One might therefore expect to
be able to grow high-quality Fe/Si superlattices on these substrates.
Figure~\ref{epihigh}a) shows high-angle x-ray diffraction spectra for
a purely (001)-oriented (Fe40\AA/Si\AA)x60 multilayer grown on
MgO(001). The spectrum in Figure~\ref{epihigh}b) is data for a highly
(011)-oriented (Fe40\AA/Si14)x46 multilayer grown on $\rm Al_2O_3$.
Both multilayers were deposited at +200$^{\circ}$C.
Figure~\ref{epihigh}c) shows a $\phi$ scan for the MgO (110) and Fe
(110) peaks for the film on the MgO substrate. These sets of peaks
are offset from one another by 45$^{\circ}$ in $\phi$, confirming the
well-known epitaxial relation Fe(001) $\|$ MgO(001) and Fe(110) $\|$
MgO(100).\cite{metoki} The $\phi$ scans for the $\rm Al_2O_3$
substrate show that this film is only weakly oriented in-plane.
Mattson {et al.} have previously grown Fe/FeSi multilayers on $\rm
Al_2O_3$, but they did not comment on the orientation of the
multilayer.\cite{mattson2} Rocking curves widths for both films are
about 1$^{\circ}$ wide, indicating a considerably smaller mosaic than
for the multilayers grown on glass. SAXS data for the multilayers on
single-crystal substrates are comparable to the data for films grown
on glass.
The films grown on MgO are the only purely (001)-textured Fe/Si
multilayers produced by IBS so far. Dekoster {et al.} have grown
epitaxial Fe/FeSi multilayers on MgO(001) by MBE, but they do not
present any x-ray diffraction data or magnetization
curves.\cite{dekoster} Magnetization curves of films grown on
single-crystal substrates (not shown) are qualitatively similar to
those grown on glass or oxidized Si substrates. The only differences
are that the saturation fields are higher for the epitaxial samples
and that magnetocrystalline anisotropy effects are observed. The
magnetocrystalline anisotropy energies of epitaxial trilayers grown on
MgO and Ge are similar to bulk Fe.\cite{michel}
The shape of the high-angle peaks plus superlattice satellites are
described by a theory due to Fullerton {\it et al.}\cite{fullerton3}
Application of this theory to the Fe/Si multilayers is difficult
because the silicide lattice constant, the thickness of the remaining
pure Fe and the thickness of the silicide spacer can be estimated only
roughly. A precise determination of the silicide lattice constant
should make a quantitative analysis of these satellite features possible.
\section{Discussion}
Fe and Si appear to be the only known transition-metal/semiconductor
combination in which the two elements interdiffuse to form a
crystalline spacer layer with coherent interfaces. The reasons why
this unusual morphology occurs in the Fe/Si system are unknown but
likely involve a high rate of Fe diffusion into a-Si and a low heat of
crystallization of the iron silicide compound. A detailed discussion
of these issues is beyond the scope of this paper.
Three different crystal structures have been proposed for the
crystalline spacer layer of the Fe/Si multilayers. The $\epsilon$
phase can be eliminated on the basis of the electron diffraction
patterns and TEM dark field images presented here. The B2 and DO$_3$
crystal structures are better lattice-matched to Fe than
$\epsilon$-FeSi or $\alpha$- and $\beta$-FeSi$_2$. The lattice
constant of the B2 phase was reported by M\"ader and coworkers to be
2.77\AA, only 3.1\% different from Fe.\cite{mader} The lattice
constant of the $\epsilon$ phase is 4.46\AA,\cite{mader} which matches
the Fe(110) plane only in the energetically unfavorable (210)
direction.\cite{mattson}
Recent conversion-electron M\"ossbauer data are interpreted in
support of the B2 crystal structure, although the possibility of the
DO$_3$ phase was not considered in that study.\cite{dekoster} It is
plausible that the B2 or DO$_3$ structures form in rapid,
far-from-equilibrium growth conditions because of their small unit
cells. Since silicon deposited at low substrate temperatures is
amorphous, the most likely scenario is the following. Silicon
deposited on a crystalline Fe layer goes down amorphous and diffuses
only slightly into the Fe. Subsequently deposited Fe atoms diffuse
rapidly into the amorphous Si, analogous to what happens during the
growth of Mo/Si multilayers.\cite{stearns,holloway} During the
diffusion of Fe into Si, crystallization of the silicide occurs,
possibly driven by the heat of mixing or by the kinetic energy of the
incident Fe atoms. Growth of the crystalline phase may proceed upward
from the lattice-matched Fe template, or downward from the atomically
bombarded film surface. If the growth of the crystalline silicide
phase proceeds downward from the film surface, one might expect to
see some crystalline silicide in the high-resolution TEM image for the
t$\rm _{Si}$ = 20\AA\ film (Figure~\ref{TEMhires}b). The lack of any
evidence for crystalline silicide in this image suggest that the
crystallization proceeds upward from the iron/silicide interface, not
downward from the film surface.
It is difficult to determine how realistic this model for growth of
the crystalline silicide is since the Fe/Fe-Si and Si-Fe/Fe interfaces
appear identical in Figure~\ref{TEMhires}b. In contrast, the Mo/Si
and Si/Mo interfaces in Mo/Si multilayers appear quite different from
one another.\cite{stearns,holloway} In the Mo/Si multilayers, an
amorphous MoSi$_2$ region appears which is thicker at the Mo/Si
interface than at the Si/Mo interface. Detailed TEM studies of
multilayers with t$\rm _{Si}$ larger than 20\AA\ may help to answer
whether amorphous silicides can occur in IBS-grown Fe/Si multilayers.
Using the B2 phase lattice constant reported by the Z\"urich
group,\cite{mader} we can estimate the expected bilayer period of a
nominal Fe/Si multilayer in which Fe atoms diffuse into the Si layer
up to a 1:1 stoichiometry. The spacing between the Fe and Fe$\rm
_x$Si$\rm _{1-x}$ layers is taken as the average of the interplanar
spacings of the two materials. The result of this rough calculation
is that an (Fe40\AA/Si14\AA) multilayer which interdiffuses up to the
1:1 stoichiometry should form a (Fe33.2\AA/FeSi16.3\AA) multilayer
with a bilayer period of 49.4\AA. The missing bilayer period
predicted from this model is 4.6\AA, in the middle of values on the
x-axis of Figure~\ref{missmom}. One can also calculate the expected
magnetic moment reduction assuming that Fe atoms in the silicide layer
have no moment and those in the Fe layer have their full moment.
Under this assumption a calculation predicts 8.2\AA\ of missing Fe
moment, slightly lower than indicated in Figure~\ref{missmom}. This
calculation neglects the possibility that some Fe atoms in the Fe
layer with Si near neighbors may have reduced magnetic moments.
In the discussion above the possibility has not been mentioned that
the missing bilayer period and magnetic moment are due to an
inaccurate thickness calibration. This explanation is contradicted by
magnetization and x-ray diffraction measurements on Fe/Ge multilayers,
where measured magnetic moments and bilayer periods are in much closer
agreement with nominal values than for Fe/Si.\cite{michel} The
improved agreement in the case of Fe/Ge multilayers suggests that
interdiffusion is less important in multilayers with Ge spacer layers
than in multilayers with Si spacers.
The main point is that the formation of the B2 silicide does
qualitatively explain the bilayer period reduction observed in the
Fe/Si multilayers. The underlying reason for the bilayer period
reduction is that the silicide which forms is denser than both Fe and
Si. This situation is similar to that observed in other metal/Si
multilayers\cite{fullerton2,holloway} except that in the other
multilayers the silicide remains amorphous.
Confirmation that the spacer layer phase has the B2 or DO$\rm _3$
structure is important for understanding the coupling mechanism in
these compounds. Both the B2 and DO$\rm _3$ phases are known to be
metallic for some ranges of composition.\cite{vonkanel,kudrnovsky}
Thus the present results and those of other
workers\cite{fullerton,dekoster} suggest that Fe/Si is really a
metal/metal multilayer. The origin of the interlayer coupling is then
likely to be described by the same theories as describe coupling in
Co/Cu and Fe/Cr multilayers.\cite{bruno2,stiles} Fe/Si multilayers may
therefore not be a good test case for theories which model interlayer
exchange coupling across insulators.\cite{bruno,singh}
In the discussion above the possibility has been neglected that the
amorphous spacer layer in the thick-Si films may also be metallic. If
both the thick amorphous spacers and the thin crystalline spacers are
metallic silicides, then it must be the crystallinity that is the
essential feature for the existence of AF interlayer coupling. Up to
now there have been no reports of AF coupling across amorphous
metallic spacer layers. Toscano {\rm et al.} have reported AF
coupling across amorphous silicon spacer layers.\cite{toscano} These
Fe/a-Si/Fe trilayers were prepared at low temperature so as to
suppress interdiffusion.\cite{toscano} The character of AF coupling in
the a-Si spacer trilayers is likely quite different than in the
multilayers described in this study, where substrate heating increases
the strength of coupling.
At the moment there is no direct evidence regarding the metallic or
insulating nature of the amorphous spacer layers found in the
(Fe30\AA/Si20\AA) multilayers. Temperature-dependent current-in-plane
resistivity measurements suggest that both crystalline and amorphous
spacer layers in Fe/Si multilayers are poorly conducting.\cite{michel}
Fe$_{70}$Si$_{30}$ and Fe$_{65}$Si$_{35}$ amorphous alloys have a
temperature-independent resistivity, suggesting non-metallic
behavior.\cite{marchal} Overall the evidence suggests that the
amorphous spacer layers in (Fe30\AA/Si20\AA) multilayers are not
metallic, but spectroscopic measurements like soft x-ray
fluorescence\cite{carlisle} are needed for confirmation. The
interesting question as to whether there can be AF interlayer coupling
across an amorphous metal spacer layer must then be left for another
study.
\section{Conclusions}
An extensive study of the growth of Fe/Si multilayers by ion-beam
sputtering has been performed. The crystalline quality of the films
is better when they are grown with thick Fe layers, with thin Si
layers, at high temperature, and on single-crystal substrates.
Improved growth conditions lead to higher saturation fields and lower
remanence in magnetization curves. Measured bilayer periods are
consistently shorter in these multilayers than the nominal value,
suggesting formation of a dense silicide phase in the spacer layer.
Despite considerable interdiffusion in the multilayers, a strong
composition modulation along the growth direction is maintained as
evidenced by SAXS measurements.
There are two surprising results from this study. One is that the
films grow on glass with a mixed (011) and (001) texture near nominal
RT and with a pure (011) texture at higher and lower temperatures.
The other surprise is that the strength of the interlayer coupling
depends strongly on the number of bilayer periods in films with thin
Fe layers. This latter result is explained on the basis of substrate
surface roughness.
Unraveling the behavior of the Fe/Si multilayer system has proven to
be a considerably more complex task than understanding the Fe/Cr or
Co/Cu multilayer systems. The reason is that compound formation at
the Fe/Si interface is crucial to understanding the AF interlayer
coupling. Identification of possibly disordered phases in the spacer
layer of a multilayer continues to be an experimental challenge.
Mounting evidence suggests that the spacer layer in the AF-coupled
Fe/Si multilayers is metallic and crystalline and that the Fe/Si
interlayer coupling therefore has the same origin as in metal/metal
multilayers.
\bigskip
\paragraph*{Acknowledgements}
\noindent We would like to thank P.E.A. Turchi, T.W. Barbee Jr., T.P. Weihs,
E.E. Fullerton, Y. Huai and E.C. Honea for helpful discussions, and
B.H. O'Dell and S. Torres for technical assistance. Further thanks go
to C.-T. Wang of Stanford for the four-circle x-ray diffractometry and to
Sandia National Lab for use of their electron microscope for HREM
work. Part of this work was performed under the auspices of the
U.S. Department of Energy by LLNL under contract No. W-7405-ENG-48.
\clearpage
|
\section{Introduction}
Recently the CLEO collaboration has observed\cite{cleo1} the
exclusive radiative decay $B \rightarrow K^* \gamma $ with a
branching fraction
of $BR(B\rightarrow K^* \gamma)=(4.5 \pm1.0 \pm0.9)\times10^{-5}$.
The inclusive $b\to s \gamma$ branching ratio measured by CLEO\cite{cleo2}
is:
\begin{eqnarray}
BR(B\to X_s \gamma ) = (2.32\pm 0.57 \pm 0.35) \times 10^{-4}.
\end{eqnarray}
The newest upper and lower limits of this decay branching ratio are
\begin{eqnarray}
1.0\times 10^{-4} < BR(B \to X_s \gamma) < 4.2\times 10^{-4},
\ \ at \ \ 95\%C.L..
\end{eqnarray}
As a loop-induced flavor changing neutral current(FCNC)
process the inclusive decay(at quark level) $b\to s \gamma$ is
in particular sensitive to contributions
from those new physics beyond the Standard Model(SM)\cite{Hew1}.
There is a vast interest in this decay.
The decay $b\to s \gamma$ and its large leading log QCD corrections
have been evaluated in the SM by several groups
\cite{Ciu}. The reliability of the calculations of this
decay is improving as partial calculations of the next-to-leading
logarithmic QCD corrections to the effective Hamiltonian\cite{AJB,lcd}.
On the other hand the discovery of the top quark and the measurement
of its mass (in this paper we use the
weighted average $m_t=180\pm 12\; GeV$ from the announced
results of $m_t$ by CDF and D0\cite{cdfd0} wherever possible)
at FERMILAB basically eliminated a source of uncertainties for the
calculation of the decay $b\to s \gamma$ in the SM and in beyond theories.
The great progress in theoretical studies and in experiments achieved
recently encourage us to do more investigations about this decay in
Technicolor theories.
In this paper, we estimate the possible contributions
to the decay $b\to s \gamma$
from the exchange of the charged Pseudo-Goldstone bosons which will
appear in no-minimal Technicolor models, such as the Farhi-Susskind
one-generation Technicolor model (OGTM) \cite{Farhi}.
We know that the experimental data seems
disfavor the OGTM which generally tend to predict $S$ parameter
large and positive \cite{Peskin92}. Why we here
still choose it to do the calculations? The reasons are the following:
(1) At first, presence of the Pseudo-Goldstone bosons in the particle spectrum
is a common feature of those non-minimal TC models with ordinary or
novel ETC sectors, no matter the specific differences of structures
between those models.
The gauge couplings of the PGBs are determined by their quantum numbers,
while the Yukawa couplings of PGBs to ordinary fermions
are generally proportional to fermion masses
for many TC/ETC models.
Among the non-minimal TC models,
the OGTM \cite{Farhi} is the simplest and most frequently studied
model. Many relevant works
\cite{King95} have been done since the late 1970's.
One can use those existed results directly in further investigations.
(2) On the other hand, the constraints on the $S$ parameter could be relaxed
considerably by introducing three additional parameters
$(V, W, X)$ \cite{Burgess94a}.
A global fit to the data in which
all six oblique parameters S through X are allowed to vary
simultaneously gives the one standard deviation bound on $S$:
$ S \sim -0.93 \pm 1.7$ \cite{Burgess94b}.
This fact means that the constraint on the OGTM
from the parameter $S$ could be considerably weakened if we
consider the effects from light technifermions and light
PGBs \cite{Evans94}.
In this paper, we estimate the possible
contributions to the rare decay $b\to s \gamma$ from the charged PGBs in the
framework
of the OGTM. At least, one can
regard our results as an estimation for the ``correct'' output of
the future ``realistic'' TC models.
This paper is organized as the following: In Section 2, we present the
basic ingredients of the OGTM and then calculate the PGB contributions
to $b\to s \gamma $ decay, together with the full leading log QCD
corrections. In Section 3, we obtain the branching ratios of this decay,
and derive out the
constraints on masses of charged PGBs by phenomenological analysis.
The conclusions are also included in this
section.
\section{Charged PGBs and QCD Corrections to $b\to s \gamma$}
In the OGTM \cite{Farhi}, when the technifermion condensate
$<\overline{T}T>\neq 0$
was formed, the global flavor symmetry will break as follows:
$SU(8)_L \times SU(8)_R \rightarrow SU(8)_{L+R}$.
Consequently, 63 (Pseudo)-Goldstone bosons will
be produced from this breaking.
When all other interactions but the Technicolor are turned off, these 63
Goldstone bosons are exactly massless. Three of them are eaten by the
$W^{\pm}$ and $Z^0$ gauge bosons. The others acquire masses
when one turned on the
gauge interactions, and therefore they are Pseudo-Goldstone Bosons(PGBs).
According to previous studies, the phenomenology of those color-singlet
charged PGBs in the OGTM is very similar with that of the elementary
charged Higgs bosons $H^{\pm }$ of Type-I Two-Higgs-Doublet
Model(2HDM) \cite{Gunion}. And consequently, the contributions to the decay
$b\rightarrow s\gamma$ from the color-singlet
charged PGBs in the OGTM will be very similar with that from
charged Higgs bosons in the 2HDM.
As for the color-octet charged PGBs, the situation is more complicated
because of the involvement of the color interactions.
Other neutral PGB's don't contribute to the rare decay
$b\to s \gamma$.
The gauge couplings of the PGBs are determined by their quantum numbers.
The Yukawa couplings of PGBs to ordinary fermions are induced by ETC
interactions and hence are model dependent. However, these Yukawa couplings
are generally proportional to fermion masses with small differences in the
magnitude of the coefficients for different TC/ETC models.
The relevant couplings needed in our
calculation are directly quoted from refs.\cite{Ellis,Chivukula95,Xiao94}
and summarized in Table \ref{table1},
where the $V_{ud} $ is the corresponding element of Kobayashi-Maskawa
matrix. For the OGTM, the Goldstone boson decay constant
$F_\pi$ in Table \ref{table1}
should be $F_{\pi}=v/2=123\;GeV$,
in order to ensure the correct masses for the gauge bosons
$Z^0$ and $W^{\pm}$ \cite{King95}.
\begin{table}[htbp]
\caption{ The relevant gauge couplings and Effective Yukawa couplings for
the OGTM. }\label{table1}
\begin{center}
\vspace{.1cm}
\begin{tabular}{|c|c|} \hline
$P^+ P^- \gamma_\mu$ & $-ie(p_+ - p_-)_\mu$ \\ \hline
$P^+_{8a} P^-_{8b} \gamma_\mu$ & $-ie(p_+ - p_-)_\mu \delta_{ab}$ \\ \hline
$P^+\; u\; d$ & $i\frac{V_{ud}}{2 F_\pi}\sqrt{\frac{2}{3}}
[M_u (1-\gamma_5) - M_d (1 + \gamma_5) ]$ \\ \hline
$P^+_{8a}\; u\; d$ & $i\frac{V_{ud}}{2 F_\pi} 2 \lambda_a
[M_u (1-\gamma_5) - M_d (1 + \gamma_5) ]$ \\ \hline
$P^+_{8a} P^-_{8b} g_{c\mu}$ & $-g f_{abc}(p_a - p_b)_\mu $ \\ \hline
\end{tabular}
\end{center}
\end{table}
In ref.\cite{Randall}, Randall and Sundrum have estimated the contributions to
$b\to s \gamma$ from the exchange of ETC gauge bosons in various
ETC scenarios. In the case of ``traditional'' ETC (just the case which
will be studied here), the dominant contribution to $b\to s \gamma$
occurs when the ETC gauge boson is exchanged between purely left-handed
doublets and when the photon is emitted from the
technifermion line. But the resulted ETC contribution is strongly
suppressed with respect to the SM by a factor of $m_t/(4\pi v)< 0.09$ for
$m_t< 200\;GeV$\cite{Randall}. In short, the ETC contribution to the decay
$b\rightarrow s\gamma$ is small and will be masked by still large
experimental and theoretical uncertainties. We therefore can
neglect the ETC Contributions to $b\to s \gamma$ at present
phenomenological analysis.
In Fig.1, we draw the relevant Feynman diagrams which contribute to the decay
$b\to s \gamma$, where the half-circle lines represent the W
gauge boson of SM as well as the charged PGBs $P^{\pm}$ and
$P_8^{\pm}$ of OGTM.
In the evaluation we at first integrate
out the top quark and the weak W bosons at $\mu=M_W$ scale,
generating an effective five-quark theory. By using the
renormalization group equation, we run the effective field theory
down to b-quark scale to give the leading log QCD corrections,
then at this scale, we calculate the rate of radiative $b$ decay.
After applying the full QCD equations of motion\cite{eom}, a complete
set of dimension-6 operators relevant for $b\to s \gamma $ decay can be
chosen to be:
\begin{eqnarray}
O_1&=&(\overline{c}_{L\beta} \gamma^{\mu} b_{L\alpha})
(\overline{s}_{L\alpha} \gamma_{\mu} c_{L\beta})\;,\\
O_2&=&(\overline{c}_{L\alpha} \gamma^{\mu} b_{L\alpha})
(\overline{s}_{L\beta} \gamma_{\mu} c_{L\beta})\;,\\
O_3&=&(\overline{s}_{L\alpha} \gamma^{\mu} b_{L\alpha})
\sum_{q=u,d,s,c,b}(\overline{q}_{L\beta} \gamma_{\mu} q_{L\beta})\;,\\
O_4&=&(\overline{s}_{L\alpha} \gamma^{\mu} b_{L\beta})
\sum_{q=u,d,s,c,b}(\overline{q}_{L\beta} \gamma_{\mu} q_{L\alpha})\;,\\
O_5&=&(\overline{s}_{L\alpha} \gamma^{\mu} b_{L\alpha})
\sum_{q=u,d,s,c,b}(\overline{q}_{R\beta} \gamma_{\mu} q_{R\beta})\;,\\
O_6&=&(\overline{s}_{L\alpha} \gamma^{\mu} b_{L\beta})
\sum_{q=u,d,s,c,b}(\overline{q}_{R\beta} \gamma_{\mu} q_{R\alpha})\;,\\
O_7&=&(e/16\pi^2) m_b \overline{s}_L \sigma^{\mu\nu}
b_{R} F_{\mu\nu}\;,\\
O_8&=&(g/16\pi^2) m_b \overline{s}_{L} \sigma^{\mu\nu}
T^a b_{R} G_{\mu\nu}^a\;.
\end{eqnarray}
The effective Hamiltonian appears just below the W-scale is given as
\begin{equation}
{\cal H}_{eff} =\frac{4G_F}{\sqrt{2}} V_{tb}V_{ts}^*
\displaystyle{\sum_{i=1}^{8} }C_i (M_W^-) O_i(M_W^-).
\end{equation}
The coefficients of 8 operators are calculated from diagrams of Fig.1:
\begin{eqnarray}
C_i(M_W)&=&0, \;\; i=1,3,4,5,6, \;\;\; C_2(M_W)=-1,\\
C_7(M_W)&=& A(\delta) +\frac{B(x)}{3\sqrt{2}G_F F_{\pi}^2 }
+\frac{8 B(y)}{3\sqrt{2}G_F F_{\pi}^2 },
\label{c7}\\
C_8(M_W)&=& C(\delta)
+\frac{D(x)}{3\sqrt{2}G_F F_{\pi}^2 }
+\frac{8 D(y) + E(y)}{3\sqrt{2}G_F F_{\pi}^2 },\label{c8}
\end{eqnarray}
with $\delta=M_W^2/m_t^2$, $x=(m(P^{\pm})/m_t)^2$ and
$y=(m(P^{\pm}_8)/m_t)^2$.
The functions $A$ and $C$ arise from graphs with W boson exchange
are already known contributions from SM; while the functions $B$, $D$,
and $E$ arise from diagrams with color-singlet and color-octet charged
PGBs of OGTM. They are given by,
\begin{eqnarray}
A(\delta)&=& \frac{ \frac{1}{3} +\frac{5}{24} \delta -\frac{7}{24}
\delta^2}{(1-\delta)^3}
+\frac{ \frac{3}{4}\delta -\frac{1}{2}\delta^2}{(1-\delta)^4}
\log[\delta] \\
B(y)& =& \frac{ -\frac{11}{36}
+\frac{53}{72}y -\frac{25}{72}y^2}{(1-y)^3}
+\frac{ -\frac{1}{4}y +\frac{2}{3}y^2 -\frac{1}{3}y^3}
{(1-y)^4}\log[y],\\
C(\delta)&=& \frac{\frac{1}{8} -\frac{5}{8} \delta-\frac{1}{4}
\delta^2}{(1-\delta)^3}
-\frac{ \frac{3}{4}\delta^2}{(1-\delta)^4} \log[\delta] \\
D(y)& =& \frac{ -\frac{5}{24}
+\frac{19}{24}y -\frac{5}{6}y^2}{(1-y)^3}
+\frac{ \frac{1}{4}y^2 -\frac{1}{2}y^3}{(1-y)^4}
\log[y],\\
E(y) & =& \frac{ \frac{3}{2}-\frac{15}{8}y -\frac{5}{8}y^2 }{(1-y)^3}
+\frac{\frac{9}{4}y -\frac{9}{2}y^2}{(1-y)^4 }\log[y].
\end{eqnarray}
It is shown from these expressions that, for $\delta < 1$, $x,y >> 1$,
the OGTM contribution $B$, $D$ and $E$ have always a relative minus
sign with the SM contribution $A$ and $C$. As a result, the OGTM
contribution always destructively interferes with the SM contribution.
This can also be seen from the numerical results and discussion
in the next section.
The running of the coefficients of operators from $\mu=M_W$ to $\mu=m_b$
was well described in refs.\cite{Ciu}. After renormalization
group running we have the QCD corrected
coefficients of operators at $\mu=m_b$ scale.
\begin{equation}
C_7^{eff}(m_b) = \eta^{16/23}C_7(M_W) +\frac{8}{3}
( \eta^{14/23}-\eta^{16/23} ) C_8(M_W)
+C_2(M_W) \displaystyle \sum _{i=1}^{8} h_i \eta^{a_i}.
\end{equation}
With $\eta = \alpha_s(M_W) /\alpha_s (m_b)$,
$$ h_i =\left( \frac{626126}{272277}, -\frac{56281}{51730},
-\frac{3}{7}, -\frac{1}{14}, -0.6494, -0.0380, -0.0186, -0.0057 \right),$$
$$a_i = \left( \frac{14}{23}, \frac{16}{23}, \frac{6}{23}, -\frac{12}{23},
0.4086, -0.4230, -0.8994, 0.1456 \right).$$
\section{The $B \rightarrow X_s \gamma$ decay rate and phenomenology}
Following refs.\cite{Ciu}, applying a spectator model,
\begin{eqnarray}
BR(B \rightarrow X_s \gamma) /BR(B\rightarrow X_c e\overline{\nu})
\simeq\Gamma(b\rightarrow s\gamma)/\Gamma
(b\rightarrow ce\overline{\nu}).
\end{eqnarray}
Then when have
\begin{eqnarray}
\frac{BR(B \rightarrow X_s \gamma)}{BR(B
\rightarrow X_c e \overline{\nu})} \simeq \frac{|V_{tb} V_{ts}^*|^2}
{|V_{cb}|^2} \frac{6 \alpha_{QED}}{\pi
g (m_c/m_b)}
|C_7^{eff}(m_b)|^2 \left(1-\frac{2 \alpha_{s}(m_b)}{3 \pi} f(m_c/m_b)
\right)^{-1},
\end{eqnarray}
where the phase space factor $g(z)$ is given by:
\begin{eqnarray}
g(z)=1-8z^2+8z^6-z^8-24z^4\log z,
\end{eqnarray}
and the factor $f(m_c/m_b)$ of one-loop QCD correction to the semileptonic
decay is,
\begin{eqnarray}
f(m_c/m_b)=(\pi^2 -31/4)(1-m_c^2/m_b^2) + 3/2.
\end{eqnarray}
Afterwards one obtains the $B \rightarrow X_s
\gamma$ decay rate normalized to the quite well established
semileptonic decay rate $Br(B \to X_c e\overline{\nu} )$.
If we take experimental result $BR(B \to
X_c e\overline{\nu} ) =10.8\% $\cite{data}, the branching ratios of
$B \to X_s \gamma$ is found to be:
\begin{eqnarray}
BR(B \rightarrow X_s \gamma)
\simeq 10.8\%\times \frac{|V_{tb} V_{ts}^*|^2}
{|V_{cb}|^2} \frac{6 \alpha_{QED}\;|C_7^{eff}(m_b)|^2}
{\pi g (m_c/m_b)}
\left(1-\frac{2 \alpha_{s}(m_b)}{3 \pi} f(m_c/m_b)
\right)^{-1}.
\end{eqnarray}
In numerical calculations we always use
$M_W=80.22\;GeV$, $\alpha_s(m_Z)=0.117$,
$m_c=1.5\;GeV$, $m_b=4.8\;GeV$ and $|V_{tb} V_{ts}^*|^2/
|V_{cb}|^2= 0.95$ \cite{data} as input parameters.
Generally speaking, the contribution to the decay $b\to s \gamma$ from
color singlet $P^\pm$ is small
when compared with the contribution from the color octet $P_8^\pm$,
since there is a color enhancement factor $8$ appeared in the third terms
in eqs.(\ref{c7}, \ref{c8}) for the functions $B(y)$ and $D(y)$.
Fig.2 is the plot of the branching ratio $Br(B \rightarrow X_s \gamma)$
as a function of the top quark mass.
The upper dashed curve in Fig.2 represents
the branching ratio in the standard model, while the solid curve shows
the same ratio with the inclusion of the contributions from
$P^{\pm}$ and $P_8^{\pm}$ assuming $m(P^\pm) =300\;GeV$ and
$m(P^\pm_8) =600\;GeV$. The band between two dash-dotted lines
corresponds to the newest CLEO limits: $ 1.0\times 10^{-4} <
BR(B \to X_s \gamma) < 4.2\times 10^{-4}$ at $95\%C.L.$ \cite{cleo2}.
The branching ratio of $b \to s\gamma$ with large contribution
from OGTM, is much more sensitive with the top quark mass, compared with
the case of pure SM.
It is known from the decoupling theorem that for heavy enough
nonstandard boson, we should recover the SM result. So for
sufficiently large values of $m(P^{\pm})$, $m(P^{\pm}_8)$ (
e.g. $m(P^{\pm})>600$GeV, $m(P^{\pm}_8)>2000$GeV), the
contributions from OGTM shall be negligible. This can also be seen
from the fact that the functions $B$, $D$ and $E$ go to zero, as
$x$,$y\to \infty$. For
not so large $m(P^{\pm})$, $m(P^{\pm}_8)$, the OGTM contribution cancels much
of the SM contribution because of the relative minus sign between
their contribution. As a result, the branching of $b\to s\gamma$
reached the lower limit of the CLEO experiment. So a large region of
$m(P^{\pm})$, $m(P^{\pm}_8)$ ( i.e. $1000GeV <m(P^{\pm}_8)<2000GeV$,
for all
$m(P^{\pm})$ ) is ruled out. When $m(P^{\pm})$, $m(P^{\pm}_8)$ go on
smaller, their contribution is about two times as large as contribution
of SM (recall there is a relative minus sign), the branching ratio
of $b \to s\gamma$ resumes to experiment allowed region. But if the
$m(P^{\pm})$, $m(P^{\pm}_8)$ are smaller enough, the contribution
of OGTM is more larger, the region is also excluded by the upper limit
of CLEO experiment. The whole result is
illustrated at Fig.3, large part of the parameter space in the
$m(P^\pm)-m(P_8^\pm)$ plane can be excluded according to the current $CLEO$
95\% C.L. limits on the ratio $BR(B \rightarrow X_s \gamma)$ \cite{cleo2}.
It is easy
to see that no direct limits on $m(P^\pm)$ can be obtained at present
for free $m(P^{\pm}_8)$,
but at the same time, one can simply read out the lower
bound on the mass of color octet PGBs:
$m(P^{\pm}_8) > 440\;GeV$ for free $m(P^{\pm})$
(assuming $m_t=180\;GeV$), if we simply interpret the $CLEO$
$95\% C.L.$ limits on the ratio $BR(B \rightarrow X_s \gamma)$
as the bounds on the masses of charged PGBs.
Of cause, we have not considered the effects of other possible
uncertainties, such as that of $\alpha_s(m_Z)$, next-to-leading-log
QCD contribution\cite{AJB}, QCD correction from $m_{top}$ to $M_W$\cite{lcd}
etc. The inclusion of those
additional uncertainties will broaden the border lines between the
allowed regions and excluded regions in Fig.3. The limitations drawn
from the calculations will be surely weaken, i.e., the lower
limit will become $m(P^{\pm}_8) > 400\; GeV$ at $90\%C.L.$ if we include
an additional $20\%$ theoretical uncertainties.
As a conclusion, the size of contribution to the rare decay of $b\to s \gamma$
from the PGBs strongly depends
on the values of the masses of the top quark and the charged PGBs. This is
quite different from the SM case. By the comparison
of the theoretical prediction with the current data one can derives out
the constraints on the masses of the color octet charged PGBs:
$m(P^\pm_8) > 400\;GeV$ at $90\%C.L.$ for free $m(P^\pm)$, assuming
$m_t=180\;GeV$.
\vspace{1cm}
\noindent {\bf ACKNOWLEDGMENT}
This work was supported in part by the National Natural Science
Foundation of China, the funds from
Henan Science and Technology Committee and China Postdoctoral Science
Foundation.
\vspace{1cm}
|
\section{Introduction}
The option of
building a photon-photon interaction region
at an $e^{+} e^{-}\;$ linear collider is now taken seriously under consideration.
Based on
the idea of using laser-induced backscattered photons for
inducing high-energy photon collisions,
a $\gamma \gamma\;$ collider (PLC)
gives rise to new physics opportunity\cite{GinzburgNIM}.
The issues concerning electroweak physics
will be summarized in this talk\cite{Paris}.
Since the symmetry-breaking mechanism remains the last open
question in the standard model,
an important part of the planning at any future
collider must be devoted to that.
This obviously includes searches for the Higgs particle
and the determination of its properties.
One unique opportunity for $\gamma \gamma\;$ colliders in that respect is the
direct measurement of the $H\gamma \gamma$ coupling.
Should the Higgs searches remain fruitless, the
study
of the longitudinal $W$ sector would give
a handle on the symmetry breaking mechanism. Photon
colliders, being essentially a $W$ pair
factory, could make a useful
contribution in that respect.
Before going into the heart of the subject
and to give a first idea of the possibilities of
$\gamma \gamma\;$ colliders, I will present the main characteristics of
high-energy $\gamma \gamma\;$ collisions:
\begin{itemize}
\item
Any elementary charged particle, phase-space allowing,
can be produced in $\gamma \gamma\;$ collisions with a model-independent
predictable cross-section.
\item
$\gamma \gamma\;$ gives access to the $J_Z=0$ channel, which is chirality suppressed
in $e^{+} e^{-}\;$. To test the electroweak symmetry breaking (ESB) mechanism,
this means producing the Higgs as a resonance.
\item
$\gamma \gamma\;$ collisions feature very large cross-sections, which are
always larger than in $e^{+} e^{-}\;$ for the
same energy and luminosity.
\end{itemize}
But all is not so bright, for $\gamma \gamma\;$ colliders also have shortcomings.
First,
the Higgs resonance cannot be so prominent as the Z at LEP since the coupling
of
neutral scalars to two photons only occurs at the loop
level and is suppressed by a factor $\alpha$.
Second, $\gamma \gamma\;$ does not have the same energy as $e^{+} e^{-}\;$
(the maximum energy varies between 80-90\%) and the useful luminosity
can be smaller than in $e^{+} e^{-}\;$. The latter is true especially if
one uses beams optimized for $e^{+} e^{-}\;$ coliders rather then designed
specifically for $\gamma \gamma\;$\cite{Telnovshef}.
Finally, the photon collider is not monochromatic,
although, as was discussed by Telnov,
one can tune the parameters of the
laser such as to have a nearly monochromatic
spectrum near the maximum energy. This is done at the expense
of a drop in luminosity.
Certainly, one has a great
flexibility in choosing the energy, the spectrum and
the polarization of the beams. The
choice of spectrum will be dictated by the physics one is interested in.
For example, a ``peaked" spectrum where much the luminosity is
concentrated over a narrow
energy band would be most appropriate to study
a resonance. A ``broad" spectrum, one
with sizeable luminosity over a wide
energy range, would correspond to a multi-purpose
machine useful for many processes\cite{Paris}.
Whatever the spectrum used, a precise knowledge of it is
essential to be able to do precision measurements.
More efforts in that direction are needed since
recent studies have shown that going beyond the ideal spectrum of
Ginzburg et al.\cite{GinzburgNIM}
could significantly affect the region where the photons carry only a
small fraction of the initial beam energy.
This region is particularly sensitive to
multiple scattering and nonlinear effects\cite{Chenshef}.
For lack of a more realistic spectrum, most of the results presented
here use the ideal one.
\section{Typical electroweak cross-sections}
\renewcommand{\thesection.\arabic{equation}}{\Alph{section}-\arabic{equation}}
\setcounter{equation}{0}
A comparison of a few characteristic
cross-sections in $\gamma \gamma\;$ with the corresponding same final-state
processes at $e^{+} e^{-}\;$ clearly show the advantage of the former.
Indeed, independent of the spin of the particle
and at the same centre-of-mass
energy, $\gamma \gamma\;$-initiated processes are, at high enough energy, about an order of
magnitude larger than the corresponding $e^{+} e^{-}\;$ reactions
(Fig.~1).
Even the reaction
$\gamma \gamma \rightarrow ZZ$ \cite{Jikiazz,DicusKao},
which is purely a loop effect, rapidly overtakes the
corresponding tree-level $e^{+} e^{-}\;$ process. This is due
to the rescattering effect $\gamma \gamma \rightarrow W^+W^- \rightarrow ZZ$.
Vector-boson production
dominates in $\gamma \gamma\;$ collisions due to the t-channel spin-1 exchange. Most
prominent is the $W$ pair cross-section, which very quickly reaches a plateau
of almost $90$pb.
This process is so important that it triggers a host of higher order
processes like
triple vector production ($\approx 1$pb), 4 vector production
($\approx 100 fb$) or $H$ production via $WWH$ \cite{nousgg3v}.
\begin{figure*}[htb]
\caption{\label{allgamegam}
{\em Typical sizes of non-hadronic $\gamma \gamma\;$, $e\gamma$ and
$e^{+} e^{-}\;$ processes.
The subscripts in Higgs(top) processes refer to the mass of the
Higgs(top). }}
\begin{center}
\vspace*{.5cm}
\vspace*{-1.cm}
\end{center}
\end{figure*}
Large cross-sections are great, but more
is required to do interesting physics.
A drawback of
$\gamma \gamma\;$ processes, especially as regards ESB
tests is the small fraction of longitudinal
vs transverse $W$'s.
While there is a large sample of
$W_LW_L$ (in fact more then 5 times than in $e^+e^-$),
the extraction of these longitudinals from the transverse background is
an
arduous task. Since most transverse W's are produced quite forward, imposing
angular cuts improves the situation significantly (see
Fig.~2). Nevertheless, the ratio LL/TT remains higher in $e^{+} e^{-}\;$.
The problem just alluded to is in fact characteristic
of all processes that will be discussed for ESB tests in
$\gamma \gamma\;$: the extraction of a signal, usually in the longitudinal sector, from
a large transverse background.
\begin{figure*}[hbt]
\vspace*{-.5cm}
\begin{center}
\caption{\label{eeggwwfig2}{\em Comparing the total $WW$ cross-sections
and the longitudinal $W_L W_L$ in $e^{+} e^{-}\;$ {\it vs} $\gamma \gamma\;$.
}}
\vspace*{-1.cm}
\end{center}
\end{figure*}
\section{Electroweak symmetry breaking}
The various options for symmetry breaking can be divided into two
classes:
light Higgs ($m_H\leq 800 GeV$)
or
no Higgs (for the discussion here this is equivalent to a
heavy Higgs).
The implications for electroweak physics differ markedly
according to the option one is willing to consider.
A light Higgs probably means the existence of supersymmetry unless
one is ready to give up the naturality argument raised to avoid the
large fine-tuning necessary for the elementary
scalar to remain light to all orders.
In this option the physics issues at a collider, in particular
the PLC, would be the
search for the Higgs and measurement of its properties,
the search for other supersymmetric particles
and determination of their properties, and precision measurements
in order to see indirect effects of new physics. An example of the latter is
the
measurement of the trilinear couplings, $WW\gamma$.
These topics can be covered at moderate
energies linear colliders ($\sqrt{s_{ee}}=300-500 GeV$)
If the light Higgs does not exist, then ESB is triggered by strong forces,
the scale being set by
$\Lambda=4\pi v\approx 1$~TeV. Although the details of the model are
not known
there must be new physics at this scale (e.g. technicolour,
strongly-interacting particles).
In particular, this new physics would show up in $W$ self-interactions
or in $W_LW_L$ scattering.
The connection between heavy Higgs and longitudinal W's is best
established via the process
$W_L W_L\rightarrow W_L W_L$.
In the SM the Higgs is introduced to cure the bad high-energy behaviour
of this amplitude; nonetheless
$W_LW_L$ interactions become strong if $m_H\approx .8-1$~TeV.
The physics of the heavy Higgs is most relevant for TeV linear colliders.
Since no model has gained a consensus to
describe the strongly-interacting electroweak sector, one must
strive for a model-independent
description of this sector.
One approach, which is valid up to some
scale $\Lambda$, uses an effective chiral Lagrangian.
Assuming a custodial $SU(2)$ symmetry
to ensure that the parameter $\rho\approx 1$,
new physics in the weak boson sector is described by
nonrenormalizable terms suppressed by powers of
$1/\Lambda$,
\begin{eqnarray}
{\cal L}={\cal L}_{SM} (no Higgs)+\sum \frac{1}{\Lambda^n} {\cal L}_n
\end{eqnarray}
\noindent
The leading order chiral Lagrangian, ${\cal L}_2$,
contains only the mass terms for the vector bosons
while the Next-to-Leading order,
${\cal L}_4$, contains the self-interactions. This
includes trilinear or quartic interactions of massive vector bosons
and at most one photon\cite{Hawai}. Self-interactions including two photons
only appear at higher order, ${\cal L}_6$.
The effective Lagrangian formalism would break down if the
scale at which the experiment is performed is sufficient to produce
new resonances. These must be explicitly incorporated,
the cases
of either a scalar, vector or tensor resonance will be considered.
The scalar one ($\sigma$-like) is representative of
a heavy Higgs while the vector one ($\rho$-like)
occurs in technicolour.
To
cover all standard and non-standard manifestations of symmetry breaking,
the strategy at the future collider must include: tests of $W$
self-interactions and of longitudinal
vector-boson scattering, the search for the Higgs in the whole
range of possible masses, as well as searches for other heavy resonances.
All these aspects can be tackled at a photon collider,
as will be described in the rest of this talk.
\section{$\gamma \gamma \ra W^+ W^-\;$ and W self-interactions}
The importance of this process cannot be over-emphasized considering the
large cross-section involved.
Although the bulk of the reaction is due to the gauge transverse sector,
the fact that there are so many $W$'s around
makes this reaction the ideal place
to conduct precision tests of the electromagnetic
couplings of the $W$.
In the effective chiral Lagrangian description of Higgsless models,
the anomalous trilinear couplings invoke two C and P conserving operators at
Next-to-Leading order,
$L_{9L}$ and $L_{9R}$, and one C and P violating
operator $L_{C}$\cite{Hawai}.
The latter affects only $ZWW$ and $\gamma ZWW$ interactions while only
the combination $L_{9L}+L_{9R}$ contribute to $\gamma WW$.
The $L_i$ operators are expected to be ${\cal O} (1)$.
There is an extensive literature
\cite{Paris} on the effect of
anomalous couplings in various experiments.
Comparisons of different analyses have shown that a
$500$~GeV linear collider with a luminosity
of $10 fb^{-1}$ does significantly better then the LHC for trilinear
couplings. Furthermore, the
limits that can be obtained in $\gamma \gamma\;$ at this energy,
$
|L_{9L}+L_{9R}|< 10
$
represent a 50\% improvement over $e^{+} e^{-}\;$\cite{Hawai}. This is shown in
Fig.~3.
However, this result is not sufficient to reach the level
where one expects new physics to set in.
In $e^{+} e^{-}\;$ it was shown recently\cite{steve,barklow} that meaningful
limits could be obtained with a
luminosity ${\cal L}=50-80 fb^{-1}$.
It remains to be seen if the same
can be done at a $\gamma \gamma\;$ collider. For that one needs to
generate the four-fermion final state from the decay of the
$W$'s, while keeping the full spin-correlation.
\begin{figure*}[htb]
\caption{\label{gg19}{\em Comparison between the expected bounds on the
two-parameter space $(L_{9L},L_{9R})$ at the NLC500, LHC
and LEP2.
We also show (``bars") the limits from a single parameter fit.}}
\begin{center}
\vspace*{-1.cm}
\vspace*{-1.cm}
\end{center}
\end{figure*}
A $\gamma \gamma\;$ collider can do more than precision tests on $\gamma WW$.
It is also sensitive to $ZWW$ couplings through processes with three
particles in the final state, for example
$\gamma \gamma \rightarrow WWZ$\cite{nousggwwz}, $e\gamma \rightarrow e WW$ and $e\gamma\rightarrow \nu W Z$.
The latter
is very sensitive to the operator
$L_{C}$\cite{Dawsong5}. The limits obtained, $L_{C}<25$
at 500 GeV are comparable to the ones from
$e^+e^-\rightarrow W^+W^-$\cite{BMT}. Furthermore, the sensitivity
increases rapidly with energy.
\subsection{Effect of radiative corrections}
When doing precision tests
one must worry about the effect of radiative corrections that could
mimic those of the new couplings.
Recently the complete one-loop SM corrections for helicity amplitudes
for $\gamma \gamma \ra W^+ W^-\;$ were calculated\cite{Dennerradcor}. It turns out that the
radiative corrections for this
process are theoretically clean
due to the absence of most
universal leading corrections.
The running of $\alpha$ is irrelevant since we are dealing with on-shell
photons,
all uncertainties due to
$\log(m_q^2)$ terms in small masses disappear, and
there are no large log corrections associated with colinear
photons except at very high energies. Furthermore, the corrections are not very
sensitive to either $M_{t}$ or $M_{H}$ except near the resonance.
Although
some helicity amplitudes receive huge corrections,
they are precisely the ones
that contribute very little to the total cross-section.
Typically radiative corrections between $1-10\%$ at
$\sqrt{s_{ee}}=500$~GeV are obtained, and they tend to increase with
energy ($\approx 20\%$ at $1$~TeV).
In any case, the
inclusion of radiative corrections are not expected to change much
the previously obtained results on measurements of trilinear couplings.
Considering the large
numbers of $W$'s available, there are other interesting questions
that can be studied in $\gamma \gamma\;$ which I
have not addressed here. Among them are the possibility of
direct tests of quartic couplings involving photons\cite{nousggvv},
CP tests in $W$ decay
and measurement of the $Wtb$ coupling which could also give some clues
about symmetry breaking.
\section{Higgs searches}
One of the most attractive motivations for doing physics with very
energetic
photon beams is the unique capability of this mode for producing a
scalar particle, such as the Higgs, as a resonance.
I have already mentioned that the coupling of
the Higgs to two photons occurs only at the loop level.
It should be emphasized that a precision measurement of the
$H\gamma \gamma$ coupling is an indirect way of revealing all massive charged
particles that could be present in an extension of the standard model.
These heavy quanta would not decouple and would
contribute to the production rate in $\gamma \gamma\;$.
While many processes are sensitive to the presence of the Higgs
(see Table 1), the prime interest of the photon mode lies in the
Intermediate Mass region
\footnote{$\gamma \gamma\;$ is also very useful in the mass range below 90 GeV,
a case can be made for building a low-energy dedicated $\gamma \gamma\;$ collider
in the event of a discovery of the Higgs at LEP2.}.
For such a Higgs, the main decay mode is into $b\overline{b}$.
Although a search is feasible at LHC, it will be a difficult and long task to
extract
a signal in this case.
For heavier Higgs masses the resonance can be seen in the $WW$
\cite{Ginzburgshef} or $ZZ$ channel.
However, the usefulness of these modes is tamed by the presence of
large
backgrounds from transverse vector bosons. Ultimately,
for a Higgs above 400 GeV, and regardless
of the energy available for the PLC, one would have to resort
to other channels such as associated Higgs production
or $WW$ fusion (via the process $\gamma \gamma\rightarrow WWWW$).
\baselineskip=12pt
\begin{table*}[htb]
\caption{\label{Higgsmass}
{\em Processses for Higgs searches at PLC and other colliders}}
\vspace*{0.3cm}
\centering
\begin{tabular}{|l|l|l|l|}
\hline
Mass&Collider& PLC&$\sqrt{s_{ee}}$
\\
\hline
&&&\\
$M_{H}<65 GeV$& LEP& Ruled out&-----\\
$65 GeV<M_{H}<90GeV$&LEP2&$\gamma \gamma\rightarrow H\rightarrowb\overline{b}$&$.1-.5$~TeV\\
$90 GeV<M_{H}<140GeV$&NLC & $\gamma \gamma\rightarrow H\rightarrowb\overline{b}$&$.2-.5$~TeV \\
$140 GeV<M_{H}<200GeV$&LHC &$\gamma \gamma\rightarrow H\rightarrow WW$&$.5$~TeV \\
$200 GeV<M_{H}<400GeV$&LHC &$\gamma \gamma\rightarrow H\rightarrow ZZ$&$1$~TeV \\
$400 GeV<M_{H}<700GeV$&LHC & $\gamma \gamma\rightarrow WWWW$&$2$~TeV\\
\hline
\end{tabular}
\end{table*}
\baselineskip=14pt
\subsection{Intermediate Mass Higgs}
For the IMH, $\gamma \gamma\;$ can contribute in the discovery mode
or perform precision measurements of its properties.
A crucial point relates to the choice of the spectrum
and polarization used. Since the Higgs is produced only
in the
$J_z=0$ channel, polarization plays a crucial role in
enhancing the signal over background.
Assuming the Higgs has been found and its mass
measured, one could
tune the energy of the collider and the parameters of the laser
such that the peak of luminosity lies precisely at $\sqrt{s_{\gamma \gamma}}=M_{H}$.
This is obviously the preferred way to operate when measuring
$H\gamma \gamma$,
though one has to realise that good
luminosity is required. Early estimates
for a 500 GeV collider and a luminosity of ${\cal L}=20 fb^{-1}$
give a 10\% precision on the width,\cite{Borden} the effect of background from
one
radiated gluon is discussed by Khoze \cite{Khozeshef}.
One disadvantage of operating in that mode is that the PLC would be run
at energies much below the nominal $e^{+} e^{-}\;$ energy,
precluding the study of interesting processes such as the $W$ pair
production
and other $W$ reactions that could occur at higher
energy. Furthermore, this low-energy narrow-band scheme could
render kinematically inaccessible some of the particles
that would only be probed indirectly in $H\gamma \gamma$,
not to mention that the $\gamma \gamma\;$ mode, when operated in the full range of
energy, can access scalar particles that would be kinematically out of
reach in the $e^{+} e^{-}\;$ mode.
If one would have two interaction regions,
one devoted to $\gamma \gamma\;$ the other to $e^{+} e^{-}\;$,
and if the Higgs has not been found elsewhere,
$\gamma \gamma\;$ could be used to search for the Higgs. The
method
that allows for simultaneous studies
of processes at high energy consists of running
the PLC using a ``broad" spectrum
so that one would have reasonable luminosity over a range of energies.
The main problem in the Higgs searches, whatever the scheme used,
lies in the large background. The prominent one comes from
direct QED $\gamma \gamma \rightarrow q \bar q$ production
where $q=b$ or other light quark flavours, in particular charm.
However this background can be dealt with since
the bulk of the cross-section is in the forward direction, so that a modest
angular cut could efficiently suppressed this background
and would almost totally eliminate its
$J_Z=0$ contribution.
Therefore
a spectrum with a predominantly $J_Z=0$
component would both enhanced the signal and reduce the
background.
When the PLC is run in the ``broad" spectrum mode there are other
more important backgrounds that have to be taken into
account\cite{Halzen,Higgsreson}.
They arise from the hadronic structure of the photon
which can resolve into a gluon or a quark with some spectator jets
left over. One then has to worry about $q \bar q$ production through $\gamma g$
as well as a host of 1-resolved and 2-resolved process.
These backgrounds dominate the signal. However,
since most of the resolved events are very boosted,
judicious cuts can reduce it to a manageable
level. It was shown, using the ideal spectrum,
that at $500$~GeV with ${\cal L}=10 fb^{-1}$,
one could obtain a good signal for $M_H=110-140$~GeV
\cite{Higgsreson}. Furthermore,
the situation improves for a collider of lesser energy, due to
the reduced resolved background. For example,
at $350$~GeV a signal is easily extracted for the whole IMH range.
Of course this assumes the
ideal spectrum. However, for the masses
considered, the signal falls in the region where the spectrum is most
severely affected by effects of multiple scattering and nonlinear
effects. These questions should be reassessed taking these effects into
account as the conclusions could differ drastically.
There have been suggestions to determine directly
the parity of the Higgs using
linear polarizations of the photon\cite{Gunionparity}. Since the degree of
linear
polarization is never very large
($<30\%$), this always requires large luminosities,
${\cal L}=100 fb^{-1}$.
\subsection{Associated production}
For the IMH, it will be hard to unravel a peak formation
if the collider energy is greater than $500$~GeV.
As will be discussed in the next section, the resonance will remain
hidden
for heavier Higgs ($M_{H}\ge 400$~GeV)
even if one uses the most favourable channel, $ZZ$.
Fortunately, other efficient mechanisms for Higgs production are
available, in particular the
radiation of a Higgs from a $W$ pair.
This is to be expected since the cross section for $W$
pair production is
so large and the Higgs couples preferentially to the weak bosons.
In fact, at
1 TeV, before folding with the luminosity
spectrum, the $e\gamma$, $\gamma \gamma\;$ and Bjorken process are comparable
for all Higgs masses\cite{nousgg3v}.
However, the $\gamma \gamma\;$ production mode is suppressed
when including a more realistic photon luminosity
spectrum. Still,
at $1$~TeV one obtains a measurable cross-section ($\sigma >3fb$)
for $M_{H}<400 GeV$.
\subsection{\boldmath{ $\gamma \gamma \rightarrow ZZ$}}
At first this reaction was believed to
provide a background-free environment
for either Higgs production or non-standard physics
signals in $Z_LZ_L$ since it is purely a loop process
in the ${\cal{S}} {\cal{M}}\;$.
The first full calculation by Jikia\cite{Jikiazz}
of the one-loop process $\gamma \gamma \ra ZZ \;$ within the ${\cal{S}} {\cal{M}}\;$
dampened this enthusiasm since it
turned out that, once again, the
transverse modes are overly dominant, especially at high energy.
This is due essentially to the $W$
loops, the $WW$ produced in $\gamma \gamma\;$ rescatter
into $ZZ$.
At
$\sqrt{s_{ee}}=400$~GeV,
the Higgs resonance is clearly evident over the $TT$ continuum all the way
up to the kinematic limit. With $\sqrt{s_{ee}}=500$~GeV,
it already becomes difficult to extract a Higgs with $M_H\sim350$~GeV.
\cite{Jikiazz}
To obtain these results, Jikia used a predominantly $J_Z=0$
spectrum that is peaked towards the maximum $\hat{s}_{\gamma \gamma}$,
this is not the optimum choice.
With a broader spectrum featuring
a dominant $J_Z=0$ for small $M_{ZZ}$,
one could still see a peak in the $M_{ZZ}$
invariant mass for Higgs masses up to $400$~GeV at a $1$~TeV $e^{+} e^{-}\;$ machine
\cite{DicusKao}.
{}From the perspective of observing the Higgs resonance beyond TeV
$e^{+} e^{-}\;$ energies, the situation
becomes totally hopeless as the
transverse $ZZ$ are awesome\cite{Jikiazz}.
\section{Strongly-interacting electroweak sector (SEWS)}
If the Higgs is not found at LHC, or in the sub-TeV version of
NLC, we will be in the realm of the SEWS.
This sector would be probed most efficiently
at TeV energies through the reaction
$V_L V_L\rightarrow V_L V_L $ ($V=W$ or $Z$).
In this channel one would either search for a resonance
or, if the energy is not sufficient, for new interactions such as the ones
described by the effective chiral Lagrangian.
The $V$ pair-production processes could be regarded as the testing ground for
possible rescattering effects in $WW\rightarrow VV$ that originate from the
symmetry breaking sector.
Unfortunately,
at high energies, {\it i.e.}, at high $VV$ invariant masses, where
the effect of the New Physics would be most evident, one has to fight
extremely hard against the background for transverse $W$ and $Z$.
Indeed,
recent analyses have shown that while it might be possible to see
effect of a tensor resonance, a scalar one as well as indirect effects
are
hopeless\cite{BergerBerkeley}.
One then has to resort to the only source of longitudinal vector
bosons, the ones taking part in the fusion process and contributing
to $WWWW$ or $WWZZ$ production. This process is the analog of $e^+e^-\rightarrow
\nu\overline\nu W^+W^-$ and was originally believed to be
more favourable due to a presumed larger $W_L$ content in the
photon than in the electron.
While it is true that in the photon there is an additional
structure function corresponding to the spectator $W_L$,
the dominant contribution is from transverse
spectator $W$'s. The latter features basically the same structure
function as in the electron, except for an overall factor\cite{Paris}.
One would therefore expect that
$\gamma \gamma\;$ should be comparable to $e^{+} e^{-}\;$ at the same energy and luminosity.
This conclusion was born out
by two independent exact calculations of this ${\cal{S}} {\cal{M}}\;$ process
\cite{Jikiawwww,Cheungwwww}.
The signal of a heavy Higgs is a significant increase
in the channels with at least three $W_L$.
To extract a signal requires tagging all four $W$'s, the spectator ones
being associated with
the low $p_T$ and the longitudinal ones with the central $W$'s.
The spectators are tagged with one hadronic and one leptonic decay while
the central ones go into four jets. The results of the analysis
showed that a 2~TeV PLC (${\cal L}=10 fb^{-1}$)
would give a good signal ($S/\sqrt{B}\approx 10$)
for a heavy Higgs-like scalar of 1TeV\cite{Cheungwwww}. This is comparable to
the $e^{+} e^{-}\;$ process. However, the inclusion of the
spectrum has a dramatic effect and a linear collider of
$2$~TeV in the $e^{+} e^{-}\;$ center of mass with ${\cal L}=200fb^{-1}$
is needed to reach the same significance level.
Another interesting conclusion from these calculations is that a signal
for a Higgs of 400-700~GeV can easily be seen with
$\sqrt{s_{ee}}=1.5$~TeV and ${\cal L}=200 fb^{-1}$.
The PLC can therefore cover the whole mass range for light or
heavy Higgs searches
provided a good choice of energy and spectrum is made,
although precision measurements are possible only for light Higgs
($M_{H}< 120 GeV$).
\section{Search for new particles: supersymmetry}
As the best motivated alternative to the standard model, one should
investigate the consequence of supersymmetic models.
Supersymmetry would provide a natural framework for
light Higgses.
The three neutral scalars of supersymmetric models, $h$,$H$,$A$
(pseudoscalar)
could be produced as a resonance in $\gamma \gamma\;$. This reaction would then extend
the reach in mass of $e^{+} e^{-}\;$ since in the latter $H$ and $A$ can only be
produced together and require
$\sqrt{s_{ee}}>M_H+M_A$. At $\sqrt{s_{ee}}=500$~GeV, using the $b\overline{b}$ mode, this gives
the following discovery region for the supersymmetric
scalars\cite{GunionHaber}:
$
110$~GeV$< M_{H} < 200$~GeV and $100$~GeV$<M_A < 2 M_{t}$.
Recently, it was pointed out that this was true only if scalars decayed
primarily via ${\cal{S}} {\cal{M}}\;$ final states.
Otherwise the above limits require high luminosities ${\cal L}> 60
fb^{-1}$\cite{Gunionsusy}.
The $\gamma \gamma\;$ collider can search also for other supersymmetric particles
\cite{Murayama},the
production cross-sections being universal, were already shown.
Typically, one finds that $\gamma \gamma\;$ can have good cross-sections but offer
little advantage over the $e^{+} e^{-}\;$ mode, in part because of the
lower achievable energy. It is for selectron searches in $e\gamma\rightarrow
\tilde{e}\tilde{\chi}$ that the laser scheme becomes extremely useful,
as the discovery of a selectron of
$m_{\tilde{e}}\approx\sqrt{s_{e\gamma}}$ is possible.
\section{Conclusion}
A $\gamma \gamma\;$ collider of energy ranging from .2 to 2~TeV
should prove to be a useful tool for probing the electroweak
symmetry-breaking sector through either Higgs
searches or $W$ physics. It is unique in producing
a scalar on resonance and is complementary
to an $e^{+} e^{-}\;$ collider in many processes.
\noindent
{\bf {\large Acknowledgements}}
I am most grateful to my friends and
collaborators Marc Baillargeon and Fawzi Boudjema for all
the work on electroweak physics issues.
I also thank George Jikia for
kindly supplying the curves for $\gamma \gamma\rightarrow WWWW$.
\vspace{.5in}
|
\section{INTRODUCTION}
Spectroscopy in the superdeformed minimum has reached a certain level of
maturity to justify a phenomenological analysis of the available data (see
\cite{jan91,nol94} for recent reviews). Such an approach would be useful in
systematizing the data and would also provide a complimentary perspective
to the more microscopic theories. For this purpose, we use the interacting
boson model (IBM) \cite{iac87} which has been established as one of the
simplest and most versatile collective models. It has been especially
successful in correlating spectroscopic data in deformed nuclei in terms of a
few parameters of a quadrupole Hamiltonian \cite{cas88}.
Microscopic study of the nucleon pair structure in the superdeformed well
\cite{hon92} indicates that, compared to the deformed nuclei, they have
about three times more active pairs of nucleons (bosons), and the $L=4$ pair
($g$ boson) plays a much more significant role. As numerical diagonalization of
an $sdg$-IBM Hamiltonian is not possible for more than 10 bosons, one needs
alternative methods of solution to apply the IBM to superdeformed nuclei.
Here, we use the angular momentum projected mean field
theory which leads to a $1/N$ expansion for all matrix elements of
interest \cite{kuy88}. Accurate representation of high-spin states in the
$1/N$ expansion formalism requires terms to order $(L/N)^6$ which have been
obtained recently using computer algebra \cite{kuy95}. The extended formalism
provides an analytical method for analysis of superdeformed states which is
both accurate and efficient.
The plan of the paper is as follows. After reviewing the microscopic basis
of the IBM for superdeformed nuclei, we introduce the $1/N$ expansion formalism
and
discuss its recent extension to higher orders. We then use the $1/N$ expansion
formulas for a quadrupole Hamiltonian to study systematic features of
dynamic moment of inertia and $B(E2)$ values. The results are used in a
description the superdeformed bands in the Hg-Pb and Gd-Dy regions.
\section{MICROSCOPIC BASIS}
In this section, we study the typical structure of strongly deformed states
and investigate the relation between the superdeformation and the IBM
based on \cite{hon92}.
For this purpose, we use the Nilsson + BCS model with particle number
projection. Superdeformed states can be characterized as ground states in a
superdeformed potential well which is separated from the normal one by a
potential barrier. For such ground-like states which show strongly collective
nature, this model seems to work well. Using the experimental deformation
parameters and electric transition probabilities (or moments) as input, one can
obtain reasonable wave functions. These wave functions are analyzed from the
viewpoint of collective nucleon pairs, which leads to a natural extension of
the usual IBM.
We briefly summarize the formulation of the Nilsson +
particle-number-conserving BCS model. The single particle orbits in a deformed
potential are described well by the Nilsson Hamiltonian \cite{nil55}
\begin{eqnarray}
H_{\rm Nilsson} &=& -\frac{\hbar^{2}}{2m} \nabla ^2
+\frac{m}{2}\omega_{0}^{2}r^{2}
[1-\frac{4}{3}\delta P_{2}(\cos\theta)] \nonumber \\
&-& 2\hbar \omega_{0} \kappa \mbox{\boldmath $l$} \cdot
\mbox{\boldmath $s$} -\hbar \omega_{0} \kappa \mu
(\mbox{\boldmath $l$}^{2}-\langle \mbox{\boldmath $l$}^{2} \rangle_{N}),
\end{eqnarray}
where $\delta$ is the deformation parameter and $P_{2}(\cos\theta)$ denotes
the Legendre polynomial.
The term $\langle \mbox{\boldmath $l$}^{2} \rangle_{N} = \frac{1}{2}N(N+3)$ is
the expectation value of $\mbox{\boldmath $l$}^{2}$
averaged over one major shell with the principal quantum number $N= 2n+l$.
The value of the oscillator frequency for a mass-A nucleus is determined
from $\hbar\omega_{0} = 41A^{-\frac{1}{3}}$, which is about 7.1 MeV for
the superdeformed nuclei in the Hg-Pb region ($A \sim$ 190) and
7.7MeV for those in the Dy-Gd region ($A \sim$150).
We use the usual values for the parameters $\kappa$ and $\mu$
which are 0.0637, 0.60 for proton orbits and 0.0637, 0.42 for neutron orbits,
respectively \cite{gus67}.
In order to include short range correlations, the monopole pairing
interaction is added to the Nilsson hamiltonian \cite{nil61}
\begin{equation}
H = H_{\rm Nilsson} + G P^{\dagger} P,
\end{equation}
where $G$ denotes the pairing strength parameter, and
\begin{equation}
P^{\dagger} = \sum_{k>0} c_{k}^{\dagger} c_{\bar k}^{\dagger} \ ,
\end{equation}
is a pair creation operator.
Here $c_{k}^{\dagger}$ stands for the creation operator of a nucleon
in the spherical single particle orbit $k$,
and $\bar k$ denotes the time reversed state of $k$.
This Hamiltonian is solved by the variation using a BCS wave function
\begin{equation}
\mid \Psi \rangle = \prod_{\alpha>0} (u_{\alpha} + v_{\alpha}
a_{\alpha}^{\dagger} a_{\bar \alpha}^{\dagger}) \mid 0 \rangle,
\end{equation}
where $a_{\alpha}^{\dagger}$ denotes the creation operator
for a nucleon in the deformed canonical (Nilsson) orbit labeled by $\alpha$.
The particle number conservation has been found to be important in the case of
weak pairing correlations and also for moments of inertia of high spin states
in the cranking calculation of the superdeformed states \cite{dud88,shi90}.
Thus we carry out the particle number projection before variation
according to the method given in Ref. \cite{egi82}.
The solution corresponds to the minimum of the number projected energy
\begin{equation}
E^{P}[\Psi] = \frac{\langle \Psi \mid H P^{N} \mid \Psi \rangle}
{\langle \Psi \mid P^{N} \mid \Psi \rangle},
\end{equation}
where $P^{N}$ denotes the particle number projection operator.
The deformation parameters of the superdeformed states in the Hg-Pb (Dy-Gd)
region are given by $\delta \sim$ 0.40 (0.50), which is equivalent to the
axis ratio of 5:3 (2:1). Because of this strong deformation, it is insufficient
to take only one active major shell and take into account
the corrections due to the
core-polarization effect through renormalization in one major shell.
Thus we first seek a suitable model space for
description of superdeformed states. For simplicity, we turn off the
pairing force which is not important for this purpose. We take $^{194}$Hg
($N$=114, $Z$=80) and $^{152}$Dy ($N$=86, $Z$=66) as examples of the Hg-Pb
and the Dy-Gd regions, respectively.
In order to define the model space necessary for description of
superdeformed states, we utilize the intrinsic quadrupole moment $Q_0$.
The intrinsic quadrupole moment is calculated in the space of all spherical
orbits up to the principal quantum number $N=N_{\rm max}$. Then $N_{\rm max}$
is increased until the value of $Q_0$ is saturated to a good extent.
{}From this procedure we obtain $N_{\rm max}=12$. The corresponding values of
$Q_0$ for proton and neutron orbits are 19{\it b} and 29{\it b} for $^{194}$Hg,
and 19{\it b} and 27{\it b} for $^{152}$Dy, respectively.
Note that the experimental values of $Q_0$ are $18\pm 3~eb$ for $^{152}$Dy
\cite{ben91} and $17\pm 2~eb$ for $^{194}$Hg \cite{hug94}, which are
consistent with the present results if we take the bare charges, $e_{\rm p}=1$
and $e_{\rm n}=0$.
We now consider the inert core of superdeformed states. The Nilsson wave
function is obtained by putting all nucleons in the Nilsson orbits from the
bottom. One Nilsson orbit can be expanded as a linear combination of many
spherical harmonic oscillator orbits, and the square of expansion coefficients
gives the occupation probability of each spherical orbit. We expand all the
occupied Nilsson orbits and sum up all the occupation probabilities which
belong to the same spherical orbits, to obtain the total occupation
probability for a given spherical harmonic oscillator basis. Due to the strong
quadrupole field, one Nilsson orbit spreads over many spherical orbits. Thus
the orbits with very high single particle energy can gain some finite
occupation probabilities, while the occupation of the orbits with small single
particle energy may become incomplete. Nevertheless several lower spherical
orbits are occupied almost completely and can be considered as a new inert
spherical core for the superdeformed states. Note that we do not take the usual
``hole'' picture as it is meaningful only for states whose configuration are
well described within one major shell.
First consider the case of $^{194}$Hg. In Fig.~\ref{fig1} the occupation
probability of each spherical harmonic oscillator orbit is shown for
neutrons (a) and protons (b). The orbits are ordered according to their single
particle energy at $\delta=0$ as $1s_{1/2}$, $1p_{3/2}$, $\cdots$. The case of
$\delta=-0.13$ which simulates the deformation of normal oblate states is also
shown for comparison.
For normal deformation, it is seen from Fig.~\ref{fig1}-a that the
occupation of
the proton orbits is almost complete at $2d_{5/2}$ ($Z=64$), while the
occupation probability is almost vanished for orbits above $Z=82$.
These results suggest that we can consider the $Z=64$ subshell
as an inert core and three valence orbits ($2d_{3/2}$, $3s_{1/2}$ and
$1h_{11/2}$) as active.
This gives the valence proton number as $Z_{v}=16$.
In the case of $\delta=0.40$, the proton orbits are almost completely
occupied up to $Z=50$. Above $Z=50$, the occupation probability
drops suddenly though it remains about 10$\%$ over many orbits.
Clearly, one should incorporate the contributions of these high
energy orbits. Thus it is reasonable to take the $Z=50$ spherical inert core
and include quite many orbits above there as active valence orbits. In this
case the valence proton number becomes $Z_{v}=30$.
In the same way, it can be
seen form Fig.~\ref{fig1}-b that the spherical inert core for neutron orbits
are $N=100$ ($N_{v}=14$) and 82 ($N_{v}=32$) for $\delta=-0.13$ and
0.40, respectively. The active valence orbits for $\delta=-0.13$ are
$2f_{5/2}$, $3p_{3/2}$, $3p_{1/2}$ and $1i_{13/2}$, while for $\delta=0.40$
it is still insufficient to include only two or three major shells.
We can see a similar behaviour of occupation probabilities in the
wavefunctions of $^{152}$Dy, which is shown in Figs. \ref{fig1}-c for
proton and \ref{fig1}-d for neutron orbits. In these figures two cases of
$\delta=0.50$ (superdeformed state) and 0.25 (normal prolate state) are
compared. It is clear that $Z=50$ and $N=82$ inert cores are good for normal
states, while $Z=40$ and $N=50$ cores are suitable for superdeformed states.
Thus the inert core of superdeformed states becomes much smaller than that of
normal states in both the Hg-Pb and Dy-Gd regions.
{}From the viewpoint of the IBM, the number of bosons is determined by half of
the number of valence nucleons. Because of the small inert core the number of
bosons increases significantly for superdeformed states in comparison with
that in the usual IBM. In fact in the case of $^{194}$Hg, the boson number
in the usual IBM is $N_{\rm normal}=(82-80)/2+(128-114)/2=8$ by taking
the usual hole picture, and $N_{\rm normal}=(80-64)/2+(114-100)/2=15$ with the
particle picture mentioned above. On the other hand, the number of bosons for
superdeformation becomes $N_{\rm super}=(80-50)/2+(114-82)/2=31$.
Similarly, in the case of $^{152}$Dy, $N_{\rm normal} = (66-50)/2 +
(86-82)/2 = 10$ while $N_{\rm super} = (66-40)/2 + (86-50)/2 = 31$. In
general, $N_{\rm super}$ is about three times larger than $N_{\rm normal}$.
Note that the number of proton bosons and neutron bosons are close in these
two cases, and this approximate equality seems to be a general tendency of the
superdeformed states. This result can be naturally understood since equal
numbers of valence protons and neutrons maximizes the attractive
proton-neutron interaction.
Next we consider the effects of pairing correlations on the structure of
wave functions of superdeformed states. The strength parameter $G$ of the
pairing interaction should be chosen depending on the model space.
Since the value of $G$ for such a large space is not known empirically,
we first describe normal states within the extended valence space,
and determine the value of $G$ by requiring that the pairing gap ${\it
\Delta}$ takes a reasonable value. For $^{194}$Hg the value of $G$ has turned
out to be 0.06MeV which gives ${\it \Delta}\sim 1$ MeV. Using this value, we
investigate the effect of the pairing correlations on the structure of valence
wave functions. For this value of $G$, the gap for superdeformed states becomes
about $\Delta =0.5$ MeV for both proton and neutron orbits.
In contrast to normal deformed states, which are sensitive to changes in values
of $G$, the superdeformed states are almost insensitive to $G$ values
(the intrinsic quadrupole moment and the occupation probabilities
change very little). Thus the following discussion about the structure of
valence wave function of superdeformation
is almost independent of pairing correlations.
We can investigate the relation between the superdeformation and the IBM by
analyzing valence wave functions from the viewpoint of collective nucleon
pairs. Since the bosons in the IBM are understood as images of these
pairs, such an analysis is essential in establishing a microscopic basis for
the super IBM. We consider $^{194}$Hg as an example.
The Nilsson + particle-number-conserving BCS wave function can be expressed
as the condensed state of coherent Cooper-pairs in the deformed potential
\cite{ots86}
\begin{equation}
P^{N} \mid \Psi \rangle \propto ({\it \Lambda}_{\pi}^{\dagger})^{N_{\pi}}
({\it \Lambda}_{\nu}^{\dagger})^{N_{\nu}}
|0 \rangle ,
\end{equation}
acting on the inert core $|0 \rangle$. In this expression,
${\it \Lambda}_{\pi}^{\dagger}$ (${\it \Lambda}_{\nu}^{\dagger}$)
denotes the creation operator of a Cooper-pair in proton (neutron) orbits and
$N_{\pi}$ ($N_{\nu}$) means half of the valence proton (neutron) number.
These ${\it \Lambda}$-pairs can be decomposed into a linear combination of
collective nucleon pairs with good angular momenta
\begin{equation}
{\it \Lambda}^{\dagger} = x_{0}S^{\dagger}
+ x_{2}D_{0}^{\dagger}
+ x_{4}G_{0}^{\dagger}+\cdots ,
\end{equation}
where $S^{\dagger}$, $D^{\dagger}$, $G^{\dagger}$, $\cdots$ denote the
collective nucleon pairs with spin-parity $J^{\pi}=$ $0^{+}$, $2^{+}$,
$4^{+}$, $\cdots$ and the $x_{J}$'s are amplitudes. The probability of each
pair in the ${\it \Lambda}$-pair is given by the square of each amplitude,
and is listed in Table~\ref{table1} for two cases of $\delta=-0.13$ and
$\delta=0.40$. It is well known that in the case of normal deformation the
dominant components are the $S$- and $D$-pairs \cite{ots82,bes82}.
In fact, these two components account for 100$\%$ probability
in the case of $\delta=-0.13$. In the case of $\delta=0.40$, the total
probability of the $S$- and $D$-pairs is about 80$\%$ and we can
conclude that these pairs are still dominant in the ${\it \Lambda}$-pair.
However the probability of the $G$-pair is now sizable, and it can no
longer be neglected in a detailed description of high-spin states.
It should be noted that the ratio of the $S$-pair to the other pairs
is quite similar to that of $s$-boson to the other bosons in the SU(3) limit
of the IBM, which are shown in the same table. This suggests that the SU(3)
limit of the $sdg$-IBM could provide a reasonable phenomenological framework
for superdeformed states.
To summarize the microscopic results, we emphasize two important points
for the
description of superdeformed bands in the IBM: One is the significant increase
in the boson number, and the other is the importance of g-bosons. In addition,
it has been found that the bosons for superdeformed states carry the
collectivity over many major-shells and that the SU(3) limit is a reasonable
starting point.
\section{1/N EXPANSION FOR SUPER IBM}
A simultaneous description of the spectroscopy of normal and superdeformed
states requires rather complicated wave functions, therefore we focus on the
latter here and leave the complete picture for future work. We introduce
the superbosons ${\bf s}, {\bf d}, {\bf g}$ as the boson images of the $S, D,
G$ collective nucleon pairs in the superdeformed well (bold face notation is
used for super bosons to distinguish them from the normal ones). The quadrupole
Hamiltonian for this system of bosons has the form
\begin{equation}
H=-\kappa {\bf Q} \cdot {\bf Q},
\label{ham}
\end{equation}
where the quadrupole operator is defined as
\begin{equation}
{\bf Q} =
[{\bf s}^\dagger {\tilde {\bf d}} + {\bf d}^\dagger {\tilde {\bf s}} ]^{(2)} +
q_{22} [{\bf d}^\dagger {\tilde {\bf d}} ]^{(2)} +
q_{24} [{\bf d}^\dagger {\tilde {\bf g}} +
{\bf g}^\dagger {\tilde {\bf d}} ]^{(2)} +
q_{44} [{\bf g}^\dagger {\tilde {\bf g}} ]^{(2)}.
\label{qsdg}
\end{equation}
Here brackets denote tensor coupling of the boson operators and
$\tilde b_{lm}=(-1)^{m}b_{l-m}$. The parameters $q_{jl}$ in Eq. (\ref{qsdg})
determine strengths of boson interactions relative to the $s-d$ coupling.
Since the SU(3) limit is used as a reference
point in the rest of the paper, we quote the values for the quadrupole
parameters in this limit; $q_{22}=-1.242$, $q_{24}=1.286$, $q_{44}=-1.589$.
As stressed in the introduction, numerical diagonalization of this Hamiltonian
for $N \sim 30$ bosons is not possible even on a supercomputer.
The large number of bosons are, however, an advantage for the analytic
$1/N$ expansion technique which we employ here for solving the Hamiltonian
Eq. (\ref{ham}).
The $1/N$ expansion method has previously been discussed in detail \cite{kuy88}
and the recent extensions to higher orders are given in Ref.~\cite{kuy95}.
Therefore, we give only a short account of the formalism here, focusing
mainly on the accuracy of the results for high-spin states.
The starting point of the $1/N$ calculations is the boson condensate
\begin{equation}
|N,{\bf x}\rangle =(N!)^{-1/2}({\bf b}^\dagger)^N|0\rangle,\quad
{\bf b}^\dagger= x_0 {\bf s}^\dagger + x_2 {\bf d}_0^\dagger +
x_4 {\bf g}_0^\dagger,
\end{equation}
where $x_l$ are the mean field amplitudes to be determined by variation after
projection (VAP) from the energy expression
\begin{equation}
E_L=\langle N,{\bf x}| H P^L_{00} | N,{\bf x}\rangle /
\langle N,{\bf x}| P^L_{00} | N,{\bf x}\rangle.
\end{equation}
Here $P^L_{00}$ denotes the projection operator.
The resulting energy expression is a double expansion in $1/N$ and
$\bar L=L(L+1)$, and has the generic form
\begin{equation}
E_L = N^2 \sum_{n,m} {e_{nm}\over (aN)^m}
\Bigl({\bar L \over a^2N^2}\Bigr)^n,
\label{me1}
\end{equation}
where $a=\sum_l \bar l x_l^2$ and the expansion coefficients $e_{nm}$ involve
various quadratic forms of the mean fields $x_{l}$, e.g.,
$e_{00}=(\sum_{jl} \langle j0 l0|20 \rangle q_{jl} x_j x_l)^2$.
The coefficients $e_{nm}$ have recently been derived
up to the third order, $(\bar L/N^2)^3$, using computer algebra \cite{kuy95}.
Another observable of interest in the study of superdeformed states is the
$E2$ transitions. Assuming that the quadrupole transition operator is the same
as in the Hamiltonian, i.e. $T(E2)=e{\bf Q}$ where $e$ is an effective boson
charge, the $E2$ matrix elements are given by
\begin{equation}
\langle L \parallel T(E2) \parallel L-2 \rangle =
e N \hat L \langle L0\, 20|L-2\ 0\rangle \bigl[m_1 + m_2 L(L-1)/N^2\bigr]
\label{e2}
\end{equation}
where $\hat L = [2L+1]^{1/2}$ and the coefficients $m_n$ are given in
Ref.~\cite{kuy95}. The first term in Eq. (\ref{e2}) gives the familiar
rigid-rotor result. The second term is negative and is responsible for the
boson cutoff effect in $E2$ transitions.
Before applying the 1/$N$ expansion results, we compare them with those
obtained from an exact diagonalization of the Hamiltonian.
Diagonalization is carried out for $N=10$ which is the maximum possible boson
number for this purpose.
The quadrupole parameters $q_{22}, q_{24}, q_{44}$ are scaled down from their
SU(3) values with $q=0.7$ which gives an adequate parametrization for the Hg-Pb
region.
Fig.~\ref{fig2}-a compares exact results for the dynamic moment of inertia
${\cal J}^{(2)}$ (circles) with three different $1/N$ calculations.
The solid line shows the third order VAP results which is seen to follow the
exact results very accurately. The second order VAP (dotted line) and
the third order VBP results (dashed line) break down around spin $L\sim 2N$.
Hence for description of high-spin states, the third order $1/N$ expansion
formulas with VAP seem to be both necessary and sufficient.
In Fig.~\ref{fig2}-b, the exact $E2$ transition matrix elements (circles)
are compared with those obtained from Eq. (\ref{e2}) (line).
The agreement is again very good up to very high-spins.
Note that the accuracy of the $1/N$ expansion results improves with increasing
$N$, hence in actual applications with $N\sim 30$, one would expect an even
better agreement.
The test case discussed here indicates that the extended formalism can be
applied with confidence in the spin region $L=N$-$3N$ which covers the
presently available data range for superdeformed bands.
\section{APPLICATIONS TO SUPERDEFORMED BANDS}
In this section, we apply the $1/N$ expansion formulas first in a
systematic study of dynamic moment of inertia and $B(E2)$ values, and then
to describe the experimental data on superdeformed bands.
Since $\kappa$ is a scale parameter for energies, we need to study
the effect of the three quadrupole parameters $q_{22}$, $q_{24}$, $q_{44}$.
Fig.~\ref{fig3} shows the effect of variations in each $q_{jl}$ on
dynamic moment of inertia while the other two are held constant at the
SU(3) values. Here $q$ denotes the scaling parameter from the SU(3) values.
Thus $q=1$, corresponds to the SU(3) limit which exhibits the rigid-rotor
behaviour. To describe the variations in ${\cal J}^{(2)}$, one needs to break
the SU(3) limit. From Fig.~\ref{fig3} it is seen that ${\cal J}^{(2)}$
is most sensitive to $q_{24}$ (note the different scales in the three figures).
The other (diagonal) parameters have smaller and opposite effect on ${\cal
J}^{(2)}$. Since the amount of data does not justify use of too many
parameters, we prefer to scale all three with the same parameter $q$. The
result of this simultaneous scaling is shown in Fig.~\ref{fig4}-a which is
essentially the same as the one for $q_{24}$ in Fig.~\ref{fig3}. An interesting
feature of these results is that the quadrupole Hamiltonian has the scope to
describe both the increases and decreases in ${\cal J}^{(2)}$. For $q<1$,
the $s-d$ coupling is relatively stronger than the $d-g$ coupling which results
in loss of monopole pairing with increasing spin, and hence increase in ${\cal
J}^{(2)}$. The opposite happens for $q>1$.
Fig.~\ref{fig4}-b shows the effect the simultaneous variations in the
quadrupole parameters on the $B(E2)$ values. The curving down of lines is due
to boson cutoff which is most effective for smaller values of $q$.
In the light of the above systematic studies, we have carried out fits to the
available data on superdeformed bands in the Hg-Pb and Gd-Dy regions.
The boson number is determined from microscopics, and $\kappa$ and $q$ are
fitted to the data. The parameter values are given in the figure
captions and the data are taken from the compilation in Ref. \cite{fir94}.
Figs.~\ref{fig5} and \ref{fig6} compare the experimental dynamic moment of
inertia (circles) with the calculated ones (lines) in Hg and Pb isotopes,
respectively. In all cases ${\cal J}^{(2)}$ exhibits a smooth increase which
is well reproduced by the calculations.
The situation in the Gd-Dy region is not as favorable for our simple
collective model as the other region, because
there are definite signs indicating the importance of the
single particle degree of freedom.
For example, in $^{144-146}$Gd there are
sudden jumps in ${\cal J}^{(2)}$ which are probably
due to particle alignment effects
\cite{lun94}. In $^{148-150}$Gd, ${\cal J}^{(2)}$ behaves reasonably smoothly
so we have attempted to describe them (see Fig.~\ref{fig7}). The average
behaviour of ${\cal J}^{(2)}$ in $^{148}$Gd is reproduced but the model fails
in the case of $^{150}$Gd, underscoring the importance of single particle
degree of freedom. For a better description of the data, one needs to
incorporate particle alignment effects in the present formalism by including
two-quasiparticle states in the basis \cite{iac91}.
The dynamic moment of
inertia of superdeformed bands in Dy isotopes exhibit an entirely different
behaviour (Fig.~\ref{fig8}). They are very close to the rigid-rotor values, and
hence the SU(3) limit as reflected in the values of $q\sim 1$.
The $B(E2)$ values provide a complimentary observable to ${\cal J}^{(2)}$ which
could be used as a further test of the model.
In Fig.~\ref{fig9}, the available $B(E2)$ data in $^{192-194}$Hg (circles) are
compared with the calculations. A reasonable description is obtained using
boson effective charges $e=0.12-0.14~eb$ which are typical values used in the
normal IBM calculations. Since the $B(E2)$ values are sensitive to the boson
number (they vary as $N^2$), this provides a consistency check on the
microscopically derived boson numbers. A further $N$ dependence is provided by
the boson cutoff term in Eq. (\ref{e2}) which causes a drop in the calculated
$B(E2)$ values at high-spins. Least-square fits to the data indeed indicate a
drop in the $B(E2)$ values towards the high-spin end. However, the error bars
are too large to reach an unambiguous conclusion whether this effect is genuine
or not.
\section{CONCLUSIONS}
In this paper, we have reviewed a microscopic basis and a practical
formulation of the IBM for application to superdeformed nuclei. The
availability of analytical formulas owing to the 1/$N$ expansion technique
means fast and efficient analysis of data.
As first examples, we have considered the superdeformed bands in the Hg-Pb and
Gd-Dy regions. A good description of data is obtained in the Hg-Pb region
confirming the simple quadrupole nature of these superdeformed bands.
In the Gd-Dy region, the dynamic moment of inertia exhibits large variations
which can not be accommodated in a simple collective model. Such variations are
due to the single particle degree of freedom and require extension of the model
for a better description of the data.
Finally, the formalism can be used in investigating some other interesting
features of superdeformed nuclei such as identical bands and $C_4$ symmetry
which will be pursued in future work.
\section{ACKNOWLEDGEMENTS}
This work is supported in parts by an exchange grant from the Australian
Academy of Science/Japan Society for Promotion of Science and by the
Australian Research Council, and in parts by Grant-in-Aid for Scientific
Research on Priority Areas (No. 05243102) from the Ministry of Education,
Science and Culture.
|
\section{#1}}
\textwidth 159mm
\textheight 230mm
\def\bf Z {\bf Z}
\def\tilde{\chi} {\tilde{\chi}}
\def\sigma {\sigma}
\def\bar{q} {\bar{q}}
\def\epsilon {\epsilon}
\def\begin{equation}{\begin{equation}}
\def\end{equation}{\end{equation}}
\def\begin{array}{\begin{array}}
\def\end{array}{\end{array}}
\def\begin{equation}{\begin{equation}}
\def\end{equation}{\end{equation}}
\def\begin{eqnarray}{\begin{eqnarray}}
\def\end{eqnarray}{\end{eqnarray}}
\def\rightarrow{\rightarrow}
\def\longrightarrow{\longrightarrow}
\def\hspace{0.1in}{\hspace{0.1in}}
\def\hspace{0.1in}{\hspace{0.1in}}
\def\hspace{0.2in}{\hspace{0.2in}}
\def\hspace{0.3in}{\hspace{0.3in}}
\def\hat{S}{\hat{S}}
\def{\cal C}{{\cal C}}
\def{\cal O}{{\cal O}}
\def{\cal F}{{\cal F}}
\def{\cal K}{{\cal K}}
\def{\cal M}{{\cal M}}
\def{\cal P}{{\cal P}}
\def{\cal T}{{\cal T}}
\def\Delta{\Delta}
\defU_q({\cal G}){U_q({\cal G})}
\begin{document}
\oddsidemargin 5mm
\renewcommand{\thefootnote}{\fnsymbol{footnote}}
\newpage
\begin{flushright}
DAMTP-HEP-95/26
\end{flushright}
\vspace{0.2cm}
\begin{center}
{\large {\bf Local Operators in Massive Quantum Field Theories}
}\\
\vspace{0.3cm}
{\bf
Anni Koubek\footnote{Work supported by PPARC grant no. GR/J20661} }\\
\vspace{0.2cm}
{\em Department of Applied Mathematics and Theoretical Physics, \\
CB3 9EW Cambridge, UK }
\end{center}
\subsection*{\large\bf 1 \phantom{11} Introduction}
A fundamental problem in quantum theory is to establish a
connection between its local description (quantum field theory)
and measurable
quantities (particle masses, scattering amplitudes). In elementary
particle physics one relies mostly on approximation techniques due to
the non-integrable structure of the interactions. In order to gain a
deeper understanding of quantum theory in general these issues are
examined for integrable systems, where one hopes to gain exact
relations between the two descriptions. Though many integrable
theories in four dimensions are known (for a review see e.g. \cite{fourd})
much more knowledge has been obtained so far for two dimensional
theories, and it is thus this class of models where my investigations
focus.
In the particle picture the interaction is encoded into the scattering
matrix. The asymptotic states are described by a linear superposition
of free one-particle states $\vert Z_\epsilon (\beta)\rangle$, which are
characterised by the particle species $\epsilon$ and their momentum,
parametrised as $p^{(0)} =m \cosh \beta$, $p^{(1)} =m \sinh \beta$
($m$ denotes the mass and $\beta$ the rapidity). They are related
through the $S$-matrix as \begin{equation} \vert Z_{\epsilon_1}(\beta_1) \dots
Z_{\epsilon_n}(\beta_n) \rangle_{in} =
S_{\epsilon_1\dots\epsilon_n}^{\epsilon'_1\dots\epsilon'_m}(\beta_1,\dots,\beta_n \vert
\beta_1'\dots\beta'_m ) \, \vert Z_{\epsilon'_1}(\beta'_1) \dots
Z_{\epsilon'_m}(\beta'_m) \rangle_{out} \hspace{0.2in} .\end{equation} On the other hand the local
description of a theory consists of the space of local operators
${\cal A} = \{ {\cal O}_i \} $ and the set of multi-point correlation
functions of them, $$ \langle 0\vert {\cal O}_1(x_1 ) \dots {\cal O}_n(x_n ) \vert
0 \rangle \hspace{0.2in} .$$ The two description are linked through the
Lehmann-Symanzik-Zimmermann reduction, since the particles can be
obtained as asymptotic limits of the local fields. Another connection
is given through the form factors. Consider an arbitrary two-point
correlation function
\[
G_{ij}(x)\,=\,<{\cal O}_i(x)\,{\cal O}_j(y)> \hspace{0.2in} ,
\]
of hermitian operators. Inserting the Identity between the two
operators and expanding it into the base of asymptotic states,
it can be expressed as an infinite series
over multi-particle intermediate states,
\begin{equation}
\langle {\cal O}_i(x)\,{\cal O}_j(y)\rangle\,=\label{correlation}\end{equation}
$$ \sum_{n=0}^{\infty}
\int \frac{d\beta_1\ldots d\beta_n}{ (2\pi)^n}
<0|{\cal O}_i(x)|Z_{\epsilon_1}(\beta_1),\ldots,Z_{\epsilon_n}(\beta_n)>_{\rm in}{}^{\rm
in} <Z_{\epsilon_1}(\beta_1),\ldots,Z_{\epsilon_n}(\beta_n)|{\cal O}_j(y)|0> $$
The matrix elements
$$<0|{\cal O}_i(0)|Z_{\epsilon_1}(\beta_1),\ldots,Z_{\epsilon_n}(\beta_n)> =
F_{\epsilon_1\dots\epsilon_n}(\beta_1,\dots,\beta_n)$$
are called form factors and in the following I will try to explain how
they can be used in order to establish a link between the local
description and the particle picture of a theory.
\subsection*{\large\bf 2 \phantom{11}
Form Factors and the Space of Local Operators}
If one considers two dimensional integrable theories many simplifying
properties occur which allow to calculate many dynamical quantities
exactly. The most remarkable fact is the factorisation of the
$S$-matrix, which determines a general scattering process as a product
of two-particle scattering amplitudes. Further these two-particle
$S$-matrices are pure phase-shifts, that is the incoming and outgoing
momenta are the same. This simplification allows to calculate the
$S$-matrix exactly (see {\em e.g.} P. Kulish's lectures in these proceedings).
Also the form-factors can be determined exactly for integrable two
dimensional systems. They obey a set of constraint equations,
originating from fundamental principles of quantum theory, such as
unitarity, analyticity, relativistic covariance and locality
\cite{Karowski,nankai}.
The important fact is that the $S$-matrix is the only dynamical
information needed.
In the following I will discuss just two examples of form-factor
equations, in order to determine their overall structure, and also to
explain the solution techniques.
Since the theories here considered are defined in only one space dimension, a
scattering process can be viewed as to interchange two particles on the
real line,
\begin{equation}
Z_{\epsilon_1}(\beta_1) Z_{\epsilon_2}(\beta_2) =
S_{\epsilon_1\epsilon_2}(\beta_1-\beta_2)
Z_{\epsilon_2}(\beta_2) Z_{\epsilon_1}(\beta_1)
\hspace{0.2in} .\end{equation}
This exchange property will lead to a constraint equation for the
form-factors
\begin{eqnarray}
& & F^{\cal O}_{\epsilon_1\dots\epsilon_i\epsilon_{i+1}\dots\epsilon_n}
(\beta_1,\dots,\beta_i,\beta_{i+1},\dots\beta_n) =\nonumber\\ &
&S_{\epsilon_i\epsilon_{i+1}}(\beta_i-\beta_{i+1})\,
F^{\cal O}_{\epsilon_1\dots\epsilon_{i+1}\epsilon_{i}\dots\epsilon_n}
(\beta_1,\dots,\beta_{i+1},\beta_{i},\dots\beta_n) \hspace{0.2in} .
\label{wat1}\end{eqnarray}
Another constraint equation derives from the bound state structure of
the theory under consideration. If particles $Z_i$, $Z_j$ form a bound
state $Z_k$, the corresponding
two-particle scattering amplitude
exhibits a pole at $\beta = iu_{ij}^{k}$ with the residue
\begin{equation}
-i \lim_{\beta \rightarrow iu_{ij}^{k}}(\beta-iu_{ij}^{k})
S_{ij}(\beta) =(\Gamma_{ij}^{k})^2 \hspace{0.2in} ;
\label{residueS}\end{equation}
$\Gamma_{ij}^{k}$ is the three--particle on--shell vertex.
Corresponding to this bound state the form--factor exhibits a pole
with the
residue $$ -i \lim_{\beta'\rightarrow \beta}(\beta'-\beta)
F^{\cal O}_{\epsilon_1\dots ij\dots\epsilon_n} (\beta_1,\dots,\beta'+i(\pi-
u_{ik}^{j}),\beta-i(\pi- u_{jk}^
{i}),\dots,\beta_{n}) = $$
\begin{equation} =\Gamma_{ij}^{k} F^{\cal O}_{\epsilon_1\dots k\dots\epsilon_n}
(\beta_1,\dots,\beta,\dots,\beta_{n} )
\hspace{0.2in} . \label{bounds}\end{equation}
As mentioned before, (\ref{wat1}) and (\ref{bounds}) are only two
examples of form-factor equations. Nevertheless they are two exponents
of the two categories of the constraint equations:
\begin{enumerate}
\item Equations with fixed n ( e.g. (\ref{wat1})): they involve form
factors with the same number of particles on both sides of the
equation
\item Recursive equations ( e.g.(\ref{bounds})): They link form
factors with different particle numbers with each other - in the
example above $n+1$ particle form factors to $n$ particle form
factors.
\end{enumerate}
For theories with scalar particles there exists a well established
solution method \cite{Karowski}. It consists of the ansatz
\begin{equation}
F_{\epsilon_1\dots\epsilon_n}(\beta_1,\dots,\beta_n)
=Q_{\epsilon_1\dots\epsilon_n}(e^{\beta_1},\dots,e^{\beta_n})
\prod_{i<j}^n F_{\epsilon_i\epsilon_j} (\beta_i -\beta_j)\hspace{0.2in} . \end{equation}
The two-particle form factors $F_{\epsilon_i\epsilon_j}$
can be calculated easily from the form
factor equations. The product term satisfies all equations of the
first type (with fixed particle number) and also is designed in order
to have the correct pole structure of an $n$-particle form factor.
Through this parametrisation the form-factor equations are reduced to
recursive relations for the functions
$Q_{\epsilon_1\dots\epsilon_n}(e^{\beta_1},\dots,e^{\beta_n})$. Further properties
of these functions can be extracted from the form factor equations:
they are homogeneous polynomials, symmetric in repeated indices with a total
degree in its arguments fixed by relativistic covariance and the
partial degree determined from the recursion relations.
This information is sufficient for simple models to obtain explicit
expressions for the form factors. In all cases though it is possible
to determine the space of local operators by just considering these
general properties of the functions $Q$ \cite{mywork}. Each linear
independent solution of the form factor equations corresponds to an
independent local operator. Therefore the space of local operators
can be determined by counting the number of independent solutions of
the form factor equations. This can be done due to the
property of the recursion relations, that
the dimension of the solution space at level $n$ is the sum of the
dimension of the solution space at level $n-1$ and of the dimension of
the kernel of the recursion relation at level $n$. Symbolically this
can be written as
$$dim(Q_n) = dim(Q_{n-1}) + dim ( {\cal K}_n)\hspace{0.2in} .$$
\begin{figure}[hbt]
\begin{center}
\begin{picture}(330,200)(-30,-30)
\thicklines
\put(0,0){\circle*{8}}
\put(0,0){\line(1,0){200}}
\put(0,0){\line(0,1){180}}
\put(0,-15){\makebox(0,0){${\cal M}_{3}$}}
\put(0,-30){\makebox(0,0){$h,\bar{h} =\{ 0,\frac 1{16},\frac 12\}$}}
\put(200,-15){\makebox(0,0){$H=0$, $T>T_c$}}
\put(100,35){Free fermion theory $S=1$}
\put(100,15){Local operators: $1 \, (0,0)$, $\psi \, (0,
\frac 12)$, $\bar{\psi} \, (
\frac 12,0)$, $\epsilon \, (\frac 12, \frac 12)$}
\put(-25,170){\makebox(0,0){$H \neq 0$}}
\put(-25,150){\makebox(0,0){ $T=T_c$}}
\put(10,160){$E_8$-scattering theory}
\put(10,140){Local operators: $1 \, (0,0)$, $\sigma \, (\frac 1{16},
\frac 1{16})$, $\epsilon \, (\frac 12 ,\frac 12)$}
\end{picture}
\caption{The critical Ising model and its integrable perturbations}
\label{fig2}
\end{center}
\end{figure}
An interesting application of this counting method is in perturbed
conformal field theories. As an example consider the Ising model. Its
point of second order phase transition is described by a conformal
field theory ( the minimal model ${\cal M}_{3}$ ) and therefore the space of
local operators at the critical point is determined by Virasoro
irreducible representations. The critical point admits two relevant
perturbations which are integrable. The perturbation with the
conformal operator with conformal weight $h=\frac 12$ drives the model
into the regime $T>T_c$ and $H=0$. It is described by a free fermion
theory. The other perturbation with the operator $h=\frac 1{16}$
corresponds to the Ising model with $H \neq 0$ and $T=T_c$ and is
described by a scattering theory containing 8 scalar massive particles
\cite{Zam}. Virasoro symmetry is obviously broken by both perturbations
and it is therefore an interesting problem to determine the space of
operators for these theories.
The situation is summarised in figure \ref{fig2}. For both
perturbations it is possible to determine the space of local
operators. It is given in terms of characters of the minimal model
$M_{3}$. This is a quite remarkable fact, since conformal symmetry is
explicitly broken by the perturbation. Further, since the thermal
perturbation is described by a free fermion theory, the local
operators are just the fermions $(\psi, \bar{\psi})$ the identity
operator ($1$) and the energy density $\epsilon$. Also the spin operator
($\sigma$) and the disorder field ($\mu$) can be analysed by the
counting method, but they are semi-local operators with respect to the
asymptotic states. The magnetic perturbation breaks the $Z_2$ symmetry
of the model and only scalar operators appear in this perturbation,
namely the identity ($1$), the energy density ($\epsilon$) and the spin
operator ($\sigma$).
A general feature of the counting method is that the space of
operators is determined by fermionic sum expressions. Such expressions
also appear in the analysis of corner transfer matrices \cite{kedem}
and spinon conformal field theories \cite{spinon}. It would be
interesting to establish a more direct connection between these
methods and the form factor approach.
Finally note that the above example only constitutes a simple
application of the counting method. It can be generalised to many
other systems, including models with a massless spectrum and/or
bounderies.
|
\section{Introduction}
Quantum localization is a generalization Duistermaat-Heckman theorem
\cite{dh} to infinite dimensions. This theorem states that if the
Hamiltonian $H$ generates a global circle, or, more generally a torus
action in the phase space $\Gamma$ then the canonical partition
function is given exactly by the saddle-point approximation around the
critical points of $H$. Extensions to calculation of quantum
mechanical partition functions using phase space path integrals have
been represented in {e.g.} \cite{us}.
We shall first consider basic ideas of localization. Then we shall
carefully regularize the pertinent functional determinants arising
from the path integrals. There is an ambiguity in choosing the
regularization scheme because of the spectral asymmetry of first order
differential operators. Therefore, the result depends on the
regularization as in the case of quantum mechanical anomalies.
Finally, we are going to apply our localization to the quantization of
the simple harmonic oscillator and to the evaluation of the Weyl
character of spin. We shall notice that different regularizations give
different energy spectra for the harmonic oscillator. We also show
that the continuum coherent state path integral yields directly the
correct character for spin if we choose an appropriate
regularization. In particular, we will consider the relation of
character formulae to the Borel-Weil theory which constructs the
irreducible representations of a Lie group as holomorphic
functions. Using this theory we relate the character formulae to the
equivariant index of the Dolbeault complex. The result is that the
path integral yields directly the correct character without an
explicit Weyl shift of the highest weight.
\\
\section{Localization of Phase Space Path Integrals}
We are interested in exact evaluation of phase space path integrals
(partition functions) of the form
\begin{equation}\label{Zcanonical}
Z(T) = \int_{L \Gamma} {{\cal D}}x \ {\rm Pf \ } \Vert \omega_{ab}(x) \Vert \exp \left( \ii
\int_0^T \dd t \left[ \vartheta_a {\dot{x}}^a - H(x) \right] \right) \;.
\end{equation}
where $\{x^a\} $ are local coordinates in $\Gamma$, ${\rm Pf \ } \Vert
\omega_{ab} \Vert$ is the Liouville measure factor, $\vartheta_a$ the
symplectic potential and $\omega_{ab} = {\partial}_a
\vartheta_b - {\partial}_b \vartheta_a$. The integration is performed over the loop
space $L\Gamma$ consisting of the phase space loops. The integrability
condition \cite{woodhouse} requires that $$\int_{\Sigma} \omega = 2
\pi n$$ for any 2-cycle $\Sigma$ in $\Gamma$ so that the path integral
is single valued. We introduce anticommuting variables $\psi^a$ to
write ${\rm Pf \ }
\Vert \omega_{ab} \Vert $ as a path integral
\begin{equation}\label{ZSUSY}
Z(T) = \int {{\cal D}}x {{\cal D}}\psi \ \exp \left( \ii \int_0^T \dd t
\left[\vartheta_a {\dot{x}}^a - H(x) + \half \psi^a
\omega_{ab} \psi^b \right]
\right) \; .
\end{equation}
The boundary conditions are periodic also for the fermions, since they
are a realization of the differentials of the bosonic coordinates.
We interpret the path integral (\ref{ZSUSY}) in terms of equivariant
cohomology in $L\Gamma$. From the bosonic part of the action we get a
Hamiltonian vector field in $L\Gamma$
\begin{equation}
{\chi}_S^a = \dot{x}^a - \omega^{ab} {\partial}_b H
$$
whose zeroes define the Hamilton's equations. The equivariant
exterior derivative in $L\Gamma$ is
\begin{equation}
\dd_S = \dd + \iota_S \;,
$$
where $\iota_S$ denotes the contraction along the vector field
$\chi_S$.
The square of $\dd_S$ is the loop space Lie derivative
$$%
{\cal L}_S = \dd \iota_S + \iota_S \dd \sim {\dd \over \dd t} - {\cal L}_H \;.
$$
The action $S_{\rm B} + S_{\rm F}$ is supersymmetric under the
infinitesimal loop space supersymmetry transformations that are
parametrized by a gauge fermion $\delta \Psi$:
\begin{eqnarray}\label{SUSY}
x^a \rightarrow x^a + \delta \Psi \dd_S x^a & = & x^a + \delta \Psi
\psi^a , \cr \psi^a \rightarrow \psi^a + \delta \Psi \dd_S \psi^a & = & \psi^a +
\delta \Psi \chi_S^a \;.
\end{eqnarray}
This implies that that the action is equivariantly closed: $$\dd_S
(S_{\rm B} + S_{\rm F}) =0 \;.$$ By an analogue of Fradkin-Vilkovisky
theorem \cite{bfv} one can show that the path integral remains
intact if we modify the action by $S \rightarrow S+\dd_S \Psi$ where $\Psi$
satisfies the Lie derivative condition
\begin{equation}\label{EquivCohomology}
\dd_S^2 \Psi = {{\cal L}}_S \Psi =0 \;.
\end{equation}
In the limit $\lambda \rightarrow 0$ the path integral
\begin{equation}\label{Z_lambda}
Z_{\lambda} (T) = {\cal} \int {{\cal D}}x^a {{\cal D}}\psi^a \
\exp \left( \ii \int_0^T \dd t \left[\vartheta_a {\dot{x}}^a - H(x^a)
+ \half \psi^a \omega_{ab} \psi^b + \lambda \dd_S \Psi \right] \right)
\end{equation}
reduces to (\ref{ZSUSY}) and $\lambda \rightarrow \infty$ gives localization.
To construct a gauge fermion $\Psi$ we need a metric $g$ in the phase
space. The loop space Lie derivative condition (\ref{EquivCohomology}) is
satisfied if the metric $g$ in $\Gamma$ is invariant under the
Hamiltonian action of $H$
\begin{equation}\label{InvariantMetric}
{{\cal L}}_H g = 0 \;,
\end{equation}
which means that $\chi_H$ is a Killing vector field. This is a very
restrictive condition for the Hamiltonian: it must generate a global
U(1)-action in $\Gamma$. We can choose any metric which satisfies the
condition (\ref{InvariantMetric}) and average it over the group
action.
We will consider the following selections for the gauge fermion:
\begin{equation}
\Psi_1 = {1 \over 2} g_{ab} \dot{x}^a \psi^b
$$
gives localization to the constant modes which are points of the
manifold,
\begin{equation}
\Psi_2 = {1 \over 2} g_{ab} \chi_H^a \psi^b
$$
to the zeroes of $\chi_H$ which we assume to be nondegenerate and
isolated and
\begin{equation}
\Psi_3 = {1 \over 2} g_{ab} \chi_S^a \psi^b
$$
to the classical trajectories. For simplicity we use subscripts 1,2,3
in the actions and partition functions corresponding to the gauge
fermions $\Psi_{1,2,3}$. The actions become
\begin{eqnarray}
S_1 & = & \int_0^T \dd t \ \left[ \left( \vartheta_a - {\lambda \over
2} g_{ab} {\chi}_H^b \right) \dot{x}^a -H + {\lambda \over 2} g_{ab}
\dot{x}^a \dot{x}^b + {\lambda \over 2} \psi^a \left( g_{ab}
{{\partial}}_t + \dot{x}^c g_{bd} {\Gamma}_{ac}^d \right) \psi^b + {1 \over
2} \psi^a \omega_{ab} \psi^b \right] \;, \cr S_2 & = & \int_0^T \dd t
\ \left[ \vartheta_a \dot{x}^a - H + {\lambda \over 2} g_{ab}
{\chi}_H^a \left( \dot{x}^b - \chi_H^b \right) + {\lambda \over 2}
\psi^a {\partial}_{a }\left( g_{cb} \chi_H^c \right) \psi^b + {1 \over 2}
\psi^a \omega_{ab} \psi^b \right] \;, \cr S_3 &= & \int_0^T \dd t
\left[\vartheta_a \dot{x}^a - H +{\lambda
\over 2} g_{ab} \chi_S^a \chi_S^b + {\lambda \over 2} \psi^a {\partial}_a (g_{cb}
\chi_S^c) \psi^b + {1 \over 2} \psi^a \omega_{ab} \psi^b \right] \;.
\end{eqnarray}
To take the limit $\lambda \rightarrow \infty$ in path integrals we make the
decomposition to constant modes $x_0^a, \psi_0^a$ and to non-constant
modes $x^a_t , \psi^a_t$ and scale the non-constant modes by $1/
\sqrt{\lambda} $
\begin{eqnarray}\label{scaling}
x^a(t) & = & x_0^a + {1 \over {\sqrt{\lambda}}}x^a_t , \cr
\psi^a(t) & = &\psi^a_0 + {1 \over {\sqrt{\lambda}}} \psi^a_t .
\end{eqnarray}
The Jacobi determinant is unity. An expansion to a quadratic order
around the constant modes and the limit $\lambda \rightarrow \infty$ gives a
Gaussian path integral
\begin{equation}\label{xdot-Gaussian}
Z_1 = \int \dd x_0^a \dd \psi_0^a \ \exp \left[ -\ii T \left( H-
\half \psi^a_0 \omega_{ab} \psi^b_0 \right) \right] Z_{\rm fl,1} (T)
\end{equation}
where the fluctuation path integral $Z_{\rm fl}(T) $ is a product of
fermionic and bosonic parts:
\begin{eqnarray}\label{x-Fluctutation}
Z_{\rm F ,1} &=& \int \prod_t \dd \psi^a_t \ \exp \left\{-{ \ii \over
2}\int_0^T \dd t \ \psi^a_t g_{ab} {{\partial}}_t \psi^b_t \right\} \; , \cr
Z_{\rm B,1} & = & \int \prod_t \dd x^a_t \ \exp \left\{ { \ii \over
2} \int_0^T \dd t \ x^a_t \left[ {\cal R}_{ab} {{\partial}}_t -g_{ab} {{\partial}}_t^2
\right] x^b_t \right\} \; .
\end{eqnarray}
Here $${\cal R}_{ab} = R_{ab} + \tilde{\Omega}_{ab}$$ is the equivariant
curvature with $R_{ab} $ the Riemannian curvature 2-form and $$
\tilde{{\Omega}}_{ab} = \half \left[ {\nabla}_{b}(g_{ac} \chi_H^c) -
{\nabla}_{a}(g_{bc} \chi_H^c) \right] $$ the momentum map
\cite{arnold} corresponding to ${\chi}_H$, $\nabla$ being the
covariant derivative. $H$, $\omega$, $g$ and ${{\cal R}}$ are evaluated at
the constant modes. The path integral $Z_2$ is given by a sum over the
critical points $\{x_i \}$ of the Hamiltonian:
\begin{equation}\label{FPICritical}
Z_2 = \sum_{ x_i } \ { {\exp[-\ii TH] } \over { {\rm Pf \ } \Vert {\partial}_a
\chi_H^b \Vert } } Z_{\rm fl,2}(T) \;.
\end{equation}
$Z_{\rm fl,2}$ is also a product of fermionic and bosonic parts:
\begin{eqnarray}\label{chi-Fluctutation}
Z_{\rm F,2}(T) & = & \int \prod_t \dd \psi^a_t \ \exp \left\{ {\ii \over 2}
\int_0^T \dd t \ \psi^a_t {\partial}_a (g_{bc} \chi_H^c) \psi^b_t \right\} \;, \cr
Z_{\rm B,2}(T) & =& \int \prod_t \dd x ^a_t \ \exp \left\{ {\ii \over 2}
\int_0^T \dd t \ x^a_t {\partial}_a (g_{bc} \chi_H^c) (\delta^b_d {\partial}_t -
{\partial}_d \chi^b_H ) x^d_t \right\} \;.
\end{eqnarray}
Here $g$ and $\chi_H$ are again evaluated at the constant modes.
Finally, the path integral $Z_3$ reduces to a sum over the
$T$-periodic classical trajectories
\begin{equation} \label{Classical}
Z_3 = \sum_{x_{\rm {cl}}} { 1 \over {\rm Pf \ } \Vert \delta_b^a
{\partial}_t - {\partial}_b \chi_H^a
\Vert } \exp[\ii S_{{\rm cl }}] \;.
\end{equation}
In practice, it is usually a highly non-trivial problem to find the
$T$-periodic classical trajectories of a dynamical system
\cite{ArnoldConj}.
\\
\section{Regularization of Fluctuation Path Integrals}
In the following all the path integrals and determinants are evaluated
over periodic configurations for both the bosonic and fermionic
degrees of freedom. The primes will denote that we exclude the
constant modes. In real polarization the fluctuation parts in
$Z_{1,2}$ become
\begin{eqnarray}\label{xdot-Z_fl}
Z_{\rm fl ,1} & = & {1 \over \sqrt{{\rm Det\ } ' \Vert \delta_b^a {\partial}_t -
{{\cal R}}_b^a \Vert }} \;,
\cr
Z_{\rm fl,2} &= & {1 \over \sqrt{{\rm Det\ } ' \Vert \delta_b^a {\partial}_t -
{\partial}_b \chi_H^a \Vert}} \;.
\end{eqnarray}
It is quite important to notice that in the reduced determinants one
index is covariant and another contravariant.
In K\"ahler polarization the fluctuations parts are, using the
additional symmetries of the metric and the Riemann curvature tensor
\cite{nakahara},
\begin{eqnarray}
Z_{\rm fl,1} & = & {1 \over {\rm Det\ }' \Vert \delta_a^b {\partial}_t -{\cal R}_a^b \Vert
}\;, \cr
Z_{\rm fl,2} & = & {1 \over {\rm Det\ } ' \Vert \delta_a^b {\partial}_t -
{\partial}_a \chi_H^b \Vert } \;.
\end{eqnarray}
These determinants are taken over the holomorphic indices. By this we
mean the following: The relevant matrices can be block diagonalized
\begin{equation}
A = {\rm diag} \ (A_1 ,A_2 ,... ,A_N)
$$
with blocks
\begin{equation}
A_k = \pmatrix{
a^+_k & 0 \cr
0 & a^-_k \cr
} \equiv \pmatrix{
a_k & 0 \cr
0 & - a_k \cr
} \;.
$$
The symbols $a^+_k$ and $a^-_k$ denote the holomorphic and
antiholomorphic eigenvalues of $A$, and we consider only the
eigenvalues corresponding to the holomorphic indices to the
determinant.
We have to choose a regularization scheme for the determinants. A
standard method is to apply $\zeta$- and $\eta$-functions. The
$\zeta$-function regularization does not directly apply to first-order
operators because they have an infinite number of negative
eigenvalues. To take them into account we define the $\eta$-function
for the first-order operator $B$ by
\begin{equation}
\eta_B (s) = \sum_{b_{n} \neq 0}
{{\rm sign\ } } (b_{n}) \abs{b_{n}}^{-s} + {\rm dim \ Ker \ } B = {1 \over
{\Gamma\left({s+1} \over 2 \right)}}
\int_0^{\infty} \dd t \ t^{(s-1)/2} \ {\rm Tr\ } [B \exp(-t B^2)] \;.
$$
Analytical continuation to $s=0$ gives the
Atiyah-Patodi-Singer $\eta$-invariant \cite{egh} of $B$ that measures
the spectral asymmetry of $B$ and specifies the phase of
${\rm Det\ }(B)$. The absolute value $\abs{{\rm Det\ } (B) }$ is regularized using
the formula
$$%
\abs{{\rm Det\ } (B)}
= +\sqrt{{\rm Det\ } (B^2)} = + \exp\left[ - \half
\zeta_{B^2} ' (0)\right] \;.
$$
In real polarization we have to evaluate the square root of a
determinant of the antisymmetric operator $B = {\partial}_t - A $ where $A$
is an antisymmetric matrix. In our case $A$ is $\Vert {\cal R}_a^b \Vert$
or $\Vert {\partial}_a \chi^b_H \Vert$. By determining the spectrum of $B$
and applying $\zeta$-function regularization we get, up to an
inessential numerical normalization, the result
\begin{equation}\label{Agenus}
{1 \over {\sqrt{{\rm Det\ }'
\left( {\partial}_t - A \right) }} }
= \prod_{n= 1}^N { \abs{a_n /2 \over {\sin (a_n T/2 )}}}= {1 \over
T^N} \hat{A}(TA) \;,
\end{equation}
where we have defined the function of the matrix $X$ $$\hat{A}(X) =
\prod_n {x_n/2 \over \sin(x_n/2)} \;$$ where $x_n$ are the skew-eigenvalues
of $X$. The result is non-negative since the negative and positive
skew-eigenvalues appear in pairs. Therefore there is no ambiguity with
the spectral asymmetry.
Now we consider the determinants in K\"ahler polarization. It is
sufficient to consider the determinant of a block. Earlier we noticed
that the fluctuation path integrals reduce to the determinant of the
operator $B = \ii {\partial}_t -a$. The functional Pfaffian in
(\ref{Classical}) is also similar to this determinant. To regularize
$${\rm Det\ }'(B) = \prod_{n \neq 0} \left( {2\pi n \over T} - a
\right)$$ properly we have to take into account that $B$ has and
infinite number of negative eigenvalues. Thus there is a problem with
the spectral asymmetry.
Therefore, we have to choose a regularization prescription which has a
relation to quantum mechanical anomalies. In the regularization of the
determinants it is not possible to maintain all the symmetries that
are present in the classical theory. For example, Elitzur
et al. \cite{elirab} considered the corresponding fermionic problem
with antiperiodic boundary conditions. They evaluated the quantum
mechanical partition function for a Dirac fermion in an external gauge
field $A(t)$ in $0+1$-dimensions
\begin{equation}
Z(T) = \int {\cal D} \bar{\psi} {\cal D} \psi \exp\left[ \ii \int_0^T \dd t \
\bar{\psi} \left ( \ii {\partial}_t -a \right) \psi \right] = {\rm Det\ } ( \ii
{\partial}_t -a ) \;.
\end{equation}
where, because of the gauge invariance of the action only the constant
mode $a$ of $A(t)$ contributes. The classical action has both the
invariance under large gauge transformations
\begin{eqnarray}
a &\rightarrow& a + n 2 \pi /T \cr
\psi &\rightarrow& \psi
\end{eqnarray}
and the charge conjugation invariance
\begin{eqnarray}
a &\leftrightarrow& -a \cr
\psi &\leftrightarrow &-\bar{\psi} \;.
\end{eqnarray}
However, when regularizing the determinant one has to choose which
symmetry one wants to maintain, which leads to a global anomaly. Here
we have an analogous situation. It is not {\em a priori} clear what
the result of the regularization should be, and there is a genuine
ambiguity.
Since the zeroes of the determinant are at $aT = 2\pi n $, the
determinant must be proportional to $$ {\sin( a T/ 2 )
\over a/2} \;. $$ The proportionality factor can be any function
without zeroes that is, the exponent function. The determinant is
therefore, up to an irrelevant constant,$${\rm Det\ }(\ii {\partial}_t -a ) = {
\sin( a T/ 2) \over a/2} \exp\left(\ii \phi aT \right) $$
with a phase $\phi$ whose natural values turn out to be $0$ and $\pm
1/2$ since they yield the (anti)symmetries of the product under $a
\leftrightarrow -a $ and $a \rightarrow a + 2\pi n /T$. However, there is
a minor subtlety: in our localization formulae the zero modes are
absent and this destroys these symmetries. Nevertheless, we may still
consider the residual symmetries. The choice $\phi=0$ corresponds to
neglecting the spectral asymmetry and choosing the (anti)symmetry $a
\rightarrow -a$ to be unbroken. In this regularization scheme the inverse
determinant is simply
$$%
{ 1 \over {\rm Det\ }'( {\partial}_t - A) } = {1 \over T^N} \hat{A}(T A)
$$
This is the result that usually appears in literature. However, there
is another possibility. The values $\phi = \pm 1/2$ correspond to
maintaining the symmetry $a \rightarrow a + 2\pi n /T$ and taking into account
the spectral asymmetry by the Atiyah-Patodi-Singer
$\eta$-invariant. This yields \begin{eqnarray}\label{ComplexDet} {1 \over {{{\rm Det\ }
'}({\partial}_t - A)}} &=& \prod_{n= 1}^N {{a_n/2} \over {\sin \left( a_n T
/ 2 \right)}} \exp \left( \ii a_n T/2 \right) = {1 \over T^N} {{\rm Td } }(T
A)
\end{eqnarray}
where we have defined the following function of the matrix $X$
$$
{\rm Td } (X) = \prod_n { x_n /2 \over \sin(x_n /2) } \e^{\ii x_n /2} \;.
$$
We take only the eigenvalues corresponding to the holomorphic indices
to the determinant.
Let us now write down the resulting localization formulae. The
localization to constant modes yields the expression
\begin{equation}\label{ConstantModes}
Z_1 (T) = {1 \over T^N} \int \dd x_0^a \dd \psi_0^a \
{{\rm Ch } }\left[ -\ii T(H- \omega)
\right] \cases{
\hat{A}({T {\cal R} }) \cr {\rm Td } ({T {\cal R} }) } \;.
\end{equation}
We have defined equivariant generalizations \cite{bgv} of the
conventional characteristic classes known as equivariant $\hat{A}$-
and Todd-genus, and identified the exponential with the equivariant
Chern class. When $H=0$ they reduce to the conventional
characteristic classes and the result is a topological invariant. The
localization to the critical points $\{ x_i\}$ of the
Hamiltonian gives the result
\begin{eqnarray}\label{CriticalPoints}
Z_2 (T) = {1 \over T^N} \sum_{x_i} {\exp(-\ii TH ) \over {\rm Pf \ }({\partial}
\chi_H^{~} )}
\cases{
\hat{A}(T {\partial} \chi_H^{~} ) \cr {\rm Td } (T {\partial} \chi_H^{~} ) } \;.
\end{eqnarray}
We must use local coordinates in the evaluation of the determinants
when localizing to the critical points of the Hamiltonian.
Finally, the localization to $T$-periodic classical trajectories yields
\begin{eqnarray}\label{WKB}
Z_3 (T) = {1 \over T^N} \sum_{x_{\rm cl}} {\exp(\ii S_{\rm cl} )
\cases{
\hat{A}(T {\partial} \chi_H^{~} ) \cr {\rm Td } (T {\partial} \chi_H^{~} ) }} \;.
\end{eqnarray}
\section{ Harmonic Oscillator}
Now we show that the localization formulae yield the correct partition
function for the harmonic oscillator in a flat phase space. The path
integral for it is Gaussian and in principle there is no reason to
apply localization to it. However, it is reasonable to check by some
simple examples that our assumptions and derivations are valid. In
particular, we will show that the choice of the metric in the phase
space is not relevant, contrary to claims in literature \cite{dlr}. It
is also illustrative to consider the significance of the
regularization schemes we have used.
In real polarization the Hamiltonian is $H= \half \left( p^2 +q^2
\right)$ and the symplectic 2-form is $\dd q \wedge \dd p . $ The
coherent state representation (K\"ahler polarization) requires some
further investigation, since we have to fix an operator ordering
prescription. In terms of creation and annihilation operators the
normal and symmetric ordered Hamiltonians are, respectively,
\begin{eqnarray}\label{Ordering}
H_{\rm n} & =& : {1 \over 2} \left( a^{\dagger}a + a a^{\dagger}
\right) : = a^{\dagger}a \;, \cr H_{\rm s} &=& a^{\dagger} a + \half
\; .
\end{eqnarray}
The symmetric ordered Hamiltonian has an explicit zero point energy
$E_0 = 1/2.$
To apply the localization formulae we must choose a metric in the
phase space and calculate the equivariant curvature and the
derivatives of the Hamiltonian vector field. If the Lie-derivative
condition ${\cal L}_H g = 0$ is satisfied we can start from an any smooth
metric in the phase space and average it. So we may choose a constant
metric $$g = \pmatrix{ 1 & 0 \cr 0 & 1 \cr }.$$ The non-zero
components of the equivariant curvature are
\begin{equation}
{\cal R}^p_q = - {\cal R}^q_p = 1 \;.
$$
The localization formula (\ref{ConstantModes})
yields the result
\b
Z(T) & = & \int \dd p \ \dd q \ \dd \psi^q \ \dd \psi^p \ \exp \left[
-\ii T \left( {1 \over 2} p^2 + \half q^2 - \psi^p \psi^q \right)
\right ] {1 \over 2\sin ( T/2)} \cr & \sim & {1 \over 2 \sin( T/2) }
= \sum_{n=0}^{\infty} \exp [\ii (n + 1/2 )T ]
\end{eqnarray}
which is the correct partition function with the zero-point energy
$E_0 = 1 /2 $.
Let us now digress slightly to discuss the result. In \cite{dlr}
Dykstra, Lykken and Reiten analyzed this problem and they noticed a
dependence on the metric. What they did not notice was that the index
structure of the equivariant curvature is ${\cal R}_a^b$ and therefore it
is invariant under global scalings of the metric. Furthermore, they
used a metric which in polar coordinates near the origin behaves like
$$
\dd s^2 = \dd r^2 + c r^2 \dd \phi^2 \;.
$$
This is a metric on a cone, not on a plane when $c \neq 1$ and is not
smooth, nor even continuous at the origin. Therefore it is not
surprising that their energy levels depend on the parameter $c$ which
represents the tip angle of the cone. From this we indeed see that we
cannot choose an arbitrary invariant metric, since it has to respect
the topology of the phase space.
The localization to the critical points of the Hamiltonian
(\ref{CriticalPoints}) yields also the correct result. The only zero
of $\chi_H^{~} $ is the origin of the phase space, which gives
\begin{equation}
Z(T) = {1 \over {\rm Pf \ } \pmatrix{ 0 & 1 \cr -1 & 0 \cr }} { {1 /2}
\over {\sin ( T/2)}} = {1 \over { 2 \sin(T/2)} } \;.
$$
If we want to apply the localization the classical trajectories we
must classify all the $T$-periodic classical trajectories. If $T \neq
2 \pi n$ the problem reduces to the localization to the critical
points of the Hamiltonian. However, if $T = 2 \pi n $ the zeroes of
$\chi_S$ are not isolated and we have to use a degenerate version of
the localization formula to the classical trajectories \cite{Palo}
We now consider the harmonic oscillator in the K\"ahler
polarization. We will only discuss the localization formulae to
constant modes. The reasoning is similar with other formulae. There
are four cases to consider: the localizations with the $\hat{A}$-genus
and Todd-genus using two different orderings. We will only list the
spectra we obtain. The use of $\hat{A}$-genus yields the spectra $E_n
= n+1/2$ (normal ordering) and $E_n = n+1$ (symmetric ordering). The
Todd-genus gives the results $E_n = n$ (normal ordering) and $E_n =
n+1/2$ (symmetric ordering). The first and fourth results have the
correct zero-point energy. From this example we see that to get
correct results from the path integral we do need some additional
information other than the classical action and boundary conditions:
we must choose a regularization scheme that gives physically correct
results.
\section{ Character for SU(2)}
We shall now use our localization formulae to derive the
Kirillov and Weyl character formulae for Lie groups \cite{bgv}. The
character formula for SU(2) has been widely discussed in literature
\cite{afs, stone, nielsen}. However, there has been some
controversy about the Weyl shift problem: the path integral usually
gives almost the correct character up to the substitution
$j \rightarrow j+1/2$. We show that the coherent state path integral and the
localization formulae with the Todd-genus directly yield the correct
character. In this calculation we use the continuum version of the
coherent state path integral and show that this also yields the
correct result, contrary to discussions in literature
\cite{funahashi}.
To motivate the use of Todd-genus we relate the character of a simple
Lie group $G$ in the highest weight representation $\lambda$ to the
index of the twisted Dolbeault complex on the coadjoint orbit $O_f$
\cite{perelomov} of the group. The Borel-Weil theory \cite{ stone, alsiwi}
constructs the irreducible representations of $G$ as holomorphic
sections of a line bundle $L$ that is associated to a principal bundle
$ G \rightarrow G/T_G \sim O_f$ where $T_G$ is the Cartan torus of $G$
\cite{alsiwi}. The holomorphic sections of this line bundle
(coherent states) form the basis for the irreducible
representation. The connection 1-form on $L$ is the symplectic
potential
$$%
\vartheta = {{\partial}{F} \over {\partial} z^k } \dd z^k - {{\partial}{F} \over {\partial}
\bar{z}^k} \dd \bar{z}^k \;
$$
where $F$ is the K\"ahler potential on $O_f$. It can be shown that
the twisted Dolbeault operator $\bar{{\partial}}_L = \bar{{\partial}} +
\vartheta_{\bar{z}}$ annihilates the normalized coherent states
$\ket{z}$ and therefore $\ket{z} \in H^{0,0} (O_f ,
L)$. If we can prove that all the other cohomology
groups are trivial, {e.g.} by Lichnerowicz vanishing theorem
\cite{bgv}, we conclude that the dimension of the highest weight
representation $R_{\lambda}$ is $\dim H^{0,0} (O_f, L)$.
Consequently, this is equal to the index of the twisted Dolbeault
complex. The Riemann-Roch-Hirzebruch index theorem relates this
analytical index to the topological invariant
\begin{equation}\label{TopDolbeault}
{{\rm ind\ }} \ \bar{{\partial}}_L = {\dim}\ R_{\lambda} = \int_{O_f}
{\rm Td}(O_f) \wedge {{\rm Ch } }(L) \;.
\end{equation}
Indeed, we notice that the localization formula (\ref{ConstantModes})
with $H=0$ represents this index provided we use the Todd-class. For
SU(2) we obtain the known result for the dimension of the
spin-$j$- representation
$$
\dim R_j = {\rm ind\ } \ \bar{{\partial}}_L = 2j +1 \;,
$$
This is the correct result without the explicit Weyl shift by the
Weyl vector $\rho = 1/2$.
We shall now use an equivariant version of the index theorem to derive
the character formulae. The character of an element in the Cartan
subalgebra is the partition function for the Hamiltonian $H$ that
represents it on $O_f$:
\begin{equation}\label{character}
\chi (\beta) = {\rm Str \ } \exp[- \ii T H ] \;.
\end{equation}
To make a relation to the Dolbeault index we write this as an
equivariant index (character index, G-index, Lefschetz number)
\cite{bgv}. One can show that the Laplacians $\bar{{\partial}}_L^{\dagger}
\bar{{\partial}}_L$ and $\bar{{\partial}}_L \bar{{\partial}}_L^{\dagger}$ have
equal non-zero eigenvalues. If all other comohomology classes except
$H^{0,0}$ are trivial, as we presume, $\bar{{\partial}}_L^{\dagger} $ does
not have zero modes. Consequently, we can write the trace as an
equivariant index:
\begin{eqnarray}
{\rm ind\ }_H (\bar{{\partial}}_L, T) &\equiv& \lim_{\beta \rightarrow \infty} {\rm Tr\ }
\e^{- \ii T H} (\e^{- \beta \bar{{\partial}}_L^{\dagger} \bar{{\partial}}_L} -
\e^{- \beta \bar{{\partial}}_L \bar{{\partial}}_L^{\dagger}} )\cr
& =& \lim_{\beta \rightarrow \infty}
{\rm Str \ } \exp[- \ii T H] \exp \left[- \beta \pmatrix{ \bar{{\partial}}_L^{\dagger}
\bar{{\partial}}_L & 0 \cr 0 & \bar{{\partial}}_L \bar{{\partial}}_L^{\dagger} } \right]
\end{eqnarray}
Only the zero modes contribute to the trace. The expression is also
independent of $\beta$. Therefore, in the limit $\beta \rightarrow 0$, all we
are left with are the zero modes of $\bar{{\partial}}_L$. Consequently, the
equivariant index is equal to the character
$$
{\rm ind\ }_H (\bar{{\partial}}_L, T) = {\rm Str \ } \exp[- \ii T H]
$$
Thus, the character is the equivariant index of the twisted Dolbeault
complex and therefore we choose the localization with the Todd-class.
To derive the character formulae we apply standard methods to write
${\rm Str \ } \exp[ - \ii T H]$ as a coherent state path integral of the form
(\ref{ZSUSY}). Since we can choose an invariant metric on a
coadjoint orbit \cite{perelomov} we can localize the path integral to
classical trajectories, to constant modes or to critical points of
the Hamiltonian. The two latter cases yield the Kirillov character
formula \cite{kirillov} ($2N$ is the dimension of the orbit)
\begin{equation}\label{Kirillov}
\chi (T) = {1 \over T^N} \int_{O_f} {{\rm Ch } }\left[ -\ii T (H - \omega)
\right] {\rm Td } ({ T {\cal R}^+}) \;,
\end{equation}
and the Weyl character formula
\begin{equation}\label{Weyl}
\chi (T) = {1 \over T^N} \sum_{z_i} {\exp(-\ii T H ) \over
\det^+ ( {\partial} \chi_H^{~} ) } {\rm Td } (T {\partial} \chi_H^{~} ) \;,
\end{equation}
respectively. In (\ref{Weyl}) we have identified the Pfaffian in the
real polarization with the determinant over the holomorphic
eigenvalues of $\Vert {\partial}_a \chi^b_H \Vert $ and the summation is
over the critical points of the Hamiltonian or equivalently the Weyl
group.
As the only example we evaluate the character for SU(2). We write the
character as a coherent state path integral over the coadjoint orbit
$\rm{SU(2)/U(1)} \sim S^2$. We choose complex coordinates by
introducing the stereographic projection from the south pole. The
K\"ahler potential on the orbit with radius $j$ is $F = j
\log(1 +z \bar{z})$ from which we obtain the metric and the symplectic 1-
and 2-forms in the standard fashion. The integrability condition
requires $j$ to be a multiple of $1/2$: this is the topological
quantization of spin. The canonical realization for $H = J_3$ is
$$
J_3 = -j {{1 -z\bar{z}} \over {1+z\bar{z}}} \;
$$
and the path integral for the character becomes
(\ref{ZSUSY})
\begin{equation}\label{SU2Character}
\chi_j (T) = \int {\cal D} z {\cal D} \bar{z} {\cal D} \psi {\cal D} \bar{\psi} \
\exp\left[ \ii j \int_0^T \dd t \left( \ii {\dot{z} \bar{z} -
z\dot{\bar{z}} \over 1 + z\bar{z}} + {1 -z\bar{z} \over 1 + z \bar{z}}
+ {2\ii \psi \bar{\psi} \over (1 +z\bar{z})^2} \right) \right]
\end{equation}
with periodic boundary conditions. The Lie-derivative condition
(\ref{InvariantMetric}) is satisfied for $H =J_3$. This path integral
is given exactly by the WKB-approximation \cite{us,funahashi}.
The relevant quantities in the Kirillov formula (\ref{Kirillov}) are
\begin{eqnarray}\label{PSInvariance}
H- \omega & = & j { 1 -z \bar{z} - \psi \bar{\psi} \over 1 + z \bar{z}
+ \psi \bar{\psi}} \;, \cr
{\cal R}^+ &=& R^+ + \Omega^+ = {1 -z \bar{z} - \psi \bar{\psi} \over 1 +z \bar{z}
+ \psi \bar{\psi}} \;.
\end{eqnarray}
Now one can use the Parisi-Sourlas integration formula $$\int \dd z \ \dd
\bar{z} \ \dd \psi \ \dd \bar{\psi} F(z \bar{z} + \psi \bar{\psi}) = \pi
[F(\infty) - F(0)] $$ which gives
\begin{equation}\label{char}
\chi_j (T) = {\sin (j+1/2)T \over \sin (T/2)} = \sum_{m = -j}^j
\exp[ \ii m T] \:.
\end{equation}
This is exactly the correct result without an explicit Weyl
shift. Also the Weyl formula (\ref{Weyl}) gives the correct result
when we use local coordinate charts in the vicinity of the critical
points. To get the correct north pole contribution we invert the
coordinates $z \rightarrow 1/z, \bar{z} \rightarrow 1/\bar{z}$. This also yields the
correct character (\ref{char}):
$$
\chi_j (T) = { \exp[-\ii j T ] \over 2 \sin(T/2) } \exp [- \ii T/2 ]
+ { \exp [-\ii j T (-1) ] \over 2 \sin(-T/2) } \exp[\ii T /2 ] =
{\sin(j+1/2)T \over \sin(T/2) } = \sum_{m= -j}^j \exp[\ii m T ] \;.
$$
On the other hand, using $\hat{A}$-genus we obtain the result
$$
\chi_j (T) = {\sin (j T) \over \sin(T/2) }
$$
which is the correct result up to the Weyl shift $j \rightarrow j+1/2$. So we
see that in the character formulae we have to use the Todd-genus
instead of $\hat{A}$-genus to directly get the correct result.
\\
\section{Conclusions}
We have considered phase space path integrals with the property that
the Hamiltonian generates an isometry of the phase space. Using
equivariant cohomology in the loop space we were able to reduce the
path integrals to finite dimensional integrals and sums. We also
noticed that the results were not uniquely defined because of spectral
asymmetry. The choice of regularization yielded equivariant
$\hat{A}$- and Todd-classes.
We applied localization to the harmonic oscillator and to the
quantization of coadjoint orbits. We showed that localization produces
correct results for these systems. In addition, we derived Kirillov
and Weyl character formulae that produce correct characters for Lie
groups without the Weyl shift. We demonstrated this explicitly by
evaluating the character for SU(2). The explanation for the Weyl shift
was the same as in the case of the Coxeter shift \cite{bbrt} in
Chern-Simons theory, the $\eta$-invariant.
It would be interesting to apply our formalism to more complicated
systems such as loop groups and field theories. Also, it seems
possible to use localization and equivariant cohomology to study
quantum integrability, generic supersymmetric theories and problems
in classical mechanics, as well. \\
{\bf Acknowledgements}
\\
We thank Prof. Antti Niemi for initiating this project and for
valuable discussions. We also thank A. Alekseev, A. Hietam\"aki and
O. Tirkkonen for discussions and comments.
|
\section{Introduction}
The structure functions $G_1$ and $G_2$ describing the spin-dependent
part of deep-inelastic scattering were discussed by Feynman in his
lectures \cite{Fey} and the definition he used is very appealing due
to the simple partonic interpretation of the dimensionless function
$g_1(x)$, the only one which survives in the scaling limit for the
longitudinal polarization case. However, when one studies the
relatively low-$Q^2$ region, the alternative definition of structure
functions, proposed by Schwinger a few years later \cite{Sch} appears
to be useful as well. In fact, it was recently applied to explain the
strong $Q^2$-dependence of the generalized Gerasimov-Drell-Hearn (GDH)
sum rule\cite{SoTe} and to clarify a number of the related problems
\cite{SoTe2}.
In the present paper this approach is applied to the first
moment of the $g_1$ structure function entering the fundamental
Bjorken sum rule. This quantity was extensively studied in
perturbative QCD and the three-loop corrections were
calculated exactly
\ci{Lar}. As a result,
the additional perturbative {\it one-loop} correction,
originating from the non-smooth zero-quark-mass limit, is found.
The basic point of the approach \cite{SoTe,SoTe2} is the consideration
of the structure functions $g_T(x)=g_1(x)+g_2(x)$ and $g_2(x)$
as independent, while $g_1(x)$ (the most familiar one) is expressed as:
\begin{equation}
\label{def}
g_1(x)=g_T(x)-g_2(x)
\end{equation}
In the resonance region the strong $Q^2-$dependence of $g_2$ due to the
BC sum rule naturally provides the similar dependence of the
generalized GDH sum rule \ci{SoTe}. At the same time,
the BC sum rule in the scaling region is
\begin{equation}
\label{BC}
\int^1_0 g_2(x) dx=0,
\end{equation}
and, at first sight, absolutely does not contribute to the first moment
of $g_1$:
\begin{equation}
\label{zero}
\int^1_0 g_1(x) dx=\int^1_0 g_T(x) dx.
\end{equation}
However, (\ref{BC}) happens to be valid, if and only if the specific
perturbative correction is taken into account. Consequently, this
correction should contribute to the r.h.s of (\ref{zero}).
The problem arose a couple of years ago due to the paper R. Mertig
and W.L. van Neerven{\ci{Mer-vN}, who observed that the
BC sum rule is violated in massive perturbative on-shell QCD at
one-loop level.
However, their result was in contradiction with the
QED calculation performed almost 20 years earlier by
Wu-Yang Tsai, L.L. DeRaad, Jr. and K.A. Milton \ci{Mil}\footnote{
Probably, this fact was not well known,
because the authors of \ci{Mil} used the Schwinger definitions
for the spin-dependent structure functions.}
(the result in QCD is the same apart from a trivial
color factor).
The origin of this discreapance was identified in the papers \ci{SoTe2,Mus}.
The QED calculation \ci{Mil} treated the fermion mass exactly.
The only
difference of the QCD case \ci{Mer-vN} is that the quark mass $m$ is just a
regulator of collinear singularities, and only the terms contributing to
the limit $m\rightarrow 0$ {\it before the integration over $x$}
were taken into account. However, this leads
to the missing of the finite term at the elastic limit, when
the emitted gluon is "soft",
\begin{equation}
\delta_m^{soft} g_2 (x)=C_F{\alpha_s \over {8 \pi}} \delta_+ (1-x),
\end{equation}
restoring the validity of the BC sum rule. The same result was
obtained by G. Altarelli, B. Lampe, P. Nason and G. Ridolfi \ci{Alt}, who
derived
the exact mass-dependent formula in QCD, coinciding, up to the mentioned
colour factor, with the QED result \cite{Mil}. These authors also
established in details the origin of the extra term.
This mass correction, due to (\ref{def}) appears in the expression for
$g_1$ as well. The important point here is that the similar finite
terms for the function $g_T$ are {\it absent} \ci{Mus,Ver}.
As a result the full correction to the $g_1$ is just
\begin{equation}
\delta_m^{soft} g_1 (x) =-\delta_m^{soft} g_2 (x)=C_F{\alpha_s \over {8 \pi}}
\delta_+
(1-x).
\end{equation}
This contribution is especially important in the case of the first moment.
The partial conservation of the non-singlet axial current leads to the
zero anomalous dimension and to the manifestation of this contribution at
the leading approximation.
As a result, the correction to the Bjorken sum rule is changed:
\begin{equation}
\delta_m^{soft} \int_0^1 g_1 (x) dx=-C_F{\alpha_s \over {8 \pi}},
\end{equation}
and the full expression
for the coefficient function can be written as:
\begin{equation}
\int_0^1 g_1 (x) dx = {1 \over 2}(1-{3 \over 4}C_F{\alpha_s \over {\pi}}|_{m
\equiv 0}-{1 \over 4}C_F{\alpha_s \over {\pi}}|_{m \to 0}^{soft}).
\end{equation}
This is a principal result of this paper. However, its importance
requires further investigation.
While such a correction is due to the integration in the region of
the "soft" emitted gluons ($x \sim 1$),
there is another source of the finite mass-to-zero contribution
due to the collinear gluons. In the complete analogy to the
$\delta_+ (1-x)$ one get $\delta_+ (\theta^2)$, where $\theta$
is the gluon emission angle in the c.m. frame. As a result, there is
a finite helicity-flip cross-section in the zero-mass limit. This
effect in QED was discovered by Lee and Nauenberg in their celebrated paper
\ci{Lee} and was extensively studied and applied recently
\ci{Ein2,Seh}.
The main quantitative result is the fermion helicity-flip probability
(analogous to the GLAP kernel, except this is finite rather than logarithmic
contribution), whose straightforward generalization
to the QCD case is:
\begin{equation}
\label{P+-}
P_{+-}(x)= C_F{\alpha_s \over {2 \pi}} (1-x).
\end{equation}
There is a similar effect in the helicity-non-flip cross section as
well. This is easily recovered by, say, the expansion of the numerator
of (8) in \ci{Seh}:
\begin{equation}
{\theta^2 d \theta^2 \over {(\theta^2+m^2/E^2)^2}}
={1 \over {\theta^2+m^2/E^2}}-{m^2/E^2 \over {(\theta^2+m^2/E^2)^2}}
\end{equation}
Keeping $O(m^2)$ one immediately get the finite correction to the
standard logarithmic term, whose $x$-dependence is exactly the same:
\begin{equation}
\label{P++}
P_{++}(x)= - C_F{\alpha_s \over {2 \pi}} \cdot {1+x^2\over {1-x}}.
\end{equation}
The spin-averaged correction is just
\begin{equation}
\label{P}
P_a(x)= P_{++}(x)+P_{+-}(x)= - C_F{\alpha_s \over { \pi}} \cdot {x\over
{1-x}}
\end{equation}
and exactly coincide with the earlier result of Baier, Fadin
and Khoze \ci{Baier}. The agreement with this paper before the angular
integration was proved in \cite{Seh}, although the finite correction
to the helicity-conserving kernel was not presented explicitly.
Let us consider the longitudinally polarized quark. Its density matrix can
be represented as a difference of the density matrices with positive and
negative helicities. The $g_1$ structure function is proportional to:
\begin{equation}
g_1(x) \sim \sigma_{++}-\sigma_{+-}-\sigma_{-+}+\sigma_{--}.
\end{equation}
Taking into account that $\sigma_{--}=\sigma_{++}, \
\sigma_{+-}=\sigma_{-+}$, and that "collinear" contribution results in the
substitutions $\sigma_{++} \to \sigma_{++}(1+P_{++})+\sigma_{+-}
P_{+-}, \
\sigma_{+-} \to \sigma_{+-}(1+P_{++})+\sigma_{++} P_{+-}$, one should
get for the "collinear" correction to $g_1$
\begin{equation}
\label{g1c}
g_1(x)^{coll}={1\over 2}(\delta (1-x)+P_{++}(x)-P_{+-}(x))={1\over 2}
[\delta (1-x)-C_F{\alpha_s \over { \pi}}({1\over {1-x}}+x)],
\end{equation}
where the Born term is kept to make the normalization clear. Passing
to the calculation of the first moment one meet the infrared singularity
at $x=1$. It is very important, that the virtual contribution, providing
its cancelation, is proportional to the same combination
$(log(E^2/m^2)-1)$, appearing in the expression for the GLAP kernel and its
correction $P_{++}$. As a result, the $"+"$ prescription should be applicable
to $P_{++}$ in complete similarity to the logarithmic term.
\begin{equation}
\label{P+++}
P_{++}(x)= - C_F{\alpha_s \over {2 \pi}} ({1+x^2\over {1-x}})_+.
\end{equation}
The first moment of $P_{++}$ is then zero and the correction to the
Bjorken sum rule is completely determined by the helicity-flip kernel:
\begin{equation}
\int_0^1 g_1(x)^{coll} dx = {1\over 2}(1-\int_0^1 dx P_{+-}(x))={1\over 2}
(1-C_F{\alpha_s \over {4 \pi}}).
\end{equation}
Finally, the one-loop correction to Bjorken sum rule for the quark is:
\begin{equation}
\label{tot}
\int_0^1 g_1(x) dx = {1\over 2}
[1-C_F{\alpha_s \over {4 \pi}}(3(m \equiv 0)+1(soft)+1(coll))],
\end{equation}
One may worry, is it really possible just to add "zero-mass", "soft"
and "collinear" terms. The explicit calculation of $g_{1+2}$ and $g_2$
on mass shell keeping mass exactly \cite{Ver}
confirm this naive derivation. The "collinear" contribution to $g_1$
reproduce (\ref{g1c}) while the total correction is just $-5C_F/4 \pi$,
like (\ref{tot}). The same value was obtained in \ci{Mer-vN},
(where, however, soft correction was not present) and was interpreted
as a manifestation of the regularization dependence.
Let us turn to the physical applications of these results. The status of the
soft and collinear contributions is then quite a different. The
collinear contributions correspond to the integration in the region
of the low transverse momenta ($k_T \leq m_q^2$). For the light quarks
because of the confinement the quark mass should be replaced by the
pion one \cite{GorIo}. This contribution is normally excluded by
the cuts when the experimental
data are obtained. The situation here is quite analogous to the
anomalous gluon contribution to the $g_1$ structure function
\cite{EST}. Let us note also in this connection that this naturally
explains the results obtained in \cite{Seh2}; namely, that the spin-flip
collinear effects contribute to the "normal" piece of the
spin-dependent photon spin structure function. This is due to the
fact, that these effects are related to the low transverse momenta of
scattered quarks.
For the heavy quarks, however, this correction should
be taken into account.
From the other side, there are no reasons to exclude "soft" piece
for both heavy and light quarks. It corresponds to the integration
over x when $1-x \sim m_q^2/Q^2$. Again, the confinement effects will
change quark mass to the pion one. In principle, one can not exclude that
the accurate treatment of quark condensates instead of quark mass term
may change the actual value of the correction. Anyway, this $x$ due to the
standard convolution formula
\begin {equation}
g_{hadron}(x_B) = \int_{x_B}^1 dx \Delta q(x) g_{quark}(x/x_B)
\end{equation}
(where $ \Delta q(x)$ is a spin-dependent quark distribution)
do not correspond to the particular value of the observed $x_B$.
It is interesting, that for the $g_2$ structure function the soft and
collinear contributions cancel each other \cite{Ver}. As a result, the
Burkhardt-Cottingham sum rule is valid either if one put quark mass
equal to zero from the very beginning, or if one take into account
all the effects, surviving the zero-mass limit (like for the heavy
quarks). If only the soft contributions is taken into account
(like for the light quarks), the BC sum rule is violated to the same
extent, up to a sign, as was reported by Mertig and van Neerven.
This sign difference is due to the fact, that in \ci{Mer-vN} the
"collinear' contribution was taken into account (it is just the
compensating terms of order $q^2/m^2$ mentioned at p.491 of this
reference), contrary to the "soft" one.
Finally, inserting the experimentally known expressions for the first
moment of the spin-dependent distributions one get the total one-loop
correction to $g_1^{p-n}$
\begin {equation}
\label{fin}
\int_0^1 g_1^{p-n}(x) dx={g_A\over 6}
(1-C_F{\alpha_s \over {\pi}})
\end{equation}
Let us discuss briefly the experimental situation (see \cite{ALE} and
ref. therein). The multiplication
of the 1-loop correction to the Bjorken sum rule by the light-quark
factor $4/3$ significantly improve the correspondence of the result
with the most exact data of E143 experiment at SLAC. Namely, one get
the value $1.64 \pm 0.09$ instead of the standard one $1.72 \pm 0.09$.
If one take into account the power corrections calculated in the
framework of QCD sum rules method results in the value $1.54 \pm
0.017$ instead of $1.62 \pm 0.017$. One should compare this with the
most recent experimental value $1.51 \pm 0.013$. Although all the
theoretical values coincide with the experimental one within the
errors, the mass correction make this agreement for the central values
much better.
I especially would like to stress the perfect agreement of the
QCD sum rules calculation (note that it was obtained with only
massless perturbative corrections taken into account, when the
agreement is substantially worse).
As it is well known, the correction to the Bjorken sum rule coincide
with that for the Gross-Llewellynn Smith (GLS) sum rule
\ci{Lar,Kat}. Although
in the massive case one can not simply transform the relevant diagrams
to each other (the $\gamma^5$ moving change the signs of the mass
term \footnote{This effect spoils, in principle, the arguments of
\ci{Mer-vN}, who
used the correction to GLS sum rule calculated earlier to get the answer for
Bjorken sum rule.}),
$O(m^2)$ terms in the numerator coming from the trace are not responsible
for the extra "soft" term in $g_2$ (and, consequently, $g_1$), as it
was first mentioned in \cite{Alt}.
As a result, the coefficient function for the light quarks receives
the same contribution:
\begin {equation}
\label{GLS}
\int_0^1 F_3(x) dx = 3(1-C_F{\alpha_s \over \pi}).
\end{equation}
This relation, however, requires an additional check by the
straightforward calculation, because one can not exclude, in
principle, the appearance of such corrections coming from another
terms. This work is now in progress \cite{Ver2}.
However, the extra factor $4/3$ already completely remove
the discreapance \cite{Sidor} of the data with perturbative QCD for
all $Q^2$.
Note that such a 'compensating' factor fully respects the qualitative
nature of this
discreapance: the experimental data as a function of $Q^2$
are going above the theoretical curve and the difference is decreasing
for higher $Q^2$.
The corrections to the Bjorken and GLS sum rules are in turn related
to the correction to the $e^+ e^-$-annihilation total cross section.
This is due to the famous Crewther relation \cite{Crewter}, studied
recently for high
orders of perturbation theory\cite{Kat}. However, the zero mass limit
in this
case is known to be smooth\cite{PSF}. It seems that there is no
contradiction
here. The Crewther relation emerges due to the non-renormalization of
axial anomaly, which is not directly related to the mass contributions
studied here.
In conclusion, the new perturbative QCD correction to the partonic sum
rules is found. It comes from the quark mass contribution, having the
non-zero limit, when the quark mass tends to zero.
This correction seems to improve the agreement with the
experimental data for Bjorken sum rule. Taking into account the
similar correction removes the deficit of Gross-LLewellyn Smith sum rule.
Although the manifestation of these correction
requires further investigation, one may conclude that the study of
mass effects (even in a zero mass limit) in hard processes
(in particular, in the partonic sum rules) seems to be very interesting.
I am thankful to A.V. Efremov, J. Ho\'rej\'si, E.A. Kuraev and
S.V. Mikhailov for stimulating discussions and valuable comments.
\begin{thebibliography}} \newcommand{\eb{99}
\bi{Fey} R.P.Feynman {\it Photon-Hadron Interactions}, (Benjamin,
Reading, MA, 1972).
\bi{Sch} J.Schwinger, Proc. Natl. Acad. Sci. U.S.A. {\bf 72},
(1975) 1559.
\bi{SoTe} J.Soffer and O.Teryaev, Phys. Rev. Lett. {\bf
70}, (1993) 3373.
\bi{SoTe2} J.Soffer and O.Teryaev, Phys. Rev. {\bf D51}, (1995) 25.
\bi{Lar} S.A. Larin, J.A.M. Vermaseren, Phys. Lett {\bf B259}, (1991)
345.
\bi{Mer-vN} R. Mertig and W.L. van Neerven, Z. Phys.{\bf C60}, (1993) 489.
\bi{Mil} Wu-Yang Tsai, L.L. DeRaad, Jr. and K.A. Milton, Phys.
Rev. {\bf D11}, (1975) 3537.
\bi{Mus} I.V. Musatov, O.V. Teryaev, CEBAF Report TH-94-25(unpublished).
\bi{Alt} G. Altarelli, B. Lampe, P. Nason and G. Ridolfi, Phys. Lett
{\bf B334}, (1994) 187.
\bi{Ver} I.V. Musatov, O.V. Teryaev and O.L. Veretin, in preparation.
\bi{Lee} T.D. Lee and M. Nauenberg, Phys. Rev. {\bf B133}, (1964) 1549.
\bi{Ein2} H.F. Contopanagos and M.B. Einhorn, Nucl. Phys {\bf B377},
(1992) 20
\bi{Seh} B. Falk and L.M. Sehgal, Phys. Lett {\bf B325}, (1994) 509.
\bi{Baier} V.N. Baier, V.S. Fadin and V.A. Khoze, Nucl. Phys. {\bf
B65}, (1973) 381.
\bi{GorIo} A.S. Gorskii, B.L. Ioffe, A.Yu. Khodzhamirian,
Phys. Lett. {\bf B227}, (1989) 474.
\bi{EST} A.V.Efremov, J.Soffer, O.V.Teryaev, Nucl.Phys. {\bf B346},
(1990) 97.
\bi{Seh2} A. Freund and L.M. Sehgal, Phys. Lett {\bf B341}, (1994) 90.
\bi{ALE} M. Anselmino, A.V. Efremov and E. Leader, CERN Preprint
CERN-TH.7216/94 (to be published in Physics Reports).
\bi{Kat} D.J. Broadhurst and A.L. Kataev, Phys. Lett {\bf B315},
(1993) 179
\bi{Sidor} A.L. Kataev and A.V. Sidorov, Phys. Lett. {\bf B331},
(1994) 179
\bi{Ver2} O.V. Teryaev, O.L. Veretin, in progress.
\bi{Crewter} R.J. Crewther, Phys. Rev. Lett. {\bf 28}, (1972) 345.
\bi{PSF} J.Schwinger, {\it Particles, Sources and Fields}
(Addison-Wesley, 1989) v. 3, p.99.
\eb
\end{document}
|
\section{Introduction}
In the period leading up to the start of the HERA program
a substantial amount of work was done on radiative corrections
to observables in deep-inelastic scattering (DIS) of electrons
off protons (see e.g. \cite{HERAproc} for an overview). These corrections
are substantial enough for both HERA experiments to correct for them.
Part of the upcoming LEP2 physics programme involves the measurement
of the photon structure function $F_2^{\gamma}(x,Q^2)$.
This structure function is extracted from deep-inelastic electron--photon
scattering in the reaction ${\mathrm{e}}^+{\mathrm{e}}^- \rightarrow {\mathrm{e}}^+{\mathrm{e}}^- X$,
where one of the leptons escapes undetected down the beam pipe,
while the other is measured at rather large angle.
It is therefore important to study radiative corrections
to deep-inelastic electron--photon scattering. This is our
purpose in this paper.
Radiative corrections to ${\mathrm{e}}^+{\mathrm{e}}^- \rightarrow {\mathrm{e}}^+{\mathrm{e}}^- X$
have been calculated for $X$ a (pseudo) scalar particle
\cite{Defrise,Neerven,Landro} and $X = \mu^+\mu^-$ \cite{Landro,Berends}.
It was found that they are very small
(on the percent level) in the no-tag case,
when neither the electron nor the positron is being measured.
In such a kinematic configuration, when the
momentum transfer between the incident
and outgoing electron (or positron) is small, the vertex-
and bremsstrahlung contributions effectively cancel each other,
leaving a small correction dominated by
vacuum polarization \cite{Neerven,Landro}. This implies that
for the equivalent photon spectrum, which is essential to relate
${\mathrm{e}}^{\pm}\gamma$ with ${\mathrm{e}}^+{\mathrm{e}}^-$ reactions, corrections are small.
In contrast, radiative corrections can be sizeable in the case
where one of the leptons scatters
at a large angle - single tag - and one
studies differential cross sections
which depend strongly on the energy and angle of the tagged
electron (or positron).
Surprisingly, no calculations of the size of radiative corrections
for inclusive deep-inelastic electron--photon scattering, i.e.
$e\gamma \rightarrow e X$ with both large $Q^2$ (the
absolute value of the transferred momentum squared) and $W$ (the mass
of state $X$), have been performed yet. Correspondingly,
in experimental analyses of the (hadronic) photon structure function
$F_2^\gamma(x,Q^2)$ radiative corrections have so far not been included
(as usual, $x=Q^2/(Q^2+W^2)$).
They are usually
assumed to be negligible. Recently, the AMY collaboration \cite{AMY}
estimated the size of radiative corrections by comparing
a Monte-Carlo event generator based on the full cross section formula
for ${\mathrm{e}}^+{\mathrm{e}}^- \rightarrow {\mathrm{e}}^+ {\mathrm{e}}^- \gamma\mu^+\mu^-$ \cite{Berends}
(with the muon mass and electric charge changed to correspond
with quark-antiquark pair production)
with a generator for ${\mathrm{e}}^+{\mathrm{e}}^- \rightarrow {\mathrm{e}}^+{\mathrm{e}}^- {\mathrm{q}\bar{\mathrm{q}}}$ where
the cross section for ${\mathrm{e}} \gamma \rightarrow {\mathrm{e}} {\mathrm{q}\bar{\mathrm{q}}}$ with
a real photon was convoluted with the equivalent photon spectrum.
A (positive) correction
of order $10$\% for the visible $x$ distribution was found, which,
however, cancelled effectively against the correction due to a non-vanishing
target-photon mass. Hence no net correction was applied.
Nevertheless, it is important to understand both corrections separately,
and their behavior as a function of the kinematic variables chosen
and phase space. Moreover, the hadronic structure of the photon
must not be neglected.
Here we therefore estimate the size of the radiative corrections
to inclusive deep-inelastic ${\mathrm{e}}\gamma$ scattering, using the full
photon structure function.
In the next section we describe the relevant kinematics and formalism
and in section 3 we present results.
\section{Formalism}
We consider the $O(\alpha)$ corrections to deep-inelastic
scattering (DIS) of electrons on (quasi-real) photons:
\begin{equation}
{\mathrm{e}}(l) + \gamma(p) \rightarrow {\mathrm{e}}(l') + \gamma(k) + X(p_X)
\ .
\label{DISreact}
\end{equation}
This process is depicted in Fig.1, in which we indicate
all momentum labels.
The target photon $\gamma(p)$ is part of the flux of equivalent
photons around the non-tagged lepton. We assume that
this flux has a momentum density given
by the Weizs\"{a}cker-Williams expression
\begin{equation}
f_{\gamma/{\mathrm{e}}}(z) = \frac{\alpha}{2\pi} \left\{
\frac{1+(1-z)^2}{z}\, \ln \frac{P^2_{max}}{P^2_{min}}
- 2 m_e^2 z(\frac{1}{P^2_{min}}
-\frac{1}{P^2_{max}})\right\}
\label{fww}
\end{equation}
where $P^2_{min} = (z^2 m_e^2)/(1-z)$ and $P^2_{max} =
(1-z) \left(E_b \theta_{max}\right)^2$.
Here $z$ is the longitudinal momentum fraction of the
target photon with respect to its parent lepton, $E_b=\sqrt{s}/2$
is the lepton beam energy, $\theta_{max}$ is the anti-tag
\footnote{i.e. all events in which the parent lepton scatters
at an angle larger than $\theta_{max}$ are rejected.} angle
and $P^2 = -p^2$.
In the following we put $P^2 = 0$ and neglect electron masses
everywhere except in (\ref{fww}). Moreover we substitute
$P^2_{max}$ by $P^2_{max}+P^2_{min}$ so that we can easily
extend the $z$ range to 1, see \cite{FMNR}.
\vglue 6.1cm
\vbox{\special{psfile=fig1.ps angle=0 hscale=60 vscale=60
hoffset=40 voffset=-180}}
{\tenrm\baselineskip=12pt
\noindent Figure 1: Photon bremsstrahlung from the tagged lepton line
in deep-inelastic scattering off an equivalent photon.}
\vglue 0.8cm
The DIS variables can be defined either from the leptonic or
the hadronic momenta:
\begin{equation}
\begin{array}{rclrcl}
q_l & = & l-l' & q_h & = & p_X - p = q_l - k\\
W_l^2 & = & (p+q_l)^2 \qquad & W_h^2 & = & (p+q_h)^2 = p_X^2\\
Q_l^2 & = & - q_l^2 & Q_h^2 & = & - q_h^2\\
x_l & = & Q_l^2/2p \cdot q_l &
x_h & = & Q_h^2/2p \cdot q_h\\
y_l & = & p \cdot q_l/p \cdot l &
y_h & = & p \cdot q_h/p \cdot l
\end{array}
\label{DISvar}
\end{equation}
Note that both $Q_l^2 = x_l y_l s_{\re \gamma}$ and $Q_h^2 = x_h y_h s_{\re \gamma}$,
where $s_{\re \gamma} = (p+l)^2$,
but that leptonic and hadronic variables agree only for nonradiative
events, i.e. if $k=0$.
We will see that the size of the corrections strongly
depends on which set of variables are used
in the measurement. In practice one determines $Q^2$ from
the tagged lepton, and $x$ from the (visible) hadronic
energy $W_h$.
The Born cross section (i.e. no $\gamma(k)$ in (\ref{DISreact}))
is given by
\begin{equation}
\frac{{\mathrm{d}}^2\sigma^{\mathrm{B}}}{{\mathrm{d}} x {\mathrm{d}} Q^2} = f^{\mathrm{B}}(x,Q^2,s)
\ ,
\label{sigBorn}
\end{equation}
where
\begin{eqnarray}
f^{\mathrm{B}}(x,Q^2,s) &= & \frac{2\pi\alpha^2}{x Q^4}\, F_2(x,Q^2)
\\ & \times & \int_{z_{min}(x,Q^2,s)}^1 {\mathrm{d}} z\, f_{\gamma/{\mathrm{e}}}(z)\,
Y_+\left(\frac{Q^2}{x z s} \right) \,
\left\{ 1 + R\left(x,Q^2,\frac{Q^2}{x z s} \right) \right\}
\label{Borncross} \ .
\end{eqnarray}
Here we have defined $z_{min}(x,Q^2,s) = Q^2/xs$,
$ Y_+(y) = 1 + (1-y)^2$ and
\begin{equation}
R(x,Q^2,y) = \frac{- y^2}{1 +(1-y)^2} \
\frac{F_L(x,Q^2)}{F_2(x,Q^2)} \ .
\label{Rdef}
\end{equation}
$F_{2,L}$ are the photon structure functions (we have dropped
the superscript $\gamma$).
The above form eq.~(\ref{Borncross}) is useful because $F_2$ can be factored
out of the $z$ integration.
For comparison of ${\mathrm{e}} \gamma$ with ${\mathrm{e}} {\mathrm{p}}$ scattering we will also
give the cross section in terms of $x$ and $y$
\begin{equation}
\frac{{\mathrm{d}}^2\sigma^{\mathrm{B}}}{{\mathrm{d}} x {\mathrm{d}} y} = g^{\mathrm{B}}(x,y,s)
\ ,
\label{nsigBorn}
\end{equation}
where
\begin{equation}
g^{\mathrm{B}}(x,y,s) = \frac{2 \pi\alpha^2 \,Y_+(y)}{x^2 y^2 s}
\int_{\epsilon(x,y,s)}^1 \frac{{\mathrm{d}} z}{z} \, f_{\gamma/{\mathrm{e}}}(z)\,
F_2(x,xyzs) \left\{1 + R(x,xyzs,y) \right\}
\ .
\label{newBorn}
\end{equation}
The lower limit $\epsilon = W^2_{\mathrm{min}}/(1-x)ys$
on the $z$-integration in eq.~(\ref{newBorn}) arises if a lower
cut is applied to the invariant hadronic mass $W$.
At the Born level expressions (\ref{sigBorn}) and (\ref{nsigBorn})
are equivalent and valid for both sets of variables in (\ref{DISvar}).
For the ${\mathrm{e}}$p case the full electroweak
corrections have been calculated for both neutral current reactions
\cite{Dubna,Wurzburg},
including elastic nucleon scattering \cite{Akhundov},
and charged-current reactions \cite{Bohm}.
It is well-known that the leading logarithmic
approximation (LLA) (in $\ln(Q^2/m_e^2)$) reproduces the
exact results to within a few percent however
\cite{Consoli, Spiesberger}.
Here we will therefore work within this approximation
and neglect furthermore $Z$ exchange.
We consider first the radiative corrections to the
differential cross section in (\ref{nsigBorn})
in terms of $x_l$ and $y_l$.
The $O(\alpha)$ corrections in LLA arise from
collinear and soft bremsstrahlung from
the initial and final electron that couple to the ``probing photon"
and from the Compton process\footnote{As stated above,
radiative effects to the Weizs\"{a}cker-Williams spectrum
are small \cite{Landro} and we therefore neglect them.}.
The latter corresponds to the case
in Fig.1 where the exchanged photon $\gamma(q)$ is quite
soft, but $\gamma(k)$ is radiated at a wide angle
causing the lepton $e(l')$ to be tagged.
The bremsstrahlung terms are given by
\begin{equation}
\frac{{\mathrm{d}}^2\sigma^{\mathrm{Br}}}{{\mathrm{d}} x {\mathrm{d}} y} =
\int_0^1 {\mathrm{d}} x_i \, D_{e/e}(x_i,Q^2)
\left\{ \Theta\left( x_i - x_i^0\right) J(x_1,x_2)
g^{\mathrm{B}}(\hat{x},\hat{y},\hat{s}) - g^{\mathrm{B}}(x,y,s) \right\}
\label{bremscross}
\end{equation}
where $\hat{s} = x_1 s$, $\hat{x} = x x_1 y/(x_1 x_2 + y -1)$,
$\hat{y} = y x/x_2 \hat{x}$, $J(x_1,x_2) = y/x_1 x_2^2 \hat{y}$,
$x_1^0 = (1-y)/(1-x y)$ and $x_2^0 = x y +1-y$, and
\begin{equation}
D_{e/e}(x,Q^2) = \frac{\alpha}{2\pi}\ln\left(\frac{Q^2}{m_e^2}\right)
\, \frac{1+x^2}{1-x}\ .
\end{equation}
In eq. (\ref{bremscross}) we have supressed the subscript $l$ on
$x$ and $y$ for clarity.
Initial-state radiation corresponds to $x_i = x_1$ and $x_2=1$
in eq.~(\ref{bremscross}) and vice versa for final-state
radiation.
The Compton contribution is given by
\begin{equation}
\frac{{\mathrm{d}}^2\sigma^{\mathrm{C}}}{{\mathrm{d}} x_l {\mathrm{d}} y_l} =
\int_{\epsilon}^1 {\mathrm{d}} z \, f_{\gamma/{\mathrm{e}}}(z)\,
h^{{\mathrm{C}}}(x_l,y_l,x_ly_l zs,zs)
\ ,
\label{Compton}
\end{equation}
where
\begin{eqnarray}
h^{{\mathrm{C}}}(x,y,Q^2,s) &=& \frac{\alpha^3}{x^2(1-y)s}\ Y_+(y)\
\ln\frac{Q^2}{M^2} \int_x^1 \frac{{\mathrm{d}} v}{v}
\left[ 1 + \left(1 - \frac{x}{v}\right)^2 \right]\nonumber\\
&\times& F_2(v,Q^2) (1+R(v,Q^2,y))
\ .
\label{hCdef}
\end{eqnarray}
The logarithm $\ln(Q^2/M^2)$ is a result of the absorption
of the collinear singularity from the quark to photon splitting
by renormalizing the photon density in the quark at scale $M$.
We take $M$ here to be the proton mass \cite{Dubna}.
Next we discuss the radiative corrections to the differential
cross section expressed in hadronic variables. Now there are,
in accordance with the KLN theorem \cite{KLN}, in LLA approximation
neither corrections from final state radiation, nor from the
Compton process because these process do not affect the
kinematic variables. We find that the correction due to
initial state radiation can simply be expressed as
\begin{eqnarray}
\frac{{\mathrm{d}}^2\sigma^{\rm corr}}{{\mathrm{d}} x_h {\mathrm{d}} Q_h^2}& = &
\frac{2 \pi\alpha^2}{x_h Q_h^4}{\Bigg\{}
F_2(x_h,Q_h^2) \, \int_{z_{min}(x_h,Q_h^2,s)}^1 {\mathrm{d}} z\, f_{\gamma/{\mathrm{e}}}(z)\,
g_h\left( x_h,Q_h^2,\frac{Q_h^2}{z x_h s}\right) \nonumber \\
& + & F_L(x_h,Q_h^2) \, \int_{z_{min}(x_h,Q_h^2,s)}^1 {\mathrm{d}} z\, f_{\gamma/{\mathrm{e}}}(z)\,
h_h\left( x_h,Q_h^2,\frac{Q_h^2}{z x_h s}\right) {\Bigg\}}
\label{hadcorr}
\end{eqnarray}
where
\begin{equation}
g_h(x,Q^2,y) = \frac{\alpha}{\pi}\ \ln \frac{Q^2}{m_e^2}\,
\left\{ Y_+(y) \ln (1-y) + y\left(1 - \frac{y}{2}\right) \ln y
+ y\left(1 - \frac{y}{4} \right) \right\}
\end{equation}
and
\begin{equation}
h_h(x,Q^2,y) = \frac{\alpha}{\pi}\ \ln \frac{Q^2}{m_e^2}\,
\left\{ -y^2 \ln (1-y) + \frac{y^2}{2}\ln y
- \frac{y}{2}\left(1 - \frac{y}{2} \right) \right\}
\end{equation}
The corrections are large at large $y$ (soft region), and can
in fact be resummed by simple exponentiation,
with the result
\begin{eqnarray}
g_h(x,Q^2,y) & = & Y_+(y)\ \left\{ \exp \left[ \frac{\alpha}{\pi}\
\left( \ln \frac{Q^2}{m_e^2} - 1 \right) \ln(1-y) \right]
- 1 \right\}
\nonumber\\ & & \quad + \frac{\alpha}{\pi}\ \ln \frac{Q^2}{m_e^2}\,
\left\{ y\left(1 - \frac{y}{2}\right) \ln y
+ y\left(1 - \frac{y}{4} \right) \right\}
\label{resummed2}
\end{eqnarray}
and
\begin{eqnarray}
h_h(x,Q^2,y) & = & -y^2\ \left\{ \exp \left[ \frac{\alpha}{\pi}\
\left( \ln \frac{Q^2}{m_e^2} - 1 \right) \ln(1-y) \right]
- 1 \right\}
\nonumber\\ & & \quad + \frac{\alpha}{\pi}\ \ln \frac{Q^2}{m_e^2}\,
\left\{ \left(\frac{y^2}{2}\right) \ln y
-\frac{y}{2}\left(1 - \frac{y}{2} \right) \right\}
\label{resummedL}
\end{eqnarray}
In the next section we use the formulae listed in the above
to estimate the size of the corrections.
\section{Results}
Here we study the radiative correction to deep-inelastic electron--photon
scattering numerically. For the results presented below
we use $\sqrt{s} = 175\,{\rm GeV}$ and $W_{min} = 2\,{\rm GeV}$.
For the parton densities in the photon we use set 1 of \cite{SaS}.
This set has an already low minimum $Q^2$ of $Q_0^2 =0.36\, {\rm GeV}^2$.
However in radiative events even lower values of $Q^2$ contribute.
We therefore extrapolate below this value by
\begin{eqnarray}
F_2^{\gamma}(x,Q^2<Q_0^2)& = &F_2^{\gamma}(x,Q_0^2)\left(\frac{Q^2}{Q_0^2}\right)^2+
\left(1-\frac{Q^2}{Q_0^2}\right) \frac{Q^2 (1-x)}{112\,{\rm GeV^2}} \nonumber \\
&\times&\left(0.211 (\frac{W^2}{\rm GeV^2})^{0.08} +
0.297 (\frac{W^2}{\rm GeV^2})^{-0.45}\right),
\label{extrapol}
\end{eqnarray}
(this expression correctly approaches the $\gamma\gamma$ total cross section
in the small $Q^2$ limit)
which vanishes as $Q^2\rightarrow 0$, as required by gauge invariance.
Furthermore we have neglected $F_L^{\gamma}$, i.e. we
put $R$ in (\ref{Borncross})
and (\ref{newBorn}) to zero. We found that its inclusion has
a negligible effect.
\vglue 0.5cm
\setlength{\unitlength}{0.240900pt}
\ifx\plotpoint\undefined\newsavebox{\plotpoint}\fi
\sbox{\plotpoint}{\rule[-0.200pt]{0.400pt}{0.400pt}}%
\begin{picture}(1200,1259)(0,0)
\font\gnuplot=cmr10 at 10pt
\gnuplot
\sbox{\plotpoint}{\rule[-0.200pt]{0.400pt}{0.400pt}}%
\put(176.0,562.0){\rule[-0.200pt]{231.264pt}{0.400pt}}
\put(176.0,113.0){\rule[-0.200pt]{0.400pt}{270.531pt}}
\put(176.0,113.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,113){\makebox(0,0)[r]{-0.8}}
\put(1116.0,113.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,225.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,225){\makebox(0,0)[r]{-0.6}}
\put(1116.0,225.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,338.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,338){\makebox(0,0)[r]{-0.4}}
\put(1116.0,338.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,450.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,450){\makebox(0,0)[r]{-0.2}}
\put(1116.0,450.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,562.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,562){\makebox(0,0)[r]{0}}
\put(1116.0,562.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,675.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,675){\makebox(0,0)[r]{0.2}}
\put(1116.0,675.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,787.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,787){\makebox(0,0)[r]{0.4}}
\put(1116.0,787.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,899.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,899){\makebox(0,0)[r]{0.6}}
\put(1116.0,899.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,1011.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,1011){\makebox(0,0)[r]{0.8}}
\put(1116.0,1011.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,1124.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,1124){\makebox(0,0)[r]{1}}
\put(1116.0,1124.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,1236.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,1236){\makebox(0,0)[r]{1.2}}
\put(1116.0,1236.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(176,68){\makebox(0,0){0}}
\put(176.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(283.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(283,68){\makebox(0,0){0.1}}
\put(283.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(389.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(389,68){\makebox(0,0){0.2}}
\put(389.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(496.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(496,68){\makebox(0,0){0.3}}
\put(496.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(603.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(603,68){\makebox(0,0){0.4}}
\put(603.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(709.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(709,68){\makebox(0,0){0.5}}
\put(709.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(816.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(816,68){\makebox(0,0){0.6}}
\put(816.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(923.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(923,68){\makebox(0,0){0.7}}
\put(923.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1029.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1029,68){\makebox(0,0){0.8}}
\put(1029.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1136.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1136,68){\makebox(0,0){0.9}}
\put(1136.0,1216.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(176.0,113.0){\rule[-0.200pt]{231.264pt}{0.400pt}}
\put(1136.0,113.0){\rule[-0.200pt]{0.400pt}{270.531pt}}
\put(176.0,1236.0){\rule[-0.200pt]{231.264pt}{0.400pt}}
\put(656,23){\makebox(0,0){$y_l$}}
\put(176.0,113.0){\rule[-0.200pt]{0.400pt}{270.531pt}}
\put(187,423){\usebox{\plotpoint}}
\multiput(187.58,423.00)(0.498,0.991){81}{\rule{0.120pt}{0.890pt}}
\multiput(186.17,423.00)(42.000,81.152){2}{\rule{0.400pt}{0.445pt}}
\multiput(229.00,506.58)(0.751,0.498){69}{\rule{0.700pt}{0.120pt}}
\multiput(229.00,505.17)(52.547,36.000){2}{\rule{0.350pt}{0.400pt}}
\multiput(283.00,542.58)(1.299,0.498){79}{\rule{1.134pt}{0.120pt}}
\multiput(283.00,541.17)(103.646,41.000){2}{\rule{0.567pt}{0.400pt}}
\multiput(389.00,583.58)(1.928,0.497){53}{\rule{1.629pt}{0.120pt}}
\multiput(389.00,582.17)(103.620,28.000){2}{\rule{0.814pt}{0.400pt}}
\multiput(496.00,611.58)(2.254,0.496){45}{\rule{1.883pt}{0.120pt}}
\multiput(496.00,610.17)(103.091,24.000){2}{\rule{0.942pt}{0.400pt}}
\multiput(603.00,635.58)(1.781,0.497){57}{\rule{1.513pt}{0.120pt}}
\multiput(603.00,634.17)(102.859,30.000){2}{\rule{0.757pt}{0.400pt}}
\multiput(709.00,665.58)(1.632,0.497){63}{\rule{1.397pt}{0.120pt}}
\multiput(709.00,664.17)(104.101,33.000){2}{\rule{0.698pt}{0.400pt}}
\multiput(816.00,698.58)(1.012,0.498){103}{\rule{0.908pt}{0.120pt}}
\multiput(816.00,697.17)(105.116,53.000){2}{\rule{0.454pt}{0.400pt}}
\multiput(923.00,751.58)(0.552,0.498){93}{\rule{0.542pt}{0.120pt}}
\multiput(923.00,750.17)(51.876,48.000){2}{\rule{0.271pt}{0.400pt}}
\multiput(976.58,799.00)(0.498,0.547){103}{\rule{0.120pt}{0.538pt}}
\multiput(975.17,799.00)(53.000,56.884){2}{\rule{0.400pt}{0.269pt}}
\multiput(1029.58,857.00)(0.498,1.096){105}{\rule{0.120pt}{0.974pt}}
\multiput(1028.17,857.00)(54.000,115.978){2}{\rule{0.400pt}{0.487pt}}
\multiput(1083.58,975.00)(0.498,2.085){103}{\rule{0.120pt}{1.760pt}}
\multiput(1082.17,975.00)(53.000,216.346){2}{\rule{0.400pt}{0.880pt}}
\put(187,370){\usebox{\plotpoint}}
\multiput(187.58,370.00)(0.498,0.931){81}{\rule{0.120pt}{0.843pt}}
\multiput(186.17,370.00)(42.000,76.251){2}{\rule{0.400pt}{0.421pt}}
\multiput(229.00,448.58)(0.731,0.498){71}{\rule{0.684pt}{0.120pt}}
\multiput(229.00,447.17)(52.581,37.000){2}{\rule{0.342pt}{0.400pt}}
\multiput(283.00,485.58)(1.331,0.498){77}{\rule{1.160pt}{0.120pt}}
\multiput(283.00,484.17)(103.592,40.000){2}{\rule{0.580pt}{0.400pt}}
\multiput(389.00,525.58)(2.000,0.497){51}{\rule{1.685pt}{0.120pt}}
\multiput(389.00,524.17)(103.502,27.000){2}{\rule{0.843pt}{0.400pt}}
\multiput(496.00,552.58)(2.463,0.496){41}{\rule{2.045pt}{0.120pt}}
\multiput(496.00,551.17)(102.755,22.000){2}{\rule{1.023pt}{0.400pt}}
\multiput(603.00,574.58)(3.174,0.495){31}{\rule{2.594pt}{0.119pt}}
\multiput(603.00,573.17)(100.616,17.000){2}{\rule{1.297pt}{0.400pt}}
\multiput(709.00,591.58)(2.463,0.496){41}{\rule{2.045pt}{0.120pt}}
\multiput(709.00,590.17)(102.755,22.000){2}{\rule{1.023pt}{0.400pt}}
\multiput(816.00,613.58)(1.928,0.497){53}{\rule{1.629pt}{0.120pt}}
\multiput(816.00,612.17)(103.620,28.000){2}{\rule{0.814pt}{0.400pt}}
\multiput(923.00,641.58)(1.490,0.495){33}{\rule{1.278pt}{0.119pt}}
\multiput(923.00,640.17)(50.348,18.000){2}{\rule{0.639pt}{0.400pt}}
\multiput(976.00,659.58)(0.951,0.497){53}{\rule{0.857pt}{0.120pt}}
\multiput(976.00,658.17)(51.221,28.000){2}{\rule{0.429pt}{0.400pt}}
\multiput(1029.00,687.58)(0.587,0.498){89}{\rule{0.570pt}{0.120pt}}
\multiput(1029.00,686.17)(52.818,46.000){2}{\rule{0.285pt}{0.400pt}}
\multiput(1083.58,733.00)(0.498,0.851){103}{\rule{0.120pt}{0.779pt}}
\multiput(1082.17,733.00)(53.000,88.383){2}{\rule{0.400pt}{0.390pt}}
\put(187,343){\usebox{\plotpoint}}
\multiput(187.58,343.00)(0.498,0.919){81}{\rule{0.120pt}{0.833pt}}
\multiput(186.17,343.00)(42.000,75.270){2}{\rule{0.400pt}{0.417pt}}
\multiput(229.00,420.58)(0.731,0.498){71}{\rule{0.684pt}{0.120pt}}
\multiput(229.00,419.17)(52.581,37.000){2}{\rule{0.342pt}{0.400pt}}
\multiput(283.00,457.58)(1.299,0.498){79}{\rule{1.134pt}{0.120pt}}
\multiput(283.00,456.17)(103.646,41.000){2}{\rule{0.567pt}{0.400pt}}
\multiput(389.00,498.58)(2.000,0.497){51}{\rule{1.685pt}{0.120pt}}
\multiput(389.00,497.17)(103.502,27.000){2}{\rule{0.843pt}{0.400pt}}
\multiput(496.00,525.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,524.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,545.58)(2.994,0.495){33}{\rule{2.456pt}{0.119pt}}
\multiput(603.00,544.17)(100.903,18.000){2}{\rule{1.228pt}{0.400pt}}
\multiput(709.00,563.58)(3.204,0.495){31}{\rule{2.618pt}{0.119pt}}
\multiput(709.00,562.17)(101.567,17.000){2}{\rule{1.309pt}{0.400pt}}
\multiput(816.00,580.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(816.00,579.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(923.00,600.58)(2.083,0.493){23}{\rule{1.731pt}{0.119pt}}
\multiput(923.00,599.17)(49.408,13.000){2}{\rule{0.865pt}{0.400pt}}
\multiput(976.00,613.58)(1.490,0.495){33}{\rule{1.278pt}{0.119pt}}
\multiput(976.00,612.17)(50.348,18.000){2}{\rule{0.639pt}{0.400pt}}
\multiput(1029.00,631.58)(0.935,0.497){55}{\rule{0.845pt}{0.120pt}}
\multiput(1029.00,630.17)(52.247,29.000){2}{\rule{0.422pt}{0.400pt}}
\multiput(1083.58,660.00)(0.498,0.509){103}{\rule{0.120pt}{0.508pt}}
\multiput(1082.17,660.00)(53.000,52.947){2}{\rule{0.400pt}{0.254pt}}
\put(187,322){\usebox{\plotpoint}}
\multiput(187.58,322.00)(0.498,0.919){81}{\rule{0.120pt}{0.833pt}}
\multiput(186.17,322.00)(42.000,75.270){2}{\rule{0.400pt}{0.417pt}}
\multiput(229.00,399.58)(0.731,0.498){71}{\rule{0.684pt}{0.120pt}}
\multiput(229.00,398.17)(52.581,37.000){2}{\rule{0.342pt}{0.400pt}}
\multiput(283.00,436.58)(1.299,0.498){79}{\rule{1.134pt}{0.120pt}}
\multiput(283.00,435.17)(103.646,41.000){2}{\rule{0.567pt}{0.400pt}}
\multiput(389.00,477.58)(2.000,0.497){51}{\rule{1.685pt}{0.120pt}}
\multiput(389.00,476.17)(103.502,27.000){2}{\rule{0.843pt}{0.400pt}}
\multiput(496.00,504.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,503.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,524.58)(3.378,0.494){29}{\rule{2.750pt}{0.119pt}}
\multiput(603.00,523.17)(100.292,16.000){2}{\rule{1.375pt}{0.400pt}}
\multiput(709.00,540.58)(3.204,0.495){31}{\rule{2.618pt}{0.119pt}}
\multiput(709.00,539.17)(101.567,17.000){2}{\rule{1.309pt}{0.400pt}}
\multiput(816.00,557.58)(3.022,0.495){33}{\rule{2.478pt}{0.119pt}}
\multiput(816.00,556.17)(101.857,18.000){2}{\rule{1.239pt}{0.400pt}}
\multiput(923.00,575.58)(2.477,0.492){19}{\rule{2.027pt}{0.118pt}}
\multiput(923.00,574.17)(48.792,11.000){2}{\rule{1.014pt}{0.400pt}}
\multiput(976.00,586.58)(1.929,0.494){25}{\rule{1.614pt}{0.119pt}}
\multiput(976.00,585.17)(49.649,14.000){2}{\rule{0.807pt}{0.400pt}}
\multiput(1029.00,600.58)(1.238,0.496){41}{\rule{1.082pt}{0.120pt}}
\multiput(1029.00,599.17)(51.755,22.000){2}{\rule{0.541pt}{0.400pt}}
\multiput(1083.00,622.58)(0.680,0.498){75}{\rule{0.644pt}{0.120pt}}
\multiput(1083.00,621.17)(51.664,39.000){2}{\rule{0.322pt}{0.400pt}}
\put(187,304){\usebox{\plotpoint}}
\multiput(187.58,304.00)(0.498,0.919){81}{\rule{0.120pt}{0.833pt}}
\multiput(186.17,304.00)(42.000,75.270){2}{\rule{0.400pt}{0.417pt}}
\multiput(229.00,381.58)(0.731,0.498){71}{\rule{0.684pt}{0.120pt}}
\multiput(229.00,380.17)(52.581,37.000){2}{\rule{0.342pt}{0.400pt}}
\multiput(283.00,418.58)(1.299,0.498){79}{\rule{1.134pt}{0.120pt}}
\multiput(283.00,417.17)(103.646,41.000){2}{\rule{0.567pt}{0.400pt}}
\multiput(389.00,459.58)(2.000,0.497){51}{\rule{1.685pt}{0.120pt}}
\multiput(389.00,458.17)(103.502,27.000){2}{\rule{0.843pt}{0.400pt}}
\multiput(496.00,486.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,485.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,506.58)(3.174,0.495){31}{\rule{2.594pt}{0.119pt}}
\multiput(603.00,505.17)(100.616,17.000){2}{\rule{1.297pt}{0.400pt}}
\multiput(709.00,523.58)(3.410,0.494){29}{\rule{2.775pt}{0.119pt}}
\multiput(709.00,522.17)(101.240,16.000){2}{\rule{1.388pt}{0.400pt}}
\multiput(816.00,539.58)(3.204,0.495){31}{\rule{2.618pt}{0.119pt}}
\multiput(816.00,538.17)(101.567,17.000){2}{\rule{1.309pt}{0.400pt}}
\multiput(923.00,556.58)(2.737,0.491){17}{\rule{2.220pt}{0.118pt}}
\multiput(923.00,555.17)(48.392,10.000){2}{\rule{1.110pt}{0.400pt}}
\multiput(976.00,566.58)(2.263,0.492){21}{\rule{1.867pt}{0.119pt}}
\multiput(976.00,565.17)(49.126,12.000){2}{\rule{0.933pt}{0.400pt}}
\multiput(1029.00,578.58)(1.519,0.495){33}{\rule{1.300pt}{0.119pt}}
\multiput(1029.00,577.17)(51.302,18.000){2}{\rule{0.650pt}{0.400pt}}
\multiput(1083.00,596.58)(0.887,0.497){57}{\rule{0.807pt}{0.120pt}}
\multiput(1083.00,595.17)(51.326,30.000){2}{\rule{0.403pt}{0.400pt}}
\put(187,287){\usebox{\plotpoint}}
\multiput(187.58,287.00)(0.498,0.919){81}{\rule{0.120pt}{0.833pt}}
\multiput(186.17,287.00)(42.000,75.270){2}{\rule{0.400pt}{0.417pt}}
\multiput(229.00,364.58)(0.731,0.498){71}{\rule{0.684pt}{0.120pt}}
\multiput(229.00,363.17)(52.581,37.000){2}{\rule{0.342pt}{0.400pt}}
\multiput(283.00,401.58)(1.299,0.498){79}{\rule{1.134pt}{0.120pt}}
\multiput(283.00,400.17)(103.646,41.000){2}{\rule{0.567pt}{0.400pt}}
\multiput(389.00,442.58)(2.000,0.497){51}{\rule{1.685pt}{0.120pt}}
\multiput(389.00,441.17)(103.502,27.000){2}{\rule{0.843pt}{0.400pt}}
\multiput(496.00,469.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,468.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,489.58)(3.174,0.495){31}{\rule{2.594pt}{0.119pt}}
\multiput(603.00,488.17)(100.616,17.000){2}{\rule{1.297pt}{0.400pt}}
\multiput(709.00,506.58)(3.410,0.494){29}{\rule{2.775pt}{0.119pt}}
\multiput(709.00,505.17)(101.240,16.000){2}{\rule{1.388pt}{0.400pt}}
\multiput(816.00,522.58)(3.204,0.495){31}{\rule{2.618pt}{0.119pt}}
\multiput(816.00,521.17)(101.567,17.000){2}{\rule{1.309pt}{0.400pt}}
\multiput(923.00,539.59)(3.465,0.488){13}{\rule{2.750pt}{0.117pt}}
\multiput(923.00,538.17)(47.292,8.000){2}{\rule{1.375pt}{0.400pt}}
\multiput(976.00,547.58)(2.477,0.492){19}{\rule{2.027pt}{0.118pt}}
\multiput(976.00,546.17)(48.792,11.000){2}{\rule{1.014pt}{0.400pt}}
\multiput(1029.00,558.58)(1.714,0.494){29}{\rule{1.450pt}{0.119pt}}
\multiput(1029.00,557.17)(50.990,16.000){2}{\rule{0.725pt}{0.400pt}}
\multiput(1083.00,574.58)(1.067,0.497){47}{\rule{0.948pt}{0.120pt}}
\multiput(1083.00,573.17)(51.032,25.000){2}{\rule{0.474pt}{0.400pt}}
\put(187,268){\usebox{\plotpoint}}
\multiput(187.58,268.00)(0.498,0.919){81}{\rule{0.120pt}{0.833pt}}
\multiput(186.17,268.00)(42.000,75.270){2}{\rule{0.400pt}{0.417pt}}
\multiput(229.00,345.58)(0.731,0.498){71}{\rule{0.684pt}{0.120pt}}
\multiput(229.00,344.17)(52.581,37.000){2}{\rule{0.342pt}{0.400pt}}
\multiput(283.00,382.58)(1.267,0.498){81}{\rule{1.110pt}{0.120pt}}
\multiput(283.00,381.17)(103.697,42.000){2}{\rule{0.555pt}{0.400pt}}
\multiput(389.00,424.58)(2.078,0.497){49}{\rule{1.746pt}{0.120pt}}
\multiput(389.00,423.17)(103.376,26.000){2}{\rule{0.873pt}{0.400pt}}
\multiput(496.00,450.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,449.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,470.58)(2.994,0.495){33}{\rule{2.456pt}{0.119pt}}
\multiput(603.00,469.17)(100.903,18.000){2}{\rule{1.228pt}{0.400pt}}
\multiput(709.00,488.58)(3.644,0.494){27}{\rule{2.953pt}{0.119pt}}
\multiput(709.00,487.17)(100.870,15.000){2}{\rule{1.477pt}{0.400pt}}
\multiput(816.00,503.58)(3.204,0.495){31}{\rule{2.618pt}{0.119pt}}
\multiput(816.00,502.17)(101.567,17.000){2}{\rule{1.309pt}{0.400pt}}
\multiput(923.00,520.59)(3.058,0.489){15}{\rule{2.456pt}{0.118pt}}
\multiput(923.00,519.17)(47.903,9.000){2}{\rule{1.228pt}{0.400pt}}
\multiput(976.00,529.58)(2.737,0.491){17}{\rule{2.220pt}{0.118pt}}
\multiput(976.00,528.17)(48.392,10.000){2}{\rule{1.110pt}{0.400pt}}
\multiput(1029.00,539.58)(2.122,0.493){23}{\rule{1.762pt}{0.119pt}}
\multiput(1029.00,538.17)(50.344,13.000){2}{\rule{0.881pt}{0.400pt}}
\multiput(1083.00,552.58)(1.274,0.496){39}{\rule{1.110pt}{0.119pt}}
\multiput(1083.00,551.17)(50.697,21.000){2}{\rule{0.555pt}{0.400pt}}
\put(187,245){\usebox{\plotpoint}}
\multiput(187.58,245.00)(0.498,0.907){81}{\rule{0.120pt}{0.824pt}}
\multiput(186.17,245.00)(42.000,74.290){2}{\rule{0.400pt}{0.412pt}}
\multiput(229.00,321.58)(0.731,0.498){71}{\rule{0.684pt}{0.120pt}}
\multiput(229.00,320.17)(52.581,37.000){2}{\rule{0.342pt}{0.400pt}}
\multiput(283.00,358.58)(1.267,0.498){81}{\rule{1.110pt}{0.120pt}}
\multiput(283.00,357.17)(103.697,42.000){2}{\rule{0.555pt}{0.400pt}}
\multiput(389.00,400.58)(2.078,0.497){49}{\rule{1.746pt}{0.120pt}}
\multiput(389.00,399.17)(103.376,26.000){2}{\rule{0.873pt}{0.400pt}}
\multiput(496.00,426.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,425.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,446.58)(2.994,0.495){33}{\rule{2.456pt}{0.119pt}}
\multiput(603.00,445.17)(100.903,18.000){2}{\rule{1.228pt}{0.400pt}}
\multiput(709.00,464.58)(3.410,0.494){29}{\rule{2.775pt}{0.119pt}}
\multiput(709.00,463.17)(101.240,16.000){2}{\rule{1.388pt}{0.400pt}}
\multiput(816.00,480.58)(3.204,0.495){31}{\rule{2.618pt}{0.119pt}}
\multiput(816.00,479.17)(101.567,17.000){2}{\rule{1.309pt}{0.400pt}}
\multiput(923.00,497.59)(3.465,0.488){13}{\rule{2.750pt}{0.117pt}}
\multiput(923.00,496.17)(47.292,8.000){2}{\rule{1.375pt}{0.400pt}}
\multiput(976.00,505.58)(2.737,0.491){17}{\rule{2.220pt}{0.118pt}}
\multiput(976.00,504.17)(48.392,10.000){2}{\rule{1.110pt}{0.400pt}}
\multiput(1029.00,515.58)(2.122,0.493){23}{\rule{1.762pt}{0.119pt}}
\multiput(1029.00,514.17)(50.344,13.000){2}{\rule{0.881pt}{0.400pt}}
\multiput(1083.00,528.58)(1.580,0.495){31}{\rule{1.347pt}{0.119pt}}
\multiput(1083.00,527.17)(50.204,17.000){2}{\rule{0.674pt}{0.400pt}}
\put(187,213){\usebox{\plotpoint}}
\multiput(187.58,213.00)(0.498,0.907){81}{\rule{0.120pt}{0.824pt}}
\multiput(186.17,213.00)(42.000,74.290){2}{\rule{0.400pt}{0.412pt}}
\multiput(229.00,289.58)(0.751,0.498){69}{\rule{0.700pt}{0.120pt}}
\multiput(229.00,288.17)(52.547,36.000){2}{\rule{0.350pt}{0.400pt}}
\multiput(283.00,325.58)(1.267,0.498){81}{\rule{1.110pt}{0.120pt}}
\multiput(283.00,324.17)(103.697,42.000){2}{\rule{0.555pt}{0.400pt}}
\multiput(389.00,367.58)(2.078,0.497){49}{\rule{1.746pt}{0.120pt}}
\multiput(389.00,366.17)(103.376,26.000){2}{\rule{0.873pt}{0.400pt}}
\multiput(496.00,393.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,392.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,413.58)(2.833,0.495){35}{\rule{2.332pt}{0.119pt}}
\multiput(603.00,412.17)(101.161,19.000){2}{\rule{1.166pt}{0.400pt}}
\multiput(709.00,432.58)(3.410,0.494){29}{\rule{2.775pt}{0.119pt}}
\multiput(709.00,431.17)(101.240,16.000){2}{\rule{1.388pt}{0.400pt}}
\multiput(816.00,448.58)(3.204,0.495){31}{\rule{2.618pt}{0.119pt}}
\multiput(816.00,447.17)(101.567,17.000){2}{\rule{1.309pt}{0.400pt}}
\multiput(923.00,465.59)(3.058,0.489){15}{\rule{2.456pt}{0.118pt}}
\multiput(923.00,464.17)(47.903,9.000){2}{\rule{1.228pt}{0.400pt}}
\multiput(976.00,474.59)(3.058,0.489){15}{\rule{2.456pt}{0.118pt}}
\multiput(976.00,473.17)(47.903,9.000){2}{\rule{1.228pt}{0.400pt}}
\multiput(1029.00,483.58)(2.122,0.493){23}{\rule{1.762pt}{0.119pt}}
\multiput(1029.00,482.17)(50.344,13.000){2}{\rule{0.881pt}{0.400pt}}
\multiput(1083.00,496.58)(1.797,0.494){27}{\rule{1.513pt}{0.119pt}}
\multiput(1083.00,495.17)(49.859,15.000){2}{\rule{0.757pt}{0.400pt}}
\put(187,159){\usebox{\plotpoint}}
\multiput(187.58,159.00)(0.498,0.931){81}{\rule{0.120pt}{0.843pt}}
\multiput(186.17,159.00)(42.000,76.251){2}{\rule{0.400pt}{0.421pt}}
\multiput(229.00,237.58)(0.773,0.498){67}{\rule{0.717pt}{0.120pt}}
\multiput(229.00,236.17)(52.512,35.000){2}{\rule{0.359pt}{0.400pt}}
\multiput(283.00,272.58)(1.299,0.498){79}{\rule{1.134pt}{0.120pt}}
\multiput(283.00,271.17)(103.646,41.000){2}{\rule{0.567pt}{0.400pt}}
\multiput(389.00,313.58)(2.078,0.497){49}{\rule{1.746pt}{0.120pt}}
\multiput(389.00,312.17)(103.376,26.000){2}{\rule{0.873pt}{0.400pt}}
\multiput(496.00,339.58)(2.714,0.496){37}{\rule{2.240pt}{0.119pt}}
\multiput(496.00,338.17)(102.351,20.000){2}{\rule{1.120pt}{0.400pt}}
\multiput(603.00,359.58)(2.833,0.495){35}{\rule{2.332pt}{0.119pt}}
\multiput(603.00,358.17)(101.161,19.000){2}{\rule{1.166pt}{0.400pt}}
\multiput(709.00,378.58)(3.410,0.494){29}{\rule{2.775pt}{0.119pt}}
\multiput(709.00,377.17)(101.240,16.000){2}{\rule{1.388pt}{0.400pt}}
\multiput(816.00,394.58)(3.022,0.495){33}{\rule{2.478pt}{0.119pt}}
\multiput(816.00,393.17)(101.857,18.000){2}{\rule{1.239pt}{0.400pt}}
\multiput(923.00,412.59)(3.058,0.489){15}{\rule{2.456pt}{0.118pt}}
\multiput(923.00,411.17)(47.903,9.000){2}{\rule{1.228pt}{0.400pt}}
\multiput(976.00,421.58)(2.737,0.491){17}{\rule{2.220pt}{0.118pt}}
\multiput(976.00,420.17)(48.392,10.000){2}{\rule{1.110pt}{0.400pt}}
\multiput(1029.00,431.58)(2.122,0.493){23}{\rule{1.762pt}{0.119pt}}
\multiput(1029.00,430.17)(50.344,13.000){2}{\rule{0.881pt}{0.400pt}}
\multiput(1083.00,444.58)(1.929,0.494){25}{\rule{1.614pt}{0.119pt}}
\multiput(1083.00,443.17)(49.649,14.000){2}{\rule{0.807pt}{0.400pt}}
\end{picture}
{\tenrm\baselineskip=12pt
\noindent Figure 2: The ratio $\delta(x_l,y_l)$ (\ref{delxy}) vs.
$y_l$ for various $x_l$ values. Top curve is for
$x_l = 0.01$, the next for $x_l=0.1$. Each subsequent
curve represents an increase of $x_l$ by $0.1$.}
\vglue 0.2cm
For comparison with the ${\mathrm{e}}$p case we now show in Fig.2 the correction
$\delta(x_l,y_l)$, using (\ref{nsigBorn}), (\ref{bremscross}) and
(\ref{Compton}),
defined by
\begin{equation}
\frac{{\mathrm{d}}^2\sigma^{\mathrm{Br}}}{{\mathrm{d}} x_l {\mathrm{d}} y_l}
+ \frac{{\mathrm{d}}^2\sigma^{\mathrm{C}}}{{\mathrm{d}} x_l {\mathrm{d}} y_l}=
\frac{{\mathrm{d}}^2\sigma^{\mathrm{B}}}{{\mathrm{d}} x_l {\mathrm{d}} y_l}\,\,\delta(x_l,y_l)\ .
\label{delxy}
\end{equation}
Note that we use here the leptonic variables to conform with
the ${\mathrm{e}}$p case.
A closer examination reveals
that initial and final state radiation are similar in order
of magnitude throughout most of the $x,y$ region,
whereas the Compton contributions is
appreciable only in the small and medium $x$, large $y$ region.
Also we note that if one freezes $F_2$ at $F_2(Q_0^2)$ for
$Q^2<Q^2_0$, instead of extrapolating as in (\ref{extrapol}),
we find that only the $x_l=0.01$ curve changes significantly.
It decreases at small and medium $y_l$ by up to 50\%.
Note that inclusion of final state radiation implies a perfect
measurement of the energy and momentum of the tagged lepton,
even in the presence of collinear radiation.
We see in Fig.2 that radiative effects can in principle be
large, around 40\% for medium $x$ and small $y$.
Next we show in Fig.3 the correction factor
$\delta(x_h,Q^2_h)$, defined by
\begin{equation}
\frac{{\mathrm{d}}^2\sigma^{\mathrm{corr}}}{{\mathrm{d}} x_h {\mathrm{d}} Q^2_h} =
\frac{{\mathrm{d}}^2\sigma^{\mathrm{B}}}{{\mathrm{d}} x_h {\mathrm{d}} Q^2_h} \delta(x_h,Q^2_h)\ .
\label{delxQ2},
\end{equation}
in terms of hadronic variables,
as a function of $x_h$ for various choices of
$Q_h^2$, cf. (\ref{hadcorr}). Here only initial state lepton
bremsstrahlung is taken into account, because the scattered lepton
is not used in constructing the kinematic variables.
\vglue 0.5cm
\setlength{\unitlength}{0.240900pt}
\ifx\plotpoint\undefined\newsavebox{\plotpoint}\fi
\sbox{\plotpoint}{\rule[-0.200pt]{0.400pt}{0.400pt}}%
\begin{picture}(1500,900)(0,0)
\font\gnuplot=cmr10 at 10pt
\gnuplot
\sbox{\plotpoint}{\rule[-0.200pt]{0.400pt}{0.400pt}}%
\put(176.0,113.0){\rule[-0.200pt]{0.400pt}{184.048pt}}
\put(176.0,113.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,113){\makebox(0,0)[r]{-0.4}}
\put(1416.0,113.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,209.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,209){\makebox(0,0)[r]{-0.35}}
\put(1416.0,209.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,304.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,304){\makebox(0,0)[r]{-0.3}}
\put(1416.0,304.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,399.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,399){\makebox(0,0)[r]{-0.25}}
\put(1416.0,399.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,495.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,495){\makebox(0,0)[r]{-0.2}}
\put(1416.0,495.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,590.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,590){\makebox(0,0)[r]{-0.15}}
\put(1416.0,590.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,686.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,686){\makebox(0,0)[r]{-0.1}}
\put(1416.0,686.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,782.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,782){\makebox(0,0)[r]{-0.05}}
\put(1416.0,782.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,877.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(154,877){\makebox(0,0)[r]{0}}
\put(1416.0,877.0){\rule[-0.200pt]{4.818pt}{0.400pt}}
\put(176.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(176,68){\makebox(0,0){0}}
\put(176.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(302.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(302,68){\makebox(0,0){0.1}}
\put(302.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(428.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(428,68){\makebox(0,0){0.2}}
\put(428.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(554.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(554,68){\makebox(0,0){0.3}}
\put(554.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(680.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(680,68){\makebox(0,0){0.4}}
\put(680.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(806.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(806,68){\makebox(0,0){0.5}}
\put(806.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(932.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(932,68){\makebox(0,0){0.6}}
\put(932.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1058.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1058,68){\makebox(0,0){0.7}}
\put(1058.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1184.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1184,68){\makebox(0,0){0.8}}
\put(1184.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1310.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1310,68){\makebox(0,0){0.9}}
\put(1310.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1436.0,113.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(1436,68){\makebox(0,0){1}}
\put(1436.0,857.0){\rule[-0.200pt]{0.400pt}{4.818pt}}
\put(176.0,113.0){\rule[-0.200pt]{303.534pt}{0.400pt}}
\put(1436.0,113.0){\rule[-0.200pt]{0.400pt}{184.048pt}}
\put(176.0,877.0){\rule[-0.200pt]{303.534pt}{0.400pt}}
\put(806,23){\makebox(0,0){$x_h$}}
\put(176.0,113.0){\rule[-0.200pt]{0.400pt}{184.048pt}}
\put(189,698){\usebox{\plotpoint}}
\multiput(189.00,698.58)(0.625,0.498){77}{\rule{0.600pt}{0.120pt}}
\multiput(189.00,697.17)(48.755,40.000){2}{\rule{0.300pt}{0.400pt}}
\multiput(239.00,738.58)(2.479,0.493){23}{\rule{2.038pt}{0.119pt}}
\multiput(239.00,737.17)(58.769,13.000){2}{\rule{1.019pt}{0.400pt}}
\multiput(302.00,751.59)(6.944,0.477){7}{\rule{5.140pt}{0.115pt}}
\multiput(302.00,750.17)(52.332,5.000){2}{\rule{2.570pt}{0.400pt}}
\multiput(365.00,756.60)(9.108,0.468){5}{\rule{6.400pt}{0.113pt}}
\multiput(365.00,755.17)(49.716,4.000){2}{\rule{3.200pt}{0.400pt}}
\put(428,760.17){\rule{12.700pt}{0.400pt}}
\multiput(428.00,759.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\put(491,762.17){\rule{12.700pt}{0.400pt}}
\multiput(491.00,761.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\multiput(554.00,764.61)(13.858,0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(554.00,763.17)(45.358,3.000){2}{\rule{4.250pt}{0.400pt}}
\put(617,765.67){\rule{15.177pt}{0.400pt}}
\multiput(617.00,766.17)(31.500,-1.000){2}{\rule{7.588pt}{0.400pt}}
\multiput(680.00,766.59)(6.944,0.477){7}{\rule{5.140pt}{0.115pt}}
\multiput(680.00,765.17)(52.332,5.000){2}{\rule{2.570pt}{0.400pt}}
\put(743,771.17){\rule{12.700pt}{0.400pt}}
\multiput(743.00,770.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\put(806,772.67){\rule{15.177pt}{0.400pt}}
\multiput(806.00,772.17)(31.500,1.000){2}{\rule{7.588pt}{0.400pt}}
\put(932,773.67){\rule{15.177pt}{0.400pt}}
\multiput(932.00,773.17)(31.500,1.000){2}{\rule{7.588pt}{0.400pt}}
\multiput(995.00,773.95)(13.858,-0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(995.00,774.17)(45.358,-3.000){2}{\rule{4.250pt}{0.400pt}}
\multiput(1058.00,772.61)(13.858,0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(1058.00,771.17)(45.358,3.000){2}{\rule{4.250pt}{0.400pt}}
\put(1121,775.17){\rule{12.700pt}{0.400pt}}
\multiput(1121.00,774.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\multiput(1184.00,775.95)(13.858,-0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(1184.00,776.17)(45.358,-3.000){2}{\rule{4.250pt}{0.400pt}}
\multiput(1247.00,774.61)(13.858,0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(1247.00,773.17)(45.358,3.000){2}{\rule{4.250pt}{0.400pt}}
\put(1310,777.17){\rule{12.700pt}{0.400pt}}
\multiput(1310.00,776.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\put(869.0,774.0){\rule[-0.200pt]{15.177pt}{0.400pt}}
\put(189,543){\usebox{\plotpoint}}
\multiput(189.58,543.00)(0.498,0.993){97}{\rule{0.120pt}{0.892pt}}
\multiput(188.17,543.00)(50.000,97.149){2}{\rule{0.400pt}{0.446pt}}
\multiput(239.00,642.58)(1.132,0.497){53}{\rule{1.000pt}{0.120pt}}
\multiput(239.00,641.17)(60.924,28.000){2}{\rule{0.500pt}{0.400pt}}
\multiput(302.00,670.58)(2.139,0.494){27}{\rule{1.780pt}{0.119pt}}
\multiput(302.00,669.17)(59.306,15.000){2}{\rule{0.890pt}{0.400pt}}
\multiput(365.00,685.59)(3.640,0.489){15}{\rule{2.900pt}{0.118pt}}
\multiput(365.00,684.17)(56.981,9.000){2}{\rule{1.450pt}{0.400pt}}
\multiput(428.00,694.59)(6.944,0.477){7}{\rule{5.140pt}{0.115pt}}
\multiput(428.00,693.17)(52.332,5.000){2}{\rule{2.570pt}{0.400pt}}
\multiput(491.00,699.59)(5.644,0.482){9}{\rule{4.300pt}{0.116pt}}
\multiput(491.00,698.17)(54.075,6.000){2}{\rule{2.150pt}{0.400pt}}
\multiput(554.00,705.59)(5.644,0.482){9}{\rule{4.300pt}{0.116pt}}
\multiput(554.00,704.17)(54.075,6.000){2}{\rule{2.150pt}{0.400pt}}
\multiput(680.00,711.60)(9.108,0.468){5}{\rule{6.400pt}{0.113pt}}
\multiput(680.00,710.17)(49.716,4.000){2}{\rule{3.200pt}{0.400pt}}
\put(617.0,711.0){\rule[-0.200pt]{15.177pt}{0.400pt}}
\multiput(806.00,715.60)(9.108,0.468){5}{\rule{6.400pt}{0.113pt}}
\multiput(806.00,714.17)(49.716,4.000){2}{\rule{3.200pt}{0.400pt}}
\multiput(869.00,719.61)(13.858,0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(869.00,718.17)(45.358,3.000){2}{\rule{4.250pt}{0.400pt}}
\put(932,722.17){\rule{12.700pt}{0.400pt}}
\multiput(932.00,721.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\put(995,722.67){\rule{15.177pt}{0.400pt}}
\multiput(995.00,723.17)(31.500,-1.000){2}{\rule{7.588pt}{0.400pt}}
\multiput(1058.00,723.61)(13.858,0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(1058.00,722.17)(45.358,3.000){2}{\rule{4.250pt}{0.400pt}}
\put(1121,725.67){\rule{15.177pt}{0.400pt}}
\multiput(1121.00,725.17)(31.500,1.000){2}{\rule{7.588pt}{0.400pt}}
\put(743.0,715.0){\rule[-0.200pt]{15.177pt}{0.400pt}}
\multiput(1247.00,727.61)(13.858,0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(1247.00,726.17)(45.358,3.000){2}{\rule{4.250pt}{0.400pt}}
\put(1184.0,727.0){\rule[-0.200pt]{15.177pt}{0.400pt}}
\multiput(194.58,113.00)(0.498,3.467){87}{\rule{0.120pt}{2.856pt}}
\multiput(193.17,113.00)(45.000,304.073){2}{\rule{0.400pt}{1.428pt}}
\multiput(239.58,423.00)(0.499,0.627){123}{\rule{0.120pt}{0.602pt}}
\multiput(238.17,423.00)(63.000,77.751){2}{\rule{0.400pt}{0.301pt}}
\multiput(302.00,502.58)(0.930,0.498){65}{\rule{0.841pt}{0.120pt}}
\multiput(302.00,501.17)(61.254,34.000){2}{\rule{0.421pt}{0.400pt}}
\multiput(365.00,536.58)(1.593,0.496){37}{\rule{1.360pt}{0.119pt}}
\multiput(365.00,535.17)(60.177,20.000){2}{\rule{0.680pt}{0.400pt}}
\multiput(428.00,556.58)(1.881,0.495){31}{\rule{1.582pt}{0.119pt}}
\multiput(428.00,555.17)(59.716,17.000){2}{\rule{0.791pt}{0.400pt}}
\multiput(491.00,573.58)(2.693,0.492){21}{\rule{2.200pt}{0.119pt}}
\multiput(491.00,572.17)(58.434,12.000){2}{\rule{1.100pt}{0.400pt}}
\multiput(554.00,585.59)(4.126,0.488){13}{\rule{3.250pt}{0.117pt}}
\multiput(554.00,584.17)(56.254,8.000){2}{\rule{1.625pt}{0.400pt}}
\multiput(617.00,593.59)(3.640,0.489){15}{\rule{2.900pt}{0.118pt}}
\multiput(617.00,592.17)(56.981,9.000){2}{\rule{1.450pt}{0.400pt}}
\multiput(680.00,602.59)(5.644,0.482){9}{\rule{4.300pt}{0.116pt}}
\multiput(680.00,601.17)(54.075,6.000){2}{\rule{2.150pt}{0.400pt}}
\multiput(743.00,608.59)(6.944,0.477){7}{\rule{5.140pt}{0.115pt}}
\multiput(743.00,607.17)(52.332,5.000){2}{\rule{2.570pt}{0.400pt}}
\multiput(806.00,613.59)(6.944,0.477){7}{\rule{5.140pt}{0.115pt}}
\multiput(806.00,612.17)(52.332,5.000){2}{\rule{2.570pt}{0.400pt}}
\multiput(869.00,618.60)(9.108,0.468){5}{\rule{6.400pt}{0.113pt}}
\multiput(869.00,617.17)(49.716,4.000){2}{\rule{3.200pt}{0.400pt}}
\multiput(932.00,622.59)(5.644,0.482){9}{\rule{4.300pt}{0.116pt}}
\multiput(932.00,621.17)(54.075,6.000){2}{\rule{2.150pt}{0.400pt}}
\put(995,627.67){\rule{15.177pt}{0.400pt}}
\multiput(995.00,627.17)(31.500,1.000){2}{\rule{7.588pt}{0.400pt}}
\put(1058,629.17){\rule{12.700pt}{0.400pt}}
\multiput(1058.00,628.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\multiput(1121.00,631.59)(6.944,0.477){7}{\rule{5.140pt}{0.115pt}}
\multiput(1121.00,630.17)(52.332,5.000){2}{\rule{2.570pt}{0.400pt}}
\multiput(1184.00,636.61)(13.858,0.447){3}{\rule{8.500pt}{0.108pt}}
\multiput(1184.00,635.17)(45.358,3.000){2}{\rule{4.250pt}{0.400pt}}
\put(1247,638.67){\rule{15.177pt}{0.400pt}}
\multiput(1247.00,638.17)(31.500,1.000){2}{\rule{7.588pt}{0.400pt}}
\put(1310,640.17){\rule{12.700pt}{0.400pt}}
\multiput(1310.00,639.17)(36.641,2.000){2}{\rule{6.350pt}{0.400pt}}
\put(1310.0,730.0){\rule[-0.200pt]{15.177pt}{0.400pt}}
\end{picture}
{\tenrm\baselineskip=12pt
\noindent Figure 3: The ratio $\delta(x_h,Q^2_h)$ vs.
$x_h$ for three $Q^2_h$ values. Top curve: $Q^2_h = 1 \, {\rm GeV^2}$.
Middle curve: $Q^2_h = 10\, {\rm GeV^2}$. Lower curve:
$Q^2_h = 100\, {\rm GeV^2}$.}
\vglue 0.5cm
We see that the corrections are sizable only for large $Q^2$.
Using however the resummed version of (\ref{resummed2}) and
(\ref{resummedL}), we find
that the corrections are reduced by about an order
of magnitude.
Thus we conclude that the size of the radiative corrections depends
significantly on the set of variables chosen. These corrections
can in principle be quite large. As noted before, in practice
mixed variables are used. In view of the results obtained in this paper,
we think a more careful study, involving Monte Carlo simulation of
the full final state, is warranted.
|
\section{1. Introduction}
In this paper, we study a general class of two-dimensional higher-derivative
gravitational theories in a Riemann-Cartan spacetime. As is well known, in two
dimensions the Einstein-Hilbert action is a total derivative and hence cannot
be used to construct a two-dimensional version of general relativity. It is
however possible to construct actions whose lagrangian density is given by
an arbitrary power of the Ricci scalar [1]. These models turn out to be
equivalent to the ordinary Einstein-Hilbert action non-minimally coupled to
a scalar field with a power-law potential [2-4]. In such formalism the field
equations become second order and it is possible to obtain exact solutions
and to perform the Dirac quantization of the theory [5]. Some well-known
special cases of gravity-scalar theories in two dimensions
include the Jackiw-Teitelboim [6] and the "string" models [7-8].
Another useful generalization of two-dimensional gravity is given by the
consideration of \RC geometries with non-trivial torsion. Several authors
have studied various aspects of a model with action quadratic in
the curvature and the torsion [9-15]. It has been proved that this model is
completely integrable [9-10] and, exploiting the hamiltonian formalism, it has
been shown that its symmetries are a realization of a specific non-linear
algebra [11-12] and that its quantization can be performed exactly [13].
Finally, a further interesting aspect of some two-dimensional models is that
they can be interpreted as gauge theories
of the \poi or de Sitter group in two dimensions [16]. In ref. [14] it has
been shown that this interpretation can be extended to theories with
non-trivial torsion, provided that one generalizes the notion of gauge
invariance
to groups generated by non-linear algebras.
In this paper we try to extend the results obtained so far to the case of
actions
containing arbitrary powers of the curvature and quadratic torsion.
We show that the static solutions of the field equations can be found exactly
and discuss their geometry. We also investigate the hamiltonian formulation
of the models and find the constraint algebra, which is a non-linear
deformation of the two-dimensional \poi algebra and therefore permits their
interpretation as a non-linear gauge theory of such algebra.
Finally, we perform the Dirac quantization of the
model and define the space of the physical states.
\section{2. The action and the field equations}
We consider a 2-dimensional lorentzian manifold with signature $(-,+)$
endowed with a \RC geometry.
The geometry can be described by the zweibein field $e^a_\m$ and the Lorentz
connection $\conn{\m}$, where $a,b,..$ are tangent space indices which can
take the values $0$, $1$ and $\m,\n,...$ are world indices, whose values will
be denoted by $t$, $x$.
The curvature and torsion are defined as
$$\eqalign{&R\ba_\mn=\de_\m\conn{\n}-\de_\n\conn{\m}\cr
&T^a_\mn=\de_\m e^a_\n-\de_\n e^a_\m+\o{^a_{\ b}}}\def\bam{{_a^{\ b}}}\def\ba{^{ab}_\m e^b_\n-\o{^a_{\ b}}}\def\bam{{_a^{\ b}}}\def\ba{^{ab}_\n e^b_\m\cr}
\eqno(1)$$
and the Ricci scalar as $R=e_a^\m e_b^\n R\ba_\mn$. In two
dimensions the Ricci scalar determines uniquely the Riemann tensor by
the relation $\curv abcd=-\ha\e_{ab}\e_{cd}R$.
Moreover, the connection can be written as
$$\conn{\m}=\e\ba\o_\m$$
where $\e\ba$ is the antisymmetric tensor $\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r}=-\e\ba$, $\e_{01}=1$.
In this paper, we study higher derivative actions of the form:
$$S=\edx (R^k-\gh T^2)\eqno(2)$$
where $e=$det $e^a_\m$, $T^2=\tors abc T^{abc}$.
Here, $k$ is any real number except $0$, $1$ and $\g$ a coupling constant.
These actions generalize to \RC geometry the higher-derivative actions
introduced in [1] and further investigated in [2-5]. The special case $k=2$ has
already been studied by several authors [9-15].
By defining a scalar field $\y=kR^{k-1}$, it is possible, by a standard
argument [2-4] to reduce the action (2) to a form which is linear in the
curvature:
$$S'=\edx\left(\y R-\gh T^2+\L\y^h\right)\eqno(3)$$
where $\L=(1-k)k^{-k/(k-1)}$ and $h={k\over k-1}$.
The action can be further reduced to a fully first order form, by introducing
a doublet of scalar fields $\y_a$ [13]:
$$S''=\edx\left(\y R+\y_a\, ^*T^a-\gi\y_a\y^a+\L\y^h\right)\eqno(4)$$
where $^*T^a=\e^{bc}\tors abc$.
The field equations obtained by varying (3) with respect to the fields
$\y$, $e^a$ and $\o$ can be written as:
$$\eqalign{&R+h\L\y^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}=0\cr
&-\na^b\tors cdb +T\ba_{\ \ c}\tors abd-{1\over 4}}\def\di{{\rm d}}\def\pro{\propto}\def\app{\approx g_{cd}(T^2-{2\L\over\g}\y^h)=0\cr
&\e^a_{\ c}e_a^\m\de_\m\y-\g\e\ba\tors abc=0\cr}\eqno(5)$$
\section{3. The static solutions}
In the following, we shall look for the static solutions of these equations.
According to the discussion of the special case $k=2$
performed in ref. [9], it results convenient
to seek for the solutions in a conformal gauge. We therefore adopt the ansatz:
$$e^0_t=e^1_x=e^{2\r(x)}\qquad\qquad e^0_x=e^1_t=0$$
We also assume that $\o_\m=\ep\de^\n\c$, with $\c=\c(x)$, which yields
$$\o_x=0,\qquad\qquad \o_t=\c'$$
where a prime denotes derivative with respect to $x$.
In terms of these variables, one has:
$$R=2e^{-2\r}\c'',\qquad\qquad\tors 001=e^{-\r}(\r'-\c')\eqno(6)$$
and the field equations become:
$$\c''+h{\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x\y^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}\erp=0\eqno(7.a)$$
$$\y'-\g(\r'-\c')=0\eqno(7.b)$$
$$\r''-\c''-\ha(\r'-\c')(\r'+\c')+{\L\over 2\g}\y^h\erp=0\eqno(7.c)$$
$$\ha(\r'-\c')(\r'+\c')+{\L\over 2\g}\y^h\erp=0\eqno(7.d)$$
These equations admit a special solution $\c=\r=\y=0$,
corresponding to a manifold with vanishing
curvature and torsion (we are cosidering the case of vanishing cosmological
constant). The zero torsion solutions do not therefore reduce to the extremals
of the action (3) with $\g=0$, which have been discussed in [5]. This fact has
been observed in [9] for the special case $k=2$.
One can however obtain more general solutions to (7).
Combining (7.c) and (7.d) one gets
$$\r''-\c''=(\r'-\c')(\r'+\c')=-{\L\over\g}\y^h\erp\eqno(8)$$
A first integral of the first equation (8) is
$$\r=\ha(f+\ln(Ef'))\eqno(9)$$
where $f\equiv\r-\c$ and $E$ is an integration constant.
{}From (7.b), one has $\y=\g f$ and thus (8) yields
$$f''=\L E\g^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g} e^ff^hf'\eqno(10)$$
which can be integrated to give
$$f'=\L E\g^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}\int^f_0 g^he^gdg\ +A\eqno(11)$$
where the integral on the r.h.s. is proportional to the incomplete gamma
function $\G(h+1,-f)$ and $A$ is an integration constant. The equation (11)
can then be integrated numerically
to give $f$. This solution generalizes that obtained in ref. [9] for $h=2$.
One can now express the curvature and the torsion in terms of the function
$f$:
$$R=-\L(\g f)^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}\qquad\qquad T^2=e^{-f}f'\eqno(12)$$
It is then easy to study the qualitative behaviour of the solutions of (11).
We shall impose the positivity of $f$ in order to avoid problems when $h$ is
not integer. From the asymptotics of (11), follows that $f\to\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$ for a finite
value of $x$. If $h>1$, a curvature singularity is therefore always present
at finite $x$ in these coordinates, while the scalar $T^2$ is singular if
$h>0$ as $f\to\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$.
A numerical study permits to distinguish three different possible behaviours
for $f$ (see fig. 1-3):
a) If $h>-1$ and $A>0$, $f$ grows monotonically from $0$ to $\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$ between
two finite values \xt\ and $x_1$}\def\xt{$x_2$}\def\conn#1{\o\ba_{\ \ #1}\ of $x$.
b) If $h>-1$ and $A<0$, the solution has two branches: one of them decreases
monotonically between a constant value $f_0$ at $x=-\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$ and 0 at $x=$ $x_1$}\def\xt{$x_2$}\def\conn#1{\o\ba_{\ \ #1},
while the other grows monotonically between $f_0$ at $-\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$ and infinity at
$x=$\xt.
c) If $h<-1$, for any $A$, the behaviour of $f$ is similar to the case b),
but $f'(x_2)\to-\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$.
In order to investigate the properties of the solutions near the critical
point, one must study the behavior of the functions $R$, $T^2$ and $\erp$
near $x_1$}\def\xt{$x_2$}\def\conn#1{\o\ba_{\ \ #1}\ and \xt. It is easy to check that for $f\to\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$, $R\sim f^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}$,
$T^2\sim f^h$, $\erp\sim e^{2f}$, while for $f\to 0$, $R\sim f^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}$,
$T^2\sim\const +f^{h+1}$, $\erp\sim\const +f^{h+1}$.
Depending on the value of $h$, one can then distinguish several cases
which are summarized in the tables 1-3. The following general results can be
stated:
If $h>-1$ and $A>0$, a naked singularity is always present, either at $x_1$}\def\xt{$x_2$}\def\conn#1{\o\ba_{\ \ #1}\ or
at \xt.
More interesting are the cases where $h>-1$ and $A<0$ or $h>-1$. In these
cases, the two branches of the solution describe the interior and the exterior
of an asymptotically flat }\def\hd{higher derivative }\def\st{spacetime \bh with the horizon located at $x=-\infty}\def\id{\equiv}\def\mo{{-1}}\def\ha{{1\over 2}$. For $h<1$ (resp. $h>1$),
the curvature singularity is at $x_1$}\def\xt{$x_2$}\def\conn#1{\o\ba_{\ \ #1} (resp. \xt), while spatial infinity is at
\xt\ (resp. $x_1$}\def\xt{$x_2$}\def\conn#1{\o\ba_{\ \ #1}). For $0<h<1$, however, the torsion diverges at spatial
infinity.
Of course, a detailed study of the spacetime structure would require
a more explicit form of the solutions.
For the special cases $h=0$ and $h=1$, this will be afforded in a future
paper.
\section{4. First order formalism}
In terms of differential forms, $e^a=e^a_\m dx^\m$, $\o=\o_\m dx^\m$,
the first order lagrangian in (4) can be written as:
$$\ha{\cal L}}\def\de{\partial}\def\na{\nabla}\def\per{\times = \y_2 d\o+\y_aT^a+\left(\Lq\y_2^h+{1\over 2\g}\y_c\y^c\right)\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r}
e^ae^b
\eqno(13)$$
where we have renamed $\y$ as $\y_2$.
This form of the lagrangian
is especially convenient because it permits to evidentiate the
connection of our models with the formulation of 2-dimensional gravity as a
gauge theory of the Poincar\'e group ISO(1,1) or one of its generalizations
[16,8].
In this formalism, $e^a$ and $\o$ play the role of gauge connections, while
the $\y$ are considered as auxiliary fields.
Local Poincar\'e transformations }\def\ther{thermodynamical }\def\coo{coordinates with parameters $\x^2$ and $\x^a$, corresponding to
Lorentz rotations and to translations, act infinitesimally on the
fields according to:
$$\d e^a=d\x^a+\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r}(\x^b\o-\x^2e^b)\qquad\qquad\d\o=d\x^2$$
$$\d\y_a=\e\bam\x^2\y_b\qquad\qquad\d\y_2=\e\bam\x^a\y_b$$
and $R=d\o$ and $T^a$ are the field strengths corresponding to Lorentz
rotations and translations respectively. The first two terms in the lagrangian
(13) are invariant under these transformations }\def\ther{thermodynamical }\def\coo{coordinates , while the potential terms are not.
As we shall see, the full action is in fact invariant under a non-linear
generalization of the \poi group.
In first order formalism, the field equations are given by:
$$\eqalign{&T^a+\gi\y^a\e_{bc}e^be^c=0\cr
&d\o+\Lq\y_2^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r} e^ae^b=0\cr
&d\y_a+\y_b\e\bam\o+\left({\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x\y_2^h+\gi\y_c\y^c\right)\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r} e^b=0\cr
&d\y_2+\y_a\e{^a_{\ b}}}\def\bam{{_a^{\ b}}}\def\ba{^{ab}\e^b=0\cr}\eqno(14)$$
whose static solutions can be written in terms of the function $f$ defined
above as:
$$\e^0_t=e^1_x=\sqrt{Ef'e^f}\qquad\qquad$$
$$\o_t={\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x E\g^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g} f^he^f-\ha f'\qquad\qquad\o_x=0$$
$$\y_0=\g\sqrt{f'\over Ee^f}\qquad\qquad\y_1=0\qquad\qquad\y_2=\g f$$
\section{5. The hamiltonian formalism}
Another advantage of the first order formalism is that it leads naturally
to a hamiltonian formulation of the model, and hence permits a
straightforward discussion of its symmetries and quantization.
In fact, after integration by parts, the lagrangian
density can be written as:
$$\eqalign{\ha{\cal L}}\def\de{\partial}\def\na{\nabla}\def\per{\times=&\y_a\dot e^a_x+\y_2\dot\o_x\cr
+&e^a_t(\y_a'+\e\bam\y_b\o_x+{\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x\y_2^h\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r} e_x^b+\gi \y_c\y^c\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r} e_x^b)
+\o_t(\y_2'+\y_a\e{^a_{\ b}}}\def\bam{{_a^{\ b}}}\def\ba{^{ab} e^b_x)\cr}\eqno(15)$$
where a dot denotes time derivative and a prime spatial derivative.
The lagrangian (15) has a canonical structure, with coordinates $(e^a_x,
\o_x)$, conjugate momenta $(\y_a,\y_2)$ and Lagrange multipliers $(e^a_t,
\o_t)$ enforcing the constraints:
$$G_a=\y_a'+\e\bam\y_b\o_x+\left({\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x\y_2^h+\gi \y_c\y^c\right)\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r} e_x^b=0$$
$$G_2=\y_2'+\y_a\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r} e^b_x=0\eqno(16)$$
Combining the two constraints (16), one can deduce that
$$\ha(\y_a\y^a)'-{\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x\y_2^h\y_2'-\gi\y_a\y^a\y_2'=0\eqno(17)$$
which implies the existence of the conserved quantity
$$Q\id\y_a\y^ae^{-2\y_2/\g}-\L\left({\g\over 2}\right)^\hp\G\left(h+1,
{2\y_2\over\g}\right)\eqno(18)$$
with $\G$ the incomplete gamma function.
The study of the algebra of constraints permits to discuss the symmetries
of the theory. The calculation of the Poisson brackets of the constraints
yields
$$\{G_a,G_2\}=\e\bam G_b\qquad \{G_a,G_b\}=\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r}{\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x h\y_2^{h-1}G_2+\gi\y_cG_c
\eqno(19)$$
with coordinate dependent structure functions.
This algebra acts locally on the fields by the infinitesimal transformations }\def\ther{thermodynamical }\def\coo{coordinates :
$$\d e^a=d\x^a+\e{^a_{\ b}}}\def\bam{{_a^{\ b}}}\def\ba{^{ab}(\x^b\o-\x^2e^b)-\gi\e_{bc}\x^be^c\y^a\qquad\qquad
\d\o=d\x^2-{\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x h\y_2^{h-1}}\def\hp{{h+1}}\def\Lq{{\L\over 4}}\def\gi{{1\over\g}\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r}\x^a e^b$$
$$\d\y_a=\e\bam\left[\x^2\y_b+\x_b\left({\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x\y_2^h+\gi\y^a\y_a\right)\right]
\qquad\qquad\d\y_2=\e\bam\x^a\y_b$$
as can be checked by computing the commutators $\d e^a=\{G_a,e^b\}$, etc.
The lagrangian (13) is invariant under these transformations up to a total
derivative. Our model can therefore be considered as
a gauge theory of the group generated by the non-linear algebra (19),
realized by means of its action on the lagrangian (13).
The generalization of the usual gauge theories to non-linear algebras has
been introduced in [14], where also the special case $h=2$ of our model has
been examined.
It must be noticed, however, that in this form the algebra fails to close.
In order to construct a closed algebra one has to include in it also the
fields $\y_i$ ($i=0,1,2)$ and to consider the family of generators
$A(\y_i)+B(\y_i)G_i$, with $A$, $B$ analytic functions of $\y_i$ [11].
One has then:
$$\{\y_a,\y_2\}=\{\y_a,\y_b\}=0$$
$$\{G_2,\y_2\}=0\qquad\qquad\{G_a,\y_b\}=\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r}\left({\L\over 2}}\def\gh{{\g\over 2}}\def\edx{\int e\ d^2x\y_2^h+\gi\y_c\y^c
\right)\eqno(20)$$
$$\{G_2,\y_a\}=-\{G_a,\y_2\}=\e\bam\y_b$$
The resulting algebra is a nonlinear deformation of $iso(1,2)$ of the kind
discussed in [17].
\section{6. Dirac quantization}
The model can now be quantized in the Dirac formalism, by replacing the
Poisson brackets with commutators and imposing the Gauss law on the physical
states. In a momentum representation for the wave functional,
$e^a\to i{d\over d\y_a}$, $\o\to i{d\over d\y_2}$, the constraint equations
become:
$$\left[\y_a'+i\e\bam\y_b{\de\over\de\y_2}+i\e_{ab}}\def\ef{e^{-2\f}}\def\er{e^{-2\r}}\def\erp{e^{2\r}\left(\Lq\y_2^h+{1\over 2\g}
\y_c\y^c\right){\de\over\de\y_b}\right]\Q(\y_a,\y_2)=0\eqno(21)$$
$$\left(\y_2'+i\e\bam\y_a{\de\over\de\y_b}\right)\Q(\y_a,\y_2)=0
\eqno(22)$$
The solution of these equations can be written as:
$$\Q=\d(Q')e^{i\Omega}\def\P{\Pi}\def\Q{\Psi}\def\S{\Sigma}\def\U{\Upsilon}\def\X{\Xi}\q(Q)\eqno(23)$$
where $Q$ is given in (18) and
$$\Omega}\def\P{\Pi}\def\Q{\Psi}\def\S{\Sigma}\def\U{\Upsilon}\def\X{\Xi=\int{\e\ba\y_2\y_ad\y_b\over\y^c\y_c}\eqno(24)$$
The parameters $\L$ and $\g$ enter in (23) only through the parameter $Q$,
which classifies the quantum states. Some special
cases of the solution (23) have been obtained in [5,13,18].
It should be pointed out, however, that one can not straightforwardly
define a Schr\"od- inger
equation, since due to the constraints (21,22), the hamiltonian vanishes
on the physical states. This is a well-known problem in the hamiltonian
quantization of gravity and can be solved by fixing a gauge: in a
two-dimensional context it has been treated in ref. [13].
\section{7. Conclusions}
We have shown that most of the results obtained in two-dimensional gravity
with quadratic curvature and torsion can be extended to the case of an action
containing arbitrary powers of the curvature scalar. A possible generalization
of these results would be to consider actions containing arbitrary functions of
of the curvature, which can be treated essentially by the same methods used
here [3,4].
Another interesting point would be to find the most general solutions of the
field equations, including time-dependent ones, which we have not considered.
It seems plausible that this can be achieved by means of a suitable
generalization of the procedure followed in ref. [9] and [10] in the special
case of quadratic curvature.
\beginref
\ref [1] H.J. Schmidt, \JMP{32}, 1562 (1991);
\ref [2] S. Mignemi, \PR{D50}, 4733 (1994);
\ref [3] S.N. Solodukhin, \PR{D51}, 591 (1995);
\ref [4] S. Mignemi and H.-J. Schmidt, \CQG{12}, 845 (1995);
\ref [5] S. Mignemi, in press on Ann. Phys.;
\ref [6] C. Teitelboim, in {\sl Quantum Theory of gravity}, S.M. Christensen,
ed. (Adam Hilger, Bristol, 1984); R. Jackiw, {\sl ibidem};
\ref [7] G. Mandal, A.M. Sengupta and S.R. Wadia, \MPL{A6}, 1685 (1991);
\ref [8] D. Cangemi and R. Jackiw, \PRL{69}, 233 (1992);
R. Jackiw, Theor. Math. Phys. {\bf 9}, 404 (1992);
\ref [9] M.O. Katanaev and I.V. Volovich, \PL{B175}, 413 (1986); \AP{N.Y.}
{197}, 1 (1990); M.O. Katanaev, \JMP{31}, 882 (1990); {\bf 32}, 2483 (1991);
\ref [10] W. Kummer and D.J. Schwarz, \PR{D45}, 3628 (1992);
\ref [11] H.Grosse, W. Kummer, P. Pre\v snajder and D.J. Schwarz,
\JMP{33}, 3892 (1992);
\ref [12] T. Strobl, \IJMP{A8}, 1383 (1993);
\ref [13] T. Strobl, \IJMP{D3}, 281 (1994);
\ref [14] N. Ikeda, \AP{N.Y.}{235}, 435 (1994); N. Ikeda and K.I. Izawa,
\PTP{89},
223 (1993); {\bf 89}, 1077 (1993); {\bf 90}, 237 (1993);
\ref [15] K.G. Akdeniz, A. Kizilers\"u and A. Rizao\v glu, \PL{B215}, 81
(1988);
K.G. Akdeniz, \"O.F. Dayi and A. Kizilers\"u, \MPL{A7}, 1757 (1992);
\ref [16] T. Fukuyama and K. Kamimura, \PL{B160}, 259 (1985);
K. Isler and C.A. Trugenberger, \PRL{63}, 834 (1989);
A.H. Chamseddine and D. Wyler, \PL{B228}, 75 (1989);
\ref [17] K. Schoutens, A. Sevrin and P. van Nieuwenhuizen, \CMP{124}, 87
(1989);
\ref [18] D. Cangemi and R. Jackiw, \PR{D50}, 3916 (1994).
\par\endgroup
\end
|
\section{Introduction}
A few years ago, a line of investigation has been proposed by Avdeev
and Chizhov \cite{ac} that consists in treating skew-symmetric rank-2
tensor fields as matter rather than gauge degrees of freedom. The model
studied in Ref.\cite{ac} has been further reassessed from the point of
view of renormalization in the framework of BRS quantization \cite{nos}.
In view of the potential relevance of matter-like tensor fields for
phenomenology \cite{ac}, it is our purpose in this paper to discuss
some facts concerning the formulation of an $N=1$ supersymmetric
Abelian gauge model realizing the coupling of gauge fields to
matter tensor fields and their partners. One intends here to present a
superspace formulation of the model and exploit the possible relevance
of extra bosonic supersymmetric partners (complex scalars) for the
issue of symmetry breaking. We would like to mention that the
supersymmetrization of 2-forms that appear as gauge fields is already
known in connection with supergravity, and the so-called linear
superfields appear to be the most appropriate multiplets to accomodate
the 2-form gauge fields \cite{Gates}. In our case, we aim at the
supersymmetrization of matter 2-form fields coupled to Abelian gauge
fields and their supersymmetric partners.
The present work is outlined as follows: in Section $2$, one searches
for the supermultiplet that accomodates the matter tensor field and
discusses its self-interaction; the coupling to the gauge supermultiplet
is pursued in Sections $3$ and $4$; in Section $5$, one couples the
well-known O' Raifeartaigh model \cite{rai} to the tensor-field
supermultiplet and discusses some features concerning spontaneous
symmetry and
supersymmetry breaking. Finally, General Conclusions are drawn in
Section $6$.
\section{Supersymmetrizing the tensor field}
Adopting conventions for the spinor algebra and the superspace parametrization
of Ref.\cite{olie}, one finds that the superfield accomodating the
skew-symmetric rank-$2$ tensor amongst its components is a spinor multiplet
subject to the chirality constraint:
\begin{equation}\begin{array}{rl}
\S_{a} = & \psi_{a} + \t^{b}\L_{ba} + \t^{2}{\cal F}_{a} - i \t^{c}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{c\dot{c}}
\bar{\t}^{\dot{c}} \partial_{\mu} \psi_{a} \\[2mm]
& - i \t^{c}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{c\dot{c}} \bar{\t}^{\dot{c}} \partial_{\mu} \t^{b} \L_{ba}
- \frac{1}{4} \t^{2} \bar{\t}^{2} \partial_{\mu}\partial^{\mu} \psi_{a} ,
\end{array}\eqn{sig}
\begin{equation}\begin{array}{rl}
\overline{\S}_{\dot{a}} = & \overline{\psi}_{\dot{a}} + \bar{\t}_{\dot{b}}
\bar{\L}^{\dot{b}}_{\;\,\dot{a}} + \bar{\t}^{2} \overline{{\cal F}}_{\dot{a}}
+ i \t^{c}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{c\dot{c}} \bar{\t}^{\dot{c}} \partial_{\mu}
\overline{\psi}_{\dot{a}} \\[2mm]
& + i \t^{c}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{c\dot{c}} \bar{\t}^{\dot{c}} \partial_{\mu} \bar{\t}_{\dot{b}}
\bar{\L}^{\dot{b}}_{\;\,\dot{a}} - \frac{1}{4} \t^{2} \bar{\t}^{2}
\partial_{\mu}\partial^{\mu} \overline{\psi}_{\dot{a}} ,
\end{array}\eqn{sigba}
\begin{equation}\begin{array}{l}
\overline{D}_{\dot{b}}\S_{a} = D_{b}\overline{\S}_{\dot{a}} = 0,
\end{array}\eqn{chiralidade}
where $\psi_{a}$ and ${\cal F}_{a}$ are chiral spinors and $\L_{ba}$,
$\bar{\L}_{\dot{b}\dot{a}}$ are decomposed as:
\begin{equation}\begin{array}{l}
\L_{ba} = \varepsilon_{ba}\rho + \sigma} \renewcommand{\S}{\Sigma^{{\mu\nu}}_{ba} \l_{{\mu\nu}} \ , \\[2mm]
\bar{\L}_{\dot{b}\dot{a}} = - \varepsilon_{\dot{b}\dot{a}} \rho^{\ast} -
\overline{\sigma} \renewcommand{\S}{\Sigma}^{{\mu\nu}}_{\dot{b}\dot{a}} \l^{\ast}_{{\mu\nu}} \ .
\end{array}\eqn{deflambda}
According to the chiral properties of the superfield $\S_{a}$, the
$\l_{{\mu\nu}}$-tensor corresponds to the $(1,0)$-representation of Lorentz group.
On the other hand, $\l^{\ast}_{{\mu\nu}}$ yields the $(0,1)$-representation. We
then write:
\begin{equation}\begin{array}{l}
\l_{{\mu\nu}} = T_{{\mu\nu}} - i \wti T_{{\mu\nu}} \ , \\[2mm]
\l^{\ast}_{{\mu\nu}} = T_{{\mu\nu}} + i \wti T_{{\mu\nu}} \ ,
\end{array}\eqn{osT}
where $\wti T_{{\mu\nu}} = \frac 1 2 \varepsilon_{\mu\nu\a\b} T^{\a\b}$. Notice also that $\wti \l_{{\mu\nu}} =
i \l_{{\mu\nu}}$ and $\widetilde{(\l_{{\mu\nu}}^{\ast})} = -i \l_{{\mu\nu}}^{\ast}$, where the
twiddle stands for the dual.
The canonical dimensions of the component fields read as below:
\begin{equation}\begin{array}{l}
d(\psi) = d(\overline{\psi}) = \frac{1}{2} \ \\[2mm]
d(\rho) = d(\l_{{\mu\nu}}) = 1 \ \\[2mm]
d({\cal F}) = d(\overline{{\cal F}}) = \frac{3}{2} \ .
\end{array}\eqn{graus}
Based on dimensional arguments, we propose the following superspace action for
the $\S_{a}$-superfield:
\begin{equation}
{\cal S} = \int d^4 \! x \, d^{2}\t d^{2}\overline{\t} \hspace{.2cm} \frac{-1}{32}\{
D^{a}\S_{a} \overline{D}_{\dot{a}}\overline{\S}^{\dot{a}} +
q\S^{a}\S_{a}\overline{\S}_{\dot{a}}\overline{\S}^{\dot{a}} \}.
\eqn{superacao}
To check whether such an action is actually the supersymmetric extension of
the model that treats $T_{{\mu\nu}}$ as a matter field \cite{ac}, we have now to
write down eq.($2.7$) in terms of the component fields
$\psi, \rho, \l_{{\mu\nu}}$ and ${\cal F}$:
\begin{equation}\begin{array}{rl}
{\cal S} & = \int d^4 \! x \, \displaystyle{\Biggl(} + \partial^{\mu}\rho \partial_{\mu}\rho^{\ast} - 16 \partial^{\mu}\l_{{\mu\nu}}
\partial_{\a}\l^{\ast\a\nu} + i \overline{{\cal F}}^{\dot{a}} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}
a} \partial_{\mu} {\cal F}^{a} - i \overline{\psi}^{\dot{a}} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a} a}
\partial_{\mu} \partial^{\nu}\partial_{\nu} \psi^{a} \\[2mm]
& - q \frac{1}{\sqrt{2}}(\rho^{2} - 4 \l_{{\mu\nu}}\l^{{\mu\nu}} )
\frac{1}{\sqrt{2}}(\rho^{\ast 2} - 4 \l^{\ast}_{\a\b}\l^{\ast \a\b} ) +4q
\l^{{\mu\nu}}\l_{{\mu\nu}}\overline{{\cal F}}_{\dot{a}}\overline{\psi}^{\dot{a}} +4q
\l^{\ast{\mu\nu}}\l_{{\mu\nu}}^{\ast}{\cal F}^{a}\psi_{a} \\[2mm]
& - 2q {\cal F}^{a}\psi_{a}\overline{{\cal F}}_{\dot{a}}\overline{\psi}^{\dot{a}} -
q\rho^{2}\overline{{\cal F}}_{\dot{a}}\overline{\psi}^{\dot{a}} -
q(\rho^{\ast})^{2}{\cal F}^{a}\psi_{a} + \frac{q}{2}\psi^{a}\psi_{a}\partial^{\mu}\partial_{\mu}
(\overline{\psi}_{\dot{a}}\overline{\psi}^{\dot{a}}) \\[2mm] & - i
q\rho\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu}(\overline{\psi}^{\dot{a}}\rho^{\ast}) + 4 q
\rho\psi^{a}\sigma} \renewcommand{\S}{\Sigma_{a\dot{a}}^{\mu}\partial_{\b}(\l^{\ast\b}_{\:\;\:\;\mu}
\overline{\psi}^{\dot{a}}) - 4q \l_{\mu\b}\partial^{\b}(\rho^{\ast}
\overline{\psi}_{\dot{a}})\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu \dot{a}a} \psi_{a} \\[2mm] & -16iq
\l_{\mu\a}\partial_{\b}(\l^{\ast\b\a}\overline{\psi}_{\dot{a}})
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu\dot{a} a} \psi_{a} \displaystyle{\Biggr)}.
\end{array}\eqn{acaocom}
Using $\l_{{\mu\nu}} = \frac{1}{4} ( T_{{\mu\nu}} - i \wti T_{{\mu\nu}} )$, we arrive at the
relation:
\begin{equation}
16 \partial^{\a}\l_{\a\mu}\partial_{\b}\l^{\ast\b\mu} = 2 \partial^{\a}T_{\a\mu}\partial_{\b}T^{\b\mu}
- \frac 1 2 \partial^{\a}T^{{\mu\nu}}\partial_{\a}T_{{\mu\nu}} \; .
\eqn{bilinear}
The action above displays the terms proposed by Avdeev et al. in Ref.\cite{ac};
besides the anti-symmetric tensor, there appear a complex scalar and a pair of
spinors as its supersymmetric partners: $ \psi_{a}$ , a non - physical fermion,
and ${\cal F}_{a}$ , corresponding to a physical Weyl spinor. The undesirable
presence of a spinor with lower canonical dimension ( $\frac{1}{2}$ , instead
of $\frac{3}{2} $ ) generating, as expected, a higher derivative term in the
Lagrangian can be avoided by a reshuffling of the spinorial degrees of
freedom, if one keeps the interactions turned off. In fact, one can join both
$\psi_{a}$ and ${\cal F}_{a}$ in a single fundamental Dirac spinor, $\Psi $, as
follows:
\begin{equation}
\Psi \; = \;
\left(
\begin{array}{c}
{\cal F}_{a}(x) \\
\overline\sigma} \renewcommand{\S}{\Sigma^{\mu\dot{a}a}\partial_{\mu}\; \psi_{a}(x)
\end{array}
\right) \; .
\eqn{OneDirac}
The usual kinetic term, $ i\; \overline\Psi\;\gamma^{\mu}\;\partial_{\mu}\;
(\Psi) $ , provides the kinetical terms for $\psi_{a}$ and ${\cal F}_{a}$, turning
the higher derivative term $, -i\;\overline\psi^{\dot{a}}\overline
\sigma} \renewcommand{\S}{\Sigma^{\mu}_{\dot{a}a}\partial_{\mu}\partial^{2}
\psi^{a} $, into a matter of choice for the field basis. Nevertheless, this
is true only for the free theory. The interaction sector of (\ref{acaocom})
cannot be re-expressed in terms of the Dirac spinor $\Psi $ , imposing a
dissociation back to Weyl spinorial degrees of freedom. Therefore, it happens
that the full theory must carry a higher derivative term, giving birth to a
conjecture that this might be the fermionic counterpart of problems concerning
the transverse sector of the original - bosonic - model for the tensor matter
field with interactions, as discussed in \cite{ac}.
\section{The gauging of the model}
In order to perform the gauging of the model described by eq.(2.7), one
proceeds along the usual lines and introduces a chiral scalar superfield,
$\L$, to act as the gauge parameter:
\begin{equation}
\L = ( 1 - i \t^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\t}^{\dot{a}}\partial_{\mu} -
\frac{1}{4} \t^{2} \overline{\t}^{2} \partial_{\mu}\partial^{\mu} ) ( \phi + \t^{b}w_{b} +
\t^{2}\pi )
\eqn{invsemb}
\begin{equation}
\overline{\L} = ( 1 + i \t^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\t}^{\dot{a}}\partial_{\mu}
- \frac{1}{4} \t^{2} \overline{\t}^{2} \partial_{\mu}\partial^{\mu} ) ( \phi^{\ast} +
\overline{\t}_{\dot{b}}\overline{w}^{\dot{b}} + \overline{\t}^{2}
\overline{\pi} ).
\eqn{invcomb}
The infinitesimal gauge transformations of the superfields $\S$ and
$\overline{\S}$ read as:
\begin{equation}\begin{array}{rl}
\d \S_{a} &= ih \L \hspace{.1cm}\S_{a} \\[2mm]
\d \overline\S_{\dot{a}} &= -ih \overline{\L}\hspace{.1cm}
\overline{\S}_{\dot{a}},
\end{array}\eqn{invsigma}
and the behaviour of $(D^{a}\S_{a})$ and
$(\overline{D_{\dot{a}}}\hspace{.1cm}\overline{\S^{\dot{a}}})$ under finite
transformations become:
\begin{equation}\begin{array}{rl}
D^{a}\hspace{.1cm}\S_{a}^{'} &= e^{ih\L}\hspace{.1cm}(D^{a}\hspace{.1cm}\S_{a}
+ ih D^{a}\L\hspace{.1cm} \S_{a}) \\[2mm]
\overline{D_{\dot{a}}}\hspace{.1cm}\overline{\S^{'\dot{a}}} &=
e^{-ih\overline{\L}}(\overline{D_{\dot{a}}}\hspace{.1cm}\overline{\S^{\dot{a}}}
-ih\overline{D_{\dot{a}}}\hspace{.1cm}\overline{\L}\hspace{.1cm}
\overline{\S^{\dot{a}}}).
\end{array}\eqn{noninv}
To gauge-covariantize the superspace derivatives, one introduces a gauge
connection superfield:
\begin{equation}
D_{a} \rightarrow \nabla_{a} = D_{a} + ih \Gamma_{a},
\eqn{dericova}
in such a way that $\Gamma_{a}$ transforms like
\begin{equation}
\Gamma_{a}^{'} = \Gamma_{a} - D_{a}\L .
\eqn{invgamma}
This yields:
\begin{equation}
(\nabla^{a}\S_{a})^{'} = e^{ih\L} (\nabla^{a}\S_{a}).
\eqn{superinv}
To achieve a U$(1)$-invariant action, one proposes
\begin{equation}
{\cal S} = \int d^4 \! x \, d^{2}\t d^{2}\overline{\t} \displaystyle{\Biggl(}
\nabla^{a}\S_{a}\hspace{.1cm}e^{hV}\hspace{.1cm}\overline{\nabla_{\dot{a}}}
\overline{\S^{\dot{a}}} \displaystyle{\Biggr)},
\eqn{gaugeaction}
where V \hspace{1mm} is the real scalar superfield \cite{jose} that
accomplishes the gauging of supersymmetric QED \cite{jose2}:
\begin{equation}
V^{'} = V + i\hspace{.1cm}( \overline{\L} - \L ).
\eqn{invV}
At this point, the gauge sector displays more degrees of freedom than
what is actually required to perform the gauging. There are component
vector fields in both $\Gamma_{a}$ and $V$. However, we notice that the
superfield $\Gamma_{a}$ is not a true independent gauge potential. Indeed,
\begin{equation}
\Gamma_{a} = - i D_{a} V
\eqn{relacomv}
reproduces correctly the gauge tranformation of $\Gamma_{a}$ and, at the
same time, eliminates the redundant degrees of freedom that would be
otherwise present, if we were to keep $\Gamma_{a}$ and $V$ as gauge
superfields. Therefore, the locally U$(1)$- invariant action takes
over the form:
\begin{equation}\begin{array}{rl}
{\cal S} & = \int d^4 \! x \, d^{2}\t \hspace{.2cm} \frac{-1}{128} \displaystyle{\Biggl(}
\overline{D}^{2}(e^{-V} D^{a} e^{V})\overline{D}^{2}(e^{-V} D_{a}
e^{V}) \displaystyle{\Biggr)} \\[2mm] & + \int d^4 \! x \, d^{2}\t d^{2}\overline{\t}
\hspace{.2cm}\frac{-1}{32} \displaystyle{\Biggl(}
\nabla^{a}\S_{a}\hspace{.1cm}e^{hV}\hspace{.1cm}
\overline{\nabla_{\dot{a}}}\overline{\S^{\dot{a}}} +
q \S^{a}\S_{a}e^{2hV}\overline{\S}_{\dot{a}}\overline{\S}^{\dot{a}} \displaystyle{\Biggr)},
\end{array}\eqn{gaugeaction}
where
\begin{equation}\begin{array}{rl}
\nabla^{a}\S_{a} &= D^{a}\S_{a} + hD^{a}V \hspace{.1cm}\S_{a} \\[2mm]
\hspace{2mm} \overline{\nabla_{\dot{a}}}\overline{\S^{\dot{a}}} &=
\overline{D_{\dot{a}}}\overline{\S^{\dot{a}}} + h\overline{D_{\dot{a}}}
V\hspace{.1cm}\overline{\S^{\dot{a}}}.
\end{array}\eqn{superderi}
The $\t$-expansion for the superfield $V$ brings about the following
component fields:
\begin{equation}\begin{array}{rl}
V &= C + \t^{a}b_{a} + \overline{\t}_{\dot{a}}\overline{b}^{\dot{a}} + \t^{a}
\overline{\t}^{\dot{a}}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}A_{\mu} \\[2mm]
& + \hspace{.1cm}\t^{2}\l + \overline{\t}^{2}\overline{\l} +
\t^{2}\overline{\t}_{\dot{a}}\overline{\gamma}^{\dot{a}} +
\overline{\t}^{2}\t^{a}\gamma_{a} + \t^{2}\overline{\t}^{2}\Delta,
\end{array}\eqn{expancao}
where C, $\l$, $\overline{\l}$ and $\Delta$ are scalars $b_{a}$ and $\gamma_{a}$ are
spinors and $A_{\mu}$ is the U($1$)-gauge field. The gauge transformation of
these fields read as below:
\begin{equation}\begin{array}{rl}
\d C &= i( \phi^{\ast} - \phi) ,\hspace{.2cm} \d\l = -i\pi,\hspace{.2cm}
\d\overline{\l} = i\overline{\pi},\hspace{.2cm} \\[2mm]
\d b_{a} &= -iw_{a},\hspace{.2cm} \d \overline{b}_{\dot{a}} =
i\overline{w}_{\dot{a}} \\[2mm]
\d\Delta &= \frac{i}{4}\partial^{\mu}\partial_{\mu} (\phi - \phi^{\ast}),\hspace{.2cm} \d A^{\mu} =
-\partial^{\mu}(\phi + \phi^{\ast}),\hspace{.2cm} \\[2mm]
\d\gamma_{a} &=
\frac{1}{2}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu}\overline{w}^{\dot{a}},\hspace{.2cm}
\d\overline{\gamma}_{\dot{a}} = - \frac{1}{2}\overline{\sigma} \renewcommand{\S}{\Sigma}_{\dot{a}a}\partial_{\mu}w^{a}.
\end{array}\eqn{compinv}
As already known, for the sake of component-field calculations, one usually
works in the so-called Wess-Zumino gauge, where C, $b_{a}$ and $\l$ are gauged
away. The expansion for the exponential of the gauge superfield simplifies,
in this gauge, according to:
\begin{equation}
e^{hV} = 1 + h\t^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\t}^{\dot{a}} A_{\mu} +
h\t^{2}\overline{\t}_{\dot{a}}\overline{\gamma}^{\dot{a}} +
h\overline{\t}^{2}\t^{a}\gamma_{a} + h\t^{2}\overline{\t}^{2}\Delta + \frac{1}{4}
h^{2}\t^{2}\overline{\t}^{2}A^{\mu}A_{\mu}.
\eqn{expe}
Using this gauge, the transformations of the matter fields are:
\begin{equation}\begin{array}{rl}
\psi_{a} &= ih\phi\psi_{a},\hspace{.2cm} \d\rho = ih\phi\rho,\hspace{.2cm} \d\l_{{\mu\nu}} =
ih\phi\l_{{\mu\nu}},\hspace{.2cm}
\d{\cal F}_{a} = ih( \phi{\cal F}_{a} );
\end{array}\eqn{invmatter}
we get thereby the following transformations for the components $T_{{\mu\nu}}$ and
$\wti T_{{\mu\nu}}$:
\begin{equation}
\d T_{{\mu\nu}} = h\phi \wti T_{{\mu\nu}}, \hspace{.5cm}\d \wti T_{{\mu\nu}} = - h\phi T_{{\mu\nu}}
\eqn{invqui}
These are precisely the Abelian gauge transformations for the tensor field as
firstly proposed in Ref.\cite{ac}.
\section{Component-field action in the Wess-Zumino gauge}
Having adopted the component fields as defined in the previous sections,
lengthy algebraic computations yield the following action in the Wess-Zumino
gauge:
\begin{displaymath}\begin{array}{rl}
{\cal S} &= \int d^4 \! x \, \displaystyle{\Biggl(} -\frac{1}{4}F^{{\mu\nu}}F_{{\mu\nu}} +2 \Delta^{2} + i
\overline{\gamma}^{\dot{a}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a} \partial_{\mu}\gamma^{a}
+ \partial^{\mu}\rho\partial_{\mu}\rho^{\ast} - 16 \partial^{\mu}\l_{{\mu\nu}} \partial_{\a}\l^{\ast\a\nu} \\[2mm]
& + i \overline{{\cal F}}^{\dot{a}} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a} a} \partial_{\mu} {\cal F}^{a}
- i \overline{\psi}^{\dot{a}} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a} a} \partial_{\mu}
\partial^{\nu}\partial_{\nu} \psi^{a} - i\frac{h}{2}\partial^{\mu}\rho A_{\mu}\rho^{\ast} + i\frac{h}{2}
\rho A^{\mu}\partial_{\mu}\rho^{\ast} \\[2mm] & + 2h\rho\Delta\rho^{\ast} + \frac{h^{2}}{4}\rho
A^{\mu}A_{\mu}\rho^{\ast} - h\gamma^{a}{\cal F}_{a}\rho^{\ast} -
h\rho\overline{\gamma}_{\dot{a}}\overline{{\cal F}}^{\dot{a}} -
i\frac{h}{2}\rho\gamma^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu}\overline{\psi}^{\dot{a}} \\[2mm] &
-i\frac{h}{2}
\rho^{\ast}\overline{\gamma}^{\dot{a}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a}\partial_{\mu}\psi^{a} + i
\frac{h}{2} \partial_{\mu}\rho^{\ast}\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\gamma}^{\dot{a}} -
i\frac{h}{2} \partial_{\mu}\rho\gamma^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\psi}^{\dot{a}}
- \frac{h}{4}
\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}A_{\mu}\partial^{\nu}\partial_{\nu}\overline{\psi}^{\dot{a}} \\[2mm] & +
\frac{h}{4}
\overline{\psi}^{\dot{a}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a}A_{\mu}\partial^{\nu}
\partial_{\nu}\psi^{a} + i h \Delta\overline{\psi}^{\dot{a}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a}
\partial_{\mu}\psi^{a} + i h \Delta\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu}\overline{\psi}^{\dot{a}}
-\frac{h}{4}\partial^{\nu}A_{\nu}\overline{\psi}^{\dot{a}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a}
\partial_{\mu}\psi^{a} \\[2mm] & +
\frac{h}{4}\partial_{\nu}A^{\nu}\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu}
\overline{\psi}^{\dot{a}} + \frac{h}{2} {\cal F}^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}A_{\mu}
\overline{{\cal F}}^{\dot{a}} - \frac{h}{8}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}
\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}
F_{\mu\nu}\partial_{\a}\overline{\psi}^{\dot{a}} \\[2mm]
&+
\frac{h}{8}\partial_{\mu}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\mu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\a}\sigma} \renewcommand{\S}{\Sigma^{\nu})_{a\dot{a}}
F_{\nu\a}\overline{\psi}^{\dot{a}}
+ h \rho^{\ast}F_{{\mu\nu}}\l^{{\mu\nu}} + h\rho F_{\mu\nu}\l^{\ast\mu\nu} \\[2mm]
& + 2ih ( 4 \partial_{\mu}\l^{\ast\mu\nu}\l_{\nu\a}A^{\a} - 4
\partial^{\mu}\l_{{\mu\nu}}A_{\a}\l^{\ast\nu\a}) \\[2mm]
&- i
\frac{h^{2}}{2}\gamma^{a}(\sigma} \renewcommand{\S}{\Sigma^{\mu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\nu}\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}
\l_{{\mu\nu}}A_{\a}\overline{\psi}^{\dot{a}} + \frac{h}{2}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}
\overline{\gamma}^{\dot{a}}\partial_{\nu} \l^{\ast}_{\mu\a} -
\frac{h}{2}\partial_{\nu}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}
\l^{\ast}_{\mu\a} \overline{\gamma}^{\dot{a}}\\[2mm]
&+ i\frac{h^{2}}{2}
\overline{\gamma}^{\dot{a}}(\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}\sigma} \renewcommand{\S}{\Sigma^{\a}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\b})_{\dot{a}a}A_{\mu}\psi^{a}\l^{\ast}_{\a\b} - \frac{h}{2}
\gamma^{a}(\sigma} \renewcommand{\S}{\Sigma^{\a}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\b}\sigma} \renewcommand{\S}{\Sigma^{\mu})_{a\dot{a}}\partial_{\mu}\l_{\a\b}
\overline{\psi}^{\dot{a}} + \frac{h}{2}\gamma^{a}(\sigma} \renewcommand{\S}{\Sigma^{a}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\b}\sigma} \renewcommand{\S}{\Sigma^{\mu})_{a\dot{a}}\l_{\a\b}\partial_{\mu}\overline{\psi}^{\dot{a}}
\\[2mm] & + \frac{h}{2}\partial_{\mu}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\mu}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\nu}\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}A_{\nu}\partial_{\a}\overline{\psi}^{\dot{a}}
+ \frac{h}{4}\partial_{\mu}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}A_{\nu}\partial_{\a}\overline{\psi}^{\dot{a}} \\[2mm]
&+ \frac{h}{4}
\partial_{\mu}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\mu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\a}\sigma} \renewcommand{\S}{\Sigma^{\nu})_{a\dot{a}}A_{\nu}\partial_{\a}
\overline{\psi}^{\dot{a}} + h^{2}\Delta\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}A_{\mu}
\overline{\psi}^{\dot{a}} +\frac{i}{16}h^{2}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}
\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}
F_{{\mu\nu}}A_{\a}\overline{\psi}^{\dot{a}} \\[2mm]
& + i
\frac{h^{2}}{16}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}
F_{\a\mu}A_{\nu}\overline{\psi}^{\dot{a}} -i\frac{h^{2}}{2}\gamma^{a}(\sigma} \renewcommand{\S}{\Sigma^{\a}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\b}\sigma} \renewcommand{\S}{\Sigma^{\mu})_{a\dot{a}}
\l_{\a\b}A_{\mu}\overline{\psi}^{\dot{a}} \; +
\end{array}\end{displaymath}
\begin{equation}\begin{array}{rl}
& -
i\frac{h^{2}}{4}\partial_{\mu}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\mu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\nu}
\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}A_{\nu}A_{\a}\overline{\psi}^{\dot{a}} - i
\frac{h^{2}}{8}\partial_{\mu}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}
\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}A_{\nu}A_{\a}\overline{\psi}^{\dot{a}} \\[2mm]
& + i \frac{h^{2}}{4}
\psi^{a}A_{\nu}A_{\mu}(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}\sigma} \renewcommand{\S}{\Sigma^{\a})_{a\dot{a}}
\partial_{\a}\overline{\psi}^{\dot{a}} + i \frac{h^{2}}{8}\psi^{a}A_{\nu}A_{\mu}
(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\a}
\sigma} \renewcommand{\S}{\Sigma^{\mu})_{a\dot{a}}\partial_{\a}\overline{\psi}^{\dot{a}} \\[2mm] & +
\frac{h^{3}}{8}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\nu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\a}\sigma} \renewcommand{\S}{\Sigma^{\mu})_{a\dot{a}}
A_{\nu}A_{\a}A_{\mu}\overline{\psi}^{\dot{a}} - h^{2}\gamma^{a}\psi_{a}
\overline{\gamma}_{\dot{a}}\overline{\psi}^{\dot{a}} + 4 h^{2}A^{\nu}A_{\mu}
\l_{\nu\b}\l^{\ast\b\mu} \\[2mm]
&- q \frac{1}{\sqrt{2}}(\rho^{2} - 4 \l_{{\mu\nu}}\l^{{\mu\nu}} )
\frac{1}{\sqrt{2}}(\rho^{\ast 2} - 4 \l^{\ast}_{\a\b}\l^{\ast \a\b} ) +4q
\l^{{\mu\nu}}\l_{{\mu\nu}}\overline{{\cal F}}_{\dot{a}}\overline{\psi}^{\dot{a}} +4q
\l^{\ast{\mu\nu}}\l_{{\mu\nu}}^{\ast}{\cal F}^{a}\psi_{a} \\[2mm]
& - 2q {\cal F}^{a}\psi_{a}\overline{{\cal F}}_{\dot{a}}\overline{\psi}^{\dot{a}} -
q\rho^{2}\overline{{\cal F}}_{\dot{a}}\overline{\psi}^{\dot{a}} -
q(\rho^{\ast})^{2}{\cal F}^{a}\psi_{a} + \frac{q}{2}\psi^{a}\psi_{a}\partial^{\mu}\partial_{\mu}
(\overline{\psi}_{\dot{a}}\overline{\psi}^{\dot{a}}) \\[2mm] &-
iq\rho\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu}(\overline{\psi}^{\dot{a}}\rho^{\ast}) + 4q
\rho\psi^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial^{\b}(\l^{\ast}_{\b\mu}\overline{\psi}^{\dot{a}}) -
4q
\l_{\mu\b}\partial^{\b}(\rho^{\ast}\overline{\psi}_{\dot{a}})
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu\dot{a}a}\psi_{a} \\[2mm] &- 16i q \l_{\mu\a}
\partial_{\b}(\l^{\ast\b\a}\overline{\psi}_{\dot{a}})
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu\dot{a}a}\psi_{a} - q h \Delta\psi^{a}\psi_{a}
\overline{\psi}_{\dot{a}}\overline{\psi}^{\dot{a}} -
iq\frac{h}{2} A_{\mu}\psi^{a}\psi_{a}\partial^{\mu}(\overline{\psi}_{\dot{a}}
\overline{\psi}^{\dot{a}}) \\[2mm]
& + iq\frac{h}{2}
A_{\mu}\partial^{\mu}(\psi^{a}\psi_{a})\overline{\psi}_{\dot{a}}\overline{\psi}^{\dot{a}}
- q \frac{h^{2}}{2}
A^{\mu}A_{\mu}\psi^{a}\psi_{a}\overline{\psi}_{\dot{a}}\overline{\psi}^{\dot{a}}
- q h \rho\gamma^{a}\psi_{a}\overline{\psi}_{\dot{a}}\overline{\psi}^{\dot{a}} - q h
\rho^{\ast}\overline{\gamma}_{\dot{a}}\overline{\psi}^{\dot{a}}\psi^{a}\psi_{a} \\[2mm]
&- q
h\gamma^{a}\sigma} \renewcommand{\S}{\Sigma^{{\mu\nu}}_{ab}\psi^{b}\overline{\psi}_{\dot{a}}
\overline{\psi}^{\dot{a}}\l_{{\mu\nu}} + q h \overline{\gamma^{\dot{a}}}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{{\mu\nu}}_{\dot{a}\dot{b}}\l^{\ast}_{{\mu\nu}}
\overline{\psi}^{\dot{b}}\psi^{a}\psi_{a} - q h \rho\psi^{a}
\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}A_{\mu}\overline{\psi}^{\dot{a}}\rho^{\ast} \\[2mm]
& - 2i q h
\rho\psi^{a}A_{\mu}(\sigma} \renewcommand{\S}{\Sigma^{\mu}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\a}
\sigma} \renewcommand{\S}{\Sigma^{\b})_{a\dot{a}}\l^{\ast}_{\a\b}\overline{\psi}^{\dot{a}} + 2 i q h
\rho^{\ast}\psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\b}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\a}\sigma} \renewcommand{\S}{\Sigma^{\mu})_{a\dot{a}}
A_{\mu}\l_{\a\b}\overline{\psi}^{\dot{a}} \\[2mm] & + 2 q h \psi^{a}(\sigma} \renewcommand{\S}{\Sigma^{\a}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\b}\sigma} \renewcommand{\S}{\Sigma^{\gamma}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}\sigma} \renewcommand{\S}{\Sigma^{\nu})_{a\dot{a}}\l_{\a\b}A_{\gamma}\l^{\ast}_{\mu\nu}
\overline{\psi}^{\dot{a}} \displaystyle{\Biggr)}. \end{array}\eqn{wsacao}
We should stress here a remarkable difference with respect to the case
of the chiral and anti-chiral \underline{scalar} superfields (Wess-Zumino
model \cite{WZM}), namely, the minimal coupling of $\S$ and $\overline{\S}$
to the gauge sector necessarily affects the $\S$ - superfield
self-interaction terms as one reads off from eq ($3.11$). The gauging of
the $U(1)$ - symmetry enriches the self-interactions of the tensor field
not only through its fermionic supersymmetric partners, but also through
the introduction of the gauge boson and the gaugino at the level of the matter
self-interaction terms (these new couplings are compatible with
power-counting renormalizability). This is so because the
model presented here is based on a single spinor superfield. Had we
introduced a couple of spinor superfields, $\S_{a}$ and ${\cal T}_{a}$,
with opposite $U(1)$
charges:
\begin{equation}\begin{array}{rl}
\S^{'}_{a} &= e^{ih\L}\S_{a}, \\[2mm]
{\cal T}^{'}_{a} &= e^{-ih\L}{\cal T}_{a},
\end{array}\eqn{taus}
a self-interacting term of the form
$(\S^{a}{\cal T}_{a}\overline{\S}_{\dot{a}}\overline{{\cal T}}^{\dot{a}})$ would
automatically be invariant whenever the symmetry is gauged, and there would be
no need for introducing the vector superfield to ensure local invariance. Such
a mixed self-interacting term could, in principle, be thought of as a possible
source for a mass term for the spinor superfields, whenever the physical scalar
component $\rho$ develops a non-trivial vacuum expectation value. Nevertheless,
by analysing the $\rho$ - field interactions in the scalar potential, one
concludes that there is no room for spontaneous symmetry breaking as induced
by such a component field (and, similarly, for its counterpart inside ${\cal T}$ ).
On the other hand, we could think to introduce a gauge-invariant mass term of
the form
\begin{equation}
{\cal S}_{mass} = \int d^4 \! x \, (d^{2}\t \hspace{.2cm}i\frac{m}{16}\S^{a}{\cal T}_{a} -
d^{2}\overline{\t} \hspace{.2cm}i\frac{m}{16}
\overline{\S}_{\dot{a}}\overline{{\cal T}}^{\dot{a}});
\eqn{qusemass}
however, a mixed mass term like the one above introduces two massive
excitations of the type $k^{4} = m^{4}$ in the
spectrum. So, regardless the sign of $m^{2}$, a tachyon shall always be
present; hence such a mass term is disregarded.
\section{On Spontaneous Symmetry Breaking}
Due to the spinorial character of the superfield $\S_{a}$, it cannot be used
to accomplish a spontaneous breaking. Indeed, Lorentz invariance is
lost whenever $\S_{a}$ acquires a non-trivial vacuum expectation value. The
idea in the present section is to couple, in a gauge-invariant manner, the
well-known O' Raifeartaigh model \cite{rai} to the spinor superfield
$\S_{a}$, so as to understand the issue of an eventual mass generation for
$\S_{a}$ via spontaneous internal symmetry or supersymmetry breakingdown.
For the sake of concreteness, the model we adopt to discuss such a
matter is defined by the action below:
\begin{equation}\begin{array}{rl}
{\cal S} &= \int d^4 \! x \, d^{2}\t d^{2}\overline{\t} \displaystyle{\Biggl(} \overline{\phi}\phi +
\overline{\phi_{+}}e^{hV}\phi_{+} + \overline{\phi_{-}}e^{-hV}\phi_{-} \displaystyle{\Biggr)} \\[2mm]
&+ \int d^4 \! x \, d^{2}\t \displaystyle{\Biggl(} \frac 1 2 m\phi^{2} + \mu \phi_{+}\phi_{-} + f \phi + g \phi\phi_{+}\phi_{-}
+ G\S^{a}\S_{a}\phi_{-}\fme \displaystyle{\Biggr)} \\[2mm]
&+ \int d^4 \! x \, d^{2}\overline{\t} \displaystyle{\Biggl(} \frac 1 2 m\overline{\phi}^{2} + \mu \overline{\phi_{+}}
\overline{\phi_{-}} + f \overline{\phi} + g \overline{\phi} \overline{\phi_{+}}
\overline{\phi_{-}} + G\overline{\S}_{\dot{a}}
\overline{\S}^{\dot{a}}\overline{\phi_{-}}
\overline{\phi_{-}} \displaystyle{\Biggr)} \\[2mm]
&+ \int d^4 \! x \, d^{2}\t \hspace{.2cm} \frac{-1}{128} \displaystyle{\Biggl(} \overline{D}^{2}(e^{-V}
D^{a} e^{V})\overline{D}^{2}(e^{-V} D_{a} e^{V}) \displaystyle{\Biggr)} \\[2mm] & + \int d^4 \! x \,
d^{2}\t d^{2}\overline{\t} \hspace{.2cm}\frac{-1}{32} \displaystyle{\Biggl(}
\nabla^{a}\S_{a}\hspace{.1cm}e^{hV}\hspace{.1cm}\overline{\nabla_{\dot{a}}}
\overline{\S^{\dot{a}}} + q \S^{a}\S_{a}e^{2hV}\overline{\S}_{\dot{a}}
\overline{\S}^{\dot{a}} \displaystyle{\Biggr)},
\end{array}\eqn{oraios}
where the chiral scalar superfields $\phi$, $\phi_{+}$ and $\phi_{-}$ are
parametrized as follows:
\begin{equation}\begin{array}{rl}
\phi &= ( 1 - i \t^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\t}^{\dot{a}}\partial_{\mu} -
\frac{1}{4} \t^{2} \overline{\t}^{2} \partial_{\mu}\partial^{\mu} ) ( A + \t^{b}\xi_{b} +
\t^{2}b ) \\[2mm]
\phi_{+} &= ( 1 - i \t^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\t}^{\dot{a}}\partial_{\mu} -
\frac{1}{4} \t^{2} \overline{\t}^{2} \partial_{\mu}\partial^{\mu} ) ( A_{+} + \t^{b}\xi_{+b}
+ \t^{2}b_{+} ) \\[2mm]
\phi_{-} &= ( 1 - i \t^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\t}^{\dot{a}}\partial_{\mu} -
\frac{1}{4}\t^{2} \overline{\t}^{2} \partial_{\mu}\partial^{\mu} ) ( A_{-} + \t^{b}\xi_{-b}
+ \t^{2}b_{-} ) \; ;
\end{array}\eqn{fis}
m and $\mu$ are mass parameters, f has dimension of ${mass}^{2}$, whereas g and
G are dimensionless coupling constants.
$\S_{a}$ and $\phi_{-}$ have opposite U$(1)$ - charges. This action, in terms of
components, reads:
\begin{displaymath}\begin{array}{rl}
{\cal S} &= \int d^4 \! x \, 4\displaystyle{\Biggl(} 4\{ \partial^{\mu}A^{\ast}\partial_{\mu}A +
\partial^{\mu}A^{\ast}_{-}\partial_{\mu}A_{-} + \partial^{\mu}A^{\ast}_{+}\partial_{\mu}A_{+} \} + 4\{
b^{\ast}b + b^{\ast}_{-}b_{-} + b^{\ast}_{+}b_{+} \} \\[2mm]
&+ 2i \{ \overline{\xi}^{\dot{a}} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a} \partial_{\mu} \xi^{a}
+ \overline{\xi}^{\dot{a}}_{-} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a} \partial_{\mu} \xi^{a}_{-}
+ \overline{\xi}^{\dot{a}}_{+} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{a}a} \partial_{\mu} \xi^{a}_{+}
\} \displaystyle{\Biggr)} \; +
\end{array}\end{displaymath}
\begin{equation}\begin{array}{rl}
&+ \int d^4 \! x \, \displaystyle{\Biggl(} 16h A^{\ast}_{+}\Delta A_{+} + 8ih A^{\ast}_{+}\partial^{\mu}A_{\mu}A_{+} +
4h^{2} A^{\ast}_{+}A^{\mu}A_{\mu}A_{+} - 8h A^{\ast}_{+}\gamma^{a}\xi_{+a} \\[2mm]
&- 8h\overline{\xi}_{+\dot{a}}\overline{\gamma}^{\dot{a}}A_{+} + 4h
\xi^{+a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}A_{\mu}\overline{\xi}^{\dot{a}}_{+} +
16ih\partial^{\mu}A^{\ast}_{+}A_{\mu}A_{+} \displaystyle{\Biggr)} \\[2mm]
&+ \int d^4 \! x \, \displaystyle{\Biggl(} -16h A^{\ast}_{-}\Delta A_{-} - 8ih A^{\ast}_{-}\partial^{\mu}A_{\mu}A_{-} +
4h^{2}A^{\ast}_{-}A^{\mu}A_{\mu}A_{-} + 8h A^{\ast}_{-}\gamma^{a}\xi_{-a} \\[2mm]
&+ 8h\overline{\xi}_{-\dot{a}}\overline{\gamma}^{\dot{a}}A_{-} - 4h
\xi^{-a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}A_{\mu}\overline{\xi}^{\dot{a}}_{-} -
16ih\partial^{\mu}A^{\ast}_{-}A_{\mu}A_{-} \displaystyle{\Biggr)} \\[2mm]
&+ \int d^4 \! x \, \displaystyle{\Biggl(} m(-4bA + \xi^{a}\xi_{a}) + m(-4b^{\ast}A^{\ast} +
\overline{\xi}_{\dot{a}}\overline{\xi}^{\dot{a}}) - 4fb - 4fb^{\ast} \displaystyle{\Biggr)} \\[2mm]
& + \int d^4 \! x \, 2\displaystyle{\Biggl(} \mu(-2b_{+}A_{-} - 2b_{-}A_{+} + \xi^{a}_{+}\xi_{a -}) +
\mu(-2b^{\ast}_{+}A^{\ast}_{-} - 2b^{\ast}_{-}A^{\ast}_{+} +
\overline{\xi}_{\dot{a}+}\overline{\xi}^{\dot{a}}_{-}) \displaystyle{\Biggr)} \\[2mm]
& + \int d^4 \! x \, 2\displaystyle{\Biggl(} g(-2bA_{+}A_{-} - 2Ab_{+}A_{-} - 2AA_{+}b_{-} + A\xi^{a}_{+}\xi_{a
-} + A_{-}\xi^{a}\xi_{a +} + A_{+}\xi^{a}\xi_{a -}) \\[2mm]
& + g (-2b^{\ast}A^{\ast}_{+}A^{\ast}_{-} - 2A^{\ast}b^{\ast}_{+}A^{\ast}_{-} -
2A^{\ast}A^{\ast}_{+}b^{\ast}_{-} + A^{\ast}\overline{\xi}_{\dot{a}
+}\overline{\xi}^{\dot{a} -} +
A^{\ast}_{-}\overline{\xi}_{\dot{a}}\overline{\xi}^{\dot{a}}_{+} +
A^{\ast}_{+}\overline{\xi}_{\dot{a}}\overline{\xi}^{\dot{a}}_{-}) \displaystyle{\Biggr)} \\[2mm]
& + \int d^4 \! x \, 4G\displaystyle{\Biggl(} (-2{\cal F}^{a}\psi_{a} + \rho^{2} +4 \l_{{\mu\nu}}\l^{{\mu\nu}})(A_{-})^{2} +
(-2\overline{{\cal F}}_{\dot{a}}\overline{\psi}^{\dot{a}} + (\rho^{\ast})^{2} +4
\l^{\ast}_{{\mu\nu}}\l^{\ast {\mu\nu}})(A^{\ast}_{-})^{2} \\[2mm]
& + 2\psi^{b}\psi_{b}(-4b_{-}A_{-} + \xi^{a}_{-}\xi_{a -}) +
2\overline{\psi}_{\dot{b}}\overline{\psi}^{\dot{b}}(-4b^{\ast}_{-}A^{\ast}_{-} +
\overline{\xi}_{\dot{a} -}\overline{\xi}^{\dot{a} -}) \\[2mm]
& - 2(\rho\psi^{b} - \sigma} \renewcommand{\S}{\Sigma^{{\mu\nu}\,ba}\l_{{\mu\nu}}\psi_{a})(\xi_{b -}A_{-})
+2(\rho^{\ast}\overline{\psi}_{\dot{b}} -
\l^{\ast}_{{\mu\nu}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{{\mu\nu}}_{\dot{b}\dot{a}}
\overline{\psi}^{\dot{a}})(\overline{\xi}^{\dot{b}}_{-} A^{\ast}_{-}) \displaystyle{\Biggr)} \\[2mm]
& + \int d^4 \! x \, \displaystyle{\Biggl(} + 2h\rho\Delta\rho^{\ast} +2 \Delta^{2} + \frac{h^{2}}{4}\rho
A^{\mu}A_{\mu}\rho^{\ast} + \ldots \displaystyle{\Biggr)} ,
\end{array} \eqn{bAxi2}
where the dots stand for spinorial partners and derivative terms that are
completely irrelevant for discussing spontaneous symmetry breaking. Also,
the $\rho $-field does not acquire a non-trivial v.e.v., as already mentioned
in the previous section. The scalars that could, in principle, trigger
spontaneous symmetry breaking are only $ A $, $ A_{+} $ and $ A_{-} $, if
one starts from action (\ref{bAxi2}).
The only possible way to endow the tensor field $ \l_{\mu\nu} $ with a mass,
via internal symmetry or supersymmetry breaking, would be by means of a
coupling of the form
\begin{equation}
\S^{a}\S_{a} \phi^{2} \; ,
\eqn{acopla}
as dictated by supersymmetry, through the chirality constraints on $ \S $
and $ \phi $. No matter the number of scalar superfields present in the
model, whenever the breaking takes place and a mass is generated for
$ \l_{\mu\nu} $ as a byproduct, one notices that both the imaginary part of
$ \rho $ and one longitudinal degree of freedom of $ \l_{\mu\nu} $ provide the
spectrum with a tachyonic excitation, without any chance of avoiding this
fact at the expenses of the $\Delta $-field coupling, enriched by an additional
Fayet-Iliopoulos term \cite{Fayet} (actually, a $\Delta $-type term does not
break supersymmetry whenever it is addded to the O' Raifeartaigh model.
Moreover, a $\Delta $-term never couples to $ \l_{\mu\nu} $, since
$ \l_{\mu\nu} \l^{\ast}_{\mu\nu} $ is identically vanishing). Therefore, our
final conclusion is that the masslessness of $ \rho $ and $ \l_{\mu\nu} $
cannot be avoided in a consistent way, just by invoking internal symmetry or
supersymmetry breaking as realised by a set of scalar superfields.
\section{General Conclusions}
The supersymmetrization of the matter tensor field first investigated in
Ref.{\cite{ac}} has been worked out here in terms of a spinor chiral
superfield, $\S_{a}$, whose kinetic and self-interacting terms have been found
in $N=1$ - superspace. The gauging of the model reveals some peculiarities,
such as the need of gauge fields to appear in the matter self-interactions.
Scalar degrees of freedom that accompany the fermionic partners of
$ \l_{\mu\nu} $ cannot be the source for spontaneous symmetry or
supersymmetry breaking, as it could in principle be thought. The reason is
that Lorentz invariance prevents $\S_{a}$ from developing a non-trivial
vacuum expectation value.
A thorough analysis of the coupling between $ \S_{a} $ and chiral and
anti-chiral scalar superfields indicate that no spontaneous breaking takes
place. In components, the scalar $ \rho $ and the tensor $ \l_{\mu\nu} $
display a quartic coupling to the physical scalar components of the
additional matter superfields, namely, $ \rho^{2} A_{-}^{2} $ and
$ \l_{\mu\nu}^{2} A_{-}^{2} $, and no non-trivial minimum with
$<A_{-}> \;\neq 0 $ shows up (a non-trivial minimum with non-zero vacuum
expectation value restricted to the neutral scalar A is possible, but has no
consequence for mass generation). So, spontaneous breaking does not happen
to be a possibility for inducing a mass for the tensor $ \l_{\mu\nu} $.
As a next step, the analysis of matter tensor fields in the framework of
$ N=2 $ extended supersymmetries and the non-Abelian version of the $ N=1 $
model are to be pursued.
\section{Conventions}
\begin{equation}
\sigma} \renewcommand{\S}{\Sigma^{\mu} = ( 1,\sigma} \renewcommand{\S}{\Sigma ), \hspace{.5cm} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu} = ( 1,-\sigma} \renewcommand{\S}{\Sigma ),
\hspace{.5cm} \overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu}_{\dot{b}a} = \sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{b}}\;\; ,
\eqn{pauli}
where $\sigma} \renewcommand{\S}{\Sigma = (\sigma} \renewcommand{\S}{\Sigma_{1},\sigma} \renewcommand{\S}{\Sigma_{2},\sigma} \renewcommand{\S}{\Sigma_{3})$ are the Pauli matrices
\begin{equation}
\sigma} \renewcommand{\S}{\Sigma_{1} = \left( \begin{array}{cc} 0 & 1 \\ 1 & 0 \end{array} \right) ,
\hspace{.5cm}
\sigma} \renewcommand{\S}{\Sigma_{2} = \left( \begin{array}{cc} 0 & -i \\ i & 0 \end{array} \right),
\hspace{.5cm}
\sigma} \renewcommand{\S}{\Sigma_{3} = \left( \begin{array}{cc} 1 & 0 \\ 0 & -1 \end{array} \right) .
\eqn{dirac-pauli}
In addition, the matrices $\sigma} \renewcommand{\S}{\Sigma^{{\mu\nu}}$ and $\overline{\sigma} \renewcommand{\S}{\Sigma}^{{\mu\nu}}$
( $ (\frac{1}{2},0)$ and $ (0,\frac{1}{2})$ SO(1,3) generators) are given by
\begin{equation}\begin{array}{rl}
\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\nu \dot{a}b} &= \eta^{{\mu\nu}}\d_{a}^{\; b} - i
(\sigma} \renewcommand{\S}{\Sigma^{{\mu\nu}})_{a}^{\; b} \\[2mm]
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\mu\dot{a} a}\sigma} \renewcommand{\S}{\Sigma^{\nu}_{a\dot{b}} &= \eta^{{\mu\nu}}\d^{\dot{a}}_{\;
\dot{b}} - i (\overline{\sigma} \renewcommand{\S}{\Sigma}^{{\mu\nu}})^{\dot{a}}_{\; \dot{b}} \; ,
\end{array}\eqn{matrizes}
and the trace is
\begin{equation}\begin{array}{rl}
\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\sigma} \renewcommand{\S}{\Sigma}^{\nu\dot{a}b}\sigma} \renewcommand{\S}{\Sigma^{\a}_{b\dot{b}}
\overline{\sigma} \renewcommand{\S}{\Sigma}^{\b\dot{b}a} = 2\displaystyle{\Biggl(} \eta^{{\mu\nu}}\eta^{\a\b} -
\eta^{\mu\a}\eta^{\nu\b} + \eta^{\mu\b}\eta^{\nu\a} + i \varepsilon^{\mu\nu\a\b} \displaystyle{\Biggr)},
\end{array}\eqn{traco}
where $\varepsilon^{0123} = -\varepsilon_{0123} = 1$. \\[2mm]
The summation convention is:
\begin{equation}
\t\eta = \t^{a}\eta_{a}, \hspace{.5cm} \overline{\t}\overline{\eta} =
\overline{\t}_{\dot{a}}\overline{\eta}^{\dot{a}}\; ,
\eqn{sobedesce}
where lowering and raising of indices are effected through
\begin{equation}
\t^{a} = \varepsilon^{ab}\t_{b}, \hspace{.5cm} \t_{a} = \varepsilon_{ab}\t^{b},
\eqn{lorai}
with $ \varepsilon_{ab} = - \varepsilon_{ba} $ , (the same for dotted indices).
Differentation with respect to the anticommuting parameters $\t_{a}$ ,
$\overline{\t}_{\dot{a}}$ is defined by
\begin{equation}
\frac{\partial}{\partial\t^{a}}\t^{b} = \d_{a}^{\, b} \hspace{.5cm}
\frac{\partial}{\partial\overline{\t}^{\dot{a}}}\overline{\t}^{\dot{b}} =
\d_{\dot{a}}^{\, \dot{b}}.
\eqn{delta1}
Covariant derivatives with respect to the supersymmetry transformations are:
\begin{equation}\begin{array}{rl}
D_{a} & = \frac{\partial}{\partial\t^{a}} -
i\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\overline{\t}^{\dot{a}}\partial_{\mu} \\[2mm]
\overline{D}_{\dot{a}} & = - \frac{\partial}{\partial\overline{\t}^{\dot{a}}} + i
\t^{a}\sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu},
\end{array}\eqn{superderivadas}
and they obey the anticommutation relations
\begin{equation}
\{ D_{a},\overline{D}_{\dot{a}} \} = 2 i \sigma} \renewcommand{\S}{\Sigma^{\mu}_{a\dot{a}}\partial_{\mu} \; ;
\hspace{.5cm} \{D_{a},D_{b} \} = 0 = \{ \overline{D}_{a},\overline{D}_{\dot{b}}
\}.
\eqn{relations}
The Dirac matrices, $ \gamma^{\mu} $ , playing a role in the purely
physical spinorial kinetic term,
$ i\;\overline\Psi \gamma^{\mu}\partial_{\mu} \Psi $ , have the
following expression in terms of the Pauli matrices:
\begin{equation}
\gamma^{\mu} \; = \;
\left(
\begin{array}{cc}
0 & \sigma} \renewcommand{\S}{\Sigma^{\mu} \\
\overline\sigma} \renewcommand{\S}{\Sigma^{\mu} & 0
\end{array}
\right) \; .
\eqn{Gamma}
The spinor $\overline\Psi $ is defined as usually:
\begin{equation}
\overline\Psi = \Psi^{\dagger}\gamma^{0}\; , \;\mbox{where} \;\;\gamma^{0}
\; = \;
\left(
\begin{array}{cc}
0 & 1 \\
1 & 0
\end{array}
\right) \; .
\eqn{Barrado}
\vspace{2cm}
\noindent{\large{\bf Acknowledgements}}
We wish to thank Dr. M. A. de Andrade for helpful discussions on an earlier
manuscript.
The {\it Conselho Nacional de Desenvolvimento Cientifico e Tecnologico},
$CNPq$-Brasil is gratefully acknowledged for the financial support.
|
\section{Introduction}
The strong CP and U(1) problems are two outstanding problems in
the QCD sector of the standard model. The former consists in the
gross disagreement of theoretical estimates of electric dipole
moment of neutron (NEDM), which are invariably larger, by nearly
nine to ten orders of magnitude, than the experimental upper
limit [1]. The U(1) problem epitomises the difficulties in
formulating a theoretically consistent framework to interpret the
mass of the flavour singlet pseudoscalar meson $\eta^{'}$ [2],
which, unlike the other Goldstone bosons, is very heavy.
The genesis of both the problems is the anomaly for space-time
indepedent `global' chiral rotation of fermi (quark) fields
\begin{equation}
q(x) \rightarrow e^{i\alpha\gamma_{5}} q(x), \hspace{.5cm}
\bar{q}(x) \rightarrow \bar{q}(x) e^{i\alpha\gamma_{5}}
\end{equation}
In perturbation theory there is no trace of this `global' anomaly
[3]. A space-time indepedent chiral phase in fermion mass drops
out from all amplitudes diagram by diagram if the interactions
are chirally invariant. This is not in conflict with the ABJ
anomaly in the four divergence of the axial vector current which
arises from the triangle diagram in perturbation theory. The
`global' anomaly is just the space-time integral of the ABJ
anomaly, which, in path integral approach, corresponds to
space-time dependent `local' chiral transformations
\begin{equation}
q(x) \rightarrow e^{i\alpha(x)\gamma_{5}} q(x), \hspace{.5in}
\bar{q}(x) \rightarrow \bar{q}(x) e^{i\alpha(x)\gamma_{5}}
\end{equation}
In perturbation theory the integrand is nontrivial but the
integral vanishes.
The sine qua non of a nonvanishing integral and hence of `global'
chiral anomaly is the zero modes of Euclidean Dirac operator
which live in compactified space-time and are inaccessible in
perturbative framework. The carrier of the virus of fermion zero
modes in popular path integral approach is identified to be the
Osterwalder-Schrader (OS) [4] recipe for the conjugate Dirac field
in Euclidean metric, viz., $\bar{q}(x)$ is indepedent of $q(x)$.
In relativistic metric the relation $\bar{q}(x) = (\gamma_{o}
q(x))^{+}$ relates the Dirac field to its conjugate. The OS
recipe, therefore, requires that the degrees of freedom are
doubled in Euclidean metric. This is neither natural nor
necessary.
We propose a prescription which is just the opposite of the OS
recipe. To be precise, we require $\bar{q}(x)$ to be antilinear
in $q(x)$ and the relationship to obey reciprocity [5]. The
representation of $\bar{q}(x)$ is unique if it is required that
$\bar{q}(x)$ has the correct chiral properties, i.e., obeys
eq.(1). The novel representation reproduces the correct two-point
correlation function, and hence, by antisymmetry, all the 2n-point
fermion correlation functions of perturbation theory [5].This
assures that the novel representation is consistent with all the
standard axioms of Euclidean field theory [6]. The point of
interest is that the fermion zero modes are evaded and the novel
representation leads to a formulation of QCD which is free from
`global' chiral anomaly. This solves the strong CP and the U(1)
problems [5].
It is remarkable that the novel representation yields, in path
integral approach, a nonvanishing determinent of the Dirac
operator $D\!\!\!\!/\,$ in the nontrivial sector $\nu \not= 0$ of gauge fields
\begin{equation}
\nu \equiv \frac{g^{2}}{16\pi^{2}} \int tr F_{\mu\nu}
\tilde{F}_{\mu\nu} d^{4}x
\end{equation}
whereas, the OS recipe gives a vanishing result. This reflects
an inherent ambiguity in the definition of the determinant of
Dirac operator in non-trivial sector. The ambiguity becomes
transparent in Weyl space where one has the option to write
Dirac determinant either as
\begin{equation}
{\em det} D\!\!\!\!/\, = {\em det} (-DD^{+}) {\em
or}\hspace{.1in} {\em det} (-D^{+}D)
\end{equation}
in terms of Weyl operators $D, D^{+}$ defined through
\begin{eqnarray}
D\!\!\!\!/\, &\equiv &\gamma_{\mu} (i\partial_{\mu} - g A_{\mu}) \nonumber
\\
&= &(\begin{array}{cc} 0 &D\\
D^{+} &0\end{array})
\end{eqnarray}
In nontrivial sector $\nu \not= 0$ one of the options in (4)
vanishes while the other does not. This follows from the index theorem
\begin{eqnarray}
\nu &= &{\em dim} {\em ker} (DD^{+}) - {\em dim} {\em ker} (D^{+}D)
\nonumber \\
&= &n_{+} - n_{-}
\end{eqnarray}
and the theorem that there are no `wrong chirality' solutions of
Dirac operator [7], i.e., for $\nu \geq 0$ $(\nu \leq 0)$, the
number of normalisable negative (positive) chirality solution
$n_{-} (n_{+})$ is zero. The novel representation chooses the
nonvanishing option for the Dirac determinant while the OS
recipe leads to the vanishing option in nontrivial sector.
Resolution of the strong CP and the U(1) problems, and hence, it
appears, physics chooses the option of the novel representation.
\section{Global Chiral Anomaly}
In Euclidean metric the QCD action is given by
\begin{equation}
S_{QCD} = S_{G} + S_{F} + \theta_{ew} \Delta^{'}S +
\gamma_{QCD} \Delta S
\end{equation}
where $S_{F}$ is the fermionic piece
\begin{equation}
S_{F} \equiv \int \bar{q}(x) (D\!\!\!\!/\, - iM) q(x) d^{4}x
\end{equation}
and $S_{G}$ is the gluon action. The (light) quarks have three
flavours. For convenience, we assume the mass matrix M to be
diagonal in flavour space and suppress the flavour indices.
The two terms $\Delta S$ and $\Delta^{'}S$ are potential sources
of CP violation. While the $\theta_{QCD}$ term is attributed to
the topological structure of the QCD vacuum
\begin{equation}
\Delta S \equiv \frac{g^{2}}{16\pi^{2}} \int tr F_{\mu\nu}
\bar{F}_{\mu\nu} d^{4}x
\end{equation}
the chiral phase $\theta_{ew}$ in quark mass
\begin{equation}
\Delta^{'}S \equiv \int \bar{q}(x) M\gamma_{5} q(x) d^{4}x
\end{equation}
arises from Higgs interactions in the electroweak sector. In
compactified space-time $\Delta S$ assumes integral values $\nu$.
The degrees of freedom of $q(x)$ are the Grassmann gennators
which appear as coefficients in the expansion of $q(x)$ in a
complete set of basis functions.For convenience, the orthonormal
set of eigenfunctions of the Dirac operator are chosen as basis
functions,
\begin{equation}
q(x) = \sum_{r} (a_{r} + a_{-r} \gamma_{5}) \varphi_{r}(x) +
\sum_{i} a_{oi} \varphi_{oi}(x)
\end{equation}
where the normalised eigenfunctions obey the equations,
\begin{eqnarray}
D\!\!\!\!/\, \varphi_{r}(x) = \lambda_{r} \varphi_{r} (x) &, &\hspace{.5in}
D\!\!\!\!/\, \gamma_{5}\varphi_{r}(x) = -
\lambda_{r}\gamma_{5}\varphi_{r}(x) \nonumber \\
D\!\!\!\!/\, \varphi_{oi}(x) = 0 &, &\hspace{.5in} \gamma_{5}\varphi_{oi}(x) =
\epsilon_{i}\varphi_{oi}
\end{eqnarray}
The zero eigenmodes $\varphi_{oi}$ have definite chiralities,
positive $(\epsilon_{i} = 1)$ or negative $(\epsilon_{i} = -1)$.
The OS recipe is implemented by choosing an independent set of
Grassmann generators for $\bar{q}(x)$
\begin{equation}
\bar{q}(x) = \sum (b_{r} + b_{-r} \gamma_{5}) \varphi_{r}(x) +
\sum b_{oi}\varphi_{oi}(x)
\end{equation}
The Jacobian of the measure in the fermionic partition function $Z_{F}$
\begin{equation}
Z_{F} \equiv \int d\mu {\em exp} [- S_{F}]
\end{equation}
has the form
\begin{equation}
ln J(\alpha) = - 2i\alpha \int A(x)d^{4}x
\end{equation}
for `global' chiral rotation (1). The integrand A(x) is
identified as the `local' chiral anomaly.
The measure corresponding to OS recipe for $\bar{q}(x)$ is given
by
$$d\mu^{I} = \Pi_{r} da_{r} db_{r} \Pi da_{oi} db_{oi}$$
The `local' chiral anomaly for this measure was obtained by
Fujikawa [8],
\begin{eqnarray}
A^{I}(x) &=
&2\sum\varphi_{r}^{+}(x)\gamma_{5}\varphi_{r}(x)+\sum\epsilon_{i}\varphi^{+}_{oi}(x)\varphi_{oi}(x)\nonumber\\
&= &\frac{g^{2}}{16\pi^{2}} trF_{\mu\nu}(x) \bar{F}_{\mu\nu}(x)
\end{eqnarray}
Nonzero eigenmodes drop out because of orthogonality of
$\varphi_{r}(x)$ and $\gamma_{5}\varphi_{r}(x)$ and only zero
modes of $A^{I}(x)$ survive in the integral (15) for the Jacobian
\begin{eqnarray}
ln J^{I}(\alpha) &= &-2i\alpha(n_{+} - n_{-}) \nonumber \\
&= &-2i\alpha\frac{g^{2}}{16\pi^{2}} \int tr
F_{\mu\nu}\bar{F}_{\mu\nu}d^{4}x
\end{eqnarray}
This means that under `global' chiral rotation QCD action
changes according to the formula
\begin{equation}
S_{QCD} \rightarrow S^{I}_{QCD}(\alpha) = S_{G} + S_{F} +
(\theta_{ew} + 2\alpha) \Delta^{'}S + (\theta_{QCD} -
6\alpha)\Delta S
\end{equation}
In effective Lagrangians for chiral models there is no scope for
a nontrivial Jacobian. The global chiral U(1) anomaly (17) in
underlying QCD can, therefore, be reproduced in effective
Lagrangians through an `anomaly term' which breaks chiral
symmetry explicitly. A popular representation of the anomaly
term is [9]
\begin{equation}
\Delta S^{I}_{eff} = - m^{2}_{\eta^{'}} f^{2}_{\pi} \int
[tr ln(\frac{U}{U^{+}}) - \theta_{QCD}]^{2}d^{4}x
\end{equation}
where $f_{\pi}$ and $m^{'}_{\eta}$ are respectively the pion
decay constant and the mass of the flavour singlet Goldstone
boson. The meson matrix U transforms as $U \rightarrow
Ue^{i\alpha}$ under global chiral rotation (1). The problems
with this `anomaly term' are (a) its chiral variation depends
explicitly on $\theta_{QCD}$, and (b) its second order variation
does not vanish. Neither of these properties hold in the
underlying QCD. This is the crux of the controversy between 't
Hooft and Crewther [2], and the reason why the popular
resolution (19) of the U(1) problem is regarded as unsatisfactory.
The transformation law (18) shows that neither $\theta_{ew}$ nor
$\theta_{QCD}$ can be physical. Only the chirally invariant
combination $\bar{\theta} \equiv (\theta_{QCD} + 3\theta_{ew})$
can appear in CP violating effects. Theoretical estimates in
various chiral models suggests [1] NEDM in the range
$$d_{n} \approx \bar{\theta} \times 10^{-15 \pm 1} e.cm$$
Experimental upper limit $\vert d_{n}\vert < 10^{-25} e.cm$,
therefore, puts a stringent constraint $\bar{\theta} <
10^{-10}$. This is the crux of the strong CP problems. The
strong CP problem is, therefore, a serious problem of fine
tuning. Two parameters $\theta_{QCD}$ and $3\theta_{ew}$ which
are of completely different origins in the standard model must
be so fine tuned as to cancel each other completely.
\section{Novel representation of Euclidean Dirac fermion}
We start from the ansatz which is just the opposite of the OS
recipe, i.e., we assume that in Euclidean metric the conjugate
field $\chi(x) \equiv [\bar{q}(x)]^{+}$ is linear in $q(x)$.
This means that the Grassmann generators defining the degrees of
freedom of $\chi(x)$ are a subset of the generators $\{a_{r},
a_{oi}\}$ appearingin $q(x)$. The resulting representation of
$\chi(x)$ is unique, modulo an overall sign, if one requires
that, (a) chiral charge of $\chi(x)$ is opposite to that of
$q(x)$, i.e., if $q(x) \rightarrow e^{i\alpha\gamma_{5}} q(x)$
then $\chi(x) \rightarrow e^{-i\alpha\gamma_{5}} q(x)$, and
(b) the linear relation obeys reciprocity,
\begin{equation}
\chi(x) = \sum_{r}[a_{r} - a_{-r} \gamma_{5}]\varphi_{r}(x)
\end{equation}
The crucial point to note is that $\chi(x)$ cannot contain the
zero mode generators $a_{oi}$ which transform as $a_{oi}
\rightarrow a_{oi} e^{i\epsilon_{i}\alpha}$ under chiral
rotation (1). This will be in conflict with the chiral charge of
$\chi(x)$. As a result, the fermion action $S_{F}$ is devoid of
the zero mode generators
\begin{eqnarray}
S_{F} &= &\int \chi^{+}(x) (D\!\!\!\!/\, - iM) q(x)d^{4}x \nonumber \\
&= &\sum_{r} [\lambda_{r}(a^{*}_{r}a_{r} +
a^{*}_{-r}a_{-r}) - iM (a^{*}_{r}a_{r} - a^{*}_{-r}a_{-r})]
\end{eqnarray}
The measure appropriate for this action
$$d\mu = \Pi_{r} da^{*}_{r} da_{r} da^{*}_{-r} da_{-r}$$
leads to the partition function
\begin{equation}
Z_{F} = \Pi_{\lambda_{r}>o} (\lambda_{r}^{2} + M^{2})
\end{equation}
whose chiral limit (M=0) does not vanish in the nontrivial
sector $(\nu \not= 0)$ of gauge fields.
The two-point correlation function, obtained in the usual path
integral approach, coincides with the familiar formula
\begin{equation}
< q(x) \bar{q}(y) > = < x\vert\frac{1}{D\!\!\!\!/\, - iM}\vert y > -
\frac{\sum_{i}\varphi_{oi}(x)\varphi^{+}_{oi}(y)}{-iM}
\end{equation}
except that the zero mode contributions are subtracted out. In
the limit of zero coupling $g = 0$, the momentum representation
of the correlation function coincides with the Wick-rotated
relativistic Feynman propagator
\begin{equation}
< q(x) \bar{q}(y) >_{g=0} = \frac{1}{(2\pi)^{4}} \int d^{4}p
\frac{p\!\!\!/ + iM}{p^{2} + M^{2}} e^{-ip(x-y)}
\end{equation}
The remaining 2n-point correlation functions follow, in path
integral framework, from the anticommutation of $q(x)$ and $\bar{q}(y)$
\begin{equation}
< q(x_{1}) ... q(x_{m}) \bar{q}(y_{1}) ... \bar{q}(y_{n}) > =
\delta_{mn} {\em det} [< q(x_{i}) \bar{q}(y_{i})]
\end{equation}
The `local' chiral anomaly is given by an expression analogous
to that in the OS formulation (16) except that zero modes are
excluded from the sum on the right hand side,
\begin{eqnarray}
A^{II(x)} &= &2\sum_{\lambda_{r}>0} \varphi^{+}_{r}(x) \gamma_{5}
\varphi_{r}(x) \nonumber \\
&= &\frac{g^{2}}{16\pi^{2}} \int tr F_{\mu\nu}
\bar{F}_{\mu\nu} d^{4}x - \epsilon_{i}\varphi^{+}_{oi}(x) \varphi_{oi}(x)
\end{eqnarray}
Thus the four divergence of the U(1) axial current has a
nontrivial anomaly $A^{II}(x)$
\begin{equation}
\partial_{\mu} J_{\mu 5} (x) = 2\bar{q}(x) M \gamma_{5} q(x) + A^{II}(x)
\end{equation}
which coincides, as it must, with the perturbative (absence of
zero modes) ABJ anomaly. However, the `global' chiral anomaly vanishes
\begin{eqnarray}
ln J^{II}(\alpha) &= &- 2i\alpha \int A^{II}(x) d^{4}x \nonumber
\\
&= &0
\end{eqnarray}
and our desired goal is achieved. The vanishing is the direct
consequence of the orthogonality of nonzero eigenmodes
$\varphi_{r}(x)$ and $\gamma_{5}\varphi_{r}(x)$, or, if one
prefers, the index theorem.
The novel representation (20) is equivalent to the functional relation
\begin{equation}
\chi(x) = \frac{1}{[D\!\!\!\!/\, ^{2}]^{\frac{1}{2}}}D\!\!\!\!/\, q(x)
\end{equation}
which is nonlocal. This is not a disability in Euclidean field
theory. Locality is a field theoretic axiom only in rleativistic
metric. This translates in Euclidean field theory into the axiom
of (anti) symmetry of correlation functions under permutations
[6]. The fact that all the correlation functions are reproduced
correctly through the eqs.(24) and (25) assures us that the
novel representation (20, 29) is consistent not only with the
axiom of symmetry but with the other axioms of Euclidean field
theory, e.g., reflection positivity, cluster decomposition etc.,
as well.
\section{Resolution of strong CP and U(1) problems}
In the scenario corresponding to the novel representation (20,
29) the QCD action is invariant under `global' chiral rotation
in the chiral (M = 0) limit
\begin{equation}
S_{QCD} \rightarrow S^{II}_{QCD}(\alpha) = S_{G} + S_{F} +
(\theta_{ew} + 2\alpha) Delta^{'}S + \theta_{QCD} \Delta S
\end{equation}
In effective Lagrangians, the `anomaly' term, which is invariant
under `global' chiral rotation but reproduces the ABJ anomaly in
axial Ward identity, is easily constructed
\begin{equation}
\Delta S^{II}_{eff} = - m^{2}_{\eta^{'}} f^{2}_{\pi} \int [tr ln
(\frac{U}{U^{+}}) - < tr ln (\frac{U}{U^{+}})>]^{2} d^{4}x
\end{equation}
The controversial features of the popular construction (19) thus
disappear and the U(1) problem satisfactorily resolved [5].
The law of transformation (30) shows that the parameter
$\theta_{QCD}$, the coefficient of $\Delta S$, is invariant,
while $\theta_{ew}$ is unphysical and can be eliminated
trivially by `global' chiral rotation (1). There is no longer
any problem of fine tuning and CP symmetry of QCD is simply
ensured through the `natural' choice $\theta_{QCD} = 0$ [5].
We conclude that the strong CP and the U(1) problems both are
legacies of the OS recipe. The alternative scenario of QCD with
the novel representation (20, 29) for the conjugate Dirac field
is not afflicted with these blemishes [5].
|
\section{\bf 1. Introduction}
The theory of spinors in spaces with boundaries is of interest physically in
connection with quantum cosmology and supergravity. (See D'Eath and Esposito
[\putref{DandE}] and Esposito [\putref{Esposito}] for some history of these
questions.) In mathematics it is encountered in the spin-index theorem and the
Atiyah, Patodi and Singer $\eta$ spectral asymmetry function, the standard
reference being Gilkey's book, [\putref{gilk}].
As explained in [\putref{DandE}], for self-adjointness of the Dirac operator,
there is a choice between spectral and local (mixed) boundary conditions,
the former being of relevance for the spin-index and the latter having more
physical significance in connection with supersymmetry, string theory
and quantum gravity, [\putref{MandP2,Luck}], although in the guise of relative
conditions they do have a cohomological importance, [\putref{gilk,BandG2}].
In the special case of the Euclidean 4-ball, it was shown [\putref{DandE2,Kam,
KandM}] that the value of $\zeta(0)$, which determines the scaling of the theory,
was the same for both sets of conditions. In this note we report on the
same question for the one-loop effective action, which is, up to factors,
$\zeta'(0)$. Our method will be that explained in [\putref{Dow9}].
\section{\bf 2. Mode properties and calculation}
The analysis of the modes of the massless Dirac equation on the 4-ball
was carried out by D'Eath and Esposito [\putref{DandE,DandE2}] and we will do
no more here on this matter than use their results. For local
boundary conditions they found that the eigenvalues, $\alpha^2$, are the roots of
the equation
$$ F^L_p(\alpha)=J_{p-1}^2(\alpha)-J_p^2(\alpha)=0
\eql{locb}$$with a degeneracy, for a given $p$, of $p^2-p$, $p=1,2,\ldots$.
For spectral conditions, there is the simpler, scalar-like condition,
$$
F_p^S(\alpha)=J_p(\alpha)=0
\eql{specb}$$
with degeneracy $2(p^2+p)$, $p=1,2,\ldots$.
Our approach is based on the Mittag-Leffler decomposition,
$$
z^{-\beta}F_p(z)=\gamma\prod_\alpha\bigg(1-{z^2\over\alpha^2}\bigg),
\eql{ml1}$$
where
$$\eqalign{
&\beta=p,\quad\gamma={1\over2^pp!},\quad{\rm spectral},\cr
&\beta=2(p-1),\quad\gamma={1\over\big(2^{p-1}(p-1)!\big)^2},\quad{\rm local}.
\cr}
$$
This standard decomposition was earlier employed by Moss [\putref{Moss}] and by
D'Eath and Esposito [\putref{DandE}] when looking at the heat-kernel expansion
and $\zeta(0)$. Here, when finding $\zeta'(0)$, we need the normalising factor,
$\gamma$, which follows from the small-$z$ behaviour of $F_p(z)$.
A few details of the calculation will be given but, for brevity,
some of our previous work must be utilised.
Bypassing a number of steps, which are fully explained in [\putref{Dow9,Dow8}],
we define the quantities
$$
G_N\sim\sum_{p=1}^\infty p^N\bigg[\big(p-{1\over2}\big)
\ln{2p\over p+\epsilon}+(\epsilon-p)+\sum_{n=1}^{N+1}\bigg({E_n(t)\over \epsilon^n}
-{E_n(1)\over p^n}\bigg)+I_N(p)\bigg]
\eql{logdetl}$$
and
$$
H_N\sim\sum_{p=1}^\infty p^N\bigg[p\ln{2p
\over p+\epsilon}+\epsilon-p-{1\over2}\ln{\epsilon\over p}+\sum_{n=1}^{N+1}
\bigg({T_n(t)\over \epsilon^n}-{T_n(1)\over p^n}\bigg)+I_N(p)\bigg],
\eql{logdets}$$
with
$$
I_N(p)=\int_0^\infty\bigg({1\over2}
-{1\over\tau}+\sum_{k=1}^{[N/2]+1}(-1)^kB_{2k}{\tau^{2k-1}\over(2k)!}
+{1\over e^\tau-1}\bigg){e^{-\tau p}\over\tau}\,d\tau,
$$
in terms of which we can write the spin-half quantities,
$$\eqalign{
&{\zeta_{1/2}^L}'(0)=2\big(G_2-G_1\big),\cr
&{\zeta^S_{1/2}}'(0)=2\big(H_2+H_1\big).\cr}
\eql{zedashes}$$
The labels $S$ and $L$ refer to spectral and local boundary conditions
respectively.
In equations (\puteqn{logdetl}) and (\puteqn{logdets}) the $\sim$ symbol signifies
that the mass-independent part of the large-mass asymptotic limit is to be
taken. The $E_n(t)$ are the polynomials in $t=p/\epsilon$, $\epsilon=(m^2+p^2)^{1/2}$,
that occur in the asymptotic expansion of $F_p^L(im)$ of (\puteqn{locb}) derived
by D'Eath and Esposito (they call them $A_n/2$) from Olver's series. The
$T_n(t)$ are the corresponding polynomials for the scalar case,
[\putref{Moss,Dow8}]. The condition that makes equation (\puteqn{logdetl})
possible is $E_n(1)=T_n(1)$ which can be proved from the explicit definition
of the $E_n$. We note that $T_n(1)$ is zero for $n$ even and that
$T_{2k-1}(1)=(-1)^kB_{2k}/2k(2k-1)$ in terms of Bernoulli numbers.
We have made use of the algebraic results of D'Eath and Esposito,
[\putref{DandE}] section IV, in deriving (\puteqn{logdetl}).
Expression (\puteqn{logdets}) is identical to one occurring for scalar fields
on the even ball, except that $N$, there being the power of
$p$ in the expansion of the degeneracy, is even. Hence for $N=2$,
our previous result in [\putref{Dow8,DandA2}] for the 4-ball (see also
[\putref{BGKE}]) could be used without change.
{}From the technique outlined in [\putref{Dow9}] the following useful limits
can be deduced,\marginnote{work out ln m part}
$$\eqalign{
&\sum_{p=1}^\infty p^N(\epsilon-p)\sim -\zeta_R(-N-1)+O(\ln m),\cr
&\sum_{p=1}^\infty p^N\ln\big({2p\over p+\epsilon}\big)\sim-\zeta_R'(-N)+\ln2\,
\zeta_R(-N)+O(\ln m),\cr
&\sum_{p=1}^\infty p^N\ln\big({\epsilon\over p}\big)\sim\zeta_R'(-N)+O(\ln m).\cr}
\eql{aslims}$$
It is necessary to state that a hidden regularisation has been employed to
render the summations finite. This consists of removing sufficient of the
Taylor expansion of the summand and will not be indicated. Since the entire
expression is finite, the divergent terms so introduced must all cancel.
These limits enable some of the terms in (\puteqn{logdetl}) and (\puteqn{logdets})
to be dealt with quickly. The rest, {\it i.e. } the polynomial and integral
contributions, need a little more work. We write them as in
[\putref{Dow8,Dow9}],
$$\eqalign{
\sum_{p=1}^\infty&\, p^N\bigg[\sum_{n=1,3,\ldots}^{N+1}
P_n(1) \bigg({1\over\epsilon^n}-{1\over p^n}\bigg)
+\sum_{n=1}^{N+1}{P'_n(t)\over\epsilon^n}\bigg]\cr
&+\lim_{s\to0}\!\int_0^\infty\!\!\bigg({1\over2}\!
-\!{1\over\tau}\!+\!\sum_{k=1}^{[N/2]+1}\!\!(-1)^kB_{2k}{\tau^{2k-1}
\over(2k)!}\!
+\!{1\over e^\tau-1}\bigg)\tau^{s-1}(-1)^N{d^N\over d\tau^N}
{1\over e^\tau\!-\!1}d\tau,\cr}
\eql{rem2}$$ where $P_n$ stands for either $E_n$ or $T_n$ and $P'_n(t)=P_n(t)
-P_n(1)$.
A recursion is developed for the multiple derivative in (\puteqn{rem2}) and the
contribution from the integral found to be, after some algebra,
$$
\zeta_R'(-N-1)+{1\over2}\zeta_R'(-N)+\zeta_R(-N-1)+
\sum_{k=1}^{N+1}M_k^{(N)}\zeta_R'(-k),
\eql{intcont}$$
where the coefficient matrix $M$ is defined by
$$
M_k^{(N)}=\sum_{l=k}^{N+1}A_l^{(N)}{S_{l+1}^{(k+1)}\over l!}
$$
in terms of easily evaluated recursion constants $A_l^{(j)}$ and Stirling
numbers $S_l^{(k)}$, [\putref{Dow8,Dow9}].
Assembling the various pieces, and using special values for the $M_k^{(N)}$,
we find
$$\eqalign{
G_N&={\zeta_R'(-N)\over2}\!+\!{\zeta_R'(-N-1)\over N+1}\!
+\!\sum_{k=1}^{N-1}\!M_k^{(N)}\zeta_R'(-k)\!\cr
&+\!\big({1\over2}\zeta_R(-N)-\zeta_R(\!-\!N\!-1)\big)\,\ln2
+\int_0^1t^NE''_{N+1}(t)\,dt+L_N,\quad N\ge1,\cr}
\eql{totl}$$
where $P_n''(t)=P'_n(t)/(1-t^2)$ and
$$
L_N=T_{N+1}(1)\bigg(\ln2
+\sum_{k=1}^N{1\over k}+\sum_{q=1}^{N/2}
{(-1)^q\sqrt\pi\,(N/2)!\over2q(N/2-q)!\Gamma(q+1/2)}\bigg).
$$
The last two terms in (\puteqn{totl}) come from the first line of (\puteqn{rem2}).
Explicitly for $N=0$, a case needed later,
$$
G_0=-{1\over24}+{1\over12}\ln2+\zeta_R'(-1).
\eql{gee0}$$
For spectral conditions,
$$\eqalign{
H_N&=-{\zeta_R'(-N)\over2}+{\zeta_R'(-N-1)\over N+1}
+\sum_{k=1}^{N-1}M_k^{(N)}\zeta_R'(-k)\cr
&+\zeta_R(-N-1)\,\ln2
+\int_0^1t^NT''_{N+1}(t)\,dt+L_N,\quad N\ge1,\cr}
\eql{tots}$$
and
$$
H_0= {5\over24}-{1\over6}\ln2+\zeta_R'(-1)-\zeta_R'(0).
\eql{aitch0}$$
Making the constructions (\puteqn{zedashes}), one finds for local spin-half,
\marginnote{SPINHALF.MTH}
$$
{\zeta^L_{1/2}}'(0)={251\over15120}-{11\over180}\ln2
+{2\over3}\big(\zeta_R'(-3)-\zeta_R'(-1)\big)\approx0.088108
\eql{dashl}$$
and for spectral,
$$
{\zeta^S_{1/2}}'(0)=-{2489\over30240}+{1\over45}\ln2
+{2\over3}\big(\zeta_R'(-3)-\zeta_R'(-1)\big)\approx0.046962
\eql{dashs}$$
which are the main results of this note.
The specific forms of the $E_n$ polynomials given in [\putref{DandE}], have
been used to
evaluate the integrals in (\puteqn{totl}). We remark that in the corresponding
evaluation of $\zeta(0)$ (=$11/360$), one needs only the particular value
$P_N(1)$, which equals $\zeta_R(-N)/N$, a non-transcendental, local quantity.
\section{\bf 3. Higher spins}
The eigenvalue conditions for some higher-spin theories are summarised in
[\putref{DandE2}] section VI. A mechanical application of the present technique
yields the following results.
For real spin-0 with Dirichlet conditions,
$$\eqalign{
2{\zeta^D_0}'(0)&=2H_2\cr
\noalign{\vskip5truept}
&={173\over15120}+{1\over45}\ln2+{2\over3}\zeta_R'(-3)-\zeta_R'(-2)
+{1\over3}\zeta_R'(-1)\cr
\noalign{\vskip5truept}
&\approx0.005738.\cr}
$$
For spin-1 (Maxwell) with Dirichlet (magnetic) conditions,
$$\eqalign{
\zeta'_{\rm TV}(0)&=2(H_2-2H_0)\cr
\noalign{\vskip5truept}
&=-{6127\over15120}+{16\over45}\ln2+{2\over3}
\zeta_R'(-3)-\zeta_R'(-2)-{5\over3}\zeta_R'(-1)+2\zeta_R'(0)\cr
\noalign{\vskip5truept}
&\approx-1.68691.\cr}
\eql{TV}$$
For spin-3/2 physical degrees of freedom with spectral conditions,
$$\eqalign{
{\zeta^S_{3/2}}'(0)&=2(H_2+H_1-H_0)\cr
\noalign{\vskip5truept}
&=-{27689\over30240}+{31\over45}\ln2+{2\over3}\zeta_R'(-3)
-{14\over3}\zeta_R'(-1)+4\zeta_R'(0)\cr
\noalign{\vskip5truept}
&\approx-3.33834.\cr}
\eql{3/2s}$$
These results imply, rather trivially, the sum rule,
$$
{\zeta^S_{3/2}}'(0)-{\zeta^S_{1/2}}'(0)=2\big(\zeta_{\rm TV}'(0)
-2{\zeta^D_0}'(0)\big).
\eql{ident1}$$
The same relation holds also for $\zeta(0)$,
$$
{\zeta^S_{3/2}}(0)-{\zeta^S_{1/2}}(0)=2\big(\zeta_{\rm TV}(0)
-2\zeta^D_0(0)\big),
\eql{ident2}$$
and, in fact, for all coefficients in the heat-kernel expansion, as can be
checked numerically from the tables provided in [\putref{BEK}] and
[\putref{KandC}].
The specific values,
$$
\zeta^S_{3/2}(0)=-{289\over360},\quad\zeta^S_{1/2}(0)={11\over360},
\quad\zeta_{\rm TV}(0)=-{77\over180},\quad \zeta^D_0(0)=-{1\over180},
\eql{zezeros}
$$
were computed in references [\putref{DandE,DandE2,Louko}], see also
[\putref{MandP,KandC,Poletti}]. The spectral label, $S$, can be replaced
by the local one, $L$, in (\puteqn{zezeros}).
The sum rules are only special cases of the general relation
$$
{\zeta^S_{3/2}}(s)-{\zeta^S_{1/2}}(s)=2\big(\zeta_{\rm TV}(s)
-2\zeta^D_0(s)\big),
\eql{ident3}$$
which is a consequence of the eigenvalue condition, (\puteqn{specb}), and
the various quadratic degeneracies.
For spin-2 transverse-traceless modes with Dirichlet conditions,
[\putref{Schleich}], \ {\it i.e. }
$$
F_p^{\rm TT}=J_p(\alpha)=0
$$ and degeneracy $2(p^2-4)$, $p\ge3$, we find
$$\eqalign{
\zeta'_{\rm TT}(0)&=2(\overline H_2-4\overline H_0)=2(H_2-H_0)+
6\big(\zeta_R'(0)+\ln2\big)\cr
&= -{25027\over15120}+{331\over45}\ln2+{2\over3}\zeta_R'(-3)-\zeta_R'(-2)
-{23\over3}\zeta_R'(-1)+14\zeta_R'(0)\cr
\noalign{\vskip5truept}
&\approx-8.119619,\cr}
\eql{TT}$$
where the bar signifies that the $p=1$ term has been left out in
(\puteqn{logdets}). (The easiest way of doing this is to remove the overall $p=1$
term at the outset.)
For the record, the local spin-3/2 expression is
$$\eqalign{
{\zeta^L_{3/2}}'(0)&=2(\overline G_2-\overline G_1-2\overline G_0)\cr
&={2771\over15120}+{289\over180}\ln2+{2\over3}\zeta_R'(-3)
-{14\over3}\zeta_R'(-1)+4\zeta_R'(0)\cr
\noalign{\vskip5truept}
&\approx-1.60405,\cr}
\eql{3/2l}$$
which exhibits the anomaly value of $-289/360$.
Arbitrary-spin fields can be treated in exactly the same way, most easily
using the mode analysis given in [\putref{Dow10,DandC2}], and will be discussed
in a later communication.
\begin{ignore}
For massless spin-j fields, the degeneracy is $2(n^2-j^2)$, [\putref{Dow10}],
where $n=p$ for integer spins with Dirichlet conditions and $n=p+j$ for
half odd-integer ones with spectral conditions.
\end{ignore}
\section{\bf 4. Comments}
The above expressions for the $\zeta'(0)$ have also been obtained by Kirsten
and Cognola [\putref{KandC}] using the method of Bordag {\it et al},
[\putref{BGKE}].
The local result, (\puteqn{dashl}), agrees with that of Apps, reported
in [\putref{DandA2}] and found using a
conformal transformation from the 4-hemisphere. In fact, the final expression
in (\puteqn{dashl}) is $\zeta'_S(0)$ on the hemisphere, the rest coming from the
cocycle function obtained from an integration of the conformal anomaly, as in
[\putref{DandA,BandG}] for example.
Spectral conditions are also conformally invariant and it seems that
(\puteqn{dashs}) can be interpreted in a similar way. The same structure is also
apparent in (\puteqn{3/2s}) and (\puteqn{3/2l}) for spin 3/2.
This suggests that the eigenvalue problem on the hemisphere is the same,
or is equivalent, for spectral and local boundary conditions. This is
confirmed
by, and may explain, the equality of $\zeta(0)$ for these conditions found
by D'Eath and Esposito in flat space and by Kamenshchik and
Mishakov on the bounded sphere. To the author's knowledge, the cocycle
function has not been calculated for spectral conditions.
The extension to higher, even-dimensional spaces is straightforward and simply
consists of substituting (\puteqn{totl}) or (\puteqn{tots}) into the appropriate
polynomial form of the spinor degeneracy. For odd dimensions the major
difference is that the $p$-sums run over half odd-integers and presents no
problem [\putref{Dow9}]. For example, the Maxwell modes on the 3-ball are
classic, {\it e.g. } [\putref{deB}], and it is soon shown that the magnetic determinant
is obtained by doubling the scalar Dirichlet value and subtracting $-2\ln2$
to allow for the different starting point of the mode sum. Similarly, the
electric determinant is the double the scalar Robin one, with $\beta=1/2$,
again minus $-2\ln2$.
\section{\bf Acknowlegment}
I wish to thank Klaus Kirsten for helpfully communicating his results.
\vskip 10truept
\noindent{\bf{References}}
\vskip 5truept
\begin{putreferences}
\reference{Rayleigh}{Lord Rayleigh{\it Theory of Sound} vols.I and II,
MacMillan, London, 1877,78.}
\reference{KCD}{G.Kennedy, R.Critchley and J.S.Dowker \aop{125}{80}{346}.}
\reference{Donnelly} {H.Donnelly \ma{224}{1976}161.}
\reference{Fur2}{D.V.Fursaev {\sl Spectral geometry and one-loop divergences on
manifolds with conical singularities}, JINR preprint DSF-13/94,
hep-th/9405143.}
\reference{HandE}{S.W.Hawking and G.F.R.Ellis {\sl The large scale structure of
space-time} Cambridge University Press, 1973.}
\reference{DandK}{J.S.Dowker and G.Kennedy \jpa{11}{78}{895}.}
\reference{ChandD}{Peter Chang and J.S.Dowker \np{395}{93}{407}.}
\reference{FandM}{D.V.Fursaev and G.Miele \pr{D49}{94}{987}.}
\reference{Dowkerccs}{J.S.Dowker \cqg{4}{87}{L157}.}
\reference{BandH}{J.Br\"uning and E.Heintze \dmj{51}{84}{959}.}
\reference{Cheeger}{J.Cheeger \jdg{18}{83}{575}.}
\reference{SandW}{K.Stewartson and R.T.Waechter \pcps{69}{71}{353}.}
\reference{CandJ}{H.S.Carslaw and J.C.Jaeger {\it The conduction of heat
in solids} Oxford, The Clarendon Press, 1959.}
\reference{BandH}{H.P.Baltes and E.M.Hilf {\it Spectra of finite systems}.}
\reference{Epstein}{P.Epstein \ma{56}{1903}{615}.}
\reference{Kennedy1}{G.Kennedy \pr{D23}{81}{2884}.}
\reference{Kennedy2}{G.Kennedy PhD thesis, Manchester (1978).}
\reference{Kennedy3}{G.Kennedy \jpa{11}{78}{L173}.}
\reference{Luscher}{M.L\"uscher, K.Symanzik and P.Weiss \np {173}{80}{365}.}
\reference{Polyakov}{A.M.Polyakov \pl {103}{81}{207}.}
\reference{Bukhb}{L.Bukhbinder, V.P.Gusynin and P.I.Fomin {\it Sov. J. Nucl.
Phys.} {\bf 44} (1986) 534.}
\reference{Alvarez}{O.Alvarez \np {216}{83}{125}.}
\reference{DandS}{J.S.Dowker and J.P.Schofield \jmp{31}{90}{808}.}
\reference{Dow1}{J.S.Dowker \cmp{162}{94}{633}.}
\reference{Dow2}{J.S.Dowker \cqg{11}{94}{557}.}
\reference{Dow3}{J.S.Dowker \jmp{35}{94}{4989}; erratum {\it ibid}, Feb.1995.}
\reference{Dow5}{J.S.Dowker {\it Heat-kernels and polytopes} To be published}
\reference{Dow6}{J.S.Dowker \pr{D50}{94}{6369}.}
\reference{Dow7}{J.S.Dowker \pr{D39}{89}{1235}.}
\reference{Dow8}{J.S.Dowker {\it Robin conditions on the Euclidean ball}
MUTP/95/7; hep-th\break/9506042.}
\reference{Dow9}{J.S.Dowker {\it Oddball determinants} MUTP/95/12; hep-th/9507096.}
\reference{Dow10}{J.S.Dowker \pr{D28}{83}{3013}.}
\reference{BandG}{P.B.Gilkey and T.P.Branson \tams{344}{94}{479}.}
\reference{Schofield}{J.P.Schofield Ph.D.thesis, University of Manchester,
(1991).}
\reference{Barnesa}{E.W.Barnes {\it Trans. Camb. Phil. Soc.} {\bf 19} (1903)
374.}
\reference{BandG2}{T.P.Branson and P.B.Gilkey {\it Comm. Partial Diff. Equations}
{\bf 15} (1990) 245.}
\reference{Pathria}{R.K.Pathria {\it Suppl.Nuovo Cim.} {\bf 4} (1966) 276.}
\reference{Baltes}{H.P.Baltes \prA{6}{72}{2252}.}
\reference{Spivak}{M.Spivak {\it Differential Geometry} vols III, IV, Publish
or Perish, Boston, 1975.}
\reference{Eisenhart}{L.P.Eisenhart {\it Differential Geometry}, Princeton
University Press, Princeton, 1926.}
\reference{Moss}{I.G.Moss \cqg{6}{89}{659}.}
\reference{Barv}{A.O.Barvinsky, Yu.A.Kamenshchik and I.P.Karmazin \aop {219}
{92}{201}.}
\reference{Kam}{Yu.A.Kamenshchik and I.V.Mishakov \prD{47}{93}{1380}.}
\reference{KandM}{Yu.A.Kamenshchik and I.V.Mishakov {\it Int. J. Mod. Phys.}
{\bf A7} (1992) 3265.}
\reference{DandE}{P.D.D'Eath and G.V.M.Esposito \prD{43}{91}{3234}.}
\reference{DandE2}{P.D.D'Eath and G.V.M.Esposito \prD{44}{91}{1713}.}
\reference{Rich}{K.Richardson \jfa{122}{94}{52}.}
\reference{Osgood}{B.Osgood, R.Phillips and P.Sarnak \jfa{80}{88}{148}.}
\reference{BCY}{T.P.Branson, S.-Y. A.Chang and P.C.Yang \cmp{149}{92}{241}.}
\reference{Vass}{D.V.Vassilevich.{\it Vector fields on a disk with mixed
boundary conditions} gr-qc /9404052.}
\reference{MandP}{I.Moss and S.Poletti \pl{B333}{94}{326}.}
\reference{Kam2}{G.Esposito, A.Y.Kamenshchik, I.V.Mishakov and G.Pollifrone
\prD{50}{94}{6329}.}
\reference{Aurell1}{E.Aurell and P.Salomonson \cmp{165}{94}{233}.}
\reference{Aurell2}{E.Aurell and P.Salomonson {\it Further results on functional
determinants of laplacians on simplicial complexes} hep-th/9405140.}
\reference{BandO}{T.P.Branson and B.\O rsted \pams{113}{91}{669}.}
\reference{Elizalde1}{E.Elizalde, \jmp{35}{94}{3308}.}
\reference{BandK}{M.Bordag and K.Kirsten {\it Heat-kernel coefficients of
the Laplace operator on the 3-dimensional ball} hep-th/9501064.}
\reference{Waechter}{R.T.Waechter \pcps{72}{72}{439}.}
\reference{GRV}{S.Guraswamy, S.G.Rajeev and P.Vitale {\it O(N) sigma-model as
a three dimensional conformal field theory}, Rochester preprint UR-1357.}
\reference{CandC}{A.Capelli and A.Costa \np {314}{89}{707}.}
\reference{IandZ}{C.Itzykson and J.-B.Zuber \np{275}{86}{580}.}
\reference{BandH}{M.V.Berry and C.J.Howls \prs {447}{94}{527}.}
\reference{DandW}{A.Dettki and A.Wipf \np{377}{92}{252}.}
\reference{Weisbergerb} {W.I.Weisberger \cmp{112}{87}{633}.}
\reference{Voros}{A.Voros \cmp{110}{87}{110}.}
\reference{Pockels}{F.Pockels {\it \"Uber die partielle Differentialgleichung
$\Delta u+k^2u=0$}, B.G.Teubner, Leipzig 1891.}
\reference{Kober}{H.Kober \mz{39}{1935}{609}.}
\reference{Watson2}{G.N.Watson \qjm{2}{31}{300}.}
\reference{DandC1}{J.S.Dowker and R.Critchley \prD {13}{76}{3224}.}
\reference{DandC2}{J.S.Dowker and R.Critchley \prD {13}{76}{224}.}
\reference{Lamb}{H.Lamb \pm{15}{1884}{270}.}
\reference{EandR}{E.Elizalde and A.Romeo International J. of Math. and Phys.
{\bf13} (1994) 453}
\reference{DandA}{J.S.Dowker and J.S.Apps \cqg{12}{95}{1363}.}
\reference{DandA2}{J.S.Dowker and J.S.Apps, {\it Functional determinants on certain
domains}. To appear in the Proceedings of the 6th Moscow Quantum Gravity
Seminar held in Moscow, June 1995; hep-th/9506204.}
\reference{Watson1}{G.N.Watson {\it Theory of Bessel Functions} Cambridge
University Press, Cambridge, 1944.}
\reference{BGKE}{M.Bordag, B.Geyer, K.Kirsten and E.Elizalde, {\it Zeta function
determinant of the Laplace operator on the D-dimensional ball} UB-ECM-PF
95/10; hep-th /9505157.}
\reference{MandO}{W.Magnus and F.Oberhettinger {\it Formeln und S\"atze}
Springer-Verlag, Berlin, 1948.}
\reference{Olver}{F.W.J.Olver {\it Phil.Trans.Roy.Soc} {\bf A247} (1954) 328.}
\reference{Hurt}{N.E.Hurt {\it Geometric Quantization in action} Reidel,
Dordrecht, 1983.}
\reference{Esposito}{G.Esposito {\it Quantum Gravity, Quantum Cosmology and
Lorentzian Geometry}, Lecture Notes in Physics, Monographs, Vol. m12,
Springer-Verlag, Berlin 1994.}
\reference{Louko}{J.Louko \prD{38}{88}{478}.}
\reference{Schleich} {K.Schleich \prD{32}{85}{1989}.}
\reference{BEK}{M.Bordag, E.Elizalde and K.Kirsten {\it Heat kernel
coefficients of the Laplace operator on the D-dimensional ball}
UB-ECM-PF 95/3; hep-th/9503023.}
\reference{ELZ}{E.Elizalde, S.Leseduarte and S.Zerbini.}
\reference{BGV}{T.P.Branson, P.B.Gilkey and D.V.Vassilevich {\it The Asymptotics
of the Laplacian on a manifold with boundary} II, hep-th/9504029.}
\reference{Erdelyi}{A.Erdelyi,W.Magnus,F.Oberhettinger and F.G.Tricomi {\it
Higher Transcendental Functions} Vol.I McGraw-Hill, New York, 1953.}
\reference{Quine}{J.R.Quine, S.H.Heydari and R.Y.Song \tams{338}{93}{213}.}
\reference{Dikii}{L.A.Dikii {\it Usp. Mat. Nauk.} {\bf13} (1958) 111.}
\reference{DandH}{P.D.D'Eath and J.J.Halliwell \prd{35}{87}{1100}.}
\reference{KandC}{K.Kirsten and G.Cognola, {\it Heat-kernel coefficients and
functional determinants for higher spin fields on the ball} UTF354. Aug. 1995.}
\reference{Louko}{J.Louko \prD{38}{88}{478}.}
\reference{MandP}{I.G.Moss and S.J.Poletti \pl{B333}{94}{326}.}
\reference{MandP2}{I.G.Moss and S.J.Poletti \np{341}{90}{155}.}
\reference{Luck}{H.C.Luckock \jmp{32}{91}{1755}.}
\reference{Poletti}{S.J.Poletti \pl{B249}{90}{355}.}
\reference{gilk}{P.B.Gilkey {\it Invariant theory, the heat equation and the
Atiyah-Singer index theorem}, Publish or Perish, Wilmington, DE, 1984.}
\reference{deB}{Louis de Broglie {\it Probl\`emes de propagation guide\'es des
ondes electromagnetiques} 2me. \'Ed. Gauthier-Villars, Paris, 1951.}
\end{putreferences}
\bye
|
\section*{Figure captions}
\begin{figure}[h]
\caption{The anisotropy coefficient $B_\eta$ for the $\eta$--channel
at initial energies from 1.26 to 4.9~GeV.}
\label{fig1}
\caption{ The cross sections $d\sigma/dM$ for $pp$ and $pn$ interactions
at 1.26 and 2.1~GeV bombarding energies.
The ``$\eta$'' denotes the contribution of the $\eta$--channel,
the ``$\Delta$'' labels the contribution of the $\Delta$--resonance term,
while ``$Br$'' denotes the bremsstrahlung channel. The dashed curves are
the sum of bremsstrahlung and $\Delta$--channel with interference.}
\label{fig2}
\caption{The weighted anisotropy coefficients $<B_i(M)>$
for $pp$ and $pn$ collisions at 1.26 and 2.1~GeV. The notation is
the same as in Fig.~\protect\ref{fig2}. The solid line denoted by
``$all$'' represents the sum of bremsstrahlung, $\Delta$--channel with
interference and $\eta$--decay contributions. }
\label{fig3}
\caption{The weighted anisotropy coefficients
$<B_i(M)>$ for $pd$ interactions at 1.26 and 2.1~GeV.
The notation is the same as in Fig.~\protect\ref{fig3}. }
\label{fig4}
\caption{The ratio of anisotropy coefficients for $pd$ to $pp$ reactions
at energies of 1.26 and 2.1~GeV.
The solid lines correspond to calculations taking
into account the $\eta$--decay contribution while the dashed lines
do not include the $\eta$--Dalitz decay.}
\label{fig5}
\end{figure}
\end{document}
|
\section*{Figure Captions\markboth
{FIGURECAPTIONS}{FIGURECAPTIONS}}\list
{Figure \arabic{enumi}:\hfill}{\settowidth\labelwidth{Figure 999:}
\leftmargin\labelwidth
\advance\leftmargin\labelsep\usecounter{enumi}}}
\let\endfigcap\endlist \relax
\def\tablecap{\section*{Table Captions\markboth
{TABLECAPTIONS}{TABLECAPTIONS}}\list
{Table \arabic{enumi}:\hfill}{\settowidth\labelwidth{Table 999:}
\leftmargin\labelwidth
\advance\leftmargin\labelsep\usecounter{enumi}}}
\let\endtablecap\endlist \relax
\def\reflist{\section*{References\markboth
{REFLIST}{REFLIST}}\list
{[\arabic{enumi}]\hfill}{\settowidth\labelwidth{[999]}
\leftmargin\labelwidth
\advance\leftmargin\labelsep\usecounter{enumi}}}
\let\endreflist\endlist \relax
\makeatletter
\newcounter{pubctr}
\def\@ifnextchar[{\@publist}{\@@publist}{\@ifnextchar[{\@publist}{\@@publist}}
\def\@publist[#1]{\list
{[\arabic{pubctr}]\hfill}{\settowidth\labelwidth{[999]}
\leftmargin\labelwidth
\advance\leftmargin\labelsep
\@nmbrlisttrue\def\@listctr{pubctr}
\setcounter{pubctr}{#1}\addtocounter{pubctr}{-1}}}
\def\@@publist{\list
{[\arabic{pubctr}]\hfill}{\settowidth\labelwidth{[999]}
\leftmargin\labelwidth
\advance\leftmargin\labelsep
\@nmbrlisttrue\def\@listctr{pubctr}}}
\let\endpublist\endlist \relax
\makeatother
\def\hskip -.1cm \cdot \hskip -.1cm{\hskip -.1cm \cdot \hskip -.1cm}
\def\not\!{\not\!}
\catcode`\@=11
\def\section{\@startsection {section}{1}{0pt}{-3.5ex plus -1ex minus
-.2ex}{2.3ex plus .2ex}{\raggedright\large\bf}}
\catcode`\@=12
\def\mbf#1{\hbox{\boldmath $#1$}}
\newskip\humongous \humongous=0pt plus 1000pt minus 1000pt
\def\mathsurround=0pt{\mathsurround=0pt}
\def\eqalign#1{\,\vcenter{\openup1\jot \mathsurround=0pt
\ialign{\strut \hfil$\displaystyle{##}$&$
\displaystyle{{}##}$\hfil\crcr#1\crcr}}\,}
\newif\ifdtup
\def\panorama{\global\dtuptrue \openup1\jot \mathsurround=0pt
\everycr{\noalign{\ifdtup \global\dtupfalse
\vskip-\lineskiplimit \vskip\normallineskiplimit
\else \penalty\interdisplaylinepenalty \fi}}}
\def\eqalignno#1{\panorama \tabskip=\humongous
\halign to\displaywidth{\hfil$\displaystyle{##}$
\tabskip=0pt&$\displaystyle{{}##}$\hfil
\tabskip=\humongous&\llap{$##$}\tabskip=0pt
\crcr#1\crcr}}
\def\hangindent3\parindent{\hangindent3\parindent}
\def\oldreffmt#1{\rlap{[#1]} \hbox to 2\parindent{}}
\def\oldref#1{\par\noindent\hangindent3\parindent \oldreffmt{#1}
\ignorespaces}
\def\hangindent=1.25in{\hangindent=1.25in}
\def\figfmt#1{\rlap{Figure {#1}} \hbox to 1in{}}
\def\fig#1{\par\noindent\hangindent=1.25in \figfmt{#1}
\ignorespaces}
\def\hbox{\it i.e.}} \def\etc{\hbox{\it etc.}{\hbox{\it i.e.}} \def\etc{\hbox{\it etc.}}
\def\hbox{\it e.g.}} \def\cf{\hbox{\it cf.}{\hbox{\it e.g.}} \def\cf{\hbox{\it cf.}}
\def\hbox{\it et al.}{\hbox{\it et al.}}
\def\hbox{---}{\hbox{---}}
\def\mathop{\rm cok}{\mathop{\rm cok}}
\def\mathop{\rm tr}{\mathop{\rm tr}}
\def\mathop{\rm Tr}{\mathop{\rm Tr}}
\def\mathop{\rm Im}{\mathop{\rm Im}}
\def\mathop{\rm Re}{\mathop{\rm Re}}
\def\mathop{\bf R}{\mathop{\bf R}}
\def\mathop{\bf C}{\mathop{\bf C}}
\def\lie{\hbox{\it \$}}
\def\partder#1#2{{\partial #1\over\partial #2}}
\def\secder#1#2#3{{\partial^2 #1\over\partial #2 \partial #3}}
\def\bra#1{\left\langle #1\right|}
\def\ket#1{\left| #1\right\rangle}
\def\VEV#1{\left\langle #1\right\rangle}
\let\vev\VEV
\def\gdot#1{\rlap{$#1$}/}
\def\abs#1{\left| #1\right|}
\def\pri#1{#1^\prime}
\def\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$<$}{\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$<$}}
\def\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$>$}{\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$>$}}
\def\contract{\makebox[1.2em][c]{
\mbox{\rule{.6em}{.01truein}\rule{.01truein}{.6em}}}}
\def{1\over 2}{{1\over 2}}
\def\begin{equation}{\begin{equation}}
\def\end{equation}{\end{equation}}
\def\underline{\underline}
\def\begin{eqnarray}{\begin{eqnarray}}
\def\lrover#1{
\raisebox{1.3ex}{\rlap{$\leftrightarrow$}} \raisebox{ 0ex}{$#1$}}
\def\com#1#2{
\left[#1, #2\right]}
\def\end{eqnarray}{\end{eqnarray}}
\def\:\raisebox{1.3ex}{\rlap{$\vert$}}\!\rightarrow{\:\raisebox{1.3ex}{\rlap{$\vert$}}\!\rightarrow}
\def\longbent{\:\raisebox{3.5ex}{\rlap{$\vert$}}\raisebox{1.3ex}%
{\rlap{$\vert$}}\!\rightarrow}
\def\onedk#1#2{
\begin{equation}
\begin{array}{l}
#1 \\
\:\raisebox{1.3ex}{\rlap{$\vert$}}\!\rightarrow #2
\end{array}
\end{equation}
}
\def\dk#1#2#3{
\begin{equation}
\begin{array}{r c l}
#1 & \rightarrow & #2 \\
& & \:\raisebox{1.3ex}{\rlap{$\vert$}}\!\rightarrow #3
\end{array}
\end{equation}
}
\def\dkp#1#2#3#4{
\begin{equation}
\begin{array}{r c l}
#1 & \rightarrow & #2#3 \\
& & \phantom{\; #2}\:\raisebox{1.3ex}{\rlap{$\vert$}}\!\rightarrow #4
\end{array}
\end{equation}
}
\def\bothdk#1#2#3#4#5{
\begin{equation}
\begin{array}{r c l}
#1 & \rightarrow & #2#3 \\
& & \:\raisebox{1.3ex}{\rlap{$\vert$}}\raisebox{-0.5ex}{$\vert$}%
\phantom{#2}\!\:\raisebox{1.3ex}{\rlap{$\vert$}}\!\rightarrow #4 \\
& & \:\raisebox{1.3ex}{\rlap{$\vert$}}\!\rightarrow #5
\end{array}
\end{equation}
}
\hyphenation{anom-a-ly}
\hyphenation{comp-act-ifica-tion}
\def\ap#1#2#3{ {\it Ann. Phys. (NY) }{\bf #1}, #2 (19#3)}
\def\apj#1#2#3{ {\it Astrophys. J. }{\bf #1}, #2 (19#3)}
\def\apjl#1#2#3{ {\it Astrophys. J. Lett. }{\bf #1}, #2 (19#3)}
\def\app#1#2#3{ {\it Acta Phys. Polon. }{\bf #1}, #2 (19#3)}
\def\ar#1#2#3{ {\it Ann. Rev. Nucl. and Part. Sci. }{\bf #1}, #2 (19#3)}
\def\com#1#2#3{ {\it Comm. Math. Phys. }{\bf #1}, #2 (19#3)}
\def\ib#1#2#3{ {\it ibid. }{\bf #1}, #2 (19#3)}
\def\nat#1#2#3{ {\it Nature (London) }{\bf #1}, #2 (19#3)}
\def\nc#1#2#3{ {\it Nuovo Cim. }{\bf #1}, #2 (19#3)}
\def\np#1#2#3{ {\it Nucl. Phys. }{\bf #1}, #2 (19#3)}
\def\pl#1#2#3{ {\it Phys. Lett. }{\bf #1}, #2 (19#3)}
\def\pr#1#2#3{ {\it Phys. Rev. }{\bf #1}, #2 (19#3)}
\def\prep#1#2#3{ {\it Phys. Rep. }{\bf #1}, #2 (19#3)}
\def\prl#1#2#3{ {\it Phys. Rev. Lett. }{\bf #1}, #2 (19#3)}
\def\pro#1#2#3{ {\it Prog. Theor. Phys. }{\bf #1}, #2 (19#3)}
\def\rmp#1#2#3{ {\it Rev. Mod. Phys. }{\bf #1}, #2 (19#3)}
\def\sp#1#2#3{ {\it Sov. Phys.-Usp. }{\bf #1}, #2 (19#3)}
\def\sjn#1#2#3{ {\it Sov. J. Nucl. Phys. }{#1}, #2 (19#3)}
\def\srv#1#2#3{ {\it Surv. High Energy Phys. }{#1}, #2 (19#3)}
\defthese proceedings{these proceedings}
\def\zp#1#2#3{ {\it Zeit. fur Physik }{\bf #1}, #2 (19#3)}
\catcode`\@=11
\def\eqnarray{\stepcounter{equation}\let\@currentlabel=\thesection.\arabic{equation}}
\global\@eqnswtrue
\global\@eqcnt\z@\tabskip\@centering\let\\=\@eqncr
\gdef\@@fix{}\def\eqno##1{\gdef\@@fix{##1}}%
$$\halign to \displaywidth\bgroup\@eqnsel\hskip\@centering
$\displaystyle\tabskip\z@{##}$&\global\@eqcnt\@ne
\hskip 2\arraycolsep \hfil${##}$\hfil
&\global\@eqcnt\tw@ \hskip 2\arraycolsep $\displaystyle\tabskip\z@{##}$\hfil
\tabskip\@centering&\llap{##}\tabskip\z@\cr}
\def\@@eqncr{\let\@tempa\relax
\ifcase\@eqcnt \def\@tempa{& & &}\or \def\@tempa{& &}
\else \def\@tempa{&}\fi
\@tempa \if@eqnsw\@eqnnum\stepcounter{equation}\else\@@fix\gdef\@@fix{}\fi
\global\@eqnswtrue\global\@eqcnt\z@\cr}
\catcode`\@=12
\font\tenbifull=cmmib10
\font\tenbimed=cmmib10 scaled 800
\font\tenbismall=cmmib10 scaled 666
\textfont9=\tenbifull \scriptfont9=\tenbimed
\scriptscriptfont9=\tenbismall
\def\fam9 {\fam9 }
\def\fam=9{\mathchar"710B } {\fam=9{\mathchar"710B } }
\def\fam=9{\mathchar"710C } {\fam=9{\mathchar"710C } }
\def\fam=9{\mathchar"710D } {\fam=9{\mathchar"710D } }
\def\fam=9{\mathchar"710E } {\fam=9{\mathchar"710E } }
\def\fam=9{\mathchar"710F } {\fam=9{\mathchar"710F } }
\def\fam=9{\mathchar"7110 } {\fam=9{\mathchar"7110 } }
\def\fam=9{\mathchar"7111 } {\fam=9{\mathchar"7111 } }
\def\fam=9{\mathchar"7112 } {\fam=9{\mathchar"7112 } }
\def\fam=9{\mathchar"7113 } {\fam=9{\mathchar"7113 } }
\def\fam=9{\mathchar"7114 } {\fam=9{\mathchar"7114 } }
\def\fam=9{\mathchar"7115 } {\fam=9{\mathchar"7115 } }
\def\fam=9{\mathchar"7116 } {\fam=9{\mathchar"7116 } }
\def\fam=9{\mathchar"7117 } {\fam=9{\mathchar"7117 } }
\def\fam=9{\mathchar"716F } {\fam=9{\mathchar"716F } }
\def\fam=9{\mathchar"7118 } {\fam=9{\mathchar"7118 } }
\def\fam=9{\mathchar"7119 } {\fam=9{\mathchar"7119 } }
\def\fam=9{\mathchar"711A } {\fam=9{\mathchar"711A } }
\def\fam=9{\mathchar"711B } {\fam=9{\mathchar"711B } }
\def\fam=9{\mathchar"711C } {\fam=9{\mathchar"711C } }
\def\fam=9{\mathchar"711D } {\fam=9{\mathchar"711D } }
\def\fam=9{\mathchar"711E } {\fam=9{\mathchar"711E } }
\def\fam=9{\mathchar"711F } {\fam=9{\mathchar"711F } }
\def\fam=9{\mathchar"7120 } {\fam=9{\mathchar"7120 } }
\def\fam=9{\mathchar"7121 } {\fam=9{\mathchar"7121 } }
\def\fam=9{\mathchar"7122 } {\fam=9{\mathchar"7122 } }
\def\fam=9{\mathchar"7123 } {\fam=9{\mathchar"7123 } }
\def\fam=9{\mathchar"7124 } {\fam=9{\mathchar"7124 } }
\def\fam=9{\mathchar"7125 } {\fam=9{\mathchar"7125 } }
\def\fam=9{\mathchar"7126 } {\fam=9{\mathchar"7126 } }
\def\fam=9{\mathchar"7127 } {\fam=9{\mathchar"7127 } }
\def\fam=6{\mathchar"7000 } {\fam=6{\mathchar"7000 } }
\def\fam=6{\mathchar"7001 } {\fam=6{\mathchar"7001 } }
\def\fam=6{\mathchar"7002 } {\fam=6{\mathchar"7002 } }
\def\fam=6{\mathchar"7003 } {\fam=6{\mathchar"7003 } }
\def\fam=6{\mathchar"7004 } {\fam=6{\mathchar"7004 } }
\def\fam=6{\mathchar"7005 } {\fam=6{\mathchar"7005 } }
\def\fam=6{\mathchar"7006 } {\fam=6{\mathchar"7006 } }
\def\fam=6{\mathchar"7007 } {\fam=6{\mathchar"7007 } }
\def\fam=6{\mathchar"7008 } {\fam=6{\mathchar"7008 } }
\def\fam=6{\mathchar"7009 } {\fam=6{\mathchar"7009 } }
\def\fam=6{\mathchar"700A } {\fam=6{\mathchar"700A } }
\def\fam=9{\mathchar"700A } {\fam=9{\mathchar"700A } }
\def\fam=9{\mathchar"7000 } {\fam=9{\mathchar"7000 } }
\def\fam=9{\mathchar"7001 } {\fam=9{\mathchar"7001 } }
\def\fam=9{\mathchar"7002 } {\fam=9{\mathchar"7002 } }
\def\fam=9{\mathchar"7003 } {\fam=9{\mathchar"7003 } }
\def\fam=9{\mathchar"7004 } {\fam=9{\mathchar"7004 } }
\def\fam=9{\mathchar"7005 } {\fam=9{\mathchar"7005 } }
\def\fam=9{\mathchar"7006 } {\fam=9{\mathchar"7006 } }
\def\fam=9{\mathchar"7007 } {\fam=9{\mathchar"7007 } }
\def\fam=9{\mathchar"7008 } {\fam=9{\mathchar"7008 } }
\def\fam=9{\mathchar"7009 } {\fam=9{\mathchar"7009 } }
\def\fam=9{\mathchar"700A } {\fam=9{\mathchar"700A } }
\relax
\def\double{
\renewcommand{1.2}{2}
\large
\normalsize
}
\def\single {
\renewcommand{1.2}{1}
\large
\normalsize
}
\def1.2{1.2}
\def\
thefootnote{\fnsymbol{footnote}}
\def\widetilde{\widetilde}
\def\widehat{\widehat}
\renewcommand{\fam=9{\mathchar"7115 } }{\mbox{\boldmath $\lambda$}}
\renewcommand{\fam=9{\mathchar"7110 } }{\mbox{\boldmath $\zeta$}}
\begin{document}
\begin{titlepage}
\begin{center}
\today \hfill LBL-37343\\
\hfill UCB-PTH-95/17\\
\hfill hep-ph/9508288 \\
\vskip .1in
\vskip .25in
{\large \bf Flavor Mixing Signals For Realistic\\ Supersymmetric Unification.}
\footnote{This work was supported in part by the Director, Office of
Energy Research, Office of High Energy and Nuclear Physics, Division of
High Energy Physics of the U.S. Department of Energy under Contract
DE-AC03-76SF00098 and in part by the National Science Foundation under
grant PHY-90-21139.}
\vskip .25in
{\bf Nima Arkani-Hamed\\
Hsin-Chia Cheng}\\
and\\
{\bf L.J. Hall}\\
\vskip .20in
{\em Theoretical Physics Group\\
Lawrence Berkeley Laboratory\\
and\\
Department of Physics\\
University of California\\
Berkeley, California 94720}
\end{center}
\vskip .25in
\begin{abstract}
The gauge interactions of any supersymmetric extension of the standard
model involve new flavor mixing matrices.
The assumptions involved in the construction of minimal supersymmetric models,
both $SU(3) \times SU(2) \times U(1)$ and
grand unified theories, force a large degree of triviality on these matrices.
However, the requirement of realistic quark and lepton masses in supersymmetric
grand unified theories forces these matrices to be non-trivial.
This leads to important new dominant contributions to the
neutron electric dipole moment
and to the decay mode $p \to K^o\mu^+$, and suggests that there may be
important
weak scale radiative corrections to the Yukawa coupling matrix of the up
quarks.
The lepton flavor violating signal $\mu \to e\gamma$ is studied in these
theories when $\tan\beta$ is sufficiently large that radiative
effects of couplings
other than
$\lambda_t$ must be included.
The naive expectation that large $\tan\beta$ will force sleptons to
unacceptably large masses is not
borne out: radiative suppressions to the leptonic flavor mixing angles
allow regions where the
sleptons are as light as 300 GeV, provided the top Yukawa
coupling in the unified
theory is near the minimal value consistent with $m_t$.
\end{abstract}
\end{titlepage}
\renewcommand{\thepage}{\roman{page}}
\setcounter{page}{2}
\mbox{ }
\vskip 1in
\begin{center}
{\bf Disclaimer}
\end{center}
\vskip .2in
\begin{scriptsize}
\begin{quotation}
This document was prepared as an account of work sponsored by the United
States Government. While this document is believed to contain correct
information, neither the United States Government nor any agency
thereof, nor The Regents of the University of California, nor any of their
employees, makes any warranty, express or implied, or assumes any legal
liability or responsibility for the accuracy, completeness, or usefulness
of any information, apparatus, product, or process disclosed, or represents
that its use would not infringe privately owned rights. Reference herein
to any specific commercial products process, or service by its trade name,
trademark, manufacturer, or otherwise, does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United States
Government or any agency thereof, or The Regents of the University of
California. The views and opinions of authors expressed herein do not
necessarily state or reflect those of the United States Government or any
agency thereof of The Regents of the University of California and shall
not be used for advertising or product endorsement purposes.
\end{quotation}
\end{scriptsize}
\vskip 2in
\begin{center}
\begin{small}
{\it Lawrence Berkeley Laboratory is an equal opportunity employer.}
\end{small}
\end{center}
\newpage
\renewcommand{\thepage}{\arabic{page}}
\setcounter{page}{1}
\noindent{\bf I. Introduction}
\medskip
It has recently been demonstrated that flavor and CP
violation provide an important new probe of supersymmetric
grand unified theories [1-4].
These new signals, such as $\mu\to e\gamma$ and the electron electric
dipole moment $d_e$, are complementary to the classic tests of
proton decay, neutrino masses
and quark and charged lepton mass relations.
The classic tests are very dependent on the flavor interactions
and symmetry breaking sector of the unified model: it is only too
easy to construct models in which these signals are absent or unobservable.
However, they are insensitive to the hardness scale, $\Lambda_H$,
of supersymmetry breaking.\footnote{This is the highest scale at
which supersymmetry breaking squark and gluino masses appear
in the theory as local interactions.} On the other hand, the new flavor and
CP violating signals are relatively insensitive
to the form of the flavor interactions and unified gauge symmetry
breaking, but are absent if the hardness scale, $\Lambda_H$,
falls beneath the unified
scale, $M_G$. The signals are generated by the unified flavor
interactions leaving
an imprint on the form of the soft supersymmetry breaking operators
\cite{HKR}, which is only possible if supersymmetry
breaking is present in the unified theory at scales above $M_G$.
The flavor and CP violating signals have been computed in the minimal
$SU(5)$ and $SO(10)$
models for leptonic [1-3] and hadronic processes \cite{BHS2},
for moderate values of tan$\beta$, the ratio of the two Higgs
vacuum expectation values.
While rare muon decays provide an important probe of $SU(5)$,
it is the $SO(10)$ theory which is
most powerfully tested.
If the hardness scale for supersymmetry breaking is large enough, as in the
popular
supergravity models, it may be possible for the minimal $SO(10)$ theory to
be probed throughout the interesting
range of superpartner masses by searches for $\mu \to e\gamma$ and $d_e$.
The flavor changing and CP violating probes of $SO(10)$ are sufficiently
powerful
to warrant an exploration of consequences for non-minimal models,
which is the subject of this
paper. In particular, we study $SO(10)$ theories in which
(I) {\it{The Yukawa interactions are non-minimal.}}
In the minimal model the quarks and leptons lie in three
16's and the two Higgs doublets $H_U$ and
$H_D$ lie in two 10 dimensional
representations $10_U$ and $10_D$.
The quark and charged lepton masses are assumed to arise from the interactions
$16 {\fam=9{\mathchar"7115 } }_U 16\ 10_U + 16 {\fam=9{\mathchar"7115 } }_D 16\ 10_D$.
This model is a useful fiction: it is very simple to work with, but leads to
the
mass relation $m_e/m_\mu = m_d/m_s$,
which is in error by an order of magnitude.
It is clearly necessary to introduce a mechanism to insert $SO(10)$
breaking into the Yukawa interactions.
The simplest way to achieve this is to assume that at the unification scale,
$M_G$,
some of the Yukawa interactions arise from higher dimensional
operators
involving fields $A$ which break the $SO(10)$ symmetry group.
This implies that ${\fam=9{\mathchar"7115 } }_{U,D} \to {\fam=9{\mathchar"7115 } }_{U,D}(A)$.
Every realistic model of $SO(10)$ which has been constructed
has this form; hence one
should view this generalization of the minimal model as a necessity.
(II) {\it{The ratio of electroweak VEV's,}} $\tan\beta = v_U/v_D$,
{\it{ is allowed to be large,}}
$\approx m_t/m_b$.
This is certainly not a necessity; to the contrary,
a simple extrapolation of the results
of \cite{BHS} to such large values of $\tan\beta$ suggests that it is already
excluded by the present limit on $\mu \to e\gamma$.
The case of large $\tan\beta$ in $SO(10)$ has received much attention
\cite{ALS,HRS,RS,ADHRS} partly because it has important ramifications for the
origin of $m_t/m_b = (\lambda_t/\lambda_b) \tan\beta$.
To what extent is this puzzling large ratio to be understood as a large
hierarchy
of Yukawa couplings, and to what extent in terms
of a large value for $\tan\beta$?
If the third generation masses arise from a single interaction of the
form $16_3 16_3 10$
it is possible to predict $m_t$ using $m_b$ and $m_\tau$ as input
\cite{ALS}, providing the theory is
perturbative up to $M_G$.
The prediction is $175 \pm 10 $ GeV \cite{HRS}, and requires
$\tan\beta \approx m_t/m_b$.
In this paper we investigate whether this intriguing possibility is
excluded by the
$\mu \to e \gamma$ signal; or, more correctly, we determine whether
it requires a soft
origin for supersymmmetry breaking, making it
incompatible with the standard supergravity scenario \cite{CAN}.
In the next section we show that $SO(10)$ models with ${\fam=9{\mathchar"7115 } } \to
{\fam=9{\mathchar"7115 } } (A)$
possess new gaugino mixing matrices in the up-quark sector, which did
not arise in the minimal models.
In section III we set our notation for the supersymmetric standard model with
arbitrary gaugino mixing matrices, and we show which mixing matrices
are expected from unified models according to the gauge group
and the value of $\tan\beta$.
In section IV we describe the new phenomenological signatures
which are generated by the gaugino mixing
matrices in the up sector; these signatures are generic to
all models with Yukawa interactions
generated from higher dimensional operators.
The consequences of large $\tan\beta$ for the flavor and CP
violating signatures are analyzed analytically in section V
and numerically in section VI.
The analysis of the first five sections applies to a wide class of models.
In section VII we illustrate the results in the particular
models introduced by Anderson et al \cite{ADHRS}.
As well as providing illustrations, these models have features
unique to themselves.
Conclusions are drawn in section VIII.
\newpage
\noindent{\bf II. New Flavor Mixing in the Up Sector}
\medskip
In [1-4] flavor and CP violating signals are studied in
minimal $SU(5)$ and $SO(10)$ models with moderate $\tan\beta$.
In these models the
radiative corrections
to the scalar mass matrices are dominated by the top quark
Yukawa coupling $\lambda_t$ of the
unified theory, so the scalar mass matrices tend to align with the
up-type Yukawa coupling matrix and all non-trivial flavor mixing matrices
are simply related to the KM matrix.
However, as mentioned above, the minimal models do not give realistic
fermion masses.
One has to insert $SO(10)$ breaking into the Yukawa interactions.
The simplest way to achieve this is to assume that the light fermion masses
come from the non-renormalizable operators
$$
\lambda'_{ij} 16_i{A_1\over M_1}{A_2\over M_2}\ldots {A_\ell\over M_\ell}
10 {A_{\ell+1}\over M_{\ell +1}} \ldots {A_n\over M_{n}} 16_j,\eqno(2.1)
$$
where the $16_i$'s contain the three low energy families, 10
contains the Higgs doublets, and $A$'s are adjoint fields with
vacuum expectation values (VEV's)
which break the $SO(10)$ gauge group.
After substituting in the VEV's of the adjoints, they
become the usual Yukawa interactions with different Clebsch factors
associated with Yukawa couplings of fields with different quantum numbers.
For example in the models introduced by Anderson et al. \cite{ADHRS},
(hereafter
referred to as ADHRS models)
$$
{\fam=9{\mathchar"7115 } }_U = \pmatrix{0&z_uC &0\cr
z'_uC & y_uE &x_uB\cr
0 & x'_uB &A},
{\fam=9{\mathchar"7115 } }_D = \pmatrix{ 0 & z_dC & C\cr
z'_d & y_dE & x_dB\cr
0 & x'_dB &A},
{\fam=9{\mathchar"7115 } }_E = \pmatrix{ 0 & z_eC & 0\cr
z'_eC & y_eE & x_eB\cr
0 & x'_eB & A},\eqno(2.2)
$$
where the $x,y,z$'s are Clebsch factors arising from the VEV's
of the adjoint fields.
Thus realistic fermion masses and mixings can be obtained.
The radiative corrections to the soft SUSY-breaking operators
above $M_G$ are now more complicated.
{}From the interactions (2.1) the following soft supersymmetry
breaking operators are generated:
$$
\lambda^\dagger_{ik}(A) m^2_{k\ell}(A) \lambda_{\ell j}(A) \;
\phi^\dagger_i\phi_j,\eqno(2.3)
$$
where $\phi_i, \phi_j$ are scalar components of the superfields, and
$\lambda_{ij}(A)$ are adjoint dependent couplings,
$\lambda(A) = \lambda' {A_1\over M_1} \ldots {A_n\over M_n}$.
After the adjoints take their VEV's, the
$m^2_{k\ell} (A)$ become the usual soft scalar masses.
If we ignore the wavefunction renormalization of the adjoint fields
(which is valid in the one-loop approximation), this is the same as if
we had replaced the adjoints by their VEV's all the
way up to the ultraheavy scale where the ultraheavy fields are integrated out,
and treated these nonrenormalizable operators as the usual Yukawa
interactions and scalar mass operators.
This is a convenient way of thinking and we will use it in the rest
of the paper.
Above the GUT scale, in addition to the Yukawa interactions
which give the fermion masses
$$
Q{\fam=9{\mathchar"7115 } }_U U^cH_U, \; Q{\fam=9{\mathchar"7115 } }_D D^c H_D, \; E^c
{\fam=9{\mathchar"7115 } }_E L H_D, \eqno(2.4)
$$
the operators (2.1) also lead to
$$
\eqalignno{
Q{\fam=9{\mathchar"7115 } }_{qq} Q H_{U_3}, &E^c {\fam=9{\mathchar"7115 } }_{eu}U^c H_{U_3}, \;
N{\fam=9{\mathchar"7115 } }_{nd}D^cH_{U_3},\cr
Q \lambda_{q\ell} L H_{D_3}, &U^c{\fam=9{\mathchar"7115 } }_{ud}D^c H_{D_3},
N{\fam=9{\mathchar"7115 } }_{n\ell} LH_U,&(2.5)\cr}
$$
where $H_{U_3}, H_{D_3}$ are the triplet partners of the two
Higgs doublets $H_U$ and $H_D$.
Each Yukawa matrix has different Clebsch factors associated with
its elements, so they can not be diagonalized in
the same basis.
The scalar mass matrices receive radiative corrections from Yukawa
interactions of both (2.4) and (2.5), which, in the one-loop approximation,
take the form
$$
\eqalignno{
\Delta{\bf{m}}^2_Q &\propto {\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U +
{\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }_D^\dagger + 2{\fam=9{\mathchar"7115 } }_{qq}
{\fam=9{\mathchar"7115 } }^\dagger_{qq}+ {\fam=9{\mathchar"7115 } }_{q\ell}
{\fam=9{\mathchar"7115 } }^\dagger_{q\ell},\cr
\Delta{\bf{m}}^2_U &\propto 2{\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U +
{\fam=9{\mathchar"7115 } }^\dagger_{eu}{\fam=9{\mathchar"7115 } }_{eu} + 2 {\fam=9{\mathchar"7115 } }_{ud}
{\fam=9{\mathchar"7115 } }^\dagger_{ud},\cr
\Delta{\bf{m}}^2_D &\propto 2{\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D +
{\fam=9{\mathchar"7115 } }^\dagger_{nd}{\fam=9{\mathchar"7115 } }_{nd} + 2{\fam=9{\mathchar"7115 } }_{ud}^\dagger
{\fam=9{\mathchar"7115 } }_{ud},\cr
\Delta{\bf{m}}^2_L &\propto {\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E +
3{\fam=9{\mathchar"7115 } }^\dagger_{q\ell} {\fam=9{\mathchar"7115 } }_{q\ell}
+ {\fam=9{\mathchar"7115 } }^\dagger_{n\ell}{\fam=9{\mathchar"7115 } }_{n\ell},\cr
\Delta{\bf{m}}^2_E &\propto 2{\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7115 } }^\dagger_E
+ 3{\fam=9{\mathchar"7115 } }_{eu}{\fam=9{\mathchar"7115 } }^\dagger_{eu}.&(2.6)
\cr}
$$
In the minimal $SO(10)$ model, scalar mass renormalizations above $M_G$
arise from a single matrix
${\fam=9{\mathchar"7115 } }_U$.
It is therefore possible to choose a ``U-basis''
in which the scalings are purely
diagonal. This is clearly not possible in the general models.
All scalar mass matrices and Yukawa matrices are in general
diagonalized in different bases.
Therefore, flavor mixing matrices should appear in all gaugino vertices,
including in the up-quark
sector (where they are trivial in the minimal models studied in [1-4]).
The up-type quark-squark-gaugino flavor mixing is a novel
feature of the general models.
Its consequences will be discussed in Sec. IV.
Also, the flavor mixing matrices are no longer simply the KM matrix.
They are model dependent and are
different for different types of quarks and charged leptons,
and are fully described in the next section.
\newpage
\noindent {\bf III. Flavor Mixing Matrices in General Superymmetric Standard
Models.}
\medskip
In this section we set our notation
for the gaugino flavor mixing matrices in the
supersymmetric theory below $M_G$, taken to have minimal field content.
We also give general expectations for these matrices in a wide variety
of unified theories.
The most general scalar masses are $6 \times 6$
matrices for squarks and charged
sleptons and $3 \times 3$ matrix for sneutrinos,
$$
\eqalignno{
{\bf{m}}^2_{{U}} &=
\pmatrix{{\bf{ m}}^2_{{U}_L} &
({\fam=9{\mathchar"7110 } }_U+{\fam=9{\mathchar"7115 } }_U\mu\cot \beta)v_U\cr
({\fam=9{\mathchar"7110 } }^\dagger_U +
{\fam=9{\mathchar"7115 } }^\dagger_U\mu\cot\beta)v_U &
{\bf{m}}^2_{{U}_R} },\cr
{\bf{m}}^2_{{D}} &=
\pmatrix{ {\bf{m}}^2_{{D}_L} & ({\fam=9{\mathchar"7110 } }_D
+{\fam=9{\mathchar"7115 } }_D\mu\tan\beta)v_D\cr
({\fam=9{\mathchar"7110 } }^\dagger_D +
{\fam=9{\mathchar"7115 } }^\dagger_D \mu\tan\beta) v_D &
{\bf{m}}^2_{{D}_R}
},\cr
{\bf{m}}^2_{{E}} &=
\pmatrix{
{\bf{m}}^2_{{E}_L} & ({\fam=9{\mathchar"7110 } }_E +
{\fam=9{\mathchar"7115 } }_E\mu\tan\beta)v_D\cr
({\fam=9{\mathchar"7110 } }^\dagger_E +
{\fam=9{\mathchar"7115 } }^\dagger_E\mu\tan\beta)v_D
& {\bf{m}}^2_{{E}_R} },
\cr
{\bf{m}}^2_{{\nu}} &= \left( {\bf{m}}^2_{{\nu}_{ij}}
\right),&(3.1)\cr}
$$
where ${\bf{m}}^2_{{U}_L}, {\bf{m}}^2_{{D}_L},
{\bf{m}}^2_{{U}_R},
{\bf{m}}^2_{{D}_R}, {\bf{m}}^2_{{E}_L}, {\bf{m}}^2_{{E}_R}$
are $3 \times 3$ soft
SUSY-breaking mass matrices for the left-handed and right-handed squarks and
sleptons, and ${\fam=9{\mathchar"7110 } }_U, {\fam=9{\mathchar"7110 } }_D, {\fam=9{\mathchar"7110 } }_E$ are the trilinear
soft SUSY-breaking
terms. To calculate flavor violating processes,
such as $\mu \to e\gamma$, one
can diagonalize the mass matrix ${\bf{m}}^2_{{E}}$ by the
$6\times 6$ unitary
rotation matrix $V_E$ and ${\bf{m}}^2_{{\nu}}$ by the 3 $\times $ 3 unitary
rotation $V_\nu$,
$$
{\bf{m}}^2_{{E}} = V_E \overline{\bf{m}}^2_{{E}} V^\dagger_E,
\hskip .25in
{\bf{m}}^2_{{\nu}} = V_\nu \overline{\bf{m}}^2_{{\nu}}
V^\dagger_\nu,\eqno(3.2)
$$
where ${\bf{\overline{m}}}^2_{{E}},
{\bf{\overline{m}}}^2_{{\nu}}$
are diagonal.
The amplitude for $\mu \to e\gamma$ is given by the
diagrams in Fig.\ 1, summing up all the internal scalar mass eigenstates.
If the entries in the scalar mass matrices are arbitrary, they generally give
unacceptably large rates for flavor violating processes. From the
experimental limits one expects that the first two generation scalar masses
should be approximately degenerate and the chirality-changing mass matrices
${\fam=9{\mathchar"7110 } }_A$ should be approximately proportional
to the corresponding Yukawa
coupling matrices ${\fam=9{\mathchar"7115 } }_A$.
In this paper we treat the
chirality-conserving mass matrices and chirality-changing mass matrices
separately,
i.e., the mass eigenstates are assumed to be purely left-handed or
right-handed,
and the chirality-changing mass terms are treated as a perturbation.
This may not be a good approximation for the third generation where
the Yukawa
couplings are large,
the correct treatment will be used in the numerical studies of Sec.\ VI.
The superpotential contains
$$
W \supset Q^T {{\fam=9{\mathchar"7115 } }}_U U^c H_U +Q^T
{{\fam=9{\mathchar"7115 } }}_D D^c {H}_D +
E^{cT}{{\fam=9{\mathchar"7115 } }}_E L {H}_D,\eqno(3.3)
$$
where ${{\fam=9{\mathchar"7115 } }}_U, {\fam=9{\mathchar"7115 } }_D, {{\fam=9{\mathchar"7115 } }}_E$
are the Yukawa coupling matrices
which are diagonalized by the left and right rotations,
$$
\eqalignno{
{\fam=9{\mathchar"7115 } }_U &= V^*_{U_L}\overline{\fam=9{\mathchar"7115 } }_U V^\dagger_{U_R},\cr
{\fam=9{\mathchar"7115 } }_D &= V^*_{D_L}\overline{\fam=9{\mathchar"7115 } }_D V^\dagger_{D_R},\cr
{\fam=9{\mathchar"7115 } }_E &= V^*_{E_R}\overline{\fam=9{\mathchar"7115 } }_EV^\dagger_{E_L}.&(3.4)\cr}
$$
The soft SUSY-breaking interactions contain
$$
\eqalignno{
\widetilde{Q}^\dagger {\bf{m}}^{2*}_Q \widetilde{Q}
&+ \widetilde{U}^{c\dagger}
{\bf{m}}^2_U
\widetilde{U}^c + \widetilde{D}^{c\dagger} {\bf{m}}^2_D
\widetilde{D}^c + \widetilde{L}^\dagger
{\bf{m}}^2_L \widetilde{L} + \widetilde{E}^{c\dagger}
{\bf{m}}^{2*}_E \widetilde{E}^c\cr
&+ \widetilde{Q}^T {{\fam=9{\mathchar"7110 } }}_U \widetilde{U}^c H_U +
\widetilde{Q}^T {{\fam=9{\mathchar"7110 } }}_D
\widetilde{D}^c {H}_D + \widetilde{E}^{cT}{{\fam=9{\mathchar"7110 } }}_E
\widetilde{L}{H}_D.&(3.5)
\cr}
$$
Because the trilinear terms should be approximately proportional to the Yukawa
couplings, we write
$$
{{\fam=9{\mathchar"7110 } }} = {{\fam=9{\mathchar"7110 } }}_0 + \Delta{{\fam=9{\mathchar"7110 } }} = A {\fam=9{\mathchar"7115 } } +
{{\Delta}}{\fam=9{\mathchar"7110 } }.\eqno(3.6)
$$
The soft-breaking mass matrices are diagonalized by:
$$
\eqalignno{
{\bf{m}}^{2*}_Q &= U_Q \overline{\bf{m}}^{2*}_Q U^\dagger_Q, \;\;
{\bf{m}}^2_U = U_U \overline{\bf{m}}^2_U U^\dagger_U, \;\;
{\bf{m}}^2_D = U_D\overline{\bf{m}}^2_D U^\dagger_D, \cr
{\bf{m}}^2_L &= U_L\overline{\bf{m}}^2_L U^\dagger_L, \;\;
{\bf{m}}^{2*}_E = U_E\overline{\bf{m}}^{2*}_E U^\dagger_E ,
&(3.7)\cr}
$$
$$
{{\Delta{\fam=9{\mathchar"7110 } }}}_U = V^{'*}_{U_L}\Delta \overline{{\fam=9{\mathchar"7110 } }}_U
V^{'\dagger}_{U_R},\;
{{\Delta\fam=9{\mathchar"7110 } }}_D = V^{'*}_{D_L} {{\Delta}}\overline{{\fam=9{\mathchar"7110 } }}_D
V^{'\dagger}_{D_R},\;
{{\Delta\fam=9{\mathchar"7110 } }}_E = V^{'*}_{E_R} \Delta \overline{\fam=9{\mathchar"7110 } }_E
V^{'\dagger}_{E_L}.\eqno(3.8)
$$
In the mass eigenstate basis the rotation matrices $V, U$ appear in the gaugino
couplings,
\newpage
$$
\eqalignno{
{\cal{L}}_g &= \sqrt{2} g' \sum^4_{\pi = 1}
\bigg[ -{1\over 2} \overline{e}_L
W^\dagger_{E_L}\widetilde{e}_L N_n(H_{n\widetilde{B}}
+ \cot \theta_W H_{n\widetilde{w}_{3}}) +
\overline{e}^c_L W^\dagger_{E_R} \widetilde{e}_R
N_n H_{n\widetilde{B}}\cr
&+ {1\over 2} \cot \theta_W\overline{\nu}_L\widetilde{\nu}_L N_n
H_{n\widetilde{w}_{3}}\cr
&+\overline{u}_L W^\dagger_{U_L} \widetilde{u}_L N_n({1\over 6}
H_{n\widetilde{B}}+{1\over 2}\cot\theta_W H_{n\widetilde{w}_3})
+\overline{d}_LW^\dagger_{D_L} \widetilde{d}_L N_n({1\over 6}
H_{n\widetilde{B}}-{1\over 2}\cot\theta_WH_{n\widetilde{w}_3})\cr
&- {2\over 3} \overline{u}_L^c W^\dagger_{U_R}
\widetilde{u}_R N_n H_{n\widetilde{B}} +
{1\over 3} \bar{d}^c_L W^\dagger_{D_R}\widetilde{d}_R
N_n H_{n\widetilde{B}} + h.c.\bigg]\cr
&+ g\sum^2_{c=1} [ \bar{e}_L W^\dagger_{E_L} \widetilde{\nu}_L
(\chi_c K_{c\widetilde{w}})
+\bar{\nu}_L \widetilde{e}_L (\chi^\dagger_c K^*_{c\widetilde{w}})\cr
&+ \bar{d}_L W^\dagger_{D_L} \widetilde{u}_{L} (\chi_c K_{c\widetilde{w}})
+ \bar{u}_L
W^\dagger_{U_L}\widetilde{d}_L
(\chi^\dagger_c K^* _{c\widetilde{w}}) +h.c.]\cr
&+\sqrt{2} g_3 [\bar{u}_L W^\dagger_{U_L}\widetilde{u}_L\widetilde{g}
+ \bar{d}_L
W^\dagger_{D_L} \widetilde{d}_L \widetilde{g}
+ \bar{u}^c_L W^\dagger_{U_R}
\widetilde{u}_R\widetilde{g} +
\overline{d}^c_LW^\dagger_{D_R}\widetilde{d}_R\widetilde{g} + h.c.],&(3.9)\cr}
$$
where\footnote{Neutrino masses are not discussed here and we choose
the neutrino to be in the sneutrino mass eigenstate basis.}
the neutralino and chargino mass
eigenstates are related to the gauge eigenstates by e.g. $\widetilde{B} =
\sum^4_{n=1} H_{n\widetilde{B}}N_n,\, \widetilde{w}_3
= \sum^4_{n=1} H_{n\widetilde{w}_3} N_n,\,
\widetilde{w}^+ = \sum^2_{c=1} K_{c\widetilde{w}}\chi_c$, and
$$
\eqalignno{
W_{E_L} &= U_L^\dagger V_{E_L}, \;
W_{E_R} = U^\dagger_E V_{E_R}, \; W_{U_L}
= U^\dagger_Q V_{U_L}, \; W_{D_L} = U^\dagger_Q V_{D_L},\cr
W_{U_R} &= U^\dagger_U
V_{U_R},\; W_{D_R} = U^\dagger_D
V_{D_R}.\cr}
$$
\noindent There are also non-diagonal chirality-changing mass terms
$$
\eqalignno{
- {\cal{L}}^{n.d}_m &= \widetilde{e}_R^T
W^*_{E_R}(A_E + \mu \tan\beta){\fam=9{\mathchar"7115 } }_E
W^\dagger_{E_L} \widetilde{e}_L v_D + \widetilde{e}_R^T U^T_E
\Delta{{\fam=9{\mathchar"7110 } }}_E
U_L\widetilde{e}_L v_D\cr
&+ \widetilde{d}_L^T W^*_{D_L}(A_D + \mu \tan\beta)
{\fam=9{\mathchar"7115 } }_D W^\dagger_{D_R}
\widetilde{d}_R v_D + \widetilde{d}^T_L U^T_Q
{\Delta{\fam=9{\mathchar"7110 } }}_D U_D
\widetilde{d}_R v_D\cr
&+ \widetilde{u}_L^T W^*_{U_L} (A_U + \mu\cot\beta)
{\fam=9{\mathchar"7115 } }_U W^\dagger_{U_R}
\widetilde{u}_R v_U
+ \widetilde{u}_L^T U^T_Q {{\Delta\fam=9{\mathchar"7110 } }}_U
U_U\widetilde{u}_Rv_U\cr
&+ h.c.&(3.10)\cr}
$$
The lepton flavor violating couplings are summarized in Fig.\ 2.
In the rest of this section we discuss the flavor mixing matrices
in the minimal supersymmetric standard model, minimal and general $SU(5)$
and $SO(10)$ models, with moderate or large $\tan\beta$.
The results are summarized in Table 1.
For the minimal supersymmetric standard model, the radiative corrections
to the soft masses
only come from the Yukawa interactions of the MSSM:
$$
\eqalignno{
\Delta{\bf{m}}^2_Q &\propto
{\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U +\kappa
{\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }^\dagger_D,\cr
\Delta{\bf{m}}^2_U &\propto 2{\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U,\cr
\Delta{\bf{m}}^2_D &\propto 2{\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D,\cr
\Delta{\bf{m}}^2_L &\propto {\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E,\cr
\Delta{\bf{m}}^2_E &\propto 2{\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7115 } }^\dagger_E.
&(3.11)\cr}
$$
We have assumed a boundary condition on the scalar mass matrices
${\bf{m}}^2_{A} \propto I $ at $M_{PL}$,
and $\kappa \neq 1$ represents the possibility that the proportionality
constants are not universal.
For moderate $\tan\beta, {\lambda}_t \gg {\lambda}_b$
so that the radiative corrections are dominated by $\lambda_t$.
Thus one can neglect the ${\fam=9{\mathchar"7115 } }_D$ contribution and the only nontrivial
mixing is $W_{D_L}$.
For large $\tan\beta, {\lambda}_t$ and ${\lambda}_b$ are comparable, so
${\bf{m}}^2_Q$ will lie between ${\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U$
and ${\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }^\dagger_D$.
Therefore both $W_{U_L}$ and $W_{D_L}$ are non-trivial.
with KM matrix and one can ignore them.
For the minimal $SU(5)$ model, there are only two Yukawa matrices,
${\fam=9{\mathchar"7115 } }_U = {\fam=9{\mathchar"7115 } }_{10}, {\fam=9{\mathchar"7115 } }_D = {\fam=9{\mathchar"7115 } }_E=
{\fam=9{\mathchar"7115 } }_5$, and
$$
\eqalignno{
\Delta{\bf{m}}^2_Q &\propto 3{\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U
+ 2\kappa {\fam=9{\mathchar"7115 } }_D
{\fam=9{\mathchar"7115 } }^\dagger_D,\cr
\Delta{\bf{m}}^2_U &\propto 3{\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U +
2 \kappa {\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D,\cr
\Delta{\bf{m}}^2_D &\propto 4{\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D,\cr
\Delta{\bf{m}}^2_L &\propto 4{\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D,\cr
\Delta{\bf{m}}^2_E &\propto 3{\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U +
2\kappa{\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }^\dagger_D.&(3.12)\cr}
$$
For moderate $\tan\beta, {\lambda}_t \gg {\lambda}_b$,
we have non-trivial
mixings for $W_{D_L}$ and $W_{E_R}$, as found in \cite{BH,BHS}.
For large $\tan\beta$, ${\fam=9{\mathchar"7115 } }_D$ can not be ignored, giving
non-trivial mixings for $W_{U_L}$ and $W_{U_R}$.
For the minimal $SO(10)$ model considered in \cite{BHS,DH},
$$
\eqalignno{
\Delta{\bf{m}}^2_Q &\propto 5 {\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger
+ 5 \kappa {\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }_D^\dagger,\cr
\Delta{\bf{m}}^2_U &\propto 5 {\fam=9{\mathchar"7115 } }_U^\dagger{\fam=9{\mathchar"7115 } }_U
+ 5 \kappa{\fam=9{\mathchar"7115 } }_D^\dagger{\fam=9{\mathchar"7115 } }_D,\cr
\Delta{\bf{m}}^2_D &\propto 5 {\fam=9{\mathchar"7115 } }_U^\dagger{\fam=9{\mathchar"7115 } }_U
+ 5\kappa{\fam=9{\mathchar"7115 } }_D^\dagger{\fam=9{\mathchar"7115 } }_D,\cr
\Delta{\bf{m}}^2_L &\propto 5 {\fam=9{\mathchar"7115 } }_U^\dagger{\fam=9{\mathchar"7115 } }_U
+ 5\kappa{\fam=9{\mathchar"7115 } }_D^\dagger{\fam=9{\mathchar"7115 } }_D,\cr
\Delta{\bf{m}}^2_E &\propto 5 {\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger +
5\kappa{\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }_D^\dagger.&(3.13)\cr}
$$
We have non-trivial mixings $W_{D_L}, W_{D_R}, W_{E_L}$, and $W_{E_R}$ for
moderate $\tan\beta$ and
non-trivial mixings for all $W$'s for large $\tan\beta$.
For the general $SU(5)$ or $SO(10)$ models, defined in the last
section, we get non-trivial mixings for all mixing matrices in general.
However, in $SU(5)$ models with moderate $\tan\beta$, the splittings among
${\bf{m}}^2_D$ and ${\bf{m}}^2_L$ are too small (because they
are generated by the
small ${\fam=9{\mathchar"7115 } }_5(A)$) to give significant flavor changing effects.
One might expect that the mixing in the $W_U$'s are smaller than those in the
$W_D$'s because of the larger hierarchy in ${\fam=9{\mathchar"7115 } }_U$
compared with ${\fam=9{\mathchar"7115 } }_D$.
However, a given $W$ is the product of a $U^\dagger$ (which diagonalizes
the scalar
mass matrix) and a $V$ (which diagonalizes the Yukawa matrix).
Even if the mixings in $V_U$'s are smaller than those in $V_D$'s because
of the larger
hierarchies in ${\fam=9{\mathchar"7115 } }_U$, we do not have a general argument for
the size of mixings in $U$ matrices.
This is because $U$ diagonalizes (appropriate combinations of)
known Yukawa matrices and unknown Yukawa matrices appearing
above the GUT scale, (2.5).
The mixings in $U^\dagger$ and $V$ can add up or cancel each other.
Our only general expectation is that these new Yukawa matrices have
similar hierarchical patterns as ${\fam=9{\mathchar"7115 } }_U$ or ${\fam=9{\mathchar"7115 } }_D$.
Without a specific model, one can at most say that all non-trivial $W$'s are
expected to be comparable to $V_{KM}$; the argument that the mixings in
$W_U$'s should be smaller than is $W_D$'s is not valid.
In the minimal models at moderate $\tan\beta$, the leading contributions
to flavor changing processes, such as $\mu\to e\gamma$,
involve diagrams with
a virtual scalar of the third generation.
Although such contributions are highly suppressed by mixing angles, they
dominate because they have large violations of super-GIM\cite{GIM}:
the top Yukawa coupling makes $m_{\widetilde{\tau}}$ very different from
$m_{\widetilde{e}}, m_{\widetilde{\mu}}$.
At large $\tan\beta$, the strange/muon Yukawa couplings get enhanced,
so the splitting between $m_{\widetilde{e}}$ and $m_{\widetilde{\mu}}$
increases, leading
to potentially competitive contributions to flavor
changing processes which do not involve the third generation.
The importance of these new diagrams can be estimated by
comparing the contributions
to $\Delta m^2_{21}$ (in a basis where gaugino vertices are diagonal) when
the super-GIM
cancellation is between scalars of the first two generations (2-1) and
third generations (3-1):
$$
{\Delta m^2_{21}(2\hbox{-}1)\over \Delta m^2_{21} (3\hbox{-}1)}
\simeq {V_{cd}\lambda^2_2\over
V_{td}V_{ts}\lambda^2_t}
\simeq \left\{
\matrix{ 10^{-2}, & \hbox{for}\; \lambda_2 = \lambda_c,\cr
\left({\tan\beta\over 60}\right)^2,
&\hbox{for}\; \lambda_2 = \lambda_s.}\right.\eqno(3.14)
$$
We can see that for large $\tan\beta$ (or any $\tan\beta$ with small
$\lambda_s$ coming from the mixing of Higgs at $M_G$ i.e., $\lambda_s (M_G) =
{\tan\beta\over 60} \lambda_2 (M_G))$, this could be comparable to the flavor
violating effects from the large
splitting of the third generation scalar masses.
However, for the $\mu \to e\gamma$ in $SO(10)$ models, it does not
contribute to diagrams
which are proportional to $m_\tau$, (because it does not involve the third
generation scalars),
the dominant contributions are still those diagrams considered
in \cite{BHS}.
For flavor changing processes which do not need chirality flipping,
such as $K-\overline{K}$ mixing, and all flavor changing processes
in $SU(5)$ models, this non-degeneracy between the first two generations is
important.
The above discussion is summarized in Table 1.
\newpage
\centerline{\bf{\large Table 1}}
\medskip
\begin{center}
\begin{tabular}{|c|c|c|c|c|c|}
\hline
& &\multicolumn{2}{c|}{$SU(5)$} & \multicolumn{2}{c|}{$SO(10)$}\\
\hline
&MSSM&Minimal&general&minimal&general\\
\hline
\hline
$\delta m^2_3$&$\surd$&$\surd$&$\surd$&$\surd$&$\surd$\\
\hline
$\delta m^2_2$&$\bullet$&$\bullet$&$\circ$&$\bullet$&$\circ$\\
\hline
\hline
$W_{U_L}$&$\bullet$&$\bullet$&$\surd$&$\bullet$&$\surd$\\
\hline
$W_{D_L}$&$\surd$&$\surd$&$\surd$&$\surd$&$\surd$\\
\hline
$W_{U_R}$&---&$\bullet$&$\surd$&$\bullet$&$\surd$\\
\hline
$W_{D_R}$&---&---&$\surd^*$&$\surd$&$\surd$\\
\hline
$W_{E_L}$&---&---&$\surd^*$&$\surd$&$\surd$\\
\hline
$W_{E_R}$&---&$\surd$&$\surd$&$\surd$&$\surd$\\
\hline
\end{tabular}
\end{center}
\medskip
{\bf{Table 1:}} Summary table for the flavor mixing matrices:
$
\delta m^2_3$ : important effects due to some third generation
scalars not degenerate with those of first two generations.
$\delta m^2_2$ : non-negligible effects due to
non-degeneracy of the scalars; of the first two generations.
$W_i$ : fermion $i$ and scalar $\widetilde{i}$ are rotated differently to get to mass
basis.
$\surd$ : present for any value of $\tan\beta$.
$\bullet$ : present only for large $\tan\beta$.
$\circ$ : present for large $\tan\beta$, but model dependent for
moderate $\tan\beta$.
--: not present.
$*$ : although present, its effect for moderate $\tan\beta$
on flavor violation is small due to the small
non-degeneracy among different generation scalars.
\newpage
\noindent{\bf{IV. Phenomenology from up-type mixing}}
\medskip
As discussed in the previous section,
unlike the minimal models with moderate $\tan\beta$ studied in
\cite{BH,BHS,DH,BHS2} in generic
GUT's (for any $\tan\beta$) and even for minimal GUT's
(at large $\tan\beta$),
we expect mixing matrices in the up sector.
Having motivated an origin for non-trivial up mixing matrices
$W_{U_{L(R)}} \neq 1$,
we consider some effects they produce.
In the following we simply assume some $W_{u_{L(R)}}$
at the weak scale and consider their phenomenological consequences.
(See however Section
V and the appendix for a discussion of
the scaling of mixing matrices from
GUT to weak scales.)
In particular we discuss
$D - \bar{D}$ mixing, corrections to up-type quark masses,
contributions to the neutron electric dipole moment
(e.d.m.) and the possibility of different
dominant proton decay modes
than those expected from minimal models.
\noindent{\bf IVa. $D-\bar{D}$ mixing:}
To get an idea for the contribution of up-type
mixing matrices to $D-\bar{D}$ mixing, we follow
\cite{GM,NS} and employ
the mass insertion approximation.
The bounds obtained from $D-\bar{D} $ mixing on the $6\times 6$
up-squark mass matrix
$m^2_U = \pmatrix{ m^2_{U_{LL}}&m^2_{U_{LR}}\cr
m^2_{U_{RL}} & m^2_{U_{RR}} }$
(in the basis where gluino and Yukawa couplings are diagonal) are
summarized in \cite{NS}.
For average up-squark mass of $\widetilde{m} = 1 $ TeV, they are
$$
\sqrt{ {m^2_{U_{LL 12}}\over \widetilde{m}^2}
{m^2_{U_{RR 12}}\over \widetilde{m}^2} } \leq
.04, \eqno(4.1)
$$
$$
{m^2_{U_{LR 12}}\over \widetilde{m}^2} \leq .06 . \eqno(4.2)
$$
Consider first (4.1).
In the last section we estimated that the contribution to
$m^2_{12}$ from the slight non-degeneracy between the first
two generation
scalars is generically at most comparabale to that from
the non-degeneracy between the
first two and
third generation scalars.
Thus, for our calculation,
we only consider the contribution from the
splitting between first two and third generation scalars.
Then, for $A=L, R$
$$
\abs{ {m^2_{AA 12}\over \widetilde{m}^2}}
= \abs{ W_{U_{A 13}} W^\dagger_{U_{A 32}}}
\abs{m^2_{\widetilde{t}_A}-m^2_{\widetilde{u}_A}\over \widetilde{m}^2} \leq
\abs{ W_{U_{A 13}}W^\dagger_{U_{A 32}} }.\eqno(4.3)
$$
We see that for $W$'s of the same size as the
corresponding KM matrix elements,
the LHS of (4.1) is of order $4 \times 10^{-4}$, and the bound
is easily satisfied.
Turning to (4.2), note that if ${\fam=9{\mathchar"7110 } }_U
= A{\fam=9{\mathchar"7115 } }_U,\; m^2_{U_{LR 12}} =0$.
However, we expect ${\fam=9{\mathchar"7110 } }_U
= A{\fam=9{\mathchar"7115 } }_U + \Delta{\fam=9{\mathchar"7110 } }_U$,
with $\Delta{\fam=9{\mathchar"7110 } }_U$ induced in running from $M_{PL}$ to $M_G$
having primarily a third generation component in the gauge eigenstate basis.
If all relevant mixing matrix elements are of order the KM matrix elements,
we expect
$\abs{ {m^2_{U_{LR 12}}\over \widetilde{m}^2} } = O
\left(\abs{ {Am_t\over \widetilde{m}^2} V_{td}V_{ts}}\right).$
Again, we see that the bound (4.2) is generically easily satisfied, and
thus we do not in general expect significant contributions to $D-\bar{D}$
mixing.
\medskip
\noindent{\bf IVb. Weak-scale corrections to up-type quark masses:}
It is well known that there are important weak-scale
radiative corrections to the
down quark mass matrix proportional to $\tan\beta$
\cite{HRS,RS,BPR,COPW,H}.
In general unified models, with non-zero $W_U$,
there are also importrant weak
scale corrections to the up quark mass matrix.
{}From the diagram in Fig.\ 3, we have a contribution to up-type
masses proportional to
$m_t$.
We find, again assuming degeneracy between the
scalars of the first two generations,
$$
\eqalignno{
\Delta m^{ij}_u &=
{8\over 3} \left( {\alpha_s\over 4\pi}\right) m_t
\left( {A +\mu \cot\beta \over M_{\widetilde{g}} }\right)
W_{U_{L3i}} W_{U_{R3j}}
\left[ h (x_{t_L}, x_{t_R}) - h (x_{t_L}, x_{u_R})\right.\cr
& -\left.
h(x_{u_L}, x_{t_R}) + h(x_{u_L}, x_{u_R})\right],&(4.4)\cr}
$$
where
$$
x_i \equiv {\widetilde{m}^2_i\over M^2_{\widetilde{g}}},\;\; h(x,y) = {1\over x-y}
\left[ {x \log x\over 1-x} - {y\log y\over 1-y}\right].\eqno(4.5)
$$
The largest fractional change in the mass occurs for the up quark.
If $W_{U_{L(R)31}} $ is comparable to the corresponding KM
matrix element, the contribution to
${\Delta m_u\over m_u}$ is not significant.
However, if each of the $W_{U_{L(R)31}}$ are a factor 3
larger than
the corresponding KM elements we can get sizable contributions.
In Fig.\ 4, we plot ${\Delta m_u\over m_u}$ in
${m_{\tilde{u}}\over M_{\tilde{g}}}
- {m_{\tilde{t}}\over m_{\tilde{u}}}$ space, where we have assumed
$m_{\widetilde{u}_L} = m_{\widetilde{u}_R}\equiv m_{\widetilde{u}} ;\;
m_{\widetilde{t}_L} = m_{\widetilde{t}_R} \equiv m_{\widetilde{t}}$, and we have put
$|W_{U_{L 31}}|= |W_{U_{R 31}}| = 1/30,\;
(A+ \mu \cot \beta)/m_{\widetilde{t}} =3$.
Any deviations from these values
can simply be multiplied in $\Delta m_u/m_u$.
In some regions of the parameter space
it is possible to get the entire up
quark mass as a radiative effect.
\newpage
\noindent{\bf IVc. Neutron e.d.m.:}
If we attach a photon in all possible ways
to the diagram giving the contribution
to $u$-quark mass, we get a contribution to the
$u$-quark e.d.m.,
which is proportional
to $m_t$ for any value of $\tan\beta$.
Evaluating the diagram, we find
$$
d^u = e|F|\sin\phi_u\eqno(4.6)
$$
where
$$
\eqalignno{
F=
&{8\over 3}\left({\alpha_s \over 4{\pi} }\right) m_t
{A + \mu \cot \beta\over M^3_{\widetilde{g}} }
W_{U_{L31}} W_{U_{L33}}^*W_{U_{R31}} W^*_{U_{R33}}\cr
&\times \bigg[
\widetilde{G}_2 (x_{t_{L}}, x_{t_{R}}) - \widetilde{G}_2(x_{t_{L}}, x_{u_{R}})
-\widetilde{G}_2 (x_{t_{R}}, x_{u_{L}})
+ \widetilde{G}(x_{u_{L}}, x_{u_{R}} )\bigg],&(4.7)
\cr}
$$
where
$$\widetilde{G}_2 (x, y) = {g(x) - g(y)\over x-y},\; \; g(x) =
{1\over 2(x-1)^3} [x^2-1-2x \log x]\eqno(4.8)
$$
and
$$
Im \left[
m_t W_{U_{L31}} W_{U_{L33}}^* W_{U_{R31}} W^*_{U_{R33}}
\right] \equiv
| m_t W_{U_{L31}}W^*_{U_{R33}}W_{U_{R31}}
W^*_{U_{R33}}|\sin \phi_u.\eqno(4.9)
$$
In general we expect
a large non-zero sin $\phi_u$.
If the combination of $W$'s appearing in the above is comparable to the
combination giving a down quark e.d.m.,
the $u$-quark contribution will dominate
over the $d$-quark contribution to the
neutron e.d.m.\ considered in \cite{DH} by a
factor ${m_t\over 4m_b\tan\beta}$,
(the factor 4 comes from the quark model result
$d_n=4/3 d_d - 1/3 d_u$).
Hence, the neutron e.d.m.\ may be competitive
with $\mu \to e\gamma$ and $d_e$ as the most
promising flavor changing signal for supersymmetric unification.
\noindent{\bf IVd. Proton decay:}
Finally we turn briefly to the relevance of up-type mixing matrices for
proton decay; in particular to the important question of the charge of the
lepton in the final state.
We know that upon integrating out the
superheavy Higgs triplets we can generate
the baryon number violating operators
${1\over 2M_H} (QQ)(QL)$ and ${1\over M_H} (EU)(DU)$
in the superpotential.
These operators must subsequently be dressed at
the weak scale in order to
obtain four-fermion operators leading to proton decay.
The dressing may be done with neutralinos,
charginos or gluinos where possible.
Since the dressed operator grows with
gauge couplings and vanishes for vanishing
neutralino/chargino/gluino mass,
one might naively expect gluino dressing
to be most important.
However, if the up-type mixing matrices are trivial,
gluino dressed operators can only
lead to proton decay with a neutrino
in the final state.
To see this, we examine each operator separately:
$(eu_a)(d_bu_c)\epsilon^{abc}$ (where $a,b,c$ are color indices)
must involve $u$'s from two different
generations because of the $\epsilon^{abc}$.
One of them has to be a $u$, so the other is a $c$ or a $t$.
If there is no up mixing, the up flavor does not
change in the dressing process, so the final
state would have to contain a $c$ or a $t$.
Since $m_t, m_c > m_p$, this can not happen.
Next, consider $(QQ)(QL) =
u^a_Ld^b_L(u^c_Le_L-d^c_L v_L)\epsilon_{abc}$.
By exactly the same argument as the above, the
$u_L^a d^b_L u^c_Le_L\epsilon_{abc}$
operator can not contribute to proton decay.
Thus, we see that in the absence of mixing
in the up sector, gluino dressing can
only give neutrinos in the final state.
However, the above arguments break down
if up-mixing matrices are non-trivial,
since gluino dressed diagrams give a significant contribution
to the branching ratio for charged lepton modes in proton decay.
A detailed study of flavor mixing in the up sector
\cite{A} concludes that, whether
the wino or gluino dressings are dominant,
the muon final state in proton decay
is of greatly enhanced importance.
Without the mixings, one expects
${\Gamma (p\to K^o \mu^+) \over \Gamma
(p\to K^+\bar{\nu})} \approx 10^{-3}$.
The up mixing in general models increases
this by $O(100)$ making the mode
$p\to K^o\mu^+$
a favorable one for discovery of proton decay.
\newpage
\noindent{\bf Section V. Large $\tan\beta$: Analytic Treatment}
\medskip
The large $\tan\beta$ scenario is interesting for a number of reasons.
For moderate $\tan\beta$, the only way to understand
$m_t \gg m_b, m_\tau$ is to
have $\lambda_t \gg \lambda_b, \lambda_\tau$ at the weak scale.
This gives us little hope of attributing a
common origin for third generation
Yukawa couplings at a higher scale.
However, for large $\tan\beta \sim {{O}}\left({m_t\over m_b}\right)$,
the weak scale $\lambda_t, \lambda_b,
\lambda_\tau$ are comparable and the above
hope is restored. (In fact it is
realized in $SO(10)$ models like the ADHRS example
outlined in section VII).
For us, this is sufficient motivation
to study the large $\tan\beta$ case in
more detail. Also, this case was not
studied in \cite{BHS}. We shall see that
unexpected new features arise in the large $\tan\beta$ limit.
The largest contribution to the
$\mu \to e\gamma$ amplitude comes from
the diagram with $L-R$ scalar mass insertion (Fig.\ 5).
In the $L-R$ insertion approximation,
the amplitude for $\mu_{L(R)}$
decay is
$$
\eqalignno{
F_{L(R)} &= {\alpha\over 4\pi\cos^2\theta_W} m_\tau
W_{E_{L(R)_{32}}}
W_{E_{R(L)_{31}}}
W^*_{E_{L(R)_{33}}}
W^*_{E_{R(L)_{33}}}
(A_E + \mu \tan\beta)\cr
&\times \left[ G_2(m^2_{\widetilde{\tau}_L}, m^2_{\widetilde{\tau}_R})
- G_2(m^2_{\widetilde{e}_L}, m^2_{\widetilde{\tau}_R})
-G_2 (m^2_{\widetilde{\tau}_L}, m^2_{\widetilde{e}_R})+
G_2 (m^2_{\widetilde{e}_L },
m^2_{\widetilde{e}_R})\right],\cr}
$$
where
$$
\eqalignno{
G_2(m^2_1, m^2_2) &= {G_2(m^2_1) - G_2(m^2_2)\over m^2_1 - m^2_2},\cr
G_2 (m^2) &= \sum^4_{n =1} (H_{n\widetilde{B}} + \cot
\theta_W H_{n\widetilde{w}_3})
g_2\left( {m^2\over M^2_n}\right).&(5.1)\cr}
$$
Note, however, that for large $\tan\beta$
the $L-R$ insertion approximation may be a bad one, since
the chirality changing mass for the third generation becomes comparable
to the chirality
conserving masses.
A correct treatment will be used for the numerical analysis
in the next section. We still expect, however, that the
amplitude to be proportional to $W_{E_{32}}W_{E_{31}}$ because of the unitarity
of
the mixing matrices: the sum of contributions from the first
two generations is proportional
to $W_{1i}W^*_{1j} + W_{2i} W^*_{2j} =- W_{3i}W^*_{3j}$
for $i\neq j$, and the contribution from the third generation is itself
proportional to $W_{3i}W^*_{3j}$.
Two simplifications in the dependence of the $\mu\to e\gamma$ rate on parameter
space occur for large $\tan\beta$.
First, since the dominant diagram involves the $L-R$ insertion $(A +
\mu\tan\beta)m_t$, and since $\tan\beta$ is large, the amplitude does not
depend on the weak scale parameter $A$.
Second, in the large tan$\beta$ limit, the chargino mass matrix is
$$
M_\chi = \pmatrix{ M_2 & \sqrt{2} M_W \sin \beta\cr
\sqrt{2} M_W \cos\beta &-\mu} \longrightarrow
\pmatrix{ M_2 & \sqrt{2}M_W\cr
0& -\mu},\eqno(5.2)
$$
and the parameters $M_2, \mu$ have a direct interpretation as the chargino
masses. (Note that this assures us that $\mu\tan\beta$ will likely always be
much bigger than $A$;
for a $\tan\beta$ of 50, the LEP lower bound on chargino mass of 45 GeV
tells us that $\mu \tan\beta > 2 $ TeV, so for $A$ to be comparable to
$\mu\tan\beta$ we must have $A > 2$ TeV.)
In considering $\mu \to e\gamma$ for large $\tan\beta$, two factors come
immediately to mind which tend to (perhaps dangerously) enhance the rate over
the case with moderate $\tan\beta$.
(i) As we have already mentioned,
the dominant contribution to $\mu \to e\gamma$
grows
with $\tan\beta$; the diagram in
Fig.\ 5 is proportional to $\tan\beta$, a
factor of 900 in the rate for $\tan\beta = 60$
compared to $\tan\beta =2$.
(ii) For large $\tan\beta,\;
\lambda_\tau$ can be $O(1)$ and we can not neglect its
contribution to the running of the slepton
mass matrix from $M_G$ to $M_S$
(soft SUSY breaking scale).
This scaling generally splits the third
generation slepton mass even further
from the first two generations, meaning a
less effective super-GIM mechanism
and a larger amplitude for $\mu \to e\gamma$.
While both of the above effects certainly exist,
there are also two sources of
{\it{suppression}} of the amplitude for
large $\tan\beta$, which can together
largely compensate for the above factors:
(i)$'$ Large $\tan\beta$ allows $\lambda_t$ to
be smaller than for moderate
tan$\beta$.
There are two reasons for this. First,
large $\tan\beta$ allows $v_U$
to be larger and so $\lambda_t$ can be smaller to reproduce
the top mass. Secondly, $b-\tau$ unification \cite{CEG}
is achieved with a
smaller $\lambda_\tau$ in the large $\tan\beta$ regime \cite{HRS,RS}.
Since $\lambda_t$ is smaller, a smaller non-degeneracy
between the third and first two
generations is induced in running from $M_{PL} $ to $M_G$,
suppressing the
amplitude compared to the moderate $\tan\beta$ case.
(ii)$'$ In comparing large and moderate tan$\beta$,
we must know how the
mixing matrices $W_{L,R_{3i}}$ (appearing at the
vertices of the diagrams
responsible for $\mu\to e\gamma$) compare in these two cases.
In the moderate tan$\beta$ minimal models discussed in \cite{BHS},
$W_{L,R_{3i}}$ were equal to the corresponding KM matrix elements
$V_{KM3i}$ at
$M_G$, and this equality was approximately
maintained in running form $M_G$ to $M_S$.
As discussed in the previous sections, for more
general models one expects that
the $W_{L(R)_{3i}}$
at $M_G$ are equal to $V_{KM3i}$ at $M_G$ up to some
combination of Clebsches.
One might then expect (as in the minimal models)
that this relationship
continues to approximately hold at lower scales.
In fact for large $\tan\beta$ this expectation is false.
We find that often, the $W_{L(R)3i}$
{\it{decrease}} from $M_G$ to $M_S$, overcompensating
for the increased non-degeneracy between the third and first two
generation slepton masses induced by large $\lambda_\tau$
(point (ii) above).
In the following, we examine the scaling of these mixing
matrices in detail.
Consider first the lepton sector. The renormalization group
equation (RGE) for ${\fam=9{\mathchar"7115 } }_E$ (in the
following
$t ={log \mu\over 16 \pi^2}$) is
$$
- {d{\fam=9{\mathchar"7115 } }_E\over dt} =
{\fam=9{\mathchar"7115 } }_E [3{\fam=9{\mathchar"7115 } }^\dagger_E
{\fam=9{\mathchar"7115 } }_E + Tr
(3{\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D
+ {\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E) - 3 g^2_2 -
{9\over 5} \; g^2_1]\eqno(5.3)
$$
giving
$$
- {d\over dt}{\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E = 6
{\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E + 2
{\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E Tr
(3{\fam=9{\mathchar"7115 } }^\dagger_D {\fam=9{\mathchar"7115 } }_D +
{\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E) - (6g^2_2+{18\over5} \; g^2_1)
{\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E \eqno(5.4)
$$
$$ - {d\over dt} {\fam=9{\mathchar"7115 } }_E {\fam=9{\mathchar"7115 } }_E^\dagger =
6 {\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7115 } }_E^\dagger + 2
{\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7115 } }^\dagger_E Tr (3{\fam=9{\mathchar"7115 } }^\dagger_D
{\fam=9{\mathchar"7115 } }_D +
{\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E) - (6g^2_2 + {18\over 5}\; g^2_1)
{\fam=9{\mathchar"7115 } }_E {\fam=9{\mathchar"7115 } }_E^\dagger . \eqno(5.5)
$$
These in turn imply that the basis in which ${\fam=9{\mathchar"7115 } }^\dagger_{{E}}
{\fam=9{\mathchar"7115 } }_{{E}}$
is diagonal, and the (in general different) basis where
${\fam=9{\mathchar"7115 } }_{{E}}{\fam=9{\mathchar"7115 } }_{{E}}^\dagger$
is diagonal, do not change with scale.
Consider now the evolution of the left handed slepton mass matrix
${\bf{m}}^2_L$.
The RGE for ${\bf{m}}^2_L$ is
$$
-{d\over dt} {\bf{m}}^2_L =
({\bf{m}}^2_L + 2m^2_{H_d}){\fam=9{\mathchar"7115 } }^\dagger_E
{\fam=9{\mathchar"7115 } }_E + 2
{\fam=9{\mathchar"7115 } }^\dagger_E {\bf{m}}^2_E
{\fam=9{\mathchar"7115 } }_E + {\fam=9{\mathchar"7115 } }^\dagger_E
{\fam=9{\mathchar"7115 } }_E
{\bf{m}}^2_L + 2 {\fam=9{\mathchar"7110 } }_E^\dagger
{\fam=9{\mathchar"7110 } }_E + \; \hbox{Gaugino
terms} .\eqno(5.6)
$$
In the basis where ${\fam=9{\mathchar"7115 } }^\dagger_{{E}}
{\fam=9{\mathchar"7115 } }_{{E}}$ is diagonal,
keeping only the $\lambda_\tau$ contribution, the $3i$
entry $(i\neq 3)$ becomes:
$$
- {d\over dt} m^2_{L3i} = \lambda^2_{\tau} m^2_{L3i} + 2
({\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7110 } }_E)_{3i} .\eqno(5.7)
$$
In this basis, we have ${\bf{m}}^2_L
= W^\dagger_L\overline{\bf{m}}_L^2W_L$.
(Here and in the remainder of this section, we abbreviate
$W_{E_{L(R)}}\to W_{L(R)}$).
Assuming degeneracy between scalars of the first two generations,
$
m^2_{L_{3i}} = W_{L 3i}W_{L33}^\dagger
(m^2_{\tau_{L}}- m^2_{e_{L}}) \equiv W_{L3i}W^\dagger_{L 33} \Delta m^2_L.
$
Then (5.7) becomes
$$
- {d\over dt} (W_{L3i} W_{L33}^\dagger\Delta m^2_L)
= \lambda^2_{\tau} (W_{L3i}W^\dagger_{L 33}
\Delta m_L^2) + 2({\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7110 } }_E)_{3i} .\eqno(5.8)
$$
For now, we ignore the $({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}$
term in (5.8), yielding the solution:
$$
(W_{L3i}W_{L33}^\dagger \Delta m^2_L)(M_S) = e^{-I_\tau} (W_{L3i}
W_{L33}^\dagger \Delta m^2) (M_G),
\eqno(5.9)
$$
where
$$
I_i \equiv \int_0^{\log {M_G\over M_S}}
{dt\over 16\pi^2} \lambda^2_i (t).
\eqno(5.10)
$$
Thus,
$$
W_{L 3i} W^\dagger_{L33}(M_S) =
e^{-I_\tau} {\Delta m^2_R(M_G)\over \Delta
m^2_R (M_S)}
W^\dagger_{L3i}W_{L_{33}} (M_G).\eqno(5.11)
$$
Similarly, we find
$$
W_{R 3i} W^\dagger_{R33} (M_S) = e^{-2I_\tau}
{\Delta m^2_R (M_G)\over \Delta m^2_R (M_S)}
W_{R 3i}W^\dagger_{R_{33}} (M_G).\eqno(5.12)
$$
Note that, generically the quantities
${\Delta m^2_{L(R)}(M_G)\over \Delta
m^2_{L(R)} (M_S)}$ are smaller than one, since the third
generation mass
gets split even further from the first
two generations in running from $M_G$ to
$M_S$.
Thus, we find that the $W_{L(R)_{3i}}$
get smaller in magnitude as we scale from
$M_G$ to $M_S$, in contrast with the KM matrix elements $V_{KM3i}$, which scale
as
$$
V_{KM3i} (M_S) = e^{(I_t+I_b)} V_{KM 3i} (M_G).\eqno(5.13)
$$
Suppose that at $M_G$ the
$W_{L(R)}$ are related to $V_{KM}$
though some combination of Clebsches
determined by the physics above the
GUT scale.
$$
W^\dagger_{L(R)33}W_{L(R)3i}
(M_G) = z_{i{L(R)}} V_{KM 3i} (M_G) .\eqno(5.14)
$$
This relationship is not maintained at lower scales; instead we have:
$$
\hspace{-.2in}
W^\dagger_{L33}W_{L 3i} (M_S) = {\Delta m^2_L(M_G)\over \Delta m^2_L (M_S)}
e^{-(I_\tau+I_t+I_b)}
z_{i_{L}}V_{KM 3i}(M_S), \eqno(5.15)
$$
$$
W^\dagger_{R33}W_{R 3i} (M_S) = {\Delta m^2_R(M_G)\over \Delta m^2_R (M_S)}
e^{-(2I_\tau + I_t+I_b)} z_{i_{R}}V_{KM 3i}(M_S).\eqno(5.16)
$$
The dominant contribution to the
$\mu \to e\gamma$
rate is proportional to \newline
$| W^\dagger_{L33} W_{L{32}}W^\dagger_{R33} W_{R{31}} (M_S)|^2 +
| W^\dagger_{L33} W_{L{31}}W^\dagger_{R33} W_{R{32}}(M_S)|^2$,
giving
$$
\eqalignno{
Br( \mu\to e\gamma) &=
\left[
{\Delta m^2_L (M_G)\over \Delta m^2_L (M_S)}
{\Delta m^2_R (M_G)\over \Delta m^2_R (M_S)}
\right]^2
e^{-(6I_\tau + 4I_t + 4I_b)}
\times (|z_{2_L}z_{1_R}|^2 + |z_{1_L}z_{2_R}|^2) \cr
&\times
Br(\mu\to e\gamma, W_{L(R)3i}
W^\dagger_{L(R)33}(M_S)\to V_{KM3i}(M_S))\cr
&\hspace{-.7in}
\equiv \epsilon (|z_{2_L} z_{1_R}|^2 + | z_{1_L} z_{2_R}|^2)
\times Br(\mu\to e\gamma, W^\dagger_{L(R)33}
W_{L(R)33}(M_S)\to V_{KM3i}(M_S)).\cr
&&(5.17)
\cr}
$$
This $\epsilon$ represents a possibly significant suppression
of the rate for large $\tan\beta$.
At this point, the reader may object: it is true that
the $W_{L(R)3i}$ decrease from $M_G$ to $M_S$, but as already mentioned,
the non-degeneracy between the third and first two generating is increasing.
Which effect wins?
We argue that in general there is a net suppression.
This is easiest to see if in computing the $\mu \to e\gamma$ amplitude,
we use the mass insertion approximation rather than mixing
matrices at the the vertices (Fig.\ 6).
Although this may be a poor approximation,
it serves to illustrate our point.
(Of course no such approximation is made in our numerical work.)
{}From the diagram it is clear that the amplitude is proportional to
$m^2_{L32} m^2_{R31} (M_S)$.
{}From (5.7), we see that the rate scales as
$$
\left( m^2_{L32}m^2_{R31}\right)^2 (M_S) = e^{-(6I_\tau +4I_t+4I_b)}
\left(m^2_{L_{32}}m^2_{R_{31}}\right)
(M_G),\eqno(5.18)
$$
a net suppression.
In the mass insertion approximation, then, the terms
${\Delta m^2(M_G)\over \Delta m^2(M_S)}$
in (5.17) serve to exactly compensate
for the increased non-degeneracy between $m^2_{\widetilde{e}_L}$
and $m^2_{\widetilde{\tau}_L}$; what remains is
still a suppression.
This, together with (i)$'$ invalidates the naive
expectation that the theory
is ruled out in most regions of parameter space due
to the enhancing factors (i)
and (ii), (although there are still stringent
constraints on the parameter space).
The above analysis suggests that individual
lepton number conservation
is an infrared fixed point of the MSSM (whereas individual quark
number conservation is
an ultraviolet fixed point).
A more complete analysis of scaling for the lepton sector
and a discussion of scaling in the quark sector
is presented in the appendix.
\newpage
\noindent{\bf Section VI. Large $\tan\beta$: Numerical Results}
\medskip
The amplitude for $\mu \to e\gamma$ depends on the $6\times 6$ slepton mass
matrix ${\bf{M}}^2$.
In the basis where ${\bf{m}}^2_L,{\bf{ m}}^2_E$ are diagonal, we have
$$
{\bf{M}}^2 = \pmatrix{
{\bf{\overline{m}}}^2_{{E}_{L}} + {\bf{D}}_L & {\bf{k}}\cr
{\bf{k}}^\dagger &{\bf{\overline{m}}}^2_{{E}_{R}} + {\bf{D}}_R}\eqno(6.1)
$$
where in the large $\tan\beta$ limit, $D_i = - (T_{3i} - Q_i
\sin^2\theta_W)M^2_Z$
is the $D$-term contribution,
and $k_{ij} = \mu m_\tau \tan\beta W_{L 3i}W_{R 3j}$.
The amplitude from Fig.\ 1 for $\mu_L$ decay is
$$
F_L = {\alpha\over 4\pi\cos^2\theta_W}
W^\dagger_{L i2} G_2({\bf{M}}^2)_{LR ij}
W^\dagger_{Rj1}
$$
where
$$
G_2 ({\bf{M}}^2) \equiv \pmatrix{G_2(M^2)_{LL} & G_2(M^2)_{LR}\cr
G_2(M^2)_{RL} & G_2(M^2)_{RR} }, \eqno(6.3)
$$
In \cite{BHS}, ${\bf{M}}^2$ was
approximately diagonalized by the $\mu m_\tau
\tan\beta$ insertion approximation,
and $G_2 ({\bf{M}}^2)$ was calculated using
this approximate diagonalization.
Since here $\tan\beta$ is large,
we wish to avoid making such an approximation,
and numerically diagonalize the full 6$\times 6\; {\bf{M}}^2$.
Faced with a rather large parameter space,
we must decide which parameters to use in
our numerical work.
We have firstly decided to do our analysis only for large
$\tan\beta$, since the moderate
$\tan\beta$ scenario has been covered in \cite{BHS}.
Secondly, we choose to present our results
in a different way than in \cite{BHS},
where the rates for $\mu \to e\gamma$ were
plotted against a combination of
Planck scale
and weak scale parameters.
In our work, we compute $\mu \to e\gamma$
entirely in terms of weak scale parameters.
In particular, we assume that the necessary
condition for a significant
$\mu\to e\gamma$ rate exist at the weak
scale, namely non-trivial mixing
matrix $W_{L,R_{3i}}$
and non-degeneracy between third and first two generation
slepton masses.
In the previous sections, we have shown a possible way
in which these
ingredients may be produced.
Our plots for $\mu \to e \gamma$ rates are
made against low energy parameters,
and we separately plot the regions
in low energy parameter space predicted
by our particular scenario for generating
$\mu\to e \gamma$.
This way, our plots are in terms of
experimentally accessible quantities
and can be thought of as constraining
the parameter space of the effective 3-2-1
softly broken supersymmetric theory resulting from the spontaneous
breakdown of a GUT.
(We use the GUT to relate weak scale gaugino masses.)
Our low energy plots have no dependence on the physics above the GUT scale,
all the model dependence comes into the
predictions for low energy parameters the GUT
makes.
If the predicted region of low energy parameters
corresponds to a $\mu \to e\gamma$
rate exceeding experimental bounds, the theory is ruled out.
There is a more practical reason for working
directly with low-energy parameters
specific to large $\tan\beta$: the well known
difficulty in achieving electroweak symmetry breaking in this regime.
Working with high energy parameters,
and imposing universal scalar masses necessitates a fine-tune to achieve
$SU(2) \times U(1)$ breaking. However, we
have nowhere in our analysis made the
assumption of universal scalar masses,
hence the Higgs masses and squark/slepton
masses are independent in our analysis,
and therefore the $\mu$ parameter is
not tightly constrained by squark/slepton masses.
Working with weak scale parameters allows us to assume that the desired
breaking has occurred without having to know the details of the breaking.
With the aforementioned assumption about the existence of a GUT, and
assuming degeneracy between the first two generations, the rate for
$\mu \to e\gamma$
depends on the weak scale parameters
$\mu, \tan\beta, M_2, m^2_{\widetilde{e}_L}, m^2_{\widetilde{\tau}_L},
m^2_{\widetilde{e}_R}, m^2_{\widetilde{\tau}_R}, $
$W_{L 3i}, W_{R 3i}$.
We know that the amplitude depends on
$W_{L(R)3i} $ simply through the product
$W_{L 3i} W_{R 3j}$, so for normalization in our plots we put
$W_{L(R)3i} = V_{KM 3i}$.
Any deviation from this can be simply multiplied into the rate.
We also fix $\tan\beta = 60$,
and put $m_{\widetilde{\tau}_{L(R)}} = m_{\widetilde{e}_{L(R)}} -
\Delta_{L(R)}$.
Next, we use some high energy bias to relate $m_{\widetilde{e}_L}$
and $m_{\widetilde{e}_R}$:
we assume that their difference is proportional to $M_2$
( as would be the case if they
started out degenerate and were split only through
different gauge interactions),
so we put $m_{\widetilde{e}_L}= m_{\widetilde{e}_R} - rM_2$.
In all specific models we have looked at, $r$ is small
(less than about .2).
We find that, as long as $r$ is small, the rate has little dependence on
its exact value, so we put
$r=0, m_{\widetilde{e}_L} = m_{\widetilde{e}_R} \equiv \overline{m}_{\widetilde{e}}$.
We also found that as long as
${\Delta_L\over \Delta_R}$ is close to 1, there is
little dependence on its actual value either, so we put
$\Delta_L=\Delta_R\equiv \Delta$.
Now, the $\mu \to e\gamma$ rate depends
only on $\mu, M_2, \overline{m}_{\widetilde{e}}$ and
$\Delta$,
and we have the large $\tan\beta$ interpretation of $\mu$ and $M_2$
as chargino masses.
Fixing $\overline{m}_{\widetilde{e}}=300$ GeV, we make contour plots
of $Br(\mu \to e\gamma)$.
The rate scales roughly as ${\overline{m}_{\widetilde{e}}}^{-4}$ and $\mu^2$
for scalar masses heavy compared with gaugino masses.
In Fig.\ 7, we fix $\mu$ and plot in $M_2-\Delta$ space.
In Fig.\ 8, we fix $\Delta$ and plot in $\mu - M_2$ space.
In Fig.\ 9, we plot the values of
$\Delta$ predicted by the GUT against $M_2$, for various
values of $\lambda_t (M_G)$
and $A_e(M_S)$ and for two values of $b_5$, the gauge beta function
coefficient above the
GUT scale.
In Fig.\ 10, we plot the suppression
factor $\epsilon$ for the same parameter
set as in Fig.\ 9.
We see that, over a significant region in parameter space,
$\epsilon$ is small, between 0.2 and 0.01.
It is clear from Fig.\ 7 that, with no suppression,
a typical value for $\Delta$ of
0.3 ($\times$300GeV)
would give rise to rates above the current bound of
$Br(\mu \to e\gamma) < 4.9 \times 10^{-11}$\cite{B}.
However, from Fig.\ 10, the suppression from $\epsilon$
is seen to be typically 20, allowing $ \Delta$'s
of up to 0.45 ($\times$300GeV).
We see that $\epsilon $ is crucial in giving the GUT more
breathing room, as $\Delta$'s of less than 0.45 are more common.
{}From Fig.\ 8 it is also clear that
regions of small $\mu$ and $M_2$ (that is,
light chargino masses) are preferred.
Smaller $\mu$ is preferred because it decreases the L-R mass
$\mu m_\tau \tan\beta$,
small $M_2$ is preferred because
in the limit that the neutralino mass
tends to zero,
the diagrams Fig.\ 5 vanish.
We also note that smaller $\mu, M_2$
are preferred for electroweak symmetry
breaking\cite{HRS,RS}.
If $\mu$ and $M_2$ are both small, the
lightest supersymmetric particle (LSP)
can be quite light,
(but where it has significant higgsino component,
it must be heavier than 45 GeV
in order to be consistent with the precise
measurement of the $Z$ width), and it
annihilates (primarily through its higgsino components) through a $Z$
into fermion antifermion pairs much like a heavy neutrino.
The contribution of the LSP to energy density of the universe
$\Omega h^2$ then just depends on its mass, and the size of
its higgsino components, both of which only depend on
$\mu$ and $M_2$ in the large $\tan\beta$
limit.
In Fig.\ 11, we make a plot of $\Omega h^2$ in $\mu-M_2$
space. We see that it is possible to get $\Omega
\sim O(1)$ in some region of the
parameter space.
\newpage
\noindent{\bf{VII. The Example of ADHRS Models}}
\medskip
In this section, we study the ADHRS models \cite{ADHRS} which are known to
give realistic fermion masses and mixing patterns.
These models are specific enough for us to do calculations and
make some real predictions.
Although not necessarily correct, they are good representatives of
general GUT models.
We believe that by studying them,
one can see in detail the general features of generic realistic GUT models and
the differences between them and the minimal $SU(5)$ or $SO(10)$ models.
As mentioned in Sec.\ II, in ADHRS models, the three families of
quarks and leptons lie in
three 16 dimensional representations of $SO(10)$, and the two low
energy Higgs doublets lie in a single 10 dimensional representation.
Only the third generation Yukawa couplings come from a renormalizable
interaction
$$
\lambda_{33} 16_3 16_3 10 .\eqno(7.1)
$$
All other small Yukawa couplings come from nonrenormalizable interactions
after integrating out the heavy fields.
These interactions can be written in general as
$$
16_i \lambda_{ij} (A_a) 16_j 10.\eqno(7.2)
$$
The $A_{a}$'s are fields in the adjoint representation of $SO(10)$
and their
vevs break $SO(10)$ down to the Standard Model gauge group.
Therefore, these Yukawa couplings can take different values for
fermions of the same generation
with different quantum numbers under $SU(3)\times SU(2) \times U(1)$
and a realistic fermion mass pattern and nontrivial KM matrix can be generated.
In ADHRS models, the minimal number (four) of operators is assumed to
generate the up,
down-type quark and charged lepton Yukawa coupling matrices
${\fam=9{\mathchar"7115 } }_U, {\fam=9{\mathchar"7115 } }_D$
and ${\fam=9{\mathchar"7115 } }_E$, and they take the form at $M_G$
$$
{\fam=9{\mathchar"7115 } }_U = \pmatrix{
0 &z_uC &0\cr
z'_uC&y_u E & x_uB\cr
0 & x'_uB& A},
{\fam=9{\mathchar"7115 } }_D =
\pmatrix{
0 &z_dC &0\cr
z'_dC & y_dE & x_dB\cr
0 & x'_dB & A},
{\fam=9{\mathchar"7115 } }_E=\pmatrix{
0 & z_eC & 0\cr
z'_eC& y_eE & x_eB\cr
0 & x'_eB & A},
\eqno(7.3)
$$
where the $x, y, z$'s are Clebsch factors arising from the VEV's of the
adjoint Higgs fields $A_a$.
This form is known to give the successful relations $V_{ub}/V_{cb}
= \sqrt{m_u/m_c}$
and $V_{td}/V_{ts} = \sqrt{ m_d/m_s}$ \cite{HR} so it is well motivated.
Strictly speaking, the interaction (7.2) become the usual Yukawa
form only after the adjoints $A_a$ take their VEV's
at the GUT scale.
However, as we explained in Sec.\ II, they can be treated
as the usual Yukawa interactions up to the
ultraheavy scale (which
we will assume to be $M_{PL}$) where the ultraheavy fields are integrated out
if the wavefunction renormalizations of $A_a$'s are ignored.
In the one-loop approximation which we use later
in calculating radiative corrections from $M_{PL}$ to $M_G$, they give
the same results, because the wavefunction
renormalizations of the adjoints $A_a$ only contribute at the
two-loop order.
This makes our analysis much easier.
Above the GUT scale, in addition to the Yukawa interactions (2.4) which give
the fermion masses, we have the interactions (2.5) as well.
Each Yukawa matrix has different Clebsch factors $x, y, z$
associated with its elements.
All the Yukawa matrices have the ADHRS form
$$
\lambda_I = \pmatrix{ 0& z_I C & 0\cr
z'_I C & y_IE & x_IB\cr
0 & x'_I B & A },\; I = qq, eu, ud, q\ell, nd, n\ell .\eqno(7.4)
$$
If each entry of the Yukawa matrices is generated dominantly by a single
operator,
like in the ADHRS models, then the phases of the same entries of all Yukawa
matrices are identical.
One can remove all but the $\lambda_{22}$ phases by rephasing the operators.
After phase redefinition only $E$ is complex and is responsible for
CP violation.
In order to generate the realistic fermion mass and mixing pattern,
one expects the following hierarchies,
$$
\eqalignno{
{B\over A} &\sim V_{cb} \hskip .14in\sim \epsilon^2, \cr
{E\over A} &\sim {m_s\over m_b} \hskip 8pt\sim \epsilon^2, \cr
{C\over E} &\sim \sin \theta_c \sim \epsilon,
\;\;\; \hbox{where } \; \epsilon \sim 0.2.&(7.5)\cr}
$$
The hierarchical Yukawa matrices can be diagonalized approximately
\cite{HR},
the unitary rotation matrices which diagonalize them at
the GUT
scale can be approximately written as
$$
{\fam=9{\mathchar"7115 } } =\pmatrix{ 0& zC & 0\cr
z'C & y\abs{E}e^{\widetilde{\phi}}&xB\cr
0 & x'B&A } = V^*_F \bar{\fam=9{\mathchar"7115 } } V^\dagger_B,\eqno(7.6)
$$
$$
V_F \simeq \pmatrix{ e^{i\phi} & S_{F_1} e^{i\phi} & 0\cr
-S_{F_1} & 1 & S_{F_2}\cr
S_{F_1}S_{F_2} & -S_{F_2} & 1\cr},\eqno(7.7)
$$
$$
V_B \simeq \pmatrix{
1 & S_{B_1} & 0\cr
-S_{B_1} e^{-i\phi} & e^{-i\phi}& S_{B_2} \cr
S_{B_1} S_{B_2} e^{-i\phi} & -S_{B_2}e^{-i\phi} & 1},\eqno(7.8)
$$
where
$$
\eqalignno{
S_{F_2}
&= {xB \over A},\; S_{B_2} = {x'B \over A},\cr
S_{F_1}
&= {zC \over E'},\; S_{B_1} = { z'C \over E'},\cr
E' &= \abs{ y E - S_{F2} S_{B2} A }
=
\abs{ y |E| e^{i\widetilde{\phi} } - {x'x B^2\over A} },\cr
\widetilde{\phi} &= arg (E),\;
\phi = arg (y E - {x'x B^2\over A} ).
}
$$
The soft SUSY-breaking scalar masses for the three low energy generations
and trilinear $A$ terms are
assumed to be universal\footnote{If the nonrenormalizable operators
already appear in the superpotential of the underlying
supergravity theory, the $A$ terms will be different for different dimensional
operators, and will induce unacceptably large $\mu \to e\gamma$
rate because the triscalar interactions and the Yukawa interactions can not
be diagonalized in the same basis for the first two generations.
In theories where the nonremormalizable operators come from
integrating out heavy fields at $M_{PL}$ and all the relevant
interactions have the same
$A$ term, the resulting nonrenormalizable operators will also have
the same $A$ term.}
at Planck scale $M_{PL}$ as in
\cite{BHS}.
Beneath $M_{PL}$,
the radiative corrections from the Yukawa couplings destroy the
universalities and render the mixing matrices non-trivial.
In the one-loop approximation, the radiative corrections to the
soft SUSY-breaking parameters at $M_{G}$ are simply related
to the Yukawa coupling matrices and therefore the relations between
general mixing matrix elements and KM matrix elements are also simple.
This allows us to see the similar hierarchies in the general mixing
matrices and the KM matrix very clearly.
Although the one-loop approximation may not be a good approximation for
quantities involving third generation Yukawa couplings,
we will be satisfied with it since it simplifies things a lot and the
uncertainties in other quantities such as Clebsch factors are
probably much bigger than the errors made in the one-loop approximation.
The RG equations, for ${\bf{m}}^2_E$ as an example, from $M_{PL}$ to
$M_G$ are
$$
\eqalignno{
{d\over dt} {\bf{m}}^2_E &= {1\over 16\pi^2}
\bigg[2(2{\fam=9{\mathchar"7115 } }_E{\bf{m}}^2_L
{\fam=9{\mathchar"7115 } }^\dagger_E + 2 {\fam=9{\mathchar"7115 } }_E m^2_{H_D}
{\fam=9{\mathchar"7115 } }^\dagger_E
+ {\bf{m}}^2_E
{\fam=9{\mathchar"7115 } }_E {\fam=9{\mathchar"7115 } }^\dagger_E +
{\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7115 } }_E^\dagger {\bf{m}}^2_E + 2
{\fam=9{\mathchar"7110 } }_E{\fam=9{\mathchar"7110 } }^\dagger_E)\cr
&+ 3(2 {\fam=9{\mathchar"7115 } }_{eu} {\bf{m}}^2_U {\fam=9{\mathchar"7115 } }^\dagger_{eu}
+ 2 {\fam=9{\mathchar"7115 } }_{eu} m^2_{H_{U_3}} {\fam=9{\mathchar"7115 } }^\dagger_{eu}
+ {\bf{m}}^2_E {\fam=9{\mathchar"7115 } }_{eu}
{\fam=9{\mathchar"7115 } }^\dagger_{eu} + {\fam=9{\mathchar"7115 } }_{eu} {\fam=9{\mathchar"7115 } }^\dagger_{eu}
{\bf{m}}^2_E + 2{\fam=9{\mathchar"7110 } }_{eu}
{\fam=9{\mathchar"7110 } }^\dagger_{eu})\cr
&- \hbox{gaugino mass contribution}\bigg].&(7.9)\cr}
$$
In the one-loop approximation, the gaugino mass contributions are diagonal
and the same for all three generations, so they can be absorbed into
the common scalar masses and do not affect the diagonalization.
The corrections to scalar masses at $M_G$ have the following leading
flavor dependence
$$
\eqalignno{
\Delta {\bf{m}}^2_E &\propto 2 {\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7115 } }_E^\dagger + 3
{\fam=9{\mathchar"7115 } }_{eu} {\fam=9{\mathchar"7115 } }^\dagger_{eu}\cr
&= 5 \pmatrix{ \overline{z^2_e} C^2 & \overline{z_e y_e}CE^* &
\overline{ z_ex_e}CB\cr
\overline{z_ey_e} CE & \overline{z'^{2}_e}C^2 +
\overline{y^2_e} \abs{E}^2 + \overline{x^2_e} B^2 &
\overline{ y_ex'_e}EB + \bar{x}_e BA\cr
\overline{z_ex'_e} CB & \overline{y_ex'_e} E^*B +
\bar{x}_e BA &
\overline{x'^2_e} B^2 + A^2},&(7.10)\cr}
$$
where the overline represents the weighted average of the Clebsch factors,
$\overline{z^2_e} = {1\over 5} (2z^2_e + 3z^2_{eu})$ and so on.
Because $\Delta {\bf{m}}^2_E$ is hierarchical, assuming no big $x, y$
Clebsches (ADHRS models have some
big $z$ Clebsches), the rotation matrix which diagonalizes it can be
given approximately as
$$
U_E(M_G) \simeq
\pmatrix{
1 & \bar{S}_{E_1} & \bar{S}_{E3}\cr
-\bar{S}_{E_1} e^{-i\widetilde{\phi}}
& e^{-i\widetilde{\phi}} & \bar{S}_{E3}\cr
\bar{S}_{E_1}\bar{S}_{E_2}e^{-i\widetilde{\phi}} -
\bar{S}_{E3} & -\bar{S}_{E2} e^{-i\widetilde{\phi}} & 1},
\eqno(7.11)
$$
where
$$
\eqalignno{
\bar{S}_{E_2} &= {\bar{x_e}B\over A},\; \bar{S}_{E_3}
= { \overline{z_ex'_e}CB\over A^2}\cr
\bar{S}_{E_1} &= { \overline{z_ey_e} C\abs{E}\over
\overline{z'^{2}_e} C^2 + \overline{y^2_e} \abs{E}^2 +
(\overline{x^2_e} - \overline{x}^2_e)B^2 },\;
e^{-i\widetilde{\phi}} = {E\over \abs{E} }.
\cr}
$$
Similarly, for other scalar masses the leading flavor dependent corrections
at $M_G$ are
$$
\eqalignno{
\Delta {\bf{m}}^2_L &\propto {\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E
+ 3{\fam=9{\mathchar"7115 } }^\dagger_{q\ell}{\fam=9{\mathchar"7115 } }_{q\ell} +
{\fam=9{\mathchar"7115 } }^\dagger_{n\ell}{\fam=9{\mathchar"7115 } }_{n\ell},\cr
\Delta{\bf{m}}^2_Q &\propto {\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger
+ {\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }^\dagger_D
+ 2{\fam=9{\mathchar"7115 } }_{qq}{\fam=9{\mathchar"7115 } }^\dagger_{qq}
+ {\fam=9{\mathchar"7115 } }_{q\ell}{\fam=9{\mathchar"7115 } }^\dagger_{q\ell},\cr
\Delta {\bf{m}}^2_U &\propto 2 {\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U
+ {\fam=9{\mathchar"7115 } }^\dagger_{eu}{\fam=9{\mathchar"7115 } }_{eu}
+ 2 {\fam=9{\mathchar"7115 } }_{ud}{\fam=9{\mathchar"7115 } }^\dagger_{ud},\cr
\Delta{\bf{m}}^2_D &\propto 2{\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D
+ {\fam=9{\mathchar"7115 } }^\dagger_{nd}{\fam=9{\mathchar"7115 } }_{nd} +
2{\fam=9{\mathchar"7115 } }^\dagger_{ud}{\fam=9{\mathchar"7115 } }_{ud},&(7.12)
\cr}
$$
and the rotation matrices which diagonalize them are given by
expressions similar to (7.11) with Clebsches
replaced by the appropriate ones.
Then, the mixing matrices appearing at the lepton-slepton-gaugino
vertices are given by
$$
\eqalignno{
&W_{E_L}(M_G) = U^\dagger_L V_{E_L}
\cr
&\simeq
\pmatrix{
1 & S_{E_{L1}} - {S}_{L1} e^{i(\widetilde{\phi} - \phi_e)} &
\bar{S}_{L_1} (\bar{S}_{L_2} -S_{E_{L_2}}) e^{i\widetilde{\phi}}
- \bar{S}_{L_3}\cr
(\bar{S}_{L_1}- \bar{S}_{E_{L_1}}) e^{-i(\widetilde{\phi}-\phi_e)} &
e^{i(\widetilde{\phi}-\phi_e)} & -(\bar{S}_{L_2} - S_{E_{L_2}})
e^{i\widetilde{\phi} }\cr
-S_{E_{L_1}}(\bar{S}_{L_2} - S_{E_{L_2}}) e^{-i{\phi}_e} + \bar{S}_{L_3} &
(\bar{S}_{L_2}-S_{E_{L_2}}) e^{-i\phi_e} & 1
},\cr
&&(7.13)\cr}
$$
$$
\eqalignno{
&W_{E_L}(M_G) = U^\dagger_E V_{E_R}
\cr
&\simeq
\pmatrix{
e^{i\phi_e}
& S_{E_{R1}}e^{i{{\phi}}} e-\bar{S}_{E_{1}}e^{i\widetilde{\phi}}&
\bar{S}_{E_1}(\bar{S}_{E_2} - S_{E_{R2}})
e^{i\widetilde{\phi}} - \bar{S}_{E_3}\cr
\bar{S}_{E_1}e^{i\phi_e}-S_{E_{R1}}e^{i\widetilde{\phi}} &
e^{i\widetilde{\phi}} & - (\bar{S}_{E_2}- S_{E_{R_2}})e^{-i{\phi}}\cr
-{S}_{E_{R1}}(\bar{S}_{E_2} - S_{E_{R2}})+ \bar{S}_{E_3} e^{i{\phi_e}} &
\bar{S}_{E_2} - S_{E_{R2}} &1},\cr
&&(7.14)\cr}
$$
\noindent where
$$
\eqalignno{
S_{E_{L_1}} &=
{z'_e C\over E'_e},\; S_{E_{R_1}}
= {z_e C\over E'_e},\;
E'_e = |y_e E - {x'_e x_e B^2\over A}|,\cr
S_{E_{L_2}} &= {x'_e B\over A},\;
S_{E_{R_2}} = {x_e B\over A},\;
\phi_e =
arg \left( y_e\abs{E} e^{i\widetilde{\phi}} -
{x'_e x_e B^2\over A}\right)\cr
& {\hskip 1.75in}
\simeq \widetilde{\phi}\;\; (\hbox{if}\; y_e \sim x_e, x'_e),\cr
\bar{S}_{E_1} &= { \overline{z_ey_e} C |E|\over \overline{z'^2_e} C^2
+ \overline{y^2_e} \abs{E}^2
+ (\overline{x^2_e} - \overline{x}^2_e) B^2},
\; \bar{S}_{E_2}={\bar{x}_eB\over A}, \;
\bar{S}_{E_3} =
{\overline{z_e x'_e} C B\over A^2},\cr
\bar{S}_{L_1} &= { \widehat{z'_e y_e} C|E|\over \widehat{z^2_e} C^2 +
\widehat{y^2_e}
|E|^2+ (\widehat{x^2_e} - \widehat{x}^2_e)B^2 },\;
\bar{S}_{L_2} =
{\widehat{x'_e} B\over A},\;
\bar{S}_{L_3} = { \widehat{z'_e x_e} CB\over A^2}, \cr
\hbox{and}&\cr
\bar{x}_e & = {1\over 5} (2x_e + 3 x_{eu}),\;\; \widehat{x'_e}= {1\over 5}
(x'_e + 3 x'_{g\ell} + x'_{n\ell}) \;\; \hbox{etc.}
\cr}
$$
Note that
$$
\bar{S}_{L_3} \sim \ {\hbox{Clebsch}}\ \times
{CB\over A^2},\; S_{E_{L_1}} (\bar{S}_{L_2}-\bar{S}_{E_{L_2}})
= {\hbox{Clebsch}} \times {CB\over EA}.\eqno(7.15)
$$
If there is no very big or small Clebsch involved and
no accidental cancellation,
$\bar{S}_{L_3}, \bar{S}_{E_3}$ can be neglected in $W$'s.
Compared with $V_{KM}$,
$$
\eqalignno{
&V_{KM} (M_G) = V^\dagger_{U_L} V_{D_L}
\cr
&\simeq
\pmatrix{
1 & S_{D_{L_1}}-S_{U_{L_1}} e^{-i(\phi_d-\phi_u)} & -S_{U_{L_1}}
(S_{D_{L_2}}-S_{U_{L_2}}) e^{i\phi_u}\cr
S_{U_{L_1}} - S_{D_{L_1}} e^{-i(\phi_d-\phi_u)} & e^{-i(\phi_d -\phi_u)}
& (S_{D_{L_2}} - S_{U_{L_2}}) e^{i\phi_u}\cr
S_{D_{L_1}} (S_{D_{L_1}} -S_{U_{L_2}}) e^{-i\phi_d} & -(S_{D_{L_2}} -
S_{U_{L_2}}) e^{-i\phi_d} & 1 },\cr
&&(7.16)\cr}
$$
where
$$
\eqalignno{
S_{U_{L1}} &= {z_u C\over E'_u},\;
E'_u = \abs{ y_u E-{x_u x'_uB^2\over A} },
\phi_u =
arg \left( y_u E- { x_u x'_u B^2\over A }\right),\cr
S_{D_{L1}} &= {z_d C\over E'_d},\;
E'_d = \abs{ y_d E - {x_dx'_d B^2\over A} },\;
\phi_d =
arg \left( y_dE - {x_dx'_d B^2\over A }\right),\cr
S_{U_{L2}} &= {x_u B\over A},\cr
S_{D_{L2}} &= {x_d B\over A},\cr}
$$
we can see that the $W$'s and $V_{KM}$ do have similar hierarchical patterns,
but have different Clebsch factors associated with their entries.
When a specific model is given, one can calculate all the Clebsch factors and
make some definite predictions for that particular model.
For example, the ADHRS Model 6, which gives results in good agreement with
the experimental data, has the following four effective fermion mass operators
$$
\eqalignno{
O_{33} &= 16_3\; 10\; 16_3,\cr
O_{23} &= 16_2\; {A_Y\over A_X}\; 10\; {A_Y\over A_X} 16_3,\cr
O_{23} &= 16_2 {A_X\over M} \; 10\; {A_{B-L}\over A_X}\; 16_2
\;\;\; {\hbox{or other 5 choices}},\cr
O_{12} &= 16_1 \left( {A_X\over M }\right)^3\; 10\;
\left( {A_X \over M}\right)^3
16_2,&(7.17)
\cr}
$$
where $A_X, A_Y, A_{B-L}$ are adjoint's of $SO(10)$ with
VEV's in the $SU(5)$
singlet, hypercharge, and $B-L$ directions.
There are six choices of $O_{22}$ operators
which give the same predictions for
the fermion masses and mixings, but
different Clebsches for other operators
appearing above $M_G$.
Fortunately, they do not enter the leading
terms of the most important mixing
matrix elements $W_{E_{L32}}, W_{E_{L31}}, W_{E_{R32}}, W_{E_{R31}}$,
which appear in
the leading contributions to the amplitudes of
LFV processes and the electric
dipole moment.
The magnitude of the mixing matrix elements $V_{KM 32}, V_{KM 31}, W_{E_{L32}},
W_{E_{L31}},\\ W_{E_{R32}},
W_{E_{R31}}$, and the relevant Clebsch factors are listed in
Tables 2 and 3.
\begin{center}
{\large\bf{Table 2}}
\vskip .20in
\begin{tabular}{|c|c|c|c|c|c|c|}
\hline
&$u$ & $d$ & $e$ &$eu$ & $q\ell$ &$ n\ell$\cr
\hline
$x$& $-1$ & $-{1\over 6}$ & ${3\over 2}$ & $-6$ & ${1\over 4}$ & 0\cr
\hline
$x'$ & $-1$ & $-{1\over 6}$ & ${3\over 2}$ & $-6$ & ${1\over 4}$ & 0\cr
\hline
$y$ & 0 & 1 & 3 & - & -& -\cr
\hline
$z$ & $-{1\over 27}$ & 1 & 1 & $-{ 1\over 27}$ & 1& 125\cr
\hline
$z'$&$-{1\over 27}$ & 1 & 1 & $-{1\over 27}$ & 1 & 125\cr
\hline
\multicolumn{7}{|c|}{$\widehat{x'_e}
= {9\over 20},\; \overline{x_e} =- 3$}\cr
\hline
\end{tabular}
\end{center}
{\bf Table 2:} Clebsch factors for Yukawa coupling matrices in
ADHRS \mbox{model 6}.
\newpage
\medskip
\centerline{\bf{Table 3}}
\vskip .20in
\begin{center}
\begin{tabular}{|c|c|c|c|}
\hline
&ADHRS models&Model 6&Relevant process\cr
\hline
$\abs{W_{E_{L_{32}}}/V_{ts} }$&
$\abs{ {\widehat{x}'_e - x'_e\over x_d - x_u} }$ &
1.26&\cr
&&&\cr
\hline
$\abs{W_{E_{R32}}/V_{ts} }$ &
$\abs{ {\bar{x}_e -x_e\over x_d - x_u} }$
& 5.4 &\cr
&&&\cr
\hline
$\abs{ W_{E_{L31}}/V_{td} }$ & $\abs{ {z'_e y_d (\widehat{x}'_e - x'_e)\over
z_d y_e (x_d - x_u)} }$ & 0.42 &\cr
&&&\cr
\hline
$\abs{ W_{E_{R31}}/V_{td} }$ & $\abs{ {z_e y_d (\bar{x}_e - x_e)
\over z_d y_e (x_d - x_u)} }$& 1.8&\cr
&&&\cr
\hline
$\abs{ { W_{E_{L32}}W_{E_{R31}}\over V_{ts} V_{td} } }$ &
$\abs{ {z_e y_d(\bar{x}_e-x_e)(\widehat{x}'_e - x'_e)
\over z_d y_e (x_d-x_u)^2 }}$ &
2.268 & $\mu\to e\gamma$ amplitude\cr
&&&\cr
\hline
$\abs{ {W_{E_{R32}} W_{E_{L31}}\over V_{ts}V_{td} } }$ &
$\abs{ {z'_e y_d (\bar{x}_e - x_e)(\widehat{x}'_e-x'_e)\over
z_d y_e (x_d - x_u)^2 } } $& 2.268 &$\mu \to e\gamma$ amplitude\cr
&&&\cr
\hline
$
\Bigg|
\left(
{
\sqrt{2} W_{E_{L31}} W_{E_{R32}}\over
\sqrt{ | W_{E_{L32}} W_{E_{R31}}|^2 +
|W_{E_{R32}} W_{E_{L31}} |^2}
}
\right) /
\left( { V_{td}\over V_{ts} }\right)
\Bigg|$
& $ \bigg| {
\sqrt{2} z'_e z_e y_d\over \sqrt{ (z^2_e + z'^2_e)} z_d y_e
}\bigg| $ &
${1\over 3} $ & $d_e$\cr
\hline
\end{tabular}
\end{center}
{\bf Table 3:} The relevant Clebsch factors for $\mu \to e\gamma$
and $d_e$ in ADHRS \mbox{model 6}.
\vskip .4in
In ADHRS models $\tan\beta$ is large.
The $\mu \to e\gamma$ rate for large $\tan\beta$ has been calculated
in Sec.\ V and VI for $W_{E_{L32}} = W_{E_{R32}} = V_{ts}$
and $W_{E_{L31}} = W_{E_{R31}} = V_{td}$.
To obtain the predictions of
ADHRS models we only have to multiply the results
by the suitable Clebsch factors.
The relevant Clebsch factors for Model 6 are listed in Table 3.
For a generic realistic GUT model with small tan$\beta$, for example the
modified ADHRS models in which
the down type Higgs lies predominantly in some
fields which do not interact with the three low energy
generations and contain
only a small fraction of the
doublets in the 10 which interact with the low
energy generations \cite{CDW}, most of analysis should still hold.
In this case the leading contributions to $\mu \to e\gamma$ are
the same ones as in the minimal $SO(10)$ model of Ref. \cite{BHS}
(Fig.\ 10
$b_{L,R}, c_{L,R},
c'_{L,R}$ of \cite{BHS}).
The diagrams $c_{LR}, c'_{LR}$ involve
the corrections to the trilinear scalar
couplings.
In the one-loop approximation the
leading corrections to ${\fam=9{\mathchar"7110 } }_E$ at
$M_G$ contain pieces proportional to
${\fam=9{\mathchar"7115 } }_E, {\fam=9{\mathchar"7115 } }_E ({\fam=9{\mathchar"7115 } }^\dagger_E
{\fam=9{\mathchar"7115 } }_E + 3 {\fam=9{\mathchar"7115 } }^\dagger_{q\ell} {\fam=9{\mathchar"7115 } }_{q\ell} +
{\fam=9{\mathchar"7115 } }^\dagger_{n\ell}
{\fam=9{\mathchar"7115 } }_{n\ell}), (2 {\fam=9{\mathchar"7115 } }_{E} {\fam=9{\mathchar"7115 } }^\dagger_{E}+ 3
{\fam=9{\mathchar"7115 } }_{eu}
{\fam=9{\mathchar"7115 } }^\dagger_{eu}){\fam=9{\mathchar"7115 } }_E$
respectively.
The piece proportional to ${\fam=9{\mathchar"7115 } }_E$ can be absorbed into
${\fam=9{\mathchar"7110 } }_{E_0}$ by a
redefinition of $A_E$, the other two
pieces are proportional to the product of
${\fam=9{\mathchar"7115 } }_E$ and the corrections to the scalar masses,
$$
\eqalignno{
\Delta{\fam=9{\mathchar"7110 } }_E &= \Delta{\fam=9{\mathchar"7110 } }_{E_R}
+ \Delta{\fam=9{\mathchar"7110 } }_{E_L}\cr
\Delta{\fam=9{\mathchar"7110 } }_{E_R} &= {1\over \mu_{E_R}} \Delta {\bf{m}}_E^2
\;{\fam=9{\mathchar"7115 } }_E\cr
\Delta{\fam=9{\mathchar"7110 } }_{E_L} &= {1\over \mu_{E_L}} {\fam=9{\mathchar"7115 } }_E
\;\Delta{\bf{m}}_{L}^2&(7.18)\cr}
$$
where $\mu_{E_{L}}, \mu_{E_{R}}$ are proportional constants
$(\mu_{E_R} = \mu_{E_L} = {6m^2_0 + A^2_0 \over 3A_0}$
in one-loop approximation).
The LFV couplings in Fig.\ 2e,
$\: \widetilde{e}^T_R U^T_E \Delta
{\fam=9{\mathchar"7110 } }_E U_L \widetilde{e}_L v_D$, now can be
written as
$$
\eqalignno{
&{1\over \mu_{E_R}} \widetilde{e}^T_R U^T_E \Delta {\bf{m}}^2_E
{\fam=9{\mathchar"7115 } }_E U_L \widetilde{e}_L v_D + {1\over \mu_{E_L}}
\widetilde{e}^T_R U^T_E {\fam=9{\mathchar"7115 } }_E
\Delta{\bf{m}}^2_L U_L \widetilde{e}_L v_D\cr
= & {1\over \mu_{E_R}} \widetilde{e}_R^T \Delta{\overline{\bf{m}}}^2_E W^*_{E_R}
\bar{\fam=9{\mathchar"7115 } }_E W^\dagger_{E_L}
\widetilde{e}_L v_D
+ {1\over \mu_{E_L}} \widetilde{e}^T_R W^*_{E_R} \bar{\fam=9{\mathchar"7115 } }_E
W^\dagger_{E_L} \Delta{\overline{\bf{m}}}^2_L
\widetilde{e}_L v_D,&(7.19)\cr}
$$
where the overline means that the matrix is diagonal.
Again, the amplitudes are given by the same formulas as in \cite{BHS}
(eqn.\ 29, 30), except that $V^e_{32}V^e_{31}(V^{e*}_{33})^2$
has to be replaced by $W_{E_{L32}} W_{E_{R31}} W^*_{E_{L33}} W^*_{E_{R33}}$,
and $W_{E_{R32}} W_{E_{L31}}W^*_{E_{R33}}W^*_{E_{L33}}$, and ${5\over 7} I'_G$
by ${1\over \mu_{E_{R}}} {\Delta}\overline{m}^2_{E33}$ and ${1\over \mu_{E_L}}
{\Delta}\overline{m}^2_{L33}$.
The results in \cite{BHS} are only modified by some
multiplicative factors and therefore represent the
central values for the LFV processes.
It was pointed out in \cite{BHS,DH}
that the electric dipole moment of
the electron $(d_e)$ constitutes an independent and equally
important signature for the $SO(10)$ unified theory as $\mu \to e \gamma$
does.
The diagrams which contribute to the electric dipole moment of the
electron are the same as the ones which contribute to $\mu \to e\gamma$,
with $\mu_L(\mu^c_L)$ replaced by $e_L(e_L^c)$.
Thus a simple relation between $d_e$ and
the $\mu\to e\gamma$ rate was obtained in
the minimal $SO(10)$ model \cite{BHS},
$$
\Gamma (\mu \to e\gamma) = {\alpha\over 2} m^3_\mu |F_2|^2,\eqno(7.20)
$$
$$
|d_e| = e|F_2| \abs{ {V_{td}\over V_{ts}}} \sin\phi = e \sqrt{
{2\Gamma (\mu\to e\gamma)\over \alpha m^3_\mu}}\abs{
{V_{td}\over V_{ts} } } \sin\phi,\eqno(7.21)
$$
where $\phi$ is an unknown new CP violating phase defined by
$$
Im [m_\tau(V^e_{31})^2 (V^{e*}_{33})^2] = |m_\tau(V^e_{31})^2
(V^{e*}_{33})^2| \sin \phi.
$$
In a more generic $SO(10)$ model, such as the ADHRS model,
we still have
this simple relation but the mixing matrix elements have to be
replaced by the $W$'s:
$$
|d_e| = e \sqrt{ {2\Gamma (\mu \to e\gamma)\over \alpha m^3_\mu} }
{ \sqrt{2}|W_{E_{L31}}W_{E_{R31}}|\over \sqrt{
|W_{E_{L32}} W_{E_{R31}}|^2 + |W_{E_{R31}}W_{E_{L31}}|^2} }
\sin\phi',\eqno(7.22)
$$
where $\phi'$ is defined by
$$
Im[m_\tau W_{E_{L31}}W_{E_{R31}} W^*_{E_{R33}}W^*_{E_{L33}}]=
|m_\tau W_{E_{L31}}W_{E_{R31}}W^*_{E_{L33}}W^*_{E_{R33}}| \sin \phi'.
$$
In particular, in ADHRS models there is only one CP violating phase, so the
phase $\phi'$
can be related to the phase appeared in the KM matrix of the Standard Model.
{}From eqn. (7.13) (7.14) (7.16) we can see that
$\phi' \approx \phi_e, \phi_e \approx \phi_d \approx
\widetilde{\phi}, \phi_u =0$
(because $y_u =0$).
The rephrase invariant quantity $J$ of the KM matrix is given by
$$
\eqalignno{
J &= Im V_{ud} V_{tb} V^*_{td} V^*_{ub}\cr
&\simeq - S_{U_{L1}} S_{D_{L1}} (S_{D_{L2}}
- S_{U_{L2}})^2\sin \phi_d.&(7.23)\cr}
$$
Therefore the CP violating
phase appeared in $d_e$ related to the CP
violation in the Standard Model by
$$
\sin \phi' \simeq {J\over |V_{td}| |V_{ub}|}.\eqno(7.24)
$$
Finally, as mentioned in the Sec.\ III,
we consider the possibility that
the slight non-degeneracy
between the first two generation scalar masses could
give a significant contribution to the flavor
changing processes because of the larger mixing matrix elements.
We still use ADHRS models as an example to estimate this contribution to the
LFV process $\mu \to e\gamma$.
For an order of magnitude estimate, the mass insertion approximation in the
super-KM basis employed in \cite{BH}
will serve as a convenient method.
After rotating the $\Delta {\bf{m}}^2_E$ in
eqn.\ (7.10) to the charged lepton
mass eigenstate basis,
the contribution
from the first two generations to ${\Delta m^2_{E_{21}}}$ is
$$
\eqalignno{
{\Delta}{m}^2_{E_{21}} (2{\hbox{-}}1)
\simeq V_{E_{R22}} V^*_{E_{R21}} \Delta m^2_{E22}
+ V_{E_{R12}}
V^*_{E_{R11}} \Delta m^2_{E_{11}}
+ V_{E_{R22}} V^*_{E_{R11}}
\Delta m^2_{E_{21}}\cr
\simeq [- {z_e C\over E'_e} { \overline{z'^2_e} C^2 +
\overline{{y}^2_e}
|E|^2 + \overline{x^2_e} B^2\over A^2} + {z_eC\over E'_e}
{\overline{z^2_e} C^2\over A^2} +
e^{-i\phi_e} {\overline{z_e {y}_e}} {CE\over A^2} ]
\Delta m^2_{E_{33}}\cr
\simeq - {z_e\over y_e} \left[ {\overline{y^2_e} C|E|
+ \overline{x^2_e} {CB^2\over |E|}
+ \overline{z_ey_e}C|E|\over A^2}
\right]
\Delta m^2_{E_{33}} \hspace{.7in}\cr
{(\hbox{assume }} \; z_e = z'_e \;
{\hbox{as in ADHRS model}}).
&(7.25)
\cr}
$$
Compared with the result found in \cite{BH} for minimal $SU(5)$:
$$
\eqalignno{
\Delta m^2_{E_{21}} (BH)
&= V^*_{ts} V_{td} \Delta m^2_{E_{33}} \cr
&\simeq - {z_d C\over E'_d}
{(x_d - x_u)^2B^2\over A^2} \Delta m^2_{E_{33}}\cr
&\simeq - {z_d\over y_d} {(x_d - x_u)^2 {CB^2\over |E|}\over A^2}
\Delta m^2_{E_{33}}, &(7.26)
\cr}
$$
we can see that if the Clebsch factors
are $O(1)$, this contribution is comparable
to that of the minimal $SU(5)$ model.
In order for this contribution to be
competitive with the dominant diagrams (Fig.\ 10
$b_{L,R}, c_{L,R}, c'_{L,R} $ of \cite{BHS})
which are enhanced by ${m_\tau\over m_\mu}$, large
Clebsch factors are required. While it is possible to have large Clebsch
factors, we consider
them as model dependent, not generic to all realistic unified theories.
\newpage
\noindent{\bf VIII. Conclusions}
In supersymmetric theories,
the Yukawa interactions which violate flavor symmetries
not only generate the quark and lepton mass matrices, but necessarily also lead
to radiative breaking of flavor symmetries in the squark and slepton mass
matrices, leading to a variety of flavor signals. While such effects have been
well studied in the MSSM and, more recently, in minimal unified models, the
purpose of this paper
has been to explore these phenomena in a wide class of grand
unified models which have realistic fermion masses.
We have argued that, if the hardness scale $\Lambda_H$
is above $M_G$,
the expectation for all realistic grand unified
supersymmetric models is that non-trivial flavor mixing matrices should occur
at {\it all} neutral gaugino vertices.
These additional, weak scale, flavor violations
are expected to have a form similar to the Kobayashi-Maskawa matrix. However,
the precise values of the matrix elements are model dependent and have
renormalization group scalings which differ
from those of the Kobayashi-Maskawa
matrix elements.
It is the non-triviality of the flavor mixing matrices of neutral gaugino
couplings in the up quark sector which strongly distinguishes between the
general and minimal unified models, as shown in Table 1. Although the minimal
unified models provide a simple approximation to flavor physics, they are not
realistic, so we stress the important new result that flavor mixing in the up
sector couplings of neutral gauginos is a necessity in unified models. this
leads to four important phenomenological consequences. While the $D^0-
\bar{D}^0$
mixing induced by this new flavor mixing is generally not close to the
present experimental limit, it could be much larger than that predicted in the
standard model.
The new mixing in the up-quark sector implies that there may be significant
radiative contributions to the up quark mass matrix which arise when the
superpartners are integrated out of the theory. This is illustrated in Figure
4, where the new mixing matrix elements have been taken to be a factor of three
larger than the corresponding Kobayashi-Maskawa matrix elements. In this case
the entire up quark mass could be generated by such a radiative mechanism:
above the weak scale the violation of up quark flavor symmetries lies in the
squark mass matrix.
The electric dipole moment of the neutron, $d_n$, is a powerful probe of the
neutral gaugino flavor mixing induced by unified theories. In the minimal
$SO(10)$ theory, $d_n$ arises from the flavor mixing in the down sector, which
leads to a down quark dipole moment, $d_d$. However, in realistic models the
flavor mixing in the up quark sector leads to a $d_u$ which typically provides
the dominant contribution to $d_n$. Thus the neutron electric dipole moment is
a more powerful probe of unified supersymmetric theories than previously
realized.
The presence of flavor mixing in the up sector plays a very important role in
determining the branching ratio for a proton to decay to $K^0 \mu^+$. In the
minimal models, without such mixings, this branching ratio is expected to be
about $10^{-3}$: the charged lepton mode will not be seen and experimental
efforts must concentrate on the mode containing a neutrino, $K^+ \nu$.
However, including these mixings the charged lepton branching ratio is greatly
increased to about $0.1$. While this number is very model dependent, we
nevertheless think that this effect greatly changes the importance of searching
for the charged lepton mode.
These four phenomenologocal consequences are sufficiently interesting that we
stress once more that they appear as a necessity in a wide class of unified
theories. The absence of mixing in the up sector is a special feature
of the minimal models. Since the flavor sectors of the minimal models
must be augmented to obtain realistic fermion masses, any conclusions based on
the absence of flavor mixings in the up sector are specious.
A second topic addressed in this paper
is the effect of large $\tan \beta$ on the lepton process, $\mu \rightarrow e
\gamma$ which is expected in unified supersymmetric $SO(10)$ models. The
amplitude for this process has a contribution proportional to $\tan \beta$.
In this paper, we have found that the naive
expectation that large $\tan \beta$ in supersymmetric $SO(10)$ is excluded
by $\mu \rightarrow e \gamma$ is incorrect, at least for all values of the
superpartner masses of interest. Contour plots
for the $\mu \rightarrow e \gamma$
branching ratio are shown in Figures 7 and 8. It depends sensitively on the
parameter $\Delta$, which is the mass splitting between the scalar electron
and scalar tau, and is plotted in Figure 9. Lower
values of the top quark Yukawa coupling, which for large $\tan \beta$ still
give allowed predictions for the $b/\tau$ mass ratio, give a much reduced value
for $\Delta$, thereby reducing the $\mu \rightarrow e \gamma$ rate and
partially compensating the $\tan^2 \beta$ enhancement. A further significant
suppression of an order of magnitude is induced by the renormalization group
scaling of the leptonic flavor mixing angles, and is shown in Figure 10. The
net effect is that while the case of $\tan \beta \approx m_t/m_b$ is not
excluded in SO(10), the $\mu \rightarrow e \gamma$ rate is still typically
larger than for moderate $\tan \beta$, so that this process
provides a more powerful probe of the theory as $\tan \beta $ increases.
For large $\tan \beta$, $\mu$ and $M_2$ become the physical masses of the two
charginos. The $\mu \rightarrow e \gamma$ contours of Figure 8
show that $\mu$ and $M_2$ should not be too large, providing an important limit
to the chargino masses in the large $\tan \beta$ limit. Furthermore, this
constrains the LSP mass to be quite small.
We find that in this region it is still possible for the LSP to account
for the observed dark matter, and even to critically close the universe, as can
be seen from Figure 11.
However, the requirement that the LSP mass be larger than 45 GeV suggests that
the two light charginos will not be light enough to be discovered at LEP II.
As an example of theories with both a realistic flavor sector and large $\tan
\beta$ we studied the models introduced by Anderson et al. The flavor sectors
of these theories are economical: the free parameters can all be fixed from the
known quark and lepton masses and mixings. Hence the flavor mixing matrices at
all neutral gaugino vertices can be calculated. These are shown for the lepton
sector of model 6 in Table 3. The Clebsch factors enhance the $\mu \rightarrow
e \gamma$ amplitude by a factor of 2.3, and suppress $d_e$ by a factor of 3.
Even taking the top quark Yukawa coupling to have its lowest value the rate for
$\mu \rightarrow e \gamma$ in this theory is very large. Another interesting
feature of these theories is that the flavor sectors contain just a single CP
violating phase. This means that the phase which appears in the result for
$d_n$ and $d_e$ can be computed: since it is closely related to the phase of
the Kobayashi-Maskawa matrix it is not very small. That which appears in
$d_e$ is given in eqn.\ (7.24) and is numerically
about 0.2. We have computed the
radiative corrections to $m_u$ in the ADHRS models and have found that the new
mixing matrices in the up sector are not large enough to yield sizable
contributions: thus the ADHRS analysis of the quark mass matrices is not
modified. Furthermore, due to a cancellation special to these theories, there
is no contribution to $d_n$ from the up quark at one loop.
\newpage
\noindent {\bf Acknowledgements}
The authors would like to thank Hitoshi Murayama for many useful
discussions. This work was supported in part by the Director, Office of Energy
Research, Office of High Energy and Nuclear Physics, Division of High Energy
Physics of the U.S. Department of Energy under contract No. DE-AC03-76SF00098
and in part by the National Science Foundation under Grant No. PHY-90-21139.
The work of N.A-H was supported by an NSERC '67 fellowship.
\vskip .6in
\noindent Note Added:
While finalizing this work,
we received a preprint by Ciafaloni,
Romanio and Strumia\cite{CRS},
where the large $\tan\beta$ scenario is also considered. However,
unlike this work, they assume strict universality in soft scalar masses,
such that imposing electroweak symmetry breaking leads
them into a region of
parameter space with
a high mass (1 TeV) for the sleptons.
In their discussion of general models, they do not include
flavor violating RG scaling of scalar masses above $M_G$.
\newpage
\noindent{\bf{Appendix}}
In this appendix, we first give a more complete treatment of mixing
matrix scaling in the lepton sector, and then give a treatment
for the quark sector.
Let us return to (5.7) and consider the effect of including the
$({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}$
term. In general the scaling from $M_{PL}$ to $M_G$
will generate a
${\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E$
not diagonal in the same basis as ${\fam=9{\mathchar"7115 } }^\dagger_E
{\fam=9{\mathchar"7115 } }_E$,
so we
expect some non-zero $({\fam=9{\mathchar"7110 } }^\dagger_{E}{\fam=9{\mathchar"7110 } }_E)_{3i}$.
{}From the RGE for ${\fam=9{\mathchar"7110 } }_E$, neglecting gauge couplings,
$$
- {d\over dt} {\fam=9{\mathchar"7110 } }_E = {\fam=9{\mathchar"7110 } }_E [ 5{\fam=9{\mathchar"7115 } }^\dagger_E
{\fam=9{\mathchar"7115 } }_E
+ Tr (3{\fam=9{\mathchar"7115 } }^\dagger_D
{\fam=9{\mathchar"7115 } }_D + {\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E)]
+ {\fam=9{\mathchar"7115 } }_E [4{\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7110 } }_E + Tr
(6{\fam=9{\mathchar"7110 } }_D {\fam=9{\mathchar"7115 } }_D^\dagger + 2
{\fam=9{\mathchar"7110 } }_E{\fam=9{\mathchar"7115 } }_E^ \dagger)].\eqno(A.1)
$$
We have
$$
\eqalignno{
-{d\over dt} ({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E) &=5[{\fam=9{\mathchar"7110 } }^\dagger_E
{\fam=9{\mathchar"7110 } }_E
{\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E
+ {\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7110 } }^\dagger_E
{\fam=9{\mathchar"7110 } }_E]+ 2Tr(3{\fam=9{\mathchar"7115 } }^\dagger_D
{\fam=9{\mathchar"7115 } }_D + {\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E)
{\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E\cr
&+ 8{\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E{\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E
+ ({\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E
+ {\fam=9{\mathchar"7115 } }_E^\dagger{\fam=9{\mathchar"7110 } }_E)
Tr(6{\fam=9{\mathchar"7110 } }_D{\fam=9{\mathchar"7115 } }^\dagger_D
+ 2{\fam=9{\mathchar"7110 } }_E{\fam=9{\mathchar"7115 } }^\dagger_E).&(A.2)
\cr}
$$
Then, to first order in the off diagonal parts of
${\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E$ and
${\fam=9{\mathchar"7110 } }_E{\fam=9{\mathchar"7110 } }_E^\dagger$, and keeping only third
generation Yukawa couplings we have
$$
- {d\over dt} ({\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7110 } }_E)_{3i}
= ({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}[17 {\lambda}^2_\tau + 6
{\lambda}^2_b + 6\eta {\lambda}_b {\lambda}_\tau],\eqno(A.3)
$$
where $\eta \equiv {\zeta_{D_{33}}\over \zeta_{E_{33}} }$.
Because of the large numerical coefficient in front of
$\lambda^2_\tau, \lambda^2_b$
in the above equation,
(${\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}$ is driven to zero more
rapidly than $W_{L3i}$, after which it ceases to have any effect on the
running of $W_{L3i}$.
More explicitly, from (5.7) we have that
$$
{d\over dt} (m^2_{L3i}(t)e^{\int^t_0 dt'\lambda^2_\tau (t')}) =- 2
({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}(t)
e^{\int^t_0 dt' \lambda^2_\tau(t')}.
\eqno(A.4)
$$
Solving (A.3) for $({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}(t)$ and
inserting into (A.4) we get
$$
-{d\over dt} \left( m^2_{L 3i}(t)e^{\int^t_0 dt'\lambda^2_\tau (t')}\right)
= 2({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}(M_G)e^{-\int^t_0
dt'[16 \lambda^2_\tau
+6\lambda^2_b + 6\eta \lambda_b \lambda_t](t')}.\eqno(A.5)
$$
Integrating (A.5), we find
$$
\eqalignno{
m^2_{L 3i} (M_S)e^{I_\tau} - m_{L 3i}^2 (M_G) &=
-2 \int_0^{ {1\over 16\pi^2}\log
{M_G\over M_S}} dt
e^{-\int^t_0 dt'[16\lambda^2_\tau + 6 \lambda^2_b + 6\eta
\lambda_b\lambda_\tau](t')}\cr
&\times ({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}(M_G)\cr
&\equiv \delta ({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i} (M_G).&(A.6)\cr}
$$
So, we have
$$
m^2_{L 3i}(M_S) = e^{-I_\tau}[m^2_{L 3i} (M_G) + \delta
({\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7110 } }_E)
(M_G)].\eqno(A.7)
$$
We expect $m^2_{L 3i}$ and
${({\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7110 } }_{E})}_{3i}$
to be related by some
combination of Clebsches $x$ at $M_G$ as follows:
$$
({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i} =
{A^2_0\over m^2_0} xm^2_{L 3i}\eqno(A.8)
$$
Where $A_0, m^2_0$ are the universal $A$ parameter and scalar mass
at $M_{PL}$,
respectively.
Then, we have from (A.7)
$$
W^\dagger_{L33}
W_{L 3i} (M_S) = e^{-I_\tau}{\Delta m^2(M_G)\over \Delta m^2 (M_S)}
[1 + \delta {A^2_0\over m^2_0} x] W^\dagger_{L33}W_{L 3i} (M_G).\eqno(A.9)
$$
Clearly if $\delta {A^2_0\over m^2_0} x$ $\ll 1$,
inclusion of the $({\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E)_{3i}$ term in (5.7)
do not change any of our results.
If $\delta{A^2_0\over m^2_0} x$ $\sim 1 $ or $\gg 1$, we
can still of course use (A.9), but the suppression effect may disappear.
A simple estimate shows, however, that $\delta$ itself is already
small $\sim {1\over 10}$,
and so we are only in trouble if ${A^2_0\over m^2_0} x$ is big.
To see this, replace
$\lambda_\tau, \lambda_b$ and $\eta$ by some average values
$\bar{\lambda}_\tau, \bar{\lambda}_b$ and
$\bar{\eta}$ in the expression (A.6) for $\delta$.
Then,
$$
\eqalignno{
\delta &= -2 \int_0^{ {1\over 16 \pi^2}\log {M_G\over M_S}}
e^{-t(16\bar{\lambda}_\tau^2
+ 6 \bar{\lambda}^2_b + 6\bar{\lambda}_b
\bar{\lambda}_t \bar{\eta} )}\cr
&= - {1\over 8\bar{\lambda}^2_t + 3(\bar{\lambda}^2_b
+\bar{\eta}\bar{\lambda}_b\bar{\lambda}_\tau)}
\left[ e^{- {1\over 16\pi^2} \log {M_G\over M_S}
(16 \bar{\lambda}^2_\tau + 6 \bar{\lambda}^2_b +
6\bar{\eta}\bar{\lambda}_b\bar{\lambda}_\tau)} \right].
&(A.10)
\cr}
$$
So,
$$
|\delta | < {
1\over 8 \bar{\lambda}^2_\tau + 3(\bar{\lambda}^2_b +\bar{\eta}
\bar{\lambda}_b
\bar{\lambda}_\tau).
}\eqno(A.11)
$$
For the $\bar{\lambda}$'s between 0.5 and 1, and $\bar{\eta} \sim 1, |\delta |$
ranges from ${1\over 3}$ to ${1\over 15}$.
How can we qualitatively
understand the above results for the scaling of mixing
matrices?
The renormalization group equations
try to align the soft supersymmetry breaking flavor
matrices with whatever combination of flavor matrices responsible for
their renormalization.
However, because a given coupling
can only be renormalized by harder couplings, there
is a hierarchy in
which flavor matrices affect the running of others.
The Yukawa matrices,
being dimensionless, can only be affected by other Yukawa
matrices.
In the lepton sector, this is the reason that the basis in which e.g.
${\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E$ is diagonal does not change.
Next, the soft trilinear terms, having mass dimension one,
can only be affected
by other trilinear terms and Yukawa couplings.
Again in the lepton sector this means that e.g.
${\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7110 } }_E$
tries to align
itself with ${\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E$.
Finally, the scalar mass, having dimension two,
are affected by everything:
${\bf{m}}^2_L$ tries to align with
${\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E$,
but suffers
interference from ${\fam=9{\mathchar"7110 } }^\dagger_E {\fam=9{\mathchar"7110 } }_E$, unless
${\fam=9{\mathchar"7110 } }^\dagger_E
{\fam=9{\mathchar"7110 } }_E$
is diagonal in the same basis as
${\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E$.
Even if ${\fam=9{\mathchar"7110 } }^\dagger_E{\fam=9{\mathchar"7110 } }_E$ is not diagonal in the same
basis as
${\fam=9{\mathchar"7115 } }^\dagger_E {\fam=9{\mathchar"7115 } }_E$, it is trying to align itself with
${\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E$,
so ${\bf{m}}^2_L$ will still tend to align with
${\fam=9{\mathchar"7115 } }^\dagger_E{\fam=9{\mathchar"7115 } }_E$.
{}From the above discussion, it is clear that the situation
is slightly
complicated in the quark sector.
In the lepton sector, there was a fixed direction in flavor space
given by
${\fam=9{\mathchar"7115 } }_E$, with which the soft matrices aligned.
In the quark sector, we have both ${\fam=9{\mathchar"7115 } }_U$ and ${\fam=9{\mathchar"7115 } }_D$, and
${\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger$, ${\fam=9{\mathchar"7115 } }_D
{\fam=9{\mathchar"7115 } }_D^\dagger$
are misaligned $(V_{KM}\neq 1)$.
This complicates the analysis for $W_{U_L}, W_{D_L}$ so we discuss them last.
Let us now examine the scaling of $W_{U_R}, W_{D_R}$.
(Throughout the following, we assume degeneracy between first two generation
scalar masses,
we neglect all Yukawa coupling matrix
eigenvalues except those of the
third generation, and we do not
include the effect of trilinear soft terms in the scaling.
The last assumption is made for simplicity; we can make similar arguments
about the importance of these neglected trilinear terms as
we did above in the lepton sector.)
First, we show that the basis in which
${\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U$ is diagonal remains
fixed.
The RGE for ${\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U$ is
$$
- {d\over dt}{\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U =
6({\fam=9{\mathchar"7115 } }^\dagger_U
{\fam=9{\mathchar"7115 } }_U)^2
+ 2{\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_D
{\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_U +
2(3Tr{\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U
- {16\over 3} g^2_3 -3 g^2_2 - {13\over 15}
g^2_1) {\fam=9{\mathchar"7115 } }_U^\dagger {\fam=9{\mathchar"7115 } }_U.\eqno(A.12)
$$
Working in a basis where ${\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U$ is diagonal,
let us see if
${d\over dt} {\fam=9{\mathchar"7115 } }^\dagger_U{\fam=9{\mathchar"7115 } }_U$
has off-diagonal components.
We have, (recalling that in this basis
${\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D = V_{KM}\bar{{\fam=9{\mathchar"7115 } }}^2_D
V^\dagger_{KM})$,
$$
\eqalignno{
- {d\over dt} ({\fam=9{\mathchar"7115 } }^\dagger_U
{\fam=9{\mathchar"7115 } }_U)_{\stackrel{ij}{i\neq j}}
&= 2 (\bar{{\fam=9{\mathchar"7115 } }}_UV_{KM}
\bar{{\fam=9{\mathchar"7115 } }}^2_DV^\dagger_{KM}
\bar{{\fam=9{\mathchar"7115 } }}_U)_{ij}\cr
&= 2\bar{{\fam=9{\mathchar"7115 } }}_{U_i} V_{KM_{i\ell}}
\bar{{\fam=9{\mathchar"7115 } }}_{D_{\ell}}^2
V^\dagger_{KM_{\ell j}}
\bar{{\fam=9{\mathchar"7115 } }}_{U_j}\cr
&= 0 \ \hbox{for} \ i, j\neq 3 &(A.13)\cr}
$$
since we neglect all Yukawa's except the third generation.
Similarly, the basis in which ${\fam=9{\mathchar"7115 } }^\dagger_D{\fam=9{\mathchar"7115 } }_D$
is diagonal does not
change.
Thus, the discussion for the scaling of $W_{U_R}, W_{D_R}$
is completely analogous to that in the lepton sector, and we find
$$
W_{U_{R 3i}}W^\dagger_{U_{R 33}}(M_S) = e^{-2I_t} {\Delta m^2_U(M_G)\over
\Delta m^2_U(M_S)} W_{U_{R 3i}}W^\dagger_{U_{R 33}}(M_G),\eqno(A.14)
$$
$$
W_{D_{R 3i}}W^\dagger_{D_{R 33}}(M_S) = e^{-2I_b} {\Delta m^2_D(M_G)\over
\Delta m^2_D(M_S)} W_{D_{R 3i}}W^\dagger_{D_{R 3}}(M_G).\eqno(A.15)
$$
We now turn to $W_{U_L}, W_{E_L}$.
Let $V^*_{U_L}(t)$ be the matrix diagonalizing ${\fam=9{\mathchar"7115 } }_U
{\fam=9{\mathchar"7115 } }^\dagger_U(t)$:
$$
{\fam=9{\mathchar"7115 } }_U {\fam=9{\mathchar"7115 } }^\dagger_U (t)= V^{*}_{U_{L}}(t)
{\bar{{\fam=9{\mathchar"7115 } }}}^2_U (t) V^{*\dagger}_{U_{L}} (t).\eqno(A.16)
$$
In the superfield basis in which
${\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U$ is diagonal,
the squark mass matrix is
$\widetilde{\bf{m}}^{2*}_{Q 3i} = V^\dagger_{U_L}{\bf{m}}^{2*}_Q V_{U_L}$.
Note as before that $\widetilde{\bf{m}}^{2*}_{Q 3i}
= (W^\dagger_{U_L}\overline{\bf{m}}^{2*}_Q
W_{U_L})_{3i}=
W_{U_{L 3i}}W^\dagger_{U_{L 33}}\Delta m^2_Q$,
so we are interested in ${d\over dt} \widetilde{\bf{m}}^{2*}_{Q 3i}$. Now,
$$
\eqalignno{
{d\over dt} \widetilde{\bf{m}}_Q^{2*} = {d\over dt}(V^\dagger_{U_L}
{\bf{m}}^{2*}_QV_{U_L})&=
\left( {d\over dt} V^\dagger_{U_L}\right)
{\bf{m}}^{2*}_Q V_{U_L} + V^\dagger_{U_L} {d\over dt} {\bf{m}}^{2*}_Q V_{U_L}
+ V^\dagger_{U_L}
{\bf{m}}^{2*}_Q {d\over dt} V_{U_L}\cr
&= \left[ \widetilde{\bf{m}}^{2*}_Q, V^\dagger_{U_L} {d\over dt} V_{U_L}\right]
+V^\dagger_{U_L} {d\over dt} {\bf{m}}^{2*}_Q V_{U_L}.&(A.17)\cr}
$$
The second term is the analogue of what we have
already seen in the lepton and right-handed quark sector; using the RGE
for ${\bf{m}}^{2*}_Q$ we find to leading order
$$
\left( V^\dagger_{U_L} {d\over dt} {\bf{m}}^{2*}_Q V_{U_L}\right)_{3i} =-
(\lambda^2_t
+ \lambda^2_b) \widetilde{\bf{m}}^2_{Q 3i}.\eqno(A.18)
$$
Now, $V^\dagger_{U_L} {d\over dt} V_{U_L}$ is obtained
from the RGE for ${\fam=9{\mathchar"7115 } }_U
{\fam=9{\mathchar"7115 } }^\dagger_U$.
Actually, note that
$$
V^\dagger_{U_L} \left( {d\over dt}
{\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U\right)
V_{U_L} =
\left[ V^\dagger_{U_L} {d\over dt} V_{U_L}, \bar{{\fam=9{\mathchar"7115 } }}^2_U\right]
+ {d\over dt}
\bar{{\fam=9{\mathchar"7115 } }}^2_U,\eqno(A.19)
$$
so that only $[V^\dagger_{U_L} {d\over dt} V_{U_L}, \bar{{\fam=9{\mathchar"7115 } }}^2_U]$
is determined.
(This is a reflection of the fact that $V_{U_L}$ is not unique: let $X (t)$
be any unitary transformation leaving $\overline{\bf{m}}^2_Q(t)$ invariant:
$
\overline{\bf{m}}^2_Q(t) = X^\dagger (t) \overline{\bf{m}}^2_Q(t) X(t).
$
In our case, $X(t)$ is most generally a $U(2)$ matrix in the first
two generation subspace.
Then, if $V_{U_L}$ diagonalizes ${\bf{m}}^{2*}_Q$, so does $V_{U_L}X$.
Under this change, $V^\dagger_{U_L} {d\over dt} V_{U_L}$ is not invariant,
but $[ V^\dagger_{U_L} {d\over dt} V_{U_L},
\bar{{\fam=9{\mathchar"7115 } }}^2_U]$ {\it {is}} invariant).
Further, since we neglect first two generations Yukawa eigenvalues,
$[V^\dagger_{U_L} {d\over dt} V_{U_L}, \bar{{\fam=9{\mathchar"7115 } }}^2_U]_{ij}
=0 $ for $ i, j= 1,2$,
and only
$[V^\dagger_{U_L} {d\over dt} V_{U_L}, \bar{{\fam=9{\mathchar"7115 } }}^2_U]_{3i (i3)}
= (\mp) \lambda^2_t V^\dagger_{U_L} {d\over dt} V^\dagger_{U_{L i 3
(3 i)}}$
is determined, and we can choose all other components of
$V^\dagger_{U_L} {d\over dt}
V_{U_L}$ to vanish.
{}From the RGE for ${\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger,
$
$$
\eqalignno{
- {d\over dt} ({\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger) &=
6({\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U)^2
+ 2(3 Tr {\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger - {16\over 3} g^2_3 - 3g^2_2 -
{ 13\over 15}
g^2_1) {\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }_U^\dagger\cr
&+ \{ {\fam=9{\mathchar"7115 } }_U {\fam=9{\mathchar"7115 } }^\dagger_U,
{\fam=9{\mathchar"7115 } }_D {\fam=9{\mathchar"7115 } }^\dagger_D\},&(A.20)\cr}
$$
we find
$$
\eqalignno{
-\left(V^\dagger_{U_L} ({d\over dt}{\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_{U})
V_{U_L}\right)_{3i}
&= \{ \bar{{\fam=9{\mathchar"7115 } }}_U^2, V_{KM} \bar{{\fam=9{\mathchar"7115 } }}^2_D
V^\dagger_{KM}\}_{3i}\cr
&= {\lambda}^2_t{\lambda}^2_bV_{KM 33}V^\dagger_{KM 3i},
&(A.21)\cr}
$$
and thus
$$
\left( V^\dagger_{U_L} {d\over dt} V_{U_{L 3i}}\right) =- \lambda^2_b
V^\dagger_{KM 3i} V_{KM 33}.\eqno(A.22)
$$
Thus to leading order
$$
[V^\dagger_{U_L} {\bf{m}}^2_Q V_{U_L}, V^\dagger_{U_L}
{d\over dt} V_{U_L}]_{3i}
=- \Delta {{m}}^2_Q \,\lambda^2_b \,V^\dagger_{KM 3i} V_{KM 33},\eqno(A.23)
$$
and finally we have
$$
- {d\over dt} (W_{U_{L3i}} W^\dagger_{U_{L 33}} \Delta m_Q^2) =
(\lambda^2_t + \lambda_b^2) W_{U_{L 3i}}W^\dagger_{U_{L 33}}
\Delta {{m}}^2_Q + \lambda^2_b V^\dagger_{KM 3i} V_{KM 33}\Delta
{{m}}^2_Q.\eqno(A.24)
$$
Similarly we find
$$
- {d\over dt} (W_{D_{L 3i}} W^\dagger_{D_{L 33}}
\Delta {{m}}^2_Q) = (\lambda^2_t +
\lambda^2_b) W_{D_{L 3i}} W^\dagger_{D_{L 33}}
\Delta {{m}}^2_Q
+ \lambda^2_t V^\dagger_{KM 3i}V_{KM 33}\Delta {{m}}^2_{Q}.\eqno(A.25)
$$
We can formally solve the above equations, e.g.
$$
W_{U_{L 3i}} W^\dagger_{U_{L 33}} (M_S) =
e^{-\left( I_t + I_b + \int^{M_G}_{M_S} dt'
\lambda^2_b {V^\dagger_{KM 3i}
V_{KM_{ 33}}\over W_{U_{L 3i}}W^\dagger_{U_{L 33}}}
\right)}
{\Delta {{m}}^2_Q (M_G)\over \Delta {{m}}^2_Q(M_S)}
W_{U_{L 3i}} W^\dagger_{U_{L 33}}
(M_G),\eqno(A.26)
$$
and, to a good approximation, given that $W_{U_{L 3i}}$ does not
scale very significantly, we can replace
$$
\int^{M_G}_{M_S} dt' \lambda^2_b {V^{\dagger}_{KM 3i}
V_{KM 33}\over W_{U_{L 3i}}W^\dagger_{U_{L 33}} }
\approx I_b {V^\dagger_{KM 3i} V_{KM 33}\over W_{U_{L 3i}}
W^\dagger_{U_{L 33}} } (M_G).\eqno(A.27)
$$
So, an approximate solution of the RGE for $W_{U_L}, W_{D_L}$ is
$$
W_{U_{L 3i}}
W^\dagger_{U_{L 33}} (M_S) \approx
e^{ -\left( I_t + I_b \left( 1+
{ V^\dagger_{KM 3i} V_{KM_{ 33}}\over W_{U_{L 3i}}
W^\dagger_{U_{L 33}}} (M_G)\right) \right)
}
{\Delta {{m}}^2_Q (M_G)\over \Delta {{m}}^2_Q(M_G)}
W_{D_{L 3i}} W^\dagger_{U_{L 33}}
(M_G),\eqno(A.28)
$$
and similarly
$$
W_{D_{L 3i}} W^\dagger_{D_{L 33}}(M_S) \approx
e^{ -\left( I_b + I_t \left( 1+
{ V^\dagger_{KM 3i} V_{KM 33}
\over W_{U_{L 3i}} W^\dagger_{U_{L 33}}}
(M_G) \right) \right)
}
{\Delta {{m}}^2_Q(M_G)\over \Delta {{m}}^2_Q(M_S)}
W_{D_{L 3i}} W^\dagger_{D_{L 33}}
(M_G).\eqno(A.29)
$$
The above results are in agreement with qualitative expectations;
the extra terms in the exponential of (A.28) and (A.29) are a
reflection of the fact that the bases in which
${\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U$ and
${\fam=9{\mathchar"7115 } }_D{\fam=9{\mathchar"7115 } }^\dagger_D$ are diagonal change
with scale. For moderate $\tan\beta$, however, we expect
that the basis in which ${\fam=9{\mathchar"7115 } }_U{\fam=9{\mathchar"7115 } }^\dagger_U$
is diagonal should not change with scale, and in this
limit the extra term drops out of (A.28).
\newpage
|
\section{Experimental Methods}
The films used in this study were grown using ion-beam sputtering
(IBS) in a chamber with a base pressure of 2$\times$10$^{-8}$ torr.
The deposition system is described in more detail
elsewhere.\cite{fesiprb} All samples used in this study were grown at
a substrate temperature of 200$^{\circ}$. All comparisons between
films grown on glass and single-crystal substrates will be made on
samples which were deposited {\em simultaneously} so as to eliminate
any reproducibility issues.
The substrates used in this study were glass coverslips, MgO (001),
Ge(001) and Al$_2$O$_3$(11\={2}0). The MgO and Al$_2$O$_3$ substrates
were cleaned according to a recipe reported by Farrow and
coworkers.\cite{farrow} The glass and Ge substrates were rinsed in
solvents. All films are capped with a 200\AA\ Ge oxidation barrier.
The magnetic and structural properties of the films are stable for at
least one year.
\section{Structural Characterization}
Figure~\ref{epihigh} shows high-angle x-ray diffraction spectra for a
purely (001)-oriented (Fe40\AA/Si14\AA)x60 multilayer grown on
MgO(001) and a purely (011)-oriented multilayer grown simultaneously
on glass. While multiple superlattice satellites are observed around
the Fe(002) peak for the film grown on MgO, the film on glass has only
one peak corresponding to (011)-textured growth. A single
superlattice satellite is typically observed in multilayers on glass
on the low-angle side in keeping with previous
observations.\cite{foiles2} An estimate of the crystallite sizes in
these films can be derived using the Scherrer formula. This analysis
gives a coherence length of 165\AA\ for the film on glass and 188\AA\
for the film on MgO.
Since the crystalline coherence lengths of these films are similar,
the presence of high-angle satellites in the film on MgO must be due
to better layering. The small-angle x-ray scattering data shown in
Figure~\ref{epilow} confirm this hypothesis. The multilayer on MgO
has 4 low-angle peaks, indicating a moderate degree of composition
modulation. The multilayer on glass shows only two relatively broad
peaks, indicating larger interfacial roughness and less order in the
layering. The low-angle x-ray spectra are consistent with rocking
curves which are only about 1$^{\circ}$ wide for films grown on MgO
and $\rm Al_2O_3$ but are typically 10$^{\circ}$ to 15$^{\circ}$ wide
for films on glass. The bilayer periods determined from the low-angle
peak positions are (41.0 $\pm$ 0.1)\AA\ for the MgO film and (40.9
$\pm$ 0.2)\AA\ for the glass film, the same within experimental error.
$\phi$ scans of the MgO and Fe [110] peaks for the film on the MgO
substrate (not shown) demonstrate that it is oriented in-plane. While
$\theta$-2$\theta$ scans for (011)-oriented multilayers grown on $\rm
Al_2O_3$ substrates (not shown) also show multiple high-angle
superlattice satellites, the $\phi$ scans for this film indicate only
weak orientation in-plane.
The shape of the high-angle peaks and their superlattice satellites
are described by a well-known theory.\cite{fullerton3} Application of
this theory to the Fe/Si multilayers is difficult because the iron
silicide lattice constant, the thickness of the remaining pure Fe and
the thickness of the iron silicide spacer can be estimated only
roughly. A precise determination of the silicide lattice constant
should make a quantitative analysis of these satellite features
possible.
\section{Magnetic Characterization}
Figure~\ref{bvsh}a shows magnetization curves for 60-repeat
(Fe40\AA/Si14\AA) multilayers grown simultaneously on glass and
MgO(001). The saturation fields H$\rm _s$ appear to be similar for
the two films. On the other hand, the remanent magnetization is 58\%
for the film on glass and 7\% for the film grown on MgO. A remanence
as low as 1\% has been observed for other multilayers grown on MgO
substrates. SQUID magnetometer data taken up to higher fields gives a
saturation field of 9.75 kOe for the multilayer on MgO at room
temperature. Assuming for a moment that the interlayer coupling is
purely bilinear in nature, a well-known formula relates the saturation
field to the AF coupling strength: A$_{12} \, =
\, H_s M_s t_{Fe} / 4$ where M$\rm _s$ is the saturation magnetization
and t$\rm _{Fe}$ is the thickness of an individual Fe
layer.\cite{krebs} Use of this equation with H$\rm _s$ = 9.75 kOe and
the measured magnetization M$_s$ = 1271 emu/cm$^3$ gives A$\rm _{12}$
= 1.2 erg/cm$^2$. This AF coupling value is comparable in size to the
coupling measured in metal/metal multilayers multilayers\cite{parkin}.
Figure~\ref{bvsh}b shows magnetization curves for
Fe100\AA/Si14\AA/Fe100\AA trilayer films grown on Al$_2$O$_3$
(11\={2}0), Ge(001), and MgO(001) substrates. All three of the
magnetization curves in Fig.~\ref{bvsh}b were taken with the field
applied along an Fe(100) easy direction. Significant in-plane
anisotropy of the magnetization curves occurs for the films on the Ge
and MgO substrates, similar to what has been observed for Fe/Cr/Fe
trilayers.\cite{krebs} The observation of magnetocrystalline anisotropy
in the film on MgO but not in the film grown on Al$_2$O$_3$ is
consistent with expectations from the $\phi$ scans, which show that
the in-plane orientation of the film on MgO is much stronger.
Figure~\ref{bvsh}b once again demonstrates that the degree of
remanence in Fe/Si multilayers is strongly related to the quality of
layering. While the remanent magnetization of the epitaxial trilayers
on Ge and MgO is only about 5\% of the saturated value, the remanence
of the polycrystalline trilayer on Al$_2$O$_3$ is close to 50\%. The
remanent magnetization of the trilayers on Ge and MgO is about 5\% in
the in-plane hard direction (H $\|$ Fe(110)) as well.
A SQUID magnetometer has been used to measure the magnetization curves
of the IBS-grown Fe/Si multilayers at lower temperatures.\cite{michel} The
temperature dependence of the remanent magnetization of these films is
similar to that reported by other authors.\cite{mattson}
\section{Discussion}
At the moment it is not possible to tell why the in-plane
ordering of the films grown on Al$_2$O$_3$(11\={2}0) is inferior to
that grown on the (001) MgO and Ge substrates. The difficulty with
the Al$_2$O$_3$ growth may have to do with the 6$^{\circ}$ miscut of
the substrates, or it may be due to an intrinsic difficulty with (011)
growth of the Fe/Si multilayers. Previous work has shown that AF
coupling in Fe/Si multilayers is dependent upon formation of a
metastable iron silicide spacer layer phase.\cite{fesiprb} The
possibility exists that the spacer silicide does not grow well on Fe
in the (011) orientation. This question can be answered only by
further growth studies on better (011) substrates and careful
structural characterizations.
A related question is whether the larger remanent magnet moment in the
(011)-textured films might be due to a fundamental difference in
magnetic properties from the (001)-textured films. Because a
46-repeat Fe/Si multilayer grown on Al$_2$O$_3$ has a remanence of
only about 10\%, this is unlikely. Undoubtedly the trilayer on
Al$_2$O$_3$ has a higher remanence than the multilayer because the
thinner film is more greatly impacted by the poor substrate surface
quality. The staircase morphology caused by the 6$^{\circ}$ miscut of
this Al$_2$O$_3$ substrate may lead to wavy interfaces between the Fe
and iron silicide films or to pinholes through the silicide layers.
Wavy interfaces can cause increased magnetostatic coupling or even
biquadratic coupling,\cite{slonczewski} both of which would tend to
increase the remanence. The large remanence of the multilayers grown
on glass substrates is likely also due to pinholes or magnetostatic
coupling.
Pinhole-induced may explain the unusual temperature dependence of the
remanence. (Magnetostatic coupling is expected to be approximately
temperature-independent.) Fe atoms in bridges through the silicide
spacer layers are expected have a reduced Curie temperature. A larger
remanence at low temperature therefore makes sense if the remanence is
derived from pinhole coupling and is not an intrinsic effect. Low
Curie-temperature material may also be present in the iron silicide
spacer layer or in at the iron/iron silicide interfaces.
By growing on a number of substrate materials and by using different
deposition conditions, Fe/Si multilayers have been prepared with a
varying degree of ordering. A large amount of accumulated evidence
demonstrates that high remanence of the magnetization curves in Fe/Si
multilayers is associated with interface roughness. The remanence is
therefore not likely to be related to unusual exchange coupling but
instead to originate from defects, perhaps pinholes through the
silicide spacer layer. Since the remanent magnetization is caused by
extrinsic effects, future studies should concentrate instead on
measurements of the saturation field of the magnetization curves in
order to learn more about the interlayer coupling.
\bigskip
\noindent We would like to thank E.E. Fullerton, J.A. Borchers,
R.M. Osgood III and Y. Huai for helpful discussions, and B.H. O'Dell and
S. Torres for technical assistance. Part of this work was performed
under the auspices of the U.S. Department of Energy by LLNL under
contract No. W-7405-ENG-48.
\clearpage
|
\section{ \normalsize} \def\ds{\displaystyle\bf INTRODUCTION. }
\hspace*{6mm} Among different representations of a given compact Lie group $G$
the model space ${\cal M}$ plays a distinguished role.
By definition,\REF{1} the model space is a direct sum of all
irreducible representations ${\cal H}_j$ with multiplicity one
\begin{equation}\label} \def\eeq{\end{equation}{0.0}
{\cal M} = \sum\limits_j \oplus{\cal H}_j
\eeq
realized in some universal way. A most popular form of
${\cal M}$ is a space of holomorphic functions on the Borel subgroup
$\cal B$ of complexified form of the group $G$. In this construction the
Borel subgroup is considered as an affine space.
A study of model spaces provides a natural language for investigation of
physical models. For example, the popular model of $2$-dimensional quantum
gravity, introduced by Polyakov,\REF{2} may be interpreted in terms of
the model space of Virasoro algebra.\REF{3}
A finite-dimensional quantum group with deformation parameter, depending
on the central charge, naturally appears in this context.
In the present paper, which was written with an intent to find new
applications of model space in modern mathematical physics, we discuss a
$q$-analogue of the model space related to $q$-deformed Lie group $G_q$.
For this purpose we introduce and examine several "coordinatizations" of
the quantum space $(T^*\B)_q$. As a by-product we obtain some generating
matrices for the set of Clebsch-Gordan coefficients (CGC). To our
knowledge this result is new even for the non-deformed case.
Throughout the paper we systematically and intentionally make use of the
$R$-matrix formalism, which we believe is the most convenient and powerful
tool to get explicit results in the domain of quantum groups.
To avoid the known difficulties with compact forms of quantum groups we
adopt here a convention to work with complexified objects (groups,
algebras) and their finite dimensional representations on a formal
algebraic level. We also do not discuss subtleties arising in the
case of $q$ being a root of unity.
Most of formulae given in this paper in $R$-matrix form have universal
validity. However, the concrete results are illustrated on the simplest
example $G_q=SL_q(2)$. The generalization to other groups needs more
technical details such as an explicit structure of $R$-matrices and
related objects.
Mentioned above "coordinatizations" of $(T^*\B)_q$ arise from two possible
decompositions of the matrix $L$ (in usual notations $L=L_+L_-^{-1}$, it
comprises all generators of the corresponding quantum Lie algebra) :
$$ L = U\,D\,U^{-1} \;\;\; {\rm and} \;\;\; L = A\,B\,A^{-1} \ , $$
where $D$ is a diagonal unimodular matrix, $U$ is a deformation of unitary
matrix, $A$ and $B$ are unimodular upper and lower triangular matrices.
As we shall clarify below, the matrices $A$ and $B$ admit a natural
interpretation as the coordinates in the base and in the fiber of
$(T^*\B)_q$, whereas entries of the matrix $U$ will be shown to
provide basic shifts on
the model space $\cal M$ and generate $q$-analogues of Clebsch-Gordan
coefficients for the quantum group $G_q$. The explicit connection between
$U$ and $(A,B)$ will be demonstrated on the example of $SL_q(2)$.
It should be mentioned that an object like the matrix $U$ appeared first
in Refs.4,5 (later it was used also in Ref.6), where it was interpreted
as a "chiral" component of the quantum group-like element $g$.
In the present paper we give another interpretation and application of
the matrix $U$ in the context of a model space.
Let us briefly describe the contents of the present paper. In the Sec.$\!$
II the definition of the cotangent bundle for a quantum group is reminded.
Next we introduce an object of especial interest for us -- the algebra
$\cal U$ generated by the entries of the matrix $U$ which diagonalizes the
coordinate in a fiber of $(T^*G)_q$. We derive explicit relations for this
algebra in the case of $G=SL(2)$.
In the Sec.$\!$ III we consider a non-deformed limit ($q=1$) of the algebra
$\cal U$ and construct an explicit representation.
For the case of $G=SL(2)$ we show that the matrix
$U_0$ generates Clebsch-Gordan coefficients (CGC) for the corresponding
non-deformed Lie algebra. The Borel subgroup $\cal B$ and the space $T^*\B$
naturally appear here. Finally, we discuss a connection of our results
with the Wigner-Eckart theorem.
In the Sec.$\!$ IV we construct representations of the algebra
$\cal U$ (for $q\neq 1$) for the case of $SL(2)$ in two different ways.
The first one uses the language of $q$-oscillators.
The second is based on explicit realization
of $(T^*\B)_q$ and hence involves a notion of quantum model space.
Here we show that the matrix $U$ is a "generating matrix" for CGC for
deformed Lie algebra. We also give some comments on the generalized
version of the Wigner-Eckart theorem.
\section{\normalsize} \def\ds{\displaystyle\bf $\bf (T^*G)_q$ AND RELATED OBJECTS.}
\setcounter{equation}{0}
\hspace*{6mm} There exist three symplectic manifolds (from the physical point
of view they are phase spaces) naturally related to a given Lie group
$G$ and its Lie algebra ${\cal J}$:
$\langle 1 \rangle$\ \ $T^*G$ -- the cotangent bundle for the group $G$;
$\langle 2 \rangle$\ \ $T^*\B$ --
the cotangent bundle for the Borel subgroup $\B$;
$\langle 3 \rangle$\ \ $\cal O$ --
an orbit of the co-adjoint action of $G$ on ${\cal J}^*$.
For instance, in the case of $G=SL(2)$ (which will be our main example)
these spaces are six-, four-, and two-dimensional, correspondingly.
The method of geometric quantization\REF{7} provides a
representation theory for $\langle 1 \rangle$, $\langle 2 \rangle$ and
$\langle 3 \rangle$. Turning from classical to quantum
groups, one can try to construct a representation theory for the deformed
analogues of these manifolds. In the present paper we shall deal with
deformations of the spaces $\langle 1 \rangle$ and $\langle 2 \rangle$.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf A. Description of $\bf (T^*G)_q$. }
\hspace*{6mm} Let $G_q$ be a deformation of the Lie group $G$ and ${\cal J}_q$ be
a deformation of the corresponding Lie algebra ${\cal J}$.
The deformed cotangent bundle $(T^*G)_q$ is a non-commutative manifold,
i.e., according to the ideology developed by A.Connes,\REF{8} its
coordinates are (non-commuting) generators of some associative algebra.
A point on this manifold is parameterized by the pair $(g,L)$, where
$g\in G_q$ is a coordinate in the base of the bundle, and $L$ is a
coordinate in a fiber.
The structure of $(T^*G)_q$ is defined via commutation relations between the
coordinates in the base and in a fiber. An appropriate $R$-matrix form of
these relations was proposed in Ref.5:
\begin{equation}\label} \def\eeq{\end{equation}{1.1}
R_{\pm} \up{1}{g} \,\up{2}{g} \,=\, \up{2}{g} \, \up{1}{g} R_{\pm} \,,
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{1.2}
R_- \up{1}{g} \up{2}{L} \,=\, \up{2}{L} R_+
\up{1}{g} \,,
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{1.3}
\up{1}{L} R_-^{-1} \up{2}{L} R_- \,=\, R_+^{-1}
\up{2}{L} R_+ \up{1}{L} \,.
\eeq
Here and below we use the formalism developed in Ref.9, i.e., objects
like $g$ and $L$ are considered as matrices (say, $L\in {\cal J}} \def\Uc{{\cal U}_q\otimes V$,
where $V$ stands for auxiliary space). We use the standard notations
for tensor products:
$\up{1}{L}\up{}{=L\otimes I}\in{\cal J}} \def\Uc{{\cal U}_q\otimes V\otimes V$, etc.
Let us take the parameter $q$, which appears in the theory of
quantum groups, in the following form
\begin{equation}\label} \def\eeq{\end{equation}{0}
q=e^{\gamma\hbar} ,
\eeq
where $\hbar$ is the Planck constant (the parameter of quantization)
and $\ga$
is the deformation parameter. In physical applications it is most natural
to suppose that $\ga$ is either pure real ($q$ belongs to the real axis)
or pure imaginary ($q$ belongs to the unit circle at the complex plane).
The second form of $q$ is typical for the WZW theory.\REF{4,6,10}
{}For $|q|=1$ we suppose also that $q$ is not a root of unity.
It should be mentioned that for both variants of choice of $\ga$
in (\ref{0}) the definition of $q$-number
\begin{equation}\label} \def\eeq{\end{equation}{0.2}
[x] \equiv \frac{q^x - q^{-x}}{q-q^{-1}}
\eeq
is invariant with respect to complex conjugation of $q$, i.e.
$\overline{[x]}=[\overline{x}]$. This property
becomes important if one discusses involutions of deformed Lie algebras.
\begin{defn}\label{DL}
The algebra $\cal L$ is an associative algebra generated by entries of
the matrix $L$ which obeys relation (\ref{1.3}).
\end{defn}
An important fact -- the connection of algebra $\cal L$ with the
corresponding quantum Lie algebra ${\cal J}} \def\Uc{{\cal U}_q$ was established in Ref.11
in the following form.
\begin{prop}\label{P4}
Let matrices $L_+$, $L_-$ obey the following exchange relations
\begin{equation}\label} \def\eeq{\end{equation}{L2}
R_{\pm} \up{1}{L_+} \up{2}{L_+} = \up{2}{L_+} \up{1}{L_+} R_{\pm} ,\;\;
R_{\pm} \up{1}{L_-} \up{2}{L_-} = \up{2}{L_-} \up{1}{L_-} R_{\pm}, \;\;
R_{+} \up{1}{L_+} \up{2}{L_-} = \up{2}{L_-} \up{1}{L_+} R_{+} \;\; .
\eeq
Then the matrix $ L= L_+ \, L_{-}^{-1}$ satisfies the relation (\ref{1.3}).
\end{prop}
This statement implies that the algebra $\cal L$ is isomorphic
(up to some technical details which we do not discuss here) to
corresponding quantum Lie algebra $U_q({\cal J}} \def\Uc{{\cal U})$ [which is defined by (\ref{L2}),
see, e.g., Ref.9].
Consider now the relations (\ref{1.1})-(\ref{1.3})
for $g$ and $L$ being $2\times 2$ matrices
\begin{equation}\label} \def\eeq{\end{equation}{g,L}
g=\left(\begin{array}} \def\er{\end{array}{cc} g_1&g_2\\g_3&g_4\er\right),\;\;\;\;
L=\left(\begin{array}} \def\er{\end{array}{cc}A&B\\C&D\er\right)
\eeq
and the $R$-matrices taken in the form
\begin{equation}\label} \def\eeq{\end{equation}{1.4}
R_{+}=q^{-1/2}\left(\begin{array}} \def\er{\end{array}{cccc}
q & 0 & 0 & 0 \\0 & 1 & \omega & 0 \\0 & 0 & 1 & 0 \\0 & 0 & 0 & q \er
\right) ,\; \omega \equiv q-q^{-1}; \;\;\;\;\;\; R_{-}=PR_{+}^{-1}P
\eeq
($P$ denotes the permutation operator: $P\!\up{1}{g}\!P\!=\,\up{2}{g}$,
etc.). In this case (\ref{1.1})-(\ref{1.3}) define the cotangent bundle for
the quantum group $G_q=GL_q(2)$; each of $R$-matrix equations (\ref{1.1})
and (\ref{1.3}) is equivalent to six independent relations:
\begin{equation}\label} \def\eeq{\end{equation}{1.35}
\begin{array}} \def\er{\end{array}{c} q\,g_1\,g_2= g_2\,g_1,\;\; q\,g_1\,g_3= g_3\,g_1,\;\;
q\,g_2\,g_4= g_4\,g_2,\;\; q\,g_3\,g_4= g_4\,g_3,\;\; \\ \\
g_2\,g_3= g_3\,g_2,\;\;\;\; g_1\,g_4 - q^{-1}\, g_4\,g_1 =
-\omega \,g_2\,g_3 ; \er
\eeq
and
\begin{equation}\label} \def\eeq{\end{equation}{1.5}
\begin{array}} \def\er{\end{array}{c} [A,B]=-q^{-1}\omega BD,\;\;\; [A,C]=q^{-1}\omega DC,\;\;\;[A,D]=0, \\
\\ CD=q^2DC,\;\;\;BD=q^{-2}DB,\;\;\;[B,C]=q^{-1}\omega D (D-A). \er
\eeq
The equation (\ref{1.2}) gives the following relations
\begin{equation}\label} \def\eeq{\end{equation}{gL}
\begin{array}} \def\er{\end{array}{ll} g_1 A = q A g_1 + \omega B g_3,\; & g_1 B = B g_1, \\ [1mm]
g_2 A = q A g_2 + \omega B g_4,\; & g_2 B = B g_2, \\ [1mm]
g_3 A = q^{-1} A g_3 + \omega g_1 C,\; & g_3 B = B g_3
+\omega g_1 D, \\ [1mm]
g_4 A = q^{-1} A g_4 + \omega g_2 C,\; & g_4 B = B g_4
+\omega g_2 D, \\ [1mm]
g_1 C = C g_1 + q^{-1}\omega D g_3,\; & g_1 D = q^{-1}\, D g_1, \\ [1mm]
g_2 C = C g_2 + q^{-1}\omega D g_4,\; & g_2 D = q^{-1}\, D g_2, \\ [1mm]
g_3 C = C g_3, & g_3 D = q D g_3, \\ [1mm]
g_4 C = C g_4, & g_4 D = q D g_4. \er
\eeq
Next, let us recall the well-known statement (see, e.g., Ref.9):
\begin{prop}\label{P1}
The algebra generated by the entries of the matrix $g$ obeying (\ref{1.35})
possesses the central element (``deformed determinant'')
\begin{equation}\label} \def\eeq{\end{equation}{1.65}
{\det}_q g = g_1\,g_4-q^{-1}g_2\,g_3 .
\eeq
\end{prop}
Similarly, for the algebra $\cal L$ in the case of $GL_q(2)$ one can
check the following.
\begin{prop}\label{P2}
The algebra with generators $A$, $B$, $C$, $D$ obeying (\ref{1.5}) possesses
two central elements:
\begin{equation}\label} \def\eeq{\end{equation}{1.6}
K_1 = qA + q^{-1} D, \;\;\; K_2 = q^{-1} AD - q BC.
\eeq
\end{prop}
Finally, using the commutation relations (\ref{gL}), one can check that
\begin{prop}\label{P3}
The operators ${\det}_q g$ and $K_2$ commute with all entries of the
matrices $g$ and $L$.
\end{prop}
This implies that, fixing values of ${\det}_q g$ and $K_2$, one gets a
certain subalgebra of the algebra defined by (\ref{1.35})-(\ref{gL}).
\begin{defn}\label{D1}
Relations (\ref{1.35})-(\ref{gL}) for
${\det}_q g = 1$ and $ K_2 = {\rm const}$
define the cotangent bundle for the quantum group $G_q=SL_q(2)$.
\end{defn}
Let us underline that the above definitions and statements can be easily
generalized, say to the case of $SL_q(N)$.
In our case the algebra $\cal L$ is isomorphic to the quantum Lie algebra
${\cal J}_q=U_q(sl(2))$ (introduced first in Ref.12) which is defined by
the relations
\begin{equation}\label} \def\eeq{\end{equation}{L3}
[l_+ , l_- ] = \frac{q^{2l_3}-q^{-2l_3}}{q-q^{-1}}\equiv [2l_3],\;\;\;\;
q^{l_3} \, l_{\pm} = q^{\pm 1} l_{\pm} \, q^{l_3}
\eeq
and the matrices $L_{\pm}$ can be chosen as follows:
\begin{equation}\label} \def\eeq{\end{equation}{L4}
L_+ = \left( \begin{array}} \def\er{\end{array}{cc} q^{l_3} & \omega q^{1/2} l_- \\ 0 &
q^{-l_3} \er \right),\;\;\;\;
L_- = \left( \begin{array}} \def\er{\end{array}{cc} q^{-l_3} & 0 \\
-\omega q^{-1/2} l_+ & q^{l_3} \er \right) .
\eeq
Note that the matrix $L$ in the Proposition \ref{P4} is defined only up to
a scaling factor. Thus, for $L_+$, $L_-$ given in (\ref{L4}), we may choose
$L$ as follows
\begin{equation}\label} \def\eeq{\end{equation}{L5}
L= q^2 L_+ L_{-}^{-1} =
\left(\begin{array}} \def\er{\end{array}{cc} q C - q^{-2 l_3} & q^{5/2} \omega l_-
q^{-l_3} \\ q^{-1/2}\omega l_+ q^{-l_3} & q^2 q^{-2l_3} \er \right).
\eeq
Here $C$ stands for the Casimir operator of $U_q(sl(2))$ :
\begin{equation}\label} \def\eeq{\end{equation}{L8}
C = {\omega}^2 l_- l_+ + q^{2l_3+1} + q^{-(2l_3+1)} = q^{2\wh{j}+1} +
q^{-(2\wh{j}+1)} ,
\eeq
where $\wh{j}$ is the operator of spin.
According to Proposition \ref{P4}, the matrix (\ref{L5}) satisfies
(\ref{1.3}). Therefore, it provides a (fundamental) representation of the
algebra ${\cal L}$ for $U_q(sl(2))$. In this representation the central
elements (\ref{1.6}) are given by
\begin{equation}\label} \def\eeq{\end{equation}{L6}
K_1 = q^2 C, \;\;\; K_2 = q^3,
\eeq
Note that the scaling factor $q^{2}$ introduced in (\ref{L5}) has changed
the values of $K_1$ and $K_2$. The choice of such normalization in (\ref{L5})
will be explained later.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf B. Connection with quantum $\bf 6j$-symbols. }
\hspace*{6mm} Let us remind the theorem which describes an important property of
the algebra ${\cal L}$ for $U_q(sl(2))$ (this statement
appeared first in Ref.5).
\begin{theorem}
Let $D\equiv D(p)$ be the unimodular diagonal matrix
\begin{equation}\label} \def\eeq{\end{equation}{1.75}
D = \left( \begin{array}} \def\er{\end{array}{cc}
q^{p/\hbar} & \\ & q^{-p/\hbar} \er \right) ,
\eeq
and let 2$\times$2 matrix $U$ satisfy the following exchange relations
\begin{equation}\label} \def\eeq{\end{equation}{1.8}
\up{1}{D}\, \up{2}{U} = \up{2}{U} \up{1}{D} \sigma,\;\;
\; \up{2}{D}\, \up{1}{U} = \up{1}{U} \up{2}{D}
\sigma, \;\;\; \sigma= {\rm diag} (q^{-1},q,q,q^{-1}),
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{1.9}
R_+ \up{1}{U} \, \up{2}{U} = \up{2}{U} \,\up{1}{U} {\cal R}_+ (p), \;\;
R_- \up{1}{U}\, \up{2}{U} = \up{2}{U}\, \up{1}{U} {\cal R}_- (p),
\eeq
where $R_{\pm}$ are the standard $R$-matrices (\ref{1.4}) and
\begin{equation}\label} \def\eeq{\end{equation}{1.10}
{\cal R}_+ (p)= P{\cal R}_-^{-1} (p) P = q^{-1/2} \left( \begin{array}} \def\er{\end{array}{cccc}
q & & & \\
& \ds \frac{\sqrt{ [p/\hbar +1] [p/\hbar -1] }}{[p/\hbar]} & \ds
\frac{q^{p/\hbar}}{[p/\hbar]} & \\ & & \\
& \ds -\frac{q^{-p/\hbar}}{[p/\hbar]} & \ds
\frac{\sqrt{ [p/\hbar +1] [p/\hbar -1] }}{[p/\hbar]} & \\
& & & q \er \right),
\eeq
(here $[x]$ denotes a "q-number" (\ref{0.2})).
Then matrix $L$ constructed by means of the similarity transformation
\begin{equation}\label} \def\eeq{\end{equation}{1.7}
L = U D\, U^{-1}, \;\;\;
\eeq
satisfies the relation (\ref{1.3}) and therefore its entries generate an
algebra ${\cal L}$ isomorphic to $U_q(sl(2))$.
\end{theorem}
The proof is given in appendix A. It makes use of the identity
\begin{equation}\label} \def\eeq{\end{equation}{RD}
{\cal R}_- (p) = (\up{1}{D})^{-1} {\cal R}_+ (p) \sigma \up{1}{D} .
\eeq
{\it Remark:} \ } \def\proof{{\it Proof}\ A consequence of (\ref{1.8}) is the commutativity of $L$ and $D$
\begin{equation}\label} \def\eeq{\end{equation}{1.12}
\up{1}{L} \, \up{2}{D} = \up{2}{D} \, \up{1}{L},
\eeq
which implies that $p$ commutes with all elements of $\cal L$. Later
we shall interpret $p$ as the operator of spin.
{\it Remark:} \ } \def\proof{{\it Proof}\ Properly generalizing objects which enter the Theorem 1, one
can extend this theorem to the case of any quantum semisimple Lie
algebra.\REF{13}
In particular, the matrix $D$ for $U_q(sl(N))$ is found to be :
$D(\vec{p}\,)\,=\, const\cdot q^{\vec{H}\,\o\,\ds\vec{p} }$,\
where $\vec{p}$ consists of the operators corresponding to components
of the weight vector (i.e., on each irreducible representation they are
multiples of unity)
and $\vec{H}$ consists of the generators $H_i$ of the Cartan subalgebra.
An explicit form of ${\cal R}(p)$ for $U_q(sl(N))$ was obtained in Ref.14.
{\it Remark:} \ } \def\proof{{\it Proof}\ The matrix ${\cal R}(p)$ obeys the deformed Yang-Baxter
equation,\REF{14-16,5} which can be written, for example, as follows :
\begin{equation}\label} \def\eeq{\end{equation}{1.13}
\up{1}{Q} \up{23}{\cal R}_+ (p) (\up{1}{Q})^{-1} \up{13}{\cal R}_+ (p)
\up{3}{Q} \up{12}{\cal R}_+ (p) (\up{3}{Q})^{-1} =
\up{12}{\cal R}_+ (p) \up{2}{Q} \up{13}{\cal R}_+ (p) (\up{2}{Q})^{-1}
\up{23}{\cal R}_+ (p),
\eeq
where for ${\cal J}} \def\Uc{{\cal U}_q=U_q(sl(2))$ the matrix
$Q = \biggl(\! \begin{array}} \def\er{\end{array}{cc} e^{i\xi} & \\ &\! e^{-i\xi} \er\!\biggr)$
contains an extra variable $\xi$, conjugated with $p$:
\begin{equation}\label} \def\eeq{\end{equation}{1.15}
[p,\xi] = -i \hbar, \;\;\;\;\; q^{p/\hbar}\,e^{i\xi} =
q\,e^{i\xi}\,q^{p/\hbar}.
\eeq
This variable $\xi$ belongs to the algebra $\cal U$ but does not enter matrix
$L$. An explicit expression for $\xi$ will be given below. The general form
of $Q$ for $U_q(sl(N))$ can be easily found\REF{13} :
$ Q\,=\,e^{ i\,\vec{H}\,\o\,\ds\vec{\xi} }$, where components of $\xi_i$
are operators conjugated to $p_i$ : $[p_j,\xi_k]=-i\hbar\,\dl_{jk}$.
The matrix ${\cal R}(p)$ was discussed in physical literature in different
contexts. In particular, it plays significant role in studies of quantum
Liouville\REF{15,16} and WZW\REF{4-6} models; its relation to
Calogero-Moser model was recently discussed in Ref.17.
But for us more important fact is a connection of ${\cal R}(p)$ with the
quantum $6j$-symbols:
the entries of (\ref{1.10}) calculated on irreducible representations
coincide (up to some normalization) with the values of some $6j$-symbols
for $U_q(sl(2))$ (exact formulae are given in Ref.18, generalizations
are discussed in Ref.13). This connection
allows to assume that objects like the matrix $U$ should be interpreted in
terms of Clebsch-Gordan coefficients (CGC). Below we demonstrate that $U$
is indeed a ``generating matrix'' for CGC and clarify
its relation to $(T^*\B)_q$.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf C. Algebra ${\cal U}$. }
\begin{defn}\label{DU}
The algebra $\cal U$ is an associative algebra generated
by entries of matrix
$U = \left( \begin{array}} \def\er{\end{array}{cc} U_1 & U_2 \\ U_3 & U_4 \er\right)$ and the operator
$p$ such that relations (\ref{1.75})-(\ref{1.10}) hold.
\end{defn}
{\it Remark:} \ } \def\proof{{\it Proof}\ For simplicity we restricted our consideration to the case of
$\cal U$ associated with $U_q(sl(2))$. Let us stress that the case
of $\cal U$ associated with $U_q(sl(N))$ can be studied
similarly but it will involve more technical details.
On the other hand, it might be rather cumbrous to obtain
exact formulae for $\cal U$ associated with $U_q({\cal J}} \def\Uc{{\cal U})$ in the case
of ${\cal J}} \def\Uc{{\cal U}$ being generic semisimple Lie algebra.
Let us give an explicit form of the defining relations (\ref{1.9}) :
\begin{equation}\label} \def\eeq{\end{equation}{2.1}
U_1 U_3 = q^{-1} U_3 U_1, \;\;\;\;\; U_2 U_4 = q^{-1} U_4 U_2,
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.2}
U_1 U_2 = U_2 U_1 \sqrt{\frac{[p/\hbar -1]}{[p/\hbar +1]}}, \;\;\;
U_3 U_4 = U_4 U_3 \sqrt{\frac{[p/\hbar -1]}{[p/\hbar +1]}}
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.3}
U_1 U_4 = U_4 U_1 \frac{\sqrt{[p/\hbar +1] [p/\hbar -1]}}
{[p/\hbar]} - U_3 U_2 \frac{q^{p/\hbar}}{[p/\hbar ]}
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.4}
U_3 U_2 = U_2 U_3 \frac{\sqrt{[p/\hbar +1] [p/\hbar -1]}}{[p/\hbar]}
- U_1 U_4 \frac{q^{-p/\hbar}}{[p/\hbar ]}
\eeq
The rest of the relations contained in (\ref{1.9}) are not independent and
can be deduced from (\ref{2.1})-(\ref{2.4}).
Additionally, from (\ref{1.8}) one gets
\begin{equation}\label} \def\eeq{\end{equation}{2.5}
\begin{array}} \def\er{\end{array}{ll} q^{p/\hbar} \, U_1 = q^{-1} \, U_1 \, q^{p/\hbar}, \; & q^{p/\hbar} \,
U_2 = q \, U_2 \, q^{p/\hbar},\\ & \\ q^{p/\hbar} \, U_3 = q^{-1} \, U_3 \,
q^{p/\hbar},\;& q^{p/\hbar} \, U_4 = q \, U_4 \, q^{p/\hbar}.\er
\eeq
Thus, relations (\ref{2.1})-(\ref{2.5}) describe the algebra ${\cal U}$.
Using them, one may verify the following statement:
\begin{prop}
A central element of ${\cal U}$ is given by the "deformed" determinant of
the matrix $U$:
\begin{equation}\label} \def\eeq{\end{equation}{2.6}
{\rm Det} U \equiv U_1 U_4 \sqrt{\frac{[p/\hbar +1]}{[p/\hbar ]}}
- U_2 U_3 \sqrt{\frac{[p/\hbar -1]}{[p/\hbar]}}=
q U_4 U_1 \sqrt{\frac{[p/\hbar -1]}{[p/\hbar]}} -
q U_3 U_2 \sqrt{\frac{[p/\hbar +1]}{[p/\hbar ]}} .
\eeq
\end{prop}
For fixed value of ${\rm Det} U$ the algebra ${\cal U}$ contains only four
independent generators. In classical limit $(\hbar = 0)$ they become the
coordinates on 4-dimensional phase space.
For further discussion it is convenient to introduce new variables instead
of $U_i$:
\begin{equation}\label} \def\eeq{\end{equation}{2.7}
\wh{U}_i = U_i \sqrt{[p/\hbar]}.
\eeq
The coordinates $\{p$, $\wh{U}_i\}$ form a new set of generators of the algebra
${\cal U}$. The commutation relations (\ref{2.1})-(\ref{2.5}) rewritten
in terms of the new generators acquire a simpler form:
\begin{equation}\label} \def\eeq{\end{equation}{2.8}
\wh{U}_1 \wh{U}_3 = q^{-1} \wh{U}_3 \wh{U}_1, \;\;
\wh{U}_2 \wh{U}_4 = q^{-1} \wh{U}_4 \wh{U}_2, \;\;
\wh{U}_1 \wh{U}_2 = \wh{U}_2 \wh{U}_1, \;\;
\wh{U}_3 \wh{U}_4 = \wh{U}_4 \wh{U}_3
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.10}
\wh{U}_1 \wh{U}_4 = \wh{U}_4 \wh{U}_1 \frac{[p/\hbar +1]}{[p/\hbar]} -
\wh{U}_3 \wh{U}_2 \frac{q^{p/\hbar}}{[p/\hbar ]},
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.11}
\wh{U}_3 \wh{U}_2 = \wh{U}_2 \wh{U}_3 \frac{[p/\hbar +1] }{[p/\hbar]} -
\wh{U}_1 \wh{U}_4 \frac{q^{-p/\hbar}}{[p/\hbar ]},
\eeq
\vspace*{1mm}
\begin{equation}\label} \def\eeq{\end{equation}{2.9}
\begin{array}} \def\er{\end{array}{ll} q^{p/\hbar} \, {\wh U}_1 = q^{-1} \, {\wh U}_1 \, q^{p/\hbar}, \;&
q^{p/\hbar} \, {\wh U}_2 = q \, {\wh U}_2 \, q^{p/\hbar}, \\ & \\
q^{p/\hbar} \, {\wh U}_3 = q^{-1} \, {\wh U}_3 \, q^{p/\hbar},\;&
q^{p/\hbar} \, {\wh U}_4 = q \, {\wh U}_4 \, q^{p/\hbar}. \er
\eeq
The central element (\ref{2.6}) in new variables looks as follows
\begin{equation}\label} \def\eeq{\end{equation}{2.12}
{\rm Det} U \equiv ( \wh{U}_1 \wh{U}_4 - \wh{U}_2 \wh{U}_3)\frac{1}
{[p/\hbar ]} = ( \wh{U}_4 \wh{U}_1 - \wh{U}_3 \wh{U}_2 )\frac{q}{[p/\hbar]}.
\eeq
The explicit form of the matrix inverse to $\wh{U}$, which we shall need
later, is
\begin{equation}\label} \def\eeq{\end{equation}{2.125}
\wh{U}^{-1} = \frac{1}{{\rm Det}U} \left(\begin{array}} \def\er{\end{array}{cc} \wh{U}_4 & -q \wh{U}_2 \\
- \wh{U}_3 & q \wh{U}_1 \er\right)\frac{1}{[p/\hbar]}.
\eeq
Finally, from (\ref{2.7}) we conclude that the expression (\ref{1.7}) for
the matrix $L$ looks similarly in terms of new matrix $\wh{U}$:
\begin{equation}\label} \def\eeq{\end{equation}{2.130}
L \,=\, U\,D\,U^{-1} \,=\, \wh{U} \, D \, \wh{U}^{-1} \ .
\eeq
\section{ \normalsize} \def\ds{\displaystyle\bf NON-DEFORMED CASE. }
\setcounter{equation}{0}
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf A. Representation of algebra ${\cal U}_0$. }
\hspace*{6mm} First, we consider the limit $\gamma\rightarrow 0$, $\hbar \neq 0$
(note that $q$-numbers turn into ordinary numbers),i.e., here we deal
with a well understood situation -- the representation theory of $SL(2)$.
An investigation of this simple non-deformed case will make further
results more transparent.
Let us denote the corresponding limit algebra as ${\cal U}_0$. The defining
$R$-matrix relations (\ref{1.9}) now degenerate to
\begin{equation}\label} \def\eeq{\end{equation}{2.12b}
\up{1}{U_0}\, \up{2}{U_0} = \up{2}{U_0} \, \up{1}{U_0} {\cal R}_{\pm}^0 (p),
\eeq
where
\begin{equation}\label} \def\eeq{\end{equation}{1.12c}
{\cal R}_{+}^0 (p) = {\cal R}_{-}^0 (p) = \left( \begin{array}} \def\er{\end{array}{cccc} 1 & & & \\
& \ds\frac{\sqrt{ (p/\hbar +1) (p/\hbar -1) }}{(p/\hbar)} &
\ds\frac{\hbar}{p} & \\ & & \\ & -\ds\frac{\hbar}{p} &
\ds\frac{\sqrt{(p/\hbar +1) (p/\hbar -1)}}{(p/\hbar)}
& \\ & & & 1 \er\right).
\eeq
The analogues of relations (\ref{2.8})-(\ref{2.9}) for ${\cal U}_0$ are (from
now on we omit the index $0$ for the generators of ${\cal U}_0$)
\begin{equation}\label} \def\eeq{\end{equation}{2.13a}
p \wh{U}_1 = \wh{U}_1 (p-\hbar), \;\; p \wh{U}_2 = \wh{U}_2 (p+\hbar),
\;\; p \wh{U}_3 = \wh{U}_3 (p-\hbar), \;\; p \wh{U}_4 = \wh{U}_4
(p+\hbar) ,
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.13b}
[\wh{U}_1, \wh{U}_2 ] \, = \, [\wh{U}_1, \wh{U}_3 ] \, = \,
[\wh{U}_2, \wh{U}_4 ] \, = \, [\wh{U}_3, \wh{U}_4 ] = 0 ,
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.13c}
[ \wh{U}_1, \wh{U}_4 ] = {\rm Det}U_0 ,\;\;\;\;\; [\wh{U}_3, \wh{U}_2 ] = -
{\rm Det}U_0 ,
\eeq
where ${\rm Det} U_0$ stands for a limit version of (\ref{2.12}):
\begin{equation}\label} \def\eeq{\end{equation}{2.13d}
{\rm Det} U_0 = ( \wh{U}_1 \wh{U}_4 - \wh{U}_2 \wh{U}_3 ) \frac{\hbar}{p}
= ( \wh{U}_4 \wh{U}_1 - \wh{U}_3 \wh{U}_2 ) \frac{\hbar}{p} .
\eeq
\begin{prop}\label{P5}
A possible solution for (\ref{2.13a})-(\ref{2.13d}) is
\begin{equation}\label} \def\eeq{\end{equation}{2.14}
\wh{U}_1 = \partial_1, \;\; \wh{U}_2 = z_2, \;\; \wh{U}_3 =
-\partial_2, \;\; \wh{U}_4 = z_1;
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{2.15}
p = \hbar ( z_1 \partial_1 + z_2 \partial_2 +1),
\eeq
where we denote
$\partial_i \equiv \frac{\partial}{\partial z_i}$.
\end{prop}
{\it Remark:} \ } \def\proof{{\it Proof}\ The representation given by (\ref{2.14})-(\ref{2.15}) is not
unique. In particular, the rescaling
$\wh{U}_i\rightarrow} \def\lar{\leftarrow} \def\ov{\overline c_i \wh{U}_i$ (where $c_i$ are numerical
constants such that $c_1 c_4 = c_2 c_3$) is allowable.
The Proposition \ref{P5} together with the connection formula (\ref{2.7})
allows us to write out the explicit form of the matrix $U_0$
\begin{equation}\label} \def\eeq{\end{equation}{2.16}
U_0 = \left( \begin{array}} \def\er{\end{array}{cc} \partial_1 & z_2 \\ -\partial_2 & z_1 \er \right)\,
\sqrt{\frac{\hbar}{p}} .
\eeq
Note that this matrix is ``unimodular'', i.e.,
${\rm Det} U_0 = ( \partial_1 z_1 + z_2 \partial_2 )\frac{\hbar}{p} = 1$.
To describe the obtained representation of the algebra ${\cal U}_0$ completely
one has to define a space where operators (\ref{2.14})-(\ref{2.16}) act. It
is natural to think that this space is $D(z_1,z_2)$ -- a space of
holomorphic functions of two complex variables.
Let us recall that $D(z_1,z_2)$ is a space spanned on the vectors
\begin{equation}\label} \def\eeq{\end{equation}{3.2}
|j,m\rangle =\frac{z_1^{j+m}z_2^{j-m}}{\sqrt{(j+m)!(j-m)!}},\;\;\;\;
j=\begin{array}} \def\er{\end{array}{lllll}0,&\!\!\frac{1}{2},&\!\!1,&\!\!\frac{3}{2},&\!\! ...\er
\;\;\;\;\;m=-j,..,j
\eeq
and equipped with the scalar product
\begin{equation}\label} \def\eeq{\end{equation}{3.3}
\langle f,g\rangle =\frac{1}{(2\pi i)^2}\int
\overline{f(z_1, z_2)} g(z_1,z_2)
e^{-z_1\bar z_1-z_2\bar z_2}dz_1 d\bar z_1 dz_2 d\bar z_2.
\eeq
The system (\ref{3.2}) is orthonormal with respect to the scalar product
(\ref{3.3}), that is
$ <j,m|j^{\prime},m^{\prime}>=\delta_{j j^{\prime}} \delta_{m m^{\prime}}$.
For the given scalar product a rule of conjugation of operators looks as
follows
\begin{equation}\label} \def\eeq{\end{equation}{3.5}
(z_i)^{*}=\partial _i,\;\;\;(\partial _i)^{*}=z_i.
\eeq
The question concerning unitarity of the matrix $U_0$ is discussed
in Appendix B.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf B. Connection with $\bf T^*{\cal B}$. }
\hspace*{6mm} The generators of $sl(2)$ can be realized on $D(z_1,z_2)$ as
differential operators:
\begin{equation}\label} \def\eeq{\end{equation}{3.1}
l_{+}=z_1\partial _2,\;\;l_{-}=z_2\partial
_1,\;\;l_3=\begin{array}} \def\er{\end{array}{c} \frac 12 \er \! (z_1\partial _1-z_2\partial _2).
\eeq
Using these expressions we can compare the representation of the
algebra ${\cal L}$ (or, more precisely, its limit version ${\cal L}_0$)
given by Theorem 1 with the representation given by Proposition
\ref{P4}.
Indeed, in the limit $\ga\rightarrow} \def\lar{\leftarrow} \def\ov{\overline 0$ the initial formula (\ref{1.7})
acquires form
\begin{equation}\label} \def\eeq{\end{equation}{3.05}
L=I+\ga L_0 +O(\ga^2) , \;\;\; L_0 = U_0\left(\begin{array}} \def\er{\end{array}{cc} p&\\ &-p\er\right)
U_0^{-1} .
\eeq
Substituting here the explicit expressions (\ref{2.15})-(\ref{2.16})
for $p$, $U_0$ and using the representation
(\ref{3.1}) for generators of $sl(2)$, one derives the following
limit form of the $L$-operator:
\begin{equation}\label} \def\eeq{\end{equation}{3.15}
L_0=\hbar \ \left(\begin{array}} \def\er{\end{array}{cc} 2 + z_1\,\partial_1 - z_2\,\partial_2 &
2 z_2\,\partial_1 \\ 2 z_1\,\partial_2 & 2 - z_1\,\partial_1 + z_2\,
\partial_2 \er \right) =
2\hbar \ \left(\begin{array}} \def\er{\end{array}{cc} 1+l_3 & l_{-} \\ l_{+} & 1-l_3 \er \right) .
\eeq
Notice that (\ref{3.15}) exactly coincides with
(\ref{L5}) taken in the limit $\ga \rightarrow} \def\lar{\leftarrow} \def\ov{\overline 0$. This explains why we had to
introduce the factor $q^2$ in (\ref{L5}).
The next observation concerning the limit of $L$-operator reads as
follows.
\begin{prop}\label{P6}
The matrix $L_0$ in the representation (\ref{3.15}) admits
the decomposition
\begin{equation}\label} \def\eeq{\end{equation}{3.35a}
L_0 = A_0\,B_0\,A_0^{-1},
\eeq
where
\begin{equation}\label} \def\eeq{\end{equation}{3.35b}
A_0 = \left(\begin{array}} \def\er{\end{array}{cc} z_1^{-1/2} & -z_1^{-1/2} z_2 \\
0 & z_1^{1/2} \er\right), \;\;\;
B_0= \hbar\left(\begin{array}} \def\er{\end{array}{cc} p/\hbar+1/2 & 0 \\ 2\,\partial_2 &
-(p/\hbar-1/2)\er\right)
\eeq
and $p$ is defined as in (\ref{2.15}).
\end{prop}
This statement can be verified directly.
Let us comment on the meaning of this proposition. First, note that $A_0$
is a realization of a group-like element of the Borel subgroup of $SL(2)$.
Moreover, this explicit form of $A_0$ is straightly connected with the
construction of the model space $\cal M$ developed by Gelfand et al.\REF{1}
Indeed, the space $D(z_1,z_2)$ being a realization of the model
space for $SL(2)$ (compare (\ref{0.0}) and (\ref{3.2})) is spanned on
monomials with arguments which are combinations of the entries of $A_0$.
On the other hand, $B_0$ is of opposite (with respect to $A_0$)
triangularity and its entries are operators acting on a given realization
of the model space. Therefore, $B_0$ can be regarded as an element of the
space dual to the corresponding Borel subalgebra.
Thus, $A_0$ and $B_0$ are coordinates in the base and in a fiber of the
cotangent bundle $T^*\B$.
At this stage the appearance of $T^*\B$ "inside" the algebra $\cal L$ looks
somewhat mysterious, but we shall clarify it later.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf C. Clebsch-Gordan coefficients. }
\hspace*{6mm} Let us consider an action of the generators of the algebra
${\cal U}_0$ defined in (\ref{2.15})-(\ref{2.16}) on the space $D(z_1,z_2)$
(which is a realization of the model space).
The action of these operators on the basic vectors (\ref{3.2}) is given by
\begin{equation}\label} \def\eeq{\end{equation}{3.8}
p \; |j,m\rangle =(2j+1)\hbar \; |j,m\rangle ,
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{3.9} \!\!\!\! \begin{array}} \def\er{\end{array}{ll}
U_1|j,m\rangle =\left(\frac{j+m}{2j+1}\right)^{1/2}|j-\frac 12 ,m-\frac 12
\rangle , &
U_2|j,m\rangle =\left(\frac{j-m+1}{2j+1}\right)^{1/2}|j+\frac 12 ,m-\frac 12
\rangle ,\\ \\
U_3|j,m\rangle =-\left(\frac{j-m}{2j+1}\right)^{1/2}|j-\frac 12 ,m+\frac 12
\rangle , &
U_4|j,m\rangle =\left(\frac{j+m+1}{2j+1}\right)^{1/2}|j+\frac 12 ,m+\frac 12
\rangle . \er
\eeq
Formula (\ref{3.8}) allows us to identify the operator $p$ as
$p=2\wh{j}+1$, where $\wh{j}$ is the operator of spin. Hence,
invariant subspaces of $p$ on the model space are those with
fixed value of spin $j$.
Formulae (\ref{3.9}) show that $U_i$ are generators of the basic shifts
on the model space (as illustrated on Fig.1). This observation is very
important. As we shall see later, the same picture holds for $q\neq 1$.
\vspace*{1mm}
\hbox{
\begin{picture}(200,150)(-100,0)
\put(20,20){\vector(1,0){150}} \put(70,20){\vector(0,1){131}}
\multiput(90,20)(0,10){12}{\line(0,1){5}}
\multiput(110,20)(0,10){12}{\line(0,1){5}}
\multiput(130,20)(0,10){12}{\line(0,1){5}}
\multiput(70,80)(10,0){8}{\line(1,0){5}}
\multiput(70,100)(10,0){8}{\line(1,0){5}}
\multiput(70,120)(10,0){8}{\line(1,0){5}}
\put(73,147){$j$} \put(173,23){$m$}
\put(72,5){$m\!-\!\frac{1}{2}$} \put(106,5){$m$}
\put(122,5){$m\!+\!\frac{1}{2}$}
\put(42,78){$j\!-\!\frac{1}{2}$} \put(42,98){$j$}
\put(42,118){$j\!+\!\frac{1}{2}$}
\put(75,65){$U_1$} \put(75,125){$U_2$}
\put(135,65){$U_3$} \put(135,125){$U_4$}
\thicklines
\put(110,100){\vector(1,1){20}} \put(110,100){\vector(-1,1){20}}
\put(110,100){\vector(1,-1){20}} \put(110,100){\vector(-1,-1){20}}
\end{picture} } \vspace*{0.5mm}
\begin{center}
\parbox{9cm}{\small {\bf Fig.1:} \
Action of the operators $U_i$ on the model space.}
\end{center}
Now comparing the matrix elements $\langle\,j'',m''|U_i|j,m\,\rangle$
following from (\ref{3.9})
with values of the Clebsch-Gordan coefficients (CGC) for decomposition of
the tensor product of irreducible representations $V_j\otimes V_{\frac 12}$
for $sl(2)$ which are given by the Van-der-Waerden formula
\begin{equation}\label} \def\eeq{\end{equation}{3.10}
\left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 & j^{\prime \prime } \\ m &
m^{\prime } & m^{\prime \prime } \er \right\} = \delta_{m^{\prime \prime },
m+m^{\prime }} \sqrt{ \frac{(j+\frac 12-j^{\prime \prime })!
(j+j^{\prime \prime }-\frac 12 )!(j^{\prime \prime }+\frac 12-j)! }
{(j+j^{\prime \prime }+\frac 32)!} } \; \times
\eeq
\[ \!\! \times \sum\limits_{r\geq 0}
\frac{(-1)^r \sqrt{ (j+m)!(j-m)!(j^{\prime \prime}+m^{\prime \prime})!
(j^{\prime\prime }-m^{\prime \prime })!(2j^{\prime\prime }+1) } }
{r!(j+\frac 12-j^{\prime \prime}-r)!(j-m-r)!(\frac 12+m^{\prime }-r)!
(j^{\prime \prime }-\frac 12+m+r)!(j^{\prime \prime }-j-m^{\prime }+r)!}, \]
we establish the following correspondence
\begin{equation}\label} \def\eeq{\end{equation}{3.14} \! \begin{array}} \def\er{\end{array}{ll}
\langle j^{\prime \prime} , m^{\prime \prime} \mid U_1\mid j, m\rangle=
\delta_{j^{\prime \prime },j- \frac 12 } \left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 &
j^{\prime \prime} \\ m & -\frac 12 & m^{\prime \prime} \er \right\} , \\ [1mm]
\langle j^{\prime \prime} , m^{\prime \prime} \mid U_2 \mid j, m\rangle=
\delta_{j^{\prime \prime },j+ \frac 12 }\left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 &
j^{\prime \prime} \\ m & -\frac 12 & m^{\prime \prime} \er \right\} , \\ [1mm]
\langle j^{\prime \prime} , m^{\prime \prime} \mid U_3 \mid j, m\rangle=
\delta_{j^{\prime \prime },j- \frac 12 }\left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 &
j^{\prime \prime} \\ m & \frac 12 & m^{\prime \prime} \er \right\}, \\ [1mm]
\langle j^{\prime \prime} , m^{\prime \prime} \mid U_4 \mid j, m\rangle=
\delta_{j^{\prime \prime },j+ \frac 12 }\left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 &
j^{\prime \prime} \\ m & \frac 12 & m^{\prime \prime}\er \right\}. \er
\eeq
Thus, we proved the following statement:
\begin{prop}\label{P7}
The generators $U_i$ of the algebra ${\cal U}_0$ are operators of the
basic shifts on the model space for $sl(2)$ and they generate the
Clebsch-Gordan coefficients corresponding to decomposition of the product
$V_j\otimes V_{\frac 12}$ of the irreps of sl(2).
\end{prop}
This statement allows to call the matrix $U_0$ a ''generating matrix''
(by analogy with the notion of a generating function) for CGC.
{\it Remark:} \ } \def\proof{{\it Proof}\ Usually, introducing a generating object (well-known examples are
the generating functions for different sets of polynomials, e.g., for the
Legendre polynomials), one makes properties of the objects under
consideration more evident. We think that the notion of generating
matrix will be useful for calculations involving CGC of classical and
quantum algebras.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf D. Wigner-Eckart theorem.}
\hspace*{6mm} One should underline a connection of the results obtained above
(Proposition \ref{P7}) and the well-known mathematical construction --
Wigner-Eckart theorem,\REF{19} which has important applications
in quantum mechanics.
Let us remind that the Wigner-Eckart theorem gives CGC for classical
Lie algebra ${\cal J}$ as matrix elements of some set of operators. These
operators are called {\it tensor operators}. They map the corresponding
model space $\cal M$ onto itself and have special transformation properties
under adjoint action of the algebra. In the case of ${\cal J}=sl(2)$ the
Wigner-Eckart theorem reads as follows.
\begin{theorem} \label{Twe}
Let $l_{+}$, $l_{-}$ and $l_{3}$ be the generators of $sl(2)$ and
let $T^{j\;}_{\;m}$, $m=-j,..,j$ be a set of operators acting on $\cal M$
and obeying the commutation relations
\begin{equation}\label} \def\eeq{\end{equation}{we1}
[l_3,T^{j\;}_{\;m}]=m\,T^{j\;}_{\;m},\;\;\;\;
[l_{\pm},T^{j\;}_{\;m}]=\sqrt{(j\mp m)(j\pm m+1)}\; T^{j\;}_{\;m\pm 1},
\eeq
where $j(j+1)$ is an eigenvalue of the Casimir operator for $sl(2)$.
Then the matrix elements of $T^{j\;}_{\;m}$ on $\cal M$ are proportional to
Clebsch-Gordan coefficients:
\[ <j^{\prime\prime} m^{\prime\prime} |T^{j\;}_{\;m}|j^{\prime} m^{\prime} >
= C_{j^{\prime\prime}j^{\prime}}^{j}\;\left\{\begin{array}} \def\er{\end{array}{ccc}j^{\prime} & j &
j^{\prime\prime}\\ m^{\prime} & m & m^{\prime\prime} \er \right\}, \]
where the coefficients $C_{j^{\prime\prime}j^{\prime}}^j$ do not depend on
$m$, $m^{\prime}$, $m^{\prime\prime}$.
\end{theorem}
Proposition \ref{P7} says that any tensor operators of spin
$j=1/2$ (that is $\{T^{1/2}_{1/2},\, T^{1/2}_{-1/2}\}$,\
$T^{1/2}_m\, :\, V_j \mapsto V_j\otimes V_{1/2}=V_{j+1/2}\oplus V_{j-1/2}$)
may be constructed via the operators $U_i$ (in fact, it is evident from
Fig.1). Indeed, comparing the commutation relations obtained directly from
(\ref{2.16}) and (\ref{3.1})
\begin{equation}\label} \def\eeq{\end{equation}{we20} \begin{array}} \def\er{\end{array}{llll}
{}[l_{+},U_1]=U_3, & [l_{+},U_2]=U_4, & [l_{+},U_3]=0, & [l_{+},U_4]=0,\\
& & & \\ {} [l_{-},U_1]=0, & [l_{-},U_2]=0, & [l_{-},U_3]=U_1, & [l_{-},U_4]
= U_2,\\ & & & \\ {}[l_{3},U_1]=-\frac{1}{2} U_1, &[l_{3},U_2] =-\frac{1}{2}
U_2, & [l_{3},U_3] = \frac{1}{2} U_3, & [l_{3},U_4]= \frac{1}{2} U_4 \er
\eeq
with Theorem \ref{Twe}, we get the following.
\begin{prop}\label{P8}
The generators $U_i$ of the algebra ${\cal U}_0$ form a basis for tensor
operators of spin $1/2$, that is components $T^{1/2}_{1/2}$ and
$T^{1/2}_{-1/2}$ of any tensor operator of spin 1/2 can be realized as
linear combinations of $U_i$:
\begin{equation}\label} \def\eeq{\end{equation}{3.55}
T^{1/2}_{-1/2}= \mu(p) \;U_1+\nu(p) \;U_2,\;\;\;
T^{1/2}_{1/2}=\mu(p) \;U_3+\nu(p) \;U_4,
\eeq
where $\mu(p)$ and $\nu(p)$ are functions only of $p=2\wh{j}+1$.
\end{prop}
\section{ \normalsize} \def\ds{\displaystyle\bf DEFORMED CASE. }
\setcounter{equation}{0}
\hspace*{6mm} Now we want to extend the results
obtained in the previous section to the case of $q\neq 1$. In
particular, we are going to examine the representations of the algebra
${\cal U}$ (see Definition \ref{DU} above) and to show that the
corresponding matrix $U$ generates Clebsch-Gordan coefficients for the
deformed Lie algebra. For these purposes we shall exploit a natural
connection of ${\cal U}$ with $(T^*\B)_q$.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf A. The $\bf q$-oscillators approach. }
\hspace*{6mm} There exist different ways to obtain desirable representations of
the algebra ${\cal U}$. First we describe a more direct but less
instructive method, which is similar to that used in the
non-deformed case.
By analogy with the non-deformed case studied above, one can assume that
the entries of the matrix $U$ might be realized as operators
(deformations of those obtained in Proposition \ref{P5}) acting on
the space of two complex variables. Indeed, using the definition
(\ref{2.6}) of the central element of $\cal U$ and taking into account
the identity for $q$-numbers
\begin{equation}\label} \def\eeq{\end{equation}{q1} [a]\,q^b+[b]\,q^{-a}=[a+b] ,
\eeq we can rewrite (\ref{2.8})-(\ref{2.11}) in the following way:
\begin{equation}\label} \def\eeq{\end{equation}{q2}
\wh{U}_1 \wh{U}_3 = q^{-1} \wh{U}_3 \wh{U}_1, \;\;\, \wh{U}_2 \wh{U}_4
= q^{-1} \wh{U}_4 \wh{U}_2, \;\;\, \wh{U}_1 \wh{U}_2 = \wh{U}_2
\wh{U}_1, \;\;\, \wh{U}_3 \wh{U}_4 = \wh{U}_4 \wh{U}_3 , \eeq \begin{equation}\label} \def\eeq{\end{equation}{q3}
\wh{U}_1 \wh{U}_4 - q^{-1} \wh{U}_4 \wh{U}_1 = q^{-1} {\rm Det}U \,
q^{p/\hbar},\;\;\;\;\; \wh{U}_3 \wh{U}_2 - q\wh{U}_2 \wh{U}_3 =
- {\rm Det}U \,q^{-p/\hbar}.
\eeq
The relations (\ref{q3}) are well known in the theory of
$q$-oscillators ($q$-bosons).\REF{20} Recall that $q$-analogues of
creation, annihilation, and number operators form the deformed
Heisenberg algebra defined by the commutation relations
\begin{equation}\label} \def\eeq{\end{equation}{q3.1}
a\, a^+ - q \, a^+ a = N^{-1},\;\; N\, a = q^{-1}\, a\, N,\;\;
N\, a^+ = q\, a^+ N \, ,
\eeq
and they can be realized in terms of multiplication and
difference operators:
\begin{equation}\label} \def\eeq{\end{equation}{q4}
a^+=z,\;\;a=z^{-1} [z\,\partial_z],\;\; N = q^{z\,\partial_z} .
\eeq
Using two pairs of generators of the deformed Heisenberg algebra, one
can construct the generators of $U_q(sl(2))$:
$l_+ = a_1^+\,a_2$, $l_- = a_2^+\,a_1$, $q^{l_3}=N_1^{1/2}N_2^{-1/2}$.
Applying here the representation (\ref{q4}) one gets
\begin{equation}\label} \def\eeq{\end{equation}{q33}
l_+ = z_1 z_2^{-1} [z_2\partial_2 ],\;\;\;\;\;
l_- = z_2 z_1^{-1} [z_1\partial_1 ],\;\;\;\;\;
q^{l_3} = q^{\frac{1}{2} (z_1 \partial_1 - z_2 \partial_2)}.
\eeq
The Casimir operator (\ref{L8}) of $U_q(sl(2))$ in this realization is
given by
\begin{equation}\label} \def\eeq{\end{equation}{q4.5}
C= q\,N_1 N_2 + q^{-1} N_1^{-1} N_2^{-1}.
\eeq
Now, comparing, (\ref{q2})-(\ref{q3}) with (\ref{q3.1}), it is easy to
conclude that the pairs $(\wh{U}_1,\wh{U}_4)$ and
$(\wh{U}_2,\wh{U}_3)$ are similar to two pairs of $q$-boson operators.
Taking into account the Weyl-like form of relations (\ref{q2}) and
having already found explicit expressions (\ref{2.14})-(\ref{2.15})
for the generators of algebra ${\cal U}_0$, we get an answer for $D$
and $\wh{U}$ in terms of $q$-oscillators. More precisely, a
straightforward calculation allows to verify the following statement:
\begin{prop}\label{P9}
Equations (\ref{q2})-(\ref{q3}) have the family of solutions:
\begin{equation}\label} \def\eeq{\end{equation}{q6}
q^{p/\hbar} = q\,N_1\,N_2,\;\;\;\; \wh{U} = \left( \begin{array}} \def\er{\end{array}{cc} \alpha} \def\be{\beta} \def\ga{\gamma_0 \,a_1
\, N_1^{\alpha} \def\be{\beta} \def\ga{\gamma}\,N_2^{-\be} &\be_0\, a_2^+\,N_1^{\be}\,N_2^{-\alpha} \def\be{\beta} \def\ga{\gamma} \\ [1mm]
-\ga_0\, a_2 \,N_1^{-(1+\be)}\, N_2^{\alpha} \def\be{\beta} \def\ga{\gamma} &
\dl_0\,a_1^+\,N_1^{-\alpha} \def\be{\beta} \def\ga{\gamma}\,N_2^{1+\be} \er\right),
\eeq
where $\;\alpha} \def\be{\beta} \def\ga{\gamma_0\,\dl_0=q\,\be_0\,\ga_0$.
\end{prop}
Let us note that this form of $\wh{U}$ is consistent with the
condition (\ref{2.5}).
Taking into account the connection formula (\ref{2.7}) and applying to
the generators $a_i$, $a_i^+$, $N_i$ the representation (\ref{q4}),
one obtains from (\ref{q6}) a family of representations of the algebra
${\cal U}$. To select some of them, we have to impose an additional
condition.
As mentioned above (see (\ref{3.05})-(\ref{3.15})), in the non-deformed
case substitution of the generating matrix $U_0$ in the formula
(\ref{1.7}) gives the matrix $L_0$ which exactly coincides with the
limit version of the matrix (\ref{L5}). It is natural to suppose that
the generating matrix $U$ corresponding to deformed algebra produces
in the same way the matrix (\ref{L5}) itself. Bearing in mind the
property (\ref{2.130}), we obtain the following
\begin{prop}\label{P10}
The condition $\wh{U}D\wh{U}^{-1}=L$, where $L$ is the matrix
(\ref{L5}), $D$ is given by
\begin{equation}\label} \def\eeq{\end{equation}{q5.5}
D = \left(\begin{array}} \def\er{\end{array}{cc} q^{p/\hbar} & \\ & q^{-p/\hbar} \er\right) =
\left(\begin{array}} \def\er{\end{array}{cc} q\,N_1\,N_2 & \\ &
q^{-1}\,N_1^{-1}\,N_2^{-1} \er\right) \, ,
\eeq and $\wh{U}$ is given by (\ref{q6}), imposes the following
restrictions:
\begin{equation}\label} \def\eeq{\end{equation}{q8}
\alpha} \def\be{\beta} \def\ga{\gamma+\be+\frac{1}{2}=0,\;\;\; \alpha} \def\be{\beta} \def\ga{\gamma_0=q\,\ga_0 ,\;\;\; \be_0= \dl_0.
\eeq
\end{prop}
Substitution of (\ref{q8}) into (\ref{q6}) completes a description of
$\wh{U}$ in terms of $q$-oscillators.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf B. Connection with $\bf (T^*{\cal B})_q$. }
\hspace*{6mm} Now we are going to develop another approach to constructing
representations of $\cal U$. It is more universal since it is based on
the connection (which takes place for arbitrary quantum Lie algebra)
of the algebra ${\cal L}$ (see Definition \ref{DL}) with $(T^*\B)_q$
and on the interpretation of the deformed Borel subgroup $\B_q$ as a
quantum model space.
To clarify the announced connection we start with the following
theorem (this is a version of the theorem given in Ref.10 for
$L$-operators with nonultralocal relations)
\begin{theorem} \label{ABL}
Let the matrices $A$ and $B$ obey the relations of type (\ref{1.1}):
\begin{equation}\label} \def\eeq{\end{equation}{4.1}
R_{\pm }\up{1}{A} \, \up{2}{A} \,=\, \up{2}{A} \, \up{1}{A}
R_{\pm} \ , \;\;\;\;
R_{\pm } \up{1}{B} \, \up{2}{B} \,=\, \up{2}{B} \, \up{1}{B} R_{\pm}
\eeq
and the additional exchange relation
\begin{equation}\label} \def\eeq{\end{equation}{4.2}
\up{1}{A} \, \up{2}{B} \,=\, \up{2}{B} \, \up{1}{A} R_{+} \ , \;\;\;\;
\up{2}{A} \, \up{1}{B}R_{-} \,=\, \up{1}{B} \, \up{2}{A} \ .
\eeq
Then the $L$-operator constructed by means of similarity transformation
\begin{equation}\label} \def\eeq{\end{equation}{4.3}
L=ABA^{-1}
\eeq
satisfies the relation (\ref{1.3}).
\end{theorem}
{\it Remark:} \ } \def\proof{{\it Proof}\ Since (\ref{4.1}) defines a quantum group structure, $A^{-1}$ in
(\ref{4.3}) should be understood as an antipode of $A$.
\proof of Theorem \ref{ABL} is straightforward :
\[ \begin{array}} \def\er{\end{array}{c}
\up{1}{L} R^{-1}_{-} \up{2}{L}R_{-}= \up{1}{A}
\up{1}{B}(\up{1}{A})^{-1} R_{-}^{-1} \up{2}{A} \up{2}{B}
(\up{2}{A})^{-1} R_{-}= \up{1}{A} \up{1}{B} \up{2}{A} R_{-}^{-1}
(\up{1}{A})^{-1} \up{2}{B} (\up{2}{A})^{-1}R_{-}=\\ \\
=\up{1}{A}\,\up{2}{A}\,\up{1}{B}\,\up{2}{B}R_{+}^{-1}(\up{1}{A})^{-1}
(\up{2}{A})^{-1} R_{-}= R_{+}^{-1}
\up{2}{A}\,\up{1}{A}\,\up{2}{B}\,\up{1}{B} R_{-} (\up{2}{A})^{-1}
(\up{1}{A})^{-1}= \\ \\ = R_{+}^{-1} \up{2}{A} \up{2}{B}\up{1}{A}R_{+}
(\up{2}{A})^{-1} \up{1}{B} (\up{1}{A})^{-1}= R_{+}^{-1}\up{2}{A}
\up{2}{B} (\up{2}{A})^{-1} R_{+} \up{1}{A} \up{1}{B}(\up{1}{A})^{-1}=
R_{+}^{-1} \up{2}{L} R_{+}\up{1}{L}.\er \]
Thus, for a given quantum group $G_q$, the algebra $\cal L$ is
embedded into the algebra generated by entries of $A$ and $B$ obeying
(\ref{4.1})-(\ref{4.2}). To argue that (\ref{4.1})-(\ref{4.2})
describe a $q$-analogue of $T^*\B$, let us notice that the
non-symmetric (with respect to $R$-matrices) form of the relations
(\ref{4.2}) imposes some restriction on the structure of the matrices
$A$ and $B$. Say, if $R_+$ is an upper triangular matrix, then $A$ and
$B$ must be upper and lower triangular, respectively. Therefore, one
may think of $A$ and $B$ as coordinates in the deformed Borel
subgroup ${\cal B}_q$ and in the dual quantum space, respectively. In
other words, the matrices $A$ and $B$ are coordinate and momentum on
the deformed phase space $(T^*\B)_q$ respectively. Thus
(\ref{4.1})-(\ref{4.2}) may be regarded as a definition of
$(T^*\B)_q$ (for additional comments see Ref.10).
We should underline here that, although the matrices $A$ and $B$ look
similarly on quantum level, they
transform into different objects when $q\rightarrow} \def\lar{\leftarrow} \def\ov{\overline 1$. Indeed, in the limit
$q\rightarrow} \def\lar{\leftarrow} \def\ov{\overline 1$ one has $L \rightarrow} \def\lar{\leftarrow} \def\ov{\overline I + \ga\hbar L_0$ and the corresponding
limit forms of $A$ and $B$ are
\begin{equation}\label} \def\eeq{\end{equation}{4.45}
A \rightarrow} \def\lar{\leftarrow} \def\ov{\overline A_0, \;\;\; B \rightarrow} \def\lar{\leftarrow} \def\ov{\overline I+\ga\hbar B_0,
\eeq
where $A_0$ is a group-like element, whereas
$B_0$ is rather an element of algebra (see (\ref{3.35b}) as an example
of $A_0$, $B_0$ for $sl(2)$).
Comparing the statements of Theorems 1 and \ref{ABL} and
taking into account the equality (\ref{2.130}), we get the formula
\begin{equation}\label} \def\eeq{\end{equation}{4.115}
L = A\, B \, A^{-1} = \wh{U} \, D \, \wh{U}^{-1},
\eeq
which points out a possibility to construct the matrix $\wh{U}$
obeying (\ref{2.8})-(\ref{2.9}) via the generators of $(T^*\B)_q$.
This connection is very important; below we consider it for
$SL_q(2)$ in all details.
Now let us turn to the example of $SL_q(2)$. For $R_{\pm}$ defined as
in (\ref{1.4}) one can choose \begin{equation}\label} \def\eeq{\end{equation}{4.4} A=\left(\begin{array}} \def\er{\end{array}{cc} a & c \\ 0 &
a^{-1} \er \right) ,\ \ \ \ B=\left(\begin{array}} \def\er{\end{array}{cc} b & 0 \\ d & b^{-1}
\er\right) . \eeq
Explicit relations for the generators of $(T^*\B)_q$ following from
(\ref{4.1})-(\ref{4.2}) are
\begin{equation}\label} \def\eeq{\end{equation}{4.5} a\,c=q^{-1}c\,a,\ \ \ b\,c=q^{1/2}c\,b,\ \ \ a\,b=q^{1/2}b\,a;
\eeq
\begin{equation}\label} \def\eeq{\end{equation}{4.6} b\,d=q^{-1}d\,b,\ \ \ a\,d=q^{1/2}d\,a,\ \ \
c\,d=q^{-1/2}d\,c+q^{-1/2}\omega\, b^{-1}a.
\eeq
Performing the following decomposition
\begin{equation}\label} \def\eeq{\end{equation}{4.7}
d=d_0+d_1=d_0+q^{1/2}c^{-1}b^{-1}a \, ,
\eeq
we transform (\ref{4.6}) to homogeneous form:
\begin{equation}\label} \def\eeq{\end{equation}{4.8} b\,d_0=q^{-1}d_0\, b,\ \ \ \ a\, d_0=q^{1/2}d_0 \, a,
\ \ \ \ c\, d_0=q^{-1/2}d_0 \, c .
\eeq
Thus, (\ref{4.5}) and (\ref{4.8}) describe four variables obeying
Weyl-like commutation relations. Using the jargon of conformal field
theory, we shall call these formulae "free field representation" and
the generators $a$, $b$, $c$, $d_0$ "free field" variables.
{\it Remark:} \ } \def\proof{{\it Proof}\ The last of equations (\ref{4.6}) is nothing but a commutation
relation entering the definition of deformed Heisenberg algebra.
Indeed, comparing (\ref{4.5})-(\ref{4.6}) with (\ref{q3.1}), one can
establish the following correspondence ($\rho$ stands for arbitrary
numerical constant):
$$
c \sim N^{\rho}\,a^{+},\;\;\; d \sim -\omega \,
N^{-1/2-\rho}\,a,\;\;\; b^{-1}\,a \sim q^{\rho}\,N^{-3/2}.
$$
Thus, the transformation (\ref{4.7}) can be interpreted as
"bosonization" of $q$-oscillators.
Now, substituting (\ref{4.4}) in (\ref{4.3}), we get
\begin{equation}\label} \def\eeq{\end{equation}{4.9} \begin{array}} \def\er{\end{array}{c}
L = q^{1/2} \left( \begin{array}} \def\er{\end{array}{cc} a & c \\ 0 & a^{-1} \er \right) \left(
\begin{array}} \def\er{\end{array}{cc} b & 0 \\ d & b^{-1} \er \right) \left( \begin{array}} \def\er{\end{array}{cc} a^{-1} & -q\,c
\\ 0 & a \er \right) = \\ \\ = \left( \begin{array}} \def\er{\end{array}{cc} q\, (b+b^{-1}) +
a^{-1}c\, d_0 & - q^2 a\, c\, (b+q \, a^{-1} c\, d_0) \\ (a\,c)^{-1}
(b^{-1} + q^{-1} a^{-1}c\, d_0) & -q^2 a^{-1} c\, d_0 \er \right). \er
\eeq
This matrix provides a "free field" realization of the algebra
$\cal L$ for $U_q(sl(2))$. Note that the additional scaling factor
$q^{1/2}$ was introduced in (\ref{4.9}) to ensure a coincidence of the
Casimir operators calculated by formulae (\ref{1.6}) for the matrix
(\ref{4.9}):
\begin{equation}\label} \def\eeq{\end{equation}{4.10}
K_1 = q^2 (b+b^{-1}), \;\;\; K_2 = q^3
\eeq
with those for the matrix (\ref{L5}). In fact, we redefined the matrix
$B$ in (\ref{4.4}) as \begin{equation}\label} \def\eeq{\end{equation}{4.95} \wt{B} = q^{1/2} B . \eeq Comparing
the Casimir operator $K_1$ given by (\ref{4.10}) with one given by
(\ref{L6}), we identify the operator $b$ with the power of the operator
of spin $\wh{j}$ :
\begin{equation}\label} \def\eeq{\end{equation}{4.105}
b = q^{2 \wh{j} +1}.
\eeq
It follows from (\ref{4.10}) that matrix $L$ contains only three
independent variables (it is easy to see from the explicit form
(\ref{4.9}) that these are $b$, $ac$ and $a^{-1}cd_0$). Moreover,
direct calculation using (\ref{4.5}),(\ref{4.8}) shows that all
elements of the matrix $L$ commute with operator $b$. That agrees with
the property (\ref{1.12}).
Now exploiting the connection described by formula (\ref{4.115}), one
can obtain an exact expression for $\wh{U}$.
\begin{theorem} \label{main}
The algebra ${\cal U}\equiv \{\wh{U},p\}$ with defining relations
(\ref{2.8})-(\ref{2.9}) has the following realization in terms of
generators $a$, $b$, $c$, $d_0$
\begin{equation}\label} \def\eeq{\end{equation}{5.00}
b = q^{p/\hbar},\;\;\;\;
\wh{U} = \left( \begin{array}} \def\er{\end{array}{cc} \frac{1}{\omega}\, a\,(b+a^{-1}\,c\,d_0)\,
e^{-\frac{i \xi}{2}} & c\,e^{\frac{i \xi}{2}} \\ [1mm]
\frac{1}{\omega}\, c^{-1}(b^{-1} + q^{-1} a^{-1}\,c\,d_0)\,
e^{-\frac{i\xi}{2}} & a^{-1}\, e^{\frac{i \xi}{2}} \er \right),
\eeq
where $\omega\equiv q-q^{-1}$,\ $d_0$ is defined in (\ref{4.7}), and
\begin{equation}\label} \def\eeq{\end{equation}{5.15}
e^{i \xi}= a^{-1}\;b^{\ga}\;c^{-1}\;d_0^{-1}
\eeq
with $\ga$ being an arbitrary constant.
\end{theorem}
This theorem gives a "free field" representation of the algebra $\cal
U$. Let us remark that the remaining freedom in (\ref{5.15})
corresponds only to canonical transformations (since $\xi$ and $p$ are
conjugate variables).
The formulated theorem will be proved in several steps. First, we
introduce a lower-triangular matrix which diagonalizes the matrix
$\wt{B}$:
\begin{equation}\label} \def\eeq{\end{equation}{5.1}
V = \left( \begin{array}} \def\er{\end{array}{cc} v_1 & 0 \\ v_3 & v_2 \er \right), \;\;\;
\wt{B} = V \wt{B}_0 V^{-1}, \;\;\; \wt{B}_0 = \left(\begin{array}} \def\er{\end{array}{cc} q^{1/2} b
& 0 \\ 0 & q^{1/2} b^{-1} \er \right) \equiv q^{1/2} B_0 .
\eeq
\begin{prop}
A possible solution for the matrix $V$ is \begin{equation}\label} \def\eeq{\end{equation}{5.3} v_1 =v_1(b),\;\;\;
v_2 =v_2(b),\;\;\; v_3 = d \,v_1(b)\,f(b), \eeq where $v_1(b)$,
$v_2(b)$ are arbitrary functions of $b$ and
$f(b) = (b-qb^{-1})^{-1}$.
\end{prop}
Thus, matrix $L$ given by (\ref{4.9}) admits a decomposition of the form:
\begin{equation}\label} \def\eeq{\end{equation}{5.5} L=\wh{U}_0\, \wt{B}_0\, \wh{U}_0^{-1}, \;\;\;\; \wh{U}_0 =
A\,V. \eeq
However, this diagonalization is not unique. Using an arbitrary power
of the diagonal matrix $Q$, which depends on the variable conjugate to
$b$,
\begin{equation}\label} \def\eeq{\end{equation}{5.6}
Q = \left( \begin{array}} \def\er{\end{array}{cc} e^{i\xi} & \\ & e^{-i\xi} \er \right),
\;\;\;\; b\,e^{i\xi} = q\,e^{i\xi}\,b,
\eeq
we obtain a
family of diagonalizing matrices:
\begin{equation}\label} \def\eeq{\end{equation}{5.7}
L= \wh{U}_{\dl}\,\wt{B}_{\dl}\,\wh{U}_{\dl}^{-1}, \;\;\; \wh{U}_{\dl}=
A\,V\,Q^{\dl}, \;\;\; \wt{B}_{\dl} = Q^{-\dl} \wt{B}_0 \, Q^{\dl} =
q^{\dl}\,\wt{B}_0 = q^{\dl+1/2}\, B_0.
\eeq
An explicit form of the diagonalizing matrix is
\begin{equation}\label} \def\eeq{\end{equation}{5.8}
\wh{U}_{\dl} = A\,V\,Q^{\dl} = \left( \begin{array}} \def\er{\end{array}{cc} (a\,v_1 +c\,d\,v_1 f)\,
e^{i\dl \xi} & c\,v_2 \,e^{-i\dl \xi} \\ a^{-1} d\,v_1 f\, e^{i\dl \xi} &
a^{-1}v_2 \, e^{-i\dl \xi} \er \right).
\eeq
Here we should describe a new object $e^{i \xi}$ which appeared in the
matrix $\wh{U}$. We assume that the following Weyl-like relations
hold:
\begin{equation}\label} \def\eeq{\end{equation}{5.10}
a\, e^{i \xi}= q^{\alpha} \def\be{\beta} \def\ga{\gamma} e^{i \xi}\, a, \;\;\; b\, e^{i
\xi}= q\,e^{i \xi}\, b, \;\;\; c\, e^{i \xi}= q^{\be} e^{i \xi}\,
c,\;\;\; d_0\, e^{i \xi}= q^{\ga} e^{i \xi}\,d_0.
\eeq
\begin{prop}
The set of equations (\ref{5.10}) is equivalent to
\begin{equation}\label} \def\eeq{\end{equation}{5.11}
e^{i\xi}= a^{\be+(\ga-1)/2}\;b^{\ga}\;c^{(\ga-1)/2-\alpha} \def\be{\beta} \def\ga{\gamma}\;d_0^{-1}.
\eeq
\end{prop}
Now we have to remind that the matrix $U$ (and $\wh{U}$ as well)
described in the Theorem 1 has to satisfy the relation (\ref{1.8}) or,
equivalently, to the relation
\begin{equation}\label} \def\eeq{\end{equation}{5.85}
\up{1}{B_0} \,\up{2}{\wh{U}_{\dl}} =
\up{2}{\wh{U}_{\dl}} \, \up{1}{B_0} \, \sigma,
\eeq
where $\sigma$ and $B_0$ were introduced in (\ref{1.8})
and (\ref{5.1}), respectively. A straightforward calculation using
(\ref{4.5})-(\ref{4.6}) leads to the following.
\begin{prop}
The matrix $\wh{U}_{\dl}$ given by (\ref{5.8}) satisfies the relation
(\ref{5.85}) only for $ \dl = -1/2$.
\end{prop}
It is worth mentioning that such a choice of $\dl$ exactly compensates
the renormalization of the matrix $B$ in (\ref{4.95}), i.e., \
$\wt{B}_{-1/2} = B_0$.
Bearing in mind the formula (\ref{4.7}), one can
rewrite (\ref{5.8}) for $\dl = \begin{array}} \def\er{\end{array}{c}-\frac{1}{2} \er$ as follows
\begin{equation}\label} \def\eeq{\end{equation}{5.9}
\wh{U} \equiv \wh{U}_{-1/2} = \left( \begin{array}} \def\er{\end{array}{cc}
a(b+a^{-1}\,c\,d_0)\,w\, e^{-\frac{i \xi}{2}} & c\,v\,e^{\frac{i
\xi}{2}} \\ c^{-1}(b^{-1} + q^{-1} a^{-1}\,c\,d_0)\,w\,e^{-\frac{i
\xi}{2}} & a^{-1}\,v \, e^{\frac{i \xi}{2}} \er \right),
\eeq
where $w \equiv f(b) v_1(b)$,\ $v \equiv v_2(b)$.
Finally, a direct check shows (see the Appendix C) that the matrix
(\ref{5.9}) obeys eqs.$\!$ (\ref{2.8})-(\ref{2.9}) if the functions
$w$, $v$ are constant (we chose them as follows: $v(b)=1$,
$w(b)=\frac{1}{\omega}$) and the coefficients in
(\ref{5.10})-(\ref{5.11}) satisfy the conditions $\be \!=\! -\alpha} \def\be{\beta} \def\ga{\gamma$, \
$\ga \!=\! \alpha} \def\be{\beta} \def\ga{\gamma-\be-1 \!=\! 2\alpha} \def\be{\beta} \def\ga{\gamma-1$. Thus, Theorem \ref{main} is
proven.
Let us end the discussion of relation of $(T^*\B)_q$ to algebras
$\cal L$ and $\cal U$ with one more statement:
\begin{theorem}
The algebra generated by coordinates on $(T^*\B)_q$ is isomorphic to
the algebra generated by entries of the matrix $L$ and $Q$.
\end{theorem}
\proof. Formulae (\ref{4.9}) and (\ref{5.15}) provide explicit
expressions for entries of $L$ and $Q$ via the generators $a$, $b$, $c$,
$d_0$ (up to unessential canonical transformation in (\ref{5.15})).
Conversely, suppose matrix $L$ and the element $e^{i\xi}$ are given.
Then, as it follows from (\ref{4.9}), one can construct from entries of
$L$ the combinations $b$, $ac$ and $a^{-1}cd_0$. Together with (\ref{5.15})
this allows to recover the ``coordinates'' $a$, $b$, $c$, $d_0$.
Although we considered this theorem only for the case of $SL_q(2)$, there
is an evidence that it holds for the generic case. Indeed, in the case
of $G_q=SL_q(N)$ a point on the quantum bundle $(T^*\B)_q$ is parameterized
by $N\times N$ matrices $A$ and $B$. As above, the matrix $L=ABA^{-1}$
satisfies (\ref{1.3}) and therefore its entries generate the corresponding
algebra $\cal L$. However, the dimension of $(T^*\B)_q$ exceeds the
dimension of $\cal L$ : \
${\rm dim}\, (T^*\B)_q - {\rm dim}\, {\cal L} =
(N^2+N-2)-(N^2-1)=N-1$.
It us very probable that the remaining $(N-1)$ generators are exactly
those that enter the diagonal unimodular $N\times N$ matrix $Q$.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf C. Explicit representation.}
\hspace*{6mm} Now we face the problem of constructing of an explicit
representation for the generators $a$, $b$, $c$, $d_0$. A Weyl-like
form of the commutation relations (\ref{4.5}),(\ref{4.8}) points out
the possibility of getting a realization for these generators in terms
of two pairs of canonical variables. This also means (due to the
interpretation of (\ref{4.7}) as "bosonization" of $q$-oscillators)
that the generators $a$, $b$, $c$, $d$ admit a realization via
$q$-oscillators. Evidently, such a representation is not unique.
It is natural to realize $a$, $b$, $c$, $d_0$ as operators acting on
the $q$-analogue of the space $D(z_1,z_2)$. We shall denote this space
as $D_q(z_1,z_2)$. The space $D_q(z_1,z_2)$ is spanned on the basic
vectors of form (remember that $[x]$ stands for $q$-numbers)
\begin{equation}\label} \def\eeq{\end{equation}{4.01} |j,m\rangle
=\frac{z_1^{j+m}z_2^{j-m}}{\sqrt{[j+m]![j-m]!}},\;\;\;\;j=
\begin{array}} \def\er{\end{array}{lllll}0,&\!\!\frac{1}{2},&\!\!1,&\!\!\frac{3}{2},&\!\! ...\er
\;\;\;\; m=-j,..,j. \eeq One can define on $D_q(z_1,z_2)$ such a
scalar product that the system (\ref{4.01}) is orthonormal, that is $
<j,m|j^{\prime},m^{\prime}>=\delta_{j j^{\prime}} \delta_{m
m^{\prime}}$.
{\it Remark:} \ } \def\proof{{\it Proof}\ This scalar product is a deformation of (\ref{3.3}). Its explicit
form makes use of the $q$-exponent and the Jackson integral. See
Ref.20 for details.
In all formulae concerning the space $D_q(z_1,z_2)$ we suppose that
$q$ is chosen as described in Sec.$\!$ II (i.e., it belongs either
to the real axis or to the unit circle at the complex plane).
In this case an
analogue of the rule of conjugation (\ref{3.5}) is
\begin{equation}\label} \def\eeq{\end{equation}{4.03}
(z_i)^{*}=z_i^{-1}\;[z_i\partial_i],\;\;\;\;\;
(z_i\partial_i)^{*}=z_i\partial_i.
\eeq
The formulae (\ref{4.105}) and (\ref{5.00}) imply that the generator
$b$ is a power of the operator of spin. Hence, on the space
$D_q(z_1,z_2)$ it is given by \begin{equation}\label} \def\eeq{\end{equation}{5.16} b = q^{z_1\,\partial_1 +
z_2\,\partial_2 + 1 } = q\,N_1\,N_2. \eeq
Next, let us remind that we already know the limit versions of the
generators $a$, $b$, $c$, $d$ (see Proposition \ref{P6}; one should
take into account the rescaling (\ref{4.95})). Their appropriate
deformations for generic $q$ are described by
\begin{prop}\label{P15}
The set of operators (with arbitrary constants $\lambda} \def\up{\stackrel} \def\wt{\widetilde_i$, $\nu_i$)
\begin{equation}\label} \def\eeq{\end{equation}{5.17} \begin{array}} \def\er{\end{array}{ll}
a=q^{\lambda} \def\up{\stackrel} \def\wt{\widetilde_0}\,z_1^{-1/2}\,N_1^{\lambda} \def\up{\stackrel} \def\wt{\widetilde_1}, & c=q^{\nu_0}\,z_1^{-1/2}\,z_2\,
N_1^{\lambda} \def\up{\stackrel} \def\wt{\widetilde_1-2}\,N_2^{\nu_2}, \\ \\ b=q\,N_1\,N_2, &
d=-q^{\lambda} \def\up{\stackrel} \def\wt{\widetilde_0-\nu_0+\nu_2}\,z_2^{-1}\,(N_2-N_2^{-1})\,N_1\,N_2^{-\nu_2} \er
\eeq
satisfies (\ref{4.5})-(\ref{4.6}) and gives in the
limit $\ga\rightarrow} \def\lar{\leftarrow} \def\ov{\overline 0$ the generators found in (\ref{3.35b}).
\end{prop}
Although due to Theorem \ref{main} this proposition gives a family of
representations for $\cal U$,
we again should impose an additional condition using the matrix
(\ref{L5}) as a standard (justification for this trick was given
above).
\begin{prop}\label{P16}
Matrix $L$ given by (\ref{4.9}) coincides with the matrix (\ref{L5})
taken in the representation (\ref{q33}) provided that \begin{equation}\label} \def\eeq{\end{equation}{5.19} b =
q\,N_1\,N_2,\;\;\;a\,c= q^{-1/2}\,z_1^{-1}\,z_2\,N_1^{-1/2}
\,N_2^{-1/2},\;\;\; a^{-1}\,c\,d_0=-N_1^{-1}\,N_2. \eeq
\end{prop}
Comparing the statements of Propositions \ref{P15} and \ref{P16}, we
derive:
\begin{equation}\label} \def\eeq{\end{equation}{5.20}
\begin{array}} \def\er{\end{array}{ll} a=q^{\lambda} \def\up{\stackrel} \def\wt{\widetilde_0}\,z_1^{-1/2}\,N_1^{3/4},\ &
b=q\,N_1\,N_2, \\
c=q^{\nu_0}\,
z_1^{-1/2}\,z_2\,N_1^{-5/4}\,N_2^{-1/2}, & \\
d_0=-q^{\lambda} \def\up{\stackrel} \def\wt{\widetilde_0-\nu_0-1/2}\, z_2^{-1}\,N_1\,N_2^{3/2},\ &
q^{\lambda} \def\up{\stackrel} \def\wt{\widetilde_0+\nu_0}= q^{-1/8} . \er
\eeq
Substituting (\ref{5.20}) into (\ref{5.15}) (and remember that
(\ref{5.15}) is defined only up to a coefficient), we get \begin{equation}\label} \def\eeq{\end{equation}{5.21}
e^{i\xi} = q^{2\epsilon}\,z_1\,N_1^{\ga-1/2}\,N_2^{\ga-1}, \eeq where
$\ga$ and $\epsilon$ are arbitrary. Finally, substituting
(\ref{5.20})-(\ref{5.21}) into (\ref{5.9}),
we obtain (one should remember
$U$ and $\wh{U}$ are defined only up to arbitrary scaling factor)
\begin{equation}\label} \def\eeq{\end{equation}{5.22}
\wh{U} = \! \left( \begin{array}} \def\er{\end{array}{cc}
\frac{1}{\omega}\,\alpha} \def\be{\beta} \def\ga{\gamma_0\,z_1^{-1}\, N_1^{1-\ga/2}\,N_2^{3/2-\ga/2}
(N_1-N_1^{-1}) & \be_0 \,z_2\,N_1^{\ga/2-3/2} N_2^{\ga/2-1} \\ & \\ -
\frac{1}{\omega}\, q^{-1} \alpha} \def\be{\beta} \def\ga{\gamma_0\, z_2^{-1} N_1^{1/2-\ga/2}\,
N_2^{1-\ga/2} (N_2-N_2^{-1}) & \be_0\,z_1\,N_1^{\ga/2-1}
N_2^{\ga/2-1/2} \er \right).
\eeq
It is easy to check that the family of matrices (\ref{5.22}) exactly
coincides with what was obtained in $q$-oscillator approach (see
Propositions \ref{P9} and \ref{P10}).
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf D. Quantum Clebsch-Gordan coefficients. }
\hspace*{6mm} Using the connection formula (\ref{2.7}) we get from (\ref{5.22})
a family of matrices $U$ which provide possible representations of the
algebra $\cal U$. It is natural to study an action of the entries of
these matrices on the space $D_q(z_1,z_2)$ described above. On the
basic vectors (\ref{4.01}) these operators act as follows:
\begin{equation}\label} \def\eeq{\end{equation}{6.1}
\begin{array}} \def\er{\end{array}{c} U_1\, |j,m\rangle = C_1\,q^{\frac{1}{2}(j-m+1)}
\sqrt{\frac{[j+m]}{[2j+1]}}\, |j-\frac{1}{2},m-\frac{1}{2}\rangle , \\
\\ U_2\, |j,m\rangle = C_2\,q^{-\frac{1}{2}(j+m)}
\sqrt{\frac{[j-m+1]}{[2j+1}]}\, |j+\frac{1}{2},m-\frac{1}{2}\rangle ,
\\ \\ U_3\, |j,m\rangle =-C_3\,q^{-\frac{1}{2}(j+m+1)}
\sqrt{\frac{[j-m]}{[2j+1]}}\, |j-\frac{1}{2},m+\frac{1}{2}\rangle , \\
\\ U_4\, |j,m\rangle =C_4\,q^{\frac{1}{2}(j-m)}
\sqrt{\frac{[j+m+1]}{[2j+1]}}\, |j+\frac{1}{2},m+\frac{1}{2}\rangle , \er
\eeq
where the coefficients $C_i$ do not depend on $m$.
Note that, similarly to the classical case, the operators $U_i$
correspond to the basic shifts on the model space. Comparing the
matrix elements $<j^{\prime},m^{\prime}|U_i|j,m>$ following from
(\ref{6.1}) with values of CGC for $U_q(sl(2))$ given by $q$-analogue
of the Van-der-Waerden formula,\REF{18,21} which for the
decomposition of $V_j\otimes V_{1/2}$ looks like following
\[ \left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 & j^{\prime \prime } \\
m & m^{\prime } & m^{\prime \prime } \er \right\}_q =
\delta_{m^{\prime \prime},m + m^{\prime}}\; \left(\frac{[j+\frac
12-j^{\prime \prime }]![j+j^{\prime \prime }-\frac 12]![j^{\prime
\prime }+\frac 12-j]!}{[j+j^{\prime \prime }+\frac 32]!}
\right)^{1/2} \times \] \begin{equation}\label} \def\eeq{\end{equation}{6.2} \times q^{\frac{1}{2} (j+\frac{1}{2}
-j^{\prime \prime})(j+ j^{\prime \prime} +\frac{3}{2}) + jm^{\prime}
-\frac{1}{2} m} \times \eeq
\[ \times \sum\limits_{r\geq 0}\frac{(-1)^r q^{-r(j+j^{\prime \prime}+
\frac 32)} \left( [j+m]![j-m]![j^{\prime \prime}+m^{\prime\prime
}]![j^{\prime \prime }-m^{\prime \prime }]![2j^{\prime \prime
}+1]\right) ^{1/2} } {[r]![j+\frac 12-j^{\prime \prime
}-r]![j-m-r]![\frac 12+m^{\prime } -r]![j^{\prime\prime }-\frac
12+m+r]![j^{\prime \prime }-j-m^{\prime }+r]!},\] we establish the
following correspondence
\[ \! \begin{array}} \def\er{\end{array}{ll}
\langle j^{\prime \prime} , m^{\prime \prime} \mid U_1\mid j,
m\rangle= \delta_{j^{\prime \prime },j- \frac 12 }\,
\alpha} \def\be{\beta} \def\ga{\gamma_0\,q^{(1-\ga/2)j-1/2}\, \left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 & j^{\prime
\prime} \\ m & -\frac 12 & m^{\prime \prime} \er \right\}_q , \\
\langle j^{\prime \prime} , m^{\prime \prime} \mid U_2 \mid j,
m\rangle= \delta_{j^{\prime \prime },j+ \frac 12 }\,
\be_0\,q^{(\ga/2-1)j}\,\left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 & j^{\prime
\prime} \\ m & -\frac 12 & m^{\prime \prime} \er \right\}_q , \\
\langle j^{\prime \prime} , m^{\prime \prime} \mid U_3 \mid j,
m\rangle= \delta_{j^{\prime \prime },j- \frac 12 }\,
\alpha} \def\be{\beta} \def\ga{\gamma_0\,q^{(1-\ga/2)j-1/2}\, \left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 & j^{\prime
\prime} \\ m & \frac 12 & m^{\prime \prime} \er \right\}_q, \\ \langle
j^{\prime \prime} , m^{\prime \prime} \mid U_4 \mid j, m\rangle=
\delta_{j^{\prime \prime },j+ \frac 12 }\,
\be_0\,q^{(\ga/2-1)j}\,\left\{\begin{array}} \def\er{\end{array}{ccc} j & \frac 12 & j^{\prime
\prime} \\ m & \frac 12 & m^{\prime \prime}\er \right\}_q . \er \]
Thus we derive an analogue of Proposition \ref{P7}:
\begin{prop}\label{P17}
The generators $U_i$ of the algebra $\cal U$ are operators of the basic
shifts on the model space for $U_q(sl(2))$ and they generate the
$q$-Clebsch-Gordan coefficients corresponding to decomposition of the
product $V_j\otimes V_{\frac 12}$ of irreps of $U_q(sl(2))$.
\end{prop}
{\it Remark:} \ } \def\proof{{\it Proof}\ Putting $\alpha} \def\be{\beta} \def\ga{\gamma_0=q^{1/2}$, $\be_0=1$ and $\ga=2$ in (\ref{5.22}),
we get the following generating matrix
\begin{equation}\label} \def\eeq{\end{equation}{6.3}
U = \left( \begin{array}} \def\er{\end{array}{cc} z_1^{-1}\,
[z_1 \partial_1 ] q^{\frac 12 (z_2 \partial_2 +1)} & z_2 \, q^{-\frac
12 z_1 \partial_1 } \\ \\-z_2^{-1}\, [z_2 \partial_2 ] q^{-\frac 12
(z_1 \partial_1 +1)} & z_1 \, q^{\frac 12 z_2 \partial_2 } \er \right)
\,\frac{1}{\sqrt{[p/\hbar]} \def\pp{[p/\hbar+1]} \def\pu{[p/\hbar-1]}} \ , \;\;\;\;
{\rm Det} U =
q^{1/2} \ ,
\eeq
which may be called ``exact'' as it satisfies (\ref{6.1}) with
$C_i=1$.
The question about unitarity of the matrix (\ref{6.3}) is discussed
in Appendix B.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf E. Generalized Wigner-Eckart theorem.}
\hspace*{6mm} As we demonstrated in the previous section, entries of the matrix
$U_0$ are tensor operators of spin $1/2$ for ${\cal J}=sl(2)$, hence
they provide a realization of the Wigner-Eckart theorem. Let us now
consider the matrix $U$ from this point of view.
The theory of tensor operators for quantum algebras was discussed by
many authors (see, e.g., Ref.22). In particular, generalized Wigner-Eckart
theorem (in the case of ${\cal J}} \def\Uc{{\cal U}_q=U_q(sl(2))$ reads as follows.
\begin{theorem} \label{TWE}
Let $l_{+}$, $l_{-}$ and $l_{3}$ be the generators of $U_q(sl(2))$ and
let $T^{j\;}_{\;m}$, $m=-j,..,j$ be a set of operators acting on the
deformed model space $\cal M$ and obeying the commutation relations
\begin{equation}\label} \def\eeq{\end{equation}{WE1}
{} [l_3,T^{j\;}_{\;m}]=m\,T^{j\;}_{\;m},\;\;\;
{} l_{\pm}\,T^{j\;}_{\;m}\,q^{l_3} - q^{l_3\mp 1}\, T^{j\;}_{\;m}\,l_{\pm}
=\sqrt{[j\mp m][j\pm m+1]}\,T^{j\;}_{\;m\pm 1}.
\eeq
Then the matrix elements of $T^{j\;}_{\;m}$ on $\cal M$ are proportional to
$q$-Clebsch-Gordan coefficients:
\[ <j^{\prime\prime} m^{\prime\prime} |T^{j\;}_{\;m}|j^{\prime} m^{\prime} >
= C_{j^{\prime\prime}j^{\prime}}^{j}\;\left\{\begin{array}} \def\er{\end{array}{ccc}j^{\prime} & j &
j^{\prime\prime}\\ m^{\prime} & m & m^{\prime\prime} \er \right\}_q, \]
where the coefficients $C_{j^{\prime\prime}j^{\prime}}^j$ do not depend on
$m$, $m^{\prime}$, $m^{\prime\prime}$.
\end{theorem}
Proposition \ref{P17} implies that $U_i$ may be regarded as
q-tensor operators. Indeed, using (\ref{5.22}) and (\ref{q33}), one can
check that $U_i$ satisfy (\ref{WE1}) (one obtains for $U_i$
deformations of relations (\ref{we20})). Similarly to the classical case
we have the following.
\begin{prop}\label{P18}
The generators $U_i$ of the algebra ${\cal U}$ form a basis for $q$-tensor
operators of spin $1/2$,
that is components $T^{1/2}_{1/2}$ and
$T^{1/2}_{-1/2}$ of any $q$-tensor operator of spin 1/2 can be realized as
linear combinations of $U_i$:
\begin{equation}\label} \def\eeq{\end{equation}{WE5}
T^{1/2}_{-1/2}= \mu(p) \;U_1+\nu(p) \;U_2,\;\;\;
T^{1/2}_{1/2}=\mu(p) \;U_3+\nu(p) \;U_4,
\eeq
where $\mu(p)$ and $\nu(p)$ are functions only of $p$.
\end{prop}
{\it Remark:} \ } \def\proof{{\it Proof}\ Unlike the classical case, solution (\ref{5.22}) gives a
family of matrices $U$. However, the corresponding matrix elements
$<j^{\prime\prime} m^{\prime\prime} |U_i|j^{\prime} m^{\prime} > $
differ only by factors which do not depend on $m^{\prime}$,
$m^{\prime\prime}$. Thus, any representative of obtained family of
matrices $U$ may be used in Proposition \ref{P18}.
Let us end the description of the algebra $\cal U$ from the point of view
of theory of $q$-tensor operators with the following statement:
\begin{prop}\label{P19}
The matrices $U$ and $L$ defined in the Theorem 1 obey the relation
\begin{equation}\label} \def\eeq{\end{equation}{WE10}
R_- \,\up{1}{U}\,\up{2}{L} = \up{2}{L}\,R_+ \,\up{1}{U}.
\eeq
\end{prop}
The proof is straightforward
\[ \begin{array}} \def\er{\end{array}{c} R_- \,\up{1}{U}\,\up{2}{L} \, =
R_- \,\up{1}{U}\,\up{2}{U}\,\up{2}{D}\,(\up{2}{U})^{-1}= \,
\up{2}{U}\, \up{1}{U}\, {\cal R}_- \,\up{2}{D}\,(\up{2}{U})^{-1} = \\ [1mm]
=\, \up{2}{U}\, \up{1}{U}\,\up{2}{D}\,\sigma\, {\cal R}_+ \,(\up{2}{U})^{-1}
= \, \up{2}{U}\,\up{2}{D}\,\up{1}{U}\, {\cal R}_+ \,(\up{2}{U})^{-1} =
\up{2}{U}\,\up{2}{D}\,(\up{2}{U})^{-1}\, R_+ \,\up{1}{U} = \,
\up{2}{L}\,R_+ \,\up{1}{U}; \er \]
it makes use the relations (\ref{1.8})-(\ref{1.9}) and the property
(\ref{RD}).
A remarkable fact is that (\ref{WE10}) may be used for definition of
$q$-tensor operators instead of (\ref{WE1}). Indeed, in the limit
$\ga\rightarrow} \def\lar{\leftarrow} \def\ov{\overline 0$ it turns into
\begin{equation}\label} \def\eeq{\end{equation}{WE3}
[\up{1}{U_0}, \up{2}{L}_0 ] = \Lambda\, \up{1}{U_0},\;\;\;\;
\Lambda=\left(\begin{array}} \def\er{\end{array}{cccc} 1/2& & & \\ &-1/2&1& \\ &1&-1/2& \\ & & &1/2
\er\right).
\eeq
Using the explicit form of $L_0$ given in (\ref{3.05}), one can easily
check that this matrix relation is equivalent to (\ref{we20}).
More on $R$-matrix description of $q$-tensor operators is given in Ref.23.
\subsection*{ \hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf CONCLUSION.}
In this paper we have constructed the $q$-analogue of the phase space
$T^*\B$ and clarified its role in description of the model representation
of the corresponding quantum group $G_q$. We unraveled a connection between
the algebras generating by entries of matrix $(A,B)$, $(U,D)$ and $(L,Q)$.
The general formulae were concretized by the example of $G=SL(2)$.
An extension of the described scheme to the case of arbitrary group $G$
will definitely improve understanding of the role played by the matrix
${\cal R}(p)$ which so far has been discussed in the literature much
less then the standard matrix $R$.
The results of this paper can be generalized in several directions even for
the case of $SL(2)$. The first is consideration of the matrix $U$ with an
auxiliary space corresponding to the higher spin representation. It must
lead to an exact form of generating matrix for all CGC. The work in this
direction is in progress now.
The second point to be discussed is the case of $q$ being a root of unity.
The structure of ${\cal R}(p)$ allows to hope that reduction on so-called
"good" representations will be quite natural in our formalism.
However, this case is to be examined more carefully.
\subsection*{\normalsize} \def\ds{\displaystyle\bf \hspace*{6mm} Acknowledgments. }
We are grateful to A.Yu. Alekseev, P.P. Kulish and V. Schomerus for
stimulating discussions and useful comments. We would like to thank
Prof. A. Niemi for hospitality at TFT, University
of Helsinki, where this work was begun.
This work was partially supported by ISF grant R2H000 and by INTAS grant.
\subsection*{ \hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf Appendix A: Proof of Theorem 1. }
Using (\ref{1.8})-(\ref{1.9}) together with the identity (\ref{RD})
and taking into account that matrices $\up{1}{D}$, $\up{2}{D}$ and
$\sigma$ mutually commute, we check
\[ \begin{array}} \def\er{\end{array}{c}
\up{1}{L} R_-^{-1} \up{2}{L} R_- = \,\up{1}{U} \up{1}{D} (\up{1}{U})^{-1}
R_-^{-1}\up{2}{U}\, \up{2}{D} (\up{2}{U})^{-1} R_- =\\ \\
=\,\up{1}{U} \up{1}{D}\,\up{2}{U} {\cal R}_-^{-1}(p)\, (\up{1}{U})^{-1}\!
\up{2}{D} (\up{2}{U})^{-1} R_- = \, \up{1}{U}\, \up{2}{U} \up{1}{D} \sigma\,
{\cal R}_-^{-1}(p) \up{2}{D} \sigma\,
(\up{1}{U})^{-1} (\up{2}{U})^{-1} R_- =\\ \\
=R_+^{-1} \up{2}{U}\, \up{1}{U} {\cal R}_+ (p) \up{1}{D}\sigma\,
{\cal R}_-^{-1}(p) \up{2}{D} \sigma\, {\cal R}_- (p)\, (\up{2}{U})^{-1}
(\up{1}{U})^{-1} =\\ \\
=R_+^{-1}\up{2}{U}\,\up{1}{U} \up{2}{D} \sigma\up{1}{D} {\cal R}_- (p) \,
(\up{2}{U})^{-1} (\up{1}{U})^{-1} = R_+^{-1} \up{2}{U} \up{2}{D}\, \up{1}{U}
{\cal R}_+ (p)\, \sigma\up{1}{D}(\up{2}{U})^{-1} (\up{1}{U})^{-1} =\\ \\
=R_+^{-1} \up{2}{U}\up{2}{D}\, \up{1}{U} {\cal R}_+ (p)\, (\up{2}{U})^{-1}\!
\up{1}{D}(\up{1}{U})^{-1} = R_+^{-1} \up{2}{U} \up{2}{D} (\up{2}{U})^{-1}
R_+\up{1}{U} \up{1}{D} (\up{1}{U})^{-1} = R_+^{-1} \up{2}{L} R_+ \up{1}{L}.
\er\]
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf Appendix B:
On conjugation of $\bf U_0$ and $\bf U$. }
{}First we consider the matrix $U_0$. Using the rules of
conjugation (\ref{3.5}) (and taking into account that $p^*=p$), one can
check that the matrix conjugated to $U_0$ does not coincide with
$U_0^{-1}$; that is, the matrix $U_0$ itself is not unitary. However,
it turns out
that the transposed matrix (one should remember that in general
$(U^T)^{-1}\neq (U^{-1})^T$ for matrices with non-commuting entries)
\[U_0^T = \left( \begin{array}} \def\er{\end{array}{cc} \partial_1 & -\partial_2\\
z_2 & z_1 \er\right)\,\sqrt{\frac{\hbar}{p}} \]
satisfies the unitarity condition:
\[ (U_0^T)^* = \sqrt{\frac{\hbar}{p}}\,\left(\begin{array}} \def\er{\end{array}{cc} z_1 & \partial_2 \\
-z_2 & \partial_1 \er \right) = (U_0^T)^{-1}. \]
In the deformed case (recall that $q$ can be either real
or $|q|=1$) the matrix $U$ includes the operator $N$ which conjugates in
different ways for the different choices of $q$.
Let us consider the matrix $U$ given by (\ref{6.3}). The conjugated matrix
can be constructed according to the rules (\ref{4.03}). Using the formula
(\ref{q1}), one can check that the unitarity condition
$(U^T)^{*}\,U^T=\,U^T\,(U^T)^{*}= I $
(i.e., the same as in the non-deformed case) for the transposed matrix
holds only for real $q$. For $|q|=1$ see Ref.13.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf Appendix C: Proof of Theorem 4. }
Here we complete the proof of Theorem \ref{main}, i.e., we have to
prove that matrix (\ref{5.9}) satisfies (\ref{2.8})-(\ref{2.11}) if
the following conditions
\hspace*{2cm}$\begin{array}} \def\er{\end{array}{c} \\ \alpha} \def\be{\beta} \def\ga{\gamma + \be = 0,\;\;
\ga+\be-\alpha} \def\be{\beta} \def\ga{\gamma+1=0,\;\; v(b)=1,\;\;w(b)=1/\omega \\ \\ \er
$\hfill (C\,1)
$\hspace*{-6mm}$are fulfilled.
First, using relations (\ref{4.5}), (\ref{4.8}), (\ref{5.10}) and
conditions (C\,1), we check
\[\begin{array}} \def\er{\end{array}{l} \wh{U}_1 \wh{U}_2 = a\,(b+a^{-1}c\,d_0)\, e^{-i\xi}\;c\,e^{i\xi}=
q^{-1/2+\be/2} c\,a(b+ q\,a^{-1}c\,d_0) \,e^{i\xi} \, e^{-i\xi} =\\ [1.5mm]
= q^{\alpha} \def\be{\beta} \def\ga{\gamma/2+\be/2} c\,e^{i\xi}\,a(b+ q^{(\ga+\be-\alpha} \def\be{\beta} \def\ga{\gamma+1)/2}\,a^{-1}c\,d_0) \,
e^{-i\xi} = \wh{U}_2 \wh{U}_1 . \er \]
\[\begin{array}} \def\er{\end{array}{l} \wh{U}_1 \wh{U}_3 = a\,(b+a^{-1}c\,d_0)\,e^{-i\xi}\;c^{-1}\,
(b^{-1}+q^{-1}\,a^{-1}c\,d_0)\, e^{-i\xi} = \\ [1.5mm]
= q^{1/2-\be/2}c^{-1}\,a\,
(b+q^{-1}\,a^{-1}c\,d_0)\,e^{-i\xi}\,(b^{-1}+q^{-1}\,a^{-1}c\,d_0)\,
e^{-i\xi} =\\ [1.5mm]
=q^{-\be/2}c^{-1}\,a\,(b^{-1}+q^{-1}\,a^{-1}c\,d_0)\,
(b+q^{-1}\,a^{-1}c\,d_0)\,e^{-i\xi}\,e^{-i\xi} = \\ [1.5mm]
= q^{-1/2-\be/2}c^{-1} \,(b^{-1}+q^{-1}\,a^{-1}c\,d_0)\,a\,
(b+q^{-1}\,a^{-1}c\,d_0)\,e^{-i\xi}\, e^{-i\xi} = \\ [1.5mm]
= q^{-1-\alpha} \def\be{\beta} \def\ga{\gamma/2-\be/2}c^{-1}\,(b^{-1}+q^{-1}\,a^{-1}c\,d_0)\,a\,
e^{-i\xi}\,(b+a^{-1}c\,d_0)\,e^{-i\xi} = q^{-1} \wh{U}_3 \wh{U}_1 .\er \]
The rest of relations (\ref{2.8}) can be proved similarly.
Next, note that relation (\ref{2.10}) can be rewritten as follows
\hspace*{3cm}$\begin{array}} \def\er{\end{array}{c} \\ \wh{U}_1\;\wh{U}_4\,(b-b^{-1})-\wh{U}_4\;\wh{U}_1\,
(q\,b-q^{-1} \,b^{-1}) =-\omega\, \wh{U}_3 \; \wh{U}_2 \, b. \\ \\
\er $ \hfill (C\,2)
$\hspace*{-7.5mm}$ To prove this equality we transform its l.h.s.
and r.h.s. as follows
\[\begin{array}} \def\er{\end{array}{l}\wh{U}_1\;\wh{U}_4\,(b-b^{-1})-\wh{U}_4\;\wh{U}_1\,(q\,b-q^{-1}
\,b^{-1}) =
a\,(b+a^{-1}c\,d_0)\,e^{-i\xi} \; a^{-1} e^{i\xi}\,(b-b^{-1})- \\ [1.5mm]
- a^{-1} e^{i\xi} \; a\,(b+a^{-1}c\,d_0)\,e^{-i\xi} (q\,b-q^{-1}\,b^{-1}) =
q^{-\alpha} \def\be{\beta} \def\ga{\gamma/2}\,(q^{1/2}\,b+q^{-1/2}\,a^{-1}c\,d_0)\,(b-b^{-1}) - \\ [1.5mm]
- q^{-\alpha} \def\be{\beta} \def\ga{\gamma/2}\,(q^{-1/2}\,b+q^{1/2}\,a^{-1}c\,d_0)\,(q\,b-q^{-1}\,b^{-1}) =
-q^{-\alpha} \def\be{\beta} \def\ga{\gamma/2}\,\omega\,(q^{1/2}\,a^{-1}c\,d_0\,b+q^{-1/2}); \\ \\
\wh{U}_3\;\wh{U}_2 \,b =c^{-1}\,(b^{-1}+q^{-1}\,a^{-1}c\,d_0)\,
e^{-i\xi}\,c\,e^{i\xi} \,b = \\ [1.5mm]
= q^{\be/2}\,(q^{-1/2}\,b^{-1}+q^{1/2}\,
a^{-1}c\,d_0)\,b = q^{\be/2}\,(q^{-1/2}+q^{1/2}\,a^{-1}c\,d_0\,b). \er \]
Thus, the equality (C\,2) is fulfilled if conditions (C\,1) are valid.
The relation (\ref{2.11}) can be proved in the same way.
\subsection*{\hspace*{6mm} \normalsize} \def\ds{\displaystyle\bf References.}
\small
\REF{1}{I.N.Bernstein, I.M.Gelfand, S.I.Gelfand, \ {\it Models for
representations of Lie groups}. In: Proceedings of I.G.Petrovsky seminars,
{\bf 2}, 3 (1976) (in Russian);\
I.N.Bernstein, I.M.Gelfand, S.I.Gelfand, \ Funct. analysis and its
applications, {\bf 9}, No 4, 61 (1975). } \\ [1mm]
\REF{2} {A.M.Polyakov, \ Mod.Phys.Lett. {\bf A 2}, 893 (1987).} \\[1mm]
\REF{3} {A.Yu.Alekseev, S.L.Shatashvili, \ Comm. Math. Phys. {\bf 128},
197 (1990); \\
\ H.La, P.Nelson, A.S.Schwarz, \
Comm. Math. Phys. {\bf 134}, 539 (1990).} \\ [1mm]
\REF{4} {L.D.Faddeev, \
Comm. Math. Phys. {\bf 132}, 131 (1990). } \\ [1mm]
\REF{5} {A.Yu.Alekseev, L.D.Faddeev, \ Comm. Math. Phys. {\bf 141}, 413
(1991).} \\ [1mm]
\REF{6} {F.Falceto, K.Gawedzki, \ J. Geom. Phys.
{\bf 11}, 251 (1993).} \\ [1mm]
\REF{7} {A.A.Kirillov, \ {\it Elements of the theory of representations}.
Springer-Verlag (1979);\\
\ A.Yu.Alekseev, L.D.Faddeev, S.L.Shatashvili, \
J. Geom. Phys. {\bf 5}, 391 (1989). } \\ [1mm]
\REF{8} {A.Connes,\ {\it Noncommutative geometry}.
(Academic Press, 1994).} \\ [1mm]
\REF{9} {L.D.Faddeev, N.Yu.Reshetikhin, L.A.Takhtajan, \
Algebra and analysis {\bf 1}, 193 (1990). } \\ [1mm]
\REF{10} {A.Yu.Alekseev, L.D.Faddeev, M.A.Semenov-Tian-Shansky,
A.Yu.Volkov, \ {\it The unraveling of the quantum group structure in the
WZNW theory}, preprint CERN-TH-5981/91 (1991). } \\ [1mm]
\REF{11} {N.Yu.Reshetikhin, M.A.Semenov-Tian-Shansky, \ Lett. Math. Phys.
{\bf 19}, 133 (1990). } \\ [1mm]
\REF{12} {P.P.Kulish, N.Yu.Reshetikhin, \ Zapiski Nauch.Semin. LOMI
{\bf 101}, 101 (1981) (in Russian). } \\ [1mm]
\REF{13} {A.G.Bytsko, V.Schomerus,\ {\it Vertex operators -- from toy
model to lattice algebras}, preprint q-alg/9611010 (1996).} \\ [1mm]
\REF{14} {A.P. Isaev,\ J. Phys. {\bf A 29}, 6903 (1996).} \\ [1mm]
\REF{15} {J-L.Gervais, A.Neveu, \ Nucl. Phys.
{\bf B 238}, 125 (1984);\\
\ E.Cremmer, J-L.Gervais, \ Comm. Math. Phys.
{\bf 134}, 619 (1990). } \\ [1mm]
\REF{16} {O.Babelon, \ Comm. Math. Phys.
{\bf 139}, 619 (1991).} \\ [1mm]
\REF{17} {J.Avan, O.Babelon, E.Billey,\ Comm. Math. Phys. {\bf 178},
281 (1996). } \\ [1mm]
\REF{18} {A.N.Kirillov, N.Yu.Reshetikhin, \ Adv. Series in Math. Phys.
{\bf 11}, 202 (World Scientific, 1990).} \\ [1mm]
\REF{19} {A.O.Barut, R.Raczka, \ {\it Theory of group representations
and applications} (Scient. Publishers, 1977);\\ \
L.C.Biedenharn, J.D.Louck, \ {\it Angular momentum in quantum physics},
Encyclopedia of mathematics and its applications, {\bf v.8}
(Addison-Wesley, 1981). \\ [1mm]
\REF{20} {M.Arik, D.D.Coon, \ J. Math. Phys. {\bf 17}, 524 (1976);\\
\ L.C.Biedenharn, \ J. Phys. {\bf A 22}, L873 (1989);\\ \
A.J.Macfarlane, \ J. Phys. {\bf A 22}, 4581 (1989);\\ \
P.P.Kulish, E.Damaskinsky, \ J. Phys. {\bf A 23}, L415 (1990);\\ \
M.Chaichian, P.P.Kulish, \ Phys. Lett. {\bf B 234}, 72 (1990).} \\ [1mm]
\REF{21} {M.Nomura. \ J. Math. Phys. {\bf 30}, 2397 (1990);\\ \
L.Vaksman. \ Sov. Math. Dokl. {\bf 39}, 467 (1989);\\ \
H. Ruegg. \ J. Math. Phys. {\bf 31}, 1085 (1990). } \\ [1mm]
\REF{22} {L.C.Biedenharn, M.Tarlini, \ Lett. Math. Phys. {\bf 20},
271 (1990);\\ \
K.Bragiel, \, Lett. Math. Phys. {\bf 21}, 181 (1991);\\ \
G.Mack, V.Schomerus,\, Phys. Lett. {\bf B 267}, 207 (1991);\\ \
V.Rittenberg, M.Scheunert, \, J. Math. Phys.
{\bf 33}, 436 (1992).} \\ [1mm]
\REF{23} {A.G.Bytsko,\ {\it Tensor operators in $R$-matrix approach},
preprint DESY-95-254 (1995).}
\end{document}
|
\section*{Preface}
The {\tsc{Pythia}} and {\tsc{Jetset}} programs are frequently used for event generation
in high-energy physics. The emphasis is on multiparticle production
in collisions between elementary particles. This in particular means
hard interactions in $\e^+\e^-$, $\p\p$ and $\e\p$ colliders, although
also other applications are envisaged. The programs are intended to
generate complete events, in as much detail as
experimentally observable ones, within the bounds of our current
understanding of the underlying physics. Many of the components of
the programs represent original research, in the sense that models
have been developed and implemented for a number of aspects not
covered by standard theory. Although originally conceived
separately, the {\tsc{Pythia}} and {\tsc{Jetset}} programs today are so often used
together that it makes sense to present them here without too much
distinction.
Both programs have a long history, and several manuals have
come out. The former round of {\tsc{Pythia/Jetset}} program descriptions appeared
in 1987. Meanwhile a large number of additions and
changes have been made. Recently a new description therefore
appeared in \\[1mm]
T. Sj\"ostrand, Computer Physics Commun. {\bf 82} (1994) 74. \\[1mm]
This is the one and only correct reference to the current versions of
{\tsc{Pythia}} and {\tsc{Jetset}}. The long writeup that you now have before you is
an (unpublished) appendix to the publication above, and need not be
separately cited. Instead remember to cite the original
literature on the physics topics of particular relevance for your
studies. (There is no reason to omit references to good physics
papers simply because some of their contents have also been
made available as program code.)
Event generators often have a reputation for being `black boxes';
if nothing else, this report should provide you with a glimpse of
what goes on inside the programs. Some such understanding may be of
special interest for new users, who have no background in the field.
An attempt has been made to structure the report sufficiently well
that many of the sections can be read independently of each other,
so you can pick the sections that interest you. I have tried to
keep together the physics and the manual sections on specific
topics, where practicable, which represents a change of policy
compared with previous manual versions. Any feedback on this and
other aspects is welcome.
A large number of persons should be thanked for their contributions.
Hans-Uno Bengtsson is the originator of the {\tsc{Pythia}} program, and for
many years we worked in parallel on its further development. Mats
Bengtsson is the main author of the final-state parton-shower
algorithm. Bo Andersson and G\"osta Gustafson are the originators
of the Lund model, and strongly influenced the early development of
the programs. Further comments on the programs have been obtained
from users too numerous to be mentioned here, but who are all
gratefully acknowledged. To write programs of this size and
complexity would be impossible without a strong user feedback.
The moral responsibility for any remaining errors clearly rests with
me. However, kindly note that this is a `University World' product,
distributed `as is', free of charge, without any binding guarantees.
And always remember that the programs do not represent a dead
collection of established truths, but rather one of many possible
approaches to the problem of multiparticle production in high-energy
physics, at the frontline of current research. Be critical!
\newpage
\mbox{}
\cleardoublepage
\tableofcontents
\cleardoublepage
\pagestyle{plain}
\setcounter{page}{1}
\section{Introduction}
Multiparticle production is the most characteristic feature
of current high-energy physics. Today, observed particle
multiplicities are typically between ten and a hundred, and
with future machines this range will be extended upwards.
The bulk of the multiplicity is found in jets, i.e. in bunches
of hadrons (or decay products of hadrons) produced by the
hadronization of quarks and gluons.
\subsubsection*{The Complexity of High-Energy Processes}
To first approximation, all processes have a simple structure at
the level of interactions between the fundamental objects of nature,
i.e. quarks, leptons and gauge bosons. For instance, a lot can
be understood about the structure of hadronic events at LEP just
from the `skeleton' process $\e^+\e^- \to \mathrm{Z}^0 \to \mathrm{q} \overline{\mathrm{q}}$.
Corrections to this picture can be subdivided, arbitrarily but
conveniently, into three main classes.
Firstly, there are bremsstrahlung-type modifications, i.e. the
emission of additional final-state particles by branchings such as
$\mathrm{e} \to \mathrm{e} \gamma$ or $\mathrm{q} \to \mathrm{q} \mathrm{g}$. Because of the largeness of
the strong coupling constant $\alpha_{\mathrm{s}}$, and because of the presence
of the triple gluon vertex, QCD emission off quarks and gluons is
especially prolific. We therefore speak about `parton showers',
wherein a single initial parton may give rise to a whole bunch of
partons in the final state. Also photon emission may give sizeable
effects in $\e^+\e^-$ and $\e\p$ processes. The bulk of the bremsstrahlung
corrections are universal, i.e. do not depend on the details of
the process studied, but only on one or a few key numbers, such as
the momentum transfer scale of the process. Such universal
corrections may be included to arbitrarily high orders, using a
probabilistic language. Alternatively, exact calculations of
bremsstrahlung corrections may be carried out order by order in
perturbation theory, but rapidly the calculations then become
prohibitively complicated and the answers correspondingly lengthy.
Secondly, we have `true' higher-order corrections, which involve a
combination of loop graphs and the soft parts of the
bremsstrahlung graphs above, a combination needed to
cancel some divergences. In a complete description it is
therefore not possible to consider bremsstrahlung separately,
as assumed here. The
necessary perturbative calculations are usually very difficult;
only rarely have results been presented that include more than one
non-`trivial' order, i.e. more than one loop. As above, answers
are usually very lengthy, but some results are sufficiently simple
to be generally known and used, such as the running of $\alpha_{\mathrm{s}}$, or
the correction factor $1 + \alpha_{\mathrm{s}}/\pi + \cdots$ in the partial
widths of $\mathrm{Z}^0 \to \mathrm{q} \overline{\mathrm{q}}$ decay channels. For high-precision
studies it is imperative to take into account the results of
loop calculations, but usually effects are minor for the qualitative
aspects of high-energy processes.
Thirdly, quarks and gluons are confined. In the two points above,
we have used a perturbative language to describe the short-distance
interactions of quarks, leptons and gauge bosons. For leptons
and colourless bosons this language is sufficient. However, for
quarks and gluons it must be complemented with a picture for the
hadronization process (which can be subdivided into fragmentation
and decays), wherein the coloured partons are transformed
into jets of colourless hadrons, photons and leptons.
This process is still not yet understood
from first principles, but has to be based on models. In one sense,
hadronization effects are overwhelmingly large, since this is where
the bulk of the multiplicity comes from. In another sense, the
overall energy flow of a high-energy event is mainly determined by
the perturbative processes, with only a minor additional smearing
caused by the hadronization step. One may therefore pick different
levels of ambition, but in general detailed studies require a
detailed modelling of the hadronization process.
The simple structure that we started out with has now become
considerably more complex --- instead of maybe two final-state
partons we have a hundred final particles. The original physics
is not gone, but the skeleton process has been dressed up and
is no longer directly visible. A direct comparison between theory
and experiment is therefore complicated at best, and
impossible at worst.
\subsubsection*{Event Generators}
It is here that event generators come to the rescue. In an event
generator, the objective strived for is to use computers to generate
events as detailed as could be observed by a perfect detector.
This is not done in one step, but rather by `factorizing' the full
problem into a number of components, each of which can be handled
reasonably accurately. Basically, this means that the hard process
is used as input to generate bremsstrahlung corrections, and that
the result of this exercise is thereafter left to hadronize. This
sounds a bit easier than it really is --- else this report would
be a lot thinner. However, the basic idea is there: if the
full problem is too complicated to be solved in one go, try to
subdivide it into smaller tasks of manageable proportions.
In the actual generation procedure, most steps therefore involve
the branching of one object into two, or at least into a very small
number, each of which being free to branch in its turn. A lot of
bookkeeping is involved, but much is of a repetitive nature, and
can therefore be left for the computer to handle.
As the name indicates, the output of an event generator should be
in the form of `events', with the same average behaviour and the
same fluctuations as real data. In the data, fluctuations arise from
the quantum mechanics of the underlying theory. In
generators, Monte Carlo techniques are used to select all relevant
variables according to the desired probability distributions,
and thereby ensure randomness in the final events.
Clearly some loss of information is entailed: quantum mechanics is
based on amplitudes, not probabilities. However, only very rarely
do (known) interference phenomena appear that cannot be cast in a
probabilistic language. This is therefore not a more restraining
approximation than many others.
Once there, an event generator can be used in many different ways.
The five main applications are probably the following:
\begin{Itemize}
\item To give physicists a feeling for the kind
of events one may expect/hope to find, and at what rates.
\item As a help in the planning of a new detector, so that detector
performance is optimized, within other constraints, for the
study of interesting physics scenarios.
\item As a tool for devising the analysis strategies that should
be used on real data, so that signal-to-background conditions are
optimized.
\item As a method for estimating detector acceptance corrections
that have to be applied to raw data, in order to extract the
`true' physics signal.
\item As a convenient framework within which to interpret the
observed phenomena in terms of a more fundamental
underlying theory (usually the Standard Model).
\end{Itemize}
Where does a generator fit into the overall analysis chain of an
experiment? In `real life', the machine produces interactions.
These events are observed by detectors, and the interesting ones
are written to tape by the
data acquisition system. Afterwards the events may be reconstructed,
i.e. the electronics signals (from wire chambers, calorimeters, and
all the rest) may be
translated into a deduced setup of charged tracks or
neutral energy depositions, in the best of worlds with full knowledge
of momenta and particle species. Based on this cleaned-up
information, one may proceed with the physics analysis.
In the Monte Carlo world, the r\^ole of the machine, namely to produce
events, is taken by the event generators described in this report.
The behaviour of the detectors --- how particles produced by the
event generator traverse the detector, spiral in magnetic
fields, shower in calorimeters, or sneak out through cracks, etc. ---
is simulated in programs such as \tsc{Geant} \cite{Bru89}.
Traditionally, this latter activity is called event simulation,
which is somewhat unfortunate
since the same words could equally well be applied to what, here, we
call event generation. A more appropriate term is detector
simulation. Ideally, the output of this simulation has exactly the
same format as the real data recorded by the detector, and can
therefore be put through the same event reconstruction and physics
analysis chain, except that here we know what the `right answer'
should be, and so can see how well we are doing.
Since the full chain of detector simulation and event
reconstruction is very
time-consuming, one often does `quick and dirty' studies in
which these steps are skipped entirely, or at least replaced by
very simplified procedures which only take into account the geometric
acceptance of the detector and other trivial effects. One may then
use the output of the event generator directly in the physics studies.
There are still many holes in our understanding of the full event
structure, despite an impressive amount of work and detailed
calculations. To put together a generator therefore involved making
a choice on what to include, and how to include it. At best, the
spread between generators can be used to give some impression of
the uncertainties involved. A multitude of approximations will
be discussed in the main part of this report, but already here
is should be noted that many major approximations are related to
the almost complete neglect
of the second point above, i.e. of the non-`trivial'
higher-order effects. It can therefore only be hoped that
the `trivial' higher order parts give the bulk of the experimental
behaviour. By and large, this seems to be the case; for $\e^+\e^-$
annihilation it even turns out to be a very good approximation.
The necessity to make compromises has one major implication:
to write a good event generator is an art, not an exact science.
It is therefore essential not to blindly trust the
results of any single event generator, but always to make several
cross-checks. In addition, with computer programs of tens of
thousands of lines, the question is not whether bugs exist, but how
many there are, and how critical their positions.
Further, an event generator cannot be thought of as all-powerful,
or able to give intelligent answers to ill-posed questions;
sound judgement and some understanding of a
generator are necessary prerequisites for successful use. In spite
of these limitations, the event generator approach is the most
powerful tool at our disposal if we wish to gain a detailed and
realistic understanding of physics at current or future high-energy
colliders.
\subsubsection*{The Origins of the JETSET and PYTHIA Programs}
Over the years, many event generators have appeared. Surveys of
generators for $\e^+\e^-$ physics in general and LEP in particular
may be found in \cite{Kle89,Sjo89}, for high-energy hadron--hadron
($\p\p$) physics in \cite{Ans90,Sjo92,Kno93}, and for $\e\p$ physics in
\cite{HER92}. We refer the reader to those for additional details
and references. In this particular report, the two closely
connected programs {\tsc{Jetset}} and {\tsc{Pythia}} will be described.
{\tsc{Jetset}} has its roots in the efforts of the Lund
group to understand the hadronization process, starting in the late
seventies \cite{And83}. The so-called string fragmentation model
was developed as an explicit and detailed framework, within which
the long-range confinement forces are allowed to distribute the
energies and flavours of a parton configuration among a collection
of primary hadrons, which subsequently may decay further. This model,
known as the Lund string model, or `Lund' for short, contained a
number of specific predictions, which were confirmed by data from
PETRA and PEP, whence the model gained a widespread
acceptance. The Lund string model is still today the most elaborate
and widely used fragmentation model
at our disposal. It remains at the heart of the
{\tsc{Jetset/Pythia}} programs.
In order to predict the shape of events at PETRA/PEP, and to
study the fragmentation process in detail, it was necessary to start
out from the partonic configurations that were to fragment.
The generation of complete $\e^+\e^-$ hadronic events was therefore
added, originally based on simple $\gamma$ exchange and
first-order QCD matrix elements, later extended to full $\gamma^* / \Z^0$
exchange with first-order initial-state QED radiation and
second-order QCD matrix elements. A number of utility routines
were also provided early on, for everything from event listing
to jet finding.
By the mid-eighties it was clear that the matrix-element approach had
reached the limit of its usefulness, in the sense that it could not
fully describe the multijet topologies of the data. (Later on,
the use of optimized perturbation theory was to lead to a resurgence
of the matrix-element approach, but only for specific applications.)
Therefore a parton-shower description was
developed \cite{Ben87a} as an alternative to the matrix-element one.
The combination of parton showers and string fragmentation has been
very successful, and forms the main approach to the description of
hadronic $\mathrm{Z}^0$ events.
In recent years, {\tsc{Jetset}} has been a fairly stable product,
covering the four main areas of fragmentation, final-state parton
showers, $\e^+\e^-$ event generation and general utilities.
The successes of string fragmentation in $\e^+\e^-$ made it interesting
to try to extend this framework to other processes, and explore
possible physics consequences. Therefore a number of
other programs were written, which combined a process-specific
description of the hard interactions with the general fragmentation
framework of {\tsc{Jetset}}. The {\tsc{Pythia}} program
evolved out of early studies on fixed-target proton--proton
processes, addressed mainly at issues related to string drawing.
With time, the interest shifted towards higher energies, first
to the SPS $\p\pbar$ collider, and later to SSC and LHC, in the
context of a number of workshops in the USA and Europe. Parton
showers were added, for final-state radiation by making use of
the {\tsc{Jetset}} routine, for initial-state one by the development
of the concept of `backwards evolution', specifically for
{\tsc{Pythia}} \cite{Sjo85}. Also a framework was developed for
minimum-bias and underlying events \cite{Sjo87a}.
Another main change was the introduction of an increasing
number of hard processes, within the Standard Model and beyond.
A special emphasis was put on the search for the Standard Model
Higgs, in different mass ranges and in different channels, with due
respect to possible background processes.
The bulk of the machinery developed for hard processes actually
depended little on the choice of initial state, as long as the
appropriate parton distributions were there for the incoming
partons and particles. It therefore made sense to extend the
program from being only a $\p\p$ generator to working also for
$\e^+\e^-$ and $\e\p$. This process was only completed in 1991,
again spurred on by physics workshop activities. Currently
{\tsc{Pythia}} should therefore work equally well for a selection
of different possible incoming beam particles.
The tasks of including new processes, and of improving the simulation
of already present ones, are never-ending. Work therefore continues
apace.
While {\tsc{Jetset}} still is formally independent of {\tsc{Pythia}}, their ties
have grown much stronger over the years, and the border-line
between the two programs has become more and more artificial.
It is no coincidence that the two are presented together here;
this way a lot of repetition of common material can be avoided.
The price to be paid is that some differences in philosophy will
have to be discussed.
\subsubsection*{About this Report}
As we see, {\tsc{Jetset}} and {\tsc{Pythia}} started out as very
ideologically motivated programs, developed to study specific
physics questions in enough detail that explicit predictions
could be made for experimental quantities. As it was recognized
that experimental imperfections could distort the basic predictions,
the programs were made available for general use by experimentalists.
It thus became feasible to explore the models in more detail
than would otherwise have been possible. As time went by, the
emphasis came to shift somewhat, away from the original strong
coupling to a specific fragmentation model, towards a description of
high-energy multiparticle production processes in general.
Correspondingly, the use expanded from being one of just comparing
data with specific model predictions, to one of extensive use for
the understanding of detector performance, for the derivation of
acceptance correction factors, for the prediction of physics at
future high-energy accelerators, and for the design of related
detectors.
While the ideology may be less apparent, it is still there, however.
This is not something unique to the programs discussed here,
but inherent in any event generator, or at least any generator that
attempts to go beyond the simple parton level skeleton description
of a hard process. Do not accept the myth that everything available
in Monte Carlo form represents ages-old common knowledge, tested
and true. Ideology is present by commissions or omissions
in any number of details. Programs like {\tsc{Pythia}} and {\tsc{Jetset}} represent
a major amount of original physics research, often on complicated
topics where no simple answers are available.
As a (potential) program user you must be
aware of this, so that you can form your own opinion, not just about
what to trust and what not to trust, but also how much to trust a given
prediction, i.e. how uncertain it is likely to be. {\tsc{Jetset}} and
{\tsc{Pythia}} are particularly well endowed in this respect, since a
number of publications exist where most of the relevant physics is
explained in considerable detail. In fact, the problem may rather be
the opposite, to find the relevant information among all the possible
places. One main objective of the current report is therefore to
collect much of this information in one single place. Not all the
material found in specialized papers is reproduced, by a wide margin,
but at least enough should be found here to understand the general
picture and to know where to go for details.
The current report is therefore intended to replace the previous round
of published physics descriptions and program manuals
\cite{Sjo86,Sjo87,Ben87}. The formal new standard reference is
\cite{Sjo94}, which is a fairly brief summary of this report
--- for obvious reasons the full description is too long to be published
in its entirety. Further specification could include a statement of
the type `We use {\tsc{Pythia}} version X.x and {\tsc{Jetset}} version Y.y'.
(If you are a {\LaTeX} fan, you may want to know that the program
names in this report have been generated by the commands
\verb+\textsc{Jetset}+ and \verb+\textsc{Pythia}+.)
Kindly do not refer to {\tsc{Jetset/Pythia}} as
`unpublished', `private communication' or `in preparation':
such phrases are only creating unnecessary confusion.
In addition, remember that many of the individual physics components
are documented in separate publications. If some of these contain
ideas that are useful to you, there is every reason to cite them.
A reasonable selection would vary as a function of the physics you are
studying. The criterion for which to pick should be simple: imagine
that a Monte Carlo implementation had not been available. Would you
then have cited a given paper on the grounds of its physics contents
alone? If so, do not punish the extra effort of turning these ideas
into publicly available software. (Monte Carlo manuals are good for
nothing in the eyes of many theorists, so often only the acceptance
of `mainstream' publications counts.) Here follows a list of some
main areas where the $\tsc{Pythia/Jetset}$ programs contain original research:
\begin{Itemize}
\item The string fragmentation model \cite{And83}.
\item The string effect \cite{And80}.
\item Baryon production (diquark/popcorn) \cite{And82,And85}.
\item Fragmentation of multiparton systems \cite{Sjo84}.
\item Fragmentation effects on $\alpha_{\mathrm{s}}$ determinations \cite{Sjo84a}.
\item Initial state parton showers \cite{Sjo85}.
\item Final state parton showers \cite{Ben87a}.
\item Photon radiation from quarks \cite{Sjo92c}
\item Deep inelastic scattering \cite{And81a,Ben88}.
\item Photoproduction \cite{Sch93a} and $\gamma\gamma$ physics
\cite{Sch94a}.
\item Parton distributions of the photon \cite{Sch95}.
\item Colour flow in hard scatterings \cite{Ben84}.
\item Elastic and diffractive cross sections \cite{Sch94}.
\item Minijets (multiple parton--parton interactions) \cite{Sjo87a}.
\item Rapidity gaps \cite{Dok92}.
\item Jet clustering in $k_{\perp}$ \cite{Sjo83}.
\end{Itemize}
In addition to a physics survey, the current report also contains a
complete manual for the two programs. Such manuals have always been
updated and distributed jointly with the programs. To a first
approximation, we therefore do not have much new to offer here.
However, an attempt has been made to group the material more logically
according to physics topics than in previous distributions,
to tie it closer to the physics description, and to improve the layout
and therefore the readability. Any feedback is welcome.
A word of warning may be in place. The program description is fairly
lengthy, and certainly could not be absorbed in one sitting. This is
not even necessary, since all switches and parameters are provided
with sensible default values, based on our best understanding (of
the physics, and of what you expect to happen if you do not
specify any options). As a new user, you can therefore disregard all
the fancy options, and just run the program with a minimum ado.
Later on, as you gain experience, the options that seem useful can
be tried out. No single user is ever likely to find need for more than
a fraction of the total number of possibilities available,
yet many of them have been added to meet specific user requests.
In some instances, not even this report will provide you with all the
information you desire. You may wish to find out about recent versions
of the program, know about related software, pick up a few sample
main programs to get going, or get hold of related physics papers.
Some such material can be found if you link to my World Wide Web
homepage:\\
\ttt{http://thep.lu.se/tf2/staff/torbjorn/Welcome.html}\\
and study the contents there.
\subsubsection*{Disclaimer}
At all times it should be remembered that this is not a commercial
product, developed and supported by professionals. Instead it is
a `University World' product, developed by a very few physicists
(mainly the current author)
originally for their own needs, and supplied to other
physicists on an `as-is' basis, free of charge. No guarantees are
therefore given for the proper functioning of the programs, nor for
the validity of physics results. In the end, it is always up to you
to decide for yourself whether to trust a given result or not. Usually
this requires comparison either with analytical results or with
results of other programs, or with both. Even this is not necessarily
foolproof: for instance, if an error
is made in the calculation of a matrix element for a given process,
this error will be propagated both into the analytical results based
on the original calculation and into all the event generators which
subsequently make use of the published formulae. In the end, there
is no substitute for a sound physics judgement.
This does not mean that you are all on your own, with a program
nobody feels responsible for. Attempts are made to check processes as
carefully as possible, to write programs that do not invite
unnecessary errors, and to provide a detailed and accurate
documentation. All of this while maintaining the full power and
flexibility, of course, since the physics must always take precedence
in any conflict of interests. If nevertheless any errors or
unclarities are found, please do communicate them to me,
e.g. on phone +46 -- 46 -- 222 48 16 or e-mail <EMAIL>.
Every attempt will be made to solve problems as soon as is
reasonably possible, given that this support is by one person
alone, who also has other responsibilities.
\subsubsection*{Appendix: The Historical Pythia}
While the origin and connotations of the `{\tsc{Jetset}}' program name
should be commonly known, the `{\tsc{Pythia}}' label may need some
explanation.
The myth tells how Apollon, the God of Wisdom, killed the powerful
dragon-like monster Python, close to the village of Delphi in Greece.
To commemorate this victory, Apollon founded the Pythic Oracle in
Delphi, on the slopes of Mount Parnassos. Here men could come to
learn the will of the Gods and the course of the future. The oracle
plays an important r\^ole in many of the other Greek myths, such
as those of Heracles and of King Oedipus.
Questions were to be put to the Pythia, the `Priestess' or
`Prophetess' of the Oracle. In fact, she was a local woman,
usually a young maiden, of no particular religious schooling.
Seated on a tripod, she inhaled the obnoxious vapours that
seeped up through a crevice in the ground. This brought her
to a trance-like state, in which she would scream seemingly
random words and sounds. It was the task of the professional
priests in Delphi to record those utterings and edit them into
the official Oracle prophecies, which often took the form of
poems in perfect hexameter. In fact, even these edited replies
were often less than easy to interpret. The Pythic oracle
acquired a reputation for ambiguous answers.
The Oracle existed already at the beginning of the historical
era in Greece, and was universally recognized as the foremost
religious seat. Individuals and city states came to consult, on
everything from cures for childlessness to matters of war. Lavish
gifts allowed the temple area to be built and decorated. Many
states supplied their own treasury halls, where especially beautiful
gifts were on display. Sideshows included the Omphalos,
a stone reputedly marking the centre of the Earth, and the Pythic
games, second only to the Olympic ones in importance.
Strife inside Greece eventually led to a decline in the power of
the Oracle. A serious blow was dealt when the Oracle of Zeus Ammon
(see below) declared Alexander the Great
to be the son of Zeus. The Pythic Oracle lived on, however, and was
only closed by a Roman Imperial decree in 390 \tsc{ad}, at a time
when Christianity was ruthlessly destroying any religious opposition.
Pythia then had been at the service of man and Gods for a
millenium and a half.
The r\^ole of the Pythic Oracle replies on the course of history is
nowhere better described than in `The Histories' by Herodotus
\cite{HerBC}, the classical and captivating description of the
Ancient World at the time of the Great War between Greeks and
Persians. Especially famous is the episode with King Croisus
of Lydia. Contemplating a war against the upstart Persian
Empire, he resolves to ask an oracle what the outcome of a potential
battle would be. However, to have some guarantee for the
veracity of any prophecy, he decides to send embassies to all the
renowned oracles of the known World. The messengers are instructed
to inquire the various divinities, on the hundredth day after
their departure, what King Croisus is doing at that very moment.
From the Pythia the messengers bring back the reply
\begin{em}\begin{verse}
I know the number of grains of sand as well as the expanse of
the sea, \\
And I comprehend the dumb and hear him who does not speak, \\
There came to my mind the smell of the hard-shelled turtle, \\
Boiled in copper together with the lamb, \\
With copper below and copper above.
\end{verse}\end{em}
The veracity of the Pythia is thus established by the crafty ruler,
who had waited until the appointed day, slaughtered a turtle and a
lamb, and boiled them together in a copper cauldron with a
copper lid. Also the Oracle of Zeus Ammon in the Libyan desert
is able to give a correct reply (lost to posterity), while all
others fail. King Croisus now sends a second embassy to Delphi,
inquiring after the outcome of a battle against the Persians.
The Pythia answers
\begin{em}\begin{verse}
If Croisus passes over the Halys he will dissolve a great Empire.
\end{verse}\end{em}
Taking this to mean he would win, the King collects his army and
crosses the border river, only to suffer a crushing defeat and
see his Kingdom conquered. When the victorious King Cyrus allows
Croisus to send an embassy to upbraid the Oracle, the God Apollon
answers through his Prophetess that he has correctly predicted the
destruction of a great empire --- Croisus' own --- and that he
cannot be held responsible if people choose to interpret the
Oracle answers to their own liking.
The history of the {\tsc{Pythia}} program is neither as long nor as
dignified as that of its eponym. However, some points of contact
exist. You must be very careful when you formulate the questions:
any ambiguities will corrupt the reply you get. And you must be even
more careful not to misinterpret the answers; in particular not to pick
the interpretation that suits you before considering the alternatives.
Finally, even a perfect God has servants that are only human: a priest
might mishear the screams of the Pythia and therefore produce an
erroneous oracle reply; the current author might unwittingly let a
bug free in the program {\tsc{Pythia}}.
\clearpage
\section{Physics Overview}
In this section we will try to give an overview of the main physics
features of {\tsc{Jetset}} and {\tsc{Pythia}}, and also to introduce some
terminology. The details will be discussed in subsequent sections.
For the description of a typical high-energy event, an event
generator should contain a simulation of several physics aspects.
If we try to follow the evolution of an event in some semblance of
a time order, one may arrange these aspects as follows:
\begin{Enumerate}
\item Initially two beam particles are coming in towards each other.
Normally each particle is characterized by a set of parton
distribution functions, which defines the partonic substructure
in terms of flavour composition and energy sharing.
\item One shower initiator parton from each beam starts off
a sequence of branchings, such as $\mathrm{q} \to \mathrm{q} \mathrm{g}$, which build up
an initial-state shower.
\item One incoming parton from each of the two showers
enters the hard process, where then a number of
outgoing partons are produced, usually two.
It is the nature of this process that determines the main
characteristics of the event.
\item Also the outgoing partons may branch, to build up
final-state showers.
\item When a shower initiator is taken out of a beam particle,
a beam remnant is left behind. This remnant may have
an internal structure, and a net colour charge that relates
it to the rest of the final state.
\item The QCD confinement mechanism ensures that the outgoing quarks
and gluons are not observable, but instead fragment to colour
neutral hadrons.
\item Many of the produced hadrons are unstable and decay further.
\end{Enumerate}
Conventionally, only quarks and gluons are counted as partons, while
leptons and photons are not. If pushed {\it ad absurdum} this may
lead to
some unwieldy terminology. We will therefore, where it does not matter,
speak of an electron or a photon in the `partonic' substructure of an
electron, lump branchings $\mathrm{e} \to \mathrm{e} \gamma$ together with other
`parton shower' branchings such as $\mathrm{q} \to \mathrm{q} \mathrm{g}$, and so on. With
this notation, the division into the above seven points applies equally
well to an interaction between two leptons, between a lepton and a
hadron, and between two hadrons.
In the following subsections, we will survey the above seven aspects,
not in the same order as given here, but rather in the order in
which they appear in the program execution, i.e. starting with the
hard process.
\subsection{Hard Processes and Parton Distributions}
In {\tsc{Jetset}}, only two hard processes are available. The first and
main one is $\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q} \overline{\mathrm{q}}$. Here the `$*$' of
$\gamma^*$ is used to denote that the photon must be off the mass shell.
The distinction is of some importance, since a photon on the mass shell
cannot decay. Of course also the $\mathrm{Z}^0$ can be off the mass shell,
but here the distinction is less relevant (strictly speaking,
a $\mathrm{Z}^0$ is always off the mass shell). In the following we may not
always use `$*$' consistently, but the rule of thumb is to use a `$*$'
only when a process is not kinematically possible for a particle of
nominal mass. The quark $\mathrm{q}$ in the final state of
$\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q} \overline{\mathrm{q}}$
may be $\u$, $\d$, $\mathrm{s}$, $\c$, $\b$ or $\t$; the flavour in
each event is picked at random, according to the relative couplings,
evaluated at the hadronic c.m. energy. Also the angular distribution of
the final $\mathrm{q} \overline{\mathrm{q}}$ pair is included. No parton-distribution functions
are needed.
The other {\tsc{Jetset}} process is a routine to generate $\mathrm{g} \mathrm{g} \mathrm{g}$ and
$\gamma \mathrm{g} \mathrm{g}$ final states, as expected in onium 1$^{--}$ decays
such as $\Upsilon$. Given the current limits on the top mass,
toponium will decay weakly much too fast for these processes to be of
any interest, so therefore no new applications are expected.
{\tsc{Pythia}} contains a much richer selection, with close to a
hundred different hard processes. These may be classified in
many different ways.
One is according to the number of final-state objects: we speak of
`$2 \to 1$' processes, `$2 \to 2$' ones, `$2 \to 3$' ones, etc.
This aspect is very relevant from a programming point of view:
the more particles in the final state, the more complicated the
phase space and therefore the whole generation procedure. In fact,
{\tsc{Pythia}} is optimized for $2 \to 1$ and $2 \to 2$ processes.
There is currently no generic treatment of processes with three or
more particles in the final state, but rather a few different
machineries, each tailored to the pole structure of a specific class of
graphs. This may be seen as a major limitation, and indeed is so
at times. However, often one can come quite far with only one
or two particles in the final state, since showers will add the
required extra activity. The classification may also be misleading
at times, since an $s$-channel resonance is considered as a single
particle, even if it is assumed always to decay into two final-state
particles. Thus the process
$\e^+\e^- \to \mathrm{W}^+ \mathrm{W}^- \to \mathrm{q}_1 \overline{\mathrm{q}}'_1 \, \mathrm{q}_2 \overline{\mathrm{q}}'_2$ is classified
as $2 \to 2$, although the decay treatment of the $\mathrm{W}$ pair includes
the full $2 \to 4$ matrix elements.
Another classification is according to the physics scenario. This will
be the main theme of section \ref{s:pytproc}. The following major
groups may be distinguished:
\begin{Itemize}
\item Hard QCD processes, e.g. $\mathrm{q} \mathrm{g} \to \mathrm{q} \mathrm{g}$.
\item Soft QCD processes, such as diffractive and elastic scattering,
and minimum-bias events.
\item Heavy-flavour production, e.g. $\mathrm{g} \mathrm{g} \to \t \overline{\mathrm{t}}$.
\item Prompt-photon production, e.g. $\mathrm{q} \mathrm{g} \to \mathrm{q} \gamma$.
\item Photon-induced processes, e.g. $\gamma \mathrm{g} \to \mathrm{q} \overline{\mathrm{q}}$.
\item Deep inelastic scattering, e.g. $\mathrm{q} \ell \to \mathrm{q} \ell$.
\item $\mathrm{W} / \mathrm{Z}$ production, such as the $\e^+\e^- \to \gamma^* / \Z^0$ already
found in {\tsc{Jetset}}, or $\mathrm{q} \overline{\mathrm{q}} \to \mathrm{W}^+ \mathrm{W}^-$.
\item Standard model Higgs production, where the Higgs is reasonably
light and narrow, and can therefore still be considered as a resonance.
\item Gauge boson scattering processes, such as $\mathrm{W} \mathrm{W} \to \mathrm{W} \mathrm{W}$,
when the Standard Model Higgs is so heavy and broad that resonant and
non-resonant contributions have to be considered together.
\item Non-standard Higgs particle production, within the framework
of a two-Higgs-doublet scenario with three neutral and two charged
Higgs states.
\item Production of new gauge bosons, such as a $\mathrm{Z}'$.
\item Production of fourth-generation fermions.
\item Leptoquark production.
\item Deviations from Standard Model processes, e.g. due to contact
interactions or a strongly interacting gauge boson sector. These
scenarios do not always appear as separate processes, but may just
be options to some of the processes above.
\end{Itemize}
This is by no means a survey of all interesting physics.
Most notable is the absence of supersymmetric particle production
and decay, but many other examples could be found. Also, within
the scenarios studied, not all contributing graphs have always been
included, but only the more important and/or more interesting ones.
In many cases, various approximations are involved in the matrix
elements coded.
The cross section for a given process $ij \to k$ is given by
\begin{equation}
\sigma_{ij \to k} = \int \d x_1 \int \d x_2 \, f^1_i(x_1) \,
f^2_j(x_2) \, \hat{\sigma}_{ij \to k} ~.
\end{equation}
Here $\hat{\sigma}$ is the cross section for the hard partonic process,
as codified in the matrix elements for each specific process.
For processes with many particles in the final state
it would be replaced by an
integral over the allowed final-state phase space. The $f^a_i(x)$ are
the parton-distribution functions, which describe the probability to
find a parton $i$ inside beam particle $a$, with parton $i$ carrying a
fraction $x$ of the total $a$ momentum. Actually, parton distributions
also depend on some momentum scale $Q^2$ that characterizes the hard
process.
Parton distributions are most familiar for hadrons, such as the proton.
Hadrons are inherently composite objects, made up of quarks and
gluons. Since we do not understand QCD, a derivation from first
principles
of hadron parton distributions does not yet exist, although some
progress is being made in lattice QCD studies. It is therefore
necessary to rely on parametrizations, where experimental data are
used in conjunction with the evolution equations for the $Q^2$
dependence, to pin down the parton distributions. Several different
groups have therefore produced their own fits, based on slightly
different sets of data, and with some variation in the theoretical
assumptions.
Also for fundamental particles, such as the electron, is it convenient
to introduce parton distributions. The function $f^{\mathrm{e}}_{\mathrm{e}}(x)$ thus
parametrizes the probability that the electron that takes part in the
hard process retains a fraction $x$ of the original energy, the rest
being radiated (into photons) in the initial state. Of course, such
radiation could equally well be made part of the hard interaction,
but the parton-distribution approach usually is much more convenient.
If need be, a description with fundamental electrons is recovered for
the choice $f_{\mathrm{e}}^{\mathrm{e}}(x, Q^2)=\delta(x-1)$. Note that, contrary to
the proton case, electron parton distributions are calculable from first
principles, and reduce to the $\delta$ function above for $Q^2 \to 0$.
The electron may also contain photons, and the photon may in its turn
contain quarks and gluons. The internal structure of the
photon is a bit of a problem, since the photon
contains a point-like part, which is perturbatively calculable, and
a vector-meson dominance part, which is not. Normally, the photon
parton distributions are therefore parametrized, just as the hadron
ones. Since the electron ultimately contains quarks and gluons,
hard QCD processes like $\mathrm{q} \mathrm{g} \to \mathrm{q} \mathrm{g}$ therefore not only appear in
$\p\p$ collisions, but also in $\e\p$ ones (`resolved photoproduction')
and in $\e^+\e^-$ ones (`doubly resolved 2$\gamma$ events'). The parton
distribution function approach here makes it much easier to reuse one
and the same hard process in different contexts.
There is also another kind of possible generalization. The two
processes $\mathrm{q} \overline{\mathrm{q}} \to \gamma^* / \Z^0$, studied in hadron colliders,
and $\e^+\e^- \to \gamma^* / \Z^0$, studied in $\e^+\e^-$ colliders, are really
special cases of a common process, $\mathrm{f} \overline{\mathrm{f}} \to \gamma^* / \Z^0$,
where $\mathrm{f}$ denotes a fundamental fermion, i.e. a quark, lepton or
neutrino. The whole structure is therefore only coded once, and
then slightly different couplings and colour prefactors are used,
depending on the initial state considered. Usually the
interesting cross section is a sum over several different initial
states, e.g. $\u \overline{\mathrm{u}} \to \gamma^* / \Z^0$ and
$\d \overline{\mathrm{d}} \to \gamma^* / \Z^0$ in a hadron collider. This kind of
summation is always implicitly done, even when not explicitly
mentioned in the text.
\subsection{Initial- and Final-State Radiation}
In every process that contains coloured and/or charged objects
in the initial or final state, gluon and/or photon radiation
may give large corrections to the overall topology of events.
Starting from a basic $2 \to 2$ process, this kind of corrections
will generate $2 \to 3$, $2 \to 4$, and so on, final-state
topologies. As the available energies are increased, hard
emission of this kind is increasingly important, relative to
fragmentation, in determining the event structure.
Two traditional approaches exist to the modelling of perturbative
corrections. One is the matrix-element method, in
which Feynman diagrams are calculated, order by order. In principle,
this is the correct approach, which takes into account exact
kinematics, and the full interference and helicity structure. The
only problem is that calculations become increasingly difficult in
higher orders, in particular for the loop graphs.
Only in exceptional cases have therefore more than one loop been
calculated in full, and often we do not have any loop corrections
at all at our disposal. On the other hand,
we have indirect but strong evidence that, in fact, the emission
of multiple soft gluons plays a significant r\^ole in building up the
event structure, e.g. at LEP, and this sets a limit to the
applicability of matrix elements.
Since the phase space available for gluon emission increases with
the available energy, the matrix-element approach becomes less
relevant for the full structure of events at higher energies.
However, the perturbative expansion by itself is better
behaved at higher energies, owing to the running of $\alpha_{\mathrm{s}}$.
As a consequence, inclusive measurements, e.g. of the rate of
well-separated jets, should yield more reliable results.
The second possible approach is the parton-shower one. Here an
arbitrary number of branchings of one parton into two (or more)
may be put together, to yield a description of multijet events,
with no explicit upper limit on the number of partons involved.
This is possible since the full matrix-element expressions are
not used, but only approximations derived by simplifying the
kinematics, and the interference and helicity structure. Parton showers
are therefore expected to give a good description of the substructure of
jets, but in principle the shower approach has limited predictive power
for the rate of well-separated jets (i.e. the 2/3/4/5-jet composition).
In practice, shower programs may be patched up to describe the
hard-gluon emission region reasonably well, in particular for the
$\e^+\e^-$ annihilation process.
Nevertheless, the shower description is not optimal for
absolute $\alpha_{\mathrm{s}}$ determinations.
Thus the two approaches are complementary in many respects,
and both have found use. However, because of its simplicity
and flexibility, the parton-shower option is generally the
first choice, while the matrix elements one is mainly used for
$\alpha_{\mathrm{s}}$ determinations, angular distribution of jets,
triple-gluon vertex studies, and other specialized studies.
Obviously, the ultimate goal would be to have an approach where the
best aspects of the two worlds are harmoniously married.
\subsubsection{Matrix elements}
Matrix elements are especially made use of in the {\tsc{Jetset}}
implementation of the process $\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q} \overline{\mathrm{q}}$.
For initial-state QED radiation, a first order (unexponentiated)
description has been adopted. This means that events are subdivided
into two classes, those where a photon is radiated above some
minimum energy, and those without such a photon. In the latter
class, the soft and virtual corrections have been lumped together
to give a total event rate that is correct up to one loop. This
approach worked fine at PETRA/PEP energies, but does not do so well
for the $\mathrm{Z}^0$ line shape, i.e. in regions where the cross section
is rapidly varying and high precision is strived for.
For final-state QCD radiation, several options are available. The
default is the parton-shower one (see below), but the matrix-elements
options are also frequently used. In the definition of 3- or
4-jet events, a cut is introduced whereby it is required that
any two partons have an invariant mass bigger than some fraction
of the c.m. energy. 3-jet events which do not fulfill this
requirement are lumped with the 2-jet ones. The first-order
matrix-element option, which only contains 3- and 2-jet events
therefore involves no ambiguities. In second order, where also
4-jets have to be considered, a main issue is what to do with
4-jet events that fail the cuts. Depending on the choice of
recombination scheme, whereby the two nearby partons are joined
into one, different 3-jet events are produced. Therefore the
second-order differential 3-jet rate has been the subject of
some controversy, and {\tsc{Jetset}} actually contains two
different implementations.
By contrast, {\tsc{Pythia}} does not contain any full higher-order
matrix elements, with loop contributions included. There are a few
cases where higher-order matrix elements are included at the Born
level. Consider e.g. the case of $\mathrm{W}$ production at a hadron collider,
which is contained in the lowest-order process $\mathrm{q} \overline{\mathrm{q}}' \to \mathrm{W}$.
In an inclusive description, additional jets recoiling against the
$\mathrm{W}$ may be generated by parton showers. {\tsc{Pythia}} also contains
the two first-order processes $\mathrm{q} \mathrm{g} \to \mathrm{W} \mathrm{q}'$ and
$\mathrm{q} \overline{\mathrm{q}}' \to \mathrm{W} \mathrm{g}$. The cross sections for these processes
are divergent when the $p_{\perp} \to 0$. In this region a correct
treatment would therefore have to take into account loop corrections,
which are not available in {\tsc{Pythia}}. Depending on the physics
application, one could then use {\tsc{Pythia}} in one of two ways.
In the region of small $p_{\perp}$, the preferred option is lowest-order
matrix elements combined with parton showers. For the
production of a $\mathrm{W}$ at large $p_{\perp}$, on the other hand, the
shower approach is too imprecise to give the right cross section;
additionally the event selection machinery is very inefficient.
Here it is advantageous to generate first-order events, and
then add showers only to describe additional softer radiation.
\subsubsection{Parton showers}
The separation of radiation into initial- and final-state showers is
arbitrary, but very convenient. There are also situations where it
is appropriate: for instance, the process
$\e^+\e^- \to \mathrm{Z}^0 \to \mathrm{q} \overline{\mathrm{q}}$ only
contains final-state QCD radiation (QED radiation, however, is
possible both in the initial and final state), while
$\mathrm{q} \overline{\mathrm{q}} \to \mathrm{Z}^0 \to \e^+\e^-$ only contains initial-state QCD one.
Similarly, the distinction of emission as coming either from the
$\mathrm{q}$ or from the $\overline{\mathrm{q}}$ is arbitrary. In general, the assignment
of radiation to a given mother parton is a good approximation
for an emission close to the direction of motion of that parton,
but not for the wide-angle emission in between two jets, where
interference terms are expected to be important.
In both initial- and final-state showers, the structure is given in
terms of branchings $a \to bc$, specifically $\mathrm{e} \to \mathrm{e} \gamma$,
$\mathrm{q} \to \mathrm{q} \mathrm{g}$, $\mathrm{q} \to \mathrm{q} \gamma$, $\mathrm{g} \to \mathrm{g} \mathrm{g}$,
and $\mathrm{g} \to \mathrm{q} \overline{\mathrm{q}}$. Each of these processes is characterized by
a splitting kernel $P_{a \to bc}(z)$. The branching rate is
proportional to the integral $\int P_{a \to bc}(z) \, \d z$.
The $z$ value picked for a branching describes the energy sharing,
with daughter $b$ taking a fraction $z$ and daughter $c$ the
remaining $1-z$ of the $a$ energy. Once formed, the daughters
$b$ and $c$ may in turn branch, and so on.
Each parton is characterized by some virtuality scale $Q^2$,
which gives an approximate sense of time ordering
to the cascade. In the initial-state shower, $Q^2$ values are
gradually increasing as the hard scattering is approached, while
$Q^2$ is decreasing in the final-state showers.
Shower evolution is cut off at some lower
scale $Q_0$, typically around 1 GeV for QCD branchings. The same
cut-off scale is also used to regularize the soft gluon emission
divergences in the splitting kernels. From above,
a maximum scale $Q_{\mathrm{max}}$ is introduced, where the showers are
matched to the hard interaction itself. The relation between
$Q_{\mathrm{max}}$ and the kinematics of the hard scattering
is uncertain, and the choice made can strongly affect the
amount of well-separated jets.
Despite a number of common traits, the initial- and final-state
radiation machineries are in fact quite different, and are described
separately below. For historical reasons, the final-state shower
is found in {\tsc{Jetset}} and the initial-state one in {\tsc{Pythia}}.
Final-state showers are time-like,
i.e. partons have $m^2 = E^2 - \mbf{p}^2 \geq 0$. The evolution
variable $Q^2$ of the cascade is therefore in {\tsc{Jetset}}
associated with the $m^2$ of the
branching parton, but this choice is not unique.
Starting from $Q^2_{\mathrm{max}}$, an original parton is evolved
downwards in $Q^2$ until a branching occurs. The selected
$Q^2$ value defines the mass of the branching parton, and the $z$
of the splitting kernel the parton energy division between its
daughters. These daughters may now, in turn, evolve
downwards, in this case with maximum virtuality already defined by
kinematics, and so on down to the $Q_0$ cut-off.
In QCD showers, corrections to the leading-log picture, so-called
coherence effects, lead to an ordering of subsequent emissions in terms
of decreasing angles. This does not follow automatically from the
mass-ordering constraint, but is implemented as an additional
requirement on allowed emissions. Photon emission is not affected
by angular ordering. It is also possible to obtain
non-trivial correlations between azimuthal angles in the various
branchings, some of which are implemented as
options. Finally, the theoretical analysis strongly
suggests the scale choice $\alpha_{\mathrm{s}} = \alpha_{\mathrm{s}}(p_{\perp}^2) =
\alpha_{\mathrm{s}}(z(1-z)m^2)$, and this is the default in the program.
The final-state radiation machinery is applied in the c.m. frame of the
hard scattering. The total energy and momentum of the hard-scattering
subsystem is preserved, as is the direction of the outgoing partons
(in that frame).
In contrast to final-state showers, initial-state ones are space-like.
This means that, in the sequence of branchings $a \to bc$ that lead
up from the shower initiator to the hard interaction,
particles $a$ and $b$ have $m^2 = E^2 - \mbf{p}^2 <0$.
The `side branch' particle $c$, which does not participate
in the hard scattering, may be on the mass shell, or have a time-like
virtuality. In the latter case a time-like shower will evolve off
it, rather like the final-state radiation described above. To first
approximation, the evolution of the space-like main branch
is characterized by the
evolution variable $Q^2 = -m^2$, which is required to be strictly
increasing along the shower, i.e. $Q_b^2 > Q_a^2$. Corrections
to this picture have been calculated,
but are basically absent in {\tsc{Pythia}}.
Initial-state radiation is handled within the backwards evolution
scheme. In this approach, the
choice of the hard scattering is based on the use of evolved
parton distributions, which means that
the inclusive effects of initial-state radiation are already
included. What remains is therefore to construct the exclusive
showers. This is done starting from the two incoming partons
at the hard interaction, tracing the showers `backwards in time',
back to the two shower initiators. In other words,
given a parton $b$, one tries to find the parton $a$ that branched
into $b$. The evolution in the Monte Carlo is therefore in
terms of a sequence of decreasing space-like virtualities $Q^2$
and increasing momentum fractions $x$. Branchings on
the two sides are interleaved in a common sequence of
decreasing $Q^2$ values.
In the above formalism, there is no real distinction between
gluon and photon emission. Some of the details actually do differ,
as will be explained in the full description.
The initial- and final-state radiation shifts around the kinematics of
the original hard interaction. In deep inelastic scattering, this
means that the $x$ and $Q^2$ values that can be derived from the
momentum of the scattered lepton do not agree with the values
originally picked. In high-$p_{\perp}$ processes, it means that one no
longer has two jets with opposite and compensating $p_{\perp}$, but more
complicated topologies. Effects of any original kinematics selection
cuts are therefore smeared out, an unfortunate side-effect of the
parton-shower approach.
\subsection{Beam Remnants}
In a hadron--hadron collision, the initial-state radiation algorithm
reconstructs one shower initiator in each beam. This initiator only
takes some fraction of the total beam energy, leaving behind a beam
remnant which takes the rest. For a proton beam, a $\u$ quark
initiator would leave behind a $\u \d$ diquark beam remnant, with an
antitriplet colour charge. The remnant is therefore colour-connected
to the hard interaction, and forms part of the same fragmenting
system. It is further customary to assign a primordial transverse
momentum to the shower initiator, to take into account the motion
of quarks inside the original hadron, basically as required by the
uncertainty principle. This primordial $k_{\perp}$ is selected
according to some suitable distribution, and the recoil is assumed
to be taken up by the beam remnant.
Often the remnant is more complicated, e.g. a $\mathrm{g}$ initiator
would leave behind a $\u \u \d$ proton remnant system in a colour octet
state, which can conveniently be subdivided into a colour triplet
quark and a colour antitriplet diquark, each of which are
colour-connected to the hard interaction. The energy sharing between
these two remnant objects, and their relative transverse momentum,
introduces additional degrees of freedom, which are not understood
from first principles.
Na\"{\i}vely, one would expect an $\e\p$ event to have only one beam
remnant, and an $\e^+\e^-$ event none. This is not always correct, e.g.
a $\gamma \gamma \to \mathrm{q} \overline{\mathrm{q}}$ interaction in an $\e^+\e^-$ event would
leave behind the $\mathrm{e}^+$ and $\mathrm{e}^-$ as beam remnants, and a
$\mathrm{q} \overline{\mathrm{q}} \to \mathrm{g} \mathrm{g}$ interaction in resolved photoproduction in an
$\e^+\e^-$ event would leave behind one $\mathrm{e}^{\pm}$ and one $\mathrm{q} / \overline{\mathrm{q}}$
in each remnant. Corresponding complications occur for
photoproduction in $\e\p$ events.
There is another source of beam remnants. If parton distributions are
used to resolve an electron inside an electron, some of the original
energy is not used in the hard interaction, but is rather associated
with initial-state photon radiation. The initial-state shower is
in principle intended to trace this evolution and reconstruct the
original electron before any radiation at all took place. However,
because of cut-off procedures, some small amount may be left
unaccounted.
Alternatively the user may have chosen to switch off initial-state
radiation altogether, but still preserved the resolved electron
parton distributions. In either case the remaining energy is given to
a single photon of vanishing transverse momentum, which is then
considered in the same spirit as `true' beam remnants.
So far we have assumed that
each event only contains one hard interaction, i.e. that each
incoming particle has only one parton which takes part in hard
processes, and that all other constituents sail through unaffected.
This is appropriate in $\e^+\e^-$ or $\e\p$ events, but not necessarily so in
hadron--hadron collisions. Here each of the beam particles contains a
multitude of partons, and so the probability for several interactions
in one and the same event need not be negligible. In principle these
additional interactions could arise because one single parton from
one beam scatters against several different partons from the other
beam, or because several partons from each beam take place in
separate $2 \to 2$ scatterings. Both are expected, but combinatorics
should favour the latter, which is the mechanism considered in
{\tsc{Pythia}}.
The dominant $2 \to 2$ QCD cross sections are
divergent for $p_{\perp} \to 0$, and drop rapidly for larger
$p_{\perp}$. Probably the lowest-order perturbative cross sections
will be regularized at small $p_{\perp}$ by colour coherence effects:
an exchanged gluon of small $p_{\perp}$ has a large transverse
wave function and can therefore not resolve the individual colour
charges of the two incoming hadrons; it will only couple to an
average colour charge that vanishes in the limit $p_{\perp} \to 0$.
In the program, some effective $p_{\perp\mathrm{min}}$ scale is therefore
introduced, below which the perturbative cross section is either
assumed completely vanishing or at least strongly damped.
Phenomenologically, $p_{\perp\mathrm{min}}$ comes out to be a number of
the order of 1.5--2.0 GeV.
In a typical `minimum-bias' event one therefore expects to find one
or a few scatterings at scales around or a bit above
$p_{\perp\mathrm{min}}$,
while a high-$p_{\perp}$ event also may have additional scatterings
at the $p_{\perp\mathrm{min}}$ scale. The probability to have several
high-$p_{\perp}$ scatterings in the same event is small, since the
cross section drops so rapidly with $p_{\perp}$.
The understanding of multiple interaction is still very primitive,
and even the experimental evidence that it exists at all is rather
weak. {\tsc{Pythia}} therefore contains several different options,
with a fairly simple one as default. The options differ in particular
on the issue of the `pedestal' effect: is there an increased
probability or not for additional interactions in an event which
is known to contain a hard scattering, compared with one that
contains no hard interactions?
\subsection{Fragmentation}
QCD perturbation theory, formulated in terms of quarks and
gluons, is valid at short distances. At long distances, QCD
becomes strongly interacting and perturbation theory breaks
down. In this confinement regime, the coloured partons are
transformed into colourless hadrons, a process called either
hadronization or fragmentation. In this paper we reserve the
former term for the combination of fragmentation and the subsequent
decay of unstable particles.
The fragmentation process has yet to be understood from first
principles, starting from the QCD Lagrangian. This has left the
way clear for the development of a number of different
phenomenological models. Three main schools are usually
distinguished, string fragmentation (SF), independent
fragmentation (IF) and cluster fragmentation (CF),
but many variants and hybrids exist.
Being models, none of them can lay claims
to being `correct', although some may be better founded than
others. The best that can be aimed for is internal consistency,
a good representation of existing data, and a predictive power for
properties not yet studied or results at higher energies.
{\tsc{Jetset}} is intimately connected with string fragmentation,
in the form of the time-honoured `Lund model'. This is the default
for all {\tsc{Jetset/Pythia}} applications, but independent
fragmentation options also exist, for applications
where one wishes to study the importance of string effects.
All current models are of a probabilistic and iterative nature.
This means that the fragmentation process as a whole is described in
terms of one or a few simple underlying branchings, of the type
jet $\to$ hadron + remainder-jet, string $\to$
hadron + remainder-string, and so on. At each branching,
probabilistic rules are given for the production of new flavours,
and for the sharing of energy and momentum between the products.
To understand fragmentation models, it is useful to start with
the simplest possible system, a colour-singlet $\mathrm{q} \overline{\mathrm{q}}$ 2-jet
event, as produced in $\e^+\e^-$ annihilation. Here lattice QCD studies
lend support to a linear confinement picture (in the absence
of dynamical quarks), i.e. the energy stored in the colour
dipole field between a charge and an anticharge increases linearly
with the separation between the charges, if the short-distance
Coulomb term is neglected. This is quite different
from the behaviour in QED, and is related to the presence of a
triple-gluon vertex in QCD. The details are not yet well
understood, however.
The assumption of linear confinement provides the starting point for
the string model. As the $\mathrm{q}$ and $\overline{\mathrm{q}}$ partons move apart from
their common production vertex, the physical picture is that of a
colour flux tube (or maybe colour vortex line) being stretched
between the $\mathrm{q}$ and the $\overline{\mathrm{q}}$. The transverse dimensions
of the tube are of typical hadronic sizes, roughly 1 fm. If
the tube is assumed to be uniform along its length, this
automatically leads to a confinement picture with a linearly
rising potential. In order to obtain a Lorentz covariant and causal
description of the energy flow due to this linear confinement,
the most straightforward way is to use the dynamics of the massless
relativistic string with no transverse degrees of freedom.
The mathematical, one-dimensional string can
be thought of as parametrizing the position of the axis of a
cylindrically symmetric flux tube. From
hadron spectroscopy, the string constant, i.e. the amount of
energy per unit length, is deduced to be
$ \kappa \approx 1$ GeV/fm. The
expression `massless' relativistic string is somewhat of a
misnomer: $\kappa$ effectively corresponds to a `mass density' along
the string.
Let us now turn to the fragmentation process. As the $\mathrm{q}$ and
$\overline{\mathrm{q}}$ move apart, the potential energy stored in the string
increases, and the string may break by the production of a new
$\mathrm{q}' \overline{\mathrm{q}}'$ pair, so that the system splits into two
colour-singlet systems $\mathrm{q} \overline{\mathrm{q}}'$ and $\mathrm{q}' \overline{\mathrm{q}}$. If the invariant
mass of either of these string pieces is large enough, further
breaks may occur. In the Lund string model, the string break-up
process is assumed to proceed until only on-mass-shell hadrons
remain, each hadron corresponding to a small piece of string
with a quark in one end and an antiquark in the other.
In order to generate the quark--antiquark pairs $\mathrm{q}' \overline{\mathrm{q}}'$ which
lead to string break-ups, the Lund model invokes the idea of
quantum mechanical tunnelling. This leads to a flavour-independent
Gaussian spectrum for the $p_{\perp}$ of $\mathrm{q}' \overline{\mathrm{q}}'$ pairs.
Since the string is assumed to have no transverse excitations,
this $p_{\perp}$ is locally compensated between the quark and the
antiquark of the pair. The total $p_{\perp}$ of a hadron is made
up out of the $p_{\perp}$ contributions from the quark and
antiquark that together
form the hadron. Some contribution of very soft perturbative gluon
emission may also effectively be included in this description.
The tunnelling picture also implies a suppression of heavy-quark
production, $\u : \d : \mathrm{s} : \c \approx 1 : 1 : 0.3 : 10^{-11}$.
Charm and heavier quarks hence are not expected to be produced in
the soft fragmentation, but only in perturbative parton-shower
branchings $\mathrm{g} \to \mathrm{q} \overline{\mathrm{q}}$.
When the quark and antiquark from two adjacent string breakings are
combined to form a meson, it is necessary to invoke an algorithm to
choose between the different allowed possibilities, notably
between pseudoscalar and vector mesons.
Here the string model is not particularly predictive. Qualitatively one
expects a $1 : 3$ ratio, from counting the number of spin states,
multiplied by some wave-function normalization factor, which should
disfavour heavier states.
A tunnelling mechanism can also be used to explain the production of
baryons. This is still a poorly understood area. In the simplest
possible approach, a diquark in a colour antitriplet state is just
treated like an ordinary antiquark, such that a string can break
either by quark--antiquark or antidiquark--diquark pair production.
A more complex scenario is the `popcorn' one, where
diquarks as such do not exist, but rather quark--antiquark pairs
are produced one after the other. This latter picture gives a less
strong correlation in flavour and momentum space between the
baryon and the antibaryon of a pair.
In general, the different string breaks are causally disconnected.
This means that it is possible to describe the breaks in any convenient
order, e.g. from the quark end inwards. One therefore is led to write
down an iterative scheme for the fragmentation, as follows.
Assume an initial quark $\mathrm{q}$ moving out along the $+z$ axis, with the
antiquark going out in the opposite direction.
By the production of a $\mathrm{q}_1 \overline{\mathrm{q}}_1$ pair, a meson $\mathrm{q} \overline{\mathrm{q}}_1$ is
produced, leaving behind an unpaired quark $\mathrm{q}_1$.
A second pair $\mathrm{q}_2 \overline{\mathrm{q}}_2$ may now be produced,
to give a new meson $\mathrm{q}_1 \overline{\mathrm{q}}_2$, etc. At each step the produced
hadron takes some fraction of the available energy and momentum.
This process may be iterated until all energy is used up, with some
modifications close to the $\overline{\mathrm{q}}$ end of the string in order to
make total energy and momentum come out right.
The choice of starting the fragmentation from the quark end is
arbitrary, however. A fragmentation process described in terms of
starting at the $\overline{\mathrm{q}}$ end of the system and fragmenting towards
the $\mathrm{q}$ end should be equivalent.
This `left--right' symmetry constrains the allowed shape of the
fragmentation function $f(z)$, where $z$ is the fraction
of the remaining light-cone momentum $E \pm p_z$ (+ for the $\mathrm{q}$ jet,
$-$ for the $\overline{\mathrm{q}}$ one) taken by each new particle.
The resulting `Lund symmetric fragmentation function' has two free
parameters, which are determined from data.
If several partons are moving apart from a common origin, the details
of the string drawing become more complicated. For a $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$
event, a string is stretched from the
$\mathrm{q}$ end via the $\mathrm{g}$ to the $\overline{\mathrm{q}}$ end, i.e.
the gluon is a kink on the string, carrying energy and momentum.
As a consequence, the gluon has two string pieces attached, and
the ratio of gluon to quark string force is 2, a number which
can be compared with the ratio of colour charge Casimir operators,
$N_C/C_F = 2/(1-1/N_C^2) = 9/4$. In this, as in other
respects, the string model can be viewed as a variant of QCD
where the number of colours $N_C$ is not 3 but infinite.
Note that the factor 2 above does not depend on
the kinematical configuration: a smaller opening angle between
two partons corresponds to a smaller
string length drawn out per unit time, but also to an increased
transverse velocity of the string piece, which gives an exactly
compensating boost factor in the energy density per unit string
length.
The $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ string will fragment along its length. To first
approximation this means that there is
one fragmenting string piece between
$\mathrm{q}$ and $\mathrm{g}$ and a second one between $\mathrm{g}$ and $\overline{\mathrm{q}}$. One hadron
is straddling both string pieces, i.e. sitting around the gluon
corner. The rest of the particles are produced as in two simple
$\mathrm{q} \overline{\mathrm{q}}$ strings, but strings boosted with respect to the overall
c.m. frame. When considered in detail, the string motion and
fragmentation is more complicated, with the appearance of
additional string regions during the time evolution of the system.
These corrections are especially important for soft and
collinear gluons, since they provide a smooth transition between
events where such radiation took place and events where it did not.
Therefore the string fragmentation scheme is `infrared safe' with
respect to soft or collinear gluon emission.
For events that involve many partons, there may be several possible
topologies for their ordering along the string.
An example would be a $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}_1 \mathrm{g}_2$ (the gluon indices are here
used to label two different gluon-momentum vectors), where the
string can connect the partons in either of the sequences
$\mathrm{q} - \mathrm{g}_1 - \mathrm{g}_2 - \overline{\mathrm{q}}$ and $\mathrm{q} - \mathrm{g}_2 - \mathrm{g}_1 - \overline{\mathrm{q}}$.
The matrix elements that are calculable in perturbation theory
contain interference terms between these two possibilities, which
means that the colour flow is not always well-defined. Fortunately,
the interference terms are down in magnitude by a factor
$1/N_C^2$, where $N_C = 3$ is the number of colours, so
approximate recipes can be found. In the leading log shower
description, on the other hand, the rules for the colour flow are
well-defined. A final comment: in the argumentation for the
importance of colour flows there is a tacit assumption that
soft-gluon exchanges between partons will not normally mess up the
original colour assignment; this is likely the case but has not
been proven.
\subsection{Decays}
A large fraction of the particles produced by fragmentation are
unstable and subsequently decay into the observable stable (or
almost stable) ones. It is therefore important to include all
particles with their proper mass distributions and decay properties.
Although involving little deep physics, this is less trivial than
it may sound: while a lot of experimental information is available,
there is also very much that is missing. For charm mesons,
it is necessary to put together measured exclusive branching ratios
with some inclusive multiplicity distributions to obtain a consistent
and reasonably complete set of decay channels, a rather delicate
task. For bottom, so far only a rather simple phase-space type of
generator has been used for hadronic decays.
Normally it is assumed that decay products are distributed according
to phase space, i.e. that there is no dynamics involved in their
relative distribution. However, in many cases additional assumptions
are necessary, e.g. for semileptonic decays of charm and bottom
hadrons one needs to include the proper weak matrix elements.
Particles may also be produced polarized and impart a non-isotropic
distribution to their decay products. Many of these effects are
not at all treated in the program. In fact, spin information is not
at all carried along, but has to be reconstructed explicitly when
needed.
The normal decay treatment is handled by {\tsc{Jetset}}, making use
of a set of tables where branching ratios and decay modes are
stored. In {\tsc{Pythia}} a separate decay treatment exists, used
exclusively for a specific list of particles:
$\mathrm{Z}^0$, $\mathrm{W}^{\pm}$, $\H^0$, $\mathrm{Z}'^0$, $\mathrm{W}'^{\pm}$, $\H'^0$, $\mathrm{A}^0$,
$\H^{\pm}$, $\eta_{\mrm{tech}}^0$, $\mathrm{R}^0$, $\mathrm{q}^*$, $\ell^*$, and
the leptoquark $\L_{\mathrm{Q}}$.
Together we call these resonances, and contrast the `particle decay'
treatment of {\tsc{Jetset}} with the `resonance decay' one of
{\tsc{Pythia}}. Of course, this is just a matter of terminology:
a particle like the $\rho$ could also be called a resonance.
What characterizes a ({\tsc{Pythia}}) resonance is that partial widths
and branching ratios are calculated dynamically, as a function of
the actual mass of a particle. Therefore not only do branching ratios
change between an $\H^0$ of nominal mass 100 GeV and one of 200 GeV,
but also for a Higgs of nominal mass 200 GeV, the branching ratios
would change between an actual mass of 190 GeV and 210 GeV, say.
This is particularly relevant for reasonably broad resonances, and
in threshold regions. For an approach like this to work, it is
clearly necessary to have perturbative expressions available for all
partial widths, which is one reason why a corresponding treatment
would not be the same for an ordinary hadronic resonance, like the
$\rho$.
The decay products of {\tsc{Pythia}} resonances are typically quarks,
leptons, or other resonances, e.g. $\mathrm{W} \to \mathrm{q} \overline{\mathrm{q}}'$ or
$\H^0 \to \mathrm{W}^+ \mathrm{W}^-$. In decays to quarks, parton showers are
automatically added to give a more realistic multijet structure, and
one may also allow photon emission off leptons. If the decay products
in turn are resonances, further decays are necessary. Often
spin information is available in resonance decay matrix elements,
contrary to the normal state of affairs in ordinary particle decays.
This means that the angular orientations in the two decays of a
$\mathrm{W}^+ \mathrm{W}^-$ pair are properly correlated. Occasionally, the
information is not available, and then resonances decay isotropically.
The top quark is a special problem. The original machinery is based
on the assumption that the $\t$ is long-lived, so that top hadrons
have time to form in the fragmentation process, and afterwards
these mesons decay weakly. With current `best bet' mass values,
this is not correct, but one should rather consider top decay before
fragmentation. Top should then be handled like one of
the above resonances. Therefore the program now contains an
alternative along these lines, which is the preferred option.
\clearpage
\section{Program Overview}
This section contains a diverse collection of information. The first
part is an overview of previous {\tsc{Jetset}} and {\tsc{Pythia}} versions.
The second gives instructions for installation of the programs and
describes their philosophy: how they are
constructed and how they are supposed to be used. It also contains
some information on how to read this manual.
The third and final part contains several examples of pieces
of code or short programs, to illustrate the general style of
program usage. The last part is mainly intended as an introduction for
completely new users, and can be skipped by more experienced ones.
Since the {\tsc{Jetset}} and {\tsc{Pythia}} programs today are so closely
connected, and are gradually coalescing, they are presented together
in this report. However, they still appear as separate entities, with
slightly different style and emphasis.
{\tsc{Jetset}} is the older of the two, and is at the origin of the
whole `Lund' family of event generators. It can be subdivided in two
parts. The larger is a generic package for jet fragmentation,
particle decays, final-state parton showers, event-analysis routines,
and other utitilies. This package can be used in the context of any
hard process, provided one is willing to buy the underlying
assumption of jet universality, i.e. that the fragmentation process
is fundamentally the same whether one is considering an $\e^+\e^-$ or a
$\p\p$ event, and that the only differences are to be found in the
parton-level processes involved. This package is not only used by all
other `Lund' programs, but also by numerous other programs written to
study specific processes. The smaller part of {\tsc{Jetset}}
is a generator for $\e^+\e^-$ annihilation events, according to either
a parton-shower or a matrix-element approach. The
{\tsc{Jetset}} program is completely selfcontained.
{\tsc{Pythia}} is a program made to generate hard or soft processes in
collisions between leptons, hadrons and photons, especially at
$\e^+\e^-$, $\e\p$ and $\p\p$ colliders. Where {\tsc{Jetset}} is a loose
collection of routines that you can combine as desired,
{\tsc{Pythia}} is a more structured program, where you initially
set up what processes you want to study, and thereafter all events
will be generated according to this specification. Included is an
extensive library of hard subprocess differential cross sections,
a library of parton distributions, a process generation machinery,
treatment of initial-state showers and beam remnants, and a few odds
and ends. {\tsc{Jetset}} is used for final-state showers, fragmentation
and decay, but no other external libraries are needed.
An interface to external parton-distribution function libraries is
provided, however.
Many programs written by other persons make use of {\tsc{Jetset}}, and a few
also of {\tsc{Pythia}}. It is not my intention to give a complete list here.
A majority of these programs are specific to given collaborations,
and therefore not publicly distributed. Below we give a list of a
few public programs from the `Lund group', which may have a
somewhat wider application. None of them are supported by the
current author, so any requests should be directed to the persons
mentioned.
\begin{Itemize}
\item \tsc{Ariadne} is a generator for dipole emission, written
mainly by L. L\"onnblad \cite{Pet88}. The dipole provides an
alternative formulation of initial- and final-state showers. {\tsc{Jetset}}
or {\tsc{Pythia}} can be used to generate the hard process and {\tsc{Jetset}} to
do the fragmentation.
\item \tsc{Aroma} is a generator for heavy-flavour processes in
leptoproduction, written by G.~Ingelman and G. Schuler \cite{Ing88}.
It uses {\tsc{Jetset}} for fragmentation.
\item \tsc{Fritiof} is a generator for hadron--hadron, hadron--nucleus
and nucleus--nucleus collisions \cite{Nil87}, which makes use of
{\tsc{Pythia}} to generate hard QCD scatterings and of {\tsc{Jetset}} for fragmentation.
Currently H. Pi is responsible for program development.
\item \tsc{Lepto} is a leptoproduction event generator, written mainly
by G. Ingelman \cite{Ing80}. It can generate parton configurations
in deep inelastic scattering according to a number of possibilities.
It makes use of {\tsc{Jetset}} for fragmentation and additionally has a
parton-shower option based on {\tsc{Pythia}}.
\item \tsc{Lucifer} is a photoproduction generator written by
G. Ingelman and A. Weigend \cite{Ing87a}. It is a modification of an
earlier version of {\tsc{Pythia}} and makes use of {\tsc{Jetset}}.
\item \tsc{Pompyt} is a generator for pomeron interactions written by
P. Bruni and G. Ingelman \cite{Bru93}. This program defines parton
distributions, flux factors and other aspects specific to the pomeron,
which is combined with the standard {\tsc{Pythia}} machinery for process
generation.
\item \tsc{Twister} is a generator for higher-twist processes, written
by G. Ingelman \cite{Ing87}. It is a modification of an earlier
version of {\tsc{Pythia}} and makes use of {\tsc{Jetset}}.
\end{Itemize}
One should also note that a version of {\tsc{Pythia}} has been modified to
include the effects of longitudinally polarized incoming protons.
This is the work of St. G\"ullenstern et al. \cite{Gul93}.
\subsection{Update History}
Both {\tsc{Jetset}} and {\tsc{Pythia}} are by now fairly old and well-established
programs, but they are still steadily being improved
on. While evolution was especially rapid for {\tsc{Jetset}} in the
early days, that program has by now reached a certain level of
maturity, and the pace of change has dropped significantly.
{\tsc{Pythia}}, on the other hand, has been continually extended
in recent years, and may still see further growth, although
most of the basic structure should be in place by now.
In earlier days, before the advent of electronic mail,
programs were only infrequently distributed, and version numbers
corresponded to distinct new upgrades. Today, the evolutionary
process is more continuous and so is the distribution of new versions.
In particular, the introduction of a new process or feature is often
done on short notice, if no problems of backwards compatibility
are involved. With this distribution, the subversion numbers have
therefore been expanded to three digits, where the last two give
sub-subversions. For every change made in the public file, the
sub-subversion number is updated, together with the `last date of
change'. In most referencing the shorter `{\tsc{Jetset}} version 7.4'
could still be preferable to e.g. `{\tsc{Jetset}} version 7.412'.
For the record, in Tables \ref{Jetsetver} and \ref{Pythiaver} we
list the official main versions of {\tsc{Jetset}} and {\tsc{Pythia}}, respectively,
with some brief comments.
\begin{table}[tp]
\captive{The main versions of {\tsc{Jetset}}, with their date of
appearance, published manuals, and main changes from previous
versions. \label{Jetsetver}} \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c|l|@{\protect\rule{0mm}{\tablinsep}}}
\hline
No. & Date & Publ. & Main new or improved
features \\[1mm]
\hline
1 & Nov 78 & \cite{Sjo78} & single-quark jets \\
2 & May 79 & \cite{Sjo79} & heavy-flavour jets \\
3.1 & Aug 79 & --- & 2-jets in $\e^+\e^-$, preliminary 3-jets \\
3.2 & Apr 80 & \cite{Sjo80} & 3-jets in $\e^+\e^-$ with full matrix
elements, \\
& & & toponium $\to \mathrm{g} \mathrm{g} \mathrm{g}$ decays \\
3.3 & Aug 80 & --- & softer fragmentation spectrum \\
4.1 & Apr 81 & --- & baryon production and diquark
fragmentation, \\
& & & fourth-generation quarks, larger
jet systems \\
4.2 & Nov 81 & --- & low-$p_{\perp}$ physics \\
4.3 & Mar 82 & \cite{Sjo82} & 4-jets and QFD structure in $\e^+\e^-$, \\
& Jul 82 & \cite{Sjo83} & event-analysis routines \\
5.1 & Apr 83 & --- & improved string fragmentation scheme,
symmetric \\
& & & fragmentation, full 2$^{\mrm{nd}}$ order QCD
for $\e^+\e^-$ \\
5.2 & Nov 83 & --- & momentum-conservation schemes for IF, \\
& & & initial-state photon radiation in $\e^+\e^-$ \\
5.3 & May 84 & --- & `popcorn' model for baryon production \\
6.1 & Jan 85 & --- & common blocks restructured, parton showers \\
6.2 & Oct 85 & \cite{Sjo86} & error detection \\
6.3 & Oct 86 & \cite{Sjo87} & new parton-shower scheme \\
7.1 & Feb 89 & --- & new particle codes and common block
structure, \\
& & & more mesons, improved decays, vertex
information, \\
& & & Abelian gluon model, Bose--Einstein
effects \\
7.2 & Nov 89 & --- & interface to new standard common block, \\
& & & photon emission in showers \\
7.3 & May 90 & \cite{Sjo92d} & expanded support for non-standard
particles \\
7.4 & Dec 93 & \cite{Sjo94} & updated particle data and defaults
\\[1mm]
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[tp]
\captive{The main versions of {\tsc{Pythia}}, with their date of
appearance, published manuals, and main changes from previous
versions. \label{Pythiaver}} \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c|l|@{\protect\rule{0mm}{\tablinsep}}}
\hline
No. & Date & Publ. & Main new or improved features \\[1mm]
\hline
1 & Dec 82 & \cite{Ben84} & synthesis of
predecessors \tsc{Compton}, \tsc{Highpt} and \\
& & & \tsc{Kassandra} \\
2 & --- & & \\
3.1 & --- & & \\
3.2 & --- & & \\
3.3 & Feb 84 & \cite{Ben84a} & scale-breaking parton distributions \\
3.4 & Sep 84 & \cite{Ben85} & more efficient kinematics selection \\
4.1 & Dec 84 & & initial- and final-state parton showers,
$\mathrm{W}$ and $\mathrm{Z}$ \\
4.2 & Jun 85 & & multiple interactions \\
4.3 & Aug 85 & & $\mathrm{W} \mathrm{W}$, $\mathrm{W} \mathrm{Z}$, $\mathrm{Z} \mathrm{Z}$ and $\mathrm{R}$
processes \\
4.4 & Nov 85 & & $\gamma \mathrm{W}$, $\gamma \mathrm{Z}$, $\gamma \gamma$
processes \\
4.5 & Jan 86 & & $\H^0$ production, diffractive and
elastic events \\
4.6 & May 86 & & angular correlation in resonance
pair decays \\
4.7 & May 86 & & $\mathrm{Z}'^0$ and $\H^+$ processes \\
4.8 & Jan 87 & \cite{Ben87} & variable impact parameter
in multiple interactions \\
4.9 & May 87 & & $\mathrm{g} \H^+$ process \\
5.1 & May 87 & & massive matrix elements for heavy quarks \\
5.2 & Jun 87 & & intermediate boson scattering \\
5.3 & Oct 89 & & new particle and subprocess codes,
new common block \\
& & & structure, new kinematics selection,
some \\
& & & lepton--lepton and lepton--hadron
interactions, \\
& & & new subprocesses \\
5.4 & Jun 90 & & $s$-dependent widths, resonances not on the
mass shell, \\
& & & new processes, new parton distributions \\
5.5 & Jan 91 & & improved $\e^+\e^-$ and $\e\p$, several new
processes \\
5.6 & Sep 91 & \cite{Sjo92d} & reorganized parton distributions,
new processes, \\
& & & user-defined external processes \\
5.7 & Dec 93 & \cite{Sjo94} & new total cross sections,
photoproduction, top decay \\[1mm]
\hline
\end{tabular}
\end{center}
\end{table}
All versions preceding {\tsc{Jetset}} 7.3 and {\tsc{Pythia}} 5.6 should now
be considered obsolete, and are no longer supported. For stable
applications, the earlier combination {\tsc{Jetset}} 6.3 and {\tsc{Pythia}} 4.8
could still be used, however.
{\tsc{Jetset}} version 7 and {\tsc{Pythia}} version 5 have been evolved
in parallel, so some of the processes added in later versions of
{\tsc{Pythia}} make use of particle data only found in {\tsc{Jetset}}
from that time onwards. Although it would be possible to
combine {\tsc{Pythia}} 5.7 with {\tsc{Jetset}} 7.3, e.g., it is not recommended.
From the current versions onwards, checks have therefore
been introduced to detect the use of (potentially) incompatible
subversions, with warnings issued at initialization if that should
be the case.
Previous versions of the manuals have contained detailed lists of
modifications from one version to the next, see e.g. \cite{Sjo92d}.
Below we only reproduce the updates that appear with the most recent
versions of the programs. Some of them were introduced in later
editions of {\tsc{Pythia}} 5.6 with {\tsc{Jetset}} 7.3, while others are completely
new. If nothing is explicitly said, these changes do not
affect backwards compatibility, but only add new features.
\subsubsection{Updates in JETSET 7.4}
Changes from version 7.3 to 7.4 are not so large, although the impact
of the updated particle data and parameter default values may need to
be studied.
\begin{Itemize}
\item Particle data have been updated in accordance with the 1992
Review of Particle Properites \cite{PDG92}. (As usual, with a free
interpretation of inconsistencies, unclarities and other gaps in
the knowledge.) Changes are especially drastic for charm and bottom.
In the bottom sector the decay properties are now given individually
for $\mathrm{B}^0$, $\mathrm{B}^+$, $\mathrm{B}_{\mathrm{s}}^0$, $\mathrm{B}_{\c}^+$ and $\Lambda_{\b}^0$,
i.e. the generic data for `pseudoparticle' 85 are only used for
other weakly decaying $\mathrm{B}$ baryons.
\item Also a few other Standard Model parameters have been updated,
such as the $\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$ masses and widths, $\sin^2 \! \theta_W$ and
the CKM matrix elements.
\item Fragmentation and parton shower parameters
have been modified to reflect current LEP knowledge \cite{LEP90},
i.e. a minor retuning starting from an average of the `best'
parameter values obtained by the four LEP collaborations.
Bose-Einstein effects are still left out. Flavour composition is
unchanged, except for a suppression of $\eta'$ production.
Affected by the change are \ttt{MSTJ(11)}, \ttt{PARJ(21)},
\ttt{PARJ(23)}, \ttt{PARJ(24)}, \ttt{PARJ(41)}, \ttt{PARJ(42)},
\ttt{PARJ(54)}--\ttt{PARJ(58)} and \ttt{PARJ(81)}.
\item Several other default values have been changed for switches and
parameters in the $\e^+\e^-$, parton shower and fragmentation parts of
the programs. These changes are intended to reflect our current
best knowledge. See \ttt{MSTJ(26)}, \ttt{MSTJ(27)}, \ttt{MSTJ(41)},
\ttt{MSTJ(46)}, \ttt{MSTJ(50)}, \ttt{MSTJ(110)},
\ttt{PARJ(26)}, and \ttt{PARJ(121)}--\ttt{PARJ(125)}.
\item A common title page for {\tsc{Jetset}} and {\tsc{Pythia}} has been introduced
with the \ttt{LULOGO} routine. Sub-subversion numbers are also given.
\item Several options have been added for the \ttt{LUSHOW} shower
routine. See \ttt{MSTJ(41)}, \ttt{MSTJ(47)} and \ttt{MSTJ(50)}.
\item A $\b$ quark produced in the decay of a top hadron is allowed
to radiate according to the standard parton shower scheme.
\item The scalar gluon option contains the full electroweak angular
distribution of 3-jet events.
\item The \ttt{LUCOMP} routine has been modified. Among other things,
the $\mathrm{B}_{\c}^+$ now appears as a separate compressed code, further
codes for diffractive states have been added to the current list, and
the pomeron (reggeon, $\eta_{\mrm{techni}}$) has been added as particle
29 (28, 38).
\item A minimum threshold for calorimeter cell energy has been
introduced for the \ttt{LUCELL} routine.
\item All obsolescent features of the Fortran 90 standard have been
removed, i.e. the program should work well either with a Fortran 77
compiler or with a Fortran 90 one.
\item A few minor errors have been corrected.
\end{Itemize}
The following changes have been made since the beginning of 1994,
i.e. since the original distribution 7.400:
\begin{Enumerate}
\item {\tsc{Jetset}} version 7.401, 11 February 1994:
\begin{Itemize}
\item Protect against overflow in \ttt{LUZDIS} (needed on some
machines).
\end{Itemize}
\item {\tsc{Jetset}} version 7.402, 7 April 1994:
\begin{Itemize}
\item New option to suppress either hard or soft radiation in
\ttt{LUSHOW}, see \ttt{MSTJ(40)}.
\item A generic interface to an external $\tau$ decay library has been
introduced, see \ttt{MSTJ(28)} and \ttt{SUBROUTINE LUTAUD}.
\item In a few places, a dot has been moved from the end of one line
to the beginning of the next continuation line, or the other way around,
to keep together tokens such as \ttt{.EQ.} or \ttt{.AND.}, since some
debuggers may otherwise complain.
\item A source of (harmless) division by zero in \ttt{LUSHOW} has been
removed.
\end{Itemize}
\item {\tsc{Jetset}} version 7.403, 15 July 1994:
\begin{Itemize}
\item Leptons and photons which are unrelated to the system
feeling the Bose-Einstein effects do not have their energies and
momenta changed in the global rescaling step of \ttt{LUBOEI}.
(Example: $\mathrm{W}^+\mathrm{W}^-$ events, where one $\mathrm{W}$ decays leptonically;
before these lepton momenta could be slightly changed, but now not.)
Further, the \ttt{LUBOEI} routine has been changed to avoid an
unintentional gap in the limits of the very first bin.
\item The option \ttt{LUEDIT(16)} (used e.g. from \ttt{PYEVNT}) has
been improved with a more extensive search for missing daughter
pointers.
\item The \ttt{KLU(I,16)} procedure for finding rank has been rewritten
to work in the current {\tsc{Jetset}} version, which it did not before. However,
note that it will only work for \ttt{MSTU(16)=2}. As a general comment,
the options 14--17 of \ttt{KLU} were written at a time when possible
event histories were less complex, and can not be guaranteed always to
work today.
\end{Itemize}
\item {\tsc{Jetset}} version 7.404, 25 August 1994:
\begin{Itemize}
\item \ttt{LUSHOW} has been corrected, so that if $\t$, $\mathrm{l}$ or $\mathrm{h}$
quarks (or $\d^*$ or $\u^*$ quarks masked as $\mathrm{l}$ or $\mathrm{h}$ ones) are
given with masses that vary from event to event (a Breit-Wigner shape,
e.g.), the current mass rather than the nominal mass is used to define
the cut-off scales of parton shower evolution.
\item \ttt{LULOGO} has been modified to take into account that a new
{\tsc{Pythia/Jetset}} description has been published in \\
T. Sj\"ostrand, Computer Phys. Commun. {\bf 82} (1994) 74 \\
and is from now on the standard reference to these two programs.
\end{Itemize}
\item {\tsc{Jetset}} version 7.405, 27 January 1995:
\begin{Itemize}
\item \ttt{LUCELL} has been corrected, in that in the option with
smearing of energy rather than transverse energy, the conversion
factor between the two was applied in the wrong direction.
\item \ttt{LUSHOW} has been corrected in one place where the PMTH
array was addressed with the wrong order of the indices. This affected
quark mass corrections in the matching to the three-jet matrix
elements.
\item An additional check has been included in \ttt{LUBOEI} that there
are at least two particles involved in the Bose-Einstein effects. (No
problem except in some bizarre situations.)
\end{Itemize}
\item {\tsc{Jetset}} version 7.406, 20 February 1995:
\begin{Itemize}
\item A new option has been added for the behaviour of the running
$\alpha_{\mathrm{em}}(Q^2)$ in \ttt{ULALEM}. This is not added as a true physics
scenario, but only to produce results with a given, fixed value for
the hard events, while still keeping the conventional value in the
$Q^2=0$ limit. See \ttt{MSTU(101)}, \ttt{PARU(103)}, \ttt{PARU(104)}.
Additionally, the $G_{\mrm{F}}$ constant has been added to the
parameter list, see \ttt{PARU(105)}.
\item The LULOGO routine has been updated to reflect my change of
affiliation.
\end{Itemize}
\item {\tsc{Jetset}} version 7.407, 21 June 1995:
\begin{Itemize}
\item Header and \ttt{LULOGO} have been updated with respect to phone
number and WWW access.
\item The \ttt{PHEP} and \ttt{VHEP} variables in the \ttt{/HEPEVT/}
common block are now assumed to be in \ttt{DOUBLE PRECISION}, in
accord with the proposed LEP 2 workshop addendum to the standard.
\item In \ttt{LUTEST} a missing decimal point on the energy check
has been reinstated.
\item In \ttt{LUINDF} an expression has been protected against
vanishing denominator.
\end{Itemize}
\item {\tsc{Jetset}} version 7.408, 23 August 1995:
\begin{Itemize}
\item Check against division by zero in \ttt{LUSHOW}.
\end{Itemize}
\end{Enumerate}
\subsubsection{Updates in PYTHIA 5.7}
The updates from version 5.6 to 5.7 are all minor, and just about any
program that ran with version 5.6 will also work with {\tsc{Pythia}} 5.7.
However, as for {\tsc{Jetset}}, it should be noted that some important default
values have been changed.
\begin{Itemize}
\item New parametrizations of the total cross sections of
hadronic reactions, based in Donnachie--Landshoff \cite{Don92},
which replace the old ones.
\item New parametrizations of elastic and single and double
diffractive cross sections of hadronic reactions, based on
Schuler--Sj\"ostrand \cite{Sch94,Sch93a}, which replace the old ones.
Also the slope parameters, the diffractive mass distributions and other
aspects of the event generation have been changed accordingly.
\item A possibility to give own total, elastic and diffractive
cross sections.
\item The single diffractive cross section has been split into its
two constituents, $AB \to XB$ and $AB \to AX$. As a consequence, the
diffractive subprocess codes 92--94 have received changed meaning.
\item A new common block \ttt{PYINT7} has been added for the expanded
total cross section information, and this information has been partly
removed from other common blocks.
\item A much extended description of photoproduction physics, with
the possibility to simulate separately VMD, anomalous and direct
processes \cite{Sch93,Sch93a}.
\item The selection of proton parton distributions that come with the
program has been updated with the CTEQ2 ones, while some others have
been removed. New default is the leading-order fit CTEQ2L.
\item Since the \tsc{Pdflib} library now has been expanded to contain
also parton-distribution functions for the photon, the interfaces to
the \tsc{Pakpdf} and \tsc{Phopdf} libraries have been removed.
In addition, the interface to \tsc{Pdflib} has been modified, and is
now for appropriate for \tsc{Pdflib} version 4.
\item An extension of hadron parton distributions into the low-$x$ and
low-$Q^2$ region \cite{Sch93a}.
\item The top quark can be made to decay before it has time to
fragment. In view of the current best estimate for the top mass,
this is the expected behaviour, and is therefore now default.
Further, a parton shower is allowed to evolve in the top decay.
Also fourth generation quarks are allowed to decay before they
fragment, and so on.
\item It is possible to call \ttt{PYEVNT} with energies that vary
from one event to the next, without the need to reinitialize.
\item Improved scheme for post-factor conservation of $x$ and $Q^2$
in deep inelastic scattering.
\item Processes 15, 19, 30 and 35 have been expanded to cover
$\gamma^*$ production in addition to the $\mathrm{Z}^0$ one, with full
interference.
\item New process 80, $\mathrm{q}\gamma \to \mathrm{q}'\pi^{\pm}$.
\item New process 110, $\mathrm{f}\overline{\mathrm{f}} \to \gamma\H^0$.
\item New process 149, $\mathrm{g}\g \to \eta_{\mrm{techni}}$.
\item New option for initial state radiation to restrict angular
range of emission in accordance with coherence considerations.
\item Some options have been added or removed, and default values
have been changed. This includes
\ttt{KFIN} (top parton distributions off by default),
\ttt{MSTP(7)}, \ttt{MSTP(11)}, \ttt{MSTP(14)},
\ttt{MSTP(23)}, \ttt{MSTP(30)} (removed),
\ttt{MSTP(31)}, ~\ttt{MSTP(34)}, \ttt{MSTP(45)},
\ttt{MSTP(48)}, \ttt{MSTP(49)}, \ttt{MSTP(62)},
\ttt{MSTP(67)}, \ttt{MSTP(101)},
\ttt{PARP(13)}, \ttt{PARP(81)}, \ttt{PARP(82)},
\ttt{PARP(47)} and \ttt{PARP(101)}.
\item All obsolescent features of the Fortran 90 standard have been
removed, i.e. the program should work well either with a Fortran 77
compiler or with a Fortran 90 one.
\item A few minor errors have been corrected.
\end{Itemize}
The following changes have been made since the beginning of 1994,
i.e. since the original distribution 5.700:
\begin{Enumerate}
\item {\tsc{Pythia}} version 5.701, 27 January 1994:
\begin{Itemize}
\item The machinery to handle $\gamma\gamma$ interactions is expanded
to the level already available for $\gamma\mathrm{p}$. This in particular
means that a number of new options appear for \ttt{MSTP(14)}. Affected
are also \ttt{MINT(105)}, \ttt{MINT(107)}, \ttt{MINT(108)},
\ttt{MINT(109)}, \ttt{VINT(282)} (removed), \ttt{VINT(283)} and
\ttt{VINT(284)}. Parametrizations are introduced for meson--meson
total, elastic and diffractive cross sections, needed for the VMD
part of the photon. The treatment of cross sections for hard processes,
of initial state radiation, of beam remnants and of other aspects are
also expanded to cover the new possibilities. A first study of the
relevant physics aspects is found in \cite{Sch94a}.
\item An option is introduced to modify the $Q^2$ scale of the
anomalous part of the photon parton distributions, see \ttt{MSTP(59)}
and \ttt{PARP(59)}.
\item Correction of an error, where the generation of jet and
low-$p_{\perp}$ events could give incorrect cross section information with
\ttt{PYSTAT(1)} at low energies. The event generation itself was
correct. (The error was introduced as a consequence of allowing
variable energies.)
\item A rejection is introduced for top events where the top mass
(selected according to a Breit-Wigner) is too low to allow the decay
into a $\mathrm{W}$ on the mass shell.
\item The correction of a few other minor bugs, probably harmless.
\end{Itemize}
\item {\tsc{Pythia}} version 5.702, 13 February 1994:
\begin{Itemize}
\item The interface to \tsc{Pdflib} has been modified to reflect that
\ttt{TMAS} should no longer be set except in first \ttt{PDFSET} call.
(Else a huge amount of irrelevant warning messages are generated by
\tsc{Pdflib}.)
\item The \ttt{STOP} statement in a few dummy routines has been modifed
to avoid irrelevant compilation warning messages on IBM mainframes.
\item A few labels have been renumbered.
\end{Itemize}
\item {\tsc{Pythia}} version 5.703, 22 February 1994:
\begin{Itemize}
\item Removal of a bug in \ttt{PYRESD}, which could give (under some
specific conditions) errors in the colour flow.
\end{Itemize}
\item {\tsc{Pythia}} version 5.704, 7 April 1994:
\begin{Itemize}
\item Process 11 has been corrected, for the part that concerns
anomalous couplings (contact interactions) in the $\mathrm{q}\q' \to \mathrm{q}\q'$
process. The error was present in the expression for
$\u\overline{\mathrm{d}} \to \u\overline{\mathrm{d}}$ and obvious permutations, while
$\u\d \to \u\d$, $\u\overline{\mathrm{u}} \to \u\overline{\mathrm{u}}$ and the others were correct.
\item The option with post-facto $(x,Q^2)$
conservation in deep inelastic scattering can give infinite loops
when applied to process 83, in particular if one asks for the
production of a top. (Remember that the standard DIS kinematics
is defined for massless quarks.) Therefore the switch
\ttt{MSTP(23)} has been modifed so that by default only process 10 is
affected.
\item \ttt{PYRESD} is modified to ensure isotropic angular distributions
in the decays of the top or a fourth generation particle, i.e. in
$\t \to \b\mathrm{W}^+$. This may not be the correct distribution but, unless
explicit knowledge exists for a given process, this should always be
the default.
\item In processes 16, 20, 31 and 36 the $\mathrm{W}$ propagator has been
modified to include $s$-dependent widths in the Breit-Wigner shape.
The most notable effect is a suppression of the low-mass tail of the
$\mathrm{W}$ mass spectrum.
\item When \tsc{Pdflib} is used, \ttt{PDFSET} is now only called
whenever a different structure function is requested. For $\mathrm{p}\p$
events therefore only one call is made, while $\gamma\mathrm{p}$ interactions
still involve a call to \ttt{PDFSET} for each \ttt{STRUCTM} one, since
$\gamma$ and $\mathrm{p}$ structure functions have to be called alternatingly.
\ttt{MINT(93)} is used to keep track of latest structure function
called.
\item In a few places, a dot has been moved from the end of one line
to the beginning of the next continuation line, or the other way around,
to keep together tokens such as \ttt{.EQ.} or \ttt{.AND.}, since some
debuggers may otherwise complain.
\item A number of minor errors have been corrected.
\end{Itemize}
\item {\tsc{Pythia}} version 5.705, 15 July 1994:
\begin{Itemize}
\item A completely new possibility to have {\tsc{Pythia}} mix different allowed
processes (direct, VMD and anomalous) in $\gamma\mathrm{p}$ and $\gamma\gamma$
interactions. This option can be accessed with \ttt{MSTP(14)=10}.
The relevant physics description and programming details may be found
in sections \ref{sss:photoprod} and \ref{sss:photoprodclass}.
This facility is still not definitive, in that it is hoped to gradually
enhance it with further features. The cross-section output of the
\ttt{PYSTAT} has been expanded to reflect the further subdivision of the
total cross section.
\item The new facility above has required a major restructuring of some
of the code: the routine \ttt{PYEVKI} has been removed, new routines
\ttt{PYINBM}, \ttt{PYINPR} and \ttt{PYSAVE} created, and some material
has been moved to or from \ttt{PYINIT}, \ttt{PYINRE} and \ttt{PYINKI}.
New variables include \ttt{MSTI(9)}, \ttt{MINT(121)}, \ttt{MINT(122)},
\ttt{MINT(123)} and \ttt{VINT(285)}.
\item The GRV leading-order dynamically generated parton distributions
for the $\mathrm{p}$ and $\pi$ have been included as options, see
\ttt{MSTP(51)} and \ttt{MSTP(53)}.
\item A parametrization of the homogeneous solution to the anomalous
photon parton distributions have been added as an option, see
\ttt{MSTP(56)=3}.
\item The treatment of the anomalous photon component can be modified
with the new switch \ttt{MSTP(15)} and variable \ttt{PARP(17)}; at the
same time \ttt{MSTP(59)} and \ttt{PARP(59)} have been removed. The new
options are mainly intended for comparative studies and should not
normally be touched.
\item The option \ttt{MSTP(92)=5} for beam remnant treatment
erroneously missed some statements which now have been inserted.
Further, new options have been added for the beam remnant splitting of
momentum between a hadron and a quark/diquark jet, where \ttt{MSTP(94)}
should now be used rather than \ttt{MSTP(92)}.
\item In \ttt{PYDIFF} the recoiling gluon energy is calculated in a
numerically more stable fashion.
\end{Itemize}
\item {\tsc{Pythia}} version 5.706, 25 August 1994:
\begin{Itemize}
\item New processes 167 and 168, $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \d^*$ and
$\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \d^*$, respectively, have been introduced.
These contact interaction production processes of excited quarks
complement the quark--gluon fusion ones in processes 147 and 148,
and obey the same general rules, see section \ref{sss:qlstarclass}.
\item The option \ttt{MSTP(57)=3} now also allows a dampening of
$\pi^{\pm}$ parton distributions.
\item A few minor errors have been corrected.
\end{Itemize}
\item {\tsc{Pythia}} version 5.707, 20 October 1994:
\begin{Itemize}
\item A major bug discovered in processes 121 and 122 (and thus also
affecting 181, 182, 186 and 187), $\mathrm{g} \mathrm{g} (\mathrm{q} \overline{\mathrm{q}}) \to \mathrm{Q} \overline{\mathrm{Q}} \H$:
the kinematics was incorrectly handed on to the Kunszt matrix elements.
This affected the default option $\mathrm{Q} = \t$, but effects were
especially dramatic when the alternative $\mathrm{Q} = \b$ was used.
The choice of appropriate $Q^2$ scale for structure functions
introduces a further uncertainty in cross sections for the
processes above. So long as only $\t$ quarks are considered,
the $\t$ mass is a reasonable choice, but for the $\mathrm{Q} = \b$
alternative this is presumably too low. Therefore new
options have been introduced in \ttt{MSTP(39)}, with the default
behaviour changed.
\item Another important bug corrected in the calculation of the
reduction of $\t\overline{\mathrm{t}}$ cross section when decay modes are forced.
This occured when both $\t$ and $\overline{\mathrm{t}}$ produced a $\mathrm{W}$, and
$\mathrm{W}^+$ and $\mathrm{W}^-$ decay modes were set differently.
\end{Itemize}
\item {\tsc{Pythia}} version 5.708, 25 October 1994:
\begin{Itemize}
\item A few further places changed to make processes 181, 182, 186
and 187 work (see version 5.707 above).
\end{Itemize}
\item {\tsc{Pythia}} version 5.709, 26 October 1994:
\begin{Itemize}
\item The matrix element for $\mathrm{f} \overline{\mathrm{f}} \to \mathrm{W}^+ \mathrm{W}^-$ has been
replaced, using the formulae of \\
D. Bardin, M. Bilenky, D. Lehner, A. Olchevski and T. Riemann,
CERN-TH.7295/94, \\
but with the dependence on the $\hat{t}$ variable not integrated
out (D. Bardin, private communication).
This avoids some problems encountered in the old expressions when
one or both $\mathrm{W}$'s were far off the mass shell.
\item Change in calls to \tsc{Pdflib}, so that the input $Q$ is
always at least the $Q_{\mathrm{min}}$ of the respective set.
\item Extra protection against infinite loops in PYSSPA.
\end{Itemize}
\item {\tsc{Pythia}} version 5.710, 27 January 1995:
\begin{Itemize}
\item The dimensions of the \ttt{HGZ} array in \ttt{PYRESD} has been
expanded to avoid accidental writing outside the bounds.
\item \ttt{VINT(41)-VINT(66)} are saved and restored in \ttt{PYSCAT},
for use in low-$p_{\perp}$ events, when beam remnant treatment has failed
(with nonzero \ttt{MINT(57)}).
\item The routine \ttt{PYSTGH} has been replaced by the routine
\ttt{PYSTHG}. This contains an improved parametrization of the
homogeneous evolution of an anomalous photon from some given initial
scale. The argument \ttt{NF} of the \ttt{PYSTGH} routine has been
removed; now $\Lambda$ is always automatically converted to the
relevant $n_f$-flavour value from its 4-flavour one, at flavour
thresholds.
\end{Itemize}
\item {\tsc{Pythia}} version 5.711, 20 February 1995:
\begin{Itemize}
\item New possibilities have been added to switch between electroweak
couplings being expressed in terms of a running $\alpha_{\mathrm{em}}(Q^2)$ or
in terms of a fixed Fermi constant $G_{\mrm{F}}$. This affects both
decay widths and process cross sections, in the routines \ttt{PYINRE},
\ttt{PYRESD}, \ttt{PYWIDT} and \ttt{PYSIGH}. See \ttt{MSTP(8)} for
details; default corresponds to old behaviour.
\item The option \ttt{MSTP(37)=1}, with running quark masses in
couplings to Higgs bosons, only works when $\alpha_{\mathrm{s}}$ is allowed to
run (so one can define a $\Lambda$ value). Therefore a check has been
introduced in \ttt{PYWIDT} and \ttt{PYSIGH} that the option
\ttt{MSTP(37)=1} is only executed if additionally
\ttt{MSTP(2)}$\geq 1$.
\item Some non-physics changes have been made in the \ttt{RKBBV}
and \ttt{STRUCTM} codes so as to avoid some (in principle harmless)
compiler warnings.
\end{Itemize}
\item {\tsc{Pythia}} version 5.712, 15 March 1995:
\begin{Itemize}
\item A serious error has been corrected in the \ttt{MSTP(173)=1}
option, i.e. when the program is run with user-defined weights that
should compensate for a biased choice of variable beam energies.
This both affected the relative admixture of low- and high-$p_{\perp}$
events and the total cross section obtained by Monte Carlo integration.
(\ttt{PYRAND} changed.)
\item In order to improve the flexibility and efficiency of the
variable-energy option, the user should now set \ttt{PARP(174)} before
the \ttt{PYINIT} call, and thereafter not change it. This allows
\ttt{PARP(173)} weights of arbitrary size. (\ttt{PYRAND} and
\ttt{PYMAXI} changed.)
\item \ttt{MSTI(5)} (and \ttt{MINT(5)}) are now changed so they count
the number of successfully generated events, rather than the number of
tries made. This change only affects runs with variable energies,
\ttt{MSTP(171)=1} and \ttt{MSTP(172)=2}, where \ttt{MSTI(61)=1}
signals that a user-provided energy has been rejected in the
weighting. This change also affects \ttt{PARI(2)}, which becomes the
cross section per fully generated event. (\ttt{PYEVNT} changed.)
\item The option \ttt{MSTP(14)=10} has now been extended so that it
also works for deep inelastic scattering of an electron off a (real)
photon, i.e. subprocess 10. What is obtained is a mixture of the
photon acting as a vector meson and it acting as an anomalous state.
This should therefore be the sum of what can be obtained with
\ttt{MSTP(14)=2} and \ttt{=3}. It is distinct from \ttt{MSTP(14)=1}
in that different sets are used for the parton distributions --- in
\ttt{MSTP(14)=1} all the contributions to the photon distributions
are lumped together, while they are split in VMD and anomalous parts
for \ttt{MSTP(14)=10}. Also the beam remnant treatment is different,
with a simple Gaussian distribution (at least by default)
for \ttt{MSTP(14)=1} and the VMD part of \ttt{MSTP(14)=10}, but a
powerlike distribution $\d k_{\perp}^2)/k_{\perp}^2$ between
\ttt{PARP(15)} and $Q$ for the anomalous part of \ttt{MSTP(14)=10}.
(\ttt{PYINIT}, \ttt{PYINPR} and \ttt{PYSTAT} changed.)\\
To access this option for $\mathrm{e}$ and $\gamma$ as incoming beams,
it is only necessary to set \ttt{MSTP(14)=10} and keep \ttt{MSEL} at
its default value. Unlike the corresponding option for $\gamma\mathrm{p}$
and $\gamma\gamma$, no cuts are overwritten, i.e. it is still the
responsability of the user to set these appropriately. Those
especially appropriate for DIS usage are \ttt{CKIN(21)-CKIN(22)} or
\ttt{CKIN(23)-CKIN(24)} for the $x$ range (former or
latter depending on which side is the incoming real photon),
and \ttt{CKIN(35)-CKIN(36)} for the $Q^2$ range. A further new option
has been added (in \ttt{PYKLIM}) to set the $W^2$ range as well, see
\ttt{CKIN(39)-CKIN(40)}.\\
A warning about the usage of \tsc{Pdflib} for photons. So long as
\ttt{MSTP(14)=1}, i.e. the photon is not split up, \tsc{Pdflib} is
accessed by \ttt{MSTP(56)=2} and \ttt{MSTP(55)} the parton
distribution set, as described in the manual. However, when the VMD
and anomalous pieces are split, the VMD part is based on a rescaling
of pion distributions by VMD factors (except for the SaS sets, that
already come with a separate VMD piece). Therefore, to access
\tsc{Pdflib} for \ttt{MSTP(14)=10}, it is not correct to set
\ttt{MSTP(56)=2} and a photon distribution in \ttt{MSTP(55)}. Instead,
one should put \ttt{MSTP(56)=2}, \ttt{MSTP(54)=2} and a pion
distribution code in \ttt{MSTP(53)}, while \ttt{MSTP(55)} has no
function. The anomalous part is still based on the SaS
parametrization, with \ttt{PARP(15)} as main free parameter.
\item A change has been made in \ttt{PYREMN} to reduce the
possibility of infinite loops.
\end{Itemize}
\item {\tsc{Pythia}} version 5.713, 22 March 1995:
\begin{Itemize}
\item The SaS parton distributions of the photons are now available,
see \cite{Sch95}. There are four new sets. These differ in that two
use a $Q_0=0.6$~GeV and two a $Q_0=2$~GeV, and in that two use the
DIS and two the $\overline{\mathrm{MS}}$ conventions
for the dominant non-leading contributions. (However, the fits are
formally still leading-order, in that not all next-to-leading
contributions have been included.) New default is the SaS 1D set.
Furthermore, for the definition of $F_2^{\gamma}$, additional terms
appear that do not form part of the parton distributions itself.
To partly take this into account, an additional doubling of the
possibilities has been included. These eight possibilites can be
accesed with \ttt{MSTP(55)}. The default value of \ttt{PARP(15)}
has been changed from 0.5 to 0.6~GeV, for consistency with SaS 1D.\\
The generic routine \ttt{PYSTFU} has been rewritten to handle the
interfacing. The old routines \ttt{PYSTAG}, \ttt{PYSTGS}, \ttt{PYDILN}
and \ttt{PYSTHG} have been removed. Instead the routines of the
\tsc{SaSgam} library have been inserted. In order to avoid any
clashes, the routines \ttt{SAS***} have been renamed \ttt{PYG***}.
Thus new routines are \ttt{PYGGAM}, \ttt{PYGVMD}, \ttt{PYGANO},
\ttt{PYGBEH} and \ttt{PYGDIR}. The common block \ttt{SASCOM} is
renamed \ttt{PYINT8}. If you want to use the parton distributions
for standalone purposes, you are encouraged to use the original
\tsc{SaSgam} routines rather than going the way via the
{\tsc{Pythia}} adaptations.
\item \ttt{PYDOCU} has been corrected so that \ttt{PARI(2)} refers
to the full cross section for $\gamma\mathrm{p}$ and $\gamma\gamma$ processes,
rather than that of the latest subprocess considered.
\item An additional check has been inserted into PYREMN.
\end{Itemize}
\item {\tsc{Pythia}} version 5.714, 22 March 1995:
\begin{Itemize}
\item Some minor modifications to \ttt{PYSTFU} and \ttt{PYGGAM} in
the wake of the changes of the previous version.
\end{Itemize}
\item {\tsc{Pythia}} version 5.715, 24 April 1995:
\begin{Itemize}
\item An unfortunate choice of default values has been corrected:
the old \ttt{MSTP(3)=2} value implied that $\Lambda_{\mrm{QCD}}$ was
entirely based on the $\Lambda$ value of the proton structure
function; also e.g. for $\e^+\e^-$ annihilation events. Thus the $\Lambda$
in \ttt{PARJ(81)} was overwritten, i.e. did not keep the value
required by standard phenomenology, which typically gave too narrow
jets. (While switching to \ttt{MSTP(3)=1} it worked fine.) In the
modified option \ttt{MSTP(3)=2} this has been corrected, to
better agree with user expectations. Since further changes were made
in version 5.716, we refer below for additional comments.
\item The form for \ttt{PTMANO}, the $p_{\perp\mathrm{min}}$ for anomalous processes,
as used in \ttt{PYINPR} when processes are mixed for $\gamma\mathrm{p}$ or
$\gamma\gamma$ events, has been updated to match (as well as can be
expected) the SaS 1D photon distributions.
\end{Itemize}
\item {\tsc{Pythia}} version 5.716, 30 June 1995:
\begin{Itemize}
\item The strategy for the changes to $\Lambda$ in version 5.715 above
have been modified for better transparency. Now \ttt{PARJ(81)} is used
for resonance decays (including e.g. $\mathrm{Z}^0$ decay, from which it is
determined), and \ttt{PARP(72)} for other timelike showers.
\ttt{PARJ(81)} is not overwritten \ttt{for MSTP(3)=2}, but only for
\ttt{=3}. Changes affect \ttt{PYINIT}, \ttt{PYEVNT} and \ttt{PYRESD}.
\item A new multiplicative factor has been introduced for the $Q^2$
scale choice of the hard scattering in \ttt{PYSIGH}, affecting
parton distributions and $\alpha_{\mathrm{s}}$, see \ttt{PARP(34)}.
\item \ttt{PYREMN} has been corrected for occasional too large boost
factors.
\item An error in \ttt{PYSIGH} for process 148 has been corrected.
\item The \ttt{MSTP(62)=1} option of \ttt{PYSSPA} is modified to
avoid division by zero.
\item Header has been updated with WWW-information.
\end{Itemize}
\item {\tsc{Pythia}} version 5.717, 23 August 1995:
\begin{Itemize}
\item \ttt{MIN1}, \ttt{MIN2}, \ttt{MAX1}, \ttt{MAX2}, \ttt{MINA}
and \ttt{MAXA} in \ttt{PYSIGH} have had an extra \ttt{M}
prefixed to avoid confusion with Fortran functions.
\item Protect against \ttt{MDCY(0,1)} being accessed in \ttt{PYSIGH}.
\item Protect against \ttt{THB=0} in \ttt{PYRAND}.
\item Protect against \ttt{YSTMAX-YSTMIN = 0} in \ttt{PYSIGH}.
\item Check for moved leptoquark at beginning of \ttt{PYRESD} just like
for other particles with colour.
\end{Itemize}
\end{Enumerate}
\subsection{Program Installation}
\label{ss:install}
Several `authorized' sources of the programs exist. The `master
copy' of the programs is the one found on my World Wide Web homepage\\
\ttt{http://thep.lu.se/tf2/staff/torbjorn/Welcome.html}\\
There you have:
\begin{tabbing}
~~~ \= \ttt{jetset74.f}~~~~~~~ \= the {\tsc{Jetset}} code, \\
\> \ttt{pythia57.f} \> the {\tsc{Pythia}} code, \\
\> \ttt{pythia57.tex} \> this common {\tsc{Pythia/Jetset}} manual, and \\
\> \ttt{update57.notes} \> plain text update notes to the manual.
\end{tabbing}
In addition to these, one may also find older versions of the
program and manuals, sample main programs and other pieces of
related software, and other physics papers.
The lack of stable versions may make it less convenient to rely
on the above files. New versions are introduced in the general
distribution of the CERN program library, maybe once a year.
These versions are better checked before release, and should be
useful for most applications. However, clearly, they may be less
up-to-date. Read the CERN Computer Newsletter for announcements.
Copies of the programs are also available via anonymous ftp,
e.g. from the asisftp server at CERN.
The programs are written entirely in standard Fortran 77, and should
run on any machine with such a compiler. To a first approximation,
program compilation should therefore be straightforward.
Unfortunately, experience with many different compilers has been
uniform: the options available for obtaining optimized code
actually produce erroneous code (e.g. operations inside \ttt{DO}
loops are moved out before them, where some of the variables have
not yet been properly set). Therefore the general advice is to
use a low optimization level. Note that this is often not the
default setting.
\ttt{SAVE} statements have been included in accordance with the
Fortran standard. Since most ordinary machines take \ttt{SAVE} for
granted, this part is not particularly well tried out, however.
All default settings and particle and process data are stored in
\ttt{BLOCK DATA LUDATA} for {\tsc{Jetset}} and \ttt{BLOCK DATA PYDATA}
for {\tsc{Pythia}}. These subprograms must be linked for a proper
functioning of the other routines. On some machines this is not done
automatically but must be forced by you, in particular if
{\tsc{Jetset}} and {\tsc{Pythia}} are maintained
as libraries from which routines are to be loaded only when they are
needed. In this connection we note that the library approach does not
give any significant space advantages over a loading of the
packages as a whole, since a normal run will call on most of the
routines anyway, directly or indirectly.
Since most machines in current use are 32-bit ones, this is the
precision normally assumed. A few pieces of code have therefore had
to be written in double precision. As a rule of thumb,
double-precision variables have as first character \ttt{D}, but
there are a few exceptions.
For applications at very high energies, such as LHC, the use of
single precision for any real variable is a problem. It might
then be necessary to rewrite the program completely, i.e. to have a
declaration \ttt{IMPLICIT DOUBLE PRECISION(A-H,O-Z)} at the beginning
of each subprogram,
and to change all real constants to double precision. Needless to say,
the latter is a major undertaking. In some cases, shortcuts are
available. On the IBM, for instance, the \ttt{AUTODBL} compiler option
for automatic precision doubling works fine, provided only that an
even number of integers precede real numbers in common blocks.
In {\tsc{Jetset}} you therefore need to introduce an additional
integer variable (\ttt{NPAD}, say) directly after \ttt{N} in the
\ttt{LUJETS} common block, and in {\tsc{Pythia}} an additional integer
(\ttt{MSEPAD}) after \ttt{MSEL} in the \ttt{PYSUBS} common block.
Some pieces of code will then actually run in quadruple precision.
A test program, \ttt{LUTEST}\label{p:LUTEST}, is included in the
{\tsc{Jetset}} package. It is disguised as a subroutine, so you have to run a
main program
\begin{verbatim}
CALL LUTEST(1)
END
\end{verbatim}
This program will generate six hundred events of different types,
under a variety of conditions. If {\tsc{Jetset}} has not been properly
installed, this program is likely to crash, or at least generate a
number of erroneous events. This will then clearly be marked in the
output, which otherwise will just contain a few sample event listings
and a table of the number of different particles produced. To switch
off the output of normal events and final table, use \ttt{LUTEST(0)}
instead of \ttt{LUTEST(1)}. The final tally of errors detected should
read 0.
In exactly the same vein, a test program \ttt{PYTEST}\label{p:PYTEST}
comes with the {\tsc{Pythia}} package. You then have to run a program
\begin{verbatim}
CALL PYTEST(1)
END
\end{verbatim}
As before the alternative \ttt{PYTEST(0)} will give a less extensive
listing. No errors should appear during execution.
\subsection{Program Philosophy}
The Monte Carlo programs are built as slave systems, i.e. you,
the user, have to supply the main program. From this the various
subroutines are called on to execute specific tasks, after which
control is returned to the main program. Some of these tasks may
be very trivial, whereas the `high-level' routines by themselves
may make a large number of subroutine calls. Many routines are
not intended to be called directly by you, but only from
higher-level routines such as \ttt{LUEXEC}, \ttt{LUEEVT},
\ttt{PYINIT} or \ttt{PYEVNT}.
Basically, this means that there are three ways by which you
communicate with the
programs. First, by setting common block variables, you specify
the details of how the programs should perform specific tasks,
i.e. which subprocesses should be generated (for {\tsc{Pythia}}), which
particle masses should be assumed, which coupling constants used,
which fragmentation scenarios, and so on with hundreds of options
and parameters. Second, by calling subroutines you tell the programs
to generate events according to the rules established above.
Normally there are few subroutine arguments, and those are usually
related to details of the physical situation, such as what
c.m. energy to assume for events. Third, you can either look at the
common block \ttt{LUJETS} to extract information on the generated
event, or you can call on various functions and subroutines
to analyse the event further for you.
It should be noted that, while the physics content is obviously at
the centre of attention, the {\tsc{Jetset/Pythia}} package
also contains a very extensive setup of auxiliary service routines.
The hope is that this will provide a comfortable
working environment, where not only events are generated, but where
you also linger on to perform a lot of the subsequent studies.
Of course, for detailed studies, it may be necessary to interface
the output directly to a detector simulation program.
The general rule is that all routines have names that are six
characters long, beginning with \ttt{LU} for {\tsc{Jetset}} routines and
\ttt{PY} for {\tsc{Pythia}} ones. Real-valued functions in {\tsc{Jetset}} begin with
\ttt{UL} instead. There are three exceptions to both the length and
the initial character rules: \ttt{KLU}, \ttt{PLU} and \ttt{RLU}. The
former two functions are strongly coupled to the \ttt{K} and \ttt{P}
matrices in the \ttt{LUJETS} common block, the latter uses \ttt{R} to
emphasize the r\^ole as a random-number generator. Also common block
names are six characters long and start with \ttt{LU} or
\ttt{PY}.
On the issue of initialization, {\tsc{Jetset}} and {\tsc{Pythia}} behave
quite differently. Most {\tsc{Jetset}} routines work without any
initialization (except for the one implied by the presence of
\ttt{BLOCK DATA LUDATA}, see above), i.e. each event and each task
stand on their own. Current common block values are used to perform
the tasks in specific ways, and those rules can be changed from
one event to the next (or even within the generation of one and the
same event) without any penalty. The random-number generator
is initialized at the first call, but usually this is transparent.
Therefore the two {\tsc{Jetset}} routines \ttt{LUEEVT} (and some of
the routines called by it) and \ttt{LUONIA} are basically the only
ones to contain some elements of initialization, where
there are a few advantages if events are generated in a coherent
fashion, but even here the penalty for not doing it is small.
In {\tsc{Pythia}}, on the other hand, a sizeable amount of
initialization is performed in the \ttt{PYINIT} call, and thereafter
the events generated by \ttt{PYEVNT} all obey the rules established
at that point. Therefore common block variables that specify methods
to be used have to be set before the \ttt{PYINIT} call and then not
be changed afterwards, with few exceptions. Of course, it is possible
to perform several \ttt{PYINIT} calls in the same run, but there is
a significant time overhead involved, so this is not something one
would do for each new event.
Apart from writing a title page, giving a brief initialization
information, printing error messages if need be,
and responding to explicit requests for listings, all tasks of the
programs are performed `silently'. All output is directed to unit
\ttt{MSTU(11)}, by default 6, and it is up to you to set
this unit open for write. The only exceptions are
\ttt{RLUGET}, \ttt{RLUSET} and \ttt{LUUPDA} where, for obvious reasons,
the input/output file number is specified at each call. Here you
again have to see to it that proper read/write access is set.
The programs are extremely versatile, but the price to be paid for this
is having a large number of adjustable parameters and switches for
alternative modes of operation. No single user is ever likely to
need more than a fraction of the available options.
Since all these parameters and switches are assigned sensible default
values, there is no reason to worry about them until the need arises.
Unless explicitly stated (or obvious from the context) all switches and
parameters can be changed independently of each other. One should note,
however, that if only a few switches/parameters are changed, this may
result in an artificially bad agreement with data. Many disagreements
can often be cured by a subsequent retuning of some other parameters of
the model, in particular those that were once determined by
a comparison with data in the context of the default scenario.
For example, for $\e^+\e^-$ annihilation, such a retuning could involve one
QCD parameter ($\alpha_{\mathrm{s}}$ or $\Lambda$), the longitudinal fragmentation
function, and the average transverse fragmentation momentum.
The programs contain a number of checks that requested processes have
been implemented, that flavours specified for jet systems make sense,
that the energy is sufficient to allow hadronization, that the memory
space in \ttt{LUJETS} is large enough, etc. If anything goes wrong
that the program can catch (obviously this may not always be possible),
an error message will be printed and the treatment of the corresponding
event will be cut short. In serious cases, the program will abort.
As long as no error messages appear on the
output, it may not be worthwhile to look into the rules for error
checking, but if but one message appears, it should be enough cause for
alarm to receive prompt attention. Also warnings are sometimes printed.
These are less serious, and the experienced user might deliberately
do operations which go against the rules, but still can be made to
make sense in their context. Only the first few warnings will be
printed, thereafter the program will be quiet. By default, the program
is set to stop execution after ten errors, after printing
the last erroneous event.
It must be emphasized that not all errors will be caught. In particular,
one tricky question is what happens if an integer-valued common block
switch or subroutine/function argument is used with a value that is not
defined. In some subroutine calls, a prompt return will be expedited,
but in most instances the subsequent action is entirely unpredictable,
and often completely haywire. The same goes for real-valued variables
that are assigned values outside the physically sensible range. One
example will suffice here: if \ttt{PARJ(2)} is defined as the
$\mathrm{s} / \u$ suppression factor, a value $>1$ will not give more
profuse production of $\mathrm{s}$ than of $\u$, but actually a spillover
into $\c$ production. Users, beware!
\subsection{Manual Conventions}
In the manual parts of this report, some conventions are used.
All names of subprograms, common blocks and variables are given in
upper-case `typewriter' style, e.g. \ttt{MSTP(111)=0}. Also
program examples are given in this style.
If a common block variable must have a value set at the beginning of
execution, then a default value is stored in one of the block data
subprograms \ttt{LUDATA} and \ttt{PYDATA}. Such a default value is
usually indicated by a `(D=\ldots)' immediately after the variable
name, e.g.
\begin{entry}
\iteme{MSTJ(1) :} (D=1) choice of fragmentation scheme.
\end{entry}
All variables in the {\tsc{Jetset}} common blocks (with very few exceptions,
clearly marked) can be freely changed from one event to the next,
or even within the treatment of one single event. In the {\tsc{Pythia}}
common blocks the situation is more complicated. The values of many
switches and parameters are used already in the \ttt{PYINIT} call,
and cannot be changed after that. The problem is mentioned in the
preamble to the afflicted common blocks, which in particular means
\ttt{/PYPARS/} and \ttt{/PYSUBS/}. For the variables which may still
be changed from one event to the next, a `(C)' is added after
the `(D=\ldots)' statement.
Normally, variables internal to the program are kept in separate
common blocks and arrays, but in a few cases such internal variables
appear among arrays of switches and parameters, mainly for historical
reasons. These are denoted by `(R)' for variables you may want to
read, because they contain potentially interesting information, and
by `(I)' for purely internal variables. In neither case may the
variables be changed by you.
In the description of a switch, the alternatives that this
switch may take are often enumerated, e.g.
\begin{entry}
\iteme{MSTJ(1) :} (D=1) choice of fragmentation scheme.
\begin{subentry}
\iteme{= 0 :} no jet fragmentation at all.
\iteme{= 1 :} string fragmentation according to the Lund model.
\iteme{= 2 :} independent fragmentation, according to specification
in \ttt{MSTJ(2)} and \ttt{MSTJ(3)}.
\end{subentry}
\end{entry}
If you then use any value other than 0, 1 or 2, results are
unpredictable. The action could even be different in different parts
of the program, depending on the order in which the alternatives are
identified.
It is also up to you to choose physically sensible values for
parameters: there is no check on the allowed ranges of variables.
We gave an example of this at the end of the preceding section.
Subroutines you are expected to use are enclosed in a box at the
point where they are defined:
\drawbox{CALL LULIST(MLIST)}
\boxsep
This is followed by a description of input or output parameters.
The difference between input and output is not explicitly marked,
but should be obvious from the context. In fact, the event-analysis
routines of section \ref{ss:evanrout} return values,
while all the rest only have input variables.
Routines that are only used internally are not boxed in.
However, we use boxes for all common blocks, so as to
enhance the readability.
\subsection{Getting Started with JETSET}
\label{ss:JETstarted}
As a first example, assume that you want to study the production of
$\u \overline{\mathrm{u}}$ 2-jet systems at 20 GeV energy. To do this, write a main
program
\begin{verbatim}
CALL LU2ENT(0,2,-2,20.)
CALL LULIST(1)
END
\end{verbatim}
and run this program, linked together with {\tsc{Jetset}}. The routine
\ttt{LU2ENT}
is specifically intended for storing two entries (jets or particles).
The first argument (0) is a command to perform fragmentation and decay
directly after the entries have been stored, the second and third that
the two entries are $\u$ (2) and $\overline{\mathrm{u}}$ ($-2$), and the last that the
c.m. energy of the pair is 20 GeV. When this is run, the resulting event
is stored in the \ttt{LUJETS} common block. This information can then be
read out by you. No output is produced by \ttt{LU2ENT} itself,
except for a title page which appears once for every {\tsc{Jetset/Pythia}} run.
Instead the second command, to \ttt{LULIST}, provides a simple visible
summary of the information stored in \ttt{LUJETS}. The argument (1)
indicates that the short version should be used, which is suitable
for viewing the listing directly on an 80-column terminal screen.
It might look as shown here.
\begin{verbatim}
Event listing (summary)
I particle/jet KS KF orig p_x p_y p_z E m
1 (u) A 12 2 0 0.000 0.000 10.000 10.000 0.006
2 (u~) V 11 -2 0 0.000 0.000 -10.000 10.000 0.006
3 (string) 11 92 1 0.000 0.000 0.000 20.000 20.000
4 (rho+) 11 213 3 0.098 -0.154 2.710 2.856 0.885
5 (rho-) 11 -213 3 -0.227 0.145 6.538 6.590 0.781
6 pi+ 1 211 3 0.125 -0.266 0.097 0.339 0.140
7 (Sigma0) 11 3212 3 -0.254 0.034 -1.397 1.855 1.193
8 (K*+) 11 323 3 -0.124 0.709 -2.753 2.968 0.846
9 p~- 1 -2212 3 0.395 -0.614 -3.806 3.988 0.938
10 pi- 1 -211 3 -0.013 0.146 -1.389 1.403 0.140
11 pi+ 1 211 4 0.109 -0.456 2.164 2.218 0.140
12 (pi0) 11 111 4 -0.011 0.301 0.546 0.638 0.135
13 pi- 1 -211 5 0.089 0.343 2.089 2.124 0.140
14 (pi0) 11 111 5 -0.316 -0.197 4.449 4.467 0.135
15 (Lambda0) 11 3122 7 -0.208 0.014 -1.403 1.804 1.116
16 gamma 1 22 7 -0.046 0.020 0.006 0.050 0.000
17 K+ 1 321 8 -0.084 0.299 -2.139 2.217 0.494
18 (pi0) 11 111 8 -0.040 0.410 -0.614 0.751 0.135
19 gamma 1 22 12 0.059 0.146 0.224 0.274 0.000
20 gamma 1 22 12 -0.070 0.155 0.322 0.364 0.000
21 gamma 1 22 14 -0.322 -0.162 4.027 4.043 0.000
22 gamma 1 22 14 0.006 -0.035 0.422 0.423 0.000
23 p+ 1 2212 15 -0.178 0.033 -1.343 1.649 0.938
24 pi- 1 -211 15 -0.030 -0.018 -0.059 0.156 0.140
25 gamma 1 22 18 -0.006 0.384 -0.585 0.699 0.000
26 gamma 1 22 18 -0.034 0.026 -0.029 0.052 0.000
sum: 0.00 0.000 0.000 0.000 20.000 20.000
\end{verbatim}
(A few blanks have been removed between the columns to make it fit into
the format of this text.) Look in the particle/jet column and note
that the first two lines are the original $\u$ and $\overline{\mathrm{u}}$, where `bar'
is actually written `$\sim$' to save space in longer names.
The parentheses enclosing the names, `\ttt{(u)}' and `\ttt{(u\~{})}',
are there as a reminder that these jets actually have been allowed
to fragment. The jets are still retained so that event histories
can be studied. Also note that the \ttt{KF} (flavour code) column
contains 2 in the first line and $-2$ in the second. These are the
codes actually stored to denote the presence of a $\u$ and a $\overline{\mathrm{u}}$,
cf. the \ttt{LU2ENT} call, while the names written are just
conveniences used when producing visible output. The \ttt{A} and
\ttt{V} near the end of the particle/jet column indicate the beginning
and end of a string (or cluster, or independent fragmentation) parton
system; any intermediate entries belonging to the same system would
have had an \ttt{I} in that column. (This gives a poor man's
representation of an up-down arrow, $\updownarrow$.)
In the \ttt{orig} (origin) column, the zeros indicate that $\u$ and
$\overline{\mathrm{u}}$ are two initial entries. The subsequent line, number 3, denotes
the fragmenting $\u \overline{\mathrm{u}}$ string system as a whole, and has origin 1,
since the first parton of this string system is entry number 1. The
particles in lines 4--10 have origin 3 to denote that they come
directly from the fragmentation of this string. In string
fragmentation it is not meaningful to say that a particle comes from
only the $\u$ quark or only the $\overline{\mathrm{u}}$ one. It is the string system
as a whole that gives a $\rho^+$, a $\rho^-$, a $\pi^+$, a
$\Sigma^0$, a $\mathrm{K}^{*+}$, a $\overline{\mathrm{p}}^-$, and a $\pi^-$. Note that some
of the particle names are again enclosed in parentheses, indicating
that these particles are not present in the final state either, but have
decayed further. Thus the $\pi^-$ in line 13 and the $\pi^0$ in line
14 have origin 5, as an indication that they come from the decay of the
$\rho^-$ in line 5. Only the names not enclosed in parentheses
remain at the end of the fragmentation/decay chain, and
are thus experimentally observable. The actual status code used to
distinguish between different classes of entries is given in the
\ttt{KS} column; codes in the range 1--10 correspond to remaining
entries, and those above 10 to those that have fragmented or decayed.
The columns with \ttt{p\_x}, \ttt{p\_y}, \ttt{p\_z}, \ttt{E} and
\ttt{m} are quite self-explanatory. All momenta, energies and
masses are given in units of GeV, since the speed of light
is taken to be $c = 1$.
Note that energy and momentum are conserved at each step of the
fragmentation/decay process (although there exist options where this is
not true). Also note that the $z$ axis plays the r\^ole of preferred
direction, along which the original partons are placed.
The final line is
intended as a quick check that nothing funny happened. It contains the
summed charge, summed momentum, summed energy and invariant mass of the
final entries at the end of the fragmentation/decay chain, and the
values should agree with the input implied by the \ttt{LU2ENT}
arguments. (In fact, warnings would normally appear on the output if
anything untoward happened, but that is another story.)
The above example has illustrated roughly what information is to be had
in the event record, but not so much about how it is stored. This is
better seen by using a 132-column format for listing events. Try e.g.
the following program
\begin{verbatim}
CALL LU3ENT(0,1,21,-1,30.,0.9,0.7)
CALL LULIST(2)
CALL LUEDIT(3)
CALL LULIST(2)
END
\end{verbatim}
where a 3-jet $\d \mathrm{g} \overline{\mathrm{d}}$ event is generated in the first line and
listed in the second. This listing will contain the numbers as
directly stored in the common block \ttt{LUJETS}
\begin{verbatim}
COMMON/LUJETS/N,K(4000,5),P(4000,5),V(4000,5)
\end{verbatim}
For particle \ttt{I}, \ttt{K(I,1)} thus gives information on whether
or not a jet or particle has fragmented or decayed, \ttt{K(I,2)}
gives the particle code, \ttt{K(I,3)} its origin, \ttt{K(I,4)} and
\ttt{K(I,5)} the position of fragmentation/decay products, and
\ttt{P(I,1})--\ttt{P(I,5)} momentum, energy and mass. The number
of lines in current use is given by \ttt{N}, i.e.
1 $\leq$ \ttt{I} $\leq$ \ttt{N}. The \ttt{V} matrix contains decay
vertices; to view those \ttt{LULIST(3)} has to be used. It
is important to learn the rules for how information is stored in
\ttt{LUJETS}.
The third line in the program illustrates another important point
about {\tsc{Jetset}}: a number of routines are available for
manipulating the event record after the event has been generated.
Thus \ttt{LUEDIT(3)} will remove everything except stable charged
particles, as shown by the result of the second \ttt{LULIST} call.
More advanced possibilities include things like sphericity or
clustering routines.
Apart from the input arguments of subroutine calls, control on the
doings of {\tsc{Jetset}} may be imposed via the \ttt{LUDAT1},
\ttt{LUDAT2}, \ttt{LUDAT3} and \ttt{LUDAT4} common blocks. Here
sensible default values are always provided. A user might want to
switch off all particle decays by putting \ttt{MSTJ(21)=0} or
increase the $\mathrm{s} / \u$ ratio in fragmentation by putting
\ttt{PARJ(2)=0.40}, to give but two examples. It is by exploring
the possibilities offered here that {\tsc{Jetset}} can be turned into an
extremely versatile tool, even if all the nice physics is already
present in the default values.
As a final, semirealistic example, assume that the $p_{\perp}$
spectrum of $\pi^+$ particles is to be studied in 91.2 GeV
$\e^+\e^-$ annihilation events, where $p_{\perp}$ is to be defined with
respect to the sphericity axis. Using the HBOOK package (version 4,
watch out for version- or installation-specific differences) for
histogramming, a complete program might look like
\begin{verbatim}
C...Common blocks.
COMMON/LUJETS/N,K(4000,5),P(4000,5),V(4000,5)
COMMON/PAWC/HMEMOR(10000)
C...Reserve histogram memory and book histograms.
CALL HLIMIT(10000)
CALL HBOOK1(1,'pT spectrum of pi+',100,0.,5.,0.)
C...Number of events to generate. Loop over events.
NEVT=100
DO 110 IEVT=1,NEVT
C...Generate event. List first one.
CALL LUEEVT(0,91.2)
IF(IEVT.EQ.1) CALL LULIST(1)
C...Find sphericity axis and rotate event so sphericity along z axis.
CALL LUSPHE(SPH,APL)
CALL LUEDIT(31)
C...Loop over all particles, but skip if not pi+.
DO 100 I=1,N
IF(K(I,2).NE.211) GOTO 100
C...Calculate pT and fill in histogram.
PT=SQRT(P(I,1)**2+P(I,2)**2)
CALL HF1(1,PT,1.)
C...End of particle and event loops.
100 CONTINUE
110 CONTINUE
C...Normalize histogram properly and list it.
CALL HOPERA(1,'+',1,1,20./NEVT,0.)
CALL HISTDO
END
\end{verbatim}
Study this program, try to understand what happens at each step, and
run it to check that it works. You should then be ready to look
at the relevant sections of this report and start writing your own
programs.
\subsection{Getting Started with PYTHIA}
\label{ss:PYTstarted}
A {\tsc{Pythia}} run has to be more strictly organized than
a {\tsc{Jetset}} one, in that it is necessary to initialize the
generation before events can be generated, and in that it is
not possible to change switches and parameters freely during
the course of the run. A fairly precise recipe for how a run
should be structured can therefore be given.
Thus, the usage of {\tsc{Pythia}} can be subdivided into three steps.
\begin{Enumerate}
\item The initialization step. It is here that all the basic
characteristics of the coming generation are specified.
The material in this section includes the following.
\begin{Itemize}
\item Common blocks, at least the following, and maybe some more:
\begin{verbatim}
COMMON/LUJETS/N,K(4000,5),P(4000,5),V(4000,5)
COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)
COMMON/PYSUBS/MSEL,MSUB(200),KFIN(2,-40:40),CKIN(200)
COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)
\end{verbatim}
\vspace{-\baselineskip}
\item Selection of required processes. Some fixed `menus'
of subprocesses can be selected with different \ttt{MSEL}
values, but with {\tt MSEL}=0 it is possible to compose
`\`a la carte', using the subprocess numbers.
To generate processes 14, 18 and 29, for instance, one needs
\begin{verbatim}
MSEL=0
MSUB(14)=1
MSUB(18)=1
MSUB(29)=1
\end{verbatim}
\vspace{-\baselineskip}
\item Selection of kinematics cuts in the \ttt{CKIN} array.
To generate hard scatterings with 5~GeV $\leq p_{\perp} \leq$
10~GeV, for instance, use
\begin{verbatim}
CKIN(3)=5.
CKIN(4)=10.
\end{verbatim}
\vspace{-\baselineskip}
Unfortunately, initial- and final-state radiation will shift
around the kinematics of the hard scattering, making the effects
of cuts less predictable. One therefore always has to be very
careful that no desired event configurations are cut out.
\item Definition of underlying physics scenario, e.g. top mass.
\item Selection of parton-distribution sets, $Q^2$ definitions,
and all other details of the generation.
\item Switching off of generator parts not needed for toy
simulations, e.g. fragmentation for parton level studies.
\item Initialization of the event generation procedure. Here
kinematics is set up, maxima of differential cross sections
are found for future Monte Carlo generation, and a number of
other preparatory tasks carried out. Initialization is performed
by \ttt{PYINIT}, which should be called only after the switches
and parameters above have been set to their desired values. The
frame, the beam particles and the energy have to be specified.
\begin{verbatim}
CALL PYINIT('CMS','p','pbar',1800.)
\end{verbatim}
\vspace{-\baselineskip}
\item Any other initial material required by the user, e.g.
histogram booking.
\end{Itemize}
\item The generation loop. It is here that events are generated
and studied. It includes the following tasks:
\begin{Itemize}
\item Generation of the next event, with
\begin{verbatim}
CALL PYEVNT
\end{verbatim}
\vspace{-\baselineskip}
\item Printing of a few events, to check that everything is
working as planned, with
\begin{verbatim}
CALL LULIST(1)
\end{verbatim}
\vspace{-\baselineskip}
\item An analysis of the event for properties of interest,
either directly reading out information from the
\ttt{LUJETS} common block
or making use of a number of utility routines in {\tsc{Jetset}}.
\item Saving of events on tape, or interfacing to detector
simulation.
\end{Itemize}
\item The finishing step. Here the tasks are:
\begin{Itemize}
\item Printing a table of deduced cross sections, obtained as a
by-product of the Monte Carlo generation activity, with the
command
\begin{verbatim}
CALL PYSTAT(1)
\end{verbatim}
\vspace{-\baselineskip}
\item Printing histograms and other user output.
\end{Itemize}
\end{Enumerate}
To illustrate this structure, imagine a toy example, where one wants
to simulate the production of a 300 GeV Higgs particle. In
{\tsc{Pythia}}, a program for this might look something like the
following.
\begin{verbatim}
C...Common blocks.
COMMON/LUJETS/N,K(4000,5),P(4000,5),V(4000,5)
COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)
COMMON/LUDAT2/KCHG(500,3),PMAS(500,4),PARF(2000),VCKM(4,4)
COMMON/LUDAT3/MDCY(500,3),MDME(2000,2),BRAT(2000),KFDP(2000,5)
COMMON/PYSUBS/MSEL,MSUB(200),KFIN(2,-40:40),CKIN(200)
COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)
COMMON/PAWC/HBOOK(10000)
C...Number of events to generate. Switch on proper processes.
NEV=1000
MSEL=0
MSUB(102)=1
MSUB(123)=1
MSUB(124)=1
C...Select t and H masses and kinematics cuts in mass.
PMAS(6,1)=140.
PMAS(25,1)=300.
CKIN(1)=290.
CKIN(2)=310.
C...For simulation of hard process only: cut out unnecessary tasks.
MSTP(61)=0
MSTP(71)=0
MSTP(81)=0
MSTP(111)=0
C...Initialize and list partial widths.
CALL PYINIT('CMS','p','p',16000.)
CALL PYSTAT(2)
C...Book histograms.
CALL HLIMIT(10000)
CALL HBOOK1(1,'Higgs mass',50,275.,325.,0.)
C...Generate events. Look at first few.
DO 200 IEV=1,NEV
CALL PYEVNT
IF(IEV.LE.3) CALL LULIST(1)
C...Loop over particles to find Higgs and histogram its mass.
DO 100 I=1,N
100 IF(K(I,2).EQ.25) HMASS=P(I,5)
CALL HF1(1,HMASS,1.)
200 CONTINUE
C...Print cross sections and histograms.
CALL PYSTAT(1)
CALL HISTDO
END
\end{verbatim}
Here 102, 123 and 124 are the three main Higgs production
graphs $\mathrm{g} \mathrm{g} \rightarrow \H$, $\mathrm{Z} \mathrm{Z} \rightarrow \H$, and
$\mathrm{W} \mathrm{W} \rightarrow \H$, and \ttt{MSUB(ISUB)=1} is the command to
switch on process \ttt{ISUB}. Full freedom to combine subprocesses
`\`a la carte' is ensured by \ttt{MSEL=0}; ready-made `menus'
can be ordered with other \ttt{MSEL} numbers.
The \ttt{PMAS} commands set the
masses of the top quark and the Higgs itself, and the \ttt{CKIN}
variables the desired mass range of the Higgs --- a Higgs with a
300 GeV nominal mass actually has a fairly broad Breit--Wigner type
mass distribution. The \ttt{MSTP} switches that come next are there to
modify the generation procedure, in this case to switch off initial-
and final-state radiation, multiple interactions among beam jets,
and fragmentation, to give only the `parton skeleton' of the hard
process. The \ttt{PYINIT} call initializes {\tsc{Pythia}}, by finding
maxima of cross sections, recalculating the Higgs decay properties
(which depend on the Higgs mass), etc. The decay properties can
be listed with \ttt{PYSTAT(2)}.
Inside the event loop, \ttt{PYEVNT}
is called to generate an event, and \ttt{LULIST(1)} to list the event.
The information used by \ttt{LULIST(1)} is the event record, stored
in the common block \ttt{LUJETS}. Here one finds all produced particles,
both final and intermediate ones, with information on particle
species and event history (\ttt{K} array), particle momenta
(\ttt{P} array) and production vertices (\ttt{V} array).
In the loop over all particles produced, \ttt{1} through \ttt{N},
the Higgs particle is found by its code, \ttt{K(I,2)=25},
and its mass is stored in \ttt{P(I,5)}.
After all events have been generated, \ttt{PYSTAT(1)}
gives a summary of the number of events generated in the
various allowed channels, and the inferred cross sections.
In the run above, a typical event listing might look like the following.
\begin{verbatim}
Event listing (summary)
I particle/jet KF p_x p_y p_z E m
1 !p+! 2212 0.000 0.000 8000.000 8000.000 0.938
2 !p+! 2212 0.000 0.000-8000.000 8000.000 0.938
======================================================================
3 !g! 21 -0.505 -0.229 28.553 28.558 0.000
4 !g! 21 0.224 0.041 -788.073 788.073 0.000
5 !g! 21 -0.505 -0.229 28.553 28.558 0.000
6 !g! 21 0.224 0.041 -788.073 788.073 0.000
7 !H0! 25 -0.281 -0.188 -759.520 816.631 300.027
8 !W+! 24 120.648 35.239 -397.843 424.829 80.023
9 !W-! -24 -120.929 -35.426 -361.677 391.801 82.579
10 !e+! -11 12.922 -4.760 -160.940 161.528 0.001
11 !nu_e! 12 107.726 39.999 -236.903 263.302 0.000
12 !s! 3 -62.423 7.195 -256.713 264.292 0.199
13 !c~! -4 -58.506 -42.621 -104.963 127.509 1.350
======================================================================
14 (H0) 25 -0.281 -0.188 -759.520 816.631 300.027
15 (W+) 24 120.648 35.239 -397.843 424.829 80.023
16 (W-) -24 -120.929 -35.426 -361.677 391.801 82.579
17 e+ -11 12.922 -4.760 -160.940 161.528 0.001
18 nu_e 12 107.726 39.999 -236.903 263.302 0.000
19 s A 3 -62.423 7.195 -256.713 264.292 0.199
20 c~ V -4 -58.506 -42.621 -104.963 127.509 1.350
21 ud_1 A 2103 -0.101 0.176 7971.328 7971.328 0.771
22 d V 1 -0.316 0.001 -87.390 87.390 0.010
23 u A 2 0.606 0.052 -0.751 0.967 0.006
24 uu_1 V 2203 0.092 -0.042-7123.668 7123.668 0.771
======================================================================
sum: 2.00 0.00 0.00 0.00 15999.98 15999.98
\end{verbatim}
The above event listing is abnormally short, in part because some
columns of information were removed to make it fit into this text,
in part because all initial- and final-state QCD radiation, all
non-trivial beam jet structure, and all fragmentation was inhibited
in the generation. Therefore only the skeleton of the process is
visible. In lines 1 and 2 one recognizes the two incoming protons.
In lines 3 and 4 are incoming partons before initial-state radiation
and in 5 and 6 after --- since there is no such radiation they coincide
here. Line 7 shows the Higgs produced by $\mathrm{g} \mathrm{g}$ fusion, 8 and 9 its
decay products and 10--13 the second-step decay products. Up to this
point lines give a summary of the event history, indicated by the
exclamation marks that surround particle names (and also reflected in
the \ttt{K(I,1)} code, not shown). From line 14 onwards come the
particles actually produced in the final states, first in lines 14--16
particles that subsequently decayed, which have their names surrounded
by brackets, and finally the particles and jets left in the end,
including beam remnants.
Here this also includes a number of unfragmented jets, since
fragmentation was inhibited. Ordinarily, the listing would have gone
on for a few hundred more lines, with the particles produced in the
fragmentation and their decay products. The final line gives total
charge and momentum, as a convenient check that nothing unexpected
happened. The first column of the listing
is just a counter, the second gives the particle name and information
on status and string drawing (the \ttt{A} and \ttt{V}), the third
the particle-flavour code (which is used to give the name),
and the subsequent columns give the momentum components.
One of the main problems is to select kinematics efficiently. Imagine
for instance that one is interested in the production of a single
$\mathrm{Z}$ with a
transverse momentum in excess of 50 GeV. If one tries to generate
the inclusive sample of $\mathrm{Z}$ events, by the basic production graphs
$\mathrm{q} \overline{\mathrm{q}} \rightarrow \mathrm{Z}$, then most events will have low transverse
momenta and will have to be discarded. That any of the desired events
are produced at all is due to the initial-state generation machinery,
which can build up transverse momenta for the incoming
$\mathrm{q}$ and $\overline{\mathrm{q}}$. However, the amount
of initial-state radiation cannot be constrained beforehand. To
increase the efficiency, one may therefore turn to the higher-order
processes $\mathrm{q} \mathrm{g} \rightarrow \mathrm{Z} \mathrm{q}$ and $\mathrm{q} \overline{\mathrm{q}} \rightarrow \mathrm{Z} \mathrm{g}$,
where
already the hard subprocess gives a transverse momentum to the $\mathrm{Z}$.
This transverse momentum can be constrained as one wishes, but again
initial- and final-state radiation will smear the picture. If one were
to set a $p_{\perp}$ cut at 50 GeV for the hard-process generation,
those events where the $\mathrm{Z}$ was given only 40 GeV in the hard process
but got the rest from initial-state radiation would be missed.
Not only therefore would cross sections
come out wrong, but so might the typical event shapes. In the end,
it is therefore necessary to find some reasonable compromise, by
starting the generation at 30 GeV, say, if one knows that only rarely
do events below this value fluctuate up to 50 GeV. Of course, most
events will therefore not contain a $\mathrm{Z}$ above 50 GeV, and one will
have to live with some inefficiency. It is not uncommon that only one
event out of ten can be used, and occasionally it can be even worse.
If it is difficult to set kinematics, it is often easier to set the
flavour content of a process. In a Higgs study, one might wish, for
example, to consider the decay $\H^0 \rightarrow \mathrm{Z}^0 \mathrm{Z}^0$, with each
$\mathrm{Z}^0 \rightarrow \e^+\e^-$ or $\mu^+ \mu^-$. It is therefore necessary to
inhibit all other $\H^0$ and $\mathrm{Z}^0$ decay channels, and also to adjust
cross sections to take into account this change, all of which is fairly
straightforward. However, if one wanted to consider instead the
decay $\mathrm{Z}^0 \rightarrow \c \overline{\mathrm{c}}$, with a $\mathrm{D}$ meson
producing a lepton, not only
would there then be the problem of different leptonic branching ratios
for different $\mathrm{D}$:s (which means that fragmentation and decay
treatments would no longer decouple), but also that of additional
$\c \overline{\mathrm{c}}$ pair production in parton-shower evolution, at a rate
that is unknown beforehand. In practice, it is therefore impossible to
force $\mathrm{D}$ decay modes in a consistent manner.
\clearpage
\section{Monte Carlo Techniques}
Quantum mechanics introduces a concept of randomness in the behaviour
of physical processes. The virtue of event generators is that this
randomness can be simulated by the use of Monte Carlo techniques.
In the process, the program authors have to use some ingenuity to
find the most efficient way to simulate an assumed probability
distribution. A detailed description of possible techniques would
carry us too far, but
in this section some of the most frequently used approaches are
presented, since they will appear in discussions in
subsequent sections. Further examples may be found e.g. in
\cite{Jam80}.
First of all one assumes the existence of a random number generator.
This is a (Fortran) function which, each time it is called, returns
a number $R$ in the range between 0 and 1, such that the inclusive
distribution of numbers $R$ is flat in the range,
and such that different numbers $R$ are uncorrelated. The random
number generator that comes with {\tsc{Jetset}} is described at the
end of this section, and we defer the discussion until then.
\subsection{Selection From a Distribution}
\label{ss:MCdistsel}
The situation that is probably most common is that we know a
function $f(x)$ which is non-negative in the allowed $x$ range
$x_{\mathrm{min}} \leq x \leq x_{\mathrm{max}}$. We want to select an $x$
`at random' so that the probability for a given $x$ is
proportional to $f(x)$. Here $f(x)$ might be a fragmentation function,
a differential cross section, or any of a number of distributions.
One does not
have to assume that the integral of $f(x)$ is explicitly normalized
to unity: by the Monte Carlo procedure of picking exactly one
accepted $x$ value, normalization is implicit in the final result.
Sometimes the integral of $f(x)$ does carry a physics content of
its own, as part of an overall weight factor
we want to keep track of. Consider, for instance, the case when $x$
represents one or several phase-space variables and $f(x)$ a
differential cross section; here the integral has a meaning of total
cross section for the process studied. The task of a Monte Carlo is
then, on the one hand, to generate events one at a time, and, on the
other hand, to estimate the total cross section.
The discussion of this important example is
deferred to section \ref{ss:PYTcrosscalc}.
If it is possible to find a primitive function $F(x)$ which
has a known inverse $F^{-1}(x)$, an $x$ can be found as
follows (method 1):
\begin{eqnarray}
& \displaystyle{ \int_{x_{\mathrm{min}}}^{x} f(x) \, \d x =
R \int_{x_{\mathrm{min}}}^{x_{\mathrm{max}}} f(x) \, \d x } & \nonumber \\[1mm]
\Longrightarrow & x = F^{-1}(F(x_{\mathrm{min}}) +
R(F(x_{\mathrm{max}}) - F(x_{\mathrm{min}}))) ~. &
\end{eqnarray}
The statement of the first line is that a fraction $R$ of the
total area under $f(x)$ should be to the left of $x$.
However, seldom are functions of interest so nice that the
method above works. It is therefore necessary to use more complicated
schemes.
Special tricks can sometimes be found. Consider e.g. the generation
of a Gaussian $f(x) = \exp(-x^2)$. This function is not integrable,
but if we combine it with the same Gaussian distribution of a second
variable $y$, it is possible to transform to polar coordinates
\begin{equation}
f(x) \, \d x \, f(y) \, \d y = \exp(-x^2-y^2) \, \d x \, \d y =
r \exp(-r^2) \, \d r \, \d \varphi ~,
\end{equation}
and now the $r$ and $\varphi$ distributions may be easily generated
and recombined to yield $x$. At the same time we get a second number
$y$, which can also be used. For the generation of transverse momenta
in fragmentation, this is very convenient, since in fact we want to
assign two transverse degrees of freedom.
If the maximum of $f(x)$ is known, $f(x) \leq f_{\mathrm{max}}$ in the
$x$ range considered, a hit-or-miss method will always yield the
correct answer (method 2):
\begin{Enumerate}
\item select an $x$ with even probability in the allowed range, i.e.
$x = x_{\mathrm{min}} + R(x_{\mathrm{max}} - x_{\mathrm{min}})$;
\item compare a (new) $R$ with the ratio $f(x)/f_{\mathrm{max}}$;
if $f(x)/f_{\mathrm{max}} \le R$, then reject the $x$ value
and return to point 1 for a new try;
\item otherwise the most recent $x$ value is retained as final answer.
\end{Enumerate}
The probability that $f(x)/f_{\mathrm{max}} > R$ is proportional to
$f(x)$; hence the correct distribution of retained $x$ values.
The efficiency of this method, i.e. the average probability that
an $x$ will be retained, is $(\int \, f(x) \, \d x)
/ (f_{\mathrm{max}}(x_{\mathrm{max}} - x_{\mathrm{min}}))$.
The method is acceptable if this number is not too low, i.e. if
$f(x)$ does not fluctuate too wildly.
Very often $f(x)$ does have narrow spikes, and it may not even be
possible to define an $f_{\mathrm{max}}$. An example of the former
phenomenon is a
function with a singularity just outside the allowed region,
an example of the latter an integrable singularity just at the
$x_{\mathrm{min}}$ and/or $x_{\mathrm{max}}$ borders.
Variable transformations may then be used to make a function
smoother. Thus a function $f(x)$ which blows up as $1/x$ for
$x \rightarrow 0$, with an $x_{\mathrm{min}}$ close to 0, would instead
be roughly constant if transformed to the variable $y = \ln x$.
The variable transformation strategy may be seen as a combination
of methods 1 and 2, as follows. Assume the existence of a function
$g(x)$, with $f(x) \leq g(x)$ over the $x$ range of interest.
Here $g(x)$ is picked to be a `simple' function, such that the
primitive function $G(x)$ and its inverse $G^{-1}(x)$ are known.
Then (method 3):
\begin{Enumerate}
\item select an $x$ according to the distribution $g(x)$, using
method 1;
\item compare a (new) $R$ with the ratio $f(x)/g(x)$;
if $f(x)/g(x) \le R$, then reject the $x$ value
and return to point 1 for a new try;
\item otherwise the most recent $x$ value is retained as final answer.
\end{Enumerate}
This works, since the first step will select $x$ with a probability
$g(x) \, \d x = \d G(x)$ and the second retain this choice with
probability $f(x)/g(x)$. The total probability to pick a value $x$
is then just the product of the two, i.e. $f(x) \, \d x$.
If $f(x)$ has several spikes, method 3 may work for each spike
separately, but it may not be possible to find a $g(x)$ that
covers all of them at the same time, and which still has an
invertible primitive function. However, assume that
we can find a function $g(x) = \sum_i g_i(x)$,
such that $f(x) \leq g(x)$ over
the $x$ range considered, and such that the functions $g_i(x)$
each are non-negative and simple, in the sense that we can find
primitive functions and their inverses. In that case (method 4):
\begin{Enumerate}
\item select an $i$ at random, with relative probability given
by the integrals
\begin{equation}
\int_{x_{\mathrm{min}}}^{x_{\mathrm{max}}} g_i(x) \, \d x =
G_i(x_{\mathrm{max}}) - G_i(x_{\mathrm{min}}) ~; \nonumber
\end{equation}
\item for the $i$ selected, use method 1 to find an $x$, i.e.
\begin{equation}
x = G_i^{-1}(G_i(x_{\mathrm{min}}) +
R(G_i(x_{\mathrm{max}})-G_i(x_{\mathrm{min}}))) ~;
\nonumber
\end{equation}
\item compare a (new) $R$ with the ratio $f(x)/g(x)$;
if $f(x)/g(x) \le R$, then reject the $x$ value
and return to point 1 for a new try;
\item otherwise the most recent $x$ value is retained as final answer.
\end{Enumerate}
This is just a trivial extension of method 3, where steps 1 and 2
ensure that, on the average, each $x$ value picked there is
distributed according to $g(x)$: the first step picks $i$
with relative probability $\int g_i(x) \, \d x$, the second $x$ with
absolute probability $g_i(x) / \int g_i(x) \, \d x$ (this is one place
where one must remember to do normalization correctly); the
product of the two is therefore $g_i(x)$ and the sum over all $i$
gives back $g(x)$.
We have now arrived at an approach that is sufficiently powerful for
a large selection of problems. In general,
for a function $f(x)$ which is known to have sharp peaks in a few
different places, the generic behaviour at each peak separately
may be covered by one or a few simple functions $g_i(x)$, to which
one adds a few more $g_i(x)$ to cover the basic behaviour away
from the peaks. By a suitable selection
of the relative strengths of the different $g_i$'s, it is
possible to find a function $g(x)$ that matches well the general
behaviour of $f(x)$, and thus achieve a reasonable Monte Carlo
efficiency.
The major additional complication is when $x$ is a multidimensional
variable. Usually the problem is not so much $f(x)$ itself, but
rather that the phase-space boundaries may be very complicated.
If the boundaries factorize it is possible to pick phase-space
points restricted to the desired region. Otherwise the region may
have to be inscribed in a hyper-rectangle, with points picked within
the whole hyper-rectangle but only retained if they are inside the
allowed region. This may lead to a significant loss in efficiency.
Variable transformations may often make the allowed region easier to
handle.
There are two main methods to handle several dimensions, each with its
set of variations. The first method is based on a factorized ansatz,
i.e. one attempts to find a function $g(\mbf{x})$ which is
everywhere larger than $f(\mbf{x})$, and which can be factorized
into $g(\mbf{x}) = g^{(1)}(x_1) \, g^{(2)}(x_2) \cdots g^{(n)}(x_n)$,
where $\mbf{x} = (x_1,x_2,\ldots,x_n)$. Here each $g^{(j)}(x_j)$ may
in its turn be a sum of functions $g^{(j)}_i$, as in method 4 above.
First, each $x_j$ is selected independently, and afterwards the ratio
$f(\mbf{x})/g(\mbf{x})$ is used to determine whether to retain the
point.
The second method is useful if the boundaries of the allowed region
can be written in a form where the maximum range of $x_1$ is known,
the allowed range of $x_2$ only depends on $x_1$, that of $x_3$ only
on $x_1$ and $x_2$, and so on until $x_n$, whose range may depend on
all the
preceding variables. In that case it may be possible to find a function
$g(\mbf{x})$ that can be integrated over $x_2$ through $x_n$ to yield
a simple function of $x_1$, according to which $x_1$ is selected.
Having done that, $x_2$ is selected according to a distribution
which now depends on $x_1$, but with $x_3$ through $x_n$ integrated
over. In particular, the allowed range for $x_2$ is known.
The procedure is continued until $x_n$ is reached, where now the
function depends on all the preceding $x_j$ values. In the end, the
ratio $f(\mbf{x})/g(\mbf{x})$ is again used to determine whether to
retain the point.
\subsection{The Veto Algorithm}
\label{ss:vetoalg}
The `radioactive decay' type of problems is very common, in particular
in parton showers, but it is also used, e.g. in the multiple
interactions description in {\tsc{Pythia}}. In this kind of problems
there is one variable $t$, which may be thought of as giving a kind
of time axis along which different events are ordered. The
probability that `something will happen' (a nucleus decay, a
parton branch) at time $t$ is described by a function $f(t)$, which
is non-negative in the range of $t$ values to be studied. However,
this na\"{\i}ve probability is modified by the additional requirement
that something can only happen at time $t$ if it did not happen
at earlier times $t' < t$. (The original nucleus cannot decay once
again if it already did decay; possibly the decay products may decay
in their turn, but that is another question.)
The probability that nothing has happened by time $t$ is expressed by
the function ${\cal N}(t)$ and the differential probability that
something happens at time $t$ by ${\cal P}(t)$. The basic equation
then is
\begin{equation}
{\cal P}(t) = - \frac{\d {\cal N}}{\d t} = f(t) \, {\cal N}(t) ~.
\end{equation}
For simplicity, we shall assume that the process starts at time
$t = 0$, with ${\cal N}(0) = 1$.
The above equation can be solved easily if one notes that
$\d {\cal N} / {\cal N} = \d \ln {\cal N}$:
\begin{equation}
{\cal N}(t) = {\cal N}(0) \exp \left\{ - \int_0^t f(t') \, \d t'
\right\}
= \exp \left\{ - \int_0^t f(t') \, \d t' \right\} ~,
\end{equation}
and thus
\begin{equation}
{\cal P}(t) = f(t) \exp \left\{ - \int_0^t f(t') \, \d t' \right\} ~.
\label{mc:Pveto}
\end{equation}
With $f(t) = c$ this is nothing but the textbook formulae for
radioactive decay. In particular, at small times the correct
decay probability, ${\cal P}(t)$, agrees well with the input
one, $f(t)$, since the exponential factor is close to unity there.
At larger $t$, the exponential gives a dampening which ensures that
the integral of ${\cal P}(t)$ never can exceed unity, even if the
integral of $f(t)$ does. The exponential can be seen as the
probability that nothing happens between the original time 0 and
the final time $t$. In the parton-shower language, this is (almost)
the so-called Sudakov form factor.
If $f(t)$ has a primitive function with a known inverse, it is easy
to select $t$ values correctly:
\begin{equation}
\int_0^t {\cal P}(t') \, \d t' = {\cal N}(0) - {\cal N}(t) =
1 - \exp \left\{ - \int_0^t f(t') \, \d t' \right\} = 1 - R ~,
\end{equation}
which has the solution
\begin{equation}
F(0) - F(t) = \ln R ~~~ \Longrightarrow ~~~
t = F^{-1}(F(0) - \ln R) ~.
\end{equation}
If $f(t)$ is not sufficiently nice, one may again try to find a
better function $g(t)$, with $f(t) \leq g(t)$ for all $t \geq 0$.
However to use method 3 with this $g(t)$ would not work, since the
method would not correctly take
into account the effects of the exponential term in ${\cal P}(t)$.
Instead one may use the so-called veto algorithm:
\begin{Enumerate}
\item start with $i = 0$ and $t_0 = 0$;
\item add 1 to $i$ and select $t_i = G^{-1}(G(t_{i-1}) - \ln R)$,
i.e. according to $g(t)$, but with the constraint that
$t_i > t_{i-1}$,
\item compare a (new) $R$ with the ratio $f(t_i)/g(t_i)$;
if $f(t_i)/g(t_i) \le R$, then return to point 2 for a new try;
\item otherwise $t_{i}$ is retained as final answer.
\end{Enumerate}
It may not be apparent why this works. Consider, however, the various
ways in which one can select a specific time $t$. The probability that
the first try works, $t = t_1$, i.e. that no intermediate $t$ values
need be rejected, is given by
\begin{equation}
{\cal P}_0 (t) = \exp \left\{ - \int_0^t g(t') \, \d t' \right\}
\, g(t) \, \frac{f(t)}{g(t)}
= f(t) \exp \left\{ - \int_0^t g(t') \, \d t' \right\} ~,
\end{equation}
where the exponential times $g(t)$ comes from eq.~(\ref{mc:Pveto})
applied to $g$, and the ratio $f(t)/g(t)$ is the probability that
$t$ is
accepted. Now consider the case where one intermediate time $t_1$ is
rejected and $t = t_2$ is only accepted in the second step.
This gives
\begin{equation}
{\cal P}_1 (t) = \int_0^t \d t_1
\exp \left\{ - \int_0^{t_1} g(t') \, \d t' \right\} g(t_1)
\left[ 1 - \frac{f(t_1)}{g(t_1)} \right]
\exp \left\{ - \int_{t_1}^t g(t') \, \d t' \right\} g(t) \,
\frac{f(t)}{g(t)} ~,
\end{equation}
where the first exponential times $g(t_1)$ gives the probability
that $t_1$ is first selected, the square brackets the probability
that $t_1$
is subsequently rejected, the following piece the probability that
$t = t_2$ is selected when starting from $t_1$, and the final factor
that $t$ is retained. The whole is to be integrated over all
possible intermediate times $t_1$. The exponentials together give an
integral over the range from 0 to $t$, just as in ${\cal P}_0$,
and the factor for the final step being accepted is also the same,
so therefore one finds that
\begin{equation}
{\cal P}_1 (t) = {\cal P}_0 (t) \int_0^t \d t_1
\left[ g(t_1) - f(t_1) \right] ~.
\end{equation}
This generalizes.
In ${\cal P}_2$ one has to consider two intermediate times,
$0 \leq t_1 \leq t_2 \leq t_3 = t$, and so
\begin{eqnarray}
{\cal P}_2 (t) & = & {\cal P}_0 (t)
\int_0^t \d t_1 \left[ g(t_1) - f(t_1) \right]
\int_{t_1}^t \d t_2 \left[ g(t_2) - f(t_2) \right] \nonumber \\
& = & {\cal P}_0 (t) \frac{1}{2} \left(
\int_0^t \left[ g(t') - f(t') \right] \d t' \right)^2 ~.
\end{eqnarray}
The last equality is most easily seen if one also considers the
alternative region $0 \leq t_2 \leq t_1 \leq t$, where the
r\^oles of $t_1$ and $t_2$ have just been interchanged, and the
integral therefore has the same value as in the region considered.
Adding the two regions, however, the integrals over $t_1$ and
$t_2$ decouple, and become equal. In general, for ${\cal P}_i$,
the $i$ intermediate times can be ordered in $i!$ different ways.
Therefore the total probability to accept $t$, in any step, is
\begin{eqnarray}
{\cal P} (t) & = &
\displaystyle{ \sum_{i=0}^{\infty} {\cal P}_i (t)
= {\cal P}_0 (t) \sum_{i=0}^{\infty} \frac{1}{i!}
\left( \int_0^t \left[ g(t') - f(t') \right] \d t' \right)^i }
\nonumber \\
& = & \displaystyle{ f(t) \exp \left\{ - \int_0^t g(t') \, \d t'
\right\} \exp \left\{ \int_0^t \left[ g(t') - f(t') \right] \d t'
\right\} } \nonumber \\
& = & \displaystyle{ f(t) \exp \left\{ - \int_0^t f(t') \, \d t'
\right\} ~, }
\end{eqnarray}
which is the desired answer.
If the process is to be stopped at some scale $t_{\mathrm{max}}$, i.e.
if one would like to remain with a fraction ${\cal N}(t_{\mathrm{max}})$
of events where nothing happens at all, this is easy to include
in the veto algorithm: just iterate upwards in $t$ at usual, but
stop the process if no allowed branching is found before
$t_{\mathrm{max}}$.
Usually $f(t)$ is a function also of additional variables $x$. The
methods of the preceding subsection are easy to generalize if one
can find a suitable function $g(t,x)$ with $f(t,x) \leq g(t,x)$.
The $g(t)$ used in the veto algorithm is the integral of $g(t,x)$
over $x$. Each time a $t_i$ has been selected also an $x_i$ is
picked, according to $g(t_i,x) \, dx$, and the $(t,x)$ point is
accepted with probability $f(t_i,x_i)/g(t_i,x_i)$.
\subsection{The Random Number Generator}
The construction of a good, portable (pseudo)random generator is not a
trivial task. Therefore {\tsc{Jetset}} has traditionally stayed
away from that area, and just provided the routine \ttt{RLU} as an
interface, which the user could modify to call on an existing routine,
implemented on the actual machine being used.
In recent years, progress has been made in constructing portable
generators with large periods and other good properties; see the review
\cite{Jam90}. Therefore the current version contains a random number
generator based on the algorithm proposed by Marsaglia, Zaman and Tsang
\cite{Mar90}. This routine should work on any machine with a mantissa
of at least 24 digits, i.e. all common 32-bit (or more) computers.
Given the same initial state, the sequence will also be identical
on different machines. This need not mean that the same sequence of
events will be generated on an IBM and a VAX, say, since the different
treatments of roundoff errors in numerical operations will lead to
slightly different real numbers being tested against these random
numbers in IF statements. Also code optimization may lead to a
divergence. Apart from nomenclature issues, and the
coding of \ttt{RLU} as a function rather than a subroutine, the only
difference between the {\tsc{Jetset}} code and the code given in
\cite{Jam90} is that slightly different algorithms are used to ensure
that the random number is not equal to 0 or 1 within the machine
precision.
The generator has a period of over $10^{43}$, and the
possibility to obtain almost $10^9$
different and disjoint subsequences, selected
by giving an initial integer number. The price to be paid for the
long period is that the state of the generator at a given moment
cannot be described by a single integer, but requires about 100 words.
Some of these are real numbers, and are thus not correctly represented
in decimal form. The normal procedure, which makes it possible to
restart the generation from a seed value written to the run output,
is therefore
not convenient. The CERN library implementation keeps track of the
number of random numbers generated since the start. With this value
saved, in a subsequent run the random generator can be asked to skip
ahead the corresponding number of random numbers. {\tsc{Jetset}}
is a heavy user of random numbers, however: typically 30\% of the full
run time is spent on random number generation. Of this, half is
overhead coming from the function call administration, but the other
half is truly related to the speed of the algorithm. Therefore a
skipping ahead would take place with 15\% of the time cost of the
original run, i.e. an uncomfortably high figure.
Instead a different solution is chosen here. Two special routines are
provided for writing and reading the state of the random number
generator (plus some initialization information) on a sequential
file, in a machine-dependent internal representation. The file used
for this purpose
has to be specified by you, and opened for read and write.
A state is written as a single record, in free format. It is possible
to write an arbitrary number of states on a file, and a record can be
overwritten, if so desired. The event generation loop might then look
something like:
\begin{Enumerate}
\item save the state of the generator on file (using flag set in
point 3 below),
\item generate an event,
\item study the event for errors or other reasons why to regenerate
it later; set flag to overwrite previous generator state if no errors,
otherwise set flag to create new record;
\item loop back to point 1.
\end{Enumerate}
With this procedure, the file will contain the state before each of the
problematical events. An alternative approach might be to save the
state every 100 events or so. If the events are subsequently
processed through a detector simulation, you may
have to save also other sets of seeds, naturally.
In addition to the service routines, the common block which contains
the state of the generator is available for manipulation,
if you so desire. In particular, the initial seed value is by
default 19780503, i.e. different from the Marsaglia/CERN default
54217137. It is possible to change this value before any random numbers
have been generated, or to force reinitialization in mid-run with any
desired new seed. Inside {\tsc{Jetset/Pythia}}, some
initialization may take place in connection with the very first event
generated in a run, so sometimes it may be necessary to generate one
ordinary event before reading in a saved state to generate an
interesting event. In the current {\tsc{Pythia}} version, some of the
multiple interaction machinery options contain an element of learning,
which means that the event sequence may be broken.
It should be noted that, of course, the appearance of a random
number generator package inside {\tsc{Jetset}} does in no way preclude
the use of other routines. You can easily revert to the
old approach, where \ttt{RLU} is nothing but an interface to an
arbitrary external random number generator; e.g. to call a routine
\ttt{RNDM} all you need to have is
\begin{verbatim}
FUNCTION RLU(IDUMMY)
100 RLU=RNDM(IDUMMY)
IF(RLU.LE.0..OR.RLU.GE.1.) GOTO 100
RETURN
END
\end{verbatim}
The random generator subpackage consists of the following components.
\drawbox{R = RLU(IDUMMY)}\label{p:RLU}
\begin{entry}
\itemc{Purpose:} to generate a (pseudo)random number \ttt{R} uniformly
in the range 0$<$\ttt{R}$<$1, i.e. excluding the endpoints.
\iteme{IDUMMY :} dummy input argument; normally 0.
\end{entry}
\drawbox{CALL RLUGET(LFN,MOVE)}\label{p:RLUGET}
\begin{entry}
\itemc{Purpose:} to dump the current state of the random number
generator on a separate file, using internal representation for
real and integer numbers. To be precise, the full contents of
the \ttt{LUDATR} common block are written on the file, with the
exception of \ttt{MRLU(6)}.
\iteme{LFN :} (logical file number) the file number to which the state
is dumped. You
must associate this number with a true file (with a machine-dependent
name), and see to it that this file is open for write.
\iteme{MOVE :} choice of adding a new record to the file or
overwriting old record(s). Normally only options 0 or $-$1 should be
used.
\begin{subentry}
\iteme{= 0 (or > 0) :} add a new record to the end of the
file.
\iteme{= -1 :} overwrite the last record with a new one (i.e. do
one \ttt{BACKSPACE} before the new write).
\iteme{= $-n$ :} back up $n$ records before writing the
new record. The records following after the new one are lost,
i.e. the last $n$ old records are lost and one new added.
\end{subentry}
\end{entry}
\drawbox{CALL RLUSET(LFN,MOVE)}\label{p:RLUSET}
\begin{entry}
\itemc{Purpose:} to read in a state for the random number generator,
from which the subsequent generation can proceed. The state must
previously have been saved by a \ttt{RLUGET} call. Again the full
contents of the \ttt{LUDATR} common block are read, with the
exception of \ttt{MRLU(6)}.
\iteme{LFN :} (logical file number) the file number from which the
state is read. You
must associate this number with a true file previously written with
a \ttt{RLUGET} call, and see to it that this file is open for read.
\iteme{MOVE :} positioning in file before a record is read. With zero
value, records are read one after the other for each new call, while
non-zero values may be used to navigate back and forth, and e.g.
return to the same initial state several times.
\begin{subentry}
\iteme{= 0 :} read the next record.
\iteme{= $+n$ :} skip ahead $n$ records before reading the record
that sets the state of the random number generator.
\iteme{= $-n$ :} back up $n$ records before reading the record that
sets the state of the random number generator.
\end{subentry}
\end{entry}
\drawbox{COMMON/LUDATR/MRLU(6),RRLU(100)}\label{p:LUDATR}
\begin{entry}
\itemc{Purpose:} to contain the state of the random number generator
at any moment (for communication between \ttt{RLU}, \ttt{RLUGET}
and \ttt{RLUSET}), and also to provide the user with the
possibility to initialize different random number sequences, and to
know how many numbers have been generated.
\iteme{MRLU(1) :}\label{p:MRLU} (D=19780503) the integer number that
specifies
which of the possible subsequences will be initialized in the next
\ttt{RLU} call for which \ttt{MRLU(2)=0}. Allowed values are
0$\leq$\ttt{MRLU(1)}$\leq$900\,000\,000, the original Marsaglia
(and CERN library) seed is 54217137. The \ttt{MRLU(1)} value is not
changed by any of the {\tsc{Jetset}} routines.
\iteme{MRLU(2) :} (D=0) initialization flag, put to 1 in the first
\ttt{RLU} call of run. A reinitialization of the random number
generator can be made in mid-run by resetting \ttt{MRLU(2)} to 0
by hand. In addition, any time the counter \ttt{MRLU(3)} reaches
1000000000, it is reset to 0 and \ttt{MRLU(2)} is increased by 1.
\iteme{MRLU(3) :} (D=0) counter for the number of random numbers
generated from the beginning of the run. To avoid overflow when
very many numbers are generated, \ttt{MRLU(2)} is used as
described above.
\iteme{MRLU(4), MRLU(5) :} \ttt{I97} and \ttt{J97} of the CERN
library implementation; part of the state of the generator.
\iteme{MRLU(6) :} (D=0) current position, i.e. how many records after
beginning, in the file; used by \ttt{RLUGET} and \ttt{RLUSET}.
\iteme{RRLU(1) - RRLU(97) :}\label{p:RRLU} the \ttt{U} array of the
CERN library implementation; part of the state of the generator.
\iteme{RRLU(98) - RRLU(100) :} \ttt{C}, \ttt{CD} and \ttt{CM} of the
CERN library implementation; the first part of the state of the
generator, the latter two constants calculated at initialization.
\end{entry}
\clearpage
\section{The Event Record}
The event record is the central repository for information about
the particles produced in the current event: flavours, momenta,
event history, and production vertices. It plays a very central
r\^ole: without a proper understanding of what the record is and
how information is stored, it is meaningless to try to use either
{\tsc{Jetset}} or {\tsc{Pythia}}. The record is stored in the common block
\ttt{LUJETS}. Almost all the routines thatthe user calls
can be viewed as performing some action on the record: fill a
new event, let partons fragment or particles decay, boost it,
list it, find clusters, etc.
In this section we will first describe the KF flavour code,
subsequently the \ttt{LUJETS} common block, and then give a few
comments about the r\^ole of the event record in the programs.
To ease the interfacing of different event generators, a
\ttt{HEPEVT} standard common block structure for the event record
has been agreed on. For historical reasons the standard common blocks
are not directly used in {\tsc{Jetset}}, but a conversion routine comes with
the program, and is described at the end of this section.
\subsection{Particle Codes}
\label{ss:codes}
The new particle code now adopted by the Particle Data Group
\cite{PDG88,PDG92} is used consistently throughout the program, and is
referred to as the KF particle code. This code you have to be
thoroughly familiar with. It is described below.
Note that a few inconsistencies between the KF and the PDG codes
are known, which stem from differences of interpretation of
the rules agreed on when developing the standard. These rules form
the basis of the PDG tables and (independently) of the {\tsc{Jetset}} tables.
(Of course, my private opinion is that I follow the original agreement,
and the PDG deviate from it.) Hopefully, this should have few practical
consequences, since only rarely-produced particles are affected.
Anyway, here is a list of the known discrepancies:
\begin{Enumerate}
\item The PDG has not allowed for the existence of an $\eta_{\b}$,
which in {\tsc{Jetset}} is included with code 551. This code is reserved for
$\chi_{0 \b}$ by the PDG, a particle which appears as 10551 in {\tsc{Jetset}}.
(We agree to have $\eta_{\c}$ as 441, which illustrates the basic
difference: I use the additional recurrence figure to refer to a
whole multiplet, whether all particles of that multiplet have been
found or not; the PDG, on the other hand, does not reserve space for
particles which we know should be there but have not yet been
discovered, which means that members of a multiplet need not go
together.)
\item The PDG has not allowed for the existence of an $\mathrm{h}_{1 \c}$,
which in {\tsc{Jetset}} is represented by 10443. Therefore $\chi_{1 \c}$
is the PDG code 10443 but {\tsc{Jetset}} code 20443. Further $\psi'$ is
either 20443 or 30443, and $\Upsilon' = \Upsilon(2S)$ either
20553 or 30553. (Comment as for point 1.)
\item Different conventions for spin $1/2$ baryons with one heavy
flavour (charm, bottom, top), one strange flavour, and one light
($\u$ or $\d$). Here two states exist, e.g. $\Xi_{\c}^+$ and
$\Xi'^+_{\c}$, both with flavour content $\c\mathrm{s}\u$. By analogy with the
$\Lambda^0$--$\Sigma$ pair, {\tsc{Jetset}} uses the decreasing order of flavour
content for the heavier state and inversed order of the two lighter
flavours for the lighter state, while the PDG tables use the opposite
convention. Thus in {\tsc{Jetset}} $\Xi_{\c}^+$ is 4232 and $\Xi'^+_{\c}$ 4322,
while in PDG it is the other way around.
\end{Enumerate}
There are no plans to change the {\tsc{Jetset}} rules to agree with
the PDG ones in either of the cases above.
The KF code is not convenient for a direct storing of masses,
decay data, or other particle properties, since the KF
codes are so spread out. Instead a compressed code KC between
1 and 500 is used here, where the most frequently used particles
have a separate code, but many heavy-flavour hadrons are lumped
together in groups. Normally this code is only used at very specific
places in the program, not visible to the user. If need be, the
correspondence can always be obtained by using the function
\ttt{LUCOMP}, \mbox{\ttt{KC = LUCOMP(KF)}}. It is therefore not
intended that you should ever need to know any KC codes at all.
It may be useful to know, however, that for codes smaller than 80,
KF and KC agree.
The particle names printed in the tables in this section correspond
to the ones obtained
with the routine \ttt{LUNAME}, which is used extensively, e.g. in
\ttt{LULIST}. Greek characters are spelt out in full, with a capital
first letter to correspond to a capital Greek letter. Generically the
name of a particle is made up of the following pieces:
\begin{Enumerate}
\item The basic root name. This includes a * for most spin 1
($L = 0$) mesons and spin $3/2$ baryons, and a $'$ for some spin
$1/2$ baryons (where there are two states to be distinguished,
cf. $\Lambda$--$\Sigma^0$). The rules for heavy baryon naming are in
accordance with the 1986 Particle Data Group conventions \cite{PDG86}.
For mesons with one unit of orbital angular momentum, K (D, B,
\ldots) is used for quark-spin 0 and K* (D*, B*, \ldots) for
quark-spin 1 mesons; the convention for `*' may here deviate slightly
from the one used by the PDG.
\item Any lower indices, separated from the root by a \_. For
heavy hadrons, this is the additional heavy-flavour content not
inherent in the root itself. For a diquark, it is the spin.
\item The character $\sim$ (alternatively bar, see \ttt{MSTU(15)})
for an antiparticle, wherever the distinction between particle and
antiparticle is not inherent in the charge information.
\item Charge information: $++$, $+$, $0$, $-$, or $--$.
Charge is not given for quarks or diquarks. Some neutral particles
which are customarily given without a 0 also here lack it,
such as neutrinos, $\mathrm{g}$, $\gamma$,
and flavour-diagonal mesons other than $\pi^0$ and $\rho^0$. Note that
charge is included both for the proton and the neutron. While
non-standard, it is helpful in avoiding misunderstandings when
looking at an event listing.
\end{Enumerate}
Below follows a list of KF particle codes. The list is not complete;
a more extensive one may be obtained with \ttt{CALL LULIST(11)}.
Particles are grouped together, and the basic rules are described
for each group. Whenever a distinct antiparticle exists, it is given
the same KF code with a minus sign (whereas KC codes are always
positive).
\begin{Enumerate}
\item Quarks and leptons, Table \ref{t:codeone}. \\
This group contains the basic building blocks of matter, arranged
according to family, with the lower member of weak isodoublets also
having the smaller code (thus $\d$ precedes $\u$, contrary to the
ordering in previous {\tsc{Jetset}} versions). A fourth generation is
included for future reference. The quark codes are used as building
blocks for the diquark, meson and baryon codes below.
\begin{table}[ptb]
\captive{Quark and lepton codes.
\protect\label{t:codeone} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c||c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Name & Printed & KF & Name & Printed \\
\hline
1 & $\d$ & \ttt{d} & 11 & $\mathrm{e}^-$ & \ttt{e-} \\
2 & $\u$ & \ttt{u} & 12 & $\nu_{\mathrm{e}}$ & \ttt{nu\_e} \\
3 & $\mathrm{s}$ & \ttt{s} & 13 & $\mu^-$ & \ttt{mu-} \\
4 & $\c$ & \ttt{c} & 14 & $\nu_{\mu}$ & \ttt{nu\_mu} \\
5 & $\b$ & \ttt{b} & 15 & $\tau^-$ & \ttt{tau-} \\
6 & $\t$ & \ttt{t} & 16 & $\nu_{\tau}$ & \ttt{nu\_tau} \\
7 & $\mathrm{l}$ & \ttt{l} & 17 & $\chi^-$ & \ttt{chi-} \\
8 & $\mathrm{h}$ & \ttt{h} & 18 & $\nu_{\chi}$ & \ttt{nu\_chi} \\
9 & & & 19 & & \\
10 & & & 20 & & \\
\hline
\end{tabular}
\end{center}
\end{table}
\item Gauge bosons and other fundamental bosons,
Table \ref{t:codetwo}. \\
This group includes all the gauge and Higgs bosons of the standard
model, as well as some of the bosons appearing in various extensions
of it. The latter are not covered by the standard
PDG codes. They correspond to one extra {\bf U(1)}
group and one extra {\bf SU(2)} one, a further Higgs doublet, a
(scalar, colour octet) techni-$\eta$, a (scalar) leptoquark
$\L_{\mathrm{Q}}$, and a horizontal gauge boson $\mathrm{R}$
(coupling between families). Additionally, we here include the
pomeron $\mathrm{I}\!\mathrm{P}$ and reggeon $\mathrm{I}\!\mathrm{R}$ `particles', which are
important e.g. in the description of diffractive scattering, but
have no obvious position anywhere in the classification scheme.
\begin{table}[ptb]
\captive{Gauge boson and other fundamental boson codes.
\protect\label{t:codetwo} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c||c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Name & Printed & KF & Name & Printed \\
\hline
21 & $\mathrm{g}$ & \ttt{g} & 31 & & \\
22 & $\gamma$ & \ttt{gamma} & 32 & $\mathrm{Z}'^0$ & \ttt{Z'0} \\
23 & $\mathrm{Z}^0$ & \ttt{Z0} & 33 & $\mathrm{Z}''^0$ & \ttt{Z"0} \\
24 & $\mathrm{W}^+$ & \ttt{W+} & 34 & $\mathrm{W}'^+$ & \ttt{W'+} \\
25 & $\H^0$ & \ttt{H0} & 35 & $\H'^0$ & \ttt{H'0} \\
26 & & & 36 & $\mathrm{A}^0$ & \ttt{A0} \\
27 & & & 37 & $\H^+$ & \ttt{H+} \\
28 & $\mathrm{I}\!\mathrm{R}$ & \ttt{reggeon} &
38 & $\eta_{\mrm{techni}}$ & \ttt{eta\_tech0} \\
29 & $\mathrm{I}\!\mathrm{P}$ & \ttt{pomeron} & 39 & $\L_{\mathrm{Q}}$ & \ttt{LQ} \\
30 & & & 40 & $\mathrm{R}^0$ & \ttt{R0} \\
\hline
\end{tabular}
\end{center}
\end{table}
\item Free space. \\
The positions 41--80 are currently unused. In the future, they might
come to be used, e.g. for supersymmetric partners of the particles
above, or for some other kind of new physics. At the moment, they are
at your disposal.
\item Various special codes, Table \ref{t:codefour}. \\
In a Monte Carlo, it is always necessary to have codes that do not
correspond to any specific particle, but are used to lump together
groups of similar particles for decay treatment, or to specify generic
decay products. These codes, which again are non-standard, are found
between numbers 81 and 100. Several are not found in the event record,
and therefore properly belong only to the KC group of codes.
\begin{table}[ptb]
\captive{Various special codes.
\protect\label{t:codefour} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Printed & Meaning \\
\hline
81 & \ttt{specflav} & Spectator flavour; used in decay-product
listings \\
82 & \ttt{rndmflav} & A random $\u$, $\d$, or $\mathrm{s}$ flavour;
possible decay product \\
83 & \ttt{phasespa} & Simple isotropic phase-space decay \\
84 & \ttt{c-hadron} & Information on decay of generic charm
hadron \\
85 & \ttt{b-hadron} & Information on decay of generic bottom
hadron \\
86 & \ttt{t-hadron} & Information on decay of generic top
hadron \\
87 & \ttt{l-hadron} & Information on decay of generic low
hadron \\
88 & \ttt{h-hadron} & Information on decay of generic high
hadron \\
89 & \ttt{Wvirt} & Off-mass-shell $\mathrm{W}$ in weak decays
of $\t$, $\mathrm{l}$, $\mathrm{h}$ or $\chi$ \\
90 & \ttt{diquark} & Generic code for diquark colour
information \\
91 & \ttt{cluster} & Parton system in cluster
fragmentation \\
92 & \ttt{string} & Parton system in string
fragmentation \\
93 & \ttt{indep.} & Parton system in independent
fragmentation \\
94 & \ttt{CMshower} & Four-momentum of time-like showering
system \\
95 & \ttt{SPHEaxis} & Event axis found with \ttt{LUSPHE} \\
96 & \ttt{THRUaxis} & Event axis found with \ttt{LUTHRU} \\
97 & \ttt{CLUSjet} & Jet (cluster) found with \ttt{LUCLUS} \\
98 & \ttt{CELLjet} & Jet (cluster) found with \ttt{LUCELL} \\
99 & \ttt{table} & Tabular output from \ttt{LUTABU} \\
100 & & \\
\hline
\end{tabular}
\end{center}
\end{table}
\item Diquark codes, Table \ref{t:codefive}. \\
A diquark made up of a quark with code $i$ and another with code $j$,
where $i \geq j$, and with total spin $s$, is given the code
\begin{equation}
\mrm{KF} = 1000 i + 100 j + 2s + 1 ~,
\end{equation}
i.e. the tens position is left empty (cf. the baryon code below).
Some of the most frequently used codes are listed in the table. All
the lowest-lying spin 0 and 1 diquarks are included in the program.
The corresponding KC code is 90, and it is mainly used to store colour
charge.
\begin{table}[ptb]
\captive{Diquark codes.
\protect\label{t:codefive} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c||c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Name & Printed & KF & Name & Printed \\
\hline
& & & 1103 & $\d\d_1$ & \ttt{dd\_1} \\
2101 & $\u\d_0$ & \ttt{ud\_0} & 2103 & $\u\d_1$ & \ttt{ud\_1} \\
& & & 2203 & $\u\u_1$ & \ttt{uu\_1} \\
3101 & $\mathrm{s}\d_0$ & \ttt{sd\_0} & 3103 & $\mathrm{s}\d_1$ & \ttt{sd\_1} \\
3201 & $\mathrm{s}\u_0$ & \ttt{su\_0} & 3203 & $\mathrm{s}\u_1$ & \ttt{su\_1} \\
& & & 3303 & $\mathrm{s}\s_1$ & \ttt{ss\_1} \\
\hline
\end{tabular}
\end{center}
\end{table}
\item Meson codes, Tables \ref{t:codesixa} and \ref{t:codesixb}. \\
A meson made up of a quark with code $i$ and an antiquark with
code $-j$, $j \neq i$, and with total spin $s$, is given the code
\begin{equation}
\mrm{KF} = \left\{ 100 \max(i,j) + 10 \min(i,j) + 2s + 1 \right\}
\, \mrm{sign}(i-j) \, (-1)^{\max(i,j)} ~.
\end{equation}
Note the presence of an extra $-$ sign if the heaviest quark is a
down-type one. This is in accordance with the particle--antiparticle
distinction adopted in the 1986 Review of Particle Properties
\cite{PDG86}. It means for example that a $\mathrm{B}$ meson contains a
$\overline{\mathrm{b}}$ antiquark rather than a $\b$ quark.
The flavour-diagonal states are arranged in order of ascending
mass. The standard rule of having the last digit of the form
$2s+1$ is broken for the $\mathrm{K}_{\mrm{S}}^0$--$\mathrm{K}_{\mrm{L}}^0$ system,
where it is 0, and this convention should carry over to mixed states
in the $\mathrm{B}$ meson system. For higher multiplets with the same spin,
$\pm$10000, $\pm$20000, etc., are added to provide the extra
distinction needed. Some of the most frequently used codes are given
below.
The full lowest-lying pseudoscalar and vector multiplets are included
in the program, Table \ref{t:codesixa}.
\begin{table}[ptb]
\captive{Meson codes, part 1.
\protect\label{t:codesixa} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c||c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Name & Printed & KF & Name & Printed \\
\hline
211 & $\pi^+$ & \ttt{pi+} & 213 & $\rho^+$ & \ttt{rho+} \\
311 & $\mathrm{K}^0$ & \ttt{K0} & 313 & $\mathrm{K}^{*0}$ & \ttt{K*0} \\
321 & $\mathrm{K}^+$ & \ttt{K+} & 323 & $\mathrm{K}^{*+}$ & \ttt{K*+} \\
411 & $\mathrm{D}^+$ & \ttt{D+} & 413 & $\mathrm{D}^{*+}$ & \ttt{D*+} \\
421 & $\mathrm{D}^0$ & \ttt{D0} & 423 & $\mathrm{D}^{*0}$ & \ttt{D*0} \\
431 & $\mathrm{D}_s^+$ & \ttt{D\_s+} &
433 & $\mathrm{D}_{\mathrm{s}}^{*+}$ & \ttt{D*\_s+} \\
511 & $\mathrm{B}^0$ & \ttt{B0} & 513 & $\mathrm{B}^{*0}$ & \ttt{B*0} \\
521 & $\mathrm{B}^+$ & \ttt{B+} & 523 & $\mathrm{B}^{*+}$ & \ttt{B*+} \\
531 & $\mathrm{B}_s^0$ & \ttt{B\_s0} &
533 & $\mathrm{B}_{\mathrm{s}}^{*0}$ & \ttt{B*\_s0} \\
541 & $\mathrm{B}_c^+$ & \ttt{B\_c+} &
543 & $\mathrm{B}_{\c}^{*+}$ & \ttt{B*\_c+} \\
111 & $\pi^0$ & \ttt{pi0} & 113 & $\rho^0$ & \ttt{rho0} \\
221 & $\eta$ & \ttt{eta} & 223 & $\omega$ & \ttt{omega} \\
331 & $\eta'$ & \ttt{eta'} & 333 & $\phi$ & \ttt{phi} \\
441 & $\eta_{\c}$ & \ttt{eta\_c} & 443 & $\mathrm{J}/\psi$ & \ttt{J/psi} \\
551 & $\eta_{\b}$ & \ttt{eta\_b} &
553 & $\Upsilon$ & \ttt{Upsilon} \\
661 & $\eta_{\t}$ & \ttt{eta\_t} & 663 & $\Theta$ & \ttt{Theta} \\
130 & $\mathrm{K}_{\mrm{L}}^0$ & \ttt{K\_L0} & & & \\
310 & $\mathrm{K}_{\mrm{S}}^0$ & \ttt{K\_S0} & & & \\
\hline
\end{tabular}
\end{center}
\end{table}
Also the lowest-lying orbital angular momentum $L = 1$ mesons are
included, Table \ref{t:codesixb}: one pseudovector multiplet
obtained for total quark-spin 0 ($L = 1, S = 0 \Rightarrow J = 1$)
and one scalar, one pseudovector and one tensor multiplet obtained
for total quark-spin 1 ($L = 1, S = 1 \Rightarrow J = 0, 1$ or 2),
where $J$ is what is conventionally called the spin $s$ of the meson.
Any mixing between the two pseudovector multiplets is
not taken into account. Please note that some members of these
multiplets have still not been found, and are included here only based
on guesswork. Even for known ones, the information on particles
(mass, width, decay modes) is highly incomplete.
Only two radial excitations are included, the $\psi' = \psi(2S)$ and
$\Upsilon' = \Upsilon(2S)$.
\begin{table}[ptb]
\captive{Meson codes, part 2.
\protect\label{t:codesixb} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c||c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Name & Printed & KF & Name & Printed \\
\hline
10213 & $\b_1$ & \ttt{b\_1+} &
10211 & $\a_0^+$ & \ttt{a\_0+} \\
10313 & $\mathrm{K}_1^0$ & \ttt{K\_10} &
10311 & $\mathrm{K}_0^{*0}$ & \ttt{K*\_00} \\
10323 & $\mathrm{K}_1^+$ & \ttt{K\_1+} &
10321 & $\mathrm{K}_0^{*+}$ & \ttt{K*\_0+} \\
10413 & $\mathrm{D}_1^+$ & \ttt{D\_1+} &
10411 & $\mathrm{D}_0^{*+}$ & \ttt{D*\_0+} \\
10423 & $\mathrm{D}_1^0$ & \ttt{D\_10} &
10421 & $\mathrm{D}_0^{*0}$ & \ttt{D*\_00} \\
10433 & $\mathrm{D}_{1 \mathrm{s}}^+$ & \ttt{D\_1s+} &
10431 & $\mathrm{D}_{0 \mathrm{s}}^{*+}$ & \ttt{D*\_0s+} \\
10113 & $\b_1^0$ & \ttt{b\_10} &
10111 & $\a_0^0$ & \ttt{a\_00} \\
10223 & $\mathrm{h}_1^0$ & \ttt{h\_10} &
10221 & $\mathrm{f}_0^0$ & \ttt{f\_00} \\
10333 & $\mathrm{h}'^0_1$ & \ttt{h'\_10} &
10331 & $\mathrm{f}'^0_0$ & \ttt{f'\_00} \\
10443 & $\mathrm{h}_{1 \c}^0$ & \ttt{h\_1c0} &
10441 & $\chi_{0 \c}^0$ & \ttt{chi\_0c0} \\ \hline
20213 & $\a_1^+$ & \ttt{a\_1+} &
215 & $\a_2^+$ & \ttt{a\_2+} \\
20313 & $\mathrm{K}_1^{*0}$ & \ttt{K*\_10} &
315 & $\mathrm{K}_2^{*0}$ & \ttt{K*\_20} \\
20323 & $\mathrm{K}_1^{*+}$ & \ttt{K*\_1+} &
325 & $\mathrm{K}_2^{*+}$ & \ttt{K*\_2+} \\
20413 & $\mathrm{D}_1^{*+}$ & \ttt{D*\_1+} &
415 & $\mathrm{D}_2^{*+}$ & \ttt{D*\_2+} \\
20423 & $\mathrm{D}_1^{*0}$ & \ttt{D*\_10} &
425 & $\mathrm{D}_2^{*0}$ & \ttt{D*\_20} \\
20433 & $\mathrm{D}_{1 \mathrm{s}}^{*+}$ & \ttt{D*\_1s+} &
435 & $\mathrm{D}_{2 \mathrm{s}}^{*+}$ & \ttt{D*\_2s+} \\
20113 & $\a_1^0$ & \ttt{a\_10} &
115 & $\a_2^0$ & \ttt{a\_20} \\
20223 & $\mathrm{f}_1^0$ & \ttt{f\_10} &
225 & $\mathrm{f}_2^0$ & \ttt{f\_20} \\
20333 & $\mathrm{f}'^0_1$ & \ttt{f'\_10} &
335 & $\mathrm{f}'^0_2$ & \ttt{f'\_20} \\
20443 & $\chi_{1 \c}^0$ & \ttt{chi\_1c0} &
445 & $\chi_{2 \c}^0$ & \ttt{chi\_2c0} \\ \hline
30443 & $\psi'$ & \ttt{psi'} & & & \\
30553 & $\Upsilon'$ & \ttt{Upsilon'} & & & \\
\hline
\end{tabular}
\end{center}
\end{table}
The corresponding meson KC codes, used for organizing mass and decay
data, range between 101 and 240.
\item Baryon codes, Table \ref{t:codeseven}. \\
A baryon made up of quarks $i$, $j$ and $k$, with $i \geq j \geq k$,
and total spin $s$, is given the code
\begin{equation}
\mrm{KF} = 1000 i + 100 j + 10 k + 2s + 1 ~.
\end{equation}
An exception is provided by spin $1/2$ baryons made up of three
different types of quarks, where the two lightest quarks form a spin-0
diquark ($\Lambda$-like baryons). Here the order of the $j$ and $k$
quarks is reversed, so as to provide a simple means of distinction
to baryons with the lightest quarks in a spin-1 diquark
($\Sigma$-like baryons).
For hadrons with heavy flavours, the root names are Lambda or
Sigma for hadrons with two $\u$ or $\d$ quarks, Xi for those
with one, and Omega for those without $\u$ or $\d$ quarks.
Some of the most frequently used codes are given in Table
\ref{t:codeseven}. The full lowest-lying spin $1/2$ and $3/2$
multiplets are included in the program.
The corresponding KC codes, used for organizing mass and decay data,
range between 301 and 400, with some slots still free.
\begin{table}[ptb]
\captive{Baryon codes.
\protect\label{t:codeseven} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c||c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Name & Printed & KF & Name & Printed \\
\hline
& & &
1114 & $\Delta^-$ & \ttt{Delta-} \\
2112 & $\mathrm{n}$ & \ttt{n0} &
2114 & $\Delta^0$ & \ttt{Delta0} \\
2212 & $\mathrm{p}$ & \ttt{p+} &
2214 & $\Delta^+$ & \ttt{Delta+} \\
& & &
2224 & $\Delta^{++}$ & \ttt{Delta++} \\
3112 & $\Sigma^-$ & \ttt{Sigma-} &
3114 & $\Sigma^{*-}$ & \ttt{Sigma*-} \\
3122 & $\Lambda^0$ & \ttt{Lambda0} & & & \\
3212 & $\Sigma^0$ & \ttt{Sigma0} &
3214 & $\Sigma^{*0}$ & \ttt{Sigma*0} \\
3222 & $\Sigma^+$ & \ttt{Sigma+} &
3224 & $\Sigma^{*+}$ & \ttt{Sigma*+} \\
3312 & $\Xi^-$ & \ttt{Xi-} &
3314 & $\Xi^{*-}$ & \ttt{Xi*-} \\
3322 & $\Xi^0$ & \ttt{Xi0} &
3324 & $\Xi^{*0}$ & \ttt{Xi*0} \\
& & &
3334 & $\Omega^-$ & \ttt{Omega-} \\
4112 & $\Sigma_{\c}^0$ & \ttt{Sigma\_c0} &
4114 & $\Sigma_{\c}^{*0}$ & \ttt{Sigma*\_c0} \\
4122 & $\Lambda_{\c}^+$ & \ttt{Lambda\_c+} & & & \\
4212 & $\Sigma_{\c}^+$ & \ttt{Sigma\_c+} &
4214 & $\Sigma_{\c}^{*+}$ & \ttt{Sigma*\_c+} \\
4222 & $\Sigma_{\c}^{++}$ & \ttt{Sigma\_c++} &
4224 & $\Sigma_{\c}^{*++}$ & \ttt{Sigma*\_c++} \\
4132 & $\Xi_{\c}^0$ & \ttt{Xi\_c0} & & & \\
4312 & $\Xi'^0_{\c}$ & \ttt{Xi'\_c0} &
4314 & $\Xi_{\c}^{*0}$ & \ttt{Xi*\_c0} \\
4232 & $\Xi_{\c}^+$ & \ttt{Xi\_c+} & & & \\
4322 & $\Xi'^+_{\c}$ & \ttt{Xi'\_c+} &
4324 & $\Xi_{\c}^{*+}$ & \ttt{Xi*\_c+} \\
4332 & $\Omega_{\c}^0$ & \ttt{Omega\_c0} &
4334 & $\Omega_{\c}^{*0}$ & \ttt{Omega*\_c0} \\
5112 & $\Sigma_{\b}^-$ & \ttt{Sigma\_b-} &
5114 & $\Sigma_{\b}^{*-}$ & \ttt{Sigma*\_b-} \\
5122 & $\Lambda_{\b}^0$ & \ttt{Lambda\_b0} & & & \\
5212 & $\Sigma_{\b}^0$ &\ttt{Sigma\_b0} &
5214 & $\Sigma_{\b}^{*0}$ & \ttt{Sigma*\_b0} \\
5222 & $\Sigma_{\b}^+$ & \ttt{Sigma\_b+} &
5224 & $\Sigma_{\b}^{*+}$ & \ttt{Sigma*\_b+} \\
\hline
\end{tabular}
\end{center}
\end{table}
\item Diffractive states, Table \ref{t:codeeight}. \\
These codes are not standard ones: they have been defined by analogy
to be used for denoting diffractive states in {\tsc{Pythia}},
as part of the event history. The first two or three digits give
flavour content, while the last one is 0, to denote the somewhat
unusual character of the code. Only a few codes have been introduced;
depending on circumstances these also have to double up for other
diffractive states.
\begin{table}[ptb]
\captive{Diffractive state codes.
\protect\label{t:codeeight} } \\
\vspace{1ex}
\begin{center}
\begin{tabular}{|c|c|c|@{\protect\rule{0mm}{\tablinsep}}}
\hline
KF & Printed & Meaning \\
\hline
110 & \ttt{rho\_diff0} & Diffractive
$\pi^0 / \rho^0 / \gamma$ state \\
210 & \ttt{pi\_diffr+} & Diffractive $\pi^+$ state \\
220 & \ttt{omega\_di0} & Diffractive $\omega$ state \\
330 & \ttt{phi\_diff0} & Diffractive $\phi$ state \\
440 & \ttt{J/psi\_di0} & Diffractive $\mathrm{J}/\psi$ state \\
2110 & \ttt{n\_diffr} & Diffractive $\mathrm{n}$ state \\
2210 & \ttt{p\_diffr+} & Diffractive $\mathrm{p}$ state \\
\hline
\end{tabular}
\end{center}
\end{table}
\item Free compressed codes.
The positions 401--500 of mass and decay arrays are left open.
Here a user may map any new kind of particle from the ordinary
KF codes, which probably are above 10000, into a more manageable
KC range for mass and decay data information. The mapping must be
implemented in the \ttt{LUCOMP} function.
\end{Enumerate}
\subsection{The Event Record}
\label{ss:evrec}
Each new event generated is in its entirety stored in the common block
\ttt{LUJETS}, which thus forms the event record. Here each jet or
particle that appears at some stage of the fragmentation or decay
chain will occupy one line in the matrices. The different components
of this line will tell which jet/particle it is, from where it
originates, its present status (fragmented/decayed or not), its
momentum, energy and mass, and the space--time position of its
production vertex. Note that \ttt{K(I,3)}--\ttt{K(I,5)} and the
\ttt{P} and \ttt{V} vectors may take special meaning for some
specific applications (e.g. sphericity or cluster
analysis), as described in those connections.
The event history information stored in \ttt{K(I,3)}--\ttt{K(I,5)}
should not be taken too literally. In the particle decay chains, the
meaning of a mother is well-defined, but the fragmentation description
is more complicated. The primary hadrons produced in string
fragmentation come from the string as a whole, rather than from an
individual parton. Even when the string is not included in the history
(see \ttt{MSTU(16)}), the pointer from hadron to parton is deceptive.
For instance, in a $\mathrm{q}\mathrm{g}\overline{\mathrm{q}}$ event, those hadrons are pointing towards
the $\mathrm{q}$ ($\overline{\mathrm{q}}$) parton that were produced by fragmentation from that
end of the string, according to the random procedure used in the
fragmentation routine. No particles point to the $\mathrm{g}$. This assignment
seldom agrees with the visual impression, and is not intended to.
The common block \ttt{LUJETS} has expanded with time, and can now house
4000 entries. This figure may seem ridiculously large, but actually the
previous limit of 2000 was often reached in studies of high-$p_{\perp}$
processes at the LHC and SSC. This is because the event record
contains not only the final particles, but also all intermediate partons
and hadrons, which subsequenty showered, fragmented or decayed. Included
are also a wealth of photons coming from $\pi^0$ decays; the simplest
way of reducing the size of the event record is actually to switch off
$\pi^0$ decays by \ttt{MDCY(LUCOMP(111),1)=0}. Also note that some
routines, such as \ttt{LUCLUS} and \ttt{LUCELL}, use memory after the
event record proper as a working area. Still, to change the size of
the common block, upwards or downwards, is easy: just do a global
substitute in the common block and change the \ttt{MSTU(4)} value to the
new number. If more than 10000 lines are to be used, the packing of
colour information should also be changed, see \ttt{MSTU(5)}.
\drawbox{COMMON/LUJETS/N,K(4000,5),P(4000,5),V(4000,5)}\label{p:LUJETS}
\begin{entry}
\itemc{Purpose:} to contain the event record, i.e. the complete list
of all partons and particles in the current event.
\iteme{N :}\label{p:N} number of lines in the \ttt{K}, \ttt{P} and
\ttt{V} matrices occupied by the current event. \ttt{N} is continuously
updated as the definition of the original configuration and the
treatment of fragmentation and decay proceed. In the following,
the individual parton/particle number, running between 1 and
\ttt{N}, is called \ttt{I}.
\iteme{K(I,1) :}\label{p:K} status code KS, which gives the current
status of the parton/particle stored in the line. The ground rule is
that codes 1--10 correspond to currently existing partons/particles,
while larger codes contain partons/particles which no longer exist,
or other kinds of event information.
\begin{subentry}
\iteme{= 0 :} empty line.
\iteme{= 1 :} an undecayed particle or an unfragmented jet, the latter
being either a single jet or the last one of a jet system.
\iteme{= 2 :} an unfragmented jet, which is followed by more jets in the
same colour-singlet jet system.
\iteme{= 3 :} an unfragmented jet with special colour flow information
stored in \ttt{K(I,4)} and \ttt{K(I,5)}, such that adjacent partons
along the string need not follow each other in the event record.
\iteme{= 4 :} a particle which could have decayed, but did not
within the allowed volume around the original vertex.
\iteme{= 5 :} a particle which is to be forced to decay in the next
\ttt{LUEXEC} call, in the vertex position given (this code is only
set by user intervention).
\iteme{= 11 :} a decayed particle or a fragmented jet, the latter
being either a single jet or the last one of a jet system, cf. \ttt{=1}.
\iteme{= 12 :} a fragmented jet, which is followed by more jets in the
same colour-singlet jet system, cf. \ttt{=2}. Further, a $\mathrm{B}$ meson
which decayed as a $\br{\mathrm{B}}$ one, or vice versa, because of
$\mathrm{B}$--$\br{\mathrm{B}}$ mixing, is marked with this code rather than
\ttt{=11}.
\iteme{= 13 :} a jet which has been removed when special colour flow
information has been used to rearrange a jet system, cf. \ttt{=3}.
\iteme{= 14 :} a parton which has branched into further partons, with
special colour-flow information provided, cf. \ttt{=3}.
\iteme{= 15 :} a particle which has been forced to decay (by user
intervention), cf. \ttt{=5}.
\iteme{= 21 :} documentation lines used to give a compressed story of
the event at the beginning of the event record.
\iteme{= 31 :} lines with information on sphericity, thrust or cluster
search.
\iteme{= 32 :} tabular output, as generated by \ttt{LUTABU}.
\iteme{= 41 :} junction (currently not fully implemented).
\iteme{< 0 :} these codes are never used by the program, and are
therefore usually not affected by operations on the record, such as
\ttt{LUROBO}, \ttt{LULIST} and event-analysis routines (the exception
is some \ttt{LUEDIT} calls, where lines are moved but not deleted).
Such codes may therefore be useful in some connections.
\end{subentry}
\iteme{K(I,2) :} parton/particle KF code, as described in section
\ref{ss:codes}.
\iteme{K(I,3) :} line number of parent particle or jet, where known,
otherwise 0. Note that the assignment of a particle to a given jet in a
jet system is unphysical, and what is given there is only related to
the way the event was generated.
\iteme{K(I,4) :} normally the line number of the first daughter;
it is 0 for an undecayed particle or unfragmented jet.
For \ttt{K(I,1) = 3, 13} or \ttt{14}, instead, it contains special
colour-flow information (for internal use only) of the form \\
\ttt{K(I,4)} = 200000000*MCFR + 100000000*MCTO + 10000*ICFR + ICTO, \\
where ICFR and ICTO give the line numbers of the partons from which
the colour comes and to where it goes, respectively; MCFR and
MCTO originally are 0 and are set to 1 when the corresponding
colour connection has been traced in the \ttt{LUPREP} rearrangement
procedure. (The packing may be changed with \ttt{MSTU(5)}.)
The `from' colour position may indicate a parton which branched
to produce the current parton, or a parton created together with
the current parton but with matched anticolour, while the `to'
normally indicates a parton that the current parton branches
into. Thus, for setting up an initial colour configuration, it
is normally only the `from' part that is used, while the `to' part
is added by the program in a subsequent call to parton-shower
evolution (for final-state radiation; it is the other way around
for initial-state radiation).
{\bf Note:} normally most users never have to worry about the exact
rules for colour-flow storage, since this is used mainly for
internal purposes. However, when it is necessary to define this
flow, it is recommended to use the \ttt{LUJOIN} routine, since it is
likely that this would reduce the chances of making a mistake.
\iteme{K(I,5) :} normally the line number of the last daughter;
it is 0 for an undecayed particle or unfragmented jet.
For \ttt{K(I,1) = 3, 13} or \ttt{14}, instead, it contains special
colour-flow information (for internal use only) of the form \\
\ttt{K(I,5)} = 200000000*MCFR + 100000000*MCTO + 10000*ICFR + ICTO, \\
where ICFR and ICTO give the line numbers of the partons from which
the anticolour comes and to where it goes, respectively; MCFR
and MCTO originally are 0 and are set to 1 when the corresponding
colour connection has been traced in the \ttt{LUPREP} rearrangement
procedure. For further discussion, see \ttt{K(I,4)}.
\iteme{P(I,1) :}\label{p:P} $p_x$, momentum in the $x$ direction,
in GeV/$c$.
\iteme{P(I,2) :} $p_y$, momentum in the $y$ direction, in GeV/$c$.
\iteme{P(I,3) :} $p_z$, momentum in the $z$ direction, in GeV/$c$.
\iteme{P(I,4) :} $E$, energy, in GeV.
\iteme{P(I,5) :} $m$, mass, in GeV/$c^2$. In parton showers, with
space-like virtualities, i.e. where $Q^2 = - m^2 > 0$,
one puts \ttt{P(I,5)}$ = -Q$.
\iteme{V(I,1) :}\label{p:V} $x$ position of production vertex, in mm.
\iteme{V(I,2) :} $y$ position of production vertex, in mm.
\iteme{V(I,3) :} $z$ position of production vertex, in mm.
\iteme{V(I,4) :} time of production, in mm/$c$
($\approx 3.33 \times 10^{-12}$ s).
\iteme{V(I,5) :} proper lifetime of particle, in mm/$c$
($\approx 3.33 \times 10^{-12}$ s). If the particle is not expected to
decay, \ttt{V(I,5)=0}. A line with \ttt{K(I,1)=4}, i.e. a
particle that could have decayed, but did not within the
allowed region, has the proper non-zero \ttt{V(I,5)}.
In the absence of electric or magnetic fields, or other
disturbances, the decay vertex \ttt{VP} of an unstable particle
may be calculated as \\
\ttt{VP(j) = V(I,j) + V(I,5)*P(I,j)/P(I,5)},
\ttt{j} = 1--4.
\end{entry}
\subsection{How The Event Record Works}
The event record is the main repository for information about an
event. In the generation chain, it is used as a `scoreboard' for
what has already been done and what remains to be done.
This information can be studied by you, to access information
not only about the final state, but also about what came before.
\subsubsection{A simple example}
The example of section \ref{ss:JETstarted} may help to clarify what
is going on. When \ttt{LU2ENT} is called to generate a $\mathrm{q}\overline{\mathrm{q}}$
pair, the quarks are stored in lines 1 and 2 of the event record,
respectively. Colour information is set to show that they belong
together as a colour singlet. The counter \ttt{N} is also updated
to the value of 2. At no stage is the previously generated event
removed. Lines 1 and 2 are overwritten, but lines 3
onwards still contain whatever may have been there before. This does
not matter, since \ttt{N} indicates where the `real' record ends.
As \ttt{LUEXEC} is called, explicitly by you or indirectly
by \ttt{LU2ENT}, the first entry is considered and found to be
the first jet of a system. Therefore the second entry is also
found, and these two together form a jet system, which may be
allowed to fragment. The `string' that fragments is put in line 3
and the fragmentation products in lines 4 through 10 (in this
particular case). At the same time, the $\mathrm{q}$ and $\overline{\mathrm{q}}$ in the
first two lines are marked as having fragmented, and the same for
the string. At this stage, \ttt{N} is 10. Internally there is
another counter with the value 2, which indicates how far down
in the record the event has been studied.
This second counter is gradually increased by one. If the entry in
the corresponding line can fragment or decay, then fragmentation or
decay is perfomed.
The fragmentation/decay products are added at the end of the event
record, and \ttt{N} is updated accordingly. The entry is then also
marked as having been treated. For instance, when line 3 is
considered, the `string' entry of this line is seen to have been
fragmented,
and no action is taken. Line 4, a $\rho^+$, is allowed to decay to
$\pi^+ \pi^0$; the decay products are stored in lines 11 and 12,
and line 4 is marked as having decayed. Next, entry 5 is allowed to
decay. The entry in line 6, $\pi^+$, is a stable particle (by
default) and is therefore passed by without any action being taken.
In the beginning of the process, entries are usually unstable, and
\ttt{N} grows faster than the second counter of treated entries.
Later on, an increasing fraction of the entries are stable end
products, and the r\^oles are now reversed, with the second counter
growing faster. When the two coincide, the end of the record has been
reached, and the process can be stopped. All unstable objects have
now been allowed to fragment or decay. They are still present in the
record, so as to simplify the tracing of the history.
Notice that \ttt{LUEXEC} could well be called a second time.
The second counter would then start all over from the beginning, but
slide through until the end without causing any action, since
all objects that can be treated already have been.
Unless some of the relevant switches were changed meanwhile, that
is. For instance, if $\pi^0$ decays were switched off the first time
around but on the second, all the $\pi^0$'s found in the record
would be allowed to decay in the second call. A particle once
decayed is not `undecayed', however, so if the $\pi^0$ is put back
stable and \ttt{LUEXEC} is called a third time, nothing will happen.
\subsubsection{Applications to PYTHIA}
\label{sss:PYrecord}
In a full-blown event generated with {\tsc{Pythia}}, the usage of \ttt{LUJETS}
is more complicated, although the general principles survive.
\ttt{LUJETS} is used extensively both by the {\tsc{Pythia}} and the {\tsc{Jetset}}
routines; indeed it provides the bridge that allows
the general utility routines in {\tsc{Jetset}} to be used
also for {\tsc{Pythia}} events. The {\tsc{Pythia}} event listing begins (optionally)
with a few lines of event summary, specific to the hard process
simulated and thus not described in the overview above. These
specific parts are covered in the following.
In most instances, only the partons and particles actually produced
are of interest. For \ttt{MSTP(125)=0}, the event record starts
off with the parton configuration existing after hard interaction,
initial- and final-state radiation, multiple interactions and beam
remnants have been considered. The partons are arranged in colour
singlet clusters, ordered as required for string fragmentation.
Also photons and leptons produced as part of the hard interaction
(e.g. from $\mathrm{q}\overline{\mathrm{q}} \to \mathrm{g} \gamma$ or $\u\overline{\mathrm{u}} \to \mathrm{Z}^0 \to \e^+\e^-$)
appear in this part of the event record. These original entries
appear with pointer \ttt{K(I,3)=0}, whereas the products of the
subsequent fragmentation and decay have \ttt{K(I,3)} numbers
pointing back to the line of the parent.
The standard documentation, obtained with \ttt{MSTP(125)=1},
includes a few lines at the beginning of the event record, which
contain a brief summary of the process that has taken place. The
number of lines used depends on the nature of the hard process
and is stored in \ttt{MSTI(4)} for the current event. These lines
all have \ttt{K(I,1)=21}. For all processes, lines 1 and 2 give
the two incoming hadrons. When listed with \ttt{LULIST}, these two
lines will be separated from subsequent ones by a sequence of
`\ttt{======}' signs, to improve readability. For diffractive and
elastic events, the two outgoing states in lines 3 and 4 complete the
list. Otherwise, lines 3 and 4 contain the two partons that initiate
the two initial-state parton showers, and 5 and 6 the end products of
these showers, i.e. the partons that enter the hard interaction. With
initial-state radiation switched off, lines 3 and 5 and lines 4 and 6
coincide. For a simple $2 \to 2$ hard scattering, lines 7 and 8 give
the two outgoing partons/particles from the hard interaction, before
any final-state radiation. For $2 \to 2$ processes proceeding via an
intermediate resonance such as $\gamma^* / \Z^0$, $\mathrm{W}^{\pm}$ or $\H^0$, the
resonance is found in line 7 and the two outgoing partons/particles in
8 and 9. In some cases one of these may be a resonance in its own
right, or both of them, so that further pairs of lines are added for
subsequent decays. If the decay of a given resonance has been
switched off, then no decay products are listed either in this
initial summary or in the subsequent ordinary listing. Whenever partons
are listed, they are assumed to be on the mass shell for simplicity.
The fact that effective masses may be generated by initial-
and final-state radiation is taken into account in the actual parton
configuration that is allowed to fragment, however. A special case is
provided by $\mathrm{W}^+ \mathrm{W}^-$ or $\mathrm{Z}^0 \mathrm{Z}^0$ fusion to an $\H^0$. Then the
virtual $\mathrm{W}$'s or $\mathrm{Z}$'s are shown in lines 7 and 8, the $\H^0$ in
line 9, and the two recoiling quarks (that emitted the bosons) in 10
and 11, followed by the Higgs decay products. Since the $\mathrm{W}$'s and
$\mathrm{Z}$'s are space-like, what is actually listed as the mass for them
is $-\sqrt{-m^2}$. The listing of the event documentation closes with
another line made up of `\ttt{======}' signs.
A few examples may help clarify the picture. For a single diffractive
event $\mathrm{p} \overline{\mathrm{p}} \to \mathrm{p}_{\mrm{diffr}} \overline{\mathrm{p}}$, the event record will start
with \\
\verb& I K(I,1) K(I,2) K(I,3) & comment \\
\verb& 1 21 2212 0 & incoming $\mathrm{p}$ \\
\verb& 2 21 -2212 0 & incoming $\overline{\mathrm{p}}$ \\
\verb&========================= & not part of record; appears in
listings \\
\verb& 3 21 27 1 & outgoing $\mathrm{p}_{\mrm{diffr}}$ \\
\verb& 4 21 -2212 2 & outgoing $\overline{\mathrm{p}}$ \\
\verb&========================= & again not part of record
The typical QCD $2 \to 2$ process would be \\
\verb& I K(I,1) K(I,2) K(I,3) & comment \\
\verb& 1 21 2212 0 & incoming $\mathrm{p}$ \\
\verb& 2 21 -2212 0 & incoming $\overline{\mathrm{p}}$ \\
\verb&========================= & \\
\verb& 3 21 2 1 & $\u$ picked from incoming $\mathrm{p}$ \\
\verb& 4 21 -1 2 & $\overline{\mathrm{d}}$ picked from incoming
$\overline{\mathrm{p}}$ \\
\verb& 5 21 21 3 & $\u$ evolved to $\mathrm{g}$ at hard
scattering \\
\verb& 6 21 -1 4 & still $\overline{\mathrm{d}}$ at hard scattering \\
\verb& 7 21 21 0 & outgoing $\mathrm{g}$ from hard
scattering \\
\verb& 8 21 -1 0 & outgoing $\overline{\mathrm{d}}$ from hard
scattering \\
\verb&========================= &
Note that, where well defined, the \ttt{K(I,3)} code does contain
information as to which side the different partons come from, e.g.
above the gluon in line 5 points back to the $\u$ in line 3,
which points back to the proton in line 1. In the example above, it
would have been possible to associate the scattered g in line 7
with the incoming one in line 5, but this is not possible in the
general case, consider e.g. $\mathrm{g} \mathrm{g} \to \mathrm{g} \mathrm{g}$.
As a final example, $\mathrm{W}^+\mathrm{W}^-$ fusion to an $\H^0$ in process 8
(not process 124, which is lengthier) might look like \\
\verb& I K(I,1) K(I,2) K(I,3) & comment \\
\verb& 1 21 2212 0 & first incoming $\mathrm{p}$ \\
\verb& 2 21 2212 0 & second incoming $\mathrm{p}$ \\
\verb&========================= & \\
\verb& 3 21 2 1 & $\u$ picked from first $\mathrm{p}$ \\
\verb& 4 21 21 2 & $\mathrm{g}$ picked from second $\mathrm{p}$ \\
\verb& 5 21 2 3 & still $\u$ after initial-state
radiation \\
\verb& 6 21 -4 4 & $\mathrm{g}$ evolved to $\overline{\mathrm{c}}$ \\
\verb& 7 21 24 5 & space-like $\mathrm{W}^+$ emitted by $\u$
quark \\
\verb& 8 21 -24 6 & space-like $\mathrm{W}^-$ emitted by
$\overline{\mathrm{c}}$ quark \\
\verb& 9 21 25 0 & Higgs produced by $\mathrm{W}^+ \mathrm{W}^-$
fusion \\
\verb&10 21 1 5 & $\u$ turned into $\d$ by emission
of $\mathrm{W}^+$ \\
\verb&11 21 -3 6 & $\overline{\mathrm{c}}$ turned into $\overline{\mathrm{s}}$ by
emission of $\mathrm{W}^-$ \\
\verb&12 21 23 9 & first $\mathrm{Z}^0$ coming from decay
of $\H^0$ \\
\verb&13 21 23 9 & second $\mathrm{Z}^0$ coming from decay
of $\H^0$ \\
\verb&14 21 12 12 & $\nu_{\mathrm{e}}$ from first $\mathrm{Z}^0$
decay \\
\verb&15 21 -12 12 & $\br{\nu}_{\mathrm{e}}$ from first
$\mathrm{Z}^0$ decay \\
\verb&16 21 5 13 & $\b$ quark from second $\mathrm{Z}^0$
decay \\
\verb&17 21 -5 13 & $\overline{\mathrm{b}}$ antiquark from second
$\mathrm{Z}^0$ decay \\
\verb&========================= &
After these lines with the initial information, the event record looks
the same as for \ttt{MSTP(125)=0}, i.e. first comes the parton
configuration to be fragmented and, after another separator line
`\ttt{======}' in the output (but not the event record), the products
of subsequent fragmentation and decay chains. The \ttt{K(I,3)}
pointers for the partons, as well as leptons and photons produced
in the hard interaction, are now pointing towards the documentation
lines above, however. In particular, beam remnants point to 1 or 2,
depending on which side they belong to, and partons emitted in the
initial-state parton showers point to 3 or 4. In the second example
above, the partons produced by final-state radiation will be pointing
back to 7 and 8; as usual, it should be remembered that a specific
assignment to 7 or 8 need not be unique. For the third example,
final-state radiation partons will come both from partons 10 and 11
and from partons 16 and 17, and additionally there will be a
neutrino--antineutrino pair pointing to 14 and 15. The extra
pairs of partons that are generated by multiple interactions do not
point back to anything, i.e. they have \ttt{K(I,3)=0}.
There exists a third documentation option, \ttt{MSTP(125)=2}. Here
the history of initial- and final-state parton branchings may be traced,
including all details on colour flow. This information has not been
optimized for user-friendliness, and cannot be recommended for
general usage. With this option, the initial documentation lines
are the same. They are followed by blank lines, \ttt{K(I,1)=0}, up to
line 20 (can be changed in \ttt{MSTP(126)}). From line 21 onwards
each parton with \ttt{K(I,1)=} 3, 13 or 14 appears with special
colour-flow information in the \ttt{K(I,4)} and \ttt{K(I,5)}
positions. For an ordinary $2 \to 2$ scattering, the two incoming
partons at the hard scattering are stored in lines 21 and 22, and the
two outgoing in 23 and 24. The colour flow between these partons has
to be chosen according to the proper relative probabilities in
cases when many alternatives are possible, see section
\ref{sss:QCDjetclass}.
If there is initial-state radiation, the two partons in lines 21 and
22 are copied down to lines 25 and 26, from which the initial-state
showers are reconstructed backwards step by step. The branching
history may be read by noting that, for a branching $a \to b c$,
the \ttt{K(I,3)} codes of $b$ and $c$ point towards the line number
of $a$. Since the showers are reconstructed backwards, this actually
means that parton $b$ would appear in the listing before parton
$a$ and $c$, and hence have a pointer to a position below
itself in the list. Associated time-like partons $c$ may initiate
time-like showers, as may the partons of the hard scattering. Again
a showering parton or pair of partons will be copied down towards
the end of the list and allowed to undergo successive branchings
$c \to d e$, with $d$ and $e$ pointing towards $c$. The mass of
time-like partons is properly stored in \ttt{P(I,5)}; for space-like
partons $-\sqrt{-m^2}$ is stored instead. After this
section, containing all the branchings, comes the final parton
configuration, properly arranged in colour, followed by all
subsequent fragmentation and decay products, as usual.
\subsection{The HEPEVT Standard}
\label{ss:HEPEVT}
A set of common blocks was developed and agreed on within the
framework of the 1989 LEP physics study, see \cite{Sjo89}.
This standard defines an event record structure which should make
the interfacing of different event generators much simpler.
It would be a major work to rewrite {\tsc{Pythia/Jetset}} to agree with this
standard event record structure. More importantly, the standard
only covers quantities which can be defined unambiguously, i.e.
which are independent of the particular program used. There are
thus no provisions for the need for colour-flow information in
models based on string fragmentation, etc., so the standard
common blocks would anyway have to be supplemented with additional
event information. For the moment, the adopted approach is therefore
to retain the \ttt{LUJETS} event record, but supply a routine
\ttt{LUHEPC} which can convert to or from the standard event record.
Owing to a somewhat different content in the two records, some
ambiguities do exist in the translation procedure. \ttt{LUHEPC}
has therefore to be used with some judgment.
In this section, the new standard event structure is first presented,
i.e. the most important points in \cite{Sjo89} are recapitulated.
Thereafter the conversion routine is described, with particular
attention to ambiguities and limitations.
The standard event record is stored in two common blocks. The second
of these is specifically intended for spin information. Since {\tsc{Jetset}}
never (explicitly) makes use of spin information, this latter
common block is not addressed here. A third common block for colour
flow information has been discussed, but never formalized.
In order to make the components of the standard more distinguishable
in user programs, the three characters \ttt{HEP} (for High Energy
Physics) have been chosen to be a part of all names.
Originally it was not specified whether real variables should be in
single or double precision. At the time, this meant that single
precision became the default choice, but since then the trend has been
towards increasing precision. In connection with the 1995 LEP~2
workshop, it was therefore agreed to adopt \ttt{DOUBLE PRECISION}
real variables as part of the standard.
\drawboxfour{~PARAMETER (NMXHEP=2000)}
{~COMMON/HEPEVT/NEVHEP,NHEP,ISTHEP(NMXHEP),IDHEP(NMXHEP),}
{\&JMOHEP(2,NMXHEP),JDAHEP(2,NMXHEP),PHEP(5,NMXHEP),VHEP(4,NMXHEP)}
{~DOUBLE PRECISION PHEP, VHEP}
\label{p:HEPEVT}\begin{entry}
\itemc{Purpose:} to contain an event record in a
Monte Carlo-independent format.
\iteme{NMXHEP:} maximum numbers of entries (partons/particles) that can
be stored in the common block. The default value of 2000 can be changed
via the parameter construction. In the translation, it is
checked that this value is not exceeded.
\iteme{NEVHEP:} is normally the event number, but may have special
meanings, according to the description below:
\begin{subentry}
\iteme{> 0 :} event number, sequentially increased by 1 for each call
to the main event generation routine, starting with 1 for the
first event generated.
\iteme{= 0 :} for a program which does not keep track of event numbers,
as {\tsc{Jetset}}.
\iteme{= -1 :} special initialization record; not used by {\tsc{Jetset}}.
\iteme{= -2 :} special final record; not used by {\tsc{Jetset}}.
\end{subentry}
\iteme{NHEP:} the actual number of entries stored in the current event.
These are found in the first \ttt{NHEP} positions of the respective
arrays below. Index \ttt{IHEP}, 1$\leq$\ttt{IHEP}$\leq$\ttt{NHEP},
is used below to denote a given entry.
\iteme{ISTHEP(IHEP):} status code for entry \ttt{IHEP}, with the
following meanings:
\begin{subentry}
\iteme{= 0 :} null entry.
\iteme{= 1 :} an existing entry, which has not decayed or fragmented.
This is the main class of entries, which represents the
`final state' given by the generator.
\iteme{= 2 :} an entry which has decayed or fragmented and is
therefore not appearing in the final state, but is retained for
event history information.
\iteme{= 3 :} a documentation line, defined separately from the event
history. This could include the two incoming reacting particles, etc.
\iteme{= 4 - 10 :} undefined, but reserved for future standards.
\iteme{= 11 - 200 :} at the disposal of each model builder for
constructs specific to his program, but equivalent to a null line
in the context of any other program.
\iteme{= 201 - :} at the disposal of users, in particular for event
tracking in the detector.
\end{subentry}
\iteme{IDHEP(IHEP) :} particle identity, according to the PDG
standard. The four additional codes 91--94 have been introduced
to make the event history more legible, see section \ref{ss:codes}
and the \ttt{MSTU(16)} description.
\iteme{JMOHEP(1,IHEP) :} pointer to the position where the mother
is stored. The value is 0 for initial entries.
\iteme{JMOHEP(2,IHEP) :} pointer to position of second mother.
Normally only one mother exists, in which case the value 0 is to be
used. In {\tsc{Jetset}}, entries with codes 91--94 are the only ones to have
two mothers. The flavour contents of these objects, as well as
details of momentum sharing, have to be found by looking at the
mother partons, i.e. the two partons in positions \ttt{JMOHEP(1,IHEP)}
and \ttt{JMOHEP(2,IHEP)} for a cluster or a shower system, and the
range
\ttt{JMOHEP(1,IHEP)}--\ttt{JMOHEP(2,IHEP)} for a string or an
independent fragmentation parton system.
\iteme{JDAHEP(1,IHEP) :} pointer to the position of the first daughter.
If an entry has not decayed, this is 0.
\iteme{JDAHEP(2,IHEP) :} pointer to the position of the last daughter.
If an entry has not decayed, this is 0. It is assumed that daughters are
stored sequentially, so that the whole range
\ttt{JDAHEP(1,IHEP)}--\ttt{JDAHEP(2,IHEP)} contains daughters. This
variable should be set also when only one daughter is present, as in
$\mathrm{K}^0 \to \mathrm{K}_{\mrm{S}}^0$ decays, so that looping from the first
daughter to the last one works transparently.
Normally daughters are stored after mothers, but in backwards
evolution of initial-state radiation the opposite may appear,
i.e. that mothers are found below the daughters they branch into.
Also, the two daughters then need not appear one after the other,
but may be separated in the event record.
\iteme{PHEP(1,IHEP) :} momentum in the $x$ direction, in GeV/$c$.
\iteme{PHEP(2,IHEP) :} momentum in the $y$ direction, in GeV/$c$.
\iteme{PHEP(3,IHEP) :} momentum in the $z$ direction, in GeV/$c$.
\iteme{PHEP(4,IHEP) :} energy, in GeV.
\iteme{PHEP(5,IHEP) :} mass, in GeV/$c^2$. For space-like partons,
it is allowed to use a negative mass, according to
\ttt{PHEP(5,IHEP)}$ = -\sqrt{-m^2}$.
\iteme{VHEP(1,IHEP) :} production vertex $x$ position, in mm.
\iteme{VHEP(2,IHEP) :} production vertex $y$ position, in mm.
\iteme{VHEP(3,IHEP) :} production vertex $z$ position, in mm.
\iteme{VHEP(4,IHEP) :} production time, in mm/$c$
($\approx 3.33 \times 10^{-12}$ s).
\end{entry}
\boxsep
This completes the brief description of the standard. In {\tsc{Jetset}}, the
routine \ttt{LUHEPC} is provided as an interface.
\drawbox{CALL LUHEPC(MCONV)}\label{p:LUHEPC}
\begin{entry}
\itemc{Purpose:} to convert between the \ttt{LUJETS} event record and
the \ttt{HEPEVT} event record.
\iteme{MCONV :} direction of conversion.
\begin{subentry}
\iteme{= 1 :} translates the current \ttt{LUJETS} record into the
\ttt{HEPEVT} one, while leaving the original \ttt{LUJETS} one
unaffected.
\iteme{= 2 :} translates the current \ttt{HEPEVT} record into the
\ttt{LUJETS} one, while leaving the original \ttt{HEPEVT} one
unaffected.
\end{subentry}
\end{entry}
\boxsep
The conversion of momenta is trivial: it is just a matter of exchanging
the order of the indices. The vertex information is but little more
complicated; the extra fifth component present in \ttt{LUJETS} can be
easily
reconstructed from other information for particles which have decayed.
(Some of the advanced features made possible by this component, such as
the possibility to consider decays within expanding spatial volumes in
subsequent \ttt{LUEXEC} calls, cannot be used if the record is
translated back and forth, however.) Also, the particle codes
\ttt{K(I,2)} and \ttt{IDHEP(I)}
are identical, since they are both based on the PDG codes.
The remaining, non-trivial areas deal with the status codes and the
event history. In moving from \ttt{LUJETS} to \ttt{HEPEVT},
information on colour flow is lost. On the other hand, the position
of a second mother, if any, has to be found; this only affects lines
with \ttt{K(I,2)=} 91--94. Also, for lines with \ttt{K(I,1)=} 13
or 14, the daughter pointers have to be found. By and large,
however, the translation from \ttt{LUJETS} to \ttt{HEPEVT}
should cause little problem, and there should never be any need for
user intervention. (We assume that {\tsc{Jetset}} is run with the default
\ttt{MSTU(16)=1}, otherwise some discrepancies with respect to the
proposed standard event history description will be present.)
In moving from \ttt{HEPEVT} to \ttt{LUJETS}, information on a second
mother is lost. Any codes \ttt{IDHEP(I)} not equal to 1, 2 or 3 are
translated into \ttt{K(I,1)=0}, and so all entries with
\ttt{K(I,1)}$\geq 30$ are effectively lost in a translation back and
forth. All entries with \ttt{IDHEP(I)=2} are translated
into \ttt{K(I,1)=11}, and so entries of type
\ttt{K(I,1) = 12, 13, 14} or \ttt{15} are never found. There is thus
no colour-flow information available for partons which have
fragmented. For partons with \ttt{IDHEP(I)=1},
i.e. which have not fragmented, an attempt is made to subdivide the
partonic system into colour singlets, as required for subsequent
string fragmentation. To this end, it is assumed that partons are
stored sequentially along strings. Normally, a string would then start
at a $\mathrm{q}$ ($\overline{\mathrm{q}}$) or $\overline{\mathrm{q}}\qbar$ ($\mathrm{q}\q$) entry, cover a number
of intermediate gluons, and end at a $\overline{\mathrm{q}}$ ($\mathrm{q}$) or $\mathrm{q}\q$
($\overline{\mathrm{q}}\qbar$) entry. Particles could be interspersed in this list with
no adverse effects, i.e. a $\u-\mathrm{g}-\gamma-\overline{\mathrm{u}}$
sequence would be interpreted as a $\u-\mathrm{g}-\overline{\mathrm{u}}$ string plus an
additional photon. A closed gluon loop would be assumed to be made up
of a sequential listing of the gluons, with the string continuing from
the last gluon up back to the first one. Contrary to the previous, open
string case, the appearance of any particle but a gluon would therefore
signal the end of the gluon loop. For example, a $\mathrm{g}-\mathrm{g}-\mathrm{g}-\mathrm{g}$ sequence
would be interpreted as one single four-gluon loop, while a
$\mathrm{g}-\mathrm{g}-\gamma-\mathrm{g}-\mathrm{g}$ sequence would be seen as composed of two
2-gluon systems.
If these interpretations, which are not unique, are not to your liking,
it is up to you to correct them, e.g. by using
\ttt{LUJOIN} to tell exactly which partons should be joined,
in which sequence, to give a string. Calls to \ttt{LUJOIN}
(or the equivalent) are also necessary if \ttt{LUSHOW} is to be used
to have some partons develop a shower.
For practical applications, one should note that {\tsc{Jetset}} $\e^+\e^-$ events,
which have been allowed to shower but not to fragment, do have partons
arranged in the order assumed above, so that a translation to
\ttt{HEPEVT} and back does not destroy the possibility to perform
fragmentation by a simple \ttt{LUEXEC} call. Also the hard interactions
in {\tsc{Pythia}} fulfil this condition, while problems may appear in the
multiple interaction scenario, where several closed $\mathrm{g}\g$ loops may
appear directly following one another, and thus would be
interpreted as a single multigluon loop after translation back and
forth.
\clearpage
\section{Hard Processes in JETSET}
\label{s:JETSETproc}
{\tsc{Jetset}} contains the simulation of two hard processes.
The process of main interest is $\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q} \overline{\mathrm{q}}$.
Higher-order QCD corrections can be obtained either with parton
showers or with second-order matrix elements. The details of the
parton-shower evolution are given in section \ref{s:showinfi},
while this section contains the matrix-element description, including
a summary of the {\tsc{Jetset}} algorithm for initial-state
photon radiation. Also {\tsc{Pythia}} can be used to simulate the
process $\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q} \overline{\mathrm{q}}$,
but without the options of using second-order matrix
elements or polarized incoming beams. Some other differences
between the two algorithms are described.
The other hard process in {\tsc{Jetset}} is $\Upsilon$ decay to
$\mathrm{g} \mathrm{g} \mathrm{g}$ or $\gamma \mathrm{g} \mathrm{g}$, which is briefly commented on.
The main sources of information for this chapter are
refs. \cite{Sjo83,Sjo86,Sjo89}.
\subsection{Annihilation Events in the Continuum}
\label{ss:eematrix}
The description of $\e^+\e^-$ annihilation into hadronic events involves a
number of components: the $s$ dependence of the total cross section
and flavour composition, multijet matrix elements, angular
orientation of events, initial-state photon bremsstrahlung
and effects of initial-state electron polarization.
Many of the published formulae
have been derived for the case of massless outgoing quarks. For each
of the components described in the following, we will begin by
discussing the massless case, and then comment on what is done to
accommodate massive quarks.
\subsubsection{Electroweak cross sections}
In the standard theory, fermions have the following couplings
(illustrated here for the first generation):
\begin{center}
\begin{tabular}{lll@{\protect\rule{0mm}{\tablinsep}}}
$e_{\nu} = 0$, & $v_{\nu} = 1$, & $a_{\nu} = 1$, \\
$e_{\mathrm{e}} = -1$, & $v_{\mathrm{e}} = -1 + 4\sin^2 \! \theta_W$, & $a_{\mathrm{e}} = -1$, \\
$e_{\u} = 2/3$, & $v_{\u} = 1 - 8\sin^2 \! \theta_W /3$, & $a_{\nu} = 1$, \\
$e_{\d} = -1/3$, & $v_{\d} = -1 + 4\sin^2 \! \theta_W /3$, & $a_{\d} = -1$, \\
\end{tabular}
\end{center}
with $e$ the electric charge, and $v$ and $a$ the vector and axial
couplings to the $\mathrm{Z}^0$. The relative energy dependence of the weak
neutral current to the electromagnetic one is given by
\begin{equation}
\chi(s) = \frac{1}{4\sin^2 \! \theta_W\cos^2 \! \theta_W} \;
\frac{s}{s - m_{\mathrm{Z}}^2 + i m_{\mathrm{Z}}\Gamma_{\mathrm{Z}}} ~,
\label{ee:chis}
\end{equation}
where $s = E_{\mrm{cm}}^2$.
In {\tsc{Jetset}} the electroweak mixing parameter $\sin^2 \! \theta_W$ and the
$\mathrm{Z}^0$ mass $m_{\mathrm{Z}}$ and width $\Gamma_{\mathrm{Z}}$ are considered as
constants to be given by you (while
{\tsc{Pythia}} itself calculates an $s$-dependent width).
Although the incoming $\mathrm{e}^+$ and $\mathrm{e}^-$ beams are normally
unpolarized, we have included the possibility of polarized beams,
following the formalism of \cite{Ols80}. Thus the incoming
$\mathrm{e}^+$ and $\mathrm{e}^-$ are characterized by polarizations
$\mbf{P}^{\pm}$ in the rest frame of the particles:
\begin{equation}
\mbf{P}^{\pm} = P_{\mrm{T}}^{\pm} \hat{\mbf{s}}^{\pm} +
P_{\mrm{L}}^{\pm} \hat{\mbf{p}}^{\pm} ~,
\end{equation}
where $0 \leq P_{\mrm{T}}^{\pm} \leq 1$ and
$-1 \leq P_{\mrm{L}}^{\pm} \leq 1$, with the constraint
\begin{equation}
(\mbf{P}^{\pm})^2 = (P_{\mrm{T}}^{\pm})^2 + (P_{\mrm{L}}^{\pm})^2
\leq 1 ~.
\end{equation}
Here $\hat{\mbf{s}}^{\pm}$ are unit vectors perpendicular to the beam
directions $\hat{\mbf{p}}^{\pm}$. To be specific, we choose a
right-handed coordinate frame with
$\hat{\mbf{p}}^{\pm} = (0,0, \mp 1)$,
and standard transverse polarization directions (out of the machine
plane for storage rings) $\hat{\mbf{s}}^{\pm} = (0, \pm 1,0)$, the
latter corresponding to azimuthal angles $\varphi^{\pm} = \pm \pi /2$.
As free parameters in the program we choose $P_{\mrm{L}}^+$,
$P_{\mrm{L}}^-$, $P_{\mrm{T}} = \sqrt{P_{\mrm{T}}^+ P_{\mrm{T}}^-}$
and $\Delta \varphi = (\varphi^+ + \varphi^-) /2$.
In the massless QED case, the probability to produce a flavour $\mathrm{f}$ is
proportional to $e_{\mathrm{f}}^2$, i.e up-type quarks are four times as likely
as down-type ones. In lowest-order massless QFD the corresponding
relative probabilities are given by \cite{Ols80}
\begin{eqnarray}
h_{\mathrm{f}}(s) & = & e_{\mathrm{e}}^2 \, (1 - P_{\mrm{L}}^+ P_{\mrm{L}}^-)
\, e_{\mathrm{f}}^2 \, + \, 2 e_{\mathrm{e}} \left\{ v_{\mathrm{e}}
(1 - P_{\mrm{L}}^+ P_{\mrm{L}}^-) - a_{\mathrm{e}}
(P_{\mrm{L}}^- - P_{\mrm{L}}^+)
\right\} \, \Re\chi(s) \, e_{\mathrm{f}} v_{\mathrm{f}} \, + \nonumber \\
& & + \, \left\{ (v_{\mathrm{e}}^2 + a_{\mathrm{e}}^2) (1 - P_{\mrm{L}}^+
P_{\mrm{L}}^-) - 2 v_{\mathrm{e}} a_{\mathrm{e}} (P_{\mrm{L}}^- - P_{\mrm{L}}^+)
\right\} \,
\left| \chi(s) \right|^2 \, \left\{ v_{\mathrm{f}}^2 + a_{\mathrm{f}}^2 \right\} ~,
\label{ee:hf}
\end{eqnarray}
where $\Re\chi(s)$ denotes the real part of $\chi(s)$.
The $h_{\mathrm{f}}(s)$ expression depends both on the $s$ value and on the
longitudinal polarization of the $\mathrm{e}^{\pm}$ beams in a non-trivial way.
The cross section for the process $\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{f} \overline{\mathrm{f}}$
may now be written as
\begin{equation}
\sigma_{\mathrm{f}}(s) = \frac{4 \pi \alpha_{\mathrm{em}}^2}{3 s} R_{\mathrm{f}}(s) ~,
\end{equation}
where $R_{\mathrm{f}}$ gives the ratio to the lowest-order QED cross section for
the process $\e^+\e^- \to \mu^+ \mu^-$,
\begin{equation}
R_{\mathrm{f}}(s) = N_C \, R_{\mrm{QCD}} \, h_{\mathrm{f}}(s) ~.
\end{equation}
The factor of $N_C = 3$ counts the number of colour states available
for the $\mathrm{q}\overline{\mathrm{q}}$ pair. The $R_{\mrm{QCD}}$
factor takes into account QCD loop corrections to the cross section.
For $n_f$ effective flavours (normally $n_f =5$)
\begin{equation}
R_{\mrm{QCD}} \approx 1 + \frac{\alpha_{\mathrm{s}}}{\pi} + (1.986 - 0.115 n_f)
\left( \frac{\alpha_{\mathrm{s}}}{\pi} \right)^2 + \cdots
\label{ee:RQCD}
\end{equation}
in the $\br{\mrm{MS}}$ renormalization scheme \cite{Din79}.
Note that $R_{\mrm{QCD}}$ does not affect the relative quark-flavour
composition, and so is of peripheral interest in {\tsc{Jetset}}.
(For leptons the $N_C$ and $R_{\mrm{QCD}}$ factors would be absent,
i.e. $N_C \, R_{\mrm{QCD}} = 1$, but leptonic final states are not
generated in {\tsc{Jetset}}.)
Neglecting higher-order QCD and QFD effects, the corrections for
massive quarks are given in terms of the velocity $v_{\mathrm{q}}$ of a quark
with mass $m_{\mathrm{q}}$, $v_{\mathrm{q}} = \sqrt{ 1 - 4 m_{\mathrm{q}}^2 /s}$, as follows.
The vector quark current terms in $h_{\mathrm{f}}$ (proportional to
$e_{\mathrm{f}}^2$, $e_{\mathrm{f}} v_{\mathrm{f}}$, or $v_{\mathrm{f}}^2$) are multiplied by a
threshold factor $v_{\mathrm{q}} (3 - v_{\mathrm{q}}^2) /2$, while the axial
vector quark current term (proportional to $a_{\mathrm{f}}^2$) is
multiplied by $v_{\mathrm{q}}^3$. While inclusion of quark masses in the
QFD formulae decreases the total cross section, first-order QCD
corrections tend in the opposite direction \cite{Jer81}. Na\"{\i}vely,
one would expect one factor of $v_{\mathrm{q}}$ to get cancelled. So far,
the available options are either to include threshold factors
in full or not at all.
Given that all five quarks are light at
the scale of the $\mathrm{Z}^0$, the issue of quark masses is not really
of interest at LEP. Here, however, purely weak corrections are
important, in particular since they change the $\b$ quark
partial width differently from that of the other ones \cite{Kuh89}.
No such effects are included in the program.
\subsubsection{First-order QCD matrix elements}
The Born process $\e^+\e^- \to \mathrm{q} \overline{\mathrm{q}}$ is modified in first-order
QCD by the probability for the $\mathrm{q}$ or $\overline{\mathrm{q}}$ to
radiate a gluon, i.e. by the process $\e^+\e^- \to \mathrm{q} \overline{\mathrm{q}} \mathrm{g}$.
The matrix element is conveniently given in terms of scaled energy
variables in the c.m. frame of the event,
$x_1 = 2E_{\mathrm{q}}/E_{\mrm{cm}}$,
$x_2 = 2E_{\overline{\mathrm{q}}}/E_{\mrm{cm}}$,
and $x_3 = 2E_{\mathrm{g}}/E_{\mrm{cm}}$,
i.e. $x_1 + x_2 + x_3 = 2$. For massless
quarks the matrix element reads \cite{Ell76}
\begin{equation}
\frac{1}{\sigma_0} \, \frac{\d \sigma}{\d x_1 \, \d x_2} =
\frac{\alpha_{\mathrm{s}}}{2\pi} \, C_F \,
\frac{x_1^2 + x_2^2}{(1-x_1)(1-x_2)} ~,
\label{ee:ME3j}
\end{equation}
where $\sigma_0$ is the lowest-order cross section, $C_F = 4/3$ is the
appropriate colour factor, and
the kinematically allowed region is $0 \leq x_i \leq 1, i = 1, 2, 3$.
By kinematics, the $x_k$ variable for parton $k$ is related to the
invariant mass $m_{ij}$ of the other two partons $i$ and $j$ by
$y_{ij} = m_{ij}^2/E_{\mrm{cm}}^2 = 1 - x_k$.
The strong coupling constant $\alpha_{\mathrm{s}}$ is in first order given by
\begin{equation}
\alpha_{\mathrm{s}}(Q^2) = \frac{12\pi}{(33-2n_f) \, \ln(Q^2/\Lambda^2)} ~.
\label{ee:aS3j}
\end{equation}
Conventionally $Q^2 = s = E_{\mrm{cm}}^2$; we will return to this
issue below.
The number of flavours $n_f$ is 5 for LEP applications, and so the
$\Lambda$ value determined is $\Lambda_5$ (while e.g. most
deep inelastic scattering studies refer to $\Lambda_4$,
the energies for these experiments being below the bottom threshold).
The $\alpha_{\mathrm{s}}$ values are matched at flavour thresholds, i.e.
as $n_f$ is changed the $\Lambda$ value is also changed. It is
therefore the derivative of $\alpha_{\mathrm{s}}$ that changes at a
threshold, not $\alpha_{\mathrm{s}}$ itself.
In order to separate 2-jets from 3-jets, it is useful to
introduce jet-resolution parameters. This can be done in several
different ways. Most famous are the $y$ and $(\epsilon, \delta)$
procedures. We will only refer to the $y$ cut, which is the one
used in the program. Here a 3-parton configuration is called
a 2-jet event if
\begin{equation}
\min_{i,j} (y_{ij}) = \min_{i,j} \left( \frac{m_{ij}^2}{E_{\mrm{cm}}^2}
\right) < y ~.
\end{equation}
The cross section in eq.~(\ref{ee:ME3j}) diverges for
$x_1 \rightarrow 1$ or $x_2 \rightarrow 1$ but, when
first-order propagator and vertex corrections are included,
a corresponding singularity with opposite sign appears in the
$\mathrm{q} \overline{\mathrm{q}}$ cross section, so that the total cross section is finite.
In analytical calculations, the average value of any well-behaved
quantity ${\cal Q}$ can therefore be calculated as
\begin{equation}
\left\langle {\cal Q} \right\rangle =
\frac{1}{\sigma_{\mrm{tot}}} \lim_{y \rightarrow 0}
\left( {\cal Q}(\mrm{2parton}) \, \sigma_{\mrm{2parton}}(y) +
\int_{y_{ij} > y} {\cal Q}(x_1,x_2) \,
\frac{\d \sigma_{\mrm{3parton}}}{\d x_1 \, \d x_2} \,
\d x_1 \, \d x_2 \right) ~,
\label{ee:Obs}
\end{equation}
where any explicit $y$ dependence disappears in the limit
$y \rightarrow 0$.
In a Monte Carlo program, it is not possible to
work with a negative total 2-jet rate, and thus it is necessary to
introduce a fixed non-vanishing $y$ cut in the 3-jet
phase space. Experimentally, there is evidence for the need of a
low $y$ cut, i.e. a large 3-jet rate.
For LEP applications, the recommended value is $y = 0.01$,
which is about as far down as one can go and still retain a positive
2-jet rate. With $\alpha_{\mathrm{s}} = 0.12$, in full second-order QCD
(see below), the $2:3:4$ jet composition is then approximately
$11 \% : 77 \% : 12 \%$.
Note, however, that initial-state QED radiation may
occasionally lower the c.m. energy significantly, i.e. increase
$\alpha_{\mathrm{s}}$, and thereby bring the 3-jet fraction above unity
if $y$ is kept fixed at 0.01 also in those events. Therefore,
at PETRA/PEP energies, $y$ values slightly above 0.01 are needed.
In addition to the $y$ cut, the program contains a cut on the
invariant mass $m_{ij}$ between any two partons, which is typically
required to be larger than 2 GeV. This cut corresponds to the
actual merging of two nearby parton jets, i.e. where a treatment with
two separate partons rather than one would be superfluous in view
of the smearing arising from the subsequent fragmentation. Since
the cut-off mass scale $\sqrt{y} E_{\mrm{cm}}$ normally is much larger,
this additional cut only enters for events at low energies.
For massive quarks, the amount of QCD radiation is slightly reduced
\cite{Iof78}:
\begin{eqnarray}
\frac{1}{\sigma_0} \, \frac{\d \sigma}{\d x_1 \, \d x_2} & = &
\frac{\alpha_{\mathrm{s}}}{2\pi}
\, C_F \, \left\{ \frac{x_1^2 + x_2^2}{(1-x_1)(1-x_2)} -
\frac{4 m_{\mathrm{q}}^2}{s} \left( \frac{1}{1-x_1} + \frac{1}{1-x_2}
\right) \right. \nonumber \\[1mm]
& & - \left. \frac{2 m_{\mathrm{q}}^2}{s} \left( \frac{1}{(1-x_1)^2} +
\frac{1}{(1-x_2)^2} \right) - \frac{4 m_{\mathrm{q}}^4}{s^2}
\left( \frac{1}{1-x_1} + \frac{1}{1-x_2} \right)^2 \right\} ~.
\label{ee:threejMEmass}
\end{eqnarray}
In addition, the phase space for emission is reduced by the
requirement
\begin{equation}
\frac{(1-x_1)(1-x_2)(1-x_3)}{x_3^2} \geq \frac{m_{\mathrm{q}}^2}{s} ~.
\end{equation}
For $\b$ quarks at LEP energies, these corrections are fairly small.
\subsubsection{4-jet matrix elements}
Two new event types are added in second-order QCD,
$\e^+\e^- \to \mathrm{q} \overline{\mathrm{q}} \mathrm{g} \mathrm{g}$ and $\e^+\e^- \to \mathrm{q} \overline{\mathrm{q}} \mathrm{q}' \overline{\mathrm{q}}'$.
The 4-jet cross section has been calculated by several
groups \cite{Ali80a,Gae80,Ell81,Dan82}, which agree on the result.
The formulae are too lengthy to be quoted here. In one of the
calculations \cite{Ali80a}, quark masses were explicitly included,
but {\tsc{Jetset}} only includes the massless expressions, as taken
from \cite{Ell81}. Here the angular orientation of the event has been
integrated out, so that five independent internal kinematical
variables remain. These may be related to the six $y_{ij}$ and
the four $y_{ijk}$ variables,
$y_{ij} = m_{ij}^2 / s = (p_i + p_j)^2 / s$ and
$y_{ijk} = m_{ijk}^2 / s = (p_i + p_j + p_k)^2 / s$,
in terms of which the matrix elements are given.
The original calculations were for the pure $\gamma$-exchange case;
it was recently pointed out \cite{Kni89} that an additional
contribution to the $\e^+\e^- \to \mathrm{q} \overline{\mathrm{q}} \mathrm{q}' \overline{\mathrm{q}}'$ cross section
arises from the axial part of the $\mathrm{Z}^0$. This term is not included
in the program, but fortunately it is finite and small.
Whereas the way the string, i.e. the fragmenting colour flux tube,
is stretched is uniquely given in $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ event, for
$\mathrm{q} \overline{\mathrm{q}} \mathrm{g} \mathrm{g}$ events there are two possibilities:
\mbox{$\mathrm{q} - \mathrm{g}_1 - \mathrm{g}_2 - \overline{\mathrm{q}}$} or
\mbox{$\mathrm{q} - \mathrm{g}_2 - \mathrm{g}_1 - \overline{\mathrm{q}}$}.
A knowledge of quark and gluon colours, obtained by perturbation
theory, will uniquely specify the stretching of the string, as long
as the two gluons do not have the same colour. The probability for
the latter is down in magnitude by a factor $1 / N_C^2 = 1 / 9$.
One may either choose to neglect these terms entirely, or to keep
them for the choice of kinematical setup, but then drop them at the
choice of string drawing \cite{Gus82}. We have adopted the latter
procedure. Comparing the two possibilities, differences are
typically 10--20\% for a given kinematical configuration,
and less for the total 4-jet cross section, so from a practical
point of view this is not a major problem.
In higher orders, results depend on the renormalization scheme;
we will use $\br{\mrm{MS}}$ throughout. In addition to this choice,
several possible forms can be chosen for $\alpha_{\mathrm{s}}$,
all of which are equivalent to that order but differ in higher
orders. We have picked the recommended standard \cite{PDG88}
\begin{equation}
\label{ee:aS4j}
\alpha_{\mathrm{s}}(Q^2) =
\frac{12\pi}{(33-2n_f) \, \ln (Q^2 / \Lambda^2_{\br{\mrm{MS}}})}
\left\{ 1 - 6 \, \frac{153-19n_f}{(33-2n_f)^2} \,
\frac{\ln (\ln ( Q^2 / \Lambda^2_{\br{\mrm{MS}}}))}
{\ln ( Q^2 / \Lambda^2_{\br{\mrm{MS}}})}
\right\} ~.
\end{equation}
\subsubsection{Second-order 3-jet matrix elements}
As for first order, a full second-order calculation consists both of
real parton emission terms and of vertex and propagator corrections.
These modify the 3-jet and 2-jet cross sections.
Although there was some initial confusion, everybody soon agreed
on the size of the loop corrections \cite{Ell81,Ver81,Fab82}.
In analytic calculations, the procedure of eq.~(\ref{ee:Obs}),
suitably expanded, can therefore be used unambiguously for a
well-behaved variable.
For Monte Carlo event simulation, it is again necessary to impose
some finite jet-resolution criterion. This means that four-parton
events which fail the cuts should be reassigned either to the
3-jet or to the 2-jet event class. It is this area that
caused quite a lot of confusion in the past
\cite{Kun81,Got82,Ali82,Zhu83,Gut84,Gut87,Kra88},
and where full agreement does not exist. Most likely, agreement
will never be reached, since there are indeed ambiguous points
in the procedure, related to uncertainties on the theoretical
side, as follows.
For the $y$-cut case, any two partons with an invariant mass
$m_{ij}^2 < y E_{\mrm{cm}}^2$ should be recombined into one. If the
four-momenta are simply added, the sum will correspond to a parton
with a positive mass, namely the original $m_{ij}$.
The loop corrections are given in terms of final
massless partons, however. In order to perform the (partial)
cancellation between the four-parton real and the 3-parton
virtual contributions, it is therefore necessary to get rid of
the bothersome mass in the four-parton states. Several
recombinations are used in practice, which go under names such as
`E', `E0', `p' and `p0' \cite{OPA91}. In the `E'-type schemes,
the energy of a recombined parton is given by $E_{ij} = E_i + E_j$,
and three-momenta may have to be adjusted accordingly. In the
`p'-type schemes, on the other hand, three-momenta are added,
$\mbf{p}_{ij} = \mbf{p}_i + \mbf{p}_j$, and then energies may have
to be adjusted. These procedures result in different 3-jet
topologies, and therefore in different second-order differential
3-jet cross sections.
Within each scheme, a number of lesser points remain to be dealt
with, in particular what to do if a recombination of a nearby parton
pair were to give an event with a non-$\mathrm{q}\overline{\mathrm{q}}\mathrm{g}$ flavour structure.
{\tsc{Jetset}} contains two alternative second-order 3-jet
implementations, GKS and ERT(Zhu). For historical
reasons the former is default, but actually the latter is the
recommended one today. Other parametrizations have also been made
available that run together with {\tsc{Jetset}}, see
\cite{Sjo89,Mag89}.
The GKS option is based on the GKS \cite{Gut84} calculation, where
some of the original mistakes in FKSS \cite{Fab82} have been
corrected. The GKS formulae have the advantage of giving the
second-order corrections in closed analytic form, as not-too-long
functions of $x_1$, $x_2$, and the $y$ cut. However, it is
today recognized, also by the authors, that important
terms are still missing, and that the matrix elements
should therefore not be taken too seriously. The option is thus
kept mainly for backwards compatibility.
The ERT(Zhu) generator \cite{Zhu83} is based on the ERT matrix elements
\cite{Ell81}, with a Monte Carlo recombination procedure suggested
by Kunszt \cite{Kun81} and developed by Ali \cite{Ali82}. It has
the merit of giving corrections in a convenient, parametrized form.
For practical applications, the main limitation is that the
corrections are only given for discrete values of the cut-off
parameter $y$, namely $y$ = 0.01, 0.02, 0.03, 0.04, and 0.05.
The basic approach is the following. Without any loss of generality,
the full second-order 3-jet cross section can be written in
terms of the `ratio function' $R(X,Y;y)$, defined by
\begin{equation}
\frac{1}{\sigma_0} \frac{\d \sigma_3^{\mrm{tot}}}{\d X \, \d Y} =
\frac{\alpha_{\mathrm{s}}}{\pi} A_0(X,Y)
\left\{ 1 + \frac{\alpha_{\mathrm{s}}}{\pi} R(X,Y;y) \right\} ~,
\label{ee:Zhupar}
\end{equation}
where $X = x_1 - x_2 = x_{\mathrm{q}} - x_{\overline{\mathrm{q}}}$, $Y = x_3 = x_g$,
$\sigma_0$ is the lowest-order hadronic cross section,
and $A_0(X,Y)$ the standard first-order 3-jet cross section,
cf. eq.~(\ref{ee:ME3j}).
By Monte Carlo integration, the value of $R(X,Y;y)$ is
evaluated in bins of $(X,Y)$, and the result parametrized
by a simple function $F(X,Y;y)$.
In order to obtain the second-order 3-jet rate, a small cut
$y_0 = 10^{-7}$ was introduced. It was assumed that four-parton events
which fail this cut can be (partly) cancelled analytically against
the virtual 3-jet events, to give a net `regularized virtual'
contribution to the 3-jet rate. For a given choice
of $y$ cut, in the physical range $y \gg y_0$, an additional `soft'
contribution comes from four-parton events which survive the $y_0$ cut
but fail the $y$ one.
A large sample (9\,000\,000) of four-parton events was generated inside
the $y_0$ cut region. For events which failed the more stringent $y$
cuts, the parton pair with the smallest invariant mass was recombined
into an effective jet, using the `p0' recombination scheme.
This means that the individual three-momenta were added,
$\mbf{p}_{ij} = \mbf{p}_i + \mbf{p}_j$, the mass
of the recombined pair was set to zero for the calculation of energy,
$E_{ij} = |\mbf{p}_i + \mbf{p}_j|$, and
finally all four-momenta were rescaled by a common factor so as
to preserve the correct c.m. frame energy.
In calculating the ${\cal O}(\alpha_{\mathrm{s}}^2)$ correction functions,
care was taken
to maintain the flavour signature of the jets in the recombination
process. A quark and a gluon were recombined into a quark with the same
flavour as the original quark, two gluons were recombined to form
a gluon, etc. In some cases the
three jets of the final state were not in the
standard $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ configuration. The probability for this to happen
corresponded to less than 0.5\% of the total cross section, even
for the most stringent cuts used. For these non-$\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ final
states, the assignment of $\mathrm{q}$, $\overline{\mathrm{q}}$ and $\mathrm{g}$ was done at random.
The sum of `regularized virtual' (1\,000\,000 3-jet events were
generated, with evaluated second-order weights)
and `soft' corrections, normalized to the first-order 3-jet
cross section, was tabulated in the $(X,Y)$ plane, using bins of size
$0.05 \times 0.05$. This estimated $R$-function behaviour was then
fit with a 12-parameter function $F$,
\begin{eqnarray}
F(X,Y;y)& = & p_1 + p_2 X^2 + p_3 X^4 + (p_4+p_5 X^2)Y+
(p_6+p_7X^2)Y^2 + \nonumber\\
& & (p_8+p_9X^2)Y^3 +
p_{10}/(X^2-Y^2) + p_{11}/(1-Y) + p_{12}/Y ~.
\label{ee:ZhuFun}
\end{eqnarray}
The parameters $p_i$ are reproduced in \cite{Sjo89}.
\subsubsection{The matrix-element event generator scheme}
The program contains parametrizations, separately, of the total
first-order 3-jet rate, the total second-order 3-jet rate,
and the total 4-jet rate, all as functions of $y$ (with
$\alpha_{\mathrm{s}}$ as a separate prefactor).
These parametrizations have been obtained as follows:
\begin{Itemize}
\item
The first-order 3-jet matrix element is almost analytically
integrable; some small finite pieces were obtained by a truncated
series expansion of the relevant integrand.
\item
The GKS second-order 3-jet matrix elements were integrated for
40 different $y$-cut values, evenly distributed in $\ln y$ between
a smallest value $y = 0.001$ and the kinematical limit $y = 1/3$.
For each $y$ value, 250\,000 phase-space points were generated,
evenly in $\d \ln (1-x_i) = \d x_i/(1-x_i)$, $i = 1,2$, and the
second-order 3-jet rate in the point evaluated. The properly
normalized sum of weights in each of the 40 $y$ points were
then fitted to a polynomial in $\ln(y^{-1}-2)$. For the ERT(Zhu)
matrix elements the parametrizations in eq.~(\ref{ee:ZhuFun})
were used to perform a corresponding Monte Carlo integration for
the five $y$ values available.
\item
The 4-jet rate was integrated numerically, separately for
$\mathrm{q} \overline{\mathrm{q}} \mathrm{g} \mathrm{g}$ and $\mathrm{q} \overline{\mathrm{q}} \mathrm{q}' \overline{\mathrm{q}}'$ events, by generating large
samples of 4-jet phase-space points
within the boundary $y = 0.001$. Each point was classified according
to the actual minimum $y$ between any two partons. The same
events could then be used to update the summed weights for 40
different counters, corresponding to $y$ values evenly distributed
in $\ln y$ between $y = 0.001$ and the kinematical limit $y = 1/6$.
In fact, since
the weight sums for large $y$ values only received contributions
from few phase-space points, extra (smaller) subsamples of events were
generated with larger $y$ cuts. The summed weights,
properly normalized, were then parametrized in terms of
polynomials in $\ln(y^{-1} - 5)$.
Since it turned out to be difficult to obtain one single good fit
over the whole range of $y$ values, different parametrizations are
used above and below $y=0.018$. As originally given, the
$\mathrm{q} \overline{\mathrm{q}} \mathrm{q}' \overline{\mathrm{q}}'$ parametrization only took into account four
$\mathrm{q}'$ flavours, i.e. secondary $\b \overline{\mathrm{b}}$ pairs were not generated,
but this has been corrected for LEP.
\end{Itemize}
In the generation stage, each event is treated on its own, which means
that the $\alpha_{\mathrm{s}}$ and $y$ values may be allowed to vary from event to
event. The main steps are the following.
\begin{Enumerate}
\item
The $y$ value to be used in the current event is determined. If
possible, this is the value given by you, but additional
constraints exist from the validity of the parametrizations
($y \geq 0.001$ for GKS, $0.01 \leq y \leq 0.05$ for ERT(Zhu))
and an extra (user-modifiable) requirement of a minimum absolute
invariant mass between jets (which translates into varying $y$ cuts
due to the effects of initial-state QED radiation).
\item
The $\alpha_{\mathrm{s}}$ value is calculated.
\item
For the $y$ and $\alpha_{\mathrm{s}}$ values given, the relative
two/three/4-jet composition is determined. This is achieved by
using the parametrized functions of $y$ for 3- and 4-jet rates,
multiplied by the relevant number of factors of $\alpha_{\mathrm{s}}$.
In ERT(Zhu), where the second-order 3-jet rate is available
only at a few $y$ values, intermediate results are obtained by linear
interpolation in the ratio of second-order to first-order
3-jet rates. The 3-jet and 4-jet rates are normalized to
the analytically known second-order total event rate, i.e. divided
by $R_{\mrm{QCD}}$ of eq.~(\ref{ee:RQCD}). Finally, the 2-jet rate is
obtained by conservation of total probability.
\item
If the combination of $y$ and $\alpha_{\mathrm{s}}$ values is such that the total
3- plus 4-jet fraction is larger than unity, i.e. the remainder
2-jet fraction negative, the $y$-cut value is raised (for that event),
and the process is started over at point 3.
\item
The choice is made between generating a 2-, 3- or 4-jet event,
according to the relative probabilities.
\item
For the generation of 4-jets, it is first necessary to make a choice
between $\mathrm{q} \overline{\mathrm{q}} \mathrm{g} \mathrm{g}$ and $\mathrm{q} \overline{\mathrm{q}} \mathrm{q}' \overline{\mathrm{q}}'$ events, according to
the relative (parametrized) total cross sections. A phase-space point
is then selected, and the differential cross section at this point is
evaluated and compared with a parametrized maximum weight. If the
phase-space point is rejected, a new one is selected, until an
acceptable 4-jet event is found.
\item
For 3-jets, a phase-space point is first chosen according to the
first-order cross section. For this point, the weight
\begin{equation}
W(x_1,x_2;y) = 1 + \frac{\alpha_{\mathrm{s}}}{\pi} R(x_1,x_2;y)
\label{ee:WTJS}
\end{equation}
is evaluated. Here $R(x_1,x_2;y)$ is analytically given for GKS
\cite{Gut84}, while it is approximated by the parametrization
$F(X,Y;y)$ of eq.~(\ref{ee:ZhuFun}) for ERT(Zhu). Again, linear
interpolation of $F(X,Y;y)$ has to be applied for intermediate $y$
values. The weight $W$ is compared with a maximum weight
\begin{equation}
W_{\mathrm{max}}(y) = 1 + \frac{\alpha_{\mathrm{s}}}{\pi} R_{\mathrm{max}}(y) ~,
\end{equation}
which has been numerically determined beforehand and suitably
parametrized. If the phase-space point is rejected, a
new point is generated, etc.
\item
Massive matrix elements are not available in {\tsc{Jetset}} for second-order
QCD (but are in the first-order option). However, if a
3- or 4-jet event determined above falls outside
the phase-space region allowed for massive quarks, the event is
rejected and reassigned to be a 2-jet event. (The way the
$y_{ij}$ and $y_{ijk}$ variables of 4-jet events should be
interpreted for massive quarks is not even unique, so some latitute
has been taken here to provide a reasonable continuity from
3-jet events.) This procedure is known not to give the expected full
mass suppression, but is a reasonable first approximation.
\item
Finally, if the event is classified as a 2-jet event, either
because it was initially so assigned, or because it failed the
massive phase-space cuts for 3- and 4-jets, the
generation of 2-jets is trivial.
\end{Enumerate}
\subsubsection{Optimized perturbation theory}
Theoretically, it turns out that the second-order corrections to the
3-jet rate are large. It is therefore not unreasonable to expect
large third-order corrections to the 4-jet rate. Indeed, the
experimental 4-jet rate is much larger than second order predicts
(when fragmentation effects have been folded in),
if $\alpha_{\mathrm{s}}$ is determined based on the 3-jet rate
\cite{Sjo84a,JAD88}.
The only consistent way to resolve this issue is to go ahead and
calculate the full next order. This is a tough task, however, so
people have looked at possible shortcuts.
For example, one can try to minimize the higher-order contributions
by a suitable choice of the renormalization scale \cite{Ste81} ---
`optimized perturbation theory'. This
is equivalent to a different choice for the $Q^2$ scale in
$\alpha_{\mathrm{s}}$, a scale which is not unambiguous anyway. Indeed
the standard value $Q^2 = s = E_{\mrm{cm}}^2$ is larger than the
natural physical scale of gluon emission in events, given that most
gluons are fairly soft. One could therefore pick another scale,
$Q^2 = f s$, with $f < 1$. The
${\cal O}(\alpha_{\mathrm{s}})$ 3-jet rate would be increased by
such a scale change, and so would the number of 4-jet
events, including those which collapse into 3-jet ones. The loop
corrections depend on the $Q^2$ scale, however,
and compensate the changes above by giving a larger negative
contribution to the 3-jet rate.
The possibility of picking an optimized scale $f$ is implemented
as follows \cite{Sjo89}. Assume that the differential 3-jet
rate at scale $Q^2 = s$ is given by the expression
\begin{equation}
R_3 = r_1 \alpha_{\mathrm{s}} + r_2 \alpha_{\mathrm{s}}^2 ~,
\end{equation}
where $R_3$, $r_1$ and $r_2$ are functions of the kinematical
variables $x_1$ and $x_2$ and the $y$ cut, as described above.
When the coupling is chosen at a different scale, $Q'^2 = f s$,
the 3-jet rate has to be changed to
\begin{equation}
R_3' = r_1' \alpha_{\mathrm{s}}' + r_2 \alpha_{\mathrm{s}}'^2 ~,
\end{equation}
where $r_1' = r_1$,
\begin{equation}
r_2' = r_2 + r_1 \frac{33-2n_f}{12\pi} \ln f ~,
\label{ee:r2optim}
\end{equation}
and $\alpha_{\mathrm{s}}' = \alpha_{\mathrm{s}}(fs)$.
Since we only have the Born term for 4-jets, here the effects of a
scale change come only from the change in the coupling constant.
Finally, the 2-jet cross section can still be calculated from the
difference between the total cross section and the 3- and 4-jet
cross sections.
If an optimized scale is used in the program, the default value is
$f=0.002$, which is favoured by the studies in ref. \cite{Bet89}. (In
fact, it is also possible to use a correspondingly optimized
$R_{\mrm{QCD}}$ factor, eq.~(\ref{ee:RQCD}), but then the
corresponding $f$ is chosen independently and much closer to unity.)
The success of describing the jet rates should not hide the fact that
one is dabbling in (educated, hopefully) guesswork, and that any
conclusions based on this method have to be taken with a pinch of
salt.
One special problem associated with the use of optimized perturbation
theory is that the differential 3-jet rate may become negative
over large regions of the $(x_1, x_2)$ phase space. This problem
already exists, at least in principle, even for a scale $f = 1$,
since $r_2$ is not guaranteed to be positive definite. Indeed,
depending on the choice of $y$ cut, $\alpha_{\mathrm{s}}$ value and recombination
scheme, one may observe a small region of negative differential
3-jet rate for the full second-order expression. This region
is centred around $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ configurations, where the $\mathrm{q}$ and
$\overline{\mathrm{q}}$ are close together in one hemisphere and the $\mathrm{g}$ is alone in
the other, i.e. $x_1 \approx x_2 \approx 1/2$. It is well understood
why second-order corrections should be negative in this region
\cite{Dok89}: the $\mathrm{q}$ and $\overline{\mathrm{q}}$ of a $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ state are in a
relative colour octet state, and thus the colour force between them is
repulsive, which translates into a negative second-order term.
However, as $f$ is decreased below unity, $r_2'$ receives a negative
contribution from the $\ln f$ term, and the region of negative
differential cross section has a tendency to become larger, also
after taking into account related changes in $\alpha_{\mathrm{s}}$. In an
event-generator framework, where all events are supposed to come
with unit
weight, it is clearly not possible to simulate negative cross sections.
What happens in the program is therefore that no 3-jet events at
all are generated in the regions of negative differential cross section,
and that the 3-jet rate in regions of positive cross sections is
reduced by a constant factor, chosen so that the total number of
3-jet events comes out as it should. This is a consequence of the
way the program works, where it is first decided what kind of event to
generate, based on integrated 3-jet rates in which positive and
negative contributions are added up with sign, and only thereafter
the kinematics is chosen.
Based on our physics understanding of the origin of this negative
cross section, the approach adopted is as sensible as any, at least
to that order in perturbation theory (what one might strive for is a
properly exponentiated description of the relevant region). It can
give rise to funny results for low $f$ values, however, as observed
by OPAL \cite{OPA92} for the energy--energy correlation asymmetry.
\subsubsection{Angular orientation}
While pure $\gamma$ exchange gives a simple $1 + \cos^2\theta$
distribution for the $\mathrm{q}$ (and $\overline{\mathrm{q}}$) direction in $\mathrm{q} \overline{\mathrm{q}}$ events,
$\mathrm{Z}^0$ exchange and $\gamma^* / \Z^0$ interference results in a
forward--backward asymmetry. If one introduces
\begin{eqnarray}
h'_{\mathrm{f}}(s) & = & 2 e_{\mathrm{e}} \left\{ a_{\mathrm{e}}
(1 - P_{\mrm{L}}^+ P_{\mrm{L}}^-) -
v_{\mathrm{e}} (P_{\mrm{L}}^- - P_{\mrm{L}}^+) \right\} \,
\Re\chi(s) e_{\mathrm{f}} a_{\mathrm{f}}
\nonumber \\
& & + \, \left\{ 2 v_{\mathrm{e}} a_{\mathrm{e}} (1 - P_{\mrm{L}}^+ P_{\mrm{L}}^-) -
(v_{\mathrm{e}}^2 + a_{\mathrm{e}}^2) (P_{\mrm{L}}^- - P_{\mrm{L}}^+) \right\} \,
|\chi(s)|^2 \, v_{\mathrm{f}} a_{\mathrm{f}} ~,
\end{eqnarray}
then the angular distribution of the quark is given by
\begin{equation}
\frac{\d \sigma}{\d (\cos\theta_{\mathrm{f}})} \propto
h_{\mathrm{f}}(s)(1 + \cos^2\theta_{\mathrm{f}}) + 2 h'_{\mathrm{f}}(s) \cos\theta_{\mathrm{f}} ~.
\end{equation}
The angular orientation of a 3- or 4-jet event may be described
in terms of three angles $\chi$, $\theta$ and $\varphi$; for 2-jet
events only $\theta$ and $\varphi$ are necessary. From a standard
orientation, with the $\mathrm{q}$ along the $+z$ axis and the $\overline{\mathrm{q}}$ in the
$xz$ plane with $p_x > 0$, an arbitrary orientation may be reached by
the rotations $+\chi$ in azimuthal angle, $+\theta$ in polar angle,
and $+\varphi$ in azimuthal angle, in that order. Differential
cross sections,
including QFD effects and arbitrary beam polarizations have been given
for 2- and 3-jet events in refs. \cite{Ols80,Sch80}. We use the
formalism of ref. \cite{Ols80}, with $\chi \to \pi - \chi$ and
$\varphi^- \to - (\varphi + \pi/2)$. The resulting formulae are
tedious, but
straightforward to apply, once the internal jet configuration has been
chosen. 4-jet events are approximated by 3-jet ones, by joining
the two gluons of a $\mathrm{q} \overline{\mathrm{q}} \mathrm{g} \mathrm{g}$ event and the $\mathrm{q}'$ and $\overline{\mathrm{q}}'$
of a $\mathrm{q} \overline{\mathrm{q}} \mathrm{q}' \overline{\mathrm{q}}'$ event into one effective jet. This means
that some angular asymmetries are neglected \cite{Ali80a}, but that weak
effects are automatically included. It is assumed that the second-order
3-jet events have the same angular orientation as the first-order
ones, some studies on this issue may be found in \cite{Kor85}. Further,
the formulae normally refer to the massless case; only for the QED
2- and 3-jet cases are mass corrections available.
The main effect of the angular distribution of multijet events
is to smear the lowest-order result, i.e. to reduce any anisotropies
present in 2-jet systems. In the parton-shower option of the program,
only the initial $\mathrm{q} \overline{\mathrm{q}}$ axis is determined. The subsequent shower
evolution then {\it de facto} leads to a smearing of the jet axis,
although not necessarily in full agreement with the expectations
from multijet matrix-element treatments.
\subsubsection{Initial-state radiation}
Initial-state photon radiation has been included using the formalism of
ref. \cite{Ber82}. Here each event contains either no photon or one,
i.e. it is a first-order non-exponentiated description.
The main formula for the hard radiative photon
cross section is
\begin{equation}
\frac{\d \sigma}{\d x_{\gamma}} = \frac{\alpha_{\mathrm{em}}}{\pi} \,
\left( \ln\frac{s}{m_{\mathrm{e}}^2} -1 \right) \,
\frac{1 + (1-x_{\gamma})^2}{x_{\gamma}} \, \sigma_0 (\hat{s}) ~,
\end{equation}
where $x_{\gamma}$ is the photon energy fraction of the beam energy,
$\hat{s} = (1-x_{\gamma}) s$ is the squared reduced hadronic c.m.
energy, and $\sigma_0$ is the ordinary annihilation cross section at
the reduced energy. In particular, the selection of jet flavours
should be done according to expectations at the reduced
energy. The cross section is divergent both for $x_{\gamma} \to 1$ and
$x_{\gamma} \to 0$. The former is related to the fact that
$\sigma_0$ has a $1/\hat{s}$ singularity (the real photon pole) for
$\hat{s} \to 0$. An upper cut on $x_{\gamma}$ can here be chosen
to fit the
experimental setup. The latter is a soft photon singularity, which is
to be compensated in the no-radiation cross section. A requirement
$x_{\gamma} > 0.01$ has therefore been chosen so that the hard-photon
fraction is smaller than unity. In the total cross section, effects
from photons with $x_{\gamma} < 0.01$ are taken into account, together
with vertex and vacuum polarization corrections (hadronic vacuum
polarizations using a simple parametrization of the more complicated
formulae of ref. \cite{Ber82}).
The hard photon spectrum can be integrated analytically, for the
full $\gamma^* / \Z^0$ structure including interference terms, provided that
no new flavour thresholds are crossed and that the $R_{\mrm{QCD}}$
term in the cross section can be approximated by a constant over the
range of allowed $\hat{s}$ values. In fact, threshold effects can be
taken into account by standard rejection techniques, at the price of
not obtaining the exact cross section analytically, but only by an
effective Monte Carlo integration taking place in parallel with the
ordinary event generation. In addition to $x_{\gamma}$, the polar
angle $\theta_{\gamma}$ and azimuthal angle $\varphi_{\gamma}$ of
the photons are also to be chosen. Further, for the orientation
of the hadronic system, a choice has to be made whether the photon is
to be considered as having been radiated from the $\mathrm{e}^+$ or from the
$\mathrm{e}^-$.
Final-state photon radiation, as well as interference between initial-
and final-state radiation, has been left out of this treatment. The
formulae for $\e^+\e^- \to \mu^+ \mu^-$ cannot be simply taken over for
the case of outgoing quarks, since the quarks as such only live for
a short while before turning into hadrons. Another simplification in
our treatment is that effects of incoming polarized $\mathrm{e}^{\pm}$ beams
have been completely neglected, i.e. neither the effective shift in
azimuthal distribution of photons nor the reduction in polarization
is included. The polarization parameters of the program are to be
thought of as the effective polarization surviving after
initial-state radiation.
\subsubsection{Alternative matrix elements}
The program contains two sets of `toy model' matrix elements, one
for an Abelian vector gluon model and one for a scalar gluon model.
Clearly both of these alternatives are already excluded by data,
and are anyway not viable alternatives for a consistent theory of
strong interactions. They are therefore included more as references
to show how well the characteristic features of QCD can be measured
experimentally.
Second-order matrix elements are available for the Abelian vector
gluon model. These are easily obtained from the standard QCD matrix
elements by a substitution of the Casimir group factors:
$C_F = 4/3 \to 1$, $N_C = 3 \to 0$, and $T_R = n_{\mathrm{f}}/2 \to 3 n_{\mathrm{f}}$.
First-order matrix elements contain only $C_F$; therefore the
standard first-order QCD results may be recovered by a rescaling
of $\alpha_{\mathrm{s}}$ by a factor $4/3$. In second order the change of $N_C$
to 0 means that $\mathrm{g} \to \mathrm{g}\g$ couplings are absent from the Abelian
model, while the change of $T_R$ corresponds to an enhancement of the
$\mathrm{g} \to \mathrm{q}'\overline{\mathrm{q}}'$ coupling, i.e. to an enhancement of the
$\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ 4-jet event rate.
The second-order corrections to the 3-jet rate
turn out to be strongly negative --- if
$\alpha_{\mathrm{s}}$ is fitted to get about the right rate of 4-jet events,
the predicted differential 3-jet rate is negative almost
everywhere in the $(x_1, x_2)$ plane. Whether this unphysical
behaviour would be saved by higher orders is unclear. It has been
pointed out that the rate can be made positive by a suitable choice of
scale, since $\alpha_{\mathrm{s}}$ runs in opposite directions in an Abelian
model and in QCD \cite{Bet89}. This may be seen directly from eq.
(\ref{ee:r2optim}), where the term $33 = 11 N_C$ is absent in the
Abelian model, and therefore the scale-dependent term changes sign.
In the program, optimized scales have not been implemented for this
toy model. Therefore the alternatives provided for you are either
to generate only 4-jet events, or to neglect second-order
corrections to the 3-jet rate, or to have the total 3-jet
rate set vanishing (so that only 2- and 4-jet events are
generated). Normally we would expect the former to be the one of most
interest, since it is in angular (and flavour) distributions of 4-jet
events that the structure of QCD can be tested.
Also note that the `correct' running of $\alpha_{\mathrm{s}}$ is not included;
you are expected to use the option where $\alpha_{\mathrm{s}}$ is just given as
a constant number.
The scalar gluon model is even more excluded than the Abelian vector
one, since differences appear already in the 3-jet matrix
element \cite{Lae80}:
\begin{equation}
\frac{\d \sigma}{\d x_1 \, \d x_2} \propto \frac{x_3^2}{(1-x_1)(1-x_2)}
\end{equation}
when only $\gamma$ exchange is included. The axial part of the $\mathrm{Z}^0$
gives a slightly different shape; this is included in the program but
does not make much difference. The angular orientation does include
the full $\gamma^* / \Z^0$ interference \cite{Lae80}, but the main interest is
in the 3-jet topology as such \cite{Ell79}. No higher-order
corrections are included. It is recommended to use the option of a
fixed $\alpha_{\mathrm{s}}$ also here, since the correct running is not available.
\subsection{Decays of Onia Resonances}
\label{ss:oniadecays}
Many different possibilities are open for the decay of heavy
$J^{PC} = 1^{--}$ onia resonances. Of special interest are
the decays into three gluons or two gluons plus a photon, since these
offer unique possibilities to study a `pure sample' of gluon jets.
A routine for this purpose is included in the program. It was written
at a time where the expectations were to find toponium at PETRA
energies.
If, as now seems likely, the top mass is above 100~GeV, weak decays
will dominate, to the extent that the top quark will decay weakly even
before a bound toponium state is formed, and thus the routine will be
of no use for top. The charm system, on the other hand, is far too low
in mass for a jet language to be of any use. The only application is
therefore likely to be for $\Upsilon$, which unfortunately also is on
the low side in mass.
The matrix element for $\mathrm{q} \overline{\mathrm{q}} \to \mathrm{g} \mathrm{g} \mathrm{g}$ is (in lowest order)
\cite{Kol78}
\begin{equation}
\frac{1}{\sigma_{\mathrm{g} \mathrm{g} \mathrm{g}}}
\frac{\d \sigma_{\mathrm{g} \mathrm{g} \mathrm{g}}}{\d x_1 \, \d x_2} =
\frac{1}{\pi^2 - 9} \left\{ \left( \frac{1-x_1}{x_2 x_3} \right)^2 +
\left( \frac{1-x_2}{x_1 x_3} \right)^2 +
\left( \frac{1-x_3}{x_1 x_2} \right)^2 \right\} ~,
\label{ee:Upsilondec}
\end{equation}
where, as before, $x_i = 2 E_i / E_{\mrm{cm}}$ in the c.m. frame of
the event. This is a well-defined expression, without the kind of
singularities encountered in the $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ matrix elements.
In principle, no cuts at all would be necessary, but for reasons
of numerical simplicity we implement a $y$ cut as for continuum
jet production, with all events not fulfilling this cut considered
as (effective) $\mathrm{g} \mathrm{g}$ events. For $\mathrm{g} \mathrm{g} \mathrm{g}$ events, each $\mathrm{g} \mathrm{g}$
invariant mass is required to be at least 2 GeV.
Another process is $\mathrm{q} \overline{\mathrm{q}} \to \gamma \mathrm{g} \mathrm{g}$, obtained by replacing
a gluon in $\mathrm{q} \overline{\mathrm{q}} \to \mathrm{g} \mathrm{g} \mathrm{g}$ by a photon. This process has the
same normalized cross section as the one above, if e.g. $x_1$ is taken
to refer to the photon. The relative rate is \cite{Kol78}
\begin{equation}
\frac{\sigma_{\gamma \mathrm{g} \mathrm{g}}}{\sigma_{\mathrm{g} \mathrm{g} \mathrm{g}}} =
\frac{36}{5} \, \frac{e_{\mathrm{q}}^2 \, \alpha_{\mathrm{em}}}
{\alpha_{\mathrm{s}}(Q^2)} ~.
\end{equation}
Here $e_{\mathrm{q}}$ is the charge of the heavy quark, and the scale in
$\alpha_{\mathrm{s}}$ has been chosen as the mass of the onium state. If the
mass of the recoiling $\mathrm{g} \mathrm{g}$ system is lower than some cut-off
(by default 2 GeV), the event is rejected.
In the present implementation the angular orientation of the
$\mathrm{g} \mathrm{g} \mathrm{g}$ and $\gamma \mathrm{g} \mathrm{g}$ events is given for the
$\e^+\e^- \to \gamma^* \to$ onium case \cite{Kol78} (optionally with
beam polarization effects included), i.e. weak effects have not
been included, since they are negligible at around 10~GeV.
It is possible to start a perturbative shower evolution from either
of the two states above. However, for $\Upsilon$ the phase space
for additional evolution is so constrained that not much is to be
gained from that. We therefore do not recommend this possibility.
The shower generation machinery, when starting up from a
$\gamma \mathrm{g} \mathrm{g}$ configuration, is constructed such that the
photon energy is not changed. This means that there is currently no
possibility to use showers to bring the theoretical photon
spectrum in better agreement with the experimental one.
In string fragmentation language, a $\mathrm{g} \mathrm{g} \mathrm{g}$ state corresponds to
a closed string triangle with the three gluons at the corners. As
the partons move apart from a common origin, the string triangle
expands. Since the photon does not take part in the fragmentation,
the $\gamma \mathrm{g} \mathrm{g}$ state corresponds to a double string running
between the two gluons.
\subsection{Routines and Common Block Variables}
\label{ss:eeroutines}
\subsubsection{$\e^+\e^-$ continuum event generation}
The only routine a normal user will call to generate $\e^+\e^-$ continuum
events is \ttt{LUEEVT}. The other routines listed below, as well as
\ttt{LUSHOW} (see section \ref{ss:showrout}), are called by
\ttt{LUEEVT}.
\drawbox{CALL LUEEVT(KFL,ECM)}\label{p:LUEEVT}
\begin{entry}
\itemc{Purpose:} to generate a complete event
$\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q}\overline{\mathrm{q}} \to$ parton shower $\to$ hadrons
according to QFD and QCD cross sections. As an alternative to
parton showers, second-order matrix elements are available for
$\mathrm{q}\overline{\mathrm{q}} + \mathrm{q}\overline{\mathrm{q}}\mathrm{g} + \mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g + \mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ production.
\iteme{KFL :} flavour of events generated.
\begin{subentry}
\iteme{= 0 :} mixture of all allowed flavours according to relevant
probabilities.
\iteme{= 1 - 8 :} primary quarks are only of the specified flavour
\ttt{KFL}.
\end{subentry}
\iteme{ECM :} total c.m. energy of system.
\itemc{Remark:} Each call generates one event, which is independent
of preceding ones, with one exception, as follows. If radiative
corrections are included, the shape of the hard photon spectrum is
recalculated only with each \ttt{LUXTOT} call, which normally is done
only if \ttt{KFL}, \ttt{ECM} or \ttt{MSTJ(102)} is changed. A change
of e.g. the $\mathrm{Z}^0$ mass in mid-run has to be followed either by a user
call to \ttt{LUXTOT} or by an internal call forced e.g. by putting
\ttt{MSTJ(116)=3}.
\end{entry}
\boxsep
\begin{entry}
\iteme{SUBROUTINE LUXTOT(KFL,ECM,XTOT) :}\label{p:LUXTOT}
to calculate the total hadronic cross section,
including quark thresholds, weak, beam polarization, and QCD effects
and radiative corrections. In the process, variables necessary
for the treatment of hard photon radiation are calculated and
stored.
\begin{subentry}
\iteme{KFL, ECM :} as for \ttt{LUEEVT}.
\iteme{XTOT :} the calculated total cross section in nb.
\end{subentry}
\iteme{SUBROUTINE LURADK(ECM,MK,PAK,THEK,PHIK,ALPK) :}\label{p:LURADK}
to describe initial-state hard $\gamma$ radiation.
\iteme{SUBROUTINE LUXKFL(KFL,ECM,ECMC,KFLC) :}\label{p:LUXKFL}
to generate the primary quark flavour in case this
is not specified by the user.
\iteme{SUBROUTINE LUXJET(ECM,NJET,CUT) :}\label{p:LUXJET}
to determine the number of jets (2, 3 or 4) to be
generated within the kinematically allowed region (characterized by
\ttt{CUT} $= y_{\mrm{cut}}$) in the matrix-element approach; to be
chosen such that all probabilities are between 0 and 1.
\iteme{SUBROUTINE LUX3JT(NJET,CUT,KFL,ECM,X1,X2) :}\label{p:LUX3JT}
to generate the internal momentum variables of a
3-jet event, $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}$, according to first- or second-order
QCD matrix elements.
\iteme{SUBROUTINE LUX4JT(NJET,CUT,KFL,ECM,KFLN,X1,X2,X4,X12,X14) :}%
\label{p:LUX4JT}
to generate the internal momentum variables for a
4-jet event, $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g$ or $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$, according to
second-order QCD matrix elements.
\iteme{SUBROUTINE LUXDIF(NC,NJET,KFL,ECM,CHI,THE,PHI) :}\label{p:LUXDIF}
to describe the angular orientation of the jets.
In first-order QCD the complete QED or QFD formulae are used; in
second order 3-jets are assumed to have the same orientation
as in first, and 4-jets are approximated by 3-jets.
\end{entry}
\subsubsection{A routine for onium decay}
In \ttt{LUONIA} we have implemented the decays of heavy onia
resonances into three gluons or two gluons plus a photon, which are
the dominant non-background-like decays of $\Upsilon$.
\drawbox{CALL LUONIA(KFL,ECM)}\label{p:LUONIA}
\begin{entry}
\itemc{Purpose:} to simulate the process
$\e^+\e^- \to \gamma^* \to 1^{--}$ onium resonance $\to (\mathrm{g}\g\mathrm{g}$ or
$\mathrm{g}\g\gamma) \to$ shower $\to$ hadrons.
\iteme{KFL :} the flavour of the quark giving rise to the resonance.
\begin{subentry}
\iteme{= 0 :} generate $\mathrm{g}\g\mathrm{g}$ events alone.
\iteme{= 1 - 8 :} generate $\mathrm{g}\g\mathrm{g}$ and $\mathrm{g}\g\gamma$ events in mixture
determined by the squared charge of flavour \ttt{KFL}. Normally
\ttt{KFL=} 5 or 6.
\end{subentry}
\iteme{ECM :} total c.m. energy of system.
\end{entry}
\subsubsection{Common block variables}
The status codes and parameters relevant for the $\e^+\e^-$ routines are
found in the common block \ttt{LUDAT1}. This common block also contains
more general status codes and parameters, described elsewhere.
\drawbox{COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)}
\begin{entry}
\itemc{Purpose:} to give access to a number of status codes and
parameters regulating the performance of the $\e^+\e^-$ event generation
routines.
\iteme{MSTJ(101) :}\label{p:MSTJ101} (D=5) gives the type of QCD
corrections used for continuum events.
\begin{subentry}
\iteme{= 0 :} only $\mathrm{q}\overline{\mathrm{q}}$ events are generated.
\iteme{= 1 :} $\mathrm{q}\overline{\mathrm{q}} + \mathrm{q}\overline{\mathrm{q}}\mathrm{g}$ events are generated according
to first-order QCD.
\iteme{= 2 :} $\mathrm{q}\overline{\mathrm{q}} + \mathrm{q}\overline{\mathrm{q}}\mathrm{g} + \mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g + \mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$
events are generated according to second-order QCD.
\iteme{= 3 :} $\mathrm{q}\overline{\mathrm{q}} + \mathrm{q}\overline{\mathrm{q}}\mathrm{g} + \mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g + \mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$
events are generated, but without second-order corrections to the
3-jet rate.
\iteme{= 5 :} a parton shower is allowed to develop from an original
$\mathrm{q}\overline{\mathrm{q}}$ pair, see \ttt{MSTJ(40) - MSTJ(50)} for details.
\iteme{= -1 :} only $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}$ events are generated (within same
matrix-element cuts as for \ttt{=1}). Since the change in flavour
composition from mass cuts or radiative corrections is not
taken into account, this option is not intended for
quantitative studies.
\iteme{= -2 :} only $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g$ and $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ events are
generated (as for \ttt{=2}). The same warning as for \ttt{=-1} applies.
\iteme{= -3 :} only $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g$ events are generated (as for
\ttt{=2}). The same warning as for \ttt{=-1} applies.
\iteme{= -4 :} only $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ events are generated
(as for \ttt{=2}). The same warning as for \ttt{=-1} applies.
\itemc{Note 1:} \ttt{MSTJ(101)} is also used in \ttt{LUONIA}, with
\iteme{$\leq$ 4 :} $\mathrm{g}\g\mathrm{g} + \gamma\mathrm{g}\g$ events are generated
according to lowest-order matrix elements.
\iteme{$\geq$ 5 :} a parton shower is allowed to develop from the
original $\mathrm{g}\g\mathrm{g}$ or $\mathrm{g}\g\gamma$ configuration, see
\ttt{MSTJ(40) - MSTJ(50)} for details.
\itemc{Note 2:} The default values of fragmentation parameters have
been chosen to work well with the default parton-shower approach
above. If any of the other options are used, or if the parton
shower is used in non-default mode, it may be necessary to
retune fragmentation parameters. As an example, we note that
the second-order matrix-element approach (\ttt{MSTJ(101)=2}) at
PETRA/PEP energies gives a better description when the $a$ and
$b$ parameters of the symmetric fragmentation function are set to
$a =$\ttt{PARJ(41)=1}, $b =$\ttt{PARJ(42)=0.7}, and the
width of the transverse momentum distribution to
$\sigma =$\ttt{PARJ(21)=0.40}.
In principle, one also ought to change the joining parameter
to \ttt{PARJ(33)=PARJ(35)=1.1} to preserve a flat rapidity
plateau, but if this should be forgotten, it does not make too
much difference. For applications at TRISTAN or LEP, one
must expect to have to change the matrix-element approach
parameters even more, to make up for additional soft gluon
effects not covered in this approach.
\end{subentry}
\iteme{MSTJ(102) :} (D=2) inclusion of weak effects ($\mathrm{Z}^0$ exchange)
for flavour production, angular orientation, cross sections and
initial-state photon radiation in continuum events.
\begin{subentry}
\iteme{= 1 :} QED, i.e. no weak effects are included.
\iteme{= 2 :} QFD, i.e. including weak effects.
\iteme{= 3 :} as \ttt{=2}, but at initialization in \ttt{LUXTOT} the
$\mathrm{Z}^0$ width is calculated from $\sin^2 \! \theta_W$, $\alpha_{\mathrm{em}}$ and
$\mathrm{Z}^0$ and quark masses (including bottom and top threshold factors for
\ttt{MSTJ(103)} odd), assuming three full generations, and the
result is stored in \ttt{PARJ(124)}.
\end{subentry}
\iteme{MSTJ(103) :} (D=7) mass effects in continuum matrix elements,
in the form \ttt{MSTJ(103)} $= M_1 + 2M_2 + 4M_3$, where $M_i = 0$
if no mass effects and $M_i = 1$ if mass effects should be included.
Here;
\begin{subentry}
\iteme{$M_1$ :} threshold factor for new flavour production
according to QFD result;
\iteme{$M_2$ :} gluon emission probability (only applies for
\ttt{|MSTJ(101)|}$\leq 1$, otherwise no mass effects anyhow);
\iteme{$M_3$ :} angular orientation of event (only applies for
\ttt{|MSTJ(101)|}$\leq 1$ and
\ttt{MSTJ(102)=1}, otherwise no mass effects anyhow).
\end{subentry}
\iteme{MSTJ(104) :} (D=5) number of allowed flavours, i.e. flavours
that can be produced in a continuum event if the energy is enough.
A change to 6 makes top production allowed above the threshold, etc.
Note that in $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ events only the first five flavours
are allowed in the secondary pair, produced by a gluon breakup.
\iteme{MSTJ(105) :} (D=1) fragmentation and decay in \ttt{LUEEVT} and
\ttt{LUONIA} calls.
\begin{subentry}
\iteme{= 0 :} no \ttt{LUEXEC} calls, i.e. only matrix-element
and/or parton-shower treatment.
\iteme{= 1 :} \ttt{LUEXEC} calls are made to generate fragmentation
and decay chain.
\iteme{= -1 :} no \ttt{LUEXEC} calls and no collapse of small jet
systems into one or two particles (in \ttt{LUPREP}).
\end{subentry}
\iteme{MSTJ(106) :} (D=1) angular orientation in \ttt{LUEEVT} and
\ttt{LUONIA}.
\begin{subentry}
\iteme{= 0 :} standard orientation of events, i.e. $\mathrm{q}$ along $+z$ axis
and $\overline{\mathrm{q}}$ along $-z$ axis or in $xz$ plane with $p_x > 0$ for
continuum events, and $\mathrm{g}_1\mathrm{g}_2\mathrm{g}_3$ or $\gamma\mathrm{g}_2\mathrm{g}_3$ in $xz$ plane
with $\mathrm{g}_1$ or $\gamma$ along the $+z$ axis for onium events.
\iteme{= 1 :} random orientation according to matrix elements.
\end{subentry}
\iteme{MSTJ(107) :} (D=0) radiative corrections to continuum events.
\begin{subentry}
\iteme{= 0 :} no radiative corrections.
\iteme{= 1 :} initial-state radiative corrections (including weak
effects for \ttt{MSTJ(102)=} 2 or 3).
\end{subentry}
\iteme{MSTJ(108) :} (D=2) calculation of $\alpha_{\mathrm{s}}$ for matrix-element
alternatives. The \ttt{MSTU(111)} and \ttt{PARU(112)} values are
automatically overwritten in \ttt{LUEEVT} or \ttt{LUONIA} calls
accordingly.
\begin{subentry}
\iteme{= 0 :} fixed $\alpha_{\mathrm{s}}$ value as given in \ttt{PARU(111)}.
\iteme{= 1 :} first-order formula is always used, with
$\Lambda_{\mrm{QCD}}$ given by \ttt{PARJ(121)}.
\iteme{= 2 :} first- or second-order formula is used, depending on
value of \ttt{MSTJ(101)}, with $\Lambda_{\mrm{QCD}}$ given by
\ttt{PARJ(121)} or \ttt{PARJ(122)}.
\end{subentry}
\iteme{MSTJ(109) :} (D=0) gives a possibility to switch from QCD
matrix elements to some alternative toy models. Is not relevant for
shower evolution, \ttt{MSTJ(101)=5}, where one can use
\ttt{MSTJ(49)} instead.
\begin{subentry}
\iteme{= 0 :} standard QCD scenario.
\iteme{= 1 :} a scalar gluon model. Since no second-order corrections
are available in this scenario, one can only use this with
\ttt{MSTJ(101) = 1} or \ttt{-1}. Also note that the event-as-a-whole
angular distribution is for photon exchange only (i.e. no weak
effects), and that no higher-order corrections to the total
cross section are included.
\iteme{= 2 :} an Abelian vector gluon theory, with the colour factors
$C_F = 1$ ($= 4/3$ in QCD), $N_C = 0$ ($= 3$ in QCD) and
$T_R = 3 n_f$ ($= n_f/2$ in QCD). If one selects
$\alpha_{\mrm{Abelian}} = (4/3) \alpha_{\mrm{QCD}}$,
the 3-jet cross section will agree with
the QCD one, and differences are to be found only in 4-jets.
The \ttt{MSTJ(109)=2} option has to be run with
\ttt{MSTJ(110)=1} and \ttt{MSTJ(111)=0}; if need be, the latter
variables will be overwritten by the program. \\
{\bf Warning:} second-order corrections give a large negative
contribution to the 3-jet cross section, so large that
the whole scenario is of doubtful use. In order to make the
second-order options work at all, the 3-jet cross section
is here by hand set exactly equal to zero for \ttt{MSTJ(101)=2}.
It is here probably better to use the option \ttt{MSTJ(101)=3},
although this is not a consistent procedure either.
\end{subentry}
\iteme{MSTJ(110) :} (D=2) choice of second-order contributions
to the 3-jet rate.
\begin{subentry}
\iteme{= 1 :} the GKS second-order matrix elements, i.e. the old
{\tsc{Jetset}} standard.
\iteme{= 2 :} the Zhu parametrization of the ERT matrix elements,
based on the program of Kunszt and Ali, i.e. in historical sequence
ERT/Kunszt/Ali/Zhu. The parametrization is available for
$y =$ 0.01, 0.02, 0.03, 0.04 and 0.05. Values outside this
range are put at the nearest border, while those inside
it are given by a linear interpolation between the
two nearest points. Since this procedure is rather primitive,
one should try to work at one of the values given above.
Note that no Abelian QCD parametrization is available for
this option.
\end{subentry}
\iteme{MSTJ(111) :} (D=0) use of optimized perturbation theory for
second-order matrix elements (it can also be used for first-order
matrix elements, but here it only corresponds to a trivial
rescaling of the $\alpha_{\mathrm{s}}$ argument).
\begin{subentry}
\iteme{= 0 :} no optimization procedure; i.e. $Q^2 = E_{\mrm{cm}}^2$.
\iteme{= 1 :} an optimized $Q^2$ scale is chosen as
$Q^2 = f E_{\mrm{cm}}^2$, where $f =$\ttt{PARJ(128)} for the total
cross section $R$ factor, while $f =$\ttt{PARJ(129)} for the
3- and 4-jet rates. This $f$ value enters via the
$\alpha_{\mathrm{s}}$, and also via a term proportional to $\alpha_{\mathrm{s}}^2 \ln f$.
Some constraints are imposed; thus the optimized `3-jet'
contribution to $R$ is assumed to be positive (for \ttt{PARJ(128)}),
the total 3-jet rate is not allowed to be negative
(for \ttt{PARJ(129)}), etc.
However, there is no guarantee that the differential 3-jet
cross section is not negative (and truncated to 0) somewhere
(this can also happen with $f = 1$, but is then less frequent).
The actually obtained $f$ values are stored in \ttt{PARJ(168)} and
\ttt{PARJ(169)}, respectively.
If an optimized $Q^2$ scale is used, then the $\Lambda_{\mrm{QCD}}$
(and $\alpha_{\mathrm{s}}$) should also be changed. With the value $f = 0.002$,
it has been shown \cite{Bet89} that a $\Lambda_{\mrm{QCD}} = 0.100$
GeV gives a reasonable agreement; the parameter to be changed is
\ttt{PARJ(122)} for a second-order running $\alpha_{\mathrm{s}}$. Note that,
since the optimized $Q^2$ scale is sometimes below the charm
threshold, the effective number of flavours used in $\alpha_{\mathrm{s}}$ may
well be 4 only. If one feels that it is still appropriate to use 5
flavours (one choice might be as good as the other), it is
necessary to put \ttt{MSTU(113)=5}.
\end{subentry}
\iteme{MSTJ(115) :} (D=1) documentation of continuum or onium
events, in increasing order of completeness.
\begin{subentry}
\iteme{= 0 :} only the parton shower, the fragmenting partons and the
generated hadronic system are stored in the \ttt{LUJETS} common block.
\iteme{= 1 :} also a radiative photon is stored (for continuum events).
\iteme{= 2 :} also the original $\e^+\e^-$ are stored (with
\ttt{K(I,1)=21}).
\iteme{= 3 :} also the $\gamma$ or $\gamma^* / \Z^0$ exchanged for continuum
events, the onium state for resonance events is stored (with
\ttt{K(I,1)=21}).
\end{subentry}
\iteme{MSTJ(116) :} (D=1) initialization of total cross section and
radiative photon spectrum in \ttt{LUEEVT} calls.
\begin{subentry}
\iteme{= 0 :} never; cannot be used together with radiative
corrections.
\iteme{= 1 :} calculated at first call and then whenever \ttt{KFL}
or \ttt{MSTJ(102)} is changed or \ttt{ECM} is changed by more than
\ttt{PARJ(139)}.
\iteme{= 2 :} calculated at each call.
\iteme{= 3 :} everything is reinitialized in the next call, but
\ttt{MSTJ(116)} is afterwards automatically put \ttt{=1} for use
in subsequent calls.
\end{subentry}
\iteme{MSTJ(119) :} (I) check on need to reinitialize \ttt{LUXTOT}.
\iteme{MSTJ(120) :} (R) type of continuum event generated with the
matrix-element option (with the shower one, the result is always
\ttt{=1}).
\begin{subentry}
\iteme{= 1 :} $\mathrm{q}\overline{\mathrm{q}}$.
\iteme{= 2 :} $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}$.
\iteme{= 3 :} $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g$ from Abelian (QED-like) graphs in
matrix element.
\iteme{= 4 :} $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g$ from non-Abelian (i.e. containing
triple-gluon coupling) graphs in matrix element.
\iteme{= 5 :} $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$.
\end{subentry}
\iteme{MSTJ(121) :} (R) flag set if a negative differential
cross section was encountered in the latest \ttt{LUX3JT} call.
Events are still generated, but maybe not quite according to
the distribution one would like (the rate is set to zero in the
regions of negative cross section, and the differential rate
in the regions of positive cross section is rescaled to give
the `correct' total 3-jet rate).
\boxsep
\iteme{PARJ(121) :}\label{p:PARJ121} (D=1.0 GeV) $\Lambda$ value
used in first-order
calculation of $\alpha_{\mathrm{s}}$ in the matrix-element alternative.
\iteme{PARJ(122) :} (D=0.25 GeV) $\Lambda$ values used in second-order
calculation of $\alpha_{\mathrm{s}}$ in the matrix-element alternative.
\iteme{PARJ(123) :} (D=91.187 GeV) mass of $\mathrm{Z}^0$ as used in
propagators for the QFD case.
\iteme{PARJ(124) :} (D=2.489 GeV) width of $\mathrm{Z}^0$ as used in
propagators for the QFD case. Overwritten at initialization if
\ttt{MSTJ(102)=3}.
\iteme{PARJ(125) :} (D=0.01) $y_{\mrm{cut}}$, minimum squared scaled
invariant mass of any two partons in 3- or 4-jet events; the main
user-controlled matrix-element cut. \ttt{PARJ(126)} provides an
additional constraint. For each new event, it is additionally
checked that the total 3- plus 4-jet fraction does not
exceed unity; if so the effective $y$ cut will be dynamically
increased. The actual $y$-cut value is stored in
\ttt{PARJ(150)}, event by event.
\iteme{PARJ(126) :} (D=2. GeV) minimum invariant mass of any two
partons in 3- or 4-jet events; a cut in addition to the one above,
mainly for the case of a radiative photon lowering the hadronic c.m.
energy significantly.
\iteme{PARJ(127) :} (D=1. GeV) is used as a safety margin for small
colour-singlet jet systems, cf. \ttt{PARJ(32)}, specifically
$\mathrm{q}\overline{\mathrm{q}}'$ masses in $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ 4-jet events and $\mathrm{g}\g$ mass
in onium $\gamma\mathrm{g}\g$ events.
\iteme{PARJ(128) :} (D=0.25) optimized $Q^2$ scale for the QCD $R$
(total rate) factor for the \ttt{MSTJ(111)=1} option is given by
$Q^2 = f E_{\mrm{cm}}^2$, where $f =$\ttt{PARJ(128)}. For various
reasons the actually used $f$ value may be increased compared with
the nominal one; while \ttt{PARJ(128)} gives the nominal value,
\ttt{PARJ(168)} gives the actual one for the current event.
\iteme{PARJ(129) :} (D=0.002) optimized $Q^2$ scale for the 3-
and 4-jet rate for the \ttt{MSTJ(111)=1} option is given by
$Q^2 = f E_{\mrm{cm}}^2$, where $f =$\ttt{PARJ(129)}. For various
reasons the actually used $f$ value may be increased compared with
the nominal one; while \ttt{PARJ(129)} gives the nominal value,
\ttt{PARJ(169)} gives the actual one for the current event. The
default value is in agreement with the studies of Bethke \cite{Bet89}.
\iteme{PARJ(131), PARJ(132) :} (D=2*0.) longitudinal polarizations
$P_{\mrm{L}}^+$ and $P_{\mrm{L}}^-$ of incoming $\mathrm{e}^+$ and $\mathrm{e}^-$.
\iteme{PARJ(133) :} (D=0.) transverse polarization
$P_{\mrm{T}} = \sqrt{P_{\mrm{T}}^+ P_{\mrm{T}}^-}$, with
$P_{\mrm{T}}^+$ and $P_{\mrm{T}}^-$ transverse
polarizations of incoming $\mathrm{e}^+$ and $\mathrm{e}^-$.
\iteme{PARJ(134) :} (D=0.) mean of transverse polarization
directions of incoming $\mathrm{e}^+$ and $\mathrm{e}^-$,
$\Delta \varphi = (\varphi^+ + \varphi^-) /2$, with $\varphi$
the azimuthal angle of polarization, leading to a shift in the
$\varphi$ distribution of jets by $\Delta \varphi$.
\iteme{PARJ(135) :} (D=0.01) minimum photon energy fraction
(of beam energy) in initial-state radiation; should normally
never be changed (if lowered too much, the fraction of events
containing a radiative photon will exceed unity, leading to
problems).
\iteme{PARJ(136) :} (D=0.99) maximum photon energy fraction
(of beam energy) in initial-state radiation; may be changed
to reflect actual trigger conditions of a detector (but must
always be larger than \ttt{PARJ(135)}).
\iteme{PARJ(139) :} (D=0.2 GeV) maximum deviation of $E_{\mrm{cm}}$
from the corresponding value at last \ttt{LUXTOT} call, above which
a new call is made if \ttt{MSTJ(116)=1}.
\iteme{PARJ(141) :} (R) value of $R$, the ratio of continuum
cross section to the lowest-order muon pair production cross section,
as given in massless QED (i.e. three times the sum of active
quark squared charges, possibly modified for polarization).
\iteme{PARJ(142) :} (R) value of $R$ including quark-mass effects
(for \ttt{MSTJ(102)=1}) and/or weak propagator effects
(for \ttt{MSTJ(102)=2}).
\iteme{PARJ(143) :} (R) value of $R$ as \ttt{PARJ(142)}, but
including QCD corrections as given by \ttt{MSTJ(101)}.
\iteme{PARJ(144) :} (R) value of $R$ as \ttt{PARJ(143)}, but
additionally including corrections from initial-state photon
radiation (if \ttt{MSTJ(107)=1}). Since the effects of heavy
flavour thresholds are not simply integrable, the initial value
of \ttt{PARJ(144)} is updated during the
course of the run to improve accuracy.
\iteme{PARJ(145) - PARJ(148) :} (R) absolute cross sections in nb
as for the cases \ttt{PARJ(141) - PARJ(144)} above.
\iteme{PARJ(150) :} (R) current effective matrix element cut-off
$y_{\mrm{cut}}$, as given by \ttt{PARJ(125), PARJ(126)} and the
requirements of having non-negative cross sections for 2-,
3- and 4-jet events. Not used in parton showers.
\iteme{PARJ(151) :} (R) value of c.m. energy \ttt{ECM} at last
\ttt{LUXTOT} call.
\iteme{PARJ(152) :} (R) current first-order contribution to the
3-jet fraction; modified by mass effects. Not used in parton
showers.
\iteme{PARJ(153) :} (R) current second-order contribution to the
3-jet fraction; modified by mass effects. Not used in parton
showers.
\iteme{PARJ(154) :} (R) current second-order contribution to the
4-jet fraction; modified by mass effects. Not used in parton
showers.
\iteme{PARJ(155) :} (R) current fraction of 4-jet rate
attributable to $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ events rather than $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g$
ones; modified by mass effects. Not used in parton showers.
\iteme{PARJ(156) :} (R) has two functions when using second-order
QCD. For a 3-jet event, it gives the ratio of the second-order
to the total 3-jet cross section in the given kinematical
point. For a 4-jet event, it gives the ratio of the
modified 4-jet cross section, obtained when neglecting interference
terms whose colour flow is not well defined, to the full
unmodified one, all evaluated in the given kinematical point.
Not used in parton showers.
\iteme{PARJ(157) - PARJ(159) :} (I) used for cross-section
calculations to include mass threshold effects to radiative
photon cross section. What is stored is basic cross section,
number of events generated and number that passed cuts.
\iteme{PARJ(160) :} (R) nominal fraction of events that should
contain a radiative photon.
\iteme{PARJ(161) - PARJ(164) :} (I) give shape of radiative photon
spectrum including weak effects.
\iteme{PARJ(168) :} (R) actual $f$ value of current event in
optimized perturbation theory for $R$; see \ttt{MSTJ(111)} and
\ttt{PARJ(128)}.
\iteme{PARJ(169) :} (R) actual $f$ value of current event in
optimized perturbation theory for 3- and 4-jet rate;
see \ttt{MSTJ(111)} and \ttt{PARJ(129)}.
\iteme{PARJ(171) :} (R) fraction of cross section corresponding
to the axial coupling of quark pair to the intermediate $\gamma^* / \Z^0$
state; needed for the Abelian gluon model 3-jet matrix
element.
\end{entry}
\subsection{Examples}
An ordinary $\e^+\e^-$ annihilation event in the continuum, at a c.m.
energy of 40 GeV, may be generated with
\begin{verbatim}
CALL LUEEVT(0,40.)
\end{verbatim}
In this case a $\mathrm{q}\overline{\mathrm{q}}$ event is generated, including weak effects,
followed by parton-shower evolution and fragmentation/decay treatment.
Before a call to \ttt{LUEEVT}, however, a number of default values
may be changed, e.g. \ttt{MSTJ(101)=2} to use second-order QCD
matrix elements, giving a mixture of $\mathrm{q}\overline{\mathrm{q}}$, $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}$,
$\mathrm{q}\overline{\mathrm{q}}\mathrm{g}\g$, and $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$ events, \ttt{MSTJ(102)=1} to
have QED only, \ttt{MSTJ(104)=6} to allow $\t\overline{\mathrm{t}}$ production
as well, \ttt{MSTJ(107)=1} to include initial-state photon radiation
(including a treatment of the $\mathrm{Z}^0$ pole), \ttt{PARJ(123)=92.0} to
change the $\mathrm{Z}^0$ mass, \ttt{PARJ(81)=0.3} to change the
parton-shower $\Lambda$ value, or \ttt{PARJ(82)=1.5} to change the
parton-shower cut-off. If initial-state photon radiation is used, some
restrictions apply to how one can alternate the generation of
events at different energies or with different $\mathrm{Z}^0$ mass, etc.
These restrictions are not there for
efficiency reasons (the extra time for recalculating the extra
constants every time is small), but because it ties in with the
cross-section calculations (see \ttt{PARJ(144)}).
Most parameters can be changed independently of each other. However,
if just one or a few parameters/switches are changed, one should not
be surprised to find a rather bad agreement with the data, like e.g.
a too low or high average hadron multiplicity. It is therefore usually
necessary to retune one parameter related to the perturbative QCD
description, like $\alpha_{\mathrm{s}}$ or $\Lambda$, one of the two parameters
$a$ and $b$ of the Lund symmetric fragmentation function (since they
are so strongly correlated, it is often not necessary to retune both
of them), and the average fragmentation transverse momentum --- see
Note~2 of the \ttt{MSTJ(101)} description for an example. For very
detailed studies it may be necessary to retune even more parameters.
The three-gluon and gluon--gluon--photon decays of $\Upsilon$ may be
simulated by a call
\begin{verbatim}
CALL LUONIA(5,9.46)
\end{verbatim}
Unfortunately, with present top-mass limits, this routine will not be
of much interest for toponium studies (weak decays will dominate).
A typical program for analysis of $\e^+\e^-$ annihilation events at
100 GeV might look something like
\begin{verbatim}
COMMON/LUJETS/N,K(4000,5),P(4000,5),V(4000,5)
COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)
COMMON/LUDAT2/KCHG(500,3),PMAS(500,4),PARF(2000),VCKM(4,4)
COMMON/LUDAT3/MDCY(500,3),MDME(2000,2),BRAT(2000),KFDP(2000,5)
MDCY(LUCOMP(111),1)=0 ! put pi0 stable
MSTJ(107)=1 ! include initial-state radiation
PARU(41)=1. ! use linear sphericity
..... ! other desired changes
..... ! initialize analysis statistics
DO 100 IEVENT=1,1000 ! loop over events
CALL LUEEVT(0,100.) ! generate new event
IF(IEVENT.EQ.1) CALL LULIST(2) ! list first event
CALL LUTABU(11) ! save particle composition
! statistics
CALL LUEDIT(2) ! remove decayed particles
CALL LUSPHE(SPH,APL) ! linear sphericity analysis
IF(SPH.LT.0.) GOTO 100 ! too few particles in event for
! LUSPHE to work on it (unusual)
CALL LUEDIT(31) ! orient event along axes above
IF(IEVENT.EQ.1) CALL LULIST(2) ! list first treated event
..... ! fill analysis statistics
CALL LUTHRU(THR,OBL) ! now do thrust analysis
..... ! more analysis statistics
100 CONTINUE !
CALL LUTABU(12) ! print particle composition
! statistics
..... ! print analysis statistics
END
\end{verbatim}
\clearpage
\section{Process Generation in PYTHIA}
\label{s:PYTprocgen}
Much can be said about the hard processes in {\tsc{Pythia}} and
the way they are generated. Therefore the material has been split
into three sections. In the current one the philo\-sophy underlying
the event generation scheme is presented. Here we provide a
generic description, where some special cases are swept under the
carpet. In the next section, the existing processes are enumerated,
with some comments about applications and limitations. Finally, in
the third section the generation routines and common block switches
are described.
The section starts with a survey of parton distributions, followed
by a detailed description of the simple $2 \to 2$ and $2 \to 1$
hard subprocess generation schemes, including pairs of resonances.
This is followed by a few comments on more complicated configurations.
\subsection{Parton Distributions}
\label{ss:structfun}
The parton-distribution function $f_i^a(x,Q^2)$ parametrizes the
probability to find a parton $i$ with a fraction $x$ of the beam energy
when the beam particle $a$ is probed by a hard scattering at virtuality
scale $Q^2$. Usually the momentum-weighted combination $x f_i^a(x,Q^2)$
is used, for which the normalization condition
$\sum_i \int_0^1 dx \, x f_i^a(x,Q^2) \equiv 1$ normally applies.
The $Q^2$ dependence of parton distributions is perturbatively
calculable, see section \ref{sss:initshowstruc}.
The parton distributions in {\tsc{Pythia}} come in many shapes, as shown in the
following.
\subsubsection{Baryons}
For protons, many sets exist on the market. These are obtained by fits
to experimental data, constrained so that the $Q^2$ dependence is in
accordance with the standard QCD evolution equations. The (new) default
in \tsc{Pythia} is CTEQ2L \cite{Bot93}, a modern leading-order fit.
Ten other sets are found in \tsc{Pythia}. The complete list is:
\begin{Itemize}
\item EHLQ sets 1 and 2 \cite{Eic84};
\item DO sets 1 and 2 \cite{Duk82};
\item the other CTEQ2 fits, namely CTEQ2M, CTEQ2MS, CTEQ2MF, CTEQ2ML,
and CTEQ2D \cite{Bot93}; and
\item the dynamically generated fit GRV LO (updated version)
\cite{Glu92}.
\end{Itemize}
Of these, EHLQ, DO, CTEQ2L and GRV LO are leading-order parton
distributions,
while CTEQ2D are in the next-to-leading-order DIS scheme and the rest
in the next-to-leading order $\br{\mrm{MS}}$ scheme. The EHLQ and DO
sets are by now rather old, and are kept mainly for backwards
compatibility. Since only Born-level matrix elements
are included in the program, there is no particular reason to use
higher-order parton distributions --- the resulting combination is
anyway only good to leading-order accuracy. (Some higher-order
corrections are effectively included by the parton-shower
treatment, but there is no exact match.)
There is a steady flow of new parton-distribution sets on the market.
To keep track of all of them is a major work on its own. Therefore
{\tsc{Pythia}} contains an interface to an external library of parton
distribution functions, \tsc{Pdflib} \cite{Plo93}.
This is a truly encyclopedic collection of almost all proton, pion
and photon parton distributions proposed since the late 70's.
Two dummy routines come with the {\tsc{Pythia}} package, so as to avoid
problems with unresolved external references if \tsc{Pdflib} is not
linked. One should also note that {\tsc{Pythia}} does not check the results,
but assumes that sensible answers will be returned, also outside the
nominal $(x, Q^2)$ range of a set. Only the sets that come with
{\tsc{Pythia}} have been suitably modified to provide reasonable answers
outside their nominal domain of validity.
From the proton parton distributions, those of the neutron are obtained
by isospin conjugation, i.e. $f_{\u}^{\mathrm{n}} = f_{\d}^{\mathrm{p}}$ and
$f_{\d}^{\mathrm{n}} = f_{\u}^{\mathrm{p}}$.
The program does allow for incoming beams of a number of hyperons:
$\Lambda^0$, $\Sigma^{-,0,+}$, $\Xi^{-,0}$ and $\Omega^-$. Here
one has essentially no experimental information. One could imagine
to construct models in which valence $\mathrm{s}$ quarks are found at larger
average $x$ values than valence $\u$ and $\d$ ones, because of the
larger $\mathrm{s}$-quark mass. However, hyperon beams is a little-used part
of the program, included only for a few specific studies. Therefore
a simple approach has been taken, in which an average valence
quark distribution is constructed as
$f_{\mrm{val}} = (f_{\u,\mrm{val}}^{\mathrm{p}} + f_{\d,\mrm{val}}^{\mathrm{p}})/3$,
according to
which each valence quark in a hyperon is assumed to be distributed.
Sea-quark and gluon distributions are taken as in the proton.
Any proton parton distribution set may be used with this procedure.
\subsubsection{Mesons and photons}
Data on meson parton distributions are scarce, so only very few sets
have been constructed, and only for the $\pi^{\pm}$. {\tsc{Pythia}} contains
the Owens set 1 and 2 parton distributions \cite{Owe84}, which for a
long time were essentially the only sets on the market, and the
more recent dynamically generated GRV LO (updated version)
\cite{Glu92a}. The first one is the default in {\tsc{Pythia}}. Further sets are
found in \tsc{Pdflib} and can therefore be used by {\tsc{Pythia}}, just as
described above for protons.
Sets of photon parton distributions have been obtained as for
hadrons; an additional complication comes from the necessity to
handle the matching of the vector meson dominance (VMD) and the
perturbative pieces in a consistent manner. New sets have been
produced where this division is explicit and therefore especially
well suited for applications to event generation\cite{Sch95}.
The Schuler and Sj\"ostand set 1D is the default. Although the
vector-meson philosophy is at the base, the details of the fits do
not rely on pion data, but only on $F_2^{\gamma}$ data. Here
follows a brief summary of relevant details.
Photons obey a set of inhomogeneous evolution equations, where the
inhomogeneous term is induced by $\gamma \to \mathrm{q}\overline{\mathrm{q}}$ branchings.
The solution can be written as the sum of two terms,
\begin{equation}
f_a^{\gamma}(x,Q^2) = f_a^{\gamma,\mrm{NP}}(x,Q^2;Q_0^2)
+ f_a^{\gamma,\mrm{PT}}(x,Q^2;Q_0^2) ~,
\end{equation}
where the former term is a solution to the homogeneous evolution
with a (non-perturbative) input at $Q=Q_0$ and the latter is a
solution to the full inhomogeneous equation with boundary condition
$f_a^{\gamma,\mrm{PT}}(x,Q_0^2;Q_0^2) \equiv 0$. One possible
physics interpretation is to let $f_a^{\gamma,\mrm{NP}}$ correspond
to $\gamma \leftrightarrow V$ fluctuations, where
$V = \rho^0, \omega, \phi, \ldots$ is a set of vector mesons,
and let $f_a^{\gamma,\mrm{PT}}$ correspond to perturbative (`anomalous')
$\gamma \leftrightarrow \mathrm{q}\overline{\mathrm{q}}$ fluctuations. The discrete spectum
of vector mesons can be combined with the continuous (in virtuality
$k^2$) spectrum of $\mathrm{q}\overline{\mathrm{q}}$ fluctuations, to give
\begin{equation}
f_a^{\gamma}(x,Q^2) =
\sum_V \frac{4\pi\alpha_{\mathrm{em}}}{f_V^2} f_a^{\gamma,V}(x,Q^2)
+ \frac{\alpha_{\mathrm{em}}}{2\pi} \, \sum_{\mathrm{q}} 2 e_{\mathrm{q}}^2 \,
\int_{Q_0^2}^{Q^2} \frac{{\d} k^2}{k^2} \,
f_a^{\gamma,\mathrm{q}\overline{\mathrm{q}}}(x,Q^2;k^2) ~,
\end{equation}
where each component $f^{\gamma,V}$ and $f^{\gamma,\mathrm{q}\overline{\mathrm{q}}}$ obeys a
unit momentum sum rule.
In sets 1 the $Q_0$ scale is picked at a low value, 0.6~GeV, where
an identification of the non-perturbative component with a set of
low-lying mesons appear natural, while sets 2 use a higher value,
2~GeV, where the validity of perturbation theory is better established.
The data are not good enough to allow a precise determination of
$\Lambda_{\mrm{QCD}}$. Therefore we use a fixed value
$\Lambda^{(4)} = 200$~MeV, in agreement with conventional
results for proton distributions. In the VMD component the $\rho^0$
and $\omega$ have been added coherently, so that
$\u\overline{\mathrm{u}} : \d\overline{\mathrm{d}} = 4 : 1$ at $Q_0$.
Unlike the $\mathrm{p}$, the $\gamma$ has a direct component where the photon
acts as an unresolved probe. In the definition of $F_2^{\gamma}$ this
adds a component $C^{\gamma}$, symbolically
\begin{equation}
F_2^{\gamma}(x,Q^2) = \sum_{\mathrm{q}} e_{\mathrm{q}}^2 \left[ f_{\mathrm{q}}^{\gamma} +
f_{\overline{\mathrm{q}}}^{\gamma} \right] \otimes C_{\mathrm{q}} +
f_{\mathrm{g}}^{\gamma} \otimes C_{\mathrm{g}} + C^{\gamma} ~.
\end{equation}
Since $C^{\gamma} \equiv 0$ in leading order, and since we stay with
leading-order fits, it is permissible to neglect this complication.
Numerically, however, it makes a non-negligible difference. We
therefore make two kinds of fits, one DIS type with $C^{\gamma} = 0$
and one {$\overline{\mrm{MS}}$} type including the universal part
of $C^{\gamma}$.
When jet production is studied for real incoming photons, the standard
evolution approach is reasonable also for heavy flavours, i.e.
predominantly the $\c$, but with a lower cut-off $Q_0 \approx m_{\c}$
for $\gamma \to \c\overline{\mathrm{c}}$. Moving to deep inelastic scattering,
$\mathrm{e}\gamma \to \mathrm{e} X$, there is an extra kinematical constraint:
$W^2 = Q^2 (1-x)/x > 4 m_{\c}^2$. It is here better to use the
`Bethe-Heitler' cross section for $\gamma^* \gamma \to \c\overline{\mathrm{c}}$.
Therefore each distribution appears in two variants. For applications
to real $\gamma$'s the parton distributions are calculated as the sum
of a vector-meson part and an anomalous part including all five flavours.
For applications to DIS, the sum runs over the same vector-meson part,
an anomalous part and possibly a $C^{\gamma}$ part for the three
light flavours, and a Bethe-Heitler part for $\c$ and $\b$.
In addition to the SaS sets, {\tsc{Pythia}} also contains the Drees--Grassie set
of parton distributions \cite{Dre85} and, as for the proton, there is
an interface to the \tsc{Pdflib} library \cite{Plo93}.
However, these sets do not allow a subdivision of the photon parton
distributions into one VMD part and one anomalous part. This
subdivision is necessary a sophisticated modelling of $\gamma\mathrm{p}$ and
$\gamma\gamma$ events, see above and section \ref{sss:photoprod}.
As an alternative, for the VMD part alone, the
$\rho^0$ parton distribution can be found from the assumed equality
\begin{equation}
f^{\rho^0}_i = f^{\pi^0}_i = \frac{1}{2} \, (f^{\pi^+}_i +
f^{\pi^-}_i) ~.
\end{equation}
Thus any $\pi^+$ parton distribution set, from any library, can be
turned into a VMD $\rho^0$ set. The $\omega$ parton distribution is
assumed the same, while the $\phi$ and $\mathrm{J}/\psi$ ones are handled
in the very crude approximation
$f^{\phi}_{\mathrm{s},\mrm{val}} = f^{\pi^+}_{\u,\mrm{val}}$ and
$f^{\phi}_{\mrm{sea}} = f^{\pi^+}_{\mrm{sea}}$.
therefore is default. The VMD part needs to be complemented by an
anomalous part to make upp a full photon distribution. The latter
is fully perturbatively calculable, given the lower cut-off scale
$Q_0$. The SaS parametrization of the anomalous part is therefore used
throughout for this purpose. The $Q_0$ scale can be set freely in the
\ttt{PARP(15)} parameter.
The $f_i^{\gamma,\mrm{anom}}$ distribution can be further decomposed,
by the flavour and the $p_{\perp}$ of the original branching
$\gamma \to \mathrm{q}\overline{\mathrm{q}}$. The flavour is distributed according to squared
charge (plus flavour thresholds for heavy flavours) and the $p_{\perp}$
according to $\d p_{\perp}^2 / p_{\perp}^2$ in the range $Q_0 < p_{\perp} < Q$.
At the branching scale, the photon only consists of a $\mathrm{q}\overline{\mathrm{q}}$ pair,
with $x$ distribution $\propto x^2 + (1-x)^2$. A component
$f_a^{\gamma,\mathrm{q}\overline{\mathrm{q}}}(x,Q^2;k^2)$, characterized by its $k \approx p_{\perp}$
and flavour, then is evolved homogeneously from $p_{\perp}$ to $Q$. For
theoretical studies it is convenient to be able to access a specific
component of this kind. Therefore also leading-order parametrizations
of these decomposed distributions are available \cite{Sch95}.
\subsubsection{Leptons}
\label{sss:estructfun}
Contrary to the hadron case, there is no necessity to introduce the
parton-distribution function concept for leptons. A lepton can be
considered as a point-like particle, with initial-state radiation
handled by higher-order matrix elements. However, the parton
distribution function approach offers a slightly simplified but very
economical description of initial-state radiation effects for any hard
process, also those for which higher-order corrections are not yet
calculated.
Parton distributions for electrons have been introduced in {\tsc{Pythia}},
but not yet for muons, i.e. currently
$f_{\mu}^{\mu}(x, Q^2) = \delta (x-1)$. Also for the electron one
is free to use a simple `unresolved' $\mathrm{e}$,
$f_{\mathrm{e}}^{\mathrm{e}}(x, Q^2) = \delta(x-1)$, where the $\mathrm{e}$ retains the
full original momentum.
Electron parton distributions are calculable entirely from first
principles, but different levels of approximation may be used.
The parton-distribution formulae in {\tsc{Pythia}} are based on a
next-to-leading-order exponentiated description, see ref.
\cite{Kle89}, p. 34. The approximate behaviour is
\begin{eqnarray}
& & f_{\mathrm{e}}^{\mathrm{e}}(x,Q^2) \approx \frac{\beta}{2}
(1-x)^{\beta/2-1} ~; \nonumber \\
& & \beta = \frac{2 \alpha_{\mathrm{em}}}{\pi}
\left( \ln \frac{Q^2}{m_{\mathrm{e}}^2} -1 \right) ~.
\end{eqnarray}
The form is divergent but integrable for $x \to 1$, i.e. the electron
likes to keep most of the energy. To handle the numerical precision
problems for $x$ very close to unity, the parton distribution is set, by
hand, to zero for $x > 0.999999$, and is rescaled upwards in the range
$0.9999 < x < 0.999999$, in such a way that the total area under the
parton distribution is preserved:
\begin{equation}
\left( f_{\mathrm{e}}^{\mathrm{e}}(x,Q^2) \right)_{\mrm{mod}} =
\left\{ \begin{array}{ll}
f_{\mathrm{e}}^{\mathrm{e}}(x,Q^2) & 0 \leq x \leq 0.9999 \\[2mm]
\frac{\displaystyle 100^{\beta/2}}{\displaystyle 100^{\beta/2}-1}
\, f_{\mathrm{e}}^{\mathrm{e}}(x,Q^2) & 0.9999 < x \leq 0.999999 \\[4mm]
0 & x > 0.999999 \, ~.
\end{array} \right.
\end{equation}
The branchings $\mathrm{e} \to \mathrm{e} \gamma$, which are responsible for the
softening of the $f_{\mathrm{e}}^{\mathrm{e}}$ parton distribution, also gives rise to
a flow of photons. In photon-induced hard processes, the
$f_{\gamma}^{\mathrm{e}}$ parton distribution can be used to describe the
equivalent flow of photons. The formula used in the program is the
simple first-order expression. There is some ambiguity in
the choice of $Q^2$ range over which emissions should be included.
The na\"{\i}ve (default) choice is
\begin{equation}
f_{\gamma}^{\mathrm{e}}(x,Q^2) = \frac{\alpha_{\mathrm{em}}}{2 \pi} \,
\frac{1+(1-x)^2}{x} \, \ln \left( \frac{Q^2}{m_{\mathrm{e}}^2} \right) ~.
\end{equation}
Here it is assumed that only one scale enters the problem, namely
that of the hard interaction, and that the scale of the branching
$\mathrm{e} \to \mathrm{e} \gamma$ is bounded from above by the hard interaction scale.
For a pure QCD or pure QED shower this is an appropriate procedure,
cf. section \ref{sss:showermatching}, but in other cases it may not be
optimal. In particular, for photoproduction the alternative that is
probably most appropriate is \cite{Ali88}:
\begin{equation}
f_{\gamma}^{\mathrm{e}}(x,Q^2) =
\frac{\alpha_{\mathrm{em}}}{2 \pi} \, \frac{1+(1-x)^2}{x}
\, \ln \left( \frac{Q^2_{\mathrm{max}} (1-x)}{m_{\mathrm{e}}^2 \, x^2} \right) ~.
\end{equation}
Here $Q^2_{\mathrm{max}}$ is a user-defined cut for the range of
scattered electron kinematics that is counted as
photoproduction. Note that we now deal with two different $Q^2$
scales, one related to the hard subprocess itself, which appears
as the argument of the parton distribution, and the other related to
the scattering of the electron, which is reflected in
$Q^2_{\mathrm{max}}$.
In resolved photoproduction or resolved $\gamma\gamma$
interactions, one has to include the parton distributions for quarks
and gluons inside the photon inside the electron. There are no
published sets where results are directly presented in terms of
quark and gluon distributions inside the electron. In the
program, the $f_{\mathrm{q},\mathrm{g}}^{\mathrm{e}}$ are therefore obtained by a numerical
convolution according to
\begin{equation}
f_{\mathrm{q},\mathrm{g}}^{\mathrm{e}}(x, Q^2) =
\int_x^1 \frac{dx_{\gamma}}{x_{\gamma}} \,
f_{\gamma}^{\mathrm{e}}(x_{\gamma}, Q^2) \, f^{\gamma}_{\mathrm{q},\mathrm{g}} \!
\left( \frac{x}{x_{\gamma}}, Q^2 \right) ~,
\label{pg:foldqgine}
\end{equation}
with $f^{\mathrm{e}}_{\gamma}$ as discussed above. The necessity for numerical
convolution makes this parton distribution evaluation rather slow
compared with the others; one should therefore only have it
switched on for resolved photoproduction studies.
One can obtain the positron distribution inside an electron,
which is also the electron sea parton distribution, by a convolution
of the two branchings $\mathrm{e} \to \mathrm{e} \gamma$ and $\gamma \to \e^+\e^-$;
the result is \cite{Che75}
\begin{equation}
f_{\mathrm{e}^+}^{\mathrm{e}^-}(x,Q^2) =
\frac{1}{2} \, \left\{ \frac{\alpha_{\mathrm{em}}}{2 \pi} \,
\left( \ln \frac{Q^2}{m_{\mathrm{e}}^2} -1 \right) \right\}^2 \,
\frac{1}{x} \, \left( \frac{4}{3} - x^2 - \frac{4}{3} x^3 +
2x(1+x) \ln x \right) ~.
\end{equation}
Finally, the program also contains the distribution of a
transverse $\mathrm{W}^-$ inside an electron
\begin{equation}
f_{\mathrm{W}}^{\mathrm{e}}(x,Q^2) = \frac{\alpha_{\mathrm{em}}}{2 \pi} \,
\frac{1}{4 \sin^2 \! \theta_W} \, \frac{1+(1-x)^2}{x} \,
\ln \left( 1 + \frac{Q^2}{m_{\mathrm{W}}^2} \right) ~.
\end{equation}
\subsection{Kinematics and Cross section for a $2 \to 2$ Process}
\label{ss:kinemtwo}
In this section we begin the description of kinematics selection
and cross-section calculation. The example is for the case of a
$2 \to 2$ process, with final-state masses assumed to be vanishing.
Later on we will expand to finite fixed masses, and to resonances.
Consider two incoming beam particles in their c.m. frame, each with
energy $E_{\mrm{beam}}$. The total squared c.m. energy is then
$s = 4 E_{\mrm{beam}}^2$. The two partons that enter the hard
interaction do not carry the total beam momentum, but only fractions
$x_1$ and $x_2$, respectively, i.e. they have four-momenta
\begin{eqnarray}
p_1 & = & E_{\mrm{beam}}(x_1; 0, 0, x_1) ~, \nonumber \\
p_2 & = & E_{\mrm{beam}}(x_2; 0, 0, -x_2) ~.
\end{eqnarray}
There is no reason to put the incoming partons on the mass shell,
i.e. to have time-like incoming four-vectors,
since partons inside a particle are always virtual and thus
space-like. These space-like virtualities are introduced as
part of the initial-state parton-shower description, see section
\ref{sss:initshowtrans}, but do not affect the formalism of this
section. The one example where it would be appropriate to put a
parton on the mass shell is for an incoming lepton beam, but even
here the massless kinematics description is adequate as long as
the c.m. energy is correctly calculated with masses.
The squared invariant mass of the two partons is defined as
\begin{equation}
\hat{s} = (p_1 + p_2)^2 = x_1 \, x_2 \, s ~.
\end{equation}
Instead of $x_1$ and $x_2$, it is often customary to use $\tau$
and either $y$ or $x_{\mrm{F}}$:
\begin{eqnarray}
\tau & = & x_1 x_2 = \frac{\hat{s}}{s} ~; \\
y & = & \frac{1}{2} \ln \frac{x_1}{x_2} ~; \\
x_{\mrm{F}} & = & x_1 - x_2 ~.
\end{eqnarray}
In addition to $x_1$ and $x_2$, two additional variables are needed
to describe the kinematics of a scattering $1 + 2 \to 3 + 4$.
One corresponds to the azimuthal angle $\varphi$ of the scattering
plane around the beam axis. This angle is always isotropically
distributed for unpolarized incoming beam particles, and so need not
be considered further. The other variable can be picked as
$\hat{\theta}$, the polar angle of parton 3 in the c.m. frame of the
hard scattering. The conventional choice is to use the variable
\begin{equation}
\hat{t} = (p_1 - p_3)^2 = (p_2 - p_4)^2 =
- \frac{\hat{s}}{2} (1 - \cos \hat{\theta}) ~,
\end{equation}
with $\hat{\theta}$ defined as above. In the following, we will make
use of both $\hat{t}$ and $\hat{\theta}$. It is also customary
to define $\hat{u}$,
\begin{equation}
\hat{u} = (p_1 - p_4)^2 = (p_2 - p_3)^2 =
- \frac{\hat{s}}{2} (1 + \cos \hat{\theta}) ~,
\end{equation}
but $\hat{u}$ is not an independent variable since
\begin{equation}
\hat{s} + \hat{t} + \hat{u} = 0 ~.
\end{equation}
If the two outgoing particles have masses $m_3$ and $m_4$,
respectively, then the four-momenta in the c.m. frame of the hard
interaction are given by
\begin{equation}
\hat{p}_{3,4} = \left(
\frac{\hat{s} \pm (m_3^2-m_4^2)}{2\sqrt{\hat{s}}} ,
\pm \frac{\sqrt{\hat{s}}}{2} \, \beta_{34} \sin \hat{\theta}, 0,
\pm \frac{\sqrt{\hat{s}}}{2} \, \beta_{34} \cos \hat{\theta}
\right) ~,
\end{equation}
where
\begin{equation}
\beta_{34} = \sqrt{ \left( 1 - \frac{m_3^2}{\hat{s}} -
\frac{m_4^2}{\hat{s}} \right)^2 - 4 \, \frac{m_3^2}{\hat{s}} \,
\frac{m_4^2}{\hat{s}} } ~.
\end{equation}
Then $\hat{t}$ and $\hat{u}$ are modified to
\begin{equation}
\hat{t}, \hat{u} = - \frac{1}{2} \left\{ ( \hat{s} - m_3^2 - m_4^2 )
\mp \hat{s} \, \beta_{34} \cos \hat{\theta} \right\} ~,
\end{equation}
with
\begin{equation}
\hat{s} + \hat{t} + \hat{u} = m_3^2 + m_4^2 ~.
\end{equation}
The cross section for the process $1 + 2 \to 3 + 4$ may be written as
\begin{eqnarray}
\sigma & = & \int \int \int \d x_1 \, \d x_2 \, \d \hat{t} \,
f_1(x_1, Q^2) \, f_2(x_2, Q^2) \, \frac{\d \hat{\sigma}}{\d \hat{t}}
\nonumber \\
& = & \int \int \int \frac{\d \tau}{\tau} \, \d y \, \d \hat{t} \,
x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2) \,
\frac{\d \hat{\sigma}}{\d \hat{t}} ~.
\label{pg:sigma}
\end{eqnarray}
The choice of $Q^2$ scale is ambiguous, and several alternatives are
available in the program. For massless outgoing particles the default
is the squared transverse momentum
\begin{equation}
Q^2 = \hat{p}_{\perp}^2 = \frac{\hat{s}}{4} \sin^2\hat{\theta} =
\frac{\hat{t}\hat{u}}{\hat{s}} ~,
\end{equation}
which is modified to
\begin{equation}
Q^2 = \frac{1}{2} (m_{\perp 3}^2 + m_{\perp 4}^2) =
\frac{1}{2} (m_3^2 + m_4^2) + \hat{p}_{\perp}^2 =
\frac{1}{2} (m_3^2 + m_4^2) +
\frac{\hat{t} \hat{u} - m_3^2 m_4^2}{\hat{s}}
\end{equation}
when masses are introduced. The mass term is selected such that,
for $m_3 = m_4 = m$, the expression reduces to the squared transverse
mass, $Q^2 = \hat{m}_{\perp}^2 = m^2 + \hat{p}_{\perp}^2$.
The $\d \hat{\sigma}/\d \hat{t}$ expresses the differential
cross section for a scattering, as a function of the kinematical
quantities $\hat{s}$, $\hat{t}$ and $\hat{u}$.
It is in this function that the physics of a given process
resides.
The performance of a machine is measured in terms of its
luminosity $\cal L$, which is directly proportional to the
number of particles in each bunch and to the bunch crossing
frequency, and inversely proportional to the area of the bunches at
the collision point. For a process with a $\sigma$ as given by
eq.~(\ref{pg:sigma}), the differential event rate is given
by $\sigma {\cal L}$, and the number of events collected
over a given period of time
\begin{equation}
N = \sigma \, \int {\cal L} \, \d t ~.
\end{equation}
The program does not calculate the number of events, but only the
integrated cross sections.
\subsection{Resonance Production}
\label{ss:kinemreson}
The simplest way to produce a resonance is by a $2 \to 1$ process.
If the decay of the resonance is not considered, the cross-section
formula does not depend on $\hat{t}$, but takes the form
\begin{equation}
\sigma = \int \int \frac{\d \tau}{\tau} \, \d y \,
x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2) \,
\hat{\sigma}(\hat{s}) ~.
\label{pg:sigmares}
\end{equation}
Here the physics is contained in the cross section
$\hat{\sigma}(\hat{s})$. The $Q^2$ scale is usually taken to
be $Q^2 = \hat{s}$.
In published formulae, cross sections are often given in
the zero-width approximation, i.e.
$\hat{\sigma}(\hat{s}) \propto \delta (\hat{s} - m_R^2)$,
where $m_R$ is the mass of the resonance. Introducing the
scaled mass $\tau_R = m_R^2/s$, this corresponds to a delta
function $\delta (\tau - \tau_R)$, which can be used to
eliminate the integral over $\tau$.
However, what we normally want to do is replace the $\delta$
function by the appropriate Breit--Wigner shape. For a resonance
width $\Gamma_R$ this is achieved by the replacement
\begin{equation}
\delta (\tau - \tau_R) \to \frac{s}{\pi} \,
\frac{m_R \Gamma_R}{(s \tau - m_R^2)^2 + m_R^2 \Gamma_R^2} ~.
\label{pg:resshapeone}
\end{equation}
In this formula the resonance width $\Gamma_R$ is a constant.
An improved description of resonance shapes is obtained if
the width is made $\hat{s}$-dependent (occasionally also
referred to as mass-dependent width, since $\hat{s}$ is not
always the resonance mass), see e.g. \cite{Ber89}.
To first approximation, this means that the expression
$m_R \Gamma_R$ is to be replaced by $\hat{s} \Gamma_R / m_R$.
To be more precise, in the program the quantity $H_R(\hat{s})$
is introduced, and the Breit--Wigner is written as
\begin{equation}
\delta (\tau - \tau_R) \to \frac{s}{\pi} \,
\frac{H_R(s \tau)}{(s \tau - m_R^2)^2 + H_R^2(s \tau)} ~.
\label{pg:resshapetwo}
\end{equation}
The $H_R$ factor is evaluated as a sum over all possible final-state
channels, $H_R = \sum_f H_R^{(f)}$. Each decay channel may have its
own $\hat{s}$ dependence, as follows.
A decay to a fermion pair, $R \to \mathrm{f} \overline{\mathrm{f}}$, gives no contribution
below threshold, i.e. for $\hat{s} < 4 m_{\mathrm{f}}^2$. Above threshold,
$H_R^{(f)}$ is proportional to $\hat{s}$, multiplied by a threshold
factor $\beta (3 - \beta^2)/2$ for the vector part of a spin 1
resonance, by $\beta^3$ for the axial vector part, and again by
$\beta^3$ for a spin 0 resonance. Here
$\beta = \sqrt{1 - 4m_{\mathrm{f}}^2/\hat{s}}$.
For the decay into unequal masses, e.g. of the $\mathrm{W}^+$, corresponding
but more complicated expressions are used.
For decays into a quark pair, the universal first-order strong
correction factor $1 + \alpha_{\mathrm{s}}(\hat{s}) / \pi$ is included in
$H_R^{(f)}$.
The second-order corrections are often known, but then are specific
to each resonance, and are not included. An option exists for the
$\gamma/\mathrm{Z}^0/\mathrm{Z}'^0$ resonances, where threshold effects due to
$\mathrm{q}\overline{\mathrm{q}}$ bound-state formation are taken into account in a
smeared-out, average sense, see eq.~(\ref{pp:threshenh}).
For other decay channels, not into fermion pairs, the $\hat{s}$
dependence is typically more complicated. For instance, the decay
$\H^0 \to \mathrm{W}^+ \mathrm{W}^-$ has a partial width proportional to
$\hat{s}^2$, with a threshold factor $\beta^3$. Since a Higgs
with $m_{\H} < 2 m_{\mathrm{W}}$ could still decay in this channel, it is
in fact necessary to perform a two-dimensional integral over
the $W^{\pm}$ Breit--Wigner mass distributions to obtain the correct
result (and this has to be done numerically, at least in part).
Fortunately, a Higgs particle lighter than $2 m_{\mathrm{W}}$ is
sufficiently narrow that the integral only needs to be performed
once and for all at initialization (whereas most other partial
widths are recalculated whenever needed). Channels that proceed
via loops, such as $\H \to \mathrm{g} \mathrm{g}$, also display complicated
threshold behaviours.
The coupling structure within the electroweak sector is usually
(re)expressed in terms of gauge boson masses, $\alpha_{\mathrm{em}}$ and
$\sin^2 \! \theta_W$, i.e. factors of $G_{\mathrm{F}}$ are replaced according to
\begin{equation}
\sqrt{2} G_{\mathrm{F}} = \frac{\pi \, \alpha_{\mathrm{em}}}{\sin^2 \! \theta_W \, m_{\mathrm{W}}^2} ~.
\end{equation}
Having done that, $\alpha_{\mathrm{em}}$ is allowed to run \cite{Kle89},
and is evaluated at the $\hat{s}$ scale. Thereby the relevant
electroweak loop correction factors are recovered at the
$m_{\mathrm{W}}/m_{\mathrm{Z}}$ scale. However, the option exists to
go the other way and eliminate $\alpha_em$ in favour of $G_{\mathrm{F}}$.
Currently $\sin^2 \! \theta_W$ is not allowed to run.
For the Higgs particle, the couplings to fermions are proportional
to the fermion masses; then also the masses are evaluated at the
$\hat{s}$ scale.
In summary, we see that an $\hat{s}$ dependence may enter several
different ways into the $H_R^{(f)}$ expressions from which the
total $H_R$ is built up. Also note that, with the exception of
the term $(\mathrm{s} \tau - m_R^2)^2$ in the denominator of the Breit--Wigner,
no memory remains of the nominal $m_R$ mass: everywhere else, what
enters is the actual resonance mass $\sqrt{\hat{s}}$.
When only decays to a specific final state $f$ are considered, the
$H_R$ in the denominator remains the sum over all allowed decay
channels, but the numerator only contains the $H_R^{(f)}$ term
of the final state considered.
If the combined production and decay process $i \to R \to f$ is
considered, the same $\hat{s}$ dependence is implicit in the
coupling structure of $i \to R$ as one would have had in
$R \to i$, i.e. to first approximation there is a symmetry between
couplings of a resonance to the initial and to the final state.
The cross section $\hat{\sigma}$
is therefore, in the program, written in the form
\begin{equation}
\hat{\sigma}_{i \to R \to f}(\hat{s}) \propto \frac{\pi}{\hat{s}} \,
\frac{H_R^{(i)}(\hat{s}) \, H_R^{(f)}(\hat{s})}
{(\hat{s} - m_R^2)^2 + H_R^2(\hat{s})} ~.
\label{pg:Hinoutsym}
\end{equation}
As a simple example, the cross section for the process
$\mathrm{e}^- \br{\nu}_{\mathrm{e}} \to \mathrm{W}^- \to \mu^- \br{\nu}_{\mu}$
can be written as
\begin{equation}
\hat{\sigma}(\hat{s}) = 12 \, \frac{\pi}{\hat{s}} \,
\frac{H_{\mathrm{W}}^{(i)}(\hat{s}) \, H_{\mathrm{W}}^{(f)}(\hat{s})}
{(\hat{s} - m_{\mathrm{W}}^2)^2 + H_{\mathrm{W}}^2(\hat{s})} ~,
\end{equation}
where
\begin{equation}
H_{\mathrm{W}}^{(i)}(\hat{s}) = H_{\mathrm{W}}^{(f)}(\hat{s}) =
\frac{\alpha_{\mathrm{em}}(\hat{s})}{24 \, \sin^2 \! \theta_W} \, \hat{s} ~.
\end{equation}
If the effects of several initial and/or final states are studied,
it is straightforward to introduce an appropriate summation in the
numerator.
The analogy between the $H_R^{(f)}$ and $H_R^{(i)}$ cannot be pushed
too far, however. The two differ in several important aspects.
Firstly, colour factors appear reversed: the decay $R \to \mathrm{q}\overline{\mathrm{q}}$
contains a colour factor $N_C = 3$ enhancement, while
$\mathrm{q}\overline{\mathrm{q}} \to R$ is instead suppressed by a factor $1/N_C = 1/3$.
Secondly, the $1 + \alpha_{\mathrm{s}}(\hat{s}) / \pi$ first-order correction
factor for the final state has to be replaced by a more complicated
$K$ factor for the initial state. This factor is not usually known, or
it is known (to first non-trivial order) but too lengthy to be included
in the program. Thirdly, incoming partons as a rule are space-like.
All the threshold suppression factors of the final state expressions
are therefore irrelevant when production is considered. In sum, the
$H_R^{(f)}$--$H_R^{(i)}$ analogy is mainly useful as a
consistency cross-check, while the two usually are
calculated separately. Exceptions include the rather
messy loop structure involved in $\mathrm{g}\g \to \H^0$ and $\H^0 \to \mathrm{g}\g$,
which is only coded once.
It is of some interest to consider the observable resonance shape
when the effects of parton distributions are included. In a
hadron collider, to first approximation, parton distributions tend
to have a behaviour roughly like $f(x) \propto 1/x$ for small $x$
--- this is why $f(x)$ is replaced by $xf(x)$ in eq.
(\ref{pg:sigma}). Instead, the basic parton-distribution behaviour
is shifted into the factor of $1/\tau$ in the integration phase space
$\d \tau/\tau$, cf. eq.~(\ref{pg:sigmares}). When folded with the
Breit--Wigner shape, two effects appear. One is that the overall
resonance is tilted: the low-mass tail is enhanced and the high-mass
one suppressed. The other is that an extremely long tail develops
on the low-mass side of the resonance: when $\tau \to 0$, eq.
(\ref{pg:Hinoutsym}) with $H_R(\hat{s}) \propto \hat{s}$ gives
a $\hat{\sigma}(\hat{s}) \propto \hat{s} \propto \tau$, which
exactly cancels the $1/\tau$ factor mentioned above. Na\"{\i}vely, the
integral over $y$, $\int \d y = - \ln \tau$, therefore gives a
net logarithmic divergence of the resonance shape when $\tau \to 0$.
Clearly, it is then necessary to consider the shape of the
parton distributions in more detail. At not-too-small $Q^2$,
the evolution
equations in fact lead to parton distributions more strongly
peaked than $1/x$, typically with $xf(x) \propto x^{-0.3}$, and
therefore a divergence like $\tau^{-0.3}$ in the cross-section
expression. Eventually this divergence is regularized by a closing
of the phase space, i.e. that $H_R(\hat{s})$ vanishes faster than
$\hat{s}$, and by a less drastic small-$x$ parton-distribution
behaviour when $Q^2 \approx \hat{s} \to 0$.
The secondary peak at small $\tau$ may give a rather high
cross section, which can even rival that of the ordinary
peak around the nominal mass. This is the case, for instance,
with $\mathrm{W}$ production. Such a peak has never been observed
experimentally, but this is not surprising, since the background
from other processes is overwhelming at low $\hat{s}$.
Thus a lepton of one or a few GeV of transverse momentum is far
more likely to come from the decay of a charm or bottom hadron than
from a `$\mathrm{W}$' of a mass of a few GeV. When resonance production
is studied, it is therefore important to set limits on the mass of
the resonance, so as to agree with the experimental definition,
at least to first approximation. If not, cross-section information
given by the program may be very confusing.
Another problem is that often the matrix elements really are valid
only in the resonance region. The reason is that one usually includes
only the simplest $s$-channel graph in the calculation. It is this
`signal' graph that has a peak at the position of the resonance,
where it (usually) gives much larger cross sections than the other
`background' graphs. Away from the resonance position, `signal' and
`background' may be of comparable order, or the `background' may
even dominate. There is a quantum mechanical interference when some
of the `signal' and `background' graphs have the same initial and
final state, and this interference may be destructive or constructive.
When the interference is non-negligible, it is no longer meaningful
to speak of a `signal' cross section. As an example, consider the
scattering of longitudinal $\mathrm{W}$'s,
$\mathrm{W}^+_{\mrm{L}} \mathrm{W}^-_{\mrm{L}} \to \mathrm{W}^+_{\mrm{L}} \mathrm{W}^-_{\mrm{L}}$,
where the `signal' process is $s$-channel exchange of a Higgs.
This graph by itself is ill-behaved away from the resonance region.
Destructive interference with `background' graphs such as $t$-channel
exchange of a Higgs and $s$- and $t$-channel exchange of a $\gamma/\mathrm{Z}$
is required to save unitarity at large energies.
In $\e^+\e^-$ colliders, the $f_{\mathrm{e}}^{\mathrm{e}}$ parton distribution is peaked
at $x = 1$ rather than at $x = 0$. The situation therefore is the
opposite, if one considers e.g. $\mathrm{Z}^0$ production in a machine
running at energies above $m_{\mathrm{Z}}$: the tail towards lower masses
is suppressed and the one towards higher masses enhanced, with a
sharp secondary peak at around the nominal energy of the machine.
Also in this case, an appropriate definition of cross sections
therefore is necessary --- with additional complications due to the
interference between $\gamma^*$ and $\mathrm{Z}^0$. When other processes are
considered, problems of interference with background appears also
here. Numerically the problems may be less pressing, however,
since the secondary peak is occuring in a high-mass region, rather
than in a more complicated low-mass one. Further, in $\e^+\e^-$ there is
little uncertainty from the shape of the parton distributions.
In $2 \to 2$ processes where a pair of resonances are produced, e.g.
$\e^+\e^- \to \mathrm{Z}^0 \H^0$, cross section are almost always given
in the zero-width approximation for the resonances. Here
two substitutions of the type
\begin{equation}
1 = \int \delta (m^2 - m_R^2) \, dm^2
\to \int \frac{1}{\pi} \,
\frac{m_R \Gamma_R}{(m^2 - m_R^2)^2 + m_R^2 \Gamma_R^2} \, dm^2
\label{pg:resshapethree}
\end{equation}
are used to introduce mass distributions
for the two resonance masses, i.e. $m_3^2$ and $m_4^2$.
In the formula, $m_R$ is the nominal mass and $m$ the actually
selected one. The
phase-space integral over $x_1$, $x_1$ and $\hat{t}$ in eq.
(\ref{pg:sigma}) is then extended to involve also $m_3^2$ and
$m_4^2$. The effects of the mass-dependent width is only partly
taken into account, by replacing the nominal masses $m_3^2$ and
$m_4^2$ in the $\d \hat{\sigma} / \d \hat{t}$ expression by the
actually generated ones (also e.g. in the relation between
$\hat{t}$ and $\cos\hat{\theta}$),
while the widths are evaluated at the nominal masses. This
is the equivalent of a simple replacement of $m_R \Gamma_R$ by
$\hat{s} \Gamma_R / m_R$ in the numerator of eq.
(\ref{pg:resshapeone}), but not in the denominator.
In addition, the full threshold dependence, i.e. the
$\beta$-dependent factors, is not reproduced.
There is no particular
reason why the full mass-dependence could not be introduced,
except for the extra work and time consumption needed for each
process. In fact, the matrix elements for several $\gamma^* / \Z^0$
production processes do contain the full expressions.
On the other hand, the matrix elements given in the literature
are often valid only when the resonances are almost on the mass shell,
since some graphs have been omitted. As an example, the process
$\mathrm{q}\overline{\mathrm{q}} \to \mathrm{e}^- \br{\nu}_{\mathrm{e}} \mu^+ \nu_{\mu}$ is dominated by
$\mathrm{q}\overline{\mathrm{q}} \to \mathrm{W}^- \mathrm{W}^+$ when each of the two lepton pairs is close to
$m_{\mathrm{W}}$ in mass, but in general also receives contributions e.g. from
$\mathrm{q}\overline{\mathrm{q}} \to \mathrm{Z}^0 \to \e^+\e^-$, followed by $\mathrm{e}^+ \to \br{\nu}_{\mathrm{e}} \mathrm{W}^+$
and $\mathrm{W}^+ \to \mu^+ \nu_{\mu}$. The latter contributions are neglected
in cross sections given in the zero-width approximation.
Processes with one final-state resonance and another ordinary
final-state product, e.g. $\mathrm{q} \mathrm{g} \to \mathrm{W}^+ \mathrm{q}'$, are treated in
the same spirit as the $2 \to 2$ processes with two resonances,
except that only one mass need be selected according to a
Breit--Wigner.
\subsection{Cross-section Calculations}
\label{ss:PYTcrosscalc}
In the program, the variables used in the generation of a
$2 \to 2$ process are $\tau$, $y$ and $z = \cos\hat{\theta}$.
For a $2 \to 1$ process, the $z$ variable can be integrated out,
and need therefore not be generated as part of the hard process,
except when the allowed angular range of decays is restricted.
In unresolved lepton beams, i.e. when
$f_{\mathrm{e}}^{\mathrm{e}}(x) = \delta(x-1)$, the variables $\tau$ and/or $y$
may be integrated out. We will cover all these special cases
towards the end of the section, and here concentrate on `standard'
$2 \to 2$ and $2 \to 1$ processes.
\subsubsection{The simple $2 \to 2$ processes}
In the spirit of section \ref{ss:MCdistsel}, we want to select simple
functions such that the true $\tau$, $y$ and $z$ dependence of the
cross sections is approximately modelled. In particular, (almost) all
conceivable kinematical peaks should be represented by separate
terms in the approximate formulae. If this can be achieved,
the ratio of the correct to the approximate cross sections will
not fluctuate too much, but allow reasonable Monte Carlo efficiency.
Therefore the variables are generated according to the distributions
$h_{\tau}(\tau)$, $h_y(y)$ and $h_z(z)$, where normally
\begin{eqnarray}
h_{\tau}(\tau) & = & \frac{c_1}{{\cal I}_1} \, \frac{1}{\tau} +
\frac{c_2}{{\cal I}_2} \, \frac{1}{\tau^2} +
\frac{c_3}{{\cal I}_3} \, \frac{1}{\tau(\tau + \tau_R)} +
\frac{c_4}{{\cal I}_4} \,
\frac{1}{(s\tau - m_R^2)^2 + m_R^2 \Gamma_R^2} \nonumber \\
& & + \frac{c_5}{{\cal I}_5} \, \frac{1}{\tau(\tau + \tau_{R'})} +
\frac{c_6}{{\cal I}_6} \,
\frac{1}{(s\tau - m_{R'}^2)^2 + m_{R'}^2 \Gamma_{R'}^2} ~, \\[1mm]
h_y(y) & = & \frac{c_1}{{\cal I}_1} \, (y - y_{\mathrm{min}}) +
\frac{c_2}{{\cal I}_2} \, (y_{\mathrm{max}} - y) +
\frac{c_3}{{\cal I}_3} \, \frac{1}{\cosh y} ~, \\[1mm]
h_z(z) & = & \frac{c_1}{{\cal I}_1} +
\frac{c_2}{{\cal I}_2} \, \frac{1}{a-z} +
\frac{c_3}{{\cal I}_3} \, \frac{1}{a+z} +
\frac{c_4}{{\cal I}_4} \, \frac{1}{(a-z)^2} +
\frac{c_5}{{\cal I}_5} \, \frac{1}{(a+z)^2} ~.
\label{pg:hforms}
\end{eqnarray}
Here each term is separately integrable, with an invertible primitive
function, such that generation of $\tau$, $y$ and $z$ separately
is a standard task, as described in section \ref{ss:MCdistsel}.
In the following we describe the details of the scheme, including
the meaning of the coefficients $c_i$ and ${\cal I}_i$, which are
separate for $\tau$, $y$ and $z$.
The first variable to be selected is $\tau$. The range of allowed
values, $\tau_{\mathrm{min}} \leq \tau \leq \tau_{\mathrm{max}}$, is generally
constrained by a number of user-defined requirements. A cut on the
allowed mass range is directly reflected in $\tau$, a cut on the
$p_{\perp}$ range indirectly. The first two terms
of $h_{\tau}$ are intended to represent a smooth $\tau$ dependence,
as generally obtained in processes which do not receive contributions
from $s$-channel resonances. Also $s$-channel exchange of
essentially massless particles ($\gamma$, $\mathrm{g}$, light quarks and
leptons) are accounted for, since these do not produce any separate
peaks at non-vanishing $\tau$. The last four terms of $h_{\tau}$ are
there to catch the peaks in the cross section from resonance
production. These terms are only included when needed. Each resonance
is represented by two pieces, a first to cover the interference with
graphs which peak at $\tau =0$, plus the variation of
parton distributions, and a second to approximate the
Breit--Wigner shape of the resonance itself. The subscripts $R$ and
$R'$ denote values pertaining to the two resonances, with
$\tau_R = m_R^2/s$. Currently there is only one process where the
full structure with two resonances is used, namely
$\mathrm{f} \overline{\mathrm{f}} \to \gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0$. Otherwise either one or no
resonance peak is taken into account.
The kinematically allowed range of $y$ values is
constrained by $\tau$, $|y| \leq - \frac{1}{2} \ln\tau$, and you
may impose additional cuts. Therefore the allowed range
$y_{\mathrm{min}} \leq y \leq y_{\mathrm{max}}$ is only constructed after
$\tau$ has been selected. The first two terms of $h_y$ give a fairly
flat $y$ dependence --- for processes which are symmetric in
$y \leftrightarrow -y$, they will add to give a completely flat $y$
spectrum between the allowed limits. In
principle, the natural subdivision would have been one term flat
in $y$ and one forward--backward asymmetric, i.e. proportional to
$y$. The latter is disallowed by the requirement of positivity,
however. The $y - y_{\mathrm{min}}$ and $y_{\mathrm{max}} - y$ terms
actually used give the same amount of freedom, but respect positivity.
The third term is peaked at around $y = 0$, and represents the bias
of parton distributions towards this region.
The allowed $z = \cos\hat{\theta}$ range is na\"{\i}vely
$-1 \leq z \leq 1$. However, most cross sections are divergent for
$z \to \pm 1$, so some kind of regularization is necessary. Normally
one requires $p_{\perp} \geq p_{\perp\mathrm{min}}$, which translates into
$z^2 \leq 1 - 4 p_{\perp\mathrm{min}}^2/(\tau s)$ for massless outgoing
particles. Since again the limits depend on $\tau$,
the selection of $z$ is done after that of
$\tau$. Additional requirements may constrain the range further.
In particular, a $p_{\perp\mathrm{max}}$ constraint may split the allowed
$z$ range into two, i.e. $z_{-\mathrm{min}} \leq z \leq z_{-\mathrm{max}}$ or
$z_{+\mathrm{min}} \leq z \leq z_{+\mathrm{max}}$. An unsplit range is
represented by $z_{-\mathrm{max}} = z_{+\mathrm{min}} = 0$.
For massless outgoing particles
the parameter $a = 1$ in $h_z$, such that the five terms represent
a piece flat in angle and pieces peaked as $1/\hat{t}$, $1/\hat{u}$,
$1/\hat{t}^2$, and $1/\hat{u}^2$, respectively. For non-vanishing
masses one has $a = 1 + 2 m_3^2 m_4^2/\hat{s}^2$.
In this case, the full range $-1 \leq z \leq 1$ is therefore
available --- physically, the standard $\hat{t}$ and $\hat{u}$
singularities are regularized by the masses $m_3$ and $m_4$.
For each of the terms, the ${\cal I}_i$ coefficients represent the
integral over the quantity multiplying the coefficient $c_i$; thus,
for instance:
\begin{eqnarray}
h_{\tau}: & & {\cal I}_1 = \int \frac{\d \tau}{\tau} =
\ln \left( \frac{\tau_{\mathrm{max}}}{\tau_{\mathrm{min}}} \right) ~,
\nonumber \\
& & {\cal I}_2 = \int \frac{\d \tau}{\tau^2} =
\frac{1}{\tau_{\mathrm{min}}} - \frac{1}{\tau_{\mathrm{max}}} ~;
\nonumber \\
h_y: & & {\cal I}_1 = \int (y - y_{\mathrm{min}}) \, \d y =
\frac{1}{2} (y_{\mathrm{max}} - y_{\mathrm{min}})^2 ~; \nonumber \\
h_z: & & {\cal I}_1 = \int \d z = (z_{-\mathrm{max}} - z_{-\mathrm{min}}) +
(z_{+\mathrm{max}} - z_{+\mathrm{min}}) , \nonumber \\
& & {\cal I}_2 = \int \frac{\d z}{a-z} =
\ln \left( \frac{(a-z_{-\mathrm{min}})(a-z_{+\mathrm{min}})}
{(a-z_{-\mathrm{max}})(a-z_{-\mathrm{min}})} \right) ~.
\end{eqnarray}
The $c_i$ coefficients are normalized to unit sum for $h_{\tau}$,
$h_y$ and $h_z$ separately. They have a simple interpretation, as
the probability
for each of the terms to be used in the preliminary selection of
$\tau$, $y$ and $z$, respectively. The variation of the cross section
over the allowed phase space is explored in the initialization
procedure of a {\tsc{Pythia}} run, and based on this knowledge the $c_i$
are optimized so as to give functions $h_{\tau}$, $h_y$ and $h_z$ that
closely follow the general behaviour of the true cross section.
For instance, the coefficient $c_4$ in $h_{\tau}$ is to be made larger
the more the total cross section is dominated by the region around
the resonance mass.
The phase-space points tested at initialization
are put on a grid, with the number of
points in each dimension given by the number of terms in the
respective $h$ expression, and with the position of each point
given by the median value of the distribution of one of the terms.
For instance, the $\d \tau / \tau$ distribution gives a median point at
$\sqrt{\tau_{\mathrm{min}}\tau_{\mathrm{max}}}$, and $\d \tau / \tau^2$ has the
median $2 \tau_{\mathrm{min}} \tau_{\mathrm{max}}
/ (\tau_{\mathrm{min}} + \tau_{\mathrm{max}})$.
Since the allowed $y$ and $z$ ranges depend on the $\tau$ value
selected, then so do the median points defined for these two
variables.
With only a limited set of phase-space points
studied at the initialization, the `optimal' set of coefficients
is not uniquely defined. To be on the safe side, 40\% of the total
weight is therefore assigned evenly between all allowed $c_i$,
whereas the remaining 60\% are assigned according to the relative
importance surmised, under the constraint that no coefficient is
allowed to receive a negative contribution from this second piece.
After a preliminary choice has been made of $\tau$, $y$ and $z$,
it is necessary to find the weight of the event, which is to be used
to determine whether to keep it or generate another one.
Using the relation $\d \hat{t} = \hat{s} \, \beta_{34} \, \d z / 2$,
eq.~(\ref{pg:sigma}) may be rewritten as
\begin{eqnarray}
\sigma & = & \int \int \int \frac{\d \tau}{\tau} \, \d y \,
\frac{\hat{s} \beta_{34}}{2} \d z \,
x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2) \,
\frac{\d \hat{\sigma}}{\d \hat{t}} \nonumber \\[1mm]
& = & \frac{\pi}{s} \int h_{\tau}(\tau) \, \d \tau
\int h_y(y) \, \d y \int h_z(z) \, \d z \, \beta_{34} \,
\frac{x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2)}
{\tau^2 h_{\tau}(\tau) \, h_y(y) \, 2 h_z(z)} \,
\frac{\hat{s}^2}{\pi} \frac{\d \hat{\sigma}}{\d \hat{t}}
\nonumber \\[1mm]
& = & \left\langle \frac{\pi}{s} \,
\frac{\beta_{34}}{\tau^2 h_{\tau}(\tau) \, h_y(y) \, 2 h_z(z)} \,
x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2) \,
\frac{\hat{s}^2}{\pi} \frac{\d \hat{\sigma}}{\d \hat{t}}
\right\rangle ~.
\label{pg:sigmamap}
\end{eqnarray}
In the middle line, a factor of $1 = h_{\tau}/h_{\tau}$
has been introduced to rewrite the $\tau$ integral
in terms of a phase space of unit volume:
$\int h_{\tau}(\tau) \, \d \tau = 1$ according to the relations
above. Correspondingly for the $y$ and $z$ integrals. In addition,
factors of $1 = \hat{s}/ (\tau s)$ and $1 = \pi / \pi$ are used to
isolate the dimensionless cross section
$(\hat{s}^2/\pi) \, \d \hat{\sigma} / \d \hat{t}$.
The content of the last line is that, with $\tau$, $y$ and $z$
selected according to the expressions $h_{\tau}(\tau)$, $h_y(y)$
and $h_z(z)$, respectively, the cross section is obtained as the
average of the final expression over all events. Since the $h$'s
have been picked to give unit volume, there is no need to multiply
by the total phase-space volume.
As can be seen, the cross section for a given Monte Carlo event is
given as the product of four factors, as follows:
\begin{Enumerate}
\item The $\pi/s$ factor, which is common to all events, gives the
overall dimensions of the cross section, in GeV$^{-2}$. Since
the final cross section is given in units of mb, the conversion
factor of 1 GeV$^{-2} = 0.3894$ mb is also included here.
\item Next comes the `Jacobian', which compensates for the change
from the original to the final phase-space volume.
\item The parton-distribution function weight is obtained by making
use of the parton distribution libraries in {\tsc{Pythia}} or externally.
The $x_1$ and $x_2$ values are obtained from $\tau$ and $y$ via the
relations $x_{1,2} = \sqrt{\tau} \exp(\pm y)$.
\item Finally, the dimensionless cross section
$(\hat{s}^2/\pi) \, \d \hat{\sigma} / \d \hat{t}$
is the quantity that has to be coded for each process separately,
and where the physics content is found.
\end{Enumerate}
Of course, the expression in the last line is not strictly necessary
to obtain the cross section by Monte Carlo integration. One could
also have used eq.~(\ref{pg:sigma}) directly, selecting phase-space
points evenly in $\tau$, $y$ and $\hat{t}$, and averaging over those
Monte Carlo weights. Clearly this would be much simpler, but the price
to be paid is that the weights of individual events could fluctuate
wildly. For instance, if the cross section contains a narrow
resonance, the few phase-space points that are generated in the
resonance region obtain large weights, while the rest do not.
With our procedure, a resonance would be included in the
$h_{\tau}(\tau)$ factor, so that more events would be generated
at around the appropriate $\tau_R$ value (owing to the $h_{\tau}$
numerator in the phase-space expression), but with these events
assigned a lower, more normal weight (owing to the factor
$1/h_{\tau}$ in the weight expression).
Since the weights fluctuate less, fewer phase-space points
need be selected to get a reasonable cross-section estimate.
In the program, the cross section is obtained as the average over all
phase-space points generated. The events actually handed on to the
user should have unit weight, however (an option with weighted events
exists, but does not represent the mainstream usage). At
initialization, after the $c_i$ coefficients have been determined,
a search inside the allowed phase-space volume is therefore made
to find the maximum of the weight expression in the last line of
eq.~(\ref{pg:sigmamap}). In the subsequent generation of events,
a selected phase-space point is then retained with a probability
equal to the weight in the point divided by the maximum weight.
Only the retained phase-space points are considered further, and
generated as complete events.
The search for the maximum is begun by evaluating the weight in the
same grid of points as used to determine the $c_i$ coefficients.
The point with highest weight is used as starting point for a
search towards the maximum. In unfortunate cases, the convergence
could be towards a local maximum which is not the global one.
To somewhat reduce this risk, also the grid point with
second-highest weight is used for another search. After
initialization, when events are generated, a warning message
will be given by default at any time a phase-space
point is selected where the weight is larger than the maximum,
and thereafter the maximum weight is adjusted to reflect the new
knowledge. This means that events generated before this time have
a somewhat erroneous distribution in phase space, but if the
maximum violation is rather modest the effects should be negligible.
The estimation of the cross section is not affected by any of these
considerations, since the maximum weight does not enter into eq.
(\ref{pg:sigmamap}).
For $2 \to 2$ processes with identical final-state particles,
the symmetrization factor of $1/2$ is explicitly included at the
end of the $\d \hat{\sigma} / \d \hat{t}$ calculation. In the final
cross section, a factor of 2 is retrieved because of integration
over the full phase space (rather than only half of it). That
way, no special provisions are needed in the phase-space
integration machinery.
\subsubsection{Resonance production}
We have now covered the simple $2 \to 2$ case. In a $2 \to 1$
process, the $\hat{t}$ integral is absent, and the differential
cross section $\d \hat{\sigma} / \d \hat{t}$ is replaced by
$\hat{\sigma}(\hat{s})$. The cross section may now be written as
\begin{eqnarray}
\sigma & = & \int \int \frac{\d \tau}{\tau} \, \d y \,
x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2) \,
\hat{\sigma}(\hat{s}) \nonumber \\
& = & \frac{\pi}{s} \int h_{\tau}(\tau) \, \d \tau
\int h_y(y) \, \d y \,
\frac{x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2)}
{\tau^2 h_{\tau}(\tau) \, h_y(y)} \,
\frac{\hat{s}}{\pi} \hat{\sigma}(\hat{s}) \nonumber \\
& = & \left\langle \frac{\pi}{s} \,
\frac{1}{\tau^2 h_{\tau}(\tau) \, h_y(y)} \,
x_1 f_1(x_1, Q^2) \, x_2 f_2(x_2, Q^2) \,
\frac{\hat{s}}{\pi} \hat{\sigma}(\hat{s})
\right\rangle ~.
\label{pg:sigmamapres}
\end{eqnarray}
The structure is thus exactly the same, but the $z$-related pieces
are absent, and the r\^ole of the dimensionless cross section is
played by $(\hat{s}/\pi) \hat{\sigma}(\hat{s})$.
If the range of allowed decay angles of the resonance is restricted,
e.g. by requiring the decay products to have a minimum transverse
momentum, effectively this translates into constraints on the
$z = \cos\hat{\theta}$ variable of the $2 \to 2$ process. The
difference is that the angular dependence of a resonance decay is
trivial, and that therefore the $z$-dependent factor can be easily
evaluated. For a spin-0 resonance, which decays isotropically, the
relevant weight is simply
$(z_{-\mathrm{max}} - z_{-\mathrm{min}})/2 + (z_{+\mathrm{max}} - z_{+\mathrm{min}})/2$.
For a transversely polarized spin-1 resonance the expression is,
instead,
\begin{equation}
\frac{3}{8}(z_{-\mathrm{max}} - z_{-\mathrm{min}}) +
\frac{3}{8}(z_{+\mathrm{max}} - z_{+\mathrm{min}}) +
\frac{1}{8}(z_{-\mathrm{max}} - z_{-\mathrm{min}})^3 +
\frac{1}{8}(z_{+\mathrm{max}} - z_{+\mathrm{min}})^3 ~.
\end{equation}
Since the allowed $z$ range could depend on $\tau$ and/or $y$ (it does
for a $p_{\perp}$ cut), the factor has to be evaluated for each individual
phase-space point and included in the expression of eq.
(\ref{pg:sigmamapres}).
For $2 \to 2$ processes where either of the final-state
particles is a resonance, or both, an additional choice has to
be made for
each resonance mass, eq.~(\ref{pg:resshapethree}). Since the allowed
$\tau$, $y$ and $z$ ranges depend on $m_3^2$ and $m_4^2$,
the selection of masses have to precede the choice of the other
phase-space variables. Just as for the other variables, masses
are not selected uniformly over the allowed range, but are rather
distributed according to a function $h_m(m^2) \, dm^2$, with a
compensating factor $1/h_m(m^2)$ in the `Jacobian'. The functional
form picked is normally
\begin{equation}
h_m(m^2) = \frac{c_1}{{\cal I}_1} \, \frac{1}{\pi} \,
\frac{m_R \Gamma_R}{(m^2 - m_R^2)^2 + m_R^2 \Gamma_R^2} +
\frac{c_2}{{\cal I}_2} +
\frac{c_3}{{\cal I}_3} \, \frac{1}{m^2} +
\frac{c_4}{{\cal I}_4} \, \frac{1}{m^4} ~.
\label{pg:reshdist}
\end{equation}
The definition of the ${\cal I}_i$ integrals is analogous to the one
before. The $c_i$ coefficients are not found by optimization, but
predetermined, normally to $c_1 = 0.8$, $c_2 = c_3 =0.1$, $c_4 = 0$.
Clearly, had the phase space and the cross section been independent
of $m_3^2$ and $m_4^2$, the optimal choice would have been to put
$c_1 = 1$ and have all other $c_i$ vanishing --- then the $1/h_m$
factor of the `Jacobian' would exactly have cancelled the
Breit--Wigner of eq.~(\ref{pg:resshapethree}) in the cross section.
The second and the third terms are there to cover the possibility
that the cross section does not die away quite as fast as given by
the na\"{\i}ve Breit--Wigner shape. In particular, the third term covers
the possibility of a secondary peak at small $m^2$, in a spirit
slightly similar to the one discussed for resonance production in
$2 \to 1$ processes.
The fourth term is only used for the process
$\mathrm{f} \overline{\mathrm{f}} \to (\gamma^* / \Z^0)(\gamma^* / \Z^0)$, where the $\gamma$ propagator
guarantees that the cross section does have a significant secondary
peak for $m^2 \to 0$. Therefore here the choice is $c_1 = 0.4$,
$c_2 = 0.05$, $c_3 = 0.3$ and $c_4 = 0.25$.
A few special tricks have been included to improve efficiency
when the allowed mass range of resonances is constrained by
kinematics or by user cuts. For instance, if a pair of equal
or charge-conjugate resonances are produced, such as in
$\e^+\e^- \to \mathrm{W}^+ \mathrm{W}^-$, use is made of the constraint that the lighter
of the two has to have a mass smaller than half the c.m. energy.
\subsubsection{Lepton beams}
Lepton beams have to be handled slightly differently from what has
been described so far. One also has to distinguish between a lepton
for which parton distributions are included and one which is treated
as an unresolved point-like particle. The necessary modifications are
the same for $2 \to 2$ and $2 \to 1$ processes, however, since the
$\hat{t}$ degree of freedom is unaffected.
If one incoming beam is an unresolved lepton, the corresponding
parton-distribution piece collapses to a $\delta$ function. This
function can be used to integrate out the $y$ variable:
$\delta(x_{1,2} - 1) = \delta(y \pm (1/2) \ln \tau)$.
It is therefore only necessary to select the $\tau$ and the $z$
variables according to the proper distributions, with compensating
weight factors, and only one set of parton distributions has to be
evaluated explicitly.
If both incoming beams are unresolved leptons, both the $\tau$ and
the $y$ variables are trivially given: $\tau = 1$ and $y = 0$.
Parton-distribution weights disappear completely. For a $2 \to 2$
process, only the $z$ selection remains to be performed, while
a $2 \to 1$ process is completely specified, i.e. the cross section
is a simple number that only depends on the c.m. energy.
For a resolved electron, the $f_{\mathrm{e}}^{\mathrm{e}}$ parton distribution is
strongly peaked towards $x = 1$. This affects both the $\tau$
and the $y$ distributions, which are not well described by
either of the pieces in $h_{\tau}(\tau)$ or $h_y(y)$ in processes with
interacting $\mathrm{e}^{\pm}$. (Processes which involve e.g. the $\gamma$
content of the $\mathrm{e}$ are still well simulated, since
$f_{\gamma}^{\mathrm{e}}$ is peaked at small $x$.)
If both parton distributions
are peaked close to 1, the $h_{\tau}(\tau)$ expression in eq.
(\ref{pg:hforms}) is therefore increased with one additional term of
the form $h_{\tau}(\tau) \propto 1 / (1 - \tau)$, with coefficients
$c_7$ and ${\cal I}_7$ determined as before. The divergence when
$\tau \to 1$ is cut off by our regularization procedure for the
$f_{\mathrm{e}}^{\mathrm{e}}$ parton distribution; therefore we only need consider
$\tau < 1 - 2 \times 10^{-6}$.
Correspondingly, the $h_y(y)$ expression is expanded with a term
$1/(1 - \exp(y-y_0))$ when incoming beam number 1 consists of a
resolved $\mathrm{e}^{\pm}$, and with a term $1/(1 - \exp(-y-y_0))$
when incoming beam number 2 consists of a resolved $\mathrm{e}^{\pm}$.
Both terms are present for an $\e^+\e^-$ collider, only one for an
$\e\p$ one. The coefficient $y_0 = - (1/2) \ln \tau$ is the na\"{\i}ve
kinematical limit of the $y$ range, $|y| < y_0$. From the
definitions of $y$ and $y_0$ it is easy to see
that the two terms above correspond to $1/(1-x_1)$ and $1/(1-x_2)$,
respectively, and thus are again regularized by our
parton-distribution function cut-off. Therefore the integration ranges
are $y < y_0 -10^{-6}$ for the first term and $y > - y_0 + 10^{-6}$
for the second one.
\subsubsection{Mixing processes}
\label{sss:mixingproc}
In the cross-section formulae given so far, we have deliberately
suppressed a summation over the allowed incoming flavours. For
instance, the process $\mathrm{f}\overline{\mathrm{f}} \to \mathrm{Z}^0$ in a hadron collider
receives contributions from $\u\overline{\mathrm{u}} \to \mathrm{Z}^0$, $\d\overline{\mathrm{d}} \to \mathrm{Z}^0$,
$\mathrm{s}\overline{\mathrm{s}} \to \mathrm{Z}^0$, and so on. These contributions share the
same basic form, but differ in the parton-distribution weights
and (usually) in a few coupling constants in the hard matrix
elements. It it therefore convenient to generate the terms
together, as follows:
\begin{Enumerate}
\item A phase-space point is picked, and all
common factors related to this choice are evaluated, i.e.
the `Jacobian' and the common pieces of the matrix elements
(e.g. for a $\mathrm{Z}^0$ the basic Breit--Wigner shape, excluding
couplings to the initial flavour).
\item The parton-distribution-function library is called to produce
all the parton distributions, at the relevant $x$ and $Q^2$ values,
for the two incoming beams.
\item A loop is made over the two incoming flavours, one from each
beam particle. For each allowed set of incoming flavours, the full
matrix-element expression is put together, using the common pieces and
the flavour-dependent couplings. This is multiplied by the common
factors and the parton-distribution weights to obtain a cross-section
weight.
\item Each allowed flavour combination is stored as a separate entry
in a table, together with its weight. In addition, a summed weight
is calculated.
\item The phase-space point is kept or rejected, according to a
comparison of the summed weight with the maximum weight obtained
at initialization. Also the cross-section Monte Carlo integration
is based on the summed weight.
\item If the point is retained, one of the allowed flavour
combinations is picked according to the relative weights stored
in the full table.
\end{Enumerate}
Generally, the flavours of the final state are either completely
specified by those of the initial state, e.g. as in $\mathrm{q}\mathrm{g} \to \mathrm{q}\mathrm{g}$,
or completely decoupled from them, e.g. as in
$\mathrm{f}\overline{\mathrm{f}} \to \mathrm{Z}^0 \to \mathrm{f}'\overline{\mathrm{f}}'$. In neither case need therefore the
final-state flavours be specified in the cross-section calculation.
It is only necessary, in the latter case, to include an overall
weight factor, which takes into account the summed contribution of
all final states that are to be simulated. For instance, if only
the process $\mathrm{Z}^0 \to \e^+\e^-$ is studied, the relevant weight factor
is simply $\Gamma_{\mathrm{e}\e} / \Gamma_{\mrm{tot}}$. Once the kinematics
and the incoming flavours have been selected, the outgoing flavours
can be picked according to the appropriate relative probabilities.
In some processes, such as $\mathrm{g}\g \to \mathrm{g}\g$, several different colour
flows are allowed, each with its own kinematical dependence of the
matrix-element weight, see section \ref{sss:QCDjetclass}. Each colour
flow is then given as a separate entry in the table mentioned above,
i.e. in total an entry is characterized by the two incoming flavours,
a colour-flow index, and the weight. For an accepted phase-space
point, the colour flow is selected in the same way as the incoming
flavours.
The program can also allow the mixed generation of two or more
completely different processes, such as $\mathrm{f}\overline{\mathrm{f}} \to \mathrm{Z}^0$ and
$\mathrm{q}\overline{\mathrm{q}} \to \mathrm{g}\g$. In that case, each process is initialized
separately, with its own set of coefficients $c_i$ and so on.
The maxima obtained for the individual cross sections are all
expressed in the same units, even when the dimensionality of the
phase space is different. (This is because we always transform to
a phase space of unit volume,
$\int h_{\tau}(\tau) \, \d \tau \equiv 1$, etc.) The above
generation scheme need therefore only be generalized as follows:
\begin{Enumerate}
\item One process is selected among the allowed ones, with a relative
probability given by the maximum weight for this process.
\item A phase-space point is found, using the distributions
$h_{\tau}(\tau)$ and so on, optimized for this particular process.
\item The total weight for the phase-space point is evaluated,
again with `Jacobians', matrix elements and allowed incoming flavour
combinations that are specific to the process.
\item The point is retained with a probability given by the ratio of
the actual to the maximum weight of the process. If the point is
rejected, one has to go back to step 1 and pick a new process.
\item Once a phase-space point has been accepted, flavours may be
selected, and the event generated in full.
\end{Enumerate}
It is clear why this works: although phase-space points are selected
among the allowed processes according to relative probabilities given
by the maximum weights, the probability that a point is accepted is
proportional to the ratio of actual to maximum weight. In total,
the probability for a given process to be retained is therefore only
proportional to the average of the actual weights, and any dependence
on the maximum weight is gone.
In $\gamma\mathrm{p}$ and $\gamma\gamma$ physics, the different components
of the photon give different final states, see section
\ref{sss:photoprod}. Technically, this introduces a further level
of administration, since each event class contains a set of (partly
overlapping) processes. From an ideological point of view, however,
it just represents one more choice to be made, that of event class,
before the selection of process in step 1 above. When a weighting
fails, both class and process have to be picked anew.
\subsection{$2 \to 3$ and $2 \to 4$ Processes}
The {\tsc{Pythia}} machinery to handle $2 \to 1$ and $2 \to 2$ processes
is fairly sophisticated and generic. The same cannot be said about
the generation of hard scattering processes with more than two
final-state particles. The number of phase-space variables is
larger, and
it is therefore more difficult to find and transform away all possible
peaks in the cross section by a suitably biased choice of phase-space
points. In addition, matrix-element expressions for $2 \to 3$
processes are typically fairly lengthy. Therefore {\tsc{Pythia}} only contains
a very limited number of $2 \to 3$ and $2 \to 4$ processes, and almost
each process is a special case of its own. It is therefore less
interesting to discuss details, and we only give a very generic
overview.
If the Higgs mass is not light, interactions among longitudinal
$\mathrm{W}$ and $\mathrm{Z}$ gauge bosons are of interest. In the program,
$2 \to 1$ processes such as $\mathrm{W}_{\mrm{L}}^+ \mathrm{W}_{\mrm{L}}^- \to \H^0$
and $2 \to 2$ ones such as
$\mathrm{W}_{\mrm{L}}^+ \mathrm{W}_{\mrm{L}}^- \to \mathrm{Z}_{\mrm{L}}^0 \mathrm{Z}_{\mrm{L}}^0$
are included.
The former are for use when the $\H^0$ still is reasonably narrow,
such that a resonance description is applicable, while the latter
are intended for high energies, where different contributions have
to be added up. Since the program does not contain
$\mathrm{W}_{\mrm{L}}$ or $\mathrm{Z}_{\mrm{L}}$
distributions inside hadrons, the basic hard scattering has
to be convoluted with the $\mathrm{q} \to \mathrm{q}' \mathrm{W}_{\mrm{L}}$ and
$\mathrm{q} \to \mathrm{q} \mathrm{Z}_{\mrm{L}}$
branchings, to yield effective $2 \to 3$ and $2 \to 4$ processes.
However, it is possible to integrate out the scattering angles of
the quarks analytically, as well as one energy-sharing variable
\cite{Cha85}. Only after an event has been accepted are these other
kinematical variables selected. This involves further choices of
random variables, according to a separate selection loop.
In total, it is therefore only necessary to introduce one additional
variable in the basic phase-space selection, which is chosen to be
$\hat{s}'$, the squared invariant mass of the full $2 \to 3$ or
$2 \to 4$ process, while $\hat{s}$ is used for the squared invariant
mass of the inner $2 \to 1$ or $2 \to 2$ process. The $y$ variable
is coupled to the full process, since parton-distribution weights
have to be given for the original quarks at
$x_{1,2} = \sqrt{\tau'} \exp{(\pm y)}$. The $\hat{t}$ variable is
related to the inner process, and thus not needed for the $2 \to 3$
processes. The selection of the $\tau' = \hat{s}'/s$ variable is
done after $\tau$, but before $y$ has been chosen. To improve the
efficiency, the selection is made according to a weighted phase space
of the form $\int h_{\tau'}(\tau') \, \d \tau'$, where
\begin{equation}
h_{\tau'}(\tau') = \frac{c_1}{{\cal I}_1} \frac{1}{\tau'} +
\frac{c_2}{{\cal I}_2} \, \frac{(1- \tau / \tau')^3}{\tau'^2} +
\frac{c_3}{{\cal I}_3} \, \frac{1}{\tau' (1 - \tau')} ~,
\end{equation}
in conventional notation. The $c_i$ coefficients are optimized
at initialization. The $c_3$ term, peaked at $\tau' \approx 1$,
is only used for $\e^+\e^-$ collisions.
The choice of $h_{\tau'}$ is roughly matched to the
longitudinal gauge-boson flux factor, which is of the form
\begin{equation}
\left( 1 + \frac{\tau}{\tau'} \right) \,
\ln \left( \frac{\tau}{\tau'} \right) -
2 \left( 1 - \frac{\tau}{\tau'} \right) ~.
\end{equation}
For a light $\H$ the effective $\mathrm{W}$ approximation above breaks down,
and it is necessary to include the full structure of the
$\mathrm{q} \mathrm{q}' \to \mathrm{q} \mathrm{q}' \H^0$ (i.e. $\mathrm{Z}\Z$ fusion) and
$\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \mathrm{q}''' \H^0$ (i.e. $\mathrm{W}\W$ fusion) matrix elements.
The $\tau'$, $\tau$ and $y$ variables are here retained, and selected
according to standard procedures. The Higgs mass is represented by the
$\tau$ choice; normally the $\H^0$ is so narrow that the $\tau$
distribution effectively collapses to a $\delta$ function. In addition,
the three-body final-state phase space is rewritten as
\begin{equation}
\left( \prod_{i=3}^5 \frac{1}{(2 \pi)^3} \frac{\d^3 p_i}{2 E_i}
\right) \, (2 \pi)^4 \delta^{(4)} (p_3 + p_4 + p_5 - p_1 - p_2) =
\frac{1}{(2 \pi)^5} \, \frac{\pi^2}{4 \sqrt{\lambda_{\perp 34}}}
\, \d p_{\perp 3}^2 \, \frac{\d \varphi_3}{2 \pi} \,
\d p_{\perp 4}^2 \, \frac{\d \varphi_4}{2 \pi} \, \d y_5 ~,
\end{equation}
where $\lambda_{\perp 34} = (m_{\perp 34}^2 - m_{\perp 3}^2 -
m_{\perp 4}^2)^2 - 4 m_{\perp 3}^2 m_{\perp 4}^2$.
The outgoing quarks are labelled 3 and 4, and the outgoing Higgs 5.
The $\varphi$ angles are selected isotropically, while the two
transverse momenta are picked, with some foreknowledge of the
shape of the $\mathrm{W} / \mathrm{Z}$ propagators in the cross sections, according
to $h_{\perp} (p_{\perp}^2) \, \d p_{\perp}^2$, where
\begin{equation}
h_{\perp}(p_{\perp}^2) = \frac{c_1}{{\cal I}_1} +
\frac{c_2}{{\cal I}_2} \, \frac{1}{m_R^2 + p_{\perp}^2} +
\frac{c_3}{{\cal I}_3} \, \frac{1}{(m_R^2 + p_{\perp}^2)^2} ~,
\end{equation}
with $m_R$ the $\mathrm{W}$ or $\mathrm{Z}$ mass, depending on process, and
$c_1 = c_2 = 0.05$, $c_3 = 0.9$. Within the limits given by the
other variable choices, the rapidity $y_5$ is chosen uniformly.
A final choice remains to be made, which comes from a twofold
ambiguity of exchanging the longitudinal momenta of partons 3
and 4 (with minor modifications if they are massive). Here
the relative weight can be obtained exactly from the form of the
matrix element itself.
No good phase-space choice was found for the process
$\mathrm{g} \mathrm{g} \to \mathrm{Z}^0 \b \overline{\mathrm{b}}$. This process is therefore not so easy
to generate with {\tsc{Pythia}}. What is currently done is to use the
basic formalism of $2 \to 2$ processes, where the $\b + \overline{\mathrm{b}}$
system is considered as an effective `resonance'. Two masses are
then selected, the $\mathrm{Z}^0$ one according to eq.~(\ref{pg:reshdist})
and the $\b + \overline{\mathrm{b}}$ one according to $\d m^2 / m^2$. Both `decays'
are selected isotropically in the respective rest frame,
to give the final four fermions in
terms of which the matrix element is given. In addition,
$\tau$, $y$ and $z$ are selected according to the standard rules
for $2 \to 2$ processes.
\subsection{Resonance Decays}
\label{ss:resdecay}
Resonances can be made to decay in two different routines. One is the
standard decay treatment (in \ttt{LUDECY}) that can be used
for any unstable particle, where decay channels
are chosen according to fixed probabilities, and decay angles
usually are picked isotropically in the rest frame of the resonance,
see section \ref{ss:partdecays}.
The more sophisticated treatment (in \ttt{PYRESD}) is the default
one for resonances produced in {\tsc{Pythia}}, and is described here.
The following are included in the list of resonances: $\mathrm{Z}^0$,
$\mathrm{W}^{\pm}$, $\H^0$, $\mathrm{Z}'^0$, $\mathrm{W}'^{\pm}$, $\H'^0$, $\mathrm{A}^0$,
$\H^{\pm}$, $\eta_{\mrm{tech}}^0$, $\L_{\mathrm{Q}}$, and $\mathrm{R}^0$.
The top is also considered as a resonance if it is assumed to
decay before it has time to fragment. Likewise for the fourth
generation fermions. If the fourth generation is used to represent
excited quarks and leptons, these are also considered to be
resonances.
\subsubsection{The decay scheme}
In the beginning of the decay treatment, either one or two
resonances may be present, the former represented by processes
such as $\mathrm{q} \overline{\mathrm{q}}' \to \mathrm{W}^+$ and $\mathrm{q} \mathrm{g} \to \mathrm{W}^+ \mathrm{q}'$, the latter
by $\mathrm{q} \overline{\mathrm{q}} \to \mathrm{W}^+ \mathrm{W}^-$. If the latter is the case, the
decay of the two resonances is considered in parallel
(unlike \ttt{LUDECY}, where one particle at a time is made to decay).
First the decay channel of each resonance is selected according to
the relative weights $H_R^{(f)}$, as described above, evaluated at
the actual mass of the resonance, rather than at the nominal one.
Threshold factors are therefore fully taken into account, with
channels automatically switched off below the threshold. Normally
the masses of the decay products are well-defined, but e.g. in
decays like $\H^0 \to \mathrm{W}^+ \mathrm{W}^-$ it is also necessary to select
the decay product masses. This is done according to two Breit--Wigners
of the type in eq.~(\ref{pg:resshapethree}), multiplied by the
threshold factor, which depends on both masses.
Next the decay angles of the resonance are selected isotropically in
its rest frame. Normally the full range of decay angles is available,
but in $2 \to 1$ processes the decay angles of the original resonance
may be restrained by user cuts, e.g. on the $p_{\perp}$ of the decay
products. Based on the angles, the four-momenta of the decay products
are constructed and boosted to the correct frame. As a rule, matrix
elements are given with quark and lepton masses assumed vanishing.
Therefore the four-momentum vectors constructed at this stage are
actually massless for all quarks and leptons.
The matrix elements may now be evaluated. For a process such as
$\mathrm{q} \overline{\mathrm{q}} \to \mathrm{W}^+ \mathrm{W}^- \to \mathrm{e}^+ \nu_{\mathrm{e}} \mu^- \br{\nu}_{\mu}$,
the matrix element is a function of the four-momenta of the two
incoming fermions and of the four outgoing ones. An upper limit for
the event weight can be constructed from the cross section for the
basic process $\mathrm{q} \overline{\mathrm{q}} \to \mathrm{W}^+ \mathrm{W}^-$, as already used to select
the two $\mathrm{W}$ momenta. If the weighting fails, new resonance decay
angles are picked and the procedure is iterated until acceptance.
Based on the accepted set of angles, the correct decay product
four-momenta are constructed, including previously neglected
fermion masses. Quarks and, optionally, leptons are allowed to
radiate, using the standard final-state showering machinery, with
maximum virtuality given by the resonance mass.
In some decays new resonances are produced, and these are then
subsequently allowed to decay. Only one resonance pair is considered
at a time, i.e. it is not possible to include correlations which
involve the simultaneous decay of three or more resonances. This is
in fact all that is currently needed: in a process like
$\mathrm{q}\overline{\mathrm{q}} \to \mathrm{Z}^0 \H^0 \to \mathrm{Z}^0 \mathrm{W}^+ \mathrm{W}^- \to 6$ fermions,
the spinless nature of the $\H^0$ ensures that the $\mathrm{W}^{\pm}$
decays are decoupled from that of the $\mathrm{Z}^0$ (but not from each other).
\subsubsection{Cross-section considerations}
\label{sss:resdecaycross}
The cross section for a process which involves the production of one
or several resonances is always reduced to take into account channels
not allowed by user flags. This is trivial for a single $s$-channel
resonance, cf. eq.~(\ref{pg:Hinoutsym}), but can also be included
approximately if several layers of resonance decays are involved.
At initialization, the ratio between the user-allowed width and the
nominally possible one is evaluated and stored, starting from the
lightest resonances and moving upwards. As an example, one first finds
the reduction factors for $\mathrm{W}^+$ and for $\mathrm{W}^-$ decays, which need not
be the same if e.g. $\mathrm{W}^+$ is allowed to decay only to quarks and
$\mathrm{W}^-$ only to leptons. These factors enter together as a weight
for the $\H^0 \to \mathrm{W}^+ \mathrm{W}^-$ channel, which is thus reduced in
importance compared with other possible Higgs decay channels.
This is also reflected
in the weight factor of the $\H^0$ itself, where some channels are open
in full, others completely closed, and finally some (like the one
above) open but with reduced weight. Finally, the weight for the
process $\mathrm{q}\overline{\mathrm{q}} \to \mathrm{Z}^0 \H^0$ is evaluated as the product of the
$\mathrm{Z}^0$ weight factor and the $\H^0$ one. The standard cross section
of the process is multiplied with this weight.
Since the restriction on allowed decay modes is already included in
the hard process cross section, mixing of different event types is
greatly simplified, and the selection of decay channel chains is
straightforward. There is a price to be paid, however. The reduction
factors evaluated at initialization all refer to resonances at their
nominal masses. For instance, the $\mathrm{W}$ reduction factor is evaluated
at the nominal $\mathrm{W}$ mass, even when that factor is used, later on, in
the description of the decay of a 120 GeV Higgs, where at least one
$\mathrm{W}$ would be produced below this mass. We know of no case where this
approximation has any serious consequences, however.
The weighting procedure works because the number of resonances to be
produced, directly or in subsequent decays, can be derived
recursively already from the start. It does not work for particles
which could also be produced at later stages, such as the
parton-shower evolution and the fragmentation. For instance,
$\mathrm{D}^0$ mesons can be produced fairly late in the event generation
chain, in unknown numbers, and so weights could not be introduced
to compensate, e.g. for the forcing of decays only into $\pi^+ \mathrm{K}^-$.
For similar reasons the top is only considered as a resonance if it
is not allowed to hadronize; see discussion in section
\ref{sss:heavflavclass}.
One should note that this reduction factor is separate from the
description of the resonance shape itself, where the full
width of the resonance has to be used. This width is based on the
sum of all possible decay modes, not just the simulated ones.
{\tsc{Pythia}} does allow the possibility to change also the underlying
physics scenario, e.g. to include the decay of a $\mathrm{Z}^0$ into a
fourth-generation neutrino.
Normally the evaluation of the reduction factors is straightforward.
However, for decays into a pair of equal or charge-conjugate
resonances, such as $\mathrm{Z}^0 \mathrm{Z}^0$ or $\mathrm{W}^+ \mathrm{W}^-$, it is possible to
pick combinations in such a way that the weight of the pair does
not factorize into a product of the weight of each resonance itself.
To be precise, any decay channel can be given seven different status
codes:
\begin{Itemize}
\item $-1$: a non-existent decay mode, completely switched off and of no
concern to us;
\item 0: an existing decay channel, which is switched off;
\item 1: a channel which is switched on;
\item 2: a channel switched on for particles, but off for antiparticles;
\item 3: a channel switched on for antiparticles, but off for particles;
\item 4: a channel switched on for one of the resonances, but not for
both;
\item 5: a channel switched on for the other of the resonances, but not
for both.
\end{Itemize}
The meaning of possibilities 4 and 5 is exemplified by the statement
`in a $\mathrm{W}^+\mathrm{W}^-$ pair, one $\mathrm{W}$ decays hadronically and the other
leptonically', which thus covers the cases where either $\mathrm{W}^+$ or
$\mathrm{W}^-$ decays hadronically.
Neglecting non-existing channels, each channel belongs to either of the
classes above. If we denote the total branching ratio into channels
of type $i$ by $r_i$, this then translates into the requirement
$r_0 + r_1 + r_2 + r_3 + r_4 + r_5 = 1$. For a single particle the
weight factor is $r_1 + r_2 + r_4$, and for a single antiparticle
$r_1 + r_3 + r_4$. For a pair of identical resonances, the joint weight
is instead
\begin{equation}
(r_1 + r_2)^2 + 2 (r_1 + r_2) (r_4 + r_5) + 2 r_4 r_5 ~,
\end{equation}
and for a resonance--antiresonance pair
\begin{equation}
(r_1 + r_2)(r_1 + r_3) + (2 r_1 + r_2 + r_3) (r_4 + r_5) +
2 r_4 r_5 ~.
\label{eq:WWallchancomb}
\end{equation}
If some channels come with a reduced weight because of restrictions
on subsequent decay chains, this may be described in terms of
properly reduced $r_i$, so that the sum is less than unity.
For instance, in a $\t\overline{\mathrm{t}} \to \b\mathrm{W}^+ \, \overline{\mathrm{b}}\mathrm{W}^-$ process,
the $\mathrm{W}$ decay modes may be restricted to $\mathrm{W}^+ \to \mathrm{q}\overline{\mathrm{q}}$
and $\mathrm{W}^- \to \mathrm{e}^-\bar{\nu}_{\mathrm{e}}$, in which case
$(\sum r_i)_{\t} \approx 2/3$ and $(\sum r_i)_{\overline{\mathrm{t}}} \approx 1/9$.
With index $\pm$ denoting resonance/antiresonance,
eq.~(\ref{eq:WWallchancomb}) then generalizes to
\begin{equation}
(r_1 + r_2)^+ (r_1 + r_3)^- + (r_1 + r_2)^+ (r_4 + r_5)^- +
(r_4 + r_5)^+ (r_1 + r_3)^- + r_4^+ r_5^- + r_5^+ r_4^- ~.
\label{eq:WWallchancombgen}
\end{equation}
\subsection{Nonperturbative Processes}
\label{ss:nonpertproc}
A few processes are not covered by the discussion so far. These are
the ones that depend on the details of hadronic wave functions,
and therefore are not strictly calculable perturbatively
(although perturbation theory may often provide some guidance).
What we have primarily in mind is elastic scattering, diffractive
scattering and low-$p_{\perp}$ `minimum-bias' events in hadron--hadron
collisions, but one can also find corresponding processes in
$\gamma \mathrm{p}$ and $\gamma \gamma$ interactions. The description
of these processes is rather differently structured from that of
the other ones, as is explained below. Models for
`minimum-bias' events are discussed in detail in section
\ref{ss:multint}, to which we refer for details on this part
of the program.
\subsubsection{Hadron--hadron interactions}
In hadron--hadron interactions, the total hadronic cross section
for $AB \to$ anything, $\sigma^{AB}_{\mrm{tot}}$, is calculated
using the parametrization of Donnachie and Landshoff \cite{Don92}.
In this approach, each cross section appears as the sum of one
pomeron term and one reggeon one
\begin{equation}
\sigma^{AB}_{\mrm{tot}}(s) = X^{AB} \, s^{\epsilon} +
Y^{AB} \, s^{-\eta} ~,
\label{pg:sigtotpomreg}
\end{equation}
where $s = E_{\mrm{cm}}^2$. The powers $\epsilon = 0.0808$ and
$\eta = 0.4525$ are expected to be universal, whereas the
coefficients $X^{AB}$ and $Y^{AB}$ are specific to each initial
state. (In fact, the high-energy behaviour given by the pomeron term
is expected to be the same for particle and antiparticle interactions,
i.e. $X^{\br{A}B} = X^{AB}$.) Parametrizations not provided
in \cite{Don92} have been calculated in the same spirit, making use
of quark counting rules \cite{Sch93a}.
The total cross section is subdivided according to
\begin{equation}
\sigma^{AB}_{\mrm{tot}}(s) = \sigma^{AB}_{\mrm{el}}(s) +
\sigma^{AB}_{\mrm{sd}(XB)}(s) + \sigma^{AB}_{\mrm{sd}(AX)}(s) +
\sigma^{AB}_{\mrm{dd}}(s) + \sigma^{AB}_{\mrm{nd}}(s) ~.
\label{pg:sigtotsplit}
\end{equation}
Here `el' is the elastic process $AB \to AB$, `sd$(XB)$' the single
diffractive $AB \to XB$, `sd$(AX)$' the single diffractive
$AB \to AX$, `dd' the double diffractive $AB \to X_1 X_2$,
and `nd' the non-diffractive ones. Higher diffractive topologies,
such as central diffraction, are currently neglected.
In the following, the elastic and diffractive cross sections
and event characteristics are described, as given in the model by
Schuler and Sj\"ostrand \cite{Sch94,Sch93a}. The non-diffractive
component is identified with the `minimum bias' physics already
mentioned, a practical but not unambiguous choice. Its cross section
is given by `whatever is left' according to eq.~(\ref{pg:sigtotsplit}),
and its properties are discussed in section \ref{ss:multint}.
At not too large squared momentum transfers $t$, the elastic cross
section can be approximated by a simple exponential fall-off. If one
neglects the small real part of the cross section, the optical
theorem then gives
\begin{equation}
\frac{\d\sigma_{\mrm{el}}}{\d t} =
\frac{\sigma_{\mrm{tot}}^2}{16 \pi} \, \exp(B_{\mrm{el}} t) ~,
\end{equation}
and $\sigma_{\mrm{el}} = \sigma_{\mrm{tot}}^2 / 16 \pi B_{\mrm{el}}$.
The elastic slope parameter is parametrized by
\begin{equation}
B_{\mrm{el}} = B^{AB}_{\mrm{el}}(s) = 2 b_A + 2 b_B +
4 s^{\epsilon} -4.2 ~,
\end{equation}
with $s$ given in units of GeV and $B_{\mrm{el}}$ in GeV$^{-2}$.
The constants $b_{A,B}$ are $b_{\mathrm{p}} = 2.3$,
$b_{\pi,\rho,\omega,\phi} = 1.4$, $b_{\mathrm{J}/\psi} = 0.23$.
The increase of the slope parameter with c.m. energy is faster
than the logarithmically one conventionally assumed; that way the
ratio $\sigma_{\mrm{el}} / \sigma_{\mrm{tot}}$ remains well-behaved
at large energies.
The diffractive cross sections are given by
\begin{eqnarray}
\frac{\d\sigma_{\mrm{sd}(XB)}(s)}{\d t \, \d M^2} & = &
\frac{g_{3\mathrm{I}\!\mathrm{P}}}{16\pi} \, \beta_{A\mathrm{I}\!\mathrm{P}} \,
\beta_{B\mathrm{I}\!\mathrm{P}}^2 \, \frac{1}{M^2} \, \exp(B_{\mrm{sd}(XB)}t)
\, F_{\mrm{sd}} ~, \nonumber \\
\frac{\d\sigma_{\mrm{sd}(AX)}(s)}{\d t \, \d M^2} & = &
\frac{g_{3\mathrm{I}\!\mathrm{P}}}{16\pi} \, \beta_{A\mathrm{I}\!\mathrm{P}}^2 \,
\beta_{B\mathrm{I}\!\mathrm{P}} \, \frac{1}{M^2} \, \exp(B_{\mrm{sd}(AX)}t)
\, F_{\mrm{sd}} ~, \nonumber \\
\frac{\d\sigma_{\mrm{dd}}(s)}{\d t \, \d M_1^2 \, \d M_2^2} & = &
\frac{g_{3\mathrm{I}\!\mathrm{P}}^2}{16\pi} \, \beta_{A\mathrm{I}\!\mathrm{P}} \,
\beta_{B\mathrm{I}\!\mathrm{P}} \, \frac{1}{M_1^2} \, \frac{1}{M_2^2} \,
\exp(B_{\mrm{dd}}t) \, F_{\mrm{dd}} ~.
\end{eqnarray}
The couplings $\beta_{A\mathrm{I}\!\mathrm{P}}$ are related to the pomeron term
$X^{AB} s^{\epsilon}$ of the total cross section parametrization,
eq.~(\ref{pg:sigtotpomreg}). Picking a reference scale
$\sqrt{s_{\mrm{ref}}} = 20$ GeV, the couplings are given by
$\beta_{A\mathrm{I}\!\mathrm{P}}\beta_{B\mathrm{I}\!\mathrm{P}} =
X^{AB} \, s_{\mrm{ref}}^{\epsilon}$. The triple-pomeron coupling is
determined from single-diffractive data to be
$g_{3\mathrm{I}\!\mathrm{P}} \approx 0.318$ mb$^{1/2}$; within the context of the
formulae in this section.
The spectrum of diffractive masses $M$ is taken to begin
0.28 GeV $\approx 2 m_{\pi}$ above the mass of the respective
incoming particle and extend to the kinematical limit. The simple
$\d M^2 / M^2$ form is modified by the mass-dependence in the
diffractive slopes and in the $F_{\mrm{sd}}$ and $F_{\mrm{dd}}$
factors.
The slope parameters are assumed to be
\begin{eqnarray}
B_{\mrm{sd}(XB)}(s) & = & 2b_B + 2\alpha' \ln\left(\frac{s}{M^2}\right)
~, \nonumber \\
B_{\mrm{sd}(AX)}(s) & = & 2b_A + 2\alpha' \ln\left(\frac{s}{M^2}\right)
~, \nonumber \\
B_{\mrm{dd}}(s) & = & 2\alpha' \ln\left(e^4 + \frac{s s_0}{M_1^2 M_2^2}
\right) ~.
\end{eqnarray}
Here $\alpha' = 0.25$ GeV$^{-2}$ and conventionally $s_0$ is picked as
$s_0 = 1 / \alpha'$. The term $e^4$ in $B_{\mrm{dd}}$ is added by hand
to avoid a breakdown of the standard expression for large values of
$M_1^2 M_2^2$. The $b_{A,B}$ terms protect $B_{\mrm{sd}}$ from breaking
down; however a minimum value of 2 is still explicitly required for
$B_{\mrm{sd}}$, which comes into play e.g. for a $\mathrm{J}/\psi$ state
(as part of a VMD photon beam).
The kinematical range in $t$ depends on all the masses of the
problem. In terms of the scaled variables $\mu_1 = m_A^2/s$,
$\mu_2 = m_B^2/s$, $\mu_3 = M_{(1)}^2/s$ ($=m_A^2/s$ when $A$
scatters elastically), $\mu_4 = M_{(2)}^2/s$ ($=m_B^2/s$ when $B$
scatters elastically), and the combinations
\begin{eqnarray}
C_1 & = & 1 - (\mu_1 + \mu_2 + \mu_3 + \mu_4) +
(\mu_1 - \mu_2) (\mu_3 - \mu_4) ~, \nonumber \\
C_2 & = & \sqrt{(1 - \mu_1 -\mu_2)^2 - 4 \mu_1 \mu_2} \,
\sqrt{(1 - \mu_3 - \mu_4)^2 - 4 \mu_3 \mu_4} ~, \nonumber \\
C_3 & = & (\mu_3 - \mu_1) (\mu_4 - \mu_2) +
(\mu_1 + \mu_4 - \mu_2 - \mu_3) (\mu_1 \mu_4 - \mu_2 \mu_3) ~,
\end{eqnarray}
one has $t_{\mathrm{min}} < t < t_{\mathrm{max}}$ with
\begin{eqnarray}
t_{\mathrm{min}} & = & - \frac{s}{2} (C_1 + C_2) ~, \nonumber \\
t_{\mathrm{max}} & = & - \frac{s}{2} (C_1 - C_2)
= - \frac{s}{2} \, \frac{4C_3}{C_1 + C_2}
= \frac{s^2 C_3}{t_{\mathrm{min}}} ~.
\end{eqnarray}
The Regge formulae above for single- and double-diffractive events
are supposed to hold in certain asymptotic regions of the total phase
space. Of course, there will be diffraction also outside these
restrictive regions. Lacking a theory which predicts differential cross
sections at arbitrary $t$ and $M^2$ values, the Regge formulae are used
everywhere, but fudge factors are introduced in order to obtain
`sensible' behaviour in the full phase space. These factors are:
\begin{eqnarray}
F_{\mrm{sd}} & = & \left( 1 - \frac{M^2}{s} \right)
\left( 1 + \frac{c_{\mrm{res}} \, M_{\mrm{res}}^2}
{M_{\mrm{res}}^2 + M^2} \right) ~, \nonumber \\
F_{\mrm{dd}} & = &
\left( 1 - \frac{\left( M_1 + M_2 \right)^2}{s} \right)
\left( \frac{s\, m_{\mathrm{p}}^2}{ s\, m_{\mathrm{p}}^2 + M_1^2\, M_2^2} \right)
\nonumber \\
& \times &
\left( 1 + \frac{c_{\mrm{res}} \, M_{\mrm{res}}^2}
{M_{\mrm{res}}^2 + M_1^2} \right)
\left( 1 + \frac{c_{\mrm{res}} \, M_{\mrm{res}}^2}
{M_{\mrm{res}}^2 + M_2^2} \right) ~.
\end{eqnarray}
The first factor in either expression suppresses production close to
the kinematical limit. The second factor in $F_{dd}$ suppresses
configurations where the two diffractive systems overlap in rapidity
space. The final factors give an enhancement of the low-mass region,
where a resonance structure is observed in the data. Clearly a more
detailed modelling would have to be based on a set of exclusive states
rather than on this smeared-out averaging procedure. A reasonable fit
to $\mathrm{p}\p / \overline{\mathrm{p}}\mathrm{p}$ data is obtained for
$c_{\mrm{res}} = 2$ and $M_{\mrm{res}} = 2$~GeV,
for an arbitary particle $A$ which is diffractively excited
we use $M_{\mrm{res}}^A = m_A - m_{\mathrm{p}} + 2$~GeV.
The diffractive cross-section formulae above have been integrated
for a set of c.m. energies, starting at 10 GeV, and the results have
been parametrized. The form of these parametrizations is given in
ref. \cite{Sch94}, with explicit numbers for the $\mathrm{p}\p/\overline{\mathrm{p}}\mathrm{p}$
case. {\tsc{Pythia}} also contains similar parametrizations for
$\pi\mathrm{p}$ (assumed to be same as $\rho\mathrm{p}$ and $\omega\mathrm{p}$),
$\phi\mathrm{p}$, $\mathrm{J}/\psi\mathrm{p}$, $\rho\rho$ ($\pi\pi$ etc.), $\rho\phi$,
$\rho\mathrm{J}/\psi$, $\phi\phi$, $\phi\mathrm{J}/\psi$ and $\mathrm{J}/\psi\Jpsi$.
The processes above do not obey the ordinary event mixing strategy.
First of all, since their total cross sections are known, it is
possible to pick the appropriate process from the start, and then
remain with that choice. In other words, if the selection of
kinematical variables fails, one would not go back and pick a new
process, the way it was done in section \ref{sss:mixingproc}.
Second, it is not possible to impose any cuts or restrain allowed
incoming or outgoing flavours: if not additional information were to
be provided, it would make the whole scenario ill-defined.
Third, it is not recommended to mix generation of these processes
with that of any of the other ones: normally the other processes
have so small cross sections that they would almost never be
generated anyway. (We here exclude the cases of `underlying events'
and `pile-up events', where mixing is provided for, and even is a
central part of the formalism, see sections \ref{ss:multint} and
\ref{ss:pileup}.)
Once the cross-section parametrizations has been used to pick one
of the processes, the variables $t$ and $M$ are selected according
to the formulae given above.
A $\rho^0$ formed by $\gamma \to \rho^0$ in elastic or diffractive
scattering is polarized, and therefore its decay angular distribution
in $\rho^0 \to \pi^+ \pi^-$ is taken to be proportional to
$\sin^2 \theta$, where the reference axis is given by the $\rho^0$
direction of motion.
A light diffractive system, with a mass less than 1 GeV above the
mass of the incoming particle, is allowed to decay isotropically into
a two-body state. Single-resonance diffractive states, such as a
$\Delta^+$, are therefore not explicily generated, but are assumed
described in an average, smeared-out sense.
A more massive diffractive system is subsequently treated
as a string with the quantum numbers of the original hadron. Since the
exact nature of the pomeron exchanged between the hadrons is unknown,
two alternatives are included. In the first, the pomeron is assumed to
couple to (valence) quarks, so that the string is stretched directly
between the struck quark and the remnant diquark (antiquark) of the
diffractive state. In the second, the interaction is rather with a
gluon, giving rise to a `hairpin' configuration in which the string
is stretched from a quark to a gluon and then back to a diquark
(antiquark). Both of these scenarios could be present in the data;
the default choice is to mix them in equal proportions.
There is experimental support for more complicated scenarios
\cite{Ing85}, wherein the pomeron has a partonic substructure,
which e.g. can lead to high-$p_{\perp}$ jet production in the diffractive
system. The full machinery, wherein a pomeron spectrum is folded
with a pomeron-proton hard interaction, is not available in {\tsc{Pythia}}.
\subsubsection{Photoproduction and $\gamma\gamma$ physics}
\label{sss:photoprod}
The photoproduction part is still under active development. Currently
only interactions between a hadron and a real photon have been studied
in detail. $\gamma\gamma$ physics is under study \cite{Sch94a}, and
is now preliminarily included for real photons. Deep inelastic
scattering on a real photon is also preliminarily included.
In the future it is hoped to add interactions of mildly virtual photons
(the transition region between real photons and deep inelastic
scattering).
The total $\gamma\mathrm{p}$ and $\gamma\gamma$ cross sections can again be
parametrized in a
form like eq.~(\ref{pg:sigtotpomreg}), which is not so obvious since
the photon has more complicated structure than an ordinary hadron.
In fact, the structure is still not so well understood. The model we
outline is the one studied by Schuler and Sj\"ostrand
\cite{Sch93,Sch93a}. In this model the physical photon is represented
by
\begin{equation}
| \gamma \rangle = \sqrt{Z_3} \, | \gamma_B \rangle +
\sum_{V=\rho^0,\omega,\phi,\mathrm{J}/\psi} \frac{e}{f_V} \, | V \rangle +
\frac{e}{f_{\mathrm{q}\overline{\mathrm{q}}}} \, | \mathrm{q}\overline{\mathrm{q}} \rangle +
\sum_{\ell=\mathrm{e},\mu,\tau} \frac{e}{f_{\ell\ell}} \,
| \ell^+ \ell^- \rangle ~.
\label{pg:gammadecompo}
\end{equation}
By virtue of this superposition, one is led to a model of $\gamma\mathrm{p}$
interactions, where three different kinds of events may be
distinguished:
\begin{Itemize}
\item Direct events, wherein the bare photon $| \gamma_B \rangle$
interacts directly with a parton from the proton. The process is
perturbatively calculable, and no parton distributions of the photon
are involved. The typical event structure is two high-$p_{\perp}$ jets and
a proton remnant, while the photon does not leave behind any
remnant.
\item VMD events, in which the photon fluctuates into a vector meson,
predominantly a $\rho^0$. All the event classes known from ordinary
hadron--hadron interactions may thus occur here, such as elastic,
diffractive, low-$p_{\perp}$ and high-$p_{\perp}$ events. For the latter, one may
define (VMD) parton distributions of the photon, and the photon also
leaves behind a beam remnant. This remnant is smeared in transverse
momentum by a typical `primordial $k_{\perp}$' of a few hundred MeV.
\item Anomalous events, in which the photon fluctuates into a $\mathrm{q}\overline{\mathrm{q}}$
pair of larger virtuality than in the VMD class. This process is
perturbatively calculable, as is the subsequent QCD evolution. It
gives rise to the so-called anomalous part of the parton distributions
ofthe photon, whence the name for the class. It is assumed that only
high-$p_{\perp}$ events may occur. Either the $\mathrm{q}$ or the $\overline{\mathrm{q}}$
plays the r\^ole of a beam remnant, but this remnant has a larger
$p_{\perp}$ than in the VMD case, related to the virtuality of the
$\gamma \leftrightarrow \mathrm{q}\overline{\mathrm{q}}$ fluctuation.
\end{Itemize}
The $| \ell^+ \ell^- \rangle$ states can only interact strongly with
partons inside the hadron at higher orders, and can therefore be
neglected.
In order that the above classification is smooth and free of double
counting, one has to introduce scales that separate the three
components. The main one is $p_0$, which separates the low-mass
vector meson region from the high-mass $| \mathrm{q}\overline{\mathrm{q}} \rangle$ one,
$p_0 \approx m_{\phi}/2 \approx 0.5$ GeV. Since it is the same
$\gamma\mathrm{q}\overline{\mathrm{q}}$ vertex that is responsible for the bare $\gamma \mathrm{p}$
interactions, $p_0$ is also the lower cut-off of the
photon--parton cross sections. In addition, a $p_{\perp\mathrm{min}}$ cut-off is
needed to separate low-$p_{\perp}$ and high-$p_{\perp}$ physics; see section
\ref{ss:multint}. As it turns out, somewhat different $p_{\perp\mathrm{min}}$ values
are needed for the VMD and anomalous parts; at least qualitatively
this can be understood in terms of different sizes of the
wave functions.
The VMD and anomalous events are together called resolved ones.
In terms of high-$p_{\perp}$ jet production, the VMD and anomalous
contributions can be combined into a total resolved one, and the
same for parton-distribution functions. However, the two classes
differ in the structure of the underlying event and in the
appearance of soft processes.
In terms of cross sections, eq.\ (\ref{pg:gammadecompo}) corresponds
to
\begin{equation}
\sigma_{\mrm{tot}}^{\gamma\mathrm{p}}(s) = \sigma_{\mrm{dir}}^{\gamma\mathrm{p}}(s) +
\sigma_{\mrm{VMD}}^{\gamma\mathrm{p}}(s) + \sigma_{\mrm{anom}}^{\gamma\mathrm{p}}(s) ~.
\label{pg:gammacrossdecompo}
\end{equation}
The direct cross section is, to lowest order, the perturbative cross
section for the two processes $\gamma\mathrm{q} \to \mathrm{q}\mathrm{g}$ and
$\gamma\mathrm{g} \to \mathrm{q}\overline{\mathrm{q}}$, with a lower cut-off $p_{\perp} > p_0$.
Properly speaking, this should be multiplied by the $Z_3$ coefficient,
\begin{equation}
Z_3 = 1 -
\sum_{V=\rho^0,\omega,\phi,\mathrm{J}/\psi} \left( \frac{e}{f_V} \right)^2 -
\left( \frac{e}{f_{\mathrm{q}\overline{\mathrm{q}}}} \right)^2 -
\sum_{\ell=\mathrm{e},\mu,\tau} \left( \frac{e}{f_{\ell\ell}} \right)^2 ~,
\end{equation}
but normally $Z_3$ is so close to unity as to make no difference.
The VMD factor $(e/f_V)^2 = 4\pi\alpha_{\mathrm{em}}/f_V^2$ gives the probability
for the transition $\gamma \to V$. The coefficients $f_V^2/4\pi$ are
determined from data to be (with a non-negligible amount of
uncertainty) 2.20 for $\rho^0$, 23.6 for $\omega$, 18.4 for $\phi$
and 11.5 for $\mathrm{J}/\psi$. Together these numbers imply that the photon
can be found in a VMD state about 0.4\% of the time, dominated by the
$\rho^0$ contribution. All the properties of the VMD interactions
can be obtained by appropriately scaling down $V\mathrm{p}$ physics
predictions. Thus the whole machinery developed in the previous
subsection for hadron--hadron interactions is directly applicable.
Also parton distributions of the VMD component inside the photon
are obtained by suitable rescaling.
The contribution from the `anomalous' high-mass fluctuations depends
on the typical scale $\mu$ of the interaction
\begin{equation}
\left( \frac{e}{f_{\mathrm{q}\overline{\mathrm{q}}}} \right)^2 \approx
\frac{\alpha_{\mathrm{em}}}{2\pi} \, \frac{N_C}{3} \,
\left( 2 \sum_{\mathrm{q}} e_{\mathrm{q}}^2 \right)
\ln\left( \frac{\mu^2}{p_0^2} \right) ~,
\label{anomintfirst}
\end{equation}
where $N_C = 3$ and $\mathrm{q}$ runs over the quarks that can be taken
massless compared with $\mu$. The logarithmic increase with $\mu$
implies that the anomalous contribution to the total photoproduction
cross section ($\mu \sim m_V$) is less important than
that to high-$p_{\perp}$ jet production ($\mu \sim p_{\perp}$). To first
approximation, therefore only perturbative jet production above some
$p_{\perp\mathrm{min}}$ scale is considered. This includes the standard QCD
parton--parton scattering processes, with anomalous-photon parton
distributions that are fully perturbatively calculable \cite{Sch95}.
In order to satisfy the equality in eq.~(\ref{pg:gammacrossdecompo}),
with the total cross section known and the direct and VDM
contributions already fixed, a behaviour roughly like
\begin{equation}
p_{\perp\mathrm{min}}^{\mrm{anom}}(s) = 0.70 + 0.17 \log^2(1.+0.05\sqrt{s})
\end{equation}
is needed over the HERA energy range. This is to be seen entirely
as a pragmatic parametrization, not be given any fundamental
interpretation. It is based on SaS set~1D, another set might well
require a somewhat different form.
In $\gamma\gamma$ physics \cite{Sch94a}, the superposition in
eq.~(\ref{pg:gammadecompo}) applies separately for each of the two
incoming photons. In total there are therefore $3 \times 3 = 9$
combinations. However, trivial symmetry reduces this to six distinct
classes, written in terms of the total cross section
(cf. eq.~(\ref{pg:gammacrossdecompo})) as
\begin{eqnarray}
\sigma_{\mrm{tot}}^{\gamma\gamma}(s) & = &
\sigma_{\mrm{dir}\times\mrm{dir}}^{\gamma\gamma}(s) +
\sigma_{\mrm{VMD}\times\mrm{VMD}}^{\gamma\gamma}(s) +
\sigma_{\mrm{anom}\times\mrm{anom}}^{\gamma\gamma}(s) \nonumber \\
& + & 2 \sigma_{\mrm{dir}\times\mrm{VMD}}^{\gamma\gamma}(s) +
2 \sigma_{\mrm{dir}\times\mrm{anom}}^{\gamma\gamma}(s) +
2 \sigma_{\mrm{VMD}\times\mrm{anom}}^{\gamma\gamma}(s) ~.
\label{pg:gagacrossdecompo}
\end{eqnarray}
A parametrization of the total $\gamma\gamma$ cross section and
comments on its subdivision into the six classes is found in
\cite{Sch94a}.
The six different kinds of $\gamma\gamma$ events are thus:
\begin{Itemize}
\item The direct$\times$direct events, which correspond to the
subprocess $\gamma\gamma \to \mathrm{q}\overline{\mathrm{q}}$ (or $\ell^+\ell^-$).
The typical event structure is two high-$p_{\perp}$ jets and no beam
remnants. The lower cut-off is $p_{\perp} > p_0$.
\item The VMD$\times$VMD events, which have the same properties as
the VMD $\gamma\mathrm{p}$ events. There are four by four combinations of
the two incoming vector mesons, with one VMD factor for each meson.
\item The anomalous$\times$anomalous events, wherein each photon
fluctuates into a $\mathrm{q}\overline{\mathrm{q}}$ pair of larger virtuality than in the
VMD class. One parton of each pair gives a beam remnant, whereas
the other (or a daughter parton thereof) participates in a high-$p_{\perp}$
scattering, with $p_{\perp} > p_{\perp\mathrm{min}}^{\mrm{anom}}$.
\item The direct$\times$VMD events, which have the same properties as
the direct $\gamma\mathrm{p}$ events.
\item The direct$\times$anomalous events, in which a bare photon
interacts with a parton from the anomalous photon. The lower cut-off
for the hard scattering is given by $p_{\perp\mathrm{min}}^{\mrm{anom}}$.
The typical structure is then two high-$p_{\perp}$ jets and a beam remnant.
\item The VMD$\times$anomalous events, which have the same properties
as the anomalous $\gamma\mathrm{p}$ events.
\end{Itemize}
In much of the literature, where a coarser classification us used,
our direct$\times$direct is called direct, our direct$\times$VMD
and direct$\times$anomalous is called 1-resolved since they both
involve one resolved photon which gives a beam remnant,
and the rest are called 2-resolves since both photons are resolved
and give beam remnants.
\clearpage
\section{Physics Processes in PYTHIA}
\label{s:pytproc}
In this section we enumerate the physics processes that are available
in {\tsc{Pythia}}, introducing the ISUB code that can be used to select
desired processes. A number of comments are made about the
physics scenarios involved, in particular with respect to
underlying assumptions and domain of validity. The section closes
with a survey of interesting processes by machine.
\subsection{The Process Classification Scheme}
\label{ss:ISUBcode}
A wide selection of fundamental $2 \to 1$ and $2 \to 2$ tree
processes of the Standard Model (electroweak and strong) has been
included in {\tsc{Pythia}}, and slots are provided for many more, not yet
implemented. In addition, a few `minimum-bias'-type processes
(like elastic scattering), loop graphs, box graphs, $2 \to 3$ tree
graphs and some non-Standard Model processes are included. The
classification is not always unique. A process that proceeds only
via an $s$-channel state is classified as a $2 \to 1$ process
(e.g. $\mathrm{q} \overline{\mathrm{q}} \to \gamma^* / \Z^0 \to \e^+\e^-$), but a $2 \to 2$
cross section may well have contributions from $s$-channel diagrams
($\mathrm{g} \mathrm{g} \to \mathrm{g} \mathrm{g}$ obtains contributions from
$\mathrm{g} \mathrm{g} \to \mathrm{g}^* \to \mathrm{g} \mathrm{g}$). Also, in the program, $2 \to 1$ and
$2 \to 2$ graphs may sometimes be folded with two $1 \to 2$
splittings to form effective $2 \to 3$ or $2 \to 4$ processes
($\mathrm{W}^+ \mathrm{W}^- \to \H^0$ is folded with $\mathrm{q} \to \mathrm{q}'' \mathrm{W}^+$ and
$\mathrm{q}' \to \mathrm{q}''' \mathrm{W}^-$ to give $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \mathrm{q}''' \H^0$).
It is possible to select a combination of subprocesses to simulate,
and also afterwards to know which subprocess was actually selected in
each event. For this purpose, all subprocesses are numbered according
to an ISUB code. The list of possible codes is given in Tables
\ref{t:procone}, \ref{t:proctwo}, \ref{t:procthree} and
\ref{t:procfour}. Only processes marked with a `+' sign in the first
column have been implemented in the program to date. Although
ISUB codes were originally designed in a logical fashion,
we must admit that subsequent developments of the program have tended
to obscure the structure. For instance, the process numbers for Higgs
production are spread out, in part as a consequence of the original
classification, in part because further production mechanisms have been
added one at a time, in whatever free slots could be found. At some
future date the subprocess list will therefore be reorganized.
In the thematic descriptions that follow the main tables, the
processes of interest are repeated in a more logical order. If you
want to look for a specific process, it will be easier
to find it there.
\begin{table}[pt]
\caption{Subprocess codes, part 1. First column is `+' for processes
implemented and blank for those that are only foreseen. Second is
the subprocess number \ttt{ISUB}, and third the description of the
process. The final column gives references from which the
cross sections have been obtained. See text for further information.
\protect\label{t:procone} }
\begin{center}
\begin{tabular}{|c|r|l|l|@{\protect\rule{0mm}{\tablinsep}}}
\hline
In & No. & Subprocess & Reference \\
\hline
& & a) $2 \to 1$, tree & \\
+ & 1 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma^* / \Z^0$ & \cite{Eic84} \\
+ & 2 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+$ & \cite{Eic84} \\
+ & 3 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \H^0$ & \cite{Eic84} \\
& 4 & $\gamma \mathrm{W}^+ \to \mathrm{W}^+$ & \\
+ & 5 & $\mathrm{Z}^0 \mathrm{Z}^0 \to \H^0$ & \cite{Eic84,Cha85} \\
& 6 & $\mathrm{Z}^0 \mathrm{W}^+ \to \mathrm{W}^+$ & \\
& 7 & $\mathrm{W}^+ \mathrm{W}^- \to \mathrm{Z}^0$ & \\
+ & 8 & $\mathrm{W}^+ \mathrm{W}^- \to \H^0$ & \cite{Eic84,Cha85} \\
\hline
& & b) $2 \to 2$, tree & \\
+ & 10 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j$ (QFD) & \cite{Ing87b} \\
+ & 11 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j$ (QCD) &
\cite{Com77,Ben84,Eic84,Chi90} \\
+ & 12 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{f}_k \overline{\mathrm{f}}_k$ &
\cite{Com77,Ben84,Eic84,Chi90} \\
+ & 13 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{g} \mathrm{g}$ & \cite{Com77,Ben84} \\
+ & 14 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{g} \gamma$ & \cite{Hal78,Ben84} \\
+ & 15 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{g} \mathrm{Z}^0$ & \cite{Eic84} \\
+ & 16 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{g} \mathrm{W}^+$ & \cite{Eic84} \\
& 17 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{g} \H^0$ & \\
+ & 18 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma \gamma$ & \cite{Ber84} \\
+ & 19 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma \mathrm{Z}^0$ & \cite{Eic84} \\
+ & 20 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \gamma \mathrm{W}^+$ & \cite{Eic84,Sam91} \\
& 21 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma \H^0$ & \\
+ & 22 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Z}^0 \mathrm{Z}^0$ & \cite{Eic84,Gun86} \\
+ & 23 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{Z}^0 \mathrm{W}^+$ & \cite{Eic84,Gun86} \\
+ & 24 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Z}^0 \H^0$ & \cite{Ber85} \\
+ & 25 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{W}^+ \mathrm{W}^-$ & \cite{Bar94,Gun86} \\
+ & 26 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+ \H^0$ & \cite{Eic84} \\
& 27 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \H^0 \H^0$ & \\
+ & 28 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_i \mathrm{g}$ & \cite{Com77,Ben84} \\
+ & 29 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_i \gamma$ & \cite{Hal78,Ben84} \\
+ & 30 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_i \mathrm{Z}^0$ & \cite{Eic84} \\
+ & 31 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_k \mathrm{W}^+$ & \cite{Eic84} \\
& 32 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_i \H^0$ & \\
+ & 33 & $\mathrm{f}_i \gamma \to \mathrm{f}_i \mathrm{g}$ & \cite{Duk82} \\
+ & 34 & $\mathrm{f}_i \gamma \to \mathrm{f}_i \gamma$ & \cite{Duk82} \\
+ & 35 & $\mathrm{f}_i \gamma \to \mathrm{f}_i \mathrm{Z}^0$ & \cite{Gab86} \\
+ & 36 & $\mathrm{f}_i \gamma \to \mathrm{f}_k \mathrm{W}^+$ & \cite{Gab86} \\
& 37 & $\mathrm{f}_i \gamma \to \mathrm{f}_i \H^0$ & \\
& 38 & $\mathrm{f}_i \mathrm{Z}^0 \to \mathrm{f}_i \mathrm{g}$ & \\
& 39 & $\mathrm{f}_i \mathrm{Z}^0 \to \mathrm{f}_i \gamma$ & \\
& 40 & $\mathrm{f}_i \mathrm{Z}^0 \to \mathrm{f}_i \mathrm{Z}^0$ & \\
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[pt]
\caption{Subprocess codes, part 2. First column is `+' for processes
implemented and blank for those that are only foreseen. Second is
the subprocess number \ttt{ISUB}, and third the description of the
process. The final column gives references from which the
cross sections have been obtained. See text for further information.
\protect\label{t:proctwo} }
\begin{center}
\begin{tabular}{|c|r|l|l|@{\protect\rule{0mm}{\tablinsep}}}
\hline
In & No. & Subprocess & Reference \\
\hline
& & b) $2 \to 2$, tree (cont'd) & \\
& 41 & $\mathrm{f}_i \mathrm{Z}^0 \to \mathrm{f}_k \mathrm{W}^+$ & \\
& 42 & $\mathrm{f}_i \mathrm{Z}^0 \to \mathrm{f}_i \H^0$ & \\
& 43 & $\mathrm{f}_i \mathrm{W}^+ \to \mathrm{f}_k \mathrm{g}$ & \\
& 44 & $\mathrm{f}_i \mathrm{W}^+ \to \mathrm{f}_k \gamma$ & \\
& 45 & $\mathrm{f}_i \mathrm{W}^+ \to \mathrm{f}_k \mathrm{Z}^0$ & \\
& 46 & $\mathrm{f}_i \mathrm{W}^+ \to \mathrm{f}_k \mathrm{W}^+$ & \\
& 47 & $\mathrm{f}_i \mathrm{W}^+ \to \mathrm{f}_k \H^0$ & \\
& 48 & $\mathrm{f}_i \H^0 \to \mathrm{f}_i \mathrm{g}$ & \\
& 49 & $\mathrm{f}_i \H^0 \to \mathrm{f}_i \gamma$ & \\
& 50 & $\mathrm{f}_i \H^0 \to \mathrm{f}_i \mathrm{Z}^0$ & \\
& 51 & $\mathrm{f}_i \H^0 \to \mathrm{f}_k \mathrm{W}^+$ & \\
& 52 & $\mathrm{f}_i \H^0 \to \mathrm{f}_i \H^0$ & \\
+ & 53 & $\mathrm{g} \mathrm{g} \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \cite{Com77,Ben84} \\
+ & 54 & $\mathrm{g} \gamma \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \cite{Duk82} \\
& 55 & $\mathrm{g} \mathrm{Z}^0 \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \\
& 56 & $\mathrm{g} \mathrm{W}^+ \to \mathrm{f}_k \overline{\mathrm{f}}_l$ & \\
& 57 & $\mathrm{g} \H^0 \to \mathrm{f}_k \overline{\mathrm{f}}_l$ & \\
+ & 58 & $\gamma \gamma \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \cite{Bar90} \\
& 59 & $\gamma \mathrm{Z}^0 \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \\
& 60 & $\gamma \mathrm{W}^+ \to \mathrm{f}_k \overline{\mathrm{f}}_l$ & \\
& 61 & $\gamma \H^0 \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \\
& 62 & $\mathrm{Z}^0 \mathrm{Z}^0 \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \\
& 63 & $\mathrm{Z}^0 \mathrm{W}^+ \to \mathrm{f}_k \overline{\mathrm{f}}_l$ & \\
& 64 & $\mathrm{Z}^0 \H^0 \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \\
& 65 & $\mathrm{W}^+ \mathrm{W}^- \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \\
& 66 & $\mathrm{W}^+ \H^0 \to \mathrm{f}_k \overline{\mathrm{f}}_l$ & \\
& 67 & $\H^0 \H^0 \to \mathrm{f}_k \overline{\mathrm{f}}_k$ & \\
+ & 68 & $\mathrm{g} \mathrm{g} \to \mathrm{g} \mathrm{g}$ & \cite{Com77,Ben84} \\
+ & 69 & $\gamma \gamma \to \mathrm{W}^+ \mathrm{W}^-$ & \cite{Kat83} \\
+ & 70 & $\gamma \mathrm{W}^+ \to \mathrm{Z}^0 \mathrm{W}^+$ & \cite{Kun87} \\
+ & 71 & $\mathrm{Z}^0 \mathrm{Z}^0 \to \mathrm{Z}^0 \mathrm{Z}^0$ (longitudinal) & \cite{Abb87} \\
+ & 72 & $\mathrm{Z}^0 \mathrm{Z}^0 \to \mathrm{W}^+ \mathrm{W}^-$ (longitudinal) & \cite{Abb87} \\
+ & 73 & $\mathrm{Z}^0 \mathrm{W}^+ \to \mathrm{Z}^0 \mathrm{W}^+$ (longitudinal) & \cite{Dob91} \\
& 74 & $\mathrm{Z}^0 \H^0 \to \mathrm{Z}^0 \H^0$ & \\
& 75 & $\mathrm{W}^+ \mathrm{W}^- \to \gamma \gamma$ & \\
+ & 76 & $\mathrm{W}^+ \mathrm{W}^- \to \mathrm{Z}^0 \mathrm{Z}^0$ (longitudinal) & \cite{Ben87b} \\
+ & 77 & $\mathrm{W}^+ \mathrm{W}^{\pm} \to \mathrm{W}^+ \mathrm{W}^{\pm}$ (longitudinal) &
\cite{Dun86,Bar90a} \\
& 78 & $\mathrm{W}^+ \H^0 \to \mathrm{W}^+ \H^0$ & \\
& 79 & $\H^0 \H^0 \to \H^0 \H^0$ & \\
+ & 80 & $\mathrm{q}_i \gamma \to \mathrm{q}_k \pi^{\pm}$ & \cite{Bag82} \\
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[pt]
\caption{Subprocess codes, part 3. First column is `+' for processes
implemented and blank for those that are only foreseen. Second is
the subprocess number \ttt{ISUB}, and third the description of the
process. The final column gives references from which the
cross sections have been obtained. See text for further information.
\protect\label{t:procthree} }
\begin{center}
\begin{tabular}{|c|r|l|l|@{\protect\rule{0mm}{\tablinsep}}}
\hline
In & No. & Subprocess & Reference \\
\hline
& & c) $2 \to 2$, tree, massive final quarks & \\
+ & 81 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ & \cite{Com79} \\
+ & 82 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ & \cite{Com79} \\
+ & 83 & $\mathrm{q}_i \mathrm{f}_j \to \mathrm{Q}_k \mathrm{f}_l$ & \cite{Dic86} \\
+ & 84 & $\mathrm{g} \gamma \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ & \cite{Fon81} \\
+ & 85 & $\gamma \gamma \to \mathrm{F}_k \overline{\mathrm{F}}_k$ & \cite{Bar90} \\
+ & 86 & $\mathrm{g} \mathrm{g} \to \mathrm{J}/\psi \mathrm{g}$ & \cite{Bai83} \\
+ & 87 & $\mathrm{g} \mathrm{g} \to \chi_{0 \c} \mathrm{g}$ & \cite{Gas87} \\
+ & 88 & $\mathrm{g} \mathrm{g} \to \chi_{1 \c} \mathrm{g}$ & \cite{Gas87} \\
+ & 89 & $\mathrm{g} \mathrm{g} \to \chi_{2 \c} \mathrm{g}$ & \cite{Gas87} \\
\hline
& & d) `minimum bias' & \\
+ & 91 & elastic scattering & \cite{Sch94} \\
+ & 92 & single diffraction ($AB \to XB$) & \cite{Sch94} \\
+ & 93 & single diffraction ($AB \to AX$) & \cite{Sch94} \\
+ & 94 & double diffraction & \cite{Sch94} \\
+ & 95 & low-$p_{\perp}$ production & \cite{Sjo87} \\
\hline
& & e) $2 \to 1$, loop & \\
& 101 & $\mathrm{g} \mathrm{g} \to \mathrm{Z}^0$ & \\
+ & 102 & $\mathrm{g} \mathrm{g} \to \H^0$ & \cite{Eic84} \\
+ & 103 & $\gamma \gamma \to \H^0$ & \cite{Dre89} \\
\hline
& & f) $2 \to 2$, box & \\
+ & 110 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma \H^0$ & \cite{Ber85a} \\
+ & 111 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{g} \H^0$ & \cite{Ell88} \\
+ & 112 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_i \H^0$ & \cite{Ell88} \\
+ & 113 & $\mathrm{g} \mathrm{g} \to \mathrm{g} \H^0$ & \cite{Ell88} \\
+ & 114 & $\mathrm{g} \mathrm{g} \to \gamma \gamma$ & \cite{Con71,Ber84,Dic88} \\
+ & 115 & $\mathrm{g} \mathrm{g} \to \mathrm{g} \gamma$ & \cite{Con71,Ber84,Dic88} \\
& 116 & $\mathrm{g} \mathrm{g} \to \gamma \mathrm{Z}^0$ & \\
& 117 & $\mathrm{g} \mathrm{g} \to \mathrm{Z}^0 \mathrm{Z}^0$ & \\
& 118 & $\mathrm{g} \mathrm{g} \to \mathrm{W}^+ \mathrm{W}^-$ & \\
& 119 & $\gamma \gamma \to \mathrm{g} \mathrm{g}$ & \\
\hline
& & g) $2 \to 3$, tree & \\
+ & 121 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \H^0$ & \cite{Kun84} \\
+ & 122 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \H^0$ & \cite{Kun84} \\
+ & 123 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j \H^0$ ($\mathrm{Z} \mathrm{Z}$ fusion) &
\cite{Cah84} \\
+ & 124 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_k \mathrm{f}_l \H^0$ ($\mathrm{W}^+\mathrm{W}^-$ fusion) &
\cite{Cah84} \\
+ & 131 & $\mathrm{g} \mathrm{g} \to \mathrm{Z}^0 \mathrm{Q}_k \br{\mathrm{Q}}_k$ & \cite{Eij90} \\
\hline
\end{tabular}
\end{center}
\end{table}
\begin{table}[pt]
\caption{Subprocess codes, part 4. First column is `+' for processes
implemented and blank for those that are only foreseen. Second is
the subprocess number \ttt{ISUB}, and third the description of the
process. The final column gives references from which the
cross sections have been obtained. See text for further information.
\protect\label{t:procfour} }
\begin{center}
\begin{tabular}{|c|r|l|l|@{\protect\rule{0mm}{\tablinsep}}}
\hline
In & No. & Subprocess & Reference \\
\hline
& & h) non-Standard Model, $2 \to 1$ & \\
+ & 141 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma/\mathrm{Z}^0/\mathrm{Z}'^0$ & \cite{Alt89} \\
+ & 142 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}'^+$ & \cite{Alt89} \\
+ & 143 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \H^+$ & \cite{Gun87} \\
+ & 144 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{R}$ & \cite{Ben85a} \\
+ & 145 & $\mathrm{q}_i \ell_j \to \L_{\mathrm{Q}}$ & \cite{Wud86} \\
+ & 147 & $\d \mathrm{g} \to \d^*$ & \cite{Bau90} \\
+ & 148 & $\u \mathrm{g} \to \u^*$ & \cite{Bau90} \\
+ & 149 & $\mathrm{g} \mathrm{g} \to \eta_{\mrm{techni}}$ & \cite{Eic84,App92} \\
+ & 151 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \H'^0$ & \cite{Eic84} \\
+ & 152 & $\mathrm{g} \mathrm{g} \to \H'^0$ & \cite{Eic84} \\
+ & 153 & $\gamma \gamma \to \H'^0$ & \cite{Dre89} \\
+ & 156 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{A}^0$ & \cite{Eic84} \\
+ & 157 & $\mathrm{g} \mathrm{g} \to \mathrm{A}^0$ & \cite{Eic84} \\
+ & 158 & $\gamma \gamma \to \mathrm{A}^0$ & \cite{Dre89} \\
\hline
& & i) non-Standard Model, $2 \to 2$ and $2 \to 3$ & \\
+ & 161 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_k \H^+$ & \cite{Bar88} \\
+ & 162 & $\mathrm{q} \mathrm{g} \to \ell \L_{\mathrm{Q}}$ & \cite{Hew88} \\
+ & 163 & $\mathrm{g} \mathrm{g} \to \L_{\mathrm{Q}} \br{\L}_{\mathrm{Q}}$ &
\cite{Hew88,Eic84} \\
+ & 164 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \L_{\mathrm{Q}} \br{\L}_{\mathrm{Q}}$ &
\cite{Hew88} \\
+ & 165 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{f}_k \overline{\mathrm{f}}_k$ (via $\gamma^* / \Z^0$) &
\cite{Eic84,Lan91} \\
+ & 166 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{f}_k \overline{\mathrm{f}}_l$ (via $\mathrm{W}^{\pm}$) &
\cite{Eic84,Lan91} \\
+ & 167 & $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \d^*$ & \cite{Bau90} \\
+ & 168 & $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \u^*$ & \cite{Bau90} \\
+ & 171 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Z}^0 \H'^0$ & \cite{Eic84} \\
+ & 172 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+ \H'^0$ & \cite{Eic84} \\
+ & 173 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j \H'^0$ ($\mathrm{Z} \mathrm{Z}$ fusion) &
\cite{Cah84} \\
+ & 174 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_k \mathrm{f}_l \H'^0$ ($\mathrm{W}^+\mathrm{W}^-$ fusion) &
\cite{Cah84} \\
+ & 176 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Z}^0 \mathrm{A}^0$ & \cite{Eic84} \\
+ & 177 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+ \mathrm{A}^0$ & \cite{Eic84} \\
+ & 178 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j \mathrm{A}^0$ ($\mathrm{Z} \mathrm{Z}$ fusion) &
\cite{Cah84} \\
+ & 179 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_k \mathrm{f}_l \mathrm{A}^0$ ($\mathrm{W}^+\mathrm{W}^-$ fusion) &
\cite{Cah84} \\
+ & 181 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \H'^0$ & \cite{Kun84} \\
+ & 182 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \H'^0$ & \cite{Kun84} \\
+ & 186 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \mathrm{A}^0$ & \cite{Kun84} \\
+ & 187 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \mathrm{A}^0$ & \cite{Kun84} \\
\hline
\end{tabular}
\end{center}
\end{table}
In the following, $\mathrm{f}_i$ represents a fundamental fermion of flavour
$i$, i.e. $\d$, $\u$, $\mathrm{s}$, $\c$, $\b$, $\t$, $\mathrm{l}$,
$\mathrm{h}$, $\mathrm{e}^-$, $\nu_{\mathrm{e}}$, $\mu^-$, $\nu_{\mu}$, $\tau^-$,
$\nu_{\tau}$, $\chi^-$ or $\nu_{\chi}$. A corresponding antifermion
is denoted by $\overline{\mathrm{f}}_i$. In several cases, some classes of fermions
are explicitly excluded, since they do not couple to the $\mathrm{g}$ or
$\gamma$ (no $\e^+\e^- \to \mathrm{g} \mathrm{g}$, e.g.). When processes have only been
included for quarks, while leptons might also have been possible,
the notation $\mathrm{q}_i$ is used. A lepton is denoted by $\ell$; in a few
cases neutrinos are also lumped under this heading.
In processes where fermion masses are
explicitly included in the matrix elements, an $\mathrm{F}$ is used to denote
an arbitrary fermion and a $\mathrm{Q}$ a quark. Flavours appearing already in
the initial state are denoted by indices $i$ and $j$, whereas new
flavours in the final state are denoted by $k$ and $l$.
Charge-conjugate channels are always assumed included as well (where
separate), and processes involving a $\mathrm{W}^+$ also imply those involving
a $\mathrm{W}^-$. Wherever $\mathrm{Z}^0$ is written, it is understood that $\gamma^*$
and $\gamma^* / \Z^0$ interference should be included as well (with
possibilities to switch off either, if so desired). In some cases
this is not fully implemented, see further below.
Correspondingly, $\mathrm{Z}'^0$ denotes the complete set
$\gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0$ (or some subset of it). Thus the notation
$\gamma$ is only used for a photon on the mass shell.
In the last column of the tables below, references are given to works
from which formulae have been taken. Sometimes these references are to
the original works on the subject, sometimes only to the place where
the formulae are given in the most convenient or accessible form, or
where chance lead us. Apologies to all matrix-element calculators who
are not mentioned. However, remember that this is not a review article
on physics processes, but only a way for readers to know what is
actually found in the program, for better or worse. In several
instances, errata have been obtained from the authors. Often
the formulae given in the literature have been generalized to include
trivial radiative corrections, Breit--Wigner line shapes with
$\hat{s}$-dependent widths (see section \ref{ss:kinemreson}), etc.
The following sections contain some useful comments on the processes
included in the program, grouped by physics interest rather than
sequentially by ISUB or \ttt{MSEL} code (see \ref{ss:PYswitchkin}
for further information on the \ttt{MSEL} code). The different
ISUB and \ttt{MSEL} codes
that can be used to simulate the different groups are given. ISUB
codes within brackets indicate the kind of processes that indirectly
involve the given physics topic, although only as part of a larger
whole. Some obvious examples, such as the possibility to produce jets
in just about any process, are not spelled out in detail.
The text at times contains information on which special switches or
parameters are of particular interest to a given process. All these
switches are described in detail in section \ref{ss:PYswitchpar}, but
are alluded to here so as to provide a more complete picture of the
possibilities available for the different subprocesses. However, the
list of possibilities is certainly not exhausted by the text below.
\subsection{QCD Processes}
In this section we discuss scatterings exclusively between coloured
partons --- a process like $\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q}\overline{\mathrm{q}}$ is also
traditionally called a QCD event, but is here book-kept as
$\gamma^* / \Z^0$ production.
\subsubsection{QCD jets}
\label{sss:QCDjetclass}
\ttt{MSEL} = 1, 2 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
11 & $\mathrm{q}_i \mathrm{q}_j \to \mathrm{q}_i \mathrm{q}_j$ \\
12 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{q}_k \overline{\mathrm{q}}_k$ \\
13 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{g} \mathrm{g}$ \\
28 & $\mathrm{q}_i \mathrm{g} \to \mathrm{q}_i \mathrm{g}$ \\
53 & $\mathrm{g} \mathrm{g} \to \mathrm{q}_k \overline{\mathrm{q}}_k$ \\
68 & $\mathrm{g} \mathrm{g} \to \mathrm{g} \mathrm{g}$ \\
\end{tabular}
No higher-order processes are explicitly included,
nor any higher-order loop corrections to the $2 \to 2$ processes.
However, by initial- and final-state QCD radiation, multijet events
are being generated, starting from the above processes. The shower
rate of multijet production is clearly uncertain
by some amount, especially for well-separated jets.
A string-based fragmentation scheme such as the Lund model needs
cross sections for the different colour flows; these have been
calculated in \cite{Ben84} and differ from the usual calculations by
interference terms of the order $1/N_C^2$. By default, the standard
QCD expressions for the differential cross sections are used. In this
case, the interference terms are distributed on the various colour
flows according to the pole structure of the terms. However, the
interference terms can be excluded, by changing \ttt{MSTP(34)}
As an example, consider subprocess 28, $\mathrm{q} \mathrm{g} \to \mathrm{q} \mathrm{g}$. The total
cross section for this process, obtained by summing and squaring the
Feynman $\hat{s}$-, $\hat{t}$-, and $\hat{u}$-channel graphs, is
\cite{Com77}
\begin{equation}
2 \left( 1 - \frac{\hat{u}\hat{s}}{\hat{t}^2} \right) -
\frac{4}{9} \left( \frac{\hat{s}}{\hat{u}} +
\frac{\hat{u}}{\hat{s}} \right) - 1 ~.
\end{equation}
(An overall factor $\pi \alpha_{\mathrm{s}}^2/\hat{s}^2$ is ignored.)
Using the identity of the Mandelstam variables
for the massless case, $\hat{s} + \hat{t} + \hat{u} = 0$,
this can be rewritten as
\begin{equation}
\frac{\hat{s}^2 + \hat{u}^2}{\hat{t}^2} -
\frac{4}{9} \left( \frac{\hat{s}}{\hat{u}} +
\frac{\hat{u}}{\hat{s}} \right) ~.
\end{equation}
On the other hand, the cross sections for the two possible colour
flows of this subprocess are \cite{Ben84}
\begin{eqnarray}
A: & & \frac{4}{9} \left( 2 \frac{\hat{u}^2}{\hat{t}^2} -
\frac{\hat{u}}{\hat{s}} \right) ~; \nonumber \\
B: & & \frac{4}{9} \left( 2 \frac{\hat{s}^2}{\hat{t}^2} -
\frac{\hat{s}}{\hat{u}} \right) ~.
\end{eqnarray}
Colour configuration $A$ is one in which the original colour
of the $\mathrm{q}$ annihilates with the anticolour of the $\mathrm{g}$,
the $\mathrm{g}$ colour flows through, and a new colour--anticolour is
created between the final $\mathrm{q}$ and $\mathrm{g}$. In colour configuration
$B$, the gluon anticolour flows through, but the $\mathrm{q}$ and $\mathrm{g}$
colours are interchanged. Note that these two colour
configurations have different kinematics dependence.
For \ttt{MSTP(34)=0}, these are the cross sections actually
used.
The sum of the $A$ and $B$ contributions is
\begin{equation}
\frac{8}{9} \frac{\hat{s}^2 + \hat{u}^2}{\hat{t}^2} -
\frac{4}{9} \left( \frac{\hat{s}}{\hat{u}} +
\frac{\hat{u}}{\hat{s}} \right) ~.
\end{equation}
The difference between this expression and that of \cite{Com77},
corresponding to the interference between the two colour-flow
configurations, is then
\begin{equation}
\frac{1}{9} \frac{\hat{s}^2 + \hat{u}^2}{\hat{t}^2} ~,
\end{equation}
which can be naturally divided between colour flows $A$ and $B$:
\begin{eqnarray}
A: & & \frac{1}{9} \frac{\hat{u}^2}{\hat{t}^2} ~; \nonumber \\
B: & & \frac{1}{9} \frac{\hat{s}^2}{\hat{t}^2} ~.
\end{eqnarray}
For \ttt{MSTP(34)=1}, the standard QCD matrix element is therefore
used, with the same relative importance of the two colour configurations
as above. Similar procedures are followed also for the other QCD
subprocesses.
All the matrix elements in this group are for massless quarks
(although final-state quarks are of course put on the mass shell).
As a consequence, cross sections are divergent for $p_{\perp} \to 0$,
and some kind of regularization is required. Normally you
are expected to set the desired $p_{\perp\mathrm{min}}$ value in
\ttt{CKIN(3)}.
The new flavour produced in the annihilation processes (ISUB = 12 and
53) is determined by the flavours allowed for gluon splitting into
quark--antiquark; see switch \ttt{MDME}.
\subsubsection{Heavy flavours}
\label{sss:heavflavclass}
\ttt{MSEL} = 4, 5, 6, 7, 8 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
81 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ \\
82 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ \\
(83) & $\mathrm{q}_i \mathrm{f}_j \to Q_k f_l$ \\
\end{tabular}
The matrix elements in this group differ from the corresponding ones
in the group above in that they correctly take into account the quark
masses. As a consequence, the cross sections are finite for
$p_{\perp} \to 0$. It is therefore not necessary to introduce any special
cuts.
The two first processes that appear here are the dominant lowest-order
graphs in hadron colliders --- a few other graphs will be mentioned
later, such as process 83, which is important for a heavy top.
The choice of flavour to produce is according to a hierarchy of options:
\begin{Enumerate}
\item if \ttt{MSEL=4-8} then the flavour is set by the \ttt{MSEL} value;
\item else if \ttt{MSTP(7)=1-8} then the flavour is set by the
\ttt{MSTP(7)} value;
\item else the flavour is determined by the heaviest flavour allowed for
gluon splitting into quark--antiquark; see switch \ttt{MDME}.
\end{Enumerate}
Note that only one heavy flavour is allowed at a time; if more than one
is turned on in \ttt{MDME}, only the heaviest will be produced (as
opposed to the case for ISUB = 12 and 53 above, where more than one
flavour is allowed simultaneously).
The lowest-order processes listed above just represent one source of
heavy-flavour production. Heavy quarks can also be present in the
parton distributions at the $Q^2$ scale of the hard interaction,
leading to processes like $\mathrm{Q} \mathrm{g} \to \mathrm{Q} \mathrm{g}$,
so-called flavour excitation, or they can be
created by gluon splittings $\mathrm{g} \to \mathrm{Q} \overline{\mathrm{Q}}$ in initial- or final-state
shower evolution. In fact, as the c.m. energy is increased, these other
processes gain in importance relative to the lowest-order production
graphs above. As as example, only 10\% of the $\b$ production at LHC
energies come from the lowest-order graphs. The figure is even smaller
for charm, while it is at or above 50\% for top. At LHC energies,
the specialized treatment described in this subsection is therefore
only of interest for top (and potential fourth-generation quarks) ---
the higher-order corrections can here be approximated by an effective
$K$ factor, except possibly in some rare corners of phase space. For
charm and bottom, on the other hand, it is necessary to simulate the
full event sample (within the desired kinematics cuts), and then only
keep those events with $\b/\c$ either from lowest-order production, or
flavour excitation, or gluon splitting. Obviously this may be a
time-consuming enterprise --- although the probability for a high-$p_{\perp}$
event to contain (at least) one charm or bottom pair is fairly large,
most of these heavy flavours are carrying a small fraction of the total
$p_{\perp}$ flow of the jets, and therefore do not survive normal experimental
cuts.
As an aside, it is not only for the lowest-order graphs that events
may be generated with a guaranteed heavy-flavour content. One may also
generate the flavour excitation process by itself, in the massless
approximation, using ISUB = 28 and setting the \ttt{KFIN} array
appropriately. No trick exists to force the gluon splittings without
introducing undesirable biases, however.
The cross section for a heavy quark pair close to threshold can be
modified according to the formulae of \cite{Fad90}, see \ttt{MSTP(35)}.
Here threshold effects due to $\mathrm{Q}\overline{\mathrm{Q}}$ bound-state formation are taken
into account in a smeared-out, average sense. Then the na\"{\i}ve
cross section is multiplied by the squared wave function at the origin.
In a colour-singlet channel this gives a net enhancement of the form
\begin{equation}
|\Psi^{(s)}(0)|^2 = \frac{X_{(s)}}{1 - \exp(- X_{(s)})} ~ ,
~~~ \mrm{where}~ X_{(s)} = \frac{4}{3}
\frac{\pi \alpha_{\mathrm{s}}}{\beta} ~,
\label{pp:threshenh}
\end{equation}
while in a colour octet channel there is a net suppression given by
\begin{equation}
|\Psi^{(8)}(0)|^2 = \frac{X_{(8)}}{\exp(- X_{(8)}) -1} ~ ,
~~~ \mrm{where}~ X_{(8)} = \frac{1}{6}
\frac{\pi \alpha_{\mathrm{s}}}{\beta} ~.
\label{pp:threshsup}
\end{equation}
The $\alpha_{\mathrm{s}}$ factor in this expression is related to the energy
scale of bound-state formation; it is selected independently from
the one of the standard production cross section.
The presence of a threshold factor affects the total rate and also
kinematical distributions.
Heavy flavours, i.e. top and fourth generation, are assumed to be so
short-lived that they decay before they have time to hadronize. This
means that the light quark in the decay $\mathrm{Q} \to \mathrm{W}^{\pm} \mathrm{q}$
inherits the colour of the heavy one.
The new {\tsc{Pythia}} description represents a change of philosophy compared
to previous versions, formulated at a time when the top was thought
to be much lighter than is believed currently. However, optionally
the old description may still be used, where top hadrons are formed
and these subsequently allowed to decay; see \ttt{MSTP(48)} and
\ttt{MSTP(49)}. For event shapes the difference between the two
time orderings normally has only marginal effects \cite{Sjo92a}.
It should be noted that cross section calculations are different in
the two cases. If the top (or the fourth generation fermion) is
assumed short-lived, then it is treated like a resonance in the sense
of section \ref{sss:resdecaycross}, i.e. the cross-section is reduced
so as only to correspond to the channels left open by the user.
This also includes the restrictions on secondary decays, i.e. on the
decays of a $\mathrm{W}^+$ or a $\H^+$ produced in the top decay.
If the top is allowed to form hadrons, no such reduction takes place.
Branching ratios then have to be folded in by hand to get the correct
cross sections. The logic behind this difference is that if
hadronization takes place, one would be allowed e.g. to decay the
$\mathrm{T}^0$ and $\mathrm{T}^+$ meson according to different branching ratios.
But which $\mathrm{T}$ mesons are to be formed is not known at the top quark
creation, so one could not weight for that. For a $\t$ quark which
decays rapidly this ambiguity does not exist, and so a reduction
factor can be introduced directly coupled to the $\t$ quark production
process.
This rule about cross-section calculations applies to all the
processes explicitly set up to handle heavy flavour creation.
In addition to the ones above, this means all the ones in Tables
\ref{t:procone}--\ref{t:procfour} where the fermion final state is
given as capital letters (`$\mathrm{Q}$' and `$\mathrm{F}$') and also flavours produced
in resonance decays ($\mathrm{Z}^0$, $\mathrm{W}^{\pm}$, $\H^0$, etc., including
processes 165 and 166). However, heavy flavours can also be produced
in a process such as 31, $\mathrm{q}_i \mathrm{g} \to \mathrm{q}_k \mathrm{W}^{\pm}$, where $\mathrm{q}_k$
could be a top quark. In this case, the thrust of the description is
clearly on light flavours --- the kinematics of the process is
formulated in the massless fermion limit --- so any top production
is purely incidental. Since here the choice of scattered flavour is
only done at a later stage, the top branching ratios are not
correctly folded in to the hard scattering cross section. So, for
applications like these, it is not recommended to restrict the allowed
top decay modes. Often one might like to get rid of the possibility
of producing top together with light flavours. This can be
done by switching off (i.e. setting \ttt{MDME(I,1)=0}) the
`channels' $\d \to \mathrm{W}^- \t$, $\mathrm{s} \to \mathrm{W}^- \t$, $\b \to \mathrm{W}^- \t$,
$\mathrm{g} \to \t\overline{\mathrm{t}}$ and $\gamma \to \t\overline{\mathrm{t}}$. Also any heavy flavours
produced by parton shower evolution would not be correctly weighted
into the cross section. However, currently top production is switched
off in both initial (see \ttt{KFIN} array) and final (see
\ttt{MSTJ(45)}) state radiation.
\subsubsection{J/$\psi$}
\label{sss:Jpsiclass}
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
86 & $\mathrm{g} \mathrm{g} \to \mathrm{J}/\psi \mathrm{g}$ \\
87 & $\mathrm{g} \mathrm{g} \to \chi_{0 \c} \mathrm{g}$ \\
88 & $\mathrm{g} \mathrm{g} \to \chi_{1 \c} \mathrm{g}$ \\
89 & $\mathrm{g} \mathrm{g} \to \chi_{2 \c} \mathrm{g}$ \\
\end{tabular}
One may distinguish three main sources of $\mathrm{J}/\psi$ production.
\begin{Enumerate}
\item Decays of $\mathrm{B}$ mesons and baryons.
\item Parton-shower evolution, wherein a $\c$ and a $\overline{\mathrm{c}}$ quark
produced in adjacent branchings
(e.g. $\mathrm{g} \to \mathrm{g} \mathrm{g} \to \c \overline{\mathrm{c}} \c \overline{\mathrm{c}}$)
turn out to have so small an invariant mass that the pair collapses
to a single particle.
\item Direct production, where a $\c$ quark loop gives a coupling
between a set of gluons and a $\c\overline{\mathrm{c}}$ bound state. Higher-lying
states, like the $\chi_c$ ones, may subsequently decay to $\mathrm{J}/\psi$.
\end{Enumerate}
In this section are given the main processes for the third source,
intended for applications at hadron colliders at non-vanishing
transverse momenta --- in the limit of $p_{\perp} \to 0$ it is necessary
to include a number of $2 \to 1$ processes and to regularize
divergences in the $2 \to 2$ graphs above. The cross sections depend
on wave function values at the origin, see \ttt{PARP(38)} and
\ttt{PARP(39)}. A review of the physics issues involved may be found
in \cite{Glo88} (note, however, that the choice of $Q^2$ scale is
different in {\tsc{Pythia}}).
\subsubsection{Minimum bias}
\label{sss:minbiasclass}
\ttt{MSEL} = 1, 2 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
91 & elastic scattering \\
92 & single diffraction ($AB \to XB$) \\
93 & single diffraction ($AB \to AX$) \\
94 & double diffraction \\
95 & low-$p_{\perp}$ production \\
\end{tabular}
These processes are briefly discussed in section
\ref{ss:nonpertproc}.
Currently they are mainly intended for interactions between hadrons,
although one may also consider $\gamma \mathrm{p}$ interactions in the
option where the incoming photon is assumed resolved,
\ttt{MSTP(14)=1} or \ttt{=2}. A possible extension to
$\gamma \gamma$ interactions is not yet available.
Uncertainties come from a number of sources, e.g. from the
parametrizations of the various cross sections and slope parameters.
In diffractive scattering, the structure of the selected hadronic
system may be regulated with \ttt{MSTP(101)}. No high-$p_{\perp}$
jet production in diffractive events is included so far.
The subprocess 95, low-$p_{\perp}$ events, is somewhat unique in
that no meaningful physical border-line to high-$p_{\perp}$ events can be
defined. Even if the QCD $2 \to 2$ high-$p_{\perp}$ processes are formally
switched off, some of the generated events will be classified as
belonging to this group, with a $p_{\perp}$ spectrum of interactions to
match the `minimum-bias' event sample. Only with the option
\ttt{MSTP(82)=0} will subprocess 95 yield strictly low-$p_{\perp}$
events, events which will then probably not be compatible with any
experimental data. A number of options exist for the detailed
structure of low-$p_{\perp}$ events, see in particular \ttt{MSTP(81)} and
\ttt{MSTP(82)}. Further details on the model(s) for minimum-bias
events are found in section \ref{ss:multint}.
\subsection{Electroweak Gauge Bosons}
This section covers the production and/or exchange of $\gamma$,
$\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$ gauge bosons, singly and in pairs. The topic
of longitudinal gauge-boson scattering at high energies is deferred
to the Higgs section, since the presence or absence of a Higgs here
makes a big difference.
\subsubsection{Prompt photon production}
\label{sss:promptgammaclass}
\ttt{MSEL} = 10 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
14 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{g} \gamma$ \\
18 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma \gamma$ \\
29 & $\mathrm{q}_i \mathrm{g} \to \mathrm{q}_i \gamma$ \\
114 & $\mathrm{g} \mathrm{g} \to \gamma \gamma$ \\
115 & $\mathrm{g} \mathrm{g} \to \mathrm{g} \gamma$ \\
\end{tabular}
In hadron colliders, processes ISUB = 14 and 29 give the main source
of single-$\gamma$ production, with ISUB = 115 giving an additional
contribution which, in some kinematics regions, may become important.
For $\gamma$-pair production, the process ISUB = 18 is often
overshadowed in importance by ISUB = 114.
Another source of photons is bremsstrahlung off incoming or outgoing
quarks. This has to be treated on an equal footing with QCD parton
showering. For time-like parton-shower evolution, i.e. in the
final-state showering and in the side branches of the initial-state
showering, photon emission may be switched on or off with
\ttt{MSTJ(41)}. Photon radiation off the space-like incoming
quark legs is not yet included, but should be of lesser importance
for production at reasonably large $p_{\perp}$ values. Radiation off an
incoming electron is included in a leading-log approximation.
{\bf Warning:} the cross sections for the box graphs 114 and 115 become
very complicated, numerically unstable and slow when the
full quark mass dependence is included. For quark masses much
below the $\hat{s}$ scale, the simplified massless expressions are
therefore used --- a fairly accurate approximation. However, there
is another set of subtle numerical cancellations between different
terms in the massive matrix elements in the region of small-angle
scattering. The associated problems have not been sorted out yet.
There are therefore two possible solutions. One is to use the
massless formulae throughout. The program then becomes faster and
numerically stable, but does not give, for example, the characteristic
dip (due to destructive interference) at top threshold. This is the
current default procedure, with five flavours assumed, but this
number can be changed in \ttt{MSTP(38)}. The other possibility is
to impose cuts on the scattering angle of the hard process, see
\ttt{CKIN(27)} and \ttt{CKIN(28)}, since the numerically unstable
regions are when $|\cos\hat{\theta}|$ is close to unity. It is then
also necessary to change \ttt{MSTP(38)} to 0.
\subsubsection{Photoproduction and $\gamma\gamma$ physics}
\label{sss:photoprodclass}
\ttt{MSEL} = 1, 2, 4, 5, 6, 7, 8 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
33 & $\mathrm{q}_i \gamma \to \mathrm{q}_i \mathrm{g}$ \\
34 & $\mathrm{f}_i \gamma \to \mathrm{f}_i \gamma$ \\
54 & $\mathrm{g} \gamma \to \mathrm{q}_k \overline{\mathrm{q}}_k$ \\
58 & $\gamma \gamma \to \mathrm{f}_k \overline{\mathrm{f}}_k$ \\
80 & $\mathrm{q}_i \gamma \to \mathrm{q}_k \pi^{\pm}$ \\
84 & $\mathrm{g} \gamma \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ \\
85 & $\gamma \gamma \to \mathrm{F}_k \overline{\mathrm{F}}_k$ \\
\end{tabular}
An (almost) real photon has both a point-like component and a
hadron-like one. This means that several classes of processes
may be distinguished, see section \ref{sss:photoprod}.
\begin{Enumerate}
\item The processes listed above are possible when the photon
interacts as a point-like particle, i.e. couples directly to
quarks and leptons.
\item When the photon acts like a hadron, i.e. is resolved in a
partonic substructure, then high-$p_{\perp}$ parton--parton interactions
are possible, as already described in sections \ref{sss:QCDjetclass}
and \ref{sss:promptgammaclass}. These interactions may be further
subdivided into VMD and anomalous ones \cite{Sch93,Sch93a}.
\item A hadron-like photon can also produce the equivalent of the
minimum bias processes of section \ref{sss:minbiasclass}.
\end{Enumerate}
For $\gamma\mathrm{p}$ events, we believe that the best description can be
obtained when three separate event classes are combined, one for
direct, one for VMD and one for anomalous events, see the detailed
description in \cite{Sch93,Sch93a}.
These correspond to \ttt{MSTP(14)} being 0, 2 and 3, respectively.
The direct and anomalous components are high-$p_{\perp}$ only, while
VMD contains both high-$p_{\perp}$ and low-$p_{\perp}$ events. The option
\ttt{MSTP(14)=1} combines the VMD and anomalous parts of the photon
into one single resolved photon concept, which therefore is less
precise than the full subdivision.
When combining three runs to obtain the totality of $\gamma\mathrm{p}$
interactions, to the best of our knowledge, it is necessary to choose
the $p_{\perp}$ cut-offs with some care, so as to represent the expected
total cross section.
\begin{Itemize}
\item The direct processes only depend on the
\ttt{CKIN(3)} cut-off of the generation, with preferred value
0.5 GeV \cite{Sch93,Sch93a}. Since this value is so low, one must
remember to reduce a few other defaults values:
\ttt{CKIN(1)=2.*CKIN(3)}, \ttt{CKIN(5)=CKIN(6)=0.5*CKIN(3)}.
For the same reason it is recommended to include a dampening of
proton parton distributions, \ttt{MSTP(57)=2}.
\item The VMD processes work as ordinary
hadron--hadron ones, i.e. one obtains both low- and high-$p_{\perp}$
events by default, with dividing line set by \ttt{PARP(81)}
(or \ttt{PARP(82)}, depending on minijet unitarization scheme).
\item For the anomalous, finally, the minimal $p_{\perp}$ of the
$\gamma \to \mathrm{q}\overline{\mathrm{q}}$ branching is set in \ttt{PARP(15)}. The
default is 0.5 GeV, in agreement with the recommended cutoff for
the same vertex in direct proccesses. In addition, a lower
\ttt{CKIN(3)} cut-off should be selected for the hard interactions.
This needs some fine-tuning, which in principle should be
done separately for each c.m. energy. A good first approximation
in the HERA energy range (but not beyond 300 GeV) is
\ttt{CKIN(3)}$=1.50 + 0.0035 \, E_{\mrm{cm}}$.
\end{Itemize}
The processes in points 1 and 2 can be simulated either with a photon
beam or with an electron beam. For a photon beam it is necessary to
use option \ttt{MSTP(14)} to switch between a point-like and a resolved
photon --- it is not possible to simulate the two sets of processes
in a single run. An electron by default is assumed to contain photons,
but this can be switched off by \ttt{MSTP(11)=0}. To have quark and
gluon distributions inside the photon (itself inside the electron),
\ttt{MSTP(12)=1} must be used. For the electron, the two kinds of
processes may be generated together, unlike for the photon. It is
not possible to have also the low-$p_{\perp}$ physics (including multiple
interactions in high-$p_{\perp}$ events) for an electron beam. Kindly note
that subprocess 34 contains both the scattering of an electron off
a photon and the scattering of a quark (inside a photon inside an
electron) off a photon; the former can be switched off with the
help of the \ttt{KFIN} array.
If you are only concerned with standard QCD physics, the option
\ttt{MSTP(14)=10} gives an automatic mixture of the VMD, direct and
anomalous event classes. The mixture is properly given according to
the relative cross sections. Whenever possible, this option is therefore
preferrable in terms of user-friendliness. However, it can only work
because of a completely new layer of administration, not found anywhere
else in {\tsc{Pythia}}. For instance, a subprocess like $\mathrm{q}\mathrm{g} \to \mathrm{q}\mathrm{g}$ is
allowed in several of the classes, but appears with different sets of
parton distributions and different $p_{\perp}$ cut-offs in each of these,
so that it is necessary to switch gears between each event in the
generation. It is therefore not possible to avoid a number of
restrictions on what you can do in this case:
\begin{Itemize}
\item The \ttt{MSTP(14)=10} option can only be used for
incoming photon beams, i.e. when \ttt{'gamma'} is the argument in the
\ttt{PYINIT} call. A convolution with the bremsstrahlung photon
spectrum in an electron beam may come one day, but not in the immediate
future.
\item The machinery has only been set up to generate standard
QCD physics, specifically either `minimum-bias' one or high-$p_{\perp}$ jets.
For minimum bias, you are not allowed to use the \ttt{CKIN} variables
at all. This is not a major limitation, since it is in the spirit of
minimum-bias physics not to impose any contraints on allowed jet
production. (If you still do, these cuts will be ineffective for the
VMD processes but take effect for the other ones, giving
inconsistencies.)
The minimum-bias physics option is obtained by default; by switching
from \ttt{MSEL=1} to \ttt{MSEL=2} also the elastic and diffractive
components of the VMD part are included. High-$p_{\perp}$ jet production is
obtained by setting the \ttt{CKIN(3)} cut-off larger than each of
the (energy-dependent) cut-off scales for the VMD, direct
and anomalous components; typically this means at least 3 GeV. For
lower input \ttt{CKIN(3)} values the program will automatically switch
back to minimum-bias physics.
\item Some variables are internally recalculated and reset:
\ttt{CKIN(1)}, \ttt{CKIN(3)}, \ttt{CKIN(5)}, \ttt{CKIN(6)},
\ttt{MSTP(57)}, \ttt{MSTP(85)}, \ttt{PARP(2)}, \ttt{PARP(81)},
\ttt{PARP(82)}, \ttt{PARU(115)} and \ttt{MDME(22,J)}. This is because
they must have values that depend on the component studied.
These variables can therefore not be modified without changing
\ttt{PYINPR} and recompiling the program, which obviously is a major
exercise.
\item Pileup events are not at all allowed.
\end{Itemize}
Also, a warning about the usage of \tsc{Pdflib} for photons. So long
as \ttt{MSTP(14)=1}, i.e. the photon is not split up, \tsc{Pdflib}
is accessed by \ttt{MSTP(56)=2} and \ttt{MSTP(55)} as the parton
distribution set. However, when the VMD and anomalous pieces are split,
the VMD part is based on a rescaling of pion distributions by VMD factors
(except for the SaS sets, that already come with a separate VMD piece).
Therefore, to access \tsc{Pdflib} for \ttt{MSTP(14)=10}, it is not
correct to set \ttt{MSTP(56)=2} and a photon distribution in
\ttt{MSTP(55)}. Instead, one should put \ttt{MSTP(56)=2},
\ttt{MSTP(54)=2} and a pion distribution code in \ttt{MSTP(53)},
while \ttt{MSTP(55)} has no function. The anomalous part is still based on
the SaS parametrization, with \ttt{PARP(15)} as main free parameter.
Currently, hadrons are not defined with any photonic content. None
of the processes are therefore relevant in hadron--hadron collisions.
In $\mathrm{e}\mathrm{p}$ collisions, the electron can emit an almost real photon,
which may interact directly or be resolved. In $\e^+\e^-$ collisions,
one may have direct, singly-resolved or doubly-resolved processes.
The $\gamma\gamma$ equivalent to the $\gamma\mathrm{p}$ description involves
six different event classes, see section \ref{sss:photoprod}.
These classes can be obtained by setting \ttt{MSTP(14)} to 0, 2, 3,
5, 6 and 7, respectively. If one combines the VMD and anomalous
parts of the parton distributions of the photon, in a more coarse
description,
it is enough to use the \ttt{MSTP(14)} options 0, 1 and 4. The cut-off
procedures follows from the ones used for the $\gamma\mathrm{p}$ ones above.
Thus the direct$\times$direct and direct$\times$VMD processes require
the same cut-offs as used for direct $\gamma\mathrm{p}$ events, the
VMD$\times$VMD ones the same as used for VMD $\gamma\mathrm{p}$ events,
and the rest (anomalous$\times$anomalous, direct$\times$anomalous
and VMD$\times$anomalous) the same as used for anomalous $\gamma\mathrm{p}$
events.
As with $\gamma\mathrm{p}$ events, the option \ttt{MSTP(14)=10} gives a mixture
of the six possible $\gamma\gamma$ event classes. The same complications
and restrictions exist here as already listed above. For normal use
the advantages should outweight the disadvantages.
It is hoped to extend the formalism also to mildly virtual photons.
Currently this is not done. The interaction of a highly virtual
photon with a real photon is included in the deep inelastic scattering
formalism below, however.
Process 54 generates a mixture of quark flavours; allowed flavours
are set by the gluon \ttt{MDME values}. Process 58 can generate both
quark and lepton pairs, according to the \ttt{MDME} values of the
photon. Processes 84 and 85 are variants of these matrix elements,
with fermion masses included in the matrix elements, but where only
one flavour can be generated at a time. This flavour is selected as
described for processes 81 and 82 in section \ref{sss:heavflavclass},
with the exception that for process 85 the `heaviest' flavour allowed
for photon splitting takes to place of the heaviest flavour allowed
for gluon splitting. Since lepton KF codes come after quark ones,
they are counted as being `heavier', and thus take precedence if
they have been allowed.
Process 80 is a higher twist one. The theory for such processes
is rather shaky, so results should not be taken too literally.
The messy formulae given in \cite{Bag82} have not been programmed
in full, instead the pion form factor has been parametrized as
$Q^2 F_{\pi}(Q^Q) \approx 0.55 / \ln Q^2$, with $Q$ in GeV.
\subsubsection{Deep inelastic scattering}
\label{sss:DISclass}
\ttt{MSEL} = 1, 2, 35, 36, 37, 38 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
10 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_k \mathrm{f}_l$ \\
83 & $\mathrm{q}_i \mathrm{f}_j \to \mathrm{Q}_k \mathrm{f}_l$ \\
\end{tabular}
The `deep inelastic scattering' (DIS) processes, i.e. $t$-channel
electroweak gauge boson exchange, are traditionally associated
with interactions between a lepton or neutrino and a hadron, but
processes 10 and 83 can equally well be applied for $\mathrm{q}\q$ scattering
in hadron colliders (with a cross section much smaller than
corresponding QCD processes, however). If applied to incoming $\e^+\e^-$
beams, process 10 corresponds to Bhabha scattering.
For process 10 both $\gamma$, $\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$ exchange
contribute, including interference between $\gamma$ and $\mathrm{Z}^0$.
The switch \ttt{MSTP(21)} may be used to restrict to only some
of these, e.g. neutral or charged current only.
The option \ttt{MSTP(14)=10} (see previous section) has now been
extended so that it also works for deep inelastic sacattering of
an electron off a (real) photon, i.e. process 10. What is obtained
is a mixture of the photon acting as a vector meson and it acting
as an anomalous state. This should therefore be the sum of what can
be obtained with \ttt{MSTP(14)=2} and \ttt{=3}. It is distinct from
\ttt{MSTP(14)=1} in that different sets are used for the parton
distributions --- in \ttt{MSTP(14)=1} all the contributions to the
photon distributions are lumped together, while they are split in
VMD and anomalous parts for \ttt{MSTP(14)=10}. Also the beam remnant
treatment is different, with a simple Gaussian distribution (at least
by default) for \ttt{MSTP(14)=1} and the VMD part of \ttt{MSTP(14)=10},
but a powerlike distribution $\d k_{\perp}^2 / k_{\perp}^2$ between
\ttt{PARP(15)} and $Q$ for the anomalous part of \ttt{MSTP(14)=10}.
To access this option for $\mathrm{e}$ and $\gamma$ as incoming beams, it is
only necessary to set \ttt{MSTP(14)=10} and keep \ttt{MSEL} at its
default value. Unlike the corresponding option for $\gamma\mathrm{p}$ and
$\gamma\gamma$, no cuts are overwritten, i.e. it is still the
responsability of the user to set these appropriately.
Cuts especially appropriate for DIS usage include either
\ttt{CKIN(21)-CKIN(22)} or
\ttt{CKIN(23)-CKIN(24)} for the $x$ range (former or latter depending
on which side is the incoming real photon), \ttt{CKIN(35)-CKIN(36)} for
the $Q^2$ range, and \ttt{CKIN(39)-CKIN(40)} for the $W^2$ range.
In principle, the DIS $x$ variable of an event corresponds to the
$x$ value stored in \ttt{PARI(33)} or \ttt{PARI(34)}, depending
on which side the incoming hadron is on, while the DIS
$Q^2 = -\hat{t} = $\ttt{-PARI(15)}. However, just like initial- and
final-state radiation can shift jet momenta, they can modify
the momentum of the scattered lepton. Therefore the DIS $x$ and
$Q^2$ variables are not automatically conserved. An option, on by
default, exists in \ttt{MSTP(23)}, where the event can be `modified
back' so as to conserve $x$ and $Q^2$, but this option is still rather
primitive and should not be taken too literally.
Process 83 is the equivalent of process 10 for $\mathrm{W}^{\pm}$ exchange
only, but with the heavy-quark mass included in the matrix element.
In hadron colliders it is mainly of interest for the production of
very heavy flavours, where the possibility of producing just one
heavy quark is kinematically favoured over pair production. The
selection of the heavy flavour is already discussed in section
\ref{sss:heavflavclass}.
\subsubsection{Single $\mathrm{W}/\mathrm{Z}$ production}
\label{sss:WZclass}
\ttt{MSEL} = 11, 12, 13, 14, 15, (21) \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
1 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma^* / \Z^0$ \\
2 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+$ \\
15 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{g} (\gamma^* / \Z^0)$ \\
16 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{g} \mathrm{W}^+$ \\
19 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma (\gamma^* / \Z^0)$ \\
20 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \gamma \mathrm{W}^+$ \\
30 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_i (\gamma^* / \Z^0)$ \\
31 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_k \mathrm{W}^+$ \\
35 & $\mathrm{f}_i \gamma \to \mathrm{f}_i (\gamma^* / \Z^0)$ \\
36 & $\mathrm{f}_i \gamma \to \mathrm{f}_k \mathrm{W}^+$ \\
131 & $\mathrm{g} \mathrm{g} \to \mathrm{Z}^0 \mathrm{Q}_k \overline{\mathrm{Q}}_k$ \\
(141) & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma/\mathrm{Z}^0/\mathrm{Z}'^0$ \\
\end{tabular}
This group consists of $2 \to 1$ processes, i.e. production of a
single resonance, and $2 \to 2$ processes, where the resonance
is recoiling against a jet or a photon. The process 141, which
also is listed here, is described further elsewhere.
With initial-state showers turned on, the $2 \to 1$ processes also
generate additional jets; in order to avoid double-counting, the
corresponding $2 \to 2$ processes should therefore not be turned
on simultaneously. The basic rule is to use the $2 \to 1$ processes
for inclusive generation of $\mathrm{W}/\mathrm{Z}$, i.e. where the bulk of the
events studied have $p_{\perp} \ll m_{\mathrm{W}/\mathrm{Z}}$, which is where parton showers
may be expected to do a good job. For dedicated studies of $\mathrm{W}/\mathrm{Z}$
production at larger transverse momenta, the parton showers tend
to underestimate the event rates. It is here better to start from
the $2 \to 2$ matrix elements and add showers to these. However, the
$2 \to 2$ matrix elements are divergent for $p_{\perp} \to 0$, and should
not be used down to the low-$p_{\perp}$ region, or one may get unphysical
cross sections. The problem of double-counting applies not only
to $\mathrm{W}/\mathrm{Z}$ production in hadron colliders, but also to a process
like $\e^+\e^- \to \mathrm{Z}^0 \gamma$, which clearly is part of the
initial-state radiation corrections to $\e^+\e^- \to \mathrm{Z}^0$ obtained for
\ttt{MSTP(11)=1}. As is the case for $\mathrm{Z}$ production in association
with jets, the $2 \to 2$ process should therefore only be used for the
high-$p_{\perp}$ region.
The $\mathrm{Z}^0$ of subprocess 1 includes the full interference structure
$\gamma^* / \Z^0$; via \ttt{MSTP(43)} you can select to produce only
$\gamma^*$, only $\mathrm{Z}^0$, or the full $\gamma^* / \Z^0$. The same holds true
for the $\mathrm{Z}'^0$ of subprocess 141; via \ttt{MSTP(44)} any combination
of $\gamma^*$, $\mathrm{Z}^0$ and $\mathrm{Z}'^0$ can be selected. Thus, subprocess
141 with \ttt{MSTP(44)=4} is essentially equivalent to subprocess
1 with \ttt{MSTP(43)=3}; however, process 141 also includes the
possibility of a decay into Higgses. Also processes 15, 19, 30 and 35
contain the full mixture of $\gamma^* / \Z^0$, with \ttt{MSTP(43)} available
to change this. Only the $\mathrm{Z}^0$ that appears in process 131 does not
contain the $\gamma^*$ contribution.
Note that process 1, with only
$\mathrm{q}\overline{\mathrm{q}} \to \gamma^* \to \ell^+ \ell^-$ allowed,
and studied in the region well below the $\mathrm{Z}^0$ mass, is what is
conventionally called Drell--Yan. This latter process therefore does
not appear under a separate heading, but can be obtained by a
suitable setting of switches and parameters.
A process like $\mathrm{f}_i \overline{\mathrm{f}}_j \to \gamma \mathrm{W}^+$ is only included in
the limit that the $\gamma$ is emitted in the `initial state',
while the possibility of a final-state radiation off the $\mathrm{W}^+$
decay products is not explicitly included (but can be obtained
implicitly by the parton-shower machinery) and various interference
terms are not at all present. Some caution must therefore be
exercised; see also section \ref{sss:WZpairclass} for related
comments.
For the $2 \to 1$ processes, the Breit--Wigner includes an
$\hat{s}$-dependent width, which should provide an improved
description of line shapes. In fact, from a line-shape point of view,
process 1 should provide a more accurate simulation of $\e^+\e^-$
annihilation events than the dedicated $\e^+\e^-$ generation scheme of
{\tsc{Jetset}} (see section \ref{ss:eematrix}). However, the $p_{\perp}$
distribution of
radiated initial-state photons is probably still better modelled in
the {\tsc{Jetset}} routines. Another difference is that {\tsc{Jetset}} only allows
the generation of $\gamma^* / \Z^0 \to \mathrm{q}\overline{\mathrm{q}}$, while process 1 additionally
contains $\gamma^* / \Z^0 \to \ell^+ \ell^-$ and $\gamma^* / \Z^0 \to \nu \br{\nu}$.
The parton-shower and fragmentation descriptions are the same,
but the {\tsc{Pythia}} implementation has not been interfaced with the
first- and second-order matrix-element options available in {\tsc{Jetset}}.
Almost all processes in this group have been included with the
correct angular distribution in the subsequent $\mathrm{W}/\mathrm{Z} \to \mathrm{f} \overline{\mathrm{f}}$
decays. The exception is process 36, where currently
the $\mathrm{W}$ decays isotropically.
The process $\e^+\e^- \to \mathrm{e}^+ \mathrm{e}^- \mathrm{Z}^0$ can be simulated in two
different ways. One is to make use of the $\mathrm{e}$ `sea' distribution
inside $\mathrm{e}$, i.e. have splittings $\mathrm{e} \to \gamma \to \mathrm{e}$.
This can be obtained, together with ordinary $\mathrm{Z}^0$ production, by
using subprocess 1, with \ttt{MSTP(11)=1} and \ttt{MSTP(12)=1}. Then
the contribution of the type above is 5.0 pb for a 500 GeV $\e^+\e^-$
collider, compared with the correct 6.2 pb \cite{Hag91}. Alternatively
one may use process 35, with \ttt{MSTP(11)=1} and \ttt{MSTP(12)=0}.
To catch the singularity in the forward direction, regularized by
the electron mass, it is necessary to set
\ttt{CKIN(3)=CKIN(5)=0.01} --- using lower values will only slow
down execution, not significantly increasing the cross section. One
then obtains 5.1 pb, i.e. again 20\% below the correct value, but now
also generates a $p_{\perp}$ distribution for the $\mathrm{Z}^0$; this is therefore
to be preferred.
Process 36, $\mathrm{f} \gamma \to \mathrm{f}' \mathrm{W}^{\pm}$ may have corresponding
problems; except that in $\e^+\e^-$ the forward scattering amplitude for
$\mathrm{e} \gamma \to \nu \mathrm{W}$ is killed (radiation zero), which means
that the differential cross section is vanishing for $p_{\perp} \to 0$.
It is therefore feasible to use the default \ttt{CKIN(3)} and
\ttt{CKIN(5)} values in $\e^+\e^-$, and one also comes closer to the
correct cross section.
One single true $2 \to 3$ process is included in this class as well;
namely $\mathrm{g} \mathrm{g} \to \mathrm{Z}^0 \mathrm{Q} \overline{\mathrm{Q}}$, with full massive matrix elements.
The more complicated phase space and the lengthy matrix-element
evaluations make this process extremely slow. With the quark flavour
picked to be $\b$, it may form an important background to intermediate
mass Higgs searches in the multilepton channel. The quark flavour is
stored in \ttt{KFPR(131,2)}; the default is $5 = \b$. The kinematics
is set up in terms of a $\mathrm{Z}^0$ recoiling against the $\mathrm{Q} \overline{\mathrm{Q}}$
system, and all ordinary kinematics cut for a $2 \to 2$ process can
be used on this level, including \ttt{CKIN(43)} and \ttt{CKIN(44)} to
restrict the range of the $\mathrm{Q} \overline{\mathrm{Q}}$ invariant mass. In addition, for
this process alone, \ttt{CKIN(51) - CKIN(54)} can be used to set the
$p_{\perp}$ range of the two quarks; as is to be expected, that of the
$\mathrm{Z}^0$ is set by \ttt{CKIN(3) - CKIN(4)}. Since the optimization
procedure is not set up to probe the full multidimensional phase space
allowed in this process, maximum violations may be quite large. It may
then be useful to make a preliminary run to find how big the violations
are in total, and then use the \ttt{MSTP(121)=1} option in the full
run.
\subsubsection{$\mathrm{W}/\mathrm{Z}$ pair production}
\label{sss:WZpairclass}
\ttt{MSEL} = 15 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
22 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to (\gamma^* / \Z^0) (\gamma^* / \Z^0)$ \\
23 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{Z}^0 \mathrm{W}^+$ \\
25 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{W}^+ \mathrm{W}^-$ \\
69 & $\gamma \gamma \to \mathrm{W}^+ \mathrm{W}^-$ \\
70 & $\gamma \mathrm{W}^+ \to \mathrm{Z}^0 \mathrm{W}^+$ \\
\end{tabular}
In this section we mainly consider the production of $\mathrm{W}/\mathrm{Z}$ pairs
by fermion--antifermion annihilation, but also include two processes
which involve $\gamma/\mathrm{W}$ beams. Scatterings between gauge-boson
pairs, i.e. processes like $\mathrm{W}^+ \mathrm{W}^- \to \mathrm{Z}^0 \mathrm{Z}^0$, depend so
crucially on the assumed Higgs scenario that they are considered
separately in section \ref{sss:heavySMHclass}.
The cross sections used for the above processes are those derived
in the narrow-width limit, but have been extended to include
Breit--Wigner shapes with mass-dependent widths. However, one
should realize that other graphs, not included here, can contribute
in regions away from the $\mathrm{W}/\mathrm{Z}$ mass. This problem is especially
important if several flavours coincide in the four-fermion final
state. Consider, as an example,
$\e^+\e^- \to \mu^+ \mu^- \nu_{\mu} \br{\nu}_{\mu}$.
Not only would such a final state receive contributions from
intermediate $\mathrm{Z}^0\mathrm{Z}^0$ and $\mathrm{W}^+\mathrm{W}^-$ states, but also
from processes $\e^+\e^- \to \mathrm{Z}^0 \to \mu^+ \mu^-$, followed
either by $\mu^+ \to \mu^+ \mathrm{Z}^0 \to \mu^+ \nu_{\mu} \br{\nu}_{\mu}$,
or by
$\mu^+ \to \br{\nu}_{\mu} \mathrm{W}^+ \to \br{\nu}_{\mu} \mu^+ \nu_{\mu}$.
In addition, all possible interferences should be considered.
Since this is not done, the processes have to be used with some
sound judgement. Very often, one may wish to constrain a
lepton pair mass to be close to $m_{\mathrm{Z}}$, in which case a number
of the possible `other' processes are negligible.
Of the above processes, the first contains the full
$\mathrm{f}_i \overline{\mathrm{f}}_i \to (\gamma^* / \Z^0)(\gamma^* / \Z^0)$ structure, obtained by a
straightforward generalization of the formulae in ref. \cite{Gun86}
(done by the present author). Of course, the possibility of
there being significant contributions from graphs that are not
included is increased, in particular
if one $\gamma^*$ is very light and therefore could be a
bremsstrahlung-type photon. It is possible to use \ttt{MSTP(43)} to
recover the pure $\mathrm{Z}^0$ case, i.e. $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Z}^0 \mathrm{Z}^0$
exclusively. In processes 23 and 70, only the pure $\mathrm{Z}^0$ contribution
is included.
Full angular correlations are included for the first three processes,
i.e. the full $2 \to 2 \to 4$ matrix elements are included in the
resonance decays, including the appropriate $\gamma^* / \Z^0$ interference
in process 22. In the latter two processes no spin
information is currently preserved, i.e. the $\mathrm{W}/\mathrm{Z}$ bosons are
allowed to decay isotropically.
We remind you that the mass ranges of the two resonances may be set
with the \ttt{CKIN(41) - CKIN(44)} parameters; this is particularly
convenient, for instance, to pick one resonance almost on the mass
shell and the other not.
\subsection{Higgs Production}
A fair fraction of all the processes in {\tsc{Pythia}} deal with Higgs
production in one form or another. This multiplication is caused by
the need to consider production by several different processes,
depending on Higgs mass and machine type. Further, the program
contains a full two-Higgs-multiplet scenario, as predicted for example
in the Minimal Supersymmetric extension of the Standard Model
(MSSM). Therefore the continued discussion is, somewhat arbitrarily,
subdivided into a few different scenarios.
\subsubsection{Light Standard Model Higgs}
\label{sss:lightSMHclass}
\ttt{MSEL} = 16, 17, 18 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
3 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \H^0$ \\
24 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Z}^0 \H^0$ \\
26 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+ \H^0$ \\
102 & $\mathrm{g} \mathrm{g} \to \H^0$ \\
103 & $\gamma \gamma \to \H^0$ \\
110 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma \H^0$ \\
111 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{g} \H^0$ \\
112 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_i \H^0$ \\
113 & $\mathrm{g} \mathrm{g} \to \mathrm{g} \H^0$ \\
121 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \H^0$ \\
122 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k \H^0$ \\
123 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j \H^0$ ($\mathrm{Z}^0 \mathrm{Z}^0$ fusion) \\
124 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_k \mathrm{f}_l \H^0$ ($\mathrm{W}^+ \mathrm{W}^-$ fusion) \\
\end{tabular}
In this section we discuss the production of a reasonably light
Standard Model Higgs, below 700 GeV, say, so that the narrow
width approximation can be used with some confidence. Below 400 GeV
there would certainly be no trouble, while above that the narrow
width approximation is gradually starting to break down.
In a hadron collider, the main production processes are 102, 123
and 124, i.e. $\mathrm{g}\g$, $\mathrm{Z}^0 \mathrm{Z}^0$ and $\mathrm{W}^+ \mathrm{W}^-$ fusion. In the
latter two processes, it is also necessary to take into account
the emission of the space-like $\mathrm{W}/\mathrm{Z}$ bosons off quarks, which
in total gives the $2 \to 3$ processes above.
Further processes of lower cross sections may be of interest
because of easier signals. For instance, processes 24 and 26 give
associated production of a $\mathrm{Z}$ or a $\mathrm{W}$ together with the $\H^0$.
There is also the processes 3, 121 and 122, which involve production
of heavy flavours.
Process 3 contains contributions from all flavours, but is
completely dominated by the subprocess $\t \overline{\mathrm{t}} \to \H^0$,
i.e. by the contribution from the top sea distributions.
This process is by now known to overestimate the cross section
for Higgs production as compared with a more careful calculation
based on the subprocess $\mathrm{g} \mathrm{g} \to \t \overline{\mathrm{t}} \H^0$ (121). The
difference between the two is that in process 3 the
$\t$ and $\overline{\mathrm{t}}$ are added by the initial-state shower, while in
121 the full matrix element is used. The price to be paid is that
the complicated multibody phase space in process 121 makes the
program run slower than with most other processes. One should
therefore think twice before using it. As usual, it would be
double-counting to include both 3 and 121. Process 122 is
similar in structure to 121, but is less important. In both
process 121 and 122 the produced quark is assumed to be a $\t$;
this can be changed in \ttt{KFPR(121,2)} and \ttt{KFPR(122,2)}
before initialization, however.
A subprocess like 113, with a Higgs recoiling against a gluon jet, is
also effectively generated by initial-state corrections to subprocess
102; thus, in order to avoid double-counting, just as for the case
of $\mathrm{Z}^0/\mathrm{W}^+$ production, section \ref{sss:WZclass}, these subprocesses
should not be switched on simultaneously. Process 102 should
be used for inclusive production of Higgs, and 111--113 for the study
of the Higgs subsample with high transverse momentum.
In $\e^+\e^-$ annihilation, associated production of an $\H^0$ with a
$\mathrm{Z}^0$, process 24, is usually the dominant one close to threshold,
while the $\mathrm{Z}^0 \mathrm{Z}^0$ and $\mathrm{W}^+ \mathrm{W}^-$ fusion processes 123 and 124
win out at high energies. Process 103, $\gamma\gamma$ fusion, may
also be of interest, in particular when the possibilities of
beamstrahlung photons and backscattered photons are included.
Process 110, which gives an $\H^0$ in association with a $\gamma$,
is a loop process and is therefore suppressed in rate. Only for a
rather massive $\H^0$ (mass above 60 GeV at LEP 1) can it start to
compete with the associated production of a $\mathrm{Z}^0$, since phase space
suppression is less severe for the former than for the latter.
The branching ratios of the Higgs are very strongly dependent on
the mass. In principle, the program is set up to calculate these
correctly, as a function of the actual Higgs mass, i.e. not just
at the nominal mass. However, higher-order corrections may at
times be important and not fully unambiguous; see for instance
\ttt{MSTP(37)}.
Since the Higgs is a spin-0 particle it decays isotropically. In decay
processes such as $\H^0 \to \mathrm{W}^+ \mathrm{W}^- \to 4$ fermions angular
correlations are included. Also in processes 24 and 26, $\mathrm{Z}^0$
and $\mathrm{W}^{\pm}$ decay angular distributions are correctly taken into
account.
\subsubsection{Heavy Standard Model Higgs}
\label{sss:heavySMHclass}
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
5 & $\mathrm{Z}^0 \mathrm{Z}^0 \to \H^0$ \\
8 & $\mathrm{W}^+ \mathrm{W}^- \to \H^0$ \\
71 & $\mathrm{Z}^0 \mathrm{Z}^0 \to \mathrm{Z}^0 \mathrm{Z}^0$ (longitudinal) \\
72 & $\mathrm{Z}^0 \mathrm{Z}^0 \to \mathrm{W}^+ \mathrm{W}^-$ (longitudinal) \\
73 & $\mathrm{Z}^0 \mathrm{W}^+ \to \mathrm{Z}^0 \mathrm{W}^+$ (longitudinal) \\
76 & $\mathrm{W}^+ \mathrm{W}^- \to \mathrm{Z}^0 \mathrm{Z}^0$ (longitudinal) \\
77 & $\mathrm{W}^+ \mathrm{W}^{\pm} \to \mathrm{W}^+ \mathrm{W}^{\pm}$ (longitudinal) \\
\end{tabular}
Processes 5 and 8 are the simple $2 \to 1$ versions of what is now
available in 123 and 124 with the full $2 \to 3$ kinematics.
For low Higgs masses processes 5 and 8 overestimate the correct
cross sections and should not be used, whereas good agreement between
the $2 \to 1$ and $2 \to 3$ descriptions is observed when heavy
Higgs production is studied.
The subprocesses 5 and 8, $V V \to \H^0$, which contribute to the
processes $V V \to V' V'$, show a bad high-energy behaviour. Here
$V$ denotes a longitudinal intermediate gauge boson, $\mathrm{Z}^0$ or
$\mathrm{W}^{\pm}$. This can be cured only by the inclusion of all
$V V \to V' V'$ graphs, as is done in subprocesses 71, 72, 73, 76
and 77. In particular, subprocesses 5 and 8 give rise to a fictitious
high-mass tail of the Higgs. If this tail is thrown away, however,
the agreement between the $s$-channel graphs only (subprocesses
5 and 8) and the full set of graphs (subprocesses 71 etc.) is very
good: for a Higgs of nominal mass 300 (800) GeV, a cut at 600 (1200)
GeV retains 95\% (84\%) of the total cross section, and differs from
the exact calculation, cut at the same values, by only 2\% (11\%)
(numbers for SSC energies). With this prescription there is
therefore no need to use subprocesses 71 etc. rather than
subprocesses 5 and 8.
For subprocess 77, there is an option, see \ttt{MSTP(45)}, to select
the charge combination of the scattering $\mathrm{W}$'s: like-sign,
opposite-sign (relevant for Higgs), or both.
Process 77 contains a divergence for $p_{\perp} \to 0$ due to
$\gamma$-exchange contributions. This leads to an infinite total
cross section, which is entirely fictitious, since the simple
parton-distribution function approach to the longitudinal $\mathrm{W}$ flux
is not appropriate in this limit. For this process, it is therefore
necessary to make use of a cut, e.g. $p_{\perp} > m_{\mathrm{W}}$.
For subprocesses 71, 72, 76 and 77, an option is included (see
\ttt{MSTP(46)}) whereby the user can select only the $s$-channel
Higgs graph; this will then be essentially equivalent to running
subprocess 5 or 8 with the proper decay channels (i.e. $\mathrm{Z}^0\mathrm{Z}^0$ or
$\mathrm{W}^+\mathrm{W}^-$) set via \ttt{MDME}. The difference is that the
Breit--Wigners in subprocesses 5 and 8 contain a mass-dependent
width, whereas the width in subprocesses 71--77 is calculated at
the nominal Higgs mass; also, higher-order corrections to the widths
are treated more accurately in subprocesses 5 and 8. Further,
processes 71--77 assume the incoming $\mathrm{W}/\mathrm{Z}$ to be on the mass shell,
with associated kinematics factors, while processes 5 and 8 have
$\mathrm{W}/\mathrm{Z}$ correctly space-like. All this leads to differences in the
cross sections by up to a factor of 1.5.
In the absence of a Higgs, the sector of longitudinal $\mathrm{Z}$ and $\mathrm{W}$
scattering will become strongly interacting at energies above 1 TeV.
The models proposed by Dobado, Herrero and Terron \cite{Dob91} to
describe this kind of physics have been included as alternative matrix
elements for subprocesses 71, 72, 73, 76 and 77, selectable by
\ttt{MSTP(46)}. From the point of view of the general classification
scheme for subprocesses, this kind of models should appropriately be
included as separate subprocesses with numbers above 100, but the
current solution allows a more efficient reuse of existing code.
By a proper choice of parameters, it is also here possible to
simulate the production of a techni-$\rho$.
Currently, the scattering of transverse gauge bosons has not been
included, neither that of mixed transverse--longitudinal scatterings.
These are expected to be less important at high energies, and do not
contain an $\H^0$ resonance peak, but need not be
entirely negligible in magnitude. As a rule of thumb, processes
71--77 should not be used for $VV$ invariant masses below 500 GeV.
The decay products of the longitudinal gauge bosons are correctly
distributed in angle.
\subsubsection{Extended neutral Higgs sector}
\label{sss:extneutHclass}
\ttt{MSEL} = 19 \\
ISUB = \\
\begin{tabular}{rrrl@{\protect\rule{0mm}{\tablinsep}}}
$\H^0$ & $\H'^0$ & $\mathrm{A}^0$ & \\
3 & 151 & 156 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to X$ \\
102 & 152 & 157 & $\mathrm{g} \mathrm{g} \to X$ \\
103 & 153 & 158 & $\gamma \gamma \to X$ \\
24 & 171 & 176 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{Z}^0 X$ \\
26 & 172 & 177 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+ X$ \\
123 & 173 & 178 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j X$ ($\mathrm{Z} \mathrm{Z}$ fusion) \\
124 & 174 & 179 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_k \mathrm{f}_l X$ ($\mathrm{W}^+\mathrm{W}^-$ fusion) \\
121 & 181 & 186 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k X$ \\
122 & 182 & 187 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k X$ \\
\end{tabular} \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
(141) & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma/\mathrm{Z}^0/\mathrm{Z}'^0$ \\
\end{tabular}
In {\tsc{Pythia}}, the particle content of a two-Higgs-doublet scenario is
included: two neutral scalar particles, 25 and 35, one pseudoscalar
one, 36, and a charged doublet, $\pm 37$. (Of course, these particles
may also be associated with corresponding Higgs states in larger
multiplets.) By convention, we choose to call the lighter scalar
Higgs $\H^0$ and the heavier $\H'^0$ --- this differs from the
convention in the MSSM, where the lighter is called $\mathrm{h}^0$ and the
heavier
$\H^0$, but allows us to call the Higgs of the one-Higgs scenario
$\H^0$. The pseudoscalar is called $\mathrm{A}^0$ and the charged $\H^{\pm}$.
Charged-Higgs production is covered in section \ref{sss:chHclass}.
A number of $\H^0$ processes have been duplicated for $\H'^0$ and
$\mathrm{A}^0$. The correspondence between ISUB numbers is shown in the table
above: the first column of ISUB numbers corresponds to
$X = \H^0$, the second to $X = \H'^0$, and the third to $X = \mathrm{A}^0$.
Note that several of these processes are not expected to take
place at all, owing to vanishing Born term couplings. We have still
included them for flexibility in simulating arbitrary couplings at
the Born or loop level.
A few Standard Model Higgs processes have no correspondence in the
scheme above. These include
\begin{Itemize}
\item 5 and 8, which anyway have been superseded by 123 and 124;
\item 71, 72, 73, 76 and 77, which deal with what happens if there
is no light Higgs, and so is a scenario complementary to the one
above, where several light Higgses are assumed;
\item 110, which is mainly of interest in Standard Model Higgs
searches; and
\item 111, 112 and 113, which describe the high-$p_{\perp}$ tail of the
Higgs production, and are less interesting for most Higgs studies.
\end{Itemize}
In processes 121, 122, 181, 182, 186 and 187 the recoiling heavy
flavour is assumed to be top, which is the only one of interest in
the Standard Model, and the one where the parton-distribution-function
approach invoked in processes 3, 151 and 156 is least reliable.
However, it is possible to change the quark flavour in 121 etc.;
for each process ISUB this flavour is given by \ttt{KFPR(ISUB,2)}.
This may become relevant if couplings to $\b\overline{\mathrm{b}}$ states are
enhanced, e.g. if $\tan\beta \gg 1$ in the MSSM.
By default, the $\H^0$ has the couplings of the Standard Model
Higgs, while the $\H'^0$ and $\mathrm{A}^0$ have couplings set in
\ttt{PARU(171) - PARU(178)} and \ttt{PARU(181) - PARU(190)},
respectively. The default values for the $\H'^0$ and $\mathrm{A}^0$ have no
deep physics motivation, but are set just so that the program will
not crash due to the absence of any couplings whatsoever. You
should therefore set the above couplings to your desired values if
you want to simulate either $\H'^0$ or $\mathrm{A}^0$. Also the couplings
of the $\H^0$ particle can be modified, in
\ttt{PARU(161) - PARU(165)}, provided that \ttt{MSTP(4)} is set to
1.
For \ttt{MSTP(4)=2}, the mass of the $\H^0$ (in \ttt{PMAS(25,1)})
and the $\tan\beta$ value (in \ttt{PARU(141)}) are used to derive
the masses of the other Higgses, as well as all Higgs couplings.
\ttt{PMAS(35,1) - PMAS(37,1)} and \ttt{PARU(161) - PARU(195)} are
overwritten accordingly. The relations used are the ones of the
Born-level MSSM \cite{Gun90}. Today, loop corrections to those
expressions have been calculated, and are known to have
non-negligible effects on the resulting phenomenology. Eventually
the modified relations will be included as an additional option, but
this has not yet been done.
Note that not all combinations of $m_{\H}$ and $\tan\beta$ are
allowed; the requirement of a finite $\mathrm{A}^0$ mass imposes the
constraint
\begin{equation}
m_{\H} < m_{\mathrm{Z}} \, \frac{\tan^2\beta - 1}{\tan^2\beta + 1},
\end{equation}
or, equivalently,
\begin{equation}
\tan^2\beta > \frac{m_{\mathrm{Z}} + m_{\H}}{m_{\mathrm{Z}} - m_{\H}}.
\end{equation}
If this condition is not fulfilled, the program will crash.
Process 141 can also be used to simulate $\mathrm{Z}^0 \to \H^0 \mathrm{A}^0$ and
$\mathrm{Z}^0 \to \H'^0 \mathrm{A}^0$ for associated neutral Higgs production.
The fact that we here make use of the $\mathrm{Z}'^0$ can easily be
discounted, either by letting the relevant couplings vanish,
or by the option \ttt{MSTP(44)=4}.
Finally, heavier Higgses may decay into lighter ones, if
kinematically allowed, in processes like $\mathrm{A}^0 \to \mathrm{Z}^0 \H^0$ or
$\H^+ \to \mathrm{W}^+ \H^0$. Such modes are included as part of the
general mixture of decay channels, but they can be enhanced if
the uninteresting channels are switched off.
\subsubsection{Charged Higgs sector}
\label{sss:chHclass}
\ttt{MSEL} = 23 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
143 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \H^+$ \\
161 & $\mathrm{f}_i \mathrm{g} \to \mathrm{f}_k \H^+$ \\
(141) & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma/\mathrm{Z}^0/\mathrm{Z}'^0$ \\
\end{tabular}
A charged Higgs doublet, $\H^{\pm}$, is included in the program.
This doublet may be the one predicted in the MSSM scenario,
see section \ref{sss:extneutHclass}, or in any other scenario.
The $\tan\beta$ parameter, which is relevant also
for charged Higgs couplings, is set via \ttt{PARU(141)}.
The basic subprocess for charged Higgs production in hadron
colliders is ISUB = 143. However, this process is dominated
by $\t \overline{\mathrm{b}} \to \H^+$, and so depends on the choice of $\t$
parton distribution. A better representation is provided by
subprocess 161, $\mathrm{f} \mathrm{g} \to \mathrm{f}' \H^+$; i.e. actually
$\overline{\mathrm{b}} \mathrm{g} \to \overline{\mathrm{t}} \H^+$. It is therefore recommended to use
161 and not 143; to use both would be double-counting.
In subprocess 141, the decay $\gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0 \to \H^+ \H^-$
allows the production of a pair of charged Higgs particles.
This process is especially important in $\e^+\e^-$ colliders.
The coupling of the $\gamma^*$ to $\H^+ \H^-$ is determined by
the charge alone, while the $\mathrm{Z}^0$ coupling is regulated
by \ttt{PARU(142)}, and that of the $\mathrm{Z}'^0$ by \ttt{PARU(143)}.
The $\mathrm{Z}'^0$ piece can be switched off, e.g. by \ttt{MSTP(44)=4}.
An ordinary $\mathrm{Z}^0$, i.e. particle code 23, cannot be made to
decay into $\H^+ \H^-$, however.
A major potential source of charged Higgs production is top decay.
When the top is treated as a resonance (the default option), it is
possible to switch on the decay channel $\t \to \b \H^+$. Top will
then decay to $\H^+$ a fraction of the time, whichever way it is
produced. The branching ratio is automatically calculated, based
on the $\tan\beta$ value and masses.
It is possible to only have the $\H^+$ decay mode switched on,
in which case the cross section is reduced accordingly. If one
instead assumes that top hadrons are formed, branching ratios
are not automatically calculated. However, you can set, for
the generic top hadron 86, the branching ratios for the two main
channels $\t \to \b \H^+$ and $\t \to \b \mathrm{W}^+$. In this option
the cross section for top production will not be reduced if only
the $\t \to \b \H^+$ decay is switched on, cf. section
\ref{sss:resdecaycross}.
\subsection{Non-Standard Physics}
The number of possible non-Standard Model scenarios is essentially
infinite, but many of the studied scenarios still share a lot of
aspects. For instance, new $\mathrm{W}'$ and $\mathrm{Z}'$ gauge bosons can
arise in a number of different ways. Therefore it still makes sense
to try to cover a few basic classes of particles, with enough
freedom in couplings that many kinds of detailed scenarios can be
accommodated by suitable parameter choices. We have already seen one
example of this, in the extended Higgs sector above. In this section
a few other kinds of non-standard generic physics is discussed.
Clearly many others could have been included, but there is probably
only one glaring omission: currently no supersymmetric particle
production has been included. One main reason for this is the large
number of particles, processes, possible mass hierarchies and decay
chains.
\subsubsection{Fourth-generation fermions}
\ttt{MSEL} = 7, 8, 37, 38 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
1 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma^* / \Z^0$ \\
2 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}^+$ \\
81 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ \\
82 & $\mathrm{g} \mathrm{g} \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ \\
83 & $\mathrm{q}_i \mathrm{f}_j \to \mathrm{Q}_k \mathrm{f}_l$ \\
84 & $\mathrm{g} \gamma \to \mathrm{Q}_k \overline{\mathrm{Q}}_k$ \\
85 & $\gamma \gamma \to \mathrm{F}_k \overline{\mathrm{F}}_k$ \\
141 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma/\mathrm{Z}^0/\mathrm{Z}'^0$ \\
142 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}'^+$ \\
\end{tabular}
The prospects of a fourth generation currently seem rather dim, but
the appropriate flavour content is still found in the program. In
fact, the fourth generation is included on an equal basis with the
first three, provided \ttt{MSTP(1)=4}. Also processes other than
the ones above can therefore be used, e.g. all other processes with
gauge bosons, including non-standard ones such as the $\mathrm{Z}'^0$. We
therefore do not repeat the descriptions found elsewhere, e.g. how
to set only the desired flavour in processes 81--85. Note that it
may be convenient to set \ttt{CKIN(1)} and other cuts such that the
mass of produced gauge bosons is enough for the wanted particle
production --- in principle the program will cope even without that,
but possibly at the expense of very slow execution.
\subsubsection{New gauge bosons}
\ttt{MSEL} = 21, 22, 24 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
141 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \gamma/\mathrm{Z}^0/\mathrm{Z}'^0$ \\
142 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{W}'^+$ \\
144 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{R}$ \\
\end{tabular}
The $\mathrm{Z}'^0$ of subprocess 141 contains the full $\gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0$
interference structure for couplings to fermion pairs. With
\ttt{MSTP(44)} it is possible to pick only a subset, e.g. only the
pure $\mathrm{Z}'^0$ piece. The couplings of the $\mathrm{Z}'^0$ to quarks and leptons
can be set via \ttt{PARU(121) - PARU(128)}. The eight numbers
correspond to the vector and axial couplings of down-type quarks,
up-type quarks, leptons and neutrinos, respectively. The default
corresponds to the same couplings as that of the Standard Model
$\mathrm{Z}^0$, with axial couplings $a_{\mathrm{f}} = \pm 1$ and vector couplings
$v_{\mathrm{f}} = a_{\mathrm{f}} - 4 e_{\mathrm{f}} \sin^2 \! \theta_W$. This implies a resonance width
that increases linearly with the mass. By a suitable choice of the
parameters, it is possible to simulate just about any imaginable
$\mathrm{Z}'^0$ scenario, with full interference effects in cross sections
and decay angular distributions.
The coupling to the decay channel $\mathrm{Z}'^0 \to \mathrm{W}^+ \mathrm{W}^-$ is regulated
by \ttt{PARU(129) - PARU(130)}. The former gives the strength of the
coupling, which determines the rate. The default, \ttt{PARU(129)=1.},
corresponds to the `extended gauge model' of \cite{Alt89}, wherein
the $\mathrm{Z}^0 \to \mathrm{W}^+ \mathrm{W}^-$ coupling is used, scaled down by a factor
$m_{\mathrm{W}}^2/m_{\mathrm{Z}'}^2$, to give a $\mathrm{Z}'^0$ partial width into this channel
that again increases linearly. If this factor is cancelled, by having
\ttt{PARU(129)} proportional to $m_{\mathrm{Z}'}^2/m_{\mathrm{W}}^2$, one obtains a
partial width that goes like the fifth power of the $\mathrm{Z}'^0$ mass, the
`reference model' of \cite{Alt89}. In the decay angular distribution
one could imagine a much richer structure than is given by the one
parameter \ttt{PARU(130)}.
Other decay modes include $\mathrm{Z}'^0 \to \mathrm{Z}^0 \H^0$, predicted in
left--right symmetric models (see \ttt{PARU(145)} and ref.
\cite{Coc91}), and a number of other Higgs decay channels, see
sections \ref{sss:extneutHclass} and \ref{sss:chHclass}.
The $\mathrm{W}'^{\pm}$ of subprocess 142 so far does not contain
interference with the Standard Model $\mathrm{W}^{\pm}$ --- in practice this
should not be a major limitation. The couplings of the $\mathrm{W}'$ to
quarks and leptons are set via \ttt{PARU(131) - PARU(134)}.
Again one may set vector and axial couplings freely, separately
for the $\mathrm{q} \overline{\mathrm{q}}'$ and the $\ell \nu_{\ell}$ decay channels.
The defaults correspond to the $V-A$ structure of the Standard Model
$\mathrm{W}$, but can be changed to simulate a wide selection of models.
One possible limitation is that the same Cabibbo--Kobayashi--Maskawa
quark mixing matrix is assumed as for the standard $\mathrm{W}$.
The coupling $\mathrm{W}' \to \mathrm{Z}^0 \mathrm{W}$ can be set via
\ttt{PARU(135) - PARU(136)}. Further comments on this channel as for
$\mathrm{Z}'$; in particular, default couplings again agree with
the `extended gauge model' of \cite{Alt89}. A $\mathrm{W}' \to \mathrm{W} \H^0$
channel is also included, in analogy with the $\mathrm{Z}'^0 \to \mathrm{Z}^0 \H^0$
one, see \ttt{PARU(146)}.
The $\mathrm{R}$ boson (particle code 40) of subprocess 144 represents one
possible scenario for a horizontal gauge boson, i.e. a gauge boson
that couples between the generations, inducing processes like
$\mathrm{s} \overline{\mathrm{d}} \to \mathrm{R}^0 \to \mu^- \mathrm{e}^+$. Experimental limits on
flavour-changing neutral currents forces such a boson to be fairly
heavy. The model implemented is the one described in \cite{Ben85a}.
\subsubsection{Leptoquarks}
\label{sss:LQclass}
\ttt{MSEL} = 25 \\
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
145 & $\mathrm{q}_i \ell_j \to \L_{\mathrm{Q}}$ \\
162 & $\mathrm{q} \mathrm{g} \to \ell \L_{\mathrm{Q}}$ \\
163 & $\mathrm{g} \mathrm{g} \to \L_{\mathrm{Q}} \br{\L}_{\mathrm{Q}}$ \\
164 & $\mathrm{q}_i \overline{\mathrm{q}}_i \to \L_{\mathrm{Q}} \br{\L}_{\mathrm{Q}}$ \\
\end{tabular}
Several processes that can generate a leptoquark have been included.
Currently only one leptoquark has been implemented, as particle 39,
denoted $\L_{\mathrm{Q}}$. The leptoquark is assumed to carry specific quark
and lepton quantum numbers, by default $\u$ quark plus electron.
These flavour numbers are conserved, i.e. a process such as
$\u \mathrm{e}^- \to \L_{\mathrm{Q}} \to \d \nu_{\mathrm{e}}$ is not allowed. This may be a
bit restrictive, but it represents one of many leptoquark possibilities.
The spin of the leptoquark is assumed to be zero, i.e. its decay is
isotropical.
Although only one leptoquark is implemented, its flavours may be
changed arbitrarily to study the different possibilities. The
flavours of the leptoquark are defined by the quark and lepton
flavours in the decay mode list. Since only one decay channel is
allowed, this means that the quark flavour is stored in
\ttt{KFDP(MDCY(39,2),1)} and the lepton one in
\ttt{KFDP(MDCY(39,2),2)}. The former must always be a quark, while
the latter could be a lepton or an antilepton; a charge-conjugate
partner is automatically defined by the program. At initialization,
the charge is recalculated as a function of the flavours defined;
also the leptoquark name is redefined to be of the type
\ttt{'LQ\_(q)(l)'}, where actual quark \ttt{(q)} and lepton \ttt{(l)}
flavours are displayed.
The $\L_{\mathrm{Q}} \to \mathrm{q} \ell$ vertex contains an undetermined Yukawa
coupling strength, which affects both the width of the leptoquark
and the cross section for many of the production graphs. This
strength may be changed in \ttt{PARU(151)}. The definition of
\ttt{PARU(151)} corresponds to the $k$ factor of \cite{Hew88}, i.e.
to $\lambda^2/(4\pi\alpha_{\mathrm{em}})$, where $\lambda$ is the Yukawa
coupling strength of \cite{Wud86}. Note that \ttt{PARU(151)}
is thus quadratic in the coupling.
The leptoquark is likely to be fairly long-lived, in which case it
has time to fragment into a mesonic- or baryonic-type state, which
would decay later on. This is a bit tedious to handle; therefore
the leptoquark is always assumed to decay before fragmentation has
to be considered. This may give some imperfections in the event
generation, but should not be off by much in the final analysis.
Inside the program, the leptoquark is treated as a resonance.
Since it carries colour, some extra care is required.
In particular, it is not allowed to put the leptoquark stable,
by modifying either \ttt{MDCY(39,1)} or \ttt{MSTP(41)}: then the
leptoquark would be handed undecayed to {\tsc{Jetset}}, which would try
to fragment it (as it does with any other coloured object), and
most likely crash.
\subsubsection{Compositeness and anomalous couplings}
\label{sss:ancoupclass}
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
11 & $\mathrm{f}_i \mathrm{f}_j \to \mathrm{f}_i \mathrm{f}_j$ (QCD) \\
12 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{f}_k \overline{\mathrm{f}}_k$ \\
20 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \gamma \mathrm{W}^+$ \\
165 & $\mathrm{f}_i \overline{\mathrm{f}}_i \to \mathrm{f}_k \overline{\mathrm{f}}_k$ (via $\gamma^* / \Z^0$) \\
166 & $\mathrm{f}_i \overline{\mathrm{f}}_j \to \mathrm{f}_k \overline{\mathrm{f}}_l$ (via $\mathrm{W}^{\pm}$) \\
\end{tabular}
Some processes have been set up to allow anomalous coupling to be
introduced, in addition to the Standard Model ones. These can be
switched on by \ttt{MSTP(5)}$\geq 1$; the default \ttt{MSTP(5)=0}
corresponds to the Standard Model behaviour.
In processes 11 and 12, the quark substructure is included in the
left--left isoscalar model \cite{Eic84,Chi90} for
\ttt{MSTP(5)=1}, with compositeness
scale $\Lambda$ given in \ttt{PARU(155)} (default 1000~GeV) and sign
$\eta$ of interference term in \ttt{PARU(156)} (default $+1$; only
other alternative $-1$). The above model assumes that only $\u$ and
$\d$ quarks are composite (at least at the scale studied); with
\ttt{MSTP(5)=2} compositeness terms are included in the
interactions between all quarks.
The processes 165 and 166 are basically equivalent to 1 and 2, i.e.
$\gamma^* / \Z^0$ and $\mathrm{W}^{\pm}$ exchange, respectively, but a bit less
fancy (no mass-dependent width etc.). The reason for this duplication
is that the resonance treatment formalism of processes 1 and 2 could
not easily be extended to include other than $s$-channel graphs.
In processes 165 and 166, only one final-state flavour
is generated at the time; this flavour should be set in
\ttt{KFPR(165,1)} and \ttt{KFPR(166,1)}, respectively. For process
166 one gives the down-type flavour, and the program will associate
the up-type flavour of the same generation. Defaults are 11 in both
cases, i.e. $\e^+\e^-$ and $\mathrm{e}^+ \nu_{\mathrm{e}}$ ($\mathrm{e}^- \br{\nu}_{\mathrm{e}}$) final
states. While \ttt{MSTP(5)=0} gives the Standard Model results,
\ttt{MSTP(5)=1} contains the left--left isoscalar model (which
does not affect process 166), and \ttt{MSTP(5)=3} the
helicity-non-conserving model (which affects both) \cite{Eic84,Lan91}.
Both models above assume that only $\u$ and $\d$ quarks are
composite; with \ttt{MSTP(5)=} 2 or 4, respectively, contact terms
are included for all quarks in the initial state. Parameters are
\ttt{PARU(155)} and \ttt{PARU(156)}, as above.
Note that processes 165 and 166 are bookkept as $2 \to 2$ processes,
while 1 and 2 are $2 \to 1$ ones. This means that the default $\mathrm{Q}^2$
scale in parton distributions is $p_{\perp}^2$ for the former and $\hat{s}$
for the latter. To make contact between the two, it is recommended to
set \ttt{MSTP(32)=4}, so as to use $\hat{s}$ as scale also for
processes 165 and 166.
In process 20, for $\mathrm{W} \gamma$ pair production, it is possible to set
an anomalous magnetic moment for the $\mathrm{W}$ in \ttt{PARU(153)}
($= \eta = \kappa-1$; where $\kappa = 1$ is the Standard Model
value). The production process is affected according to the formulae
of \cite{Sam91}, while $\mathrm{W}$ decay currently remains
unaffected. It is necessary to set \ttt{MSTP(5)=1} to enable this
extension.
\subsubsection{Excited fermions}
\label{sss:qlstarclass}
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
147 & $\d \mathrm{g} \to \d^*$ \\
148 & $\u \mathrm{g} \to \u^*$ \\
167 & $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \d^*$ \\
168 & $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \u^*$ \\
\end{tabular}
Compositeness scenarios may also give rise to sharp resonances of
excited quarks and leptons. If \ttt{MSTP(6)=1}, then at
initialization the standard fourth generation of fermions will be
overwritten, and made to correspond to an excited copy of the first
generation, consisting of spin $1/2$ particles $\d^*$ (code 7),
$\u^*$ (8), $\mathrm{e}^*$ (17) and $\nu^*_{\mathrm{e}}$ (18). Since the original
fourth-generation information is lost, it is then not possible to
generate fourth-generation particles in the same run.
The current implementation contains gauge interaction production
by quark--gluon fusion (processes 147 and 148) and contact interaction
production by quark--quark or quark--antiquark scattering (processes
167 and 168) . The couplings $f$, $f'$ and $f_s$ to the
{\bf SU(2)}, {\bf U(1)} and {\bf SU(3)} groups are stored in
\ttt{PARU(157) - PARU(159)}, the scale parameter $\Lambda$ in
\ttt{PARU(155)}; you are also expected to change the $\mathrm{f}^*$ masses
in accordance with what is desired --- see \cite{Bau90}
for details on conventions. Decay processes are of
the types $\mathrm{q}^* \to \mathrm{q} \mathrm{g}$, $\mathrm{q}^* \to \mathrm{q} \gamma$,
$\mathrm{q}^* \to \mathrm{q} \mathrm{Z}^0$ or $\mathrm{q}^* \to \mathrm{q}' \mathrm{W}^{\pm}$.
A non-trivial angular dependence is included in the $\mathrm{q}^*$
decay for processes 147 and 148, but has not been included for
processes 167 and 168.
\subsubsection{Technicolor}
\label{sss:technicolorclass}
ISUB = \\
\begin{tabular}{rl@{\protect\rule{0mm}{\tablinsep}}}
149 & $\mathrm{g} \mathrm{g} \to \eta_{\mrm{techni}}$ \\
\end{tabular}
The technicolor scenario offers an alternative to the ordinary Higgs
mechanism for giving masses to the $\mathrm{W}$ and $\mathrm{Z}$. The technicolor
gauge group is an analogue of QCD, with a rich spectrum of
technimesons made out of techniquarks. Three of the technipions
assume the role of the longitudinal components of the $\mathrm{W}$ and $\mathrm{Z}$
bosons, but many other states remain as separate particles.
No fully realistic model has been found so far, however, so any
phenomenology has to be taken as indicative only.
In section \ref{sss:heavySMHclass} it is discussed how processes
71--77, in some of its options, can be used to simulate a scenario
with techni-$\rho$ resonances in longitudinal gauge boson scattering.
Here we present another process, that of the production of a
techni-$\eta$. This particle has zero spin, is a singlet under
electroweak {\bf SU(2)}$\times${\bf U(1)}, but carries octet colour
charge. It is one of the possible techni-$\pi$ particles; the name
techni-$\eta$ is part of a subclassification not used by all authors.
The techni-$\eta$ couples to ordinary fermions according to the fermion
squared mass. The dominant decay mode is therefore $\t \overline{\mathrm{t}}$, if
allowed. The coupling to a $\mathrm{g} \mathrm{g}$ state is roughly comparable with
that to $\b \overline{\mathrm{b}}$. Production at hadron colliders is therefore
predominantly through $\mathrm{g} \mathrm{g}$ fusion, as implemented in process 149.
The two main free parameters are the techni-$\eta$ mass and the decay
constant $F_{\pi}$. The latter appears inversely quadratically in all
the partial widths. Also the total cross section is affected, since
the cross section is proportional to the $\mathrm{g} \mathrm{g}$ partial width.
$F_{\pi}$ is stored in \ttt{PARP(46)} and has the default value 123
GeV, which is the number predicted in some models.
\subsection{Main Processes by Machine}
In the previous section we have already commented on which processes
have limited validity, or have different meanings (according to
conventional terminology) in different contexts. Let us just repeat
a few of the main points to be remembered for different machines.
\subsubsection{$\e^+\e^-$ collisions}
The main annihilation process is number 1, $\e^+\e^- \to \mathrm{Z}^0$,
where in fact the full $\gamma^*/\mathrm{Z}^0$ interference structure is
included. This process can be used, with some confidence, for
c.m. energies from about 4 GeV upwards, i.e. at DORIS/CESR,
PETRA/PEP, TRISTAN, LEP, and any future linear colliders.
(To get below 10 GeV, you have to change \ttt{PARP(2)}, however.)
This is the default process obtained when \ttt{MSEL=1}, i.e.
when you do not change anything yourself.
Process 141 contains a $\mathrm{Z}'^0$, including
full interference with the standard $\gamma^* / \Z^0$. With the value
\ttt{MSTP(44)=4} in fact one is back at the standard $\gamma^* / \Z^0$
structure, i.e. the $\mathrm{Z}'^0$ piece has been switched off. Even so,
this process may be useful, since it can simulate e.g.
$\e^+\e^- \to \H^0 \mathrm{A}^0$. Since the $\H^0$ may in its turn decay to
$\mathrm{Z}^0 \mathrm{Z}^0$, a decay channel of the ordinary $\mathrm{Z}^0$ to
$\H^0 \mathrm{A}^0$, although physically correct, would be technically
confusing. In particular, it would be messy to set the original
$\mathrm{Z}^0$ to decay one way and the subsequent ones another. So, in
this sense, the $\mathrm{Z}'^0$ could be used as a copy of the ordinary
$\mathrm{Z}^0$, but with a distinguishable label.
The process $\e^+\e^- \to \Upsilon$ does not exist as a separate process
in {\tsc{Pythia}}, but can be simulated by using \ttt{LUONIA}, see section
\ref{ss:oniadecays}.
At LEP 2 and even higher energy machines, the simple $s$-channel
process 1 will lose out to other processes, such as
$\e^+\e^- \to \mathrm{Z}^0 \mathrm{Z}^0$ and $\e^+\e^- \to \mathrm{W}^+ \mathrm{W}^-$, i.e. processes
22 and 25. The former process in fact includes the structure
$\e^+\e^- \to (\gamma^* / \Z^0)(\gamma^* / \Z^0)$, which means that the cross section
is singular if either of the two $\gamma^* / \Z^0$ masses is allowed to
vanish. A mass cut therefore needs to be introduced, and is
actually also used in other processes, such as $\e^+\e^- \to \mathrm{W}^+ \mathrm{W}^-$.
For practical
applications, both with respect to cross sections and to event
shapes, it is imperative to include initial-state radiation effects.
Therefore \ttt{MSTP(11)=1} is the default, wherein exponentiated
electron-inside-electron distributions are used to give the
momentum of the actually interacting electron. By radiative
corrections to process 1, such processes as $\e^+\e^- \to \gamma \mathrm{Z}^0$
are therefore automatically generated. If process 19 were to be
used at the same time, this would mean that radiation were to be
double-counted. In the alternative \ttt{MSTP(11)=0}, electrons are
assumed to deposit their full energy in
the hard process, i.e. initial-state QED radiation is not included.
This option is very useful, since it often corresponds to the
`ideal' events that one wants to correct back to.
Resolved electrons also means that one may have interactions
between photons. This opens up the whole field of $\gamma\gamma$
processes, which is described in section \ref{sss:photoprodclass}.
In particular, with \ttt{MSTP(12)=1} photons may be resolved, i.e.
photons need not only interact point-like, but can also interact
like a hadron with a partonic substructure. The whole menagerie
of hadron--hadron collider processes can then be accessed. However,
it is not yet possible to include the low-$p_{\perp}$ processes with a
variable photon energy spectrum. That is, to generate the `total'
$\gamma\gamma$ spectrum, the program also has to be initialized for
a $\gamma\gamma$ collider.
The thrust of the {\tsc{Pythia/Jetset}} programs is towards processes that involve
hadron production, one way or another. Because of generalizations
from other areas, also a few completely
non-hadronic processes are available. These include Bhabha
scattering, $\e^+\e^- \to \e^+\e^-$ in process 10, and photon pair production,
$\e^+\e^- \to \gamma \gamma$ in process 18. However, note that the
precision that could be expected in a {\tsc{Pythia}} simulation of those
processes is certainly far less than that of dedicated programs.
For one thing, electroweak loop effects are not included.
For another, nowhere is the electron mass taken into account,
which means that explicit cut-offs at some minimum $p_{\perp}$ are always
necessary.
\subsubsection{Lepton--hadron collisions}
The issue of applications to $\mathrm{e}\mathrm{p}$ colliders has been covered in a
recent report \cite{Sjo92b}.
The default process for a lepton--hadron collider is deep inelastic
scattering, $\ell \mathrm{q} \to \ell' \mathrm{q}'$, of process 10. This includes
$\gamma^0/\mathrm{Z}^0/\mathrm{W}^{\pm}$ exchange, with full interference, as
described in section \ref{sss:DISclass}. Radiation off the incoming
lepton leg is included by \ttt{MSTP(11)=1} and off the outgoing
one by \ttt{MSTJ(41)=2} (both are default). Note that both
QED and QCD radiation (off the $\mathrm{e}$ and the $\mathrm{q}$ legs, respectively)
are allowed to modify the $x$ and $Q^2$ values of the process, while
the conventional approach in the literature is to allow only the
former. Therefore an option (on by default) has been added to preserve
these values by a post-facto rescaling, \ttt{MSTP(23)=1}.
In terms of cross sections, a more important set of processes are
related to photoproduction, either with a point-like or with a
resolved photon, see section \ref{sss:photoprodclass}. A complete
description of photoproduction is available \cite{Sch93,Sch93a}, but
needs three separate runs for the three distinct behaviours of a
photon: point-like, VMD resolved and anomalous resolved.
\subsubsection{Hadron--hadron collisions}
The default is to include QCD jet production by $2 \to 2$ processes,
see section \ref{sss:QCDjetclass}. Since the differential
cross section is divergent for $p_{\perp} \to 0$, a lower cut-off has to
be introduced. Normally that cut-off is given by the user-set
$p_{\perp\mathrm{min}}$ value in \ttt{CKIN(3)}. If \ttt{CKIN(3)} is chosen
smaller than a given value of the order of 2 GeV (see \ttt{PARP(81)}
and \ttt{PARP(82)}), then low-$p_{\perp}$ events are also switched on.
The jet cross section is regularized at low $p_{\perp}$, so as to
obtain a smooth joining between the high-$p_{\perp}$ and the low-$p_{\perp}$
descriptions, see further section \ref{ss:multint}.
As \ttt{CKIN(3)} is varied, the jump from one
scenario to another is abrupt, in terms of cross section: in a
high-energy hadron collider, the cross section for jets down to
a $p_{\perp\mathrm{min}}$ scale of a few GeV can well reach values much
larger than the total inelastic, non-diffractive cross section.
Clearly this is nonsense; therefore either $p_{\perp\mathrm{min}}$ should
be picked so large that the jet cross section be only a fraction
of the total one, or else one should select $p_{\perp\mathrm{min}} = 0$ and
make use of the full description.
If one switches to \ttt{MSEL=2}, also elastic and diffractive
processes are switched on, see section \ref{sss:minbiasclass}.
However, the simulation of these processes is fairly primitive,
and should not be used for dedicated studies,
but only to estimate how much they may contaminate the class of
non-diffractive minimum bias events.
Most processes can be simulated in hadron colliders, since the
bulk of {\tsc{Pythia}} processes can be initiated by quarks or gluons.
However, there are limits. Currently we include no photon
or lepton parton distributions, which means that a process like
$\gamma \mathrm{q} \to \gamma \mathrm{q}$ is not accessible. Further, the
possibility of having $\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$ interacting in
processes such as 71--77 has been hardwired process by process,
and does not mean that there is a generic treatment of $\mathrm{Z}^0$ and
$\mathrm{W}^{\pm}$ distributions.
The emphasis in the hadron--hadron process description is on high
energy hadron colliders. The program can be used also at
fixed-target energies, but the multiple interaction model for underlying
events then breaks down and should not be used. The limit of
applicability is somewhere at around 100 GeV. Below that, one is also
recommended to change to \ttt{MSTP(92)=3}, to obtain a reasonable
amount of beam remnant particle production in the absence of
multiple interactions.
\clearpage
\section{The PYTHIA Program Elements}
In the previous two sections, the physics processes and the
event-generation schemes of {\tsc{Pythia}} have been presented. Here, finally,
the event-generation routines and the common block variables are
described. However, routines and variables related to initial- and
final-state showers, beam remnants and underlying events, and
fragmentation and decay are relegated to subsequent sections on
these topics. Further, for historical reasons, many adjustable
coupling constants are found in the \ttt{LUDAT1} common block in
{\tsc{Jetset}}, rather than somewhere in the {\tsc{Pythia}} common blocks;
these parameters are described in section \ref{ss:coupcons}.
In the presentation in this section, information less important for
an efficient use of {\tsc{Pythia}} has been put closer to the end.
We therefore begin with the main event generation routines, and
follow this by the main common block variables.
It is useful to distinguish three phases in a normal run with {\tsc{Pythia}}.
In the first phase, the initialization, the general character of the
run is determined. At a minimum, this requires the specification of the
incoming hadrons and the energies involved. At the discretion of the
user, it is also possible to select specific final states, and to make
a number of decisions about details in the subsequent generation.
This step is finished by a \ttt{PYINIT} call, at which time several
variables are initialized in accordance with the values set. The
second phase consists of the main loop over the number of events,
with each new event being generated by a \ttt{PYEVNT} call. This event
may then be analysed, using information stored in some common blocks,
and the statistics accumulated. In the final phase, results are
presented. This may often be done without the invocation of any {\tsc{Pythia}}
routines. From \ttt{PYSTAT}, however, it is possible to obtain a
useful list of cross sections for the different subprocesses.
\subsection{The Main Subroutines}
\label{ss:PYTmainroutines}
There are two routines that you must know: \ttt{PYINIT} for
initialization and \ttt{PYEVNT} for the subsequent generation of
each new event. In addition, the cross section and other kinds of
information available with \ttt{PYSTAT} are frequently useful. The
other three routines described here, \ttt{PYFRAM}, \ttt{PYKCUT},
and \ttt{PYEVWT}, are of more specialized interest.
\drawbox{CALL PYINIT(FRAME,BEAM,TARGET,WIN)}\label{p:PYINIT}
\begin{entry}
\itemc{Purpose:} to initialize the generation procedure.
\iteme{FRAME :} a character variable used to specify the frame of the
experiment. Upper-case and lower-case letters may be freely mixed.
\begin{subentry}
\iteme{= 'CMS' :} colliding beam experiment in c.m. frame, with beam
momentum in $+z$ direction and target momentum in $-z$ direction.
\iteme{= 'FIXT' :} fixed-target experiment, with beam particle
momentum pointing in $+z$ direction.
\iteme{= 'USER' :} full freedom to specify frame by giving beam
momentum in \ttt{P(1,1)}, \ttt{P(1,2)} and \ttt{P(1,3)} and target
momentum in \ttt{P(2,1)}, \ttt{P(2,2)} and \ttt{P(2,3)} in
common block \ttt{LUJETS}. Particles are assumed on the mass shell,
and energies are calculated accordingly.
\iteme{= 'FOUR' :} as \ttt{'USER'}, except also energies should be
specified, in \ttt{P(1,4)} and \ttt{P(2,4)}, respectively. The
particles need not be on the mass shell; effective masses are
calculated from energy and momentum. (But note that numerical
precision may suffer; if you know the masses the option \ttt{'FIVE'}
below is preferrable.)
\iteme{= 'FIVE' :} as \ttt{'USER'}, except also energies and masses
should be specified, i.e the full momentum information in
\ttt{P(1,1) - P(1,5)} and \ttt{P(2,1) - P(2,5)} should be given for
beam and target, respectively. Particles need not be on the mass
shell. Space-like virtualities should be stored as $-\sqrt{-m^2}$.
Four-momentum and mass information must match.
\iteme{= 'NONE' :} there will be no initialization of any
processes, but only of resonance widths and a few other
process-independent variables. Subsequent to such a call,
\ttt{PYEVNT} cannot be used to generate events, so this option
is mainly intended for those who will want to construct their own
events afterwards, but still want to have access to some of the
{\tsc{Pythia}} facilities. In this option, the \ttt{BEAM}, \ttt{TARGET} and
\ttt{WIN} arguments are dummy.
\end{subentry}
\iteme{BEAM, TARGET :} character variables to specify beam and
target particles. Upper-case and lower-case letters may be freely
mixed. An antiparticle may be denoted either by `$\sim$' or
`bar' at the end of the name. It is also possible to leave out the
underscore (`\_') directly after `nu' in neutrino names, and the
charge for proton and neutron.
\begin{subentry}
\iteme{= 'e-' :} electron.
\iteme{= 'e+' :} positron.
\iteme{= 'nu\_e' :} $\nu_{\mathrm{e}}$.
\iteme{= 'nu\_e$\sim$' :} $\br{\nu}_{\mathrm{e}}$.
\iteme{= 'mu-' :} $\mu^-$.
\iteme{= 'mu+' :} $\mu^+$.
\iteme{= 'nu\_mu' :} $\nu_{\mu}$.
\iteme{= 'nu\_mu$\sim$' :} $\br{\nu}_{\mu}$.
\iteme{= 'tau-' :} $\tau^-$.
\iteme{= 'tau+' :} $\tau^+$.
\iteme{= 'nu\_tau' :} $\nu_{\tau}$.
\iteme{= 'nu\_tau$\sim$' :} $\br{\nu}_{\tau}$.
\iteme{= 'gamma' :} photon (real, i.e. on the mass shell).
\iteme{= 'pi0' :} $\pi^0$.
\iteme{= 'pi+' :} $\pi^+$.
\iteme{= 'pi-' :} $\pi^-$.
\iteme{= 'n0' :} neutron.
\iteme{= 'n$\sim$0' :} antineutron.
\iteme{= 'p+' :} proton.
\iteme{= 'p$\sim$-' :} antiproton.
\iteme{= 'Lambda0' :} $\Lambda$ baryon.
\iteme{= 'Sigma-' :} $\Sigma^-$ baryon.
\iteme{= 'Sigma0' :} $\Sigma^0$ baryon.
\iteme{= 'Sigma+' :} $\Sigma^+$ baryon.
\iteme{= 'Xi-' :} $\Xi^-$ baryon.
\iteme{= 'Xi0' :} $\Xi^0$ baryon.
\iteme{= 'Omega-' :} $\Omega^-$ baryon.
\iteme{= 'pomeron' :} the pomeron $\mathrm{I}\!\mathrm{P}$; since pomeron parton
distribution functions have not been defined this option can not
be used currently.
\iteme{= 'reggeon' :} the reggeon $\mathrm{I}\!\mathrm{R}$, with comments as for
the pomeron above.
\end{subentry}
\iteme{WIN :} related to energy of system, exact meaning depends on
\ttt{FRAME}.
\begin{subentry}
\iteme{FRAME='CMS' :} total energy of system (in GeV).
\iteme{FRAME='FIXT' :} momentum of beam particle (in GeV/$c$).
\iteme{FRAME='USER' :} dummy (information is taken from the \ttt{P}
vectors, see above).
\end{subentry}
\end{entry}
\drawbox{CALL PYEVNT}\label{p:PYEVNT}
\begin{entry}
\itemc{Purpose:} to generate one event of the type specified by the
\ttt{PYINIT} call. (This is the main routine, which calls a number
of other routines for specific tasks.)
\end{entry}
\drawbox{CALL PYSTAT(MSTAT)}\label{p:PYSTAT}
\begin{entry}
\itemc{Purpose:} to print out cross-sections statistics, decay
widths, branching ratios, status codes and parameter values.
\ttt{PYSTAT} may be called at any time, after the \ttt{PYINIT} call,
e.g. at the end of the run, or not at all.
\iteme{MSTAT :} specification of desired information.
\begin{subentry}
\iteme{= 1 :} prints a table of how many events of the different
kinds that have been generated and the corresponding
cross sections. All numbers already include the effects of cuts
required by the user in \ttt{PYKCUT}.
\iteme{= 2 :} prints a table of the resonances defined in the
program, with their particle codes (KF), and all allowed decay
channels. (If the number of generations in \ttt{MSTP(1)} is 3,
however, channels involving fourth-generation particles are not
displayed.) For each decay channel is shown the sequential channel
number (IDC) of the {\tsc{Jetset}} decay tables, the partial decay width,
branching ratio and effective branching ratio (in the event
some channels have been excluded by the user).
\iteme{= 3 :} prints a table with the allowed hard interaction
flavours \ttt{KFIN(I,J)} for beam and target particles.
\iteme{= 4 :} prints a table of the kinematical cuts \ttt{CKIN(I)}
set by the user in the current run.
\iteme{= 5 :} prints a table with all the values of the status codes
\ttt{MSTP(I)} and the parameters \ttt{PARP(I)} used in the current
run.
\end{subentry}
\end{entry}
\drawbox{CALL PYFRAM(IFRAME)}\label{p:PYFRAM}
\begin{entry}
\itemc{Purpose:} to transform an event listing between different
reference frames, if so desired.
\iteme{IFRAME :} specification of frame the event is to be
boosted to.
\begin{subentry}
\iteme{= 1 :} frame specified by user in the \ttt{PYINIT} call.
\iteme{= 2 :} c.m. frame of incoming particles.
\iteme{= 3 :} hadronic c.m. frame of lepton--hadron interaction
events. Mainly intended for deep inelastic scattering, but can also
be used in photoproduction. Note that both the lepton and any
photons radiated off the lepton remain in the event listing,
and have to be removed separately if you only want to study the
hadronic subsystem.
\end{subentry}
\end{entry}
\drawbox{CALL PYKCUT(MCUT)}\label{p:PYKCUT}
\begin{entry}
\itemc{Purpose:} to enable you to reject a given set of
kinematic variables at an early stage of the generation procedure
(before evaluation of cross sections), so as not to spend unnecessary
time on the generation of events that are not wanted.
The routine will not be called unless you require is by
setting \ttt{MSTP(141)=1}, and never if `minimum-bias'-type events
(including elastic and diffractive scattering) are to be generated
as well. A dummy routine \ttt{PYKCUT} is included in the program file,
so as to avoid unresolved external references when the routine is
not used.
\iteme{MCUT :} flag to signal effect of user-defined cuts.
\begin{subentry}
\iteme{= 0 :} event is to be retained and generated in full.
\iteme{= 1 :} event is to be rejected and a new one generated.
\end{subentry}
\itemc{Remark :} at the time of selection, several variables in the
\ttt{MINT} and \ttt{VINT} arrays in the \ttt{PYINT1} common block
contain information that can be used to make the decision. The routine
provided in the program file explicitly reads the variables that
have been defined at the time \ttt{PYKCUT} is called, and also
calculates some derived quantities. The information available
includes subprocess type ISUB, $E_{\mrm{cm}}$, $\hat{s}$, $\hat{t}$,
$\hat{u}$, $\hat{p}_{\perp}$, $x_1$, $x_2$, $x_{\mrm{F}}$, $\tau$, $y$,
$\tau'$, $\cos\hat{\theta}$, and a few more. Some of these may not
be relevant for the process under study, and are then set to zero.
\end{entry}
\drawbox{CALL PYEVWT(WTXS)}\label{p:PYEVWT}
\begin{entry}
\itemc{Purpose:} to allow you to reweight event cross sections,
by process type and kinematics of the hard scattering. There exists
two separate modes of usage, described in the following. \\
For \ttt{MSTP(142)=1}, it is assumed that the cross section of the
process is correctly given by default in {\tsc{Pythia}}, but that one
wishes to generate events biased to a specific region of phase
space. While the \ttt{WTXS} factor therefore multiplies the na\"{\i}ve
cross section in the choice of subprocess type and kinematics,
the produced event comes with a compensating weight
\ttt{PARI(10)=1./WTXS}, which should be used when filling histograms
etc. In the \ttt{PYSTAT(1)} table, the cross sections are unchanged
(up to statistical errors) compared with the standard cross sections,
but the relative composition of events may be changed and need
no longer be in proportion to relative cross sections. A typical
example of this usage is if one wishes to enhance the production
of high-$p_{\perp}$ events; then a weight like
\ttt{WTXS}$=(p_{\perp}/p_{\perp 0})^2$ (with $p_{\perp 0}$ some fixed
number) might be appropriate. \\
For \ttt{MSTP(142)=2}, on the other hand, it is assumed that the true
cross section is really to be modifed by the multiplicative factor
\ttt{WTXS}. The generated events therefore come with unit weight, just
as usual. This option is really equivalent to replacing the basic
cross sections coded in {\tsc{Pythia}}, but allows more flexibility: no
need to recompile the whole of {\tsc{Pythia}}. \\
The routine will not be called unless \ttt{MSTP(142)}$\geq 1$, and
never if `minimum-bias'-type events (including elastic and
diffractive scattering) are to be generated as well. Further,
cross sections for additional multiple interactions or pile-up events
are never affected. A dummy routine \ttt{PYEVWT} is included in the
program file, so as to avoid unresolved external references when the
routine is not used.
\iteme{WTXS:} multiplication factor to ordinary event cross section;
to be set (by you) in \ttt{PYEVWT} call.
\itemc{Remark :} at the time of selection, several variables in the
\ttt{MINT} and \ttt{VINT} arrays in the \ttt{PYINT1} common block
contain information that can be used to make the decision. The routine
provided in the program file explicitly reads the variables that
have been defined at the time \ttt{PYEVWT} is called, and also
calculates some derived quantities. The given list of information
includes subprocess type ISUB, $E_{\mrm{cm}}$, $\hat{s}$, $\hat{t}$,
$\hat{u}$, $\hat{p}_{\perp}$, $x_1$, $x_2$, $x_{\mrm{F}}$, $\tau$, $y$,
$\tau'$, $\cos\hat{\theta}$, and a few more. Some of these may
not be relevant for the process under study, and are then set to
zero.
\itemc{Warning:} the weights only apply to the hard scattering
subprocesses. There is no way to reweight the shape of initial- and
final-state showers, fragmentation, or other aspects of the event.
\end{entry}
\subsection{Switches for Event Type and Kinematics Selection}
\label{ss:PYswitchkin}
By default, if {\tsc{Pythia}} is run for a hadron collider,
only QCD $2 \to 2$ processes are generated,
composed of hard interactions above $p_{\perp\mathrm{min}}=$\ttt{PARP(81)},
with low-$p_{\perp}$ processes added on so as to give the full
(parametrized) inelastic, non-diffractive cross section.
In an $\e^+\e^-$ collider, $\gamma^* / \Z^0$ production is the default, and in
an $\e\p$ one it is deep inelastic scattering. With the help of the
common block \ttt{PYSUBS}, it is possible to select the generation
of another process, or combination of processes. It is also allowed
to restrict the generation to specific incoming partons/particles
at the hard interaction. This often automatically also restricts
final-state flavours but, in processes such as resonance production
or QCD/QED production of new flavours, switches in the {\tsc{Jetset}}
program may be used to this end; see section \ref{ss:parapartdat}.
The \ttt{CKIN} array may be used to impose specific kinematics cuts.
You should here be warned that, if kinematical variables are
too strongly restricted, the generation time per event may become
very long. In extreme cases, where the cuts effectively close the
full phase space, the event generation may run into an infinite
loop. The generation of $2 \to 1$ resonance production is performed
in terms of the $\hat{s}$ and $y$ variables, and so the ranges
\ttt{CKIN(1) - CKIN(2)} and \ttt{CKIN(7) - CKIN(8)} may be
arbitrarily restricted without a significant loss of speed.
For $2 \to 2$ processes, $\cos\hat{\theta}$ is added as a third
generation variable, and so additionally the range
\ttt{CKIN(27) - CKIN(28)} may be restricted without any danger.
Effects from initial- and final-state radiation
are not included, since they are not known at the time the
kinematics at the hard interaction is selected. The sharp
kinematical cut-offs that can be imposed on the generation
process are therefore smeared, both by QCD radiation and by
fragmentation. A few examples of such effects follow.
\begin{Itemize}
\item Initial-state radiation implies that each of the two incoming
partons has a non-vanishing $p_{\perp}$ when they interact. The hard
scattering subsystem thus receives a net transverse boost,
and is rotated with respect to the beam directions.
In a $2 \to 2$ process, what typically happens is that one of the
scattered partons receives an increased $p_{\perp}$, while the $p_{\perp}$
of the other parton is reduced.
\item Since the initial-state radiation
machinery assigns space-like virtualities to the incoming partons,
the definitions of $x$ in terms of energy fractions and in terms of
momentum fractions no longer coincide, and so the interacting
subsystem may receive a net longitudinal boost compared with
na\"{\i}ve expectations, as part of the parton-shower machinery.
\item Initial-state radiation gives rise to
additional jets, which in extreme cases may be mistaken for either
of the jets of the hard interaction.
\item Final-state radiation gives rise to additional jets, which
smears the meaning of the basic $2 \to 2$ scattering. The assignment
of soft jets is not unique. The energy of a jet becomes dependent
on the way it is identified, e.g. what jet cone size is used.
\item The beam remnant description assigns primordial $k_{\perp}$
values, which also gives a net $p_{\perp}$ shift of the hard-interaction
subsystem; except at low energies this effect is overshadowed by
initial-state radiation, however. Beam remnants may also add
further activity under the `perturbative' event.
\item Fragmentation will further broaden jet profiles, and make
jet assignments and energy determinations even more uncertain.
\end{Itemize}
In a study of events within a given window of
experimentally defined variables, it is up to you to leave
such liberal margins that no events are missed. In other words, cuts
have to be chosen such that a negligible fraction of events migrate
from outside the simulated region to inside the interesting region.
Often this may lead to low efficiency in terms of what fraction of
the generated events are actually of interest to you.
See also section \ref{ss:PYTstarted}.
In addition to the variables found in \ttt{PYSUBS}, also those in the
\ttt{PYPARS} common block may be used to select exactly what one wants
to have simulated. These possibilities will be described in the
following subsection.
The notation used above and in the following is that `$\hat{~}$'
denotes internal variables in the hard scattering subsystem,
while `$^*$' is for variables in the c.m. frame of the
event as a whole.
\drawbox{COMMON/PYSUBS/MSEL,MSUB(200),KFIN(2,-40:40),CKIN(200)}%
\label{p:PYSUBS}
\begin{entry}
\itemc{Purpose:} to allow the user to run the program with any desired
subset of processes, or restrict flavours or kinematics. If the
default values, denoted below by (D=\ldots), are not satisfactory,
they must be changed before the \ttt{PYINIT} call.
\boxsep
\iteme{MSEL :}\label{p:MSEL} (D=1) a switch to select between full
user control and some preprogrammed alternatives.
\begin{subentry}
\iteme{= 0 :} desired subprocesses have to be switched on in
\ttt{MSUB}, i.e. full user control.
\iteme{= 1 :} depending on incoming particles, different
alternatives are used. \\
Lepton--lepton: $\mathrm{Z}$ or $\mathrm{W}$ production (ISUB = 1 or 2). \\
Lepton--hadron: deep inelastic scattering (ISUB = 10). \\
Hadron--hadron: QCD high-$p_{\perp}$ processes (ISUB = 11, 12, 13, 28,
53, 68); additionally low-$p_{\perp}$ production if
\ttt{CKIN(3)}$<$\ttt{PARP(81)} or \ttt{PARP(82)}, depending on
\ttt{MSTP(82)} (ISUB = 95). If low-$p_{\perp}$ is switched on, the other
\ttt{CKIN} cuts are not used. \\
A resolved photon counts as hadron, except that an anomalous photon
cannot have low-$p_{\perp}$ interactions. When the photon is not resolved,
the following cases are possible. \\
Photon--lepton: Compton scattering (ISUB = 34). \\
Photon--hadron: photon-parton scattering (ISUB = 33, 34, 54). \\
Photon--photon: fermion pair production (ISUB = 58).
\iteme{= 2 :} as \ttt{MSEL = 1} for lepton--lepton, lepton--hadron
and unresolved photons. For hadron--hadron (including resolved
photons) all QCD processes, including low-$p_{\perp}$, single and
double diffractive and elastic scattering, are included (ISUB =
11, 12, 13, 28, 53, 68, 91, 92, 93, 94, 95). The \ttt{CKIN} cuts are
here not used.
\iteme{= 4 :} charm ($\c\overline{\mathrm{c}}$) production with massive matrix
elements (ISUB = 81, 82, 84, 85).
\iteme{= 5 :} bottom ($\b\overline{\mathrm{b}}$) production with massive matrix
elements (ISUB = 81, 82, 84, 85).
\iteme{= 6 :} top ($\t\overline{\mathrm{t}}$) production with massive matrix
elements (ISUB = 81, 82, 84, 85).
\iteme{= 7 :} low ($\mathrm{l}\br{\mathrm{l}}$) production with
massive matrix elements (ISUB = 81, 82, 84, 85).
\iteme{= 8 :} high ($\mathrm{h}\br{\mathrm{h}}$) production with massive matrix
elements (ISUB = 81, 82, 84, 85).
\iteme{= 10 :} prompt photons (ISUB = 14, 18, 29).
\iteme{= 11 :} $\mathrm{Z}^0$ production (ISUB = 1).
\iteme{= 12 :} $\mathrm{W}^{\pm}$ production (ISUB = 2).
\iteme{= 13 :} $\mathrm{Z}^0$ + jet production (ISUB = 15, 30).
\iteme{= 14 :} $\mathrm{W}^{\pm}$ + jet production (ISUB = 16, 31).
\iteme{= 15 :} pair production of different combinations of $\gamma$,
$\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$ (except $\gamma\gamma$; see \ttt{MSEL} = 10)
(ISUB = 19, 20, 22, 23, 25).
\iteme{= 16 :} $\H^0$ production (ISUB = 3, 102, 103, 123, 124).
\iteme{= 17 :} $\H^0 \mathrm{Z}^0$ or $\H^0 \mathrm{W}^{\pm}$ (ISUB = 24, 26).
\iteme{= 18 :} $\H^0$ production, combination relevant for $\e^+\e^-$
annihilation (ISUB = 24, 103, 123, 124).
\iteme{= 19 :} $\H^0$, $\H'^0$ and $\mathrm{A}^0$ production, excepting pair
production (ISUB = 24, 103, 123, 124, 153, 158, 171, 173, 174, 176,
178, 179).
\iteme{= 21 :} $\mathrm{Z}'^0$ production (ISUB = 141).
\iteme{= 22 :} $\mathrm{W}'^{\pm}$ production (ISUB = 142).
\iteme{= 23 :} $\H^{\pm}$ production (ISUB = 143).
\iteme{= 24 :} $\mathrm{R}^0$ production (ISUB = 144).
\iteme{= 25 :} $\L_{\mathrm{Q}}$ (leptoquark) production (ISUB = 145, 162,
163, 164).
\iteme{= 35:} single bottom production by $\mathrm{W}$ exchange (ISUB = 83).
\iteme{= 36:} single top production by $\mathrm{W}$ exchange (ISUB = 83).
\iteme{= 37:} single low production by $\mathrm{W}$ exchange (ISUB = 83).
\iteme{= 38:} single high production by $\mathrm{W}$ exchange (ISUB = 83).
\end{subentry}
\boxsep
\iteme{MSUB :}\label{p:MSUB} (D=200*0) array to be set when
\ttt{MSEL=0} (for \ttt{MSEL}$\geq 1$ relevant entries are set in
\ttt{PYINIT}) to choose which subset of subprocesses to include
in the generation. The ordering follows the ISUB code given in
section \ref{ss:ISUBcode} (with comments as given there).
\begin{subentry}
\iteme{MSUB(ISUB) = 0 :} the subprocess is excluded.
\iteme{MSUB(ISUB) = 1 :} the subprocess is included.
\itemc{Note:} when \ttt{MSEL=0}, the \ttt{MSUB} values set by the
user are never changed by {\tsc{Pythia}}. If you want to combine several
different `subruns', each with its own \ttt{PYINIT} call, into one
single run, it is up to you to remember not only to switch on
the new processes before each new \ttt{PYINIT} call, but also to
switch off the old ones that are no longer desired.
\end{subentry}
\boxsep
\iteme{KFIN(I,J) :}\label{p:KFIN} provides an option to selectively
switch on and
off contributions to the cross sections from the different incoming
partons/particles at the hard interaction. In combination with the
{\tsc{Jetset}} resonance decay switches, this also allows you to set
restrictions on flavours appearing in the final state.
\begin{subentry}
\iteme{I :} is 1 for beam side of event and 2 for target side.
\iteme{J :} enumerates flavours according to the KF code; see section
\ref{ss:codes}.
\iteme{KFIN(I,J) = 0 :} the parton/particle is forbidden.
\iteme{KFIN(I,J) = 1 :} the parton/particle is allowed.
\itemc{Note:} By default, the following are switched on: $\d$, $\u$,
$\mathrm{s}$, $\c$, $\b$, $\mathrm{e}^-$, $\nu_{\mathrm{e}}$, $\mu^-$, $\nu_{\mu}$, $\tau^-$,
$\nu_{\tau}$, $\mathrm{g}$, $\gamma$, $\mathrm{Z}^0$, $\mathrm{W}^+$ and their antiparticles.
In particular, top is off, and has to be switched on explicitly if
needed.
\end{subentry}
\boxsep
\iteme{CKIN :}\label{p:CKIN} kinematics cuts that can be set by you
before the \ttt{PYINIT} call, and that affect the region of phase space
within which events are generated. Some cuts are `hardwired' while
most are `softwired'. The hardwired ones are directly related to the
kinematical variables used in the event selection procedure,
and therefore have negligible effects on program efficiency.
The most important of these are \ttt{CKIN(1) - CKIN(8)},
\ttt{CKIN(27) - CKIN(28)}, and \ttt{CKIN(31) - CKIN(32)}.
The softwired ones are most of the remaining ones, that cannot
be fully taken into account
in the kinematical variable selection, so that generation in
constrained regions of phase space may be slow. In extreme
cases the phase space may be so small that the maximization
procedure fails to find any allowed points at all (although some
small region might still exist somewhere), and therefore switches
off some subprocesses, or aborts altogether.
\iteme{CKIN(1), CKIN(2) :} (D=2.,-1. GeV) range of allowed
$\hat{m} = \sqrt{\hat{s}}$ values. If \ttt{CKIN(2)}$ < 0.$, the upper
limit is inactive.
\iteme{CKIN(3), CKIN(4) :} (D=0.,-1. GeV) range of allowed
$\hat{p}_{\perp}$ values for hard $2 \to 2$ processes, with
transverse momentum $\hat{p}_{\perp}$ defined in the rest frame of
the hard interaction. If \ttt{CKIN(4)}$ < 0.$, the upper limit is
inactive. For processes that are singular in the limit
$\hat{p}_{\perp} \to 0$
(see \ttt{CKIN(6)}), \ttt{CKIN(5)} provides an additional constraint.
The \ttt{CKIN(3)} and \ttt{CKIN(4)} limits can also be used in
$2 \to 1 \to 2$ processes. Here, however, the product
masses are not known and hence are assumed to be vanishing in the event
selection. The actual $p_{\perp}$ range for massive products is thus
shifted downwards with respect to the nominal one.
\iteme{CKIN(5) :} (D=1. GeV) lower cut-off on $\hat{p}_{\perp}$ values,
in addition to the \ttt{CKIN(3)} cut above, for processes that are
singular in the limit $\hat{p}_{\perp} \to 0$ (see \ttt{CKIN(6)}).
\iteme{CKIN(6) :} (D=1. GeV) hard $2 \to 2$ processes, which do not
proceed only via an intermediate resonance (i.e. are $2 \to 1 \to 2$
processes), are classified as singular in the limit
$\hat{p}_{\perp} \to 0$ if either or both of the two final-state
products has a mass $m < $\ttt{CKIN(6)}.
\iteme{CKIN(7), CKIN(8) :} (D=-10.,10.) range of allowed scattering
subsystem rapidities $y = y^*$ in the c.m. frame of the event,
where $y = (1/2) \ln(x_1/x_2)$. (Following the notation of this
section, the variable should be given as $y^*$, but because of its
frequent use, it was called $y$ in section \ref{ss:kinemtwo}.)
\iteme{CKIN(9), CKIN(10) :} (D=-10.,10.) range of allowed (true)
rapidities for the product with largest rapidity in a $2 \to 2$ or a
$2 \to 1 \to 2$ process, defined in the c.m. frame of the event,
i.e. $\max(y^*_3, y^*_4)$. Note that rapidities are counted with sign,
i.e. if $y^*_3 = 1$ and $y^*_4 = -2$ then $\max(y^*_3, y^*_4) = 1$.
\iteme{CKIN(11), CKIN(12) :} (D=-10.,10.) range of allowed (true)
rapidities for the product with smallest rapidity in a $2 \to 2$ or a
$2 \to 1 \to 2$ process, defined in the c.m. frame of the event,
i.e. $\min(y^*_3, y^*_4)$. Consistency thus requires
\ttt{CKIN(11)}$\leq$\ttt{CKIN(9)} and
\ttt{CKIN(12)}$\leq$\ttt{CKIN(10)}.
\iteme{CKIN(13), CKIN(14) :} (D=-10.,10.) range of allowed
pseudorapidities for the product with largest pseudorapidity
in a $2 \to 2$ or a $2 \to 1 \to 2$ process, defined in the c.m.
frame of the event, i.e. $\max(\eta^*_3, \eta^*_4)$. Note that
pseudorapidities are counted with sign, i.e. if $\eta^*_3 = 1$ and
$\eta^*_4 = -2$ then $\max(\eta^*_3, \eta^*_4) = 1$.
\iteme{CKIN(15), CKIN(16) :} (D=-10.,10.) range of allowed
pseudorapidities for the product with smallest pseudorapidity
in a $2 \to 2$ or a $2 \to 1 \to 2$ process, defined in the c.m.
frame of the event, i.e. $\min(\eta^*_3, \eta^*_4)$. Consistency
thus requires \ttt{CKIN(15)}$\leq$\ttt{CKIN(13)} and
\ttt{CKIN(16)}$\leq$\ttt{CKIN(14)}.
\iteme{CKIN(17), CKIN(18) :} (D=-1.,1.) range of allowed
$\cos\theta^*$ values for the product with largest $\cos\theta^*$
value in a $2 \to 2$ or a $2 \to 1 \to 2$ process, defined in the
c.m. frame of the event, i.e. $\max(\cos\theta^*_3,\cos\theta^*_4)$.
\iteme{CKIN(19), CKIN(20) :} (D=-1.,1.) range of allowed
$\cos\theta^*$ values for the product with smallest $\cos\theta^*$
value in a $2 \to 2$ or a $2 \to 1 \to 2$ process, defined in the
c.m. frame of the event, i.e. $\min(\cos\theta^*_3,\cos\theta^*_4)$.
Consistency thus requires \ttt{CKIN(19)}$\leq$\ttt{CKIN(17)} and
\ttt{CKIN(20)}$\leq$\ttt{CKIN(18)}.
\iteme{CKIN(21), CKIN(22) :} (D=0.,1.) range of allowed $x_1$ values
for the parton on side 1 that enters the hard interaction.
\iteme{CKIN(23), CKIN(24) :} (D=0.,1.) range of allowed $x_2$ values
for the parton on side 2 that enters the hard interaction.
\iteme{CKIN(25), CKIN(26) :} (D=-1.,1.) range of allowed Feynman-$x$
values, where $x_{\mrm{F}} = x_1 - x_2$.
\iteme{CKIN(27), CKIN(28) :} (D=-1.,1.) range of allowed
$\cos\hat{\theta}$ values in a hard $2 \to 2$ scattering, where
$\hat{\theta}$ is the scattering angle in the rest frame of the
hard interaction.
\iteme{CKIN(31), CKIN(32) :} (D=2.,-1. GeV) range of allowed
$\hat{m}' = \sqrt{\hat{s}'}$ values, where $\hat{m}'$ is the mass of
the complete three- or four-body final state in $2 \to 3$ or
$2 \to 4$ processes (while $\hat{m}$, constrained in \ttt{CKIN(1)}
and \ttt{CKIN(2)}, here corresponds to the one- or two-body central
system). If \ttt{CKIN(32)}$ < 0.$, the upper limit is inactive.
\iteme{CKIN(35), CKIN(36) :} (D=0.,-1. GeV$^2$) range of allowed
$|\hat{t}| = - \hat{t}$ values in $2 \to 2$ processes. Note that
for deep inelastic scattering this is nothing but the $Q^2$ scale,
in the limit that initial- and final-state radiation is neglected.
If \ttt{CKIN(36)}$ < 0.$, the upper limit is inactive.
\iteme{CKIN(37), CKIN(38) :} (D=0.,-1. GeV$^2$) range of allowed
$|\hat{u}| = - \hat{u}$ values in $2 \to 2$ processes. If
\ttt{CKIN(38)}$ < 0.$, the upper limit is inactive.
\iteme{CKIN(39), CKIN(40) :} (D=4., -1. GeV$^2$) the $W^2$ range
allowed in DIS processes, i.e. subprocess number 10. If
\ttt{CKIN(40)}$ < 0.$, the upper limit is inactive. Here $W^2$ is
defined in terms of $W^2 = Q^2 (1-x)/x$. This formula is not quite
correct, in that \textit{(i)} it neglects the target mass (for a
proton), and \textit{(ii)} it neglects initial-state photon radiation
off the incoming electron. It should be good enough for loose cuts,
however.
\iteme{CKIN(41) - CKIN(44) :} (D=12.,-1.,12.,-1. GeV) range of allowed
mass values of the two (or one) resonances produced in a `true'
$2 \to 2$ process, i.e. one not (only) proceeding through a single
$s$-channel resonance ($2 \to 1 \to 2$). (These are the ones listed
as $2 \to 2$ in the tables in section \ref{ss:ISUBcode}.)
Only particles with a width above \ttt{PARP(41)} are considered as
bona fide resonances and tested against the \ttt{CKIN}
limits; particles with a smaller width are put on the mass shell
without applying any cuts. The exact interpretation of the \ttt{CKIN}
variables depends on the flavours of the two produced resonances. \\
For two resonances like $\mathrm{Z}^0 \mathrm{W}^+$ (produced from
$\mathrm{f} \mathrm{f}' \to \mathrm{Z}^0\mathrm{W}^+$), which are not identical and which are not
each other's antiparticles, one has \\
\ttt{CKIN(41)}$ < m_1 < $\ttt{CKIN(42)}, and \\
\ttt{CKIN(43)}$ < m_2 < $\ttt{CKIN(44)}, \\
where $m_1$ and $m_2$ are the actually generated masses of the two
resonances, and 1 and 2 are defined by the order in which they are
given in the production process specification. \\
For two resonances like $\mathrm{Z}^0 \mathrm{Z}^0$, which are identical, or
$\mathrm{W}^+ \mathrm{W}^-$, which are each other's antiparticles, one instead
has \\
\ttt{CKIN(41)}$ < \min(m_1,m_2) < $\ttt{CKIN(42)}, and \\
\ttt{CKIN(43)}$ < \max(m_1,m_2) < $\ttt{CKIN(44)}. \\
In addition, whatever limits are set on \ttt{CKIN(1)} and, in
particular, on \ttt{CKIN(2)} obviously affect the masses actually
selected.
\begin{subentry}
\itemc{Note 1:} If \ttt{MSTP(42)=0}, so that no mass smearing is
allowed, the \ttt{CKIN} values have no effect (the same as for
particles with too narrow a width).
\itemc{Note 2:} If \ttt{CKIN(42)}$ < $\ttt{CKIN(41)} it means that
the \ttt{CKIN(42)} limit is inactive; correspondingly, if
\ttt{CKIN(44)}$ < $\ttt{CKIN(43)} then \ttt{CKIN(44)} is inactive.
\itemc{Note 3:} If limits are active and the resonances are
identical, it is up to you to ensure that
\ttt{CKIN(41)}$\leq$\ttt{CKIN(43)} and
\ttt{CKIN(42)}$\leq$\ttt{CKIN(44)}.
\itemc{Note 4:} For identical resonances, it is not possible to
preselect which of the resonances is the lighter one; if, for
instance, one
$\mathrm{Z}^0$ is to decay to leptons and the other to quarks, there is no
mechanism to guarantee that the lepton pair has a mass smaller
than the quark one.
\itemc{Note 5:} The \ttt{CKIN} values are applied to all relevant
$2 \to 2$ processes equally, which may not be what one desires if
several processes are generated simultaneously. Some caution is
therefore urged in the use of the \ttt{CKIN(41) - CKIN(44)} values.
Also in other respects, users are recommended to take proper
care: if a $\mathrm{Z}^0$ is only allowed to decay into $\b\overline{\mathrm{b}}$, for example,
setting its mass range to be 2--8 GeV is obviously not a
good idea.
\itemc{Note 6:} In principle, the machinery should work for any
$2 \to 2$ process with resonances in the final state, but so far
it has only been checked for processes 22--26, so also from this
point some caution is urged.
\end{subentry}
\iteme{CKIN(45) - CKIN(48) :} (D=12.,-1.,12.,-1. GeV) range of allowed
mass values of the two (or one) secondary resonances produced in
a $2 \to 1 \to 2$ process (like $\mathrm{g} \mathrm{g} \to \H^0 \to \mathrm{Z}^0 \mathrm{Z}^0$) or
even a $2 \to 2 \to 4$ (or 3) process (like
$\mathrm{q} \overline{\mathrm{q}} \to \mathrm{Z}^0 \H^0 \to \mathrm{Z}^0 \mathrm{W}^+ \mathrm{W}^-$). Note that these
\ttt{CKIN} values only affect the secondary resonances; the
primary ones are constrained by \ttt{CKIN(1)}, \ttt{CKIN(2)} and
\ttt{CKIN(41) - CKIN(44)} (indirectly, of course, the choice of
primary resonance masses affects the allowed mass range for the
secondary ones). What is considered to be a resonance is defined
by \ttt{PARP(41)}; particles with a width smaller than this are
automatically put on the mass shell. The description closely
parallels the one given for \ttt{CKIN(41) - CKIN(44)}. Thus, for
two resonances that are not identical or each other's
antiparticles, one has \\
\ttt{CKIN(45)}$ < m_1 < $\ttt{CKIN(46)}, and \\
\ttt{CKIN(47)}$ < m_2 < $\ttt{CKIN(48)}, \\
where $m_1$ and $m_2$ are the actually generated masses of the two
resonances, and 1 and 2 are defined by the order in which they
given in the decay channel specification in the program (see e.g.
output from \ttt{PYSTAT(2)} or \ttt{LULIST(12)}). For two
resonances that are identical or each other's antiparticles,
one instead has \\
\ttt{CKIN(45)}$ < \min(m_1,m_2) < $\ttt{CKIN(46)}, and \\
\ttt{CKIN(47)}$ < \max(m_1,m_2) < $\ttt{CKIN(48)}.
\begin{subentry}
\itemc{Notes 1 - 5:} as for \ttt{CKIN(41) - CKIN(44)}, with trivial
modifications.
\itemc{Note 6:} Setting limits on secondary resonance masses is
possible in any of the channels of the allowed types (see above).
However, so far only $\H^0 \to \mathrm{Z}^0 \mathrm{Z}^0$ and $\H^0 \to \mathrm{W}^+ \mathrm{W}^-$
have been fully implemented, such that an arbitrary mass range
below the na\"{\i}ve mass threshold may be picked. For other possible
resonances, any restrictions made on the allowed mass range
are not reflected in the cross section; and further it is not
recommendable to pick mass windows that make a decay on the
mass shell impossible. These limitations will be relaxed in future
versions.
\end{subentry}
\iteme{CKIN(51) - CKIN(56) :} (D=0.,-1.,0.,-1.,0.,-1. GeV) range of
allowed transverse momenta in a true $2 \to 3$ process. Currently two
different alternatives are around. For subprocess 131, the $p_{\perp}$
of the first product (the $\mathrm{Z}^0$) is set by \ttt{CKIN(3)} and
\ttt{CKIN(4)}, while for the quark and antiquark $p_{\perp}$'s one has \\
\ttt{CKIN(51)}$<\min(p_{\perp \mathrm{q}},p_{\perp \overline{\mathrm{q}}})<$\ttt{CKIN(52)},
and \\
\ttt{CKIN(53)}$<\max(p_{\perp \mathrm{q}},p_{\perp \overline{\mathrm{q}}})<$\ttt{CKIN(54)}. \\
Negative \ttt{CKIN(52)} and \ttt{CKIN(54)} values means that the
corresponding limits are inactive. For subprocesses 121--124, and
their $\H'^0$ and $\mathrm{A}^0$ equivalents (173, 174, 178, 179, 181, 182,
186, 187), \ttt{CKIN(51) - CKIN(54)} again corresponds
to $p_{\perp}$ ranges for scattered partons, but in order of appearance,
i.e. \ttt{CKIN(51) - CKIN(52)} for the parton scattered off the beam
and \ttt{CKIN(53) - CKIN(54)} for the one scattered off the target.
\ttt{CKIN(55)} and \ttt{CKIN(56)} here sets $p_{\perp}$ limits for the
third product, the $\H^0$, i.e. the \ttt{CKIN(3)} and \ttt{CKIN(4)}
values have no effect for this process. Since the $p_{\perp}$ of the Higgs
is not one of the primary variables selected, any constraints here
may mean reduced Monte Carlo efficiency, while for these processes
\ttt{CKIN(51) - CKIN(54)}
are `hardwired' and therefore do not cost anything.
\end{entry}
\subsection{The General Switches and Parameters}
\label{ss:PYswitchpar}
The \ttt{PYPARS} common block contains the status code and parameters
that regulate the performance of the program. All of them are
provided with sensible default values, so that a novice user can
neglect them, and only gradually explore the full range of
possibilities. Some of the switches and parameters in \ttt{PYPARS}
will be described later, in the shower and beam remnants sections.
\drawbox{COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)}%
\label{p:PYPARS1}
\begin{entry}
\itemc{Purpose:} to give access to status code and parameters that
regulate the performance of the program. If the default values,
denoted below by (D=\ldots), are not satisfactory, they must in
general be changed before the \ttt{PYINIT} call. Exceptions, i.e.
variables that can be changed for each new event, are denoted by
(C).
\iteme{MSTP(1) :}\label{p:MSTP} (D=3) maximum number of generations.
Automatically set $\leq 4$.
\iteme{MSTP(2) :} (D=1) calculation of $\alpha_{\mathrm{s}}$ at hard interaction,
in the routine \ttt{ULALPS}.
\begin{subentry}
\iteme{= 0 :} $\alpha_{\mathrm{s}}$ is fixed at value \ttt{PARU(111)}.
\iteme{= 1 :} first-order running $\alpha_{\mathrm{s}}$.
\iteme{= 2 :} second-order running $\alpha_{\mathrm{s}}$.
\end{subentry}
\iteme{MSTP(3) :} (D=2) selection of $\Lambda$ value in $\alpha_{\mathrm{s}}$
for \ttt{MSTP(2)}$\geq 1$.
\begin{subentry}
\iteme{= 1 :} $\Lambda$ is given by \ttt{PARP(1)} for hard
interactions, by \ttt{PARP(61)} for space-like showers,
by \ttt{PARP(72)} for time-like showers not from a resonance decay,
and by \ttt{PARJ(81)} for time-like ones from a resonance decay
(including e.g. $\gamma/\mathrm{Z}^0 \to \mathrm{q}\overline{\mathrm{q}}$ decays, i.e. conventional
$\e^+\e^-$ physics). This $\Lambda$ is assumed
to be valid for 5 flavours; for the hard interaction the number of
flavours assumed can be changed by \ttt{MSTU(112)}.
\iteme{= 2 :} $\Lambda$ value is chosen according to the
parton-distribution-function para\-metri\-zations,
i.e. $\Lambda$ = 0.20 GeV for EHLQ1, = 0.29 GeV for EHLQ2,
= 0.20 GeV for DO1, = 0.40 GeV for DO2, = 0.213 GeV for CTEQ2M,
= 0.208 GeV for CTEQ2MS, = 0.208 GeV for CTEQ2MF,
= 0.322 GeV for CTEQ2ML, = 0.190 GeV for CTEQ2L,
= 0.235 GeV for CTEQ2D, = 0.25 GeV for GRV LO,
and similarly for parton-distribution functions in the \tsc{Pdflib}
library (cf. (\ttt{MSTP(51)}, \ttt{MSTP(52)}). The choice is
always based on the proton parton-distribution set selected, i.e.
is unaffected by pion and photon parton-distribution selection.
All the $\Lambda$ values above
are assumed to refer to 4 flavours, and \ttt{MSTU(112)} is set
accordingly. This $\Lambda$ value is used both for the hard
scattering and the initial- and final-state radiation. The
ambiguity in the choice of the $Q^2$ argument still remains (see
\ttt{MSTP(32)}, \ttt{MSTP(64)} and \ttt{MSTJ(44)}). This $\Lambda$
value is used also for \ttt{MSTP(57)=0}, but the sensible choice
here would be to use \ttt{MSTP(2)=0} and have no initial- or
final-state radiation. This option does \textit{not} change the
\ttt{PARJ(81)} value of timelike parton showers in resonance decays,
so that LEP experience on this specific parameter is not overwritten
unwittingly. Therefore \ttt{PARJ(81)} can be updated completely
independently.
\iteme{= 3 :} as \ttt{=2}, except that here also \ttt{PARJ(81)} is
overwritten in accordance with the $\Lambda$ value of the proton
parton-distribution-function set.
\end{subentry}
\iteme{MSTP(4) :} (D=0) treatment of the Higgs sector,
predominantly the neutral one.
\begin{subentry}
\iteme{= 0 :} the $\H^0$ is given the Standard Model Higgs
couplings, while $\H'^0$ and $\mathrm{A}^0$ couplings should be set by
the user in \ttt{PARU(171) - PARU(175)} and
\ttt{PARU(181) - PARU(185)}, respectively.
\iteme{= 1 :} the user should set couplings for all three Higgses,
for the $\H^0$ in \ttt{PARU(161) - PARU(165)}, and for the
$\H'^0$ and $\mathrm{A}^0$ as above.
\iteme{= 2 :} the mass of $\H^0$ in \ttt{PMAS(25,1)} and the
$\tan\beta$ value in \ttt{PARU(141)} are used to derive
$\H'^0$, $\mathrm{A}^0$ and $\H^{\pm}$ masses, and $\H^0$, $\H'^0$,
$\mathrm{A}^0$ and $\H^{\pm}$ couplings, using the relations of the Minimal
Supersymmetric extension of the Standard Model at Born level
\cite{Gun90}. Existing masses and couplings are overwritten by the
derived values. See section \ref{sss:extneutHclass} for discussion
on parameter constraints.
\iteme{= 3:} as \ttt{=2}, but using relations at the one-loop level.
This option is not yet implemented.
\end{subentry}
\iteme{MSTP(5) :} (D=0) presence of anomalous couplings in processes.
\begin{subentry}
\iteme{= 0 :} absent.
\iteme{$\geq$1 :} present, wherever implemented. See section
\ref{sss:ancoupclass} for further details.
\end{subentry}
\iteme{MSTP(6) :} (D=0) usage of the fourth-generation fermions to
simulate other fermion kinds.
\begin{subentry}
\iteme{= 0 :} none, i.e. can be used as a standard fourth
generation.
\iteme{= 1 :} excited fermions, as present in compositeness
scenarios; see section \ref{sss:qlstarclass}.
\end{subentry}
\iteme{MSTP(7) :} (D=0) choice of heavy flavour in subprocesses
81--85. Does not apply for \ttt{MSEL=4-8}, where the MSEL value
always takes precedence.
\begin{subentry}
\iteme{= 0 :} for processes 81--84 (85) the `heaviest' flavour
allowed for gluon (photon) splitting into a quark--antiquark
(fermion--antifermion) pair, as set in the \ttt{MDME} array.
Note that `heavy' is defined as the one with largest KF code,
so that leptons take precedence if they are allowed.
\iteme{= 1 - 8 :} pick this particular quark flavour; e.g.,
\ttt{MSTP(7)=6} means that top will be produced.
\iteme{= 11 - 18 :} pick this particular lepton flavour. Note that
neutrinos are not possible, i.e. only 11, 13, 15 and 17 are
meaningful alternatives. Lepton pair production can only occur in
process 85, so if any of the other processes have been switched on
they are generated with the same flavour as would be obtained in
the option \ttt{MSTP(7)=0}.
\end{subentry}
\iteme{MSTP(8) :} (D=0) choice of electroweak parameters to use
in the decay widths of resonances ($\mathrm{W}$, $\mathrm{Z}$, $\H$, \ldots) and
cross sections (production of $\mathrm{W}$'s, $\mathrm{Z}$'s, $\H$'s, \ldots).
\begin{subentry}
\iteme{= 0 :} everything is expressed in terms of a running
$\alpha_{\mathrm{em}}(Q^2)$ and a fixed $\sin^2 \! \theta_W$, i.e. $G_{\mathrm{F}}$ is nowhere
used.
\iteme{= 1 :} a replacement is made according to
$\alpha_{\mathrm{em}}(Q^2) \to \sqrt{2} G_{\mathrm{F}} m_{\mathrm{W}}^2 \sin^2 \! \theta_W / \pi$
in all widths and cross sections. If $G_{\mathrm{F}}$ and $m_{\mathrm{Z}}$ are
considered as given, this means that $\sin^2 \! \theta_W$ and $m_{\mathrm{W}}$ are the
only free electroweak parameter.
\iteme{= 2 :} a replacement is made as for \ttt{=1}, but additionally
$\sin^2 \! \theta_W$ is constrained by the relation
$\sin^2 \! \theta_W = 1 - m_{\mathrm{W}}^2/m_{\mathrm{Z}}^2$.
This means that $m_{\mathrm{W}}$ remains as a free parameter, but that the
$\sin^2 \! \theta_W$ value in \ttt{PARU(102)} is never used, \textit{except} in
the vector couplings in the combination $v = a - 4 \sin^2 \! \theta_W e$.
This latter degree of freedom enters e.g. for forward-backward
asymmetries in $\mathrm{Z}^0$ decays.
\itemc{Note:} This option does not affect the emission of real photons
in the initial and final state, where $\alpha_{\mathrm{em}}$ is always used. However,
it does affect also purely electromagnetic hard processes, such as
$\mathrm{q} \overline{\mathrm{q}} \to \gamma \gamma$.
\end{subentry}
\iteme{MSTP(11) :} (D=1) use of electron parton distribution in
$\e^+\e^-$ and $\e\p$ interactions.
\begin{subentry}
\iteme{= 0 :} no, i.e. electron carries the whole beam energy.
\iteme{= 1 :} yes, i.e. electron carries only a fraction of beam
energy in agreement with next-to-leading electron parton-distribution
function, thereby including the effects of initial-state
bremsstrahlung.
\end{subentry}
\iteme{MSTP(12) :} (D=0) use of $\mathrm{e}^-$ (`sea', i.e. from
$\mathrm{e} \to \gamma \to \mathrm{e}$), $\mathrm{e}^+$, quark and gluon distribution
functions inside an electron.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on, provided that \ttt{MSTP(11)}$\geq 1$.
Quark and gluons distributions
are obtained by numerical convolution of the photon content
inside an electron (as given by the bremsstrahlung spectrum of
\ttt{MSTP(11)=1}) with the quark and gluon content inside a
photon. The required numerical precision is set by \ttt{PARP(14)}.
Since the need for numerical integration makes this option
somewhat more time-consuming than ordinary parton-distribution
evaluation, one should only use it when studying processes
where it is needed.
\end{subentry}
\iteme{MSTP(13) :} (D=1) choice of $Q^2$ range over which electrons
are assumed to radiate photons; affects normalization of $\mathrm{e}^-$
(sea), $\mathrm{e}^+$, $\gamma$, quark and gluon distributions inside an
electron.
\begin{subentry}
\iteme{= 1 :} range set by $Q^2$ argument of
parton-distribution-function call, i.e. by $Q^2$ scale of the hard
interaction. Therefore
parton distributions are proportional to $\ln(Q^2/m_e^2)$. This is
normally most appropriate for $\e^+\e^-$ annihilation.
\iteme{= 2 :} range set by the user-determined $Q_{\mathrm{max}}^2$, given in
\ttt{PARP(13)}. Parton distributions are assumed to be proportional to
$\ln((Q_{\mathrm{max}}^2/m_e^2)(1-x)/x^2)$. This is normally most appropriate
for photoproduction, where the electron is supposed to go
undetected, i.e. scatter less than $Q_{\mathrm{max}}^2$.
\itemc{Note:} the choice of effective range is especially touchy for
the quark and gluon distributions. An (almost) on-the-mass-shell
photon has a VMD piece that dies away for a virtual photon. A simple
folding of distribution functions does not take this into account
properly. Therefore the contribution from $Q$ values above the
$\rho$ mass should be suppressed. A choice of
$Q_{\mathrm{max}} \approx 1$~GeV is then appropriate for a
photoproduction limit description of physics.
\end{subentry}
\iteme{MSTP(14) :} (D=0) structure of incoming photon beam or
target (does not affect photon inside electron, only photons
appearing as argument in the \ttt{PYINIT} call).
\begin{subentry}
\iteme{= 0 :} a photon is assumed to be point-like (a direct photon),
i.e. can only interact in processes which explicitly contain the
incoming photon, such as $\mathrm{f}_i \gamma \to \mathrm{f}_i \mathrm{g}$ for $\gamma\mathrm{p}$
interactions. In $\gamma\gamma$ interactions both photons are
direct, i.e the main process is $\gamma \gamma \to \mathrm{f}_i \overline{\mathrm{f}}_i$.
\iteme{= 1 :} a photon is assumed to be resolved, i.e. can only interact
through its constituent quarks and gluons, giving either high-$p_{\perp}$
parton--parton scatterings or low-$p_{\perp}$ events. Hard processes are
calculated with the use of the full photon parton distributions.
In $\gamma\gamma$ interactions both photons are resolved.
\iteme{= 2 :} a photon is assumed resolved, but only the VMD piece is
included in the parton distributions, which therefore mainly
are scaled-down versions of the $\rho^0 / \pi^0$ ones. Both high-$p_{\perp}$
parton--parton scatterings and low-pT events are allowed. In
$\gamma\gamma$ interactions both photons are VMD-like.
\iteme{= 3 :} a photon is assumed resolved, but only the anomalous
piece of the photon parton distributions is included. Only high-$p_{\perp}$
parton--parton scatterings are allowed. In $\gamma\gamma$
interactions both photons are anomalous.
\iteme {= 4 :} in $\gamma\gamma$ interactions one photon is direct
and the other resolved. A typical process is thus
$\mathrm{f}_i \gamma \to \mathrm{f}_i \mathrm{g}$. Hard processes are calculated with the use
of the full photon parton distributions for the resolved photon.
Both possibilities of which photon is direct are included, in event
topologies and in cross sections. This option cannot be used in
configurations with only one incoming photon.
\iteme{= 5 :} in $\gamma\gamma$ interactions one photon is direct
and the other VMD-like. Both possibilities of which photon is direct
are included, in event topologies and in cross sections. This option
cannot be used in configurations with only one incoming photon.
\iteme{= 6 :} in $\gamma\gamma$ interactions one photon is direct
and the other anomalous. Both possibilities of which photon is direct
are included, in event topologies and in cross sections. This option
cannot be used in configurations with only one incoming photon.
\iteme{= 7 :} in $\gamma\gamma$ interactions one photon is VMD-like
and the other anomalous. Only high-$p_{\perp}$ parton--parton scatterings
are allowed. Both possibilities of which photon is VMD-like are
included, in event topologies and in cross sections. This option
cannot be used in configurations with only one incoming photon.
\iteme{= 10 :} the VMD, direct and anomalous components of the photon
are automatically mixed. For $\gamma\mathrm{p}$ interactions, this means an
automatic mixture of the three classes 0, 2 and 3 above
\cite{Sch93,Sch93a}, for $\gamma\gamma$ ones a mixture of the
six classes 0, 2, 3, 5, 6 and 7 above \cite{Sch94a}. Various
restrictions exist for this option, as discussed in section
\ref{sss:photoprodclass}.
\itemc{Note:} our best understanding of how to mix event classes is
provided by the option 10 above, which also can be obtained by
combining three (for $\gamma\mathrm{p}$) or six (for $\gamma\gamma$) separate
runs. In a simpler alternative the VMD and anomalous classes are joined
into a single resolved class. Then $\gamma\mathrm{p}$ physics only requires two
separate runs, with 0 and 1, and $\gamma\gamma$ physics requires three,
with 0, 1 and 4.
\end{subentry}
\iteme{MSTP(15) :} (D=5) possibility to modify the nature of the
anomalous photon component (as used with the appropriate \ttt{MSTP(14)}
options), in particular with respect to the scale choices and
cut-offs of hard processes. This option is mainly intended for
comparative studies and should not normally be touched.
\begin{subentry}
\iteme{= 0 :} none, i.e. the same treatment as for the VMD component.
\iteme{= 1 :} evaluate the anomalous parton distributions at a scale
$Q^2/$\ttt{PARP(17)}$^2$.
\iteme{= 2 :} as \ttt{=1}, but instead of \ttt{PARP(17)} use either
\ttt{PARP(81)/PARP(15)} or \ttt{PARP(82)/PARP(15)}, depending on
\ttt{MSTP(82)} value.
\iteme{= 3 :} evaluate the anomalous parton distribution functions
of the photon as $f^{\gamma,\mrm{anom}}(x, Q^2, p_0^2) -
f^{\gamma,\mrm{anom}}(x, Q^2, r^2 Q^2)$ with $r = $\ttt{PARP(17)}.
\iteme{= 4 :} as \ttt{=3}, but instead of \ttt{PARP(17)} use either
\ttt{PARP(81)/PARP(15)} or \ttt{PARP(82)/PARP(15)}, depending on
\ttt{MSTP(82)} value.
\iteme{= 5 :} use larger $p_{\perp\mathrm{min}}$ for the anomalous component than
for the VMD one, but otherwise no difference.
\end{subentry}
\iteme{MSTP(21) :} (D=1) nature of fermion--fermion scatterings
simulated in process 10 by $t$-channel exchange.
\begin{subentry}
\iteme{= 0 :} all off.
\iteme{= 1 :} full mixture of $\gamma^* / \Z^0$ neutral current and
$\mathrm{W}^{\pm}$ charged current.
\iteme{= 2 :} $\gamma$ neutral current only.
\iteme{= 3 :} $\mathrm{Z}^0$ neutral current only.
\iteme{= 4 :} $\gamma^* / \Z^0$ neutral current only.
\iteme{= 5 :} $\mathrm{W}^{\pm}$ charged current only.
\end{subentry}
\iteme{MSTP(22) :} (D=0) special override of normal $Q^2$ definition
used for maximum of parton-shower evolution, intended for deep
inelastic scattering) in lepton--hadron events, see section
\ref{ss:showrout}.
\iteme{MSTP(23) :} (D=1) for deep inelastic scattering processes
(10 and 83), this option allows the $x$ and $Q^2$ of the original
hard scattering to be retained by the final electron.
\begin{subentry}
\iteme{= 0 :} no correction procedure, i.e. $x$ and $Q^2$ of the
scattered electron differ from the originally generated $x$ and
$Q^2$.
\iteme{= 1 :} post facto correction, i.e. the change of electron
momentum, by initial and final QCD radiation, primordial $k_{\perp}$
and beam remnant treatment, is corrected for by a shuffling of
momentum between the electron and hadron side in the final state.
Only process 10 is corrected, while process 83 is not.
\iteme{= 2 :} as \ttt{=1}, except that both process 10 and 83 are
treated. This option is dangerous, especially for top, since it
may well be impossible to `correct' in process 83: the standard DIS
kinematics definitions are based on the assumption of massless quarks.
Therefore infinite loops are not excluded.
\itemc{Note:} the correction procedure will fail for a fraction of
the events, which are thus rejected (and new ones generated in their
place). The correction option is not unambiguous, and should
not be taken too seriously. For very small $Q^2$ values, the $x$
is not exactly preserved even after this procedure.
\end{subentry}
\iteme{MSTP(31) :} (D=1) parametrization of total, elastic and
diffractive cross sections.
\begin{subentry}
\iteme{= 0 :} everything is to be set by you yourself in
the \ttt{PYINT7} common block. For photoproduction, additionally
you need to set \ttt{VINT(281)}. Normally you would set these
values once and for all before the \ttt{PYINIT} call, but if
you run with variable energies (see \ttt{MSTP(171)}) you can
also set it before each new \ttt{PYEVNT} call.
\iteme{= 1 :} Donnachie--Landshoff for total cross section
\cite{Don92}, and Schuler--Sj\"ostrand for elastic and diffractive
cross sections \cite{Sch94,Sch93a}.
\end{subentry}
\iteme{MSTP(32) :} (D=2) $Q^2$ definition in hard scattering for
$2 \to 2$ processes. For resonance production $Q^2$ is always chosen
to be $\hat{s} = m_R^2$, where $m_R$ is the mass of the resonance.
For gauge boson scattering processes $VV \to VV$ the $\mathrm{W}$ or $\mathrm{Z}^0$
squared mass is used as scale in parton distributions. See
\ttt{PARP(34)} for a possibility to modify the choice below by
a multiplicative factor.
\begin{subentry}
\iteme{= 1 :} $Q^2 = 2 \hat{s} \hat{t} \hat{u} / (\hat{s}^2 +
\hat{t}^2 + \hat{u}^2)$.
\iteme{= 2 :} $Q^2 = (m_{\perp 1}^2 + m_{\perp 2}^2)/2$.
\iteme{= 3 :} $Q^2 = \min(-\hat{t}, -\hat{u})$.
\iteme{= 4 :} $Q^2 = \hat{s}$.
\iteme{= 5 :} $Q^2 = -\hat{t}$.
\end{subentry}
\iteme{MSTP(33) :} (D=0) inclusion of $K$ factors in hard
cross sections for parton--parton interactions (i.e. for
incoming quarks and gluons).
\begin{subentry}
\iteme{= 0 :} none, i.e. $K = 1$.
\iteme{= 1 :} a common $K$ factor is used, as stored in
\ttt{PARP(31)}.
\iteme{= 2 :} separate factors are used for ordinary
(\ttt{PARP(31)}) and colour annihilation graphs (\ttt{PARP(32)}).
\iteme{= 3 :} A $K$ factor is introduced by a shift in the
$\alpha_{\mathrm{s}}$ $Q^2$ argument,
$\alpha_{\mathrm{s}} = \alpha_{\mathrm{s}}($\ttt{PARP(33)}$Q^2)$.
\end{subentry}
\iteme{MSTP(34) :} (D=1) use of interference term in matrix
elements for QCD processes, see section \ref{sss:QCDjetclass}.
\begin{subentry}
\iteme{= 0 :} excluded (i.e. string-inspired matrix elements).
\iteme{= 1 :} included (i.e. conventional QCD matrix elements).
\itemc{Note:} for the option \ttt{MSTP(34)=1}, i.e. interference
terms included, these terms are divided between the different
possible colour configurations according to the pole structure
of the (string-inspired) matrix elements for the different
colour configurations.
\end{subentry}
\iteme{MSTP(35) :} (D=0) threshold behaviour for heavy-flavour
production, i.e. ISUB = 81, 82, 84, 85, and also for $\mathrm{Z}$ and $\mathrm{Z}'$
decays. The non-standard options are mainly intended for top, but
can be used, with less theoretical reliability, also for charm and
bottom (for $\mathrm{Z}$ and $\mathrm{Z}'$ only top and heavier flavours
are affected). The threshold factors are given in
eqs.~(\ref{pp:threshenh}) and (\ref{pp:threshsup}).
\begin{subentry}
\iteme{= 0 :} na\"{\i}ve lowest-order matrix-element behaviour.
\iteme{= 1 :} enhancement or suppression close to threshold,
according to the colour structure of the process. The
$\alpha_{\mathrm{s}}$ value appearing in the threshold factor (which is not
the same as the $\alpha_{\mathrm{s}}$ of the lowest-order $2 \to 2$ process)
is taken to be fixed at the value given in \ttt{PARP(35)}. The
threshold factor used in an event is stored in \ttt{PARI(81)}.
\iteme{= 2 :} as \ttt{=1}, but the $\alpha_{\mathrm{s}}$ value appearing in
the threshold factor is taken to be running, with argument
$Q^2 = m_{\mathrm{Q}} \sqrt{ (\hat{m} - 2m_{\mathrm{Q}})^2 + \Gamma_{\mathrm{Q}}^2}$.
Here $m_{\mathrm{Q}}$ is the nominal heavy-quark mass, $\Gamma_{\mathrm{Q}}$ is
the width of the heavy-quark-mass distribution, and $\hat{m}$ is
the invariant mass of the heavy-quark pair. The $\Gamma_{\mathrm{Q}}$ value
has to be stored by the user in \ttt{PARP(35)}. The regularization
of $\alpha_{\mathrm{s}}$ at low $Q^2$ is given by \ttt{MSTP(36)}.
\end{subentry}
\iteme{MSTP(36) :} (D=2) regularization of $\alpha_{\mathrm{s}}$ in the
limit $Q^2 \to 0$ for the threshold factor obtainable in the
\ttt{MSTP(35)=2} option; see \ttt{MSTU(115)} for a list of
the possibilities.
\iteme{MSTP(37) :} (D=1) inclusion of running quark masses in
Higgs production ($\mathrm{q} \overline{\mathrm{q}} \to \H^0$) and decay
($\H^0 \to \mathrm{q} \overline{\mathrm{q}}$) couplings. Also included for charged Higgs
production and decay, but there only for the down-type quark,
since the up-type one normally is a top quark, with
$m_{\t} \approx m_{\H}$.
\begin{subentry}
\iteme{= 0 :} not included, i.e. fixed quark masses are used
according to the values in the \ttt{PMAS} array.
\iteme{= 1 :} included, with running starting from the
value given in the \ttt{PMAS} array, at a $Q_0$
scale of \ttt{PARP(37)} times the quark mass itself,
up to a $Q$ scale given by the Higgs mass.
This option only works when $\alpha_{\mathrm{s}}$ is allowed to run (so one can
define a $\Lambda$ value). Therefore it is only applied if additionally
\ttt{MSTP(2)}$\geq 1$.
\end{subentry}
\iteme{MSTP(38) :} (D=5) handling of quark loop masses in the box
graphs $\mathrm{g} \mathrm{g} \to \gamma \gamma$ and $\mathrm{g} \mathrm{g} \to \mathrm{g} \gamma$.
\begin{subentry}
\iteme{= 0 :} the program will for each flavour automatically
choose the massless approximation for light quarks and the full
massive formulae for heavy quarks, with a dividing line
between light and heavy quarks that depends on the actual
$\hat{s}$ scale.
\iteme{$\geq$1 :} the program will use the massless approximation
throughout, assuming the presence of \ttt{MSTP(38)} effectively
massless quark species (however, at most 8). Normally one would use
\ttt{=5} for the inclusion of all quarks up to bottom, and
\ttt{=6} to include top as well.
\itemc{Warning:} for \ttt{=0}, numerical instabilities may arise
for scattering at small angles. Users are therefore recommended in
this case to set \ttt{CKIN(27)} and \ttt{CKIN(28)} so as to exclude
the range of scattering angles that are not of interest anyway.
\end{subentry}
\iteme{MSTP(39) :} (D=2) choice of $Q^2$ scale for parton distributions
and initial state parton showers in processes $\mathrm{g}\g$ or
$\mathrm{q} \overline{\mathrm{q}} \to \mathrm{Q} \overline{\mathrm{Q}} \H$.
\begin{subentry}
\iteme{= 1 :} $m_{\mathrm{Q}}^2$.
\iteme{= 2 :} $\max(m_{\perp\mathrm{Q}}^2,m_{\perp\overline{\mathrm{Q}}}^2 ) =
m_{\mathrm{Q}}^2 + \max(p_{\perp\mathrm{Q}}^2 , p_{\perp\overline{\mathrm{Q}}}^2)$.
\iteme{= 3 :} $m_{\H}^2$.
\iteme{= 4 :} $\hat{s} = (p_{\H} + p_{\mathrm{Q}} + p_{\overline{\mathrm{Q}}})^2$.
\end{subentry}
\iteme{MSTP(41) :} (D=1) master switch for all resonance decays
($\mathrm{Z}^0$, $\mathrm{W}^{\pm}$, $\H^0$, $\mathrm{Z}'^0$, $\mathrm{W}'^{\pm}$, $\H'^0$,
$\mathrm{A}^0$, $\H^{\pm}$, $\L_{\mathrm{Q}}$, $\mathrm{R}^0$, $\d^*$, $\u^*$, \ldots).
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\itemc{Note:} also for \ttt{MSTP(41)=1} it is possible to switch
off the decays of specific resonances by using the \ttt{MDCY(KC,1)}
switches in {\tsc{Jetset}}. However, since the \ttt{MDCY} values are
overwritten in the \ttt{PYINIT} call, individual resonances should
be switched off after the \ttt{PYINIT} call.
\itemc{Warning:} leptoquark decays must not be switched off if one
later on intends to let leptoquarks decay (with \ttt{LUEXEC}); see
section \ref{sss:LQclass}.
\end{subentry}
\iteme{MSTP(42) :} (D=1) mass treatment in $2 \to 2$ processes, where
the final-state resonances have finite width (see \ttt{PARP(41)}).
(Does not apply for the production of a single $s$-channel resonance,
where the mass appears explicitly in the cross section of the
process, and thus is always selected with width.)
\begin{subentry}
\iteme{= 0 :} particles are put on the mass shell.
\iteme{= 1 :} mass generated according to a Breit--Wigner.
\end{subentry}
\iteme{MSTP(43) :} (D=3) treatment of $\mathrm{Z}^0/\gamma^*$ interference
in matrix elements. So far implemented in subprocesses 1, 15, 19, 22,
30 and 35; in other processes what is called a $\mathrm{Z}^0$ is really a
$\mathrm{Z}^0$ only, without the $\gamma^*$ piece.
\begin{subentry}
\iteme{= 1 :} only $\gamma^*$ included.
\iteme{= 2 :} only $\mathrm{Z}^0$ included.
\iteme{= 3 :} complete $\mathrm{Z}^0/\gamma^*$ structure (with
interference) included.
\end{subentry}
\iteme{MSTP(44) :} (D=7) treatment of $\mathrm{Z}'^0/\mathrm{Z}^0/\gamma^*$
interference in matrix elements.
\begin{subentry}
\iteme{= 1 :} only $\gamma^*$ included.
\iteme{= 2 :} only $\mathrm{Z}^0$ included.
\iteme{= 3 :} only $\mathrm{Z}'^0$ included.
\iteme{= 4 :} only $\mathrm{Z}^0/\gamma^*$ (with interference) included.
\iteme{= 5 :} only $\mathrm{Z}'^0/\gamma^*$ (with interference) included.
\iteme{= 6 :} only $\mathrm{Z}'^0/\mathrm{Z}^0$ (with interference) included.
\iteme{= 7 :} complete $\mathrm{Z}'^0/\mathrm{Z}^0/\gamma^*$ structure
(with interference) included.
\end{subentry}
\iteme{MSTP(45) :} (D=3) treatment of $\mathrm{W}\W \to \mathrm{W}\W$ structure
(ISUB = 77).
\begin{subentry}
\iteme{= 1 :} only $\mathrm{W}^+\mathrm{W}^+ \to \mathrm{W}^+\mathrm{W}^+$ and
$\mathrm{W}^-\mathrm{W}^- \to \mathrm{W}^-\mathrm{W}^-$ included.
\iteme{= 2 :} only $\mathrm{W}^+\mathrm{W}^- \to \mathrm{W}^+\mathrm{W}^-$ included.
\iteme{= 3 :} all charge combinations $\mathrm{W}\W \to \mathrm{W}\W$ included.
\end{subentry}
\iteme{MSTP(46) :} (D=1) treatment of $VV \to V'V'$ structures
(ISUB = 71--77), where $V$ represents a longitudinal gauge boson.
\begin{subentry}
\iteme{= 0 :} only $s$-channel Higgs exchange included (where
existing). With this option, subprocesses 71--72 and 76--77
will essentially be equivalent to subprocesses 5 and 8,
respectively, with the proper decay channels (i.e. only $\mathrm{Z}^0\mathrm{Z}^0$
or $\mathrm{W}^+\mathrm{W}^-$) set via \ttt{MDME}.
The description obtained for subprocesses 5 and 8 in this case
is more sophisticated, however; see section \ref{sss:heavySMHclass}.
\iteme{= 1 :} all graphs contributing to $VV \to V'V'$
processes are included.
\iteme{= 2 :} only graphs not involving Higgs exchange
(either in $s$, $t$ or $u$ channel) are included; this option
then gives the na\"{\i}ve behaviour one would expect if no Higgs
exists, including unphysical unitarity violations at high energies.
\iteme{= 3 :} the strongly interacting Higgs-like model of
Dobado, Herrero and Terron \cite{Dob91} with Pad\'e unitarization.
Note that to use this option it is necessary to set the Higgs mass
to a large number like 20 TeV (i.e. \ttt{PMAS(25,1)=20000}). The
parameter $\nu$ is stored in \ttt{PARP(44)}, but should not have
to be changed.
\iteme{= 4 :} as \ttt{=3}, but with K-matrix unitarization.
\iteme{= 5 :} the strongly interacting QCD-like model of
Dobado, Herrero and Terron \cite{Dob91} with Pad\'e unitarization.
The parameter $\nu$ is stored in \ttt{PARP(44)}, but should not
have to be changed. The effective techni-$\rho$ mass in this model
is stored in \ttt{PARP(45)}; by default it is 2054 GeV, which is
the expected value for three technicolors, based on scaling up
the ordinary $\rho$ mass appropriately.
\iteme{= 6 :} as \ttt{=5}, but with K-matrix unitarization.
While \ttt{PARP(45)} still is a parameter of the model, this type
of unitarization does not give rise to a resonance at a mass of
\ttt{PARP(45)}.
\end{subentry}
\iteme{MSTP(47) :} (D=1) (C) angular orientation of decay products
of resonances ($\mathrm{Z}^0$, $\mathrm{W}^{\pm}$, $\H^0$, $\mathrm{Z}'^0$, $\mathrm{W}'^{\pm}$,
etc.), either when produced singly or in pairs (also
from an $\H^0$ decay), or in combination with a single quark,
gluon or photon.
\begin{subentry}
\iteme{= 0 :} independent decay of each resonance, isotropic in c.m.
frame of the resonance.
\iteme{= 1 :} correlated decay angular distributions according to
proper matrix elements, to the extent these are known.
\end{subentry}
\iteme{MSTP(48) :} (D=2) (C) possibility to switch between top decay
before or after fragmentation. As a rule of thumb, option 0 is
recommendable for top masses below 120 GeV and option 1 above that,
but clearly there is a gradual transition between the two.
\begin{subentry}
\iteme{= 0 :} top quarks fragment to top hadrons, which subsequently
decay. The $\b$ quark may be allowed to shower, see \ttt{MSTJ(27)}.
The $\mathrm{W}$ produced in the decay is code 89.
\iteme{= 1 :} top quarks decay, $\t \to \b \mathrm{W}^+$, and thereafter the
$\b$ quark fragments. Parton showering of the $\b$ is automatically
included, but can be switched off with \ttt{MSTP(71)}. The $\mathrm{W}$
has the ordinary code 24, and is allowed to decay isotropically.
\iteme{= 2 :} as \ttt{=1}, except that the $\mathrm{W}$ decay is anisotropic,
as expected from $\mathrm{W}$ polarization in the top decay.
\itemc{Note:} in options 1 and 2 the cross section is reduced to take
into account restrictions on allowed decay modes, while no such
reduction occurs for option 0. See further section
\ref{sss:heavflavclass}.
\end{subentry}
\iteme{MSTP(49) :} (D=2) (C) possibility to switch between fourth
generation decay before or after fragmentation. For the quarks $\mathrm{h}$
and $\mathrm{l}$ the meaning is exactly as \ttt{MSTP(48)} is for the $\t$
quark. For the lepton $\chi$ the difference is whether decay is
handled as part of the {\tsc{Pythia}} resonance machinery or as part of the
{\tsc{Jetset}} particle decay one. The $\nu_{\chi}$ is assumed stable, so the
option above would currently make no difference.
\begin{subentry}
\iteme{= 0 :} hadrons are first produced, which subsequently
decay (in \ttt{LUDECY}). The new quark may be allowed to shower,
see \ttt{MSTJ(27)}. The $\mathrm{W}$ produced in the decay is code $\pm89$.
\iteme{= 1 :} the heavy quark first decays (in \ttt{PYRESD}) to a
light one, and thereafter the light quark fragments. Parton showering
in the decay is automatically included, but can be switched off with
\ttt{MSTP(71)}. The $\mathrm{W}$ has the ordinary code $\pm24$, and is allowed
to decay isotropically.
\iteme{= 2 :} as \ttt{=1}, except that the $\mathrm{W}$ decay is anisotropic,
as expected from $\mathrm{W}$ polarization in the heavy flavour decay.
\itemc{Note:} in options 1 and 2 the cross section is reduced to take
into account restrictions on allowed decay modes, while no such
reduction occurs for option 0. See further section
\ref{sss:heavflavclass}.
\end{subentry}
\iteme{MSTP(51) :} (D=9) choice of proton parton-distribution set;
see also \ttt{MSTP(52)}.
\begin{subentry}
\iteme{= 1 :} EHLQ set 1 (1986 updated version).
\iteme{= 2 :} EHLQ set 2 (1986 updated version).
\iteme{= 3 :} Duke--Owens set 1.
\iteme{= 4 :} Duke--Owens set 2.
\iteme{= 5 :} CTEQ2M (best $\br{\mrm{MS}}$ fit).
\iteme{= 6 :} CTEQ2MS (singular at small $x$).
\iteme{= 7 :} CTEQ2MF (flat at small $x$).
\iteme{= 8 :} CTEQ2ML (large $\Lambda$).
\iteme{= 9 :} CTEQ2L (best leading order fit).
\iteme{= 10 :} CTEQ2D (best DIS fit).
\iteme{= 11 :} GRV LO (1992 updated version).
\itemc{Note:} since all parametrizations have some region of
applicability, the parton distributions are assumed frozen below
the lowest $Q^2$ covered by the parametrizations; the CTEQ2 ones
have been allowed to extend down to $Q_{\mathrm{min}} = 1$~GeV. For the
former four, evolution is also frozed above the maximum $Q^2$.
The extrapolation of EHLQ to low $x$ is covered by \ttt{PARP(51)}.
\end{subentry}
\iteme{MSTP(52) :} (D=1) choice of proton
parton-distribution-function library.
\begin{subentry}
\iteme{= 1 :} the internal {\tsc{Pythia}} one, with parton distributions
according to the \ttt{MSTP(51)} above.
\iteme{= 2 :} the \tsc{Pdflib} one \cite{Plo93}, with the
\tsc{Pdflib} (version 4) \ttt{NGROUP} and \ttt{NSET} numbers to be
given as \ttt{MSTP(51) = 1000}$\times$\ttt{NGROUP + NSET}.
\itemc{Note:} to make use of option 2, it is necessary to link
\tsc{Pdflib}. Additionally, on most computers, the two dummy routines
\ttt{PDFSET} and \ttt{STRUCTM} at the end of the {\tsc{Pythia}} file should be
removed or commented out.
\itemc{Warning:} For external parton distribution libraries,
{\tsc{Pythia}} does not check whether \ttt{MSTP(51)} corresponds to a
valid code, or if special $x$ and $Q^2$ restrictions exist
for a given set, such that crazy values could be returned.
This puts an extra responsibility on you.
\end{subentry}
\iteme{MSTP(53) :} (D=1) choice of pion parton-distribution set;
see also \ttt{MSTP(54)}.
\begin{subentry}
\iteme{= 1 :} Owens set 1.
\iteme{= 2 :} Owens set 2.
\iteme{= 3 :} GRV LO (updated version).
\end{subentry}
\iteme{MSTP(54) :} (D=1) choice of pion parton-distribution-function
library.
\begin{subentry}
\iteme{= 1 :} the internal {\tsc{Pythia}} one, with parton distributions
according to the \ttt{MSTP(53)} above.
\iteme{= 2 :} the \tsc{Pdflib} one \cite{Plo93}, with the
\tsc{Pdflib} (version 4) \ttt{NGROUP} and \ttt{NSET} numbers to be
given as \ttt{MSTP(53) = 1000}$\times$\ttt{NGROUP + NSET}.
\itemc{Note:} to make use of option 2, it is necessary to link
\tsc{Pdflib}. Additionally, on most computers, the two dummy routines
\ttt{PDFSET} and \ttt{STRUCTM} at the end of the {\tsc{Pythia}} file should be
removed or commented out.
\itemc{Warning:} For external parton distribution libraries,
{\tsc{Pythia}} does not check whether \ttt{MSTP(53)} corresponds to a valid
code, or if special $x$ and $Q^2$ restrictions exist for a given
set, such that crazy values could be returned. This puts an extra
responsibility on you.
\end{subentry}
\iteme{MSTP(55)} : (D=5) choice of the parton-distribution
set of the photon; see also \ttt{MSTP(56)}.
\begin{subentry}
\iteme{= 1 :} Drees--Grassie.
\iteme{= 5 :} SaS 1D (in DIS scheme, with $Q_0=0.6$~GeV).
\iteme{= 6 :} SaS 1M (in {$\overline{\mrm{MS}}$} scheme, with $Q_0=0.6$~GeV).
\iteme{= 7 :} SaS 2D (in DIS scheme, with $Q_0=2$~GeV).
\iteme{= 8 :} SaS 2M (in {$\overline{\mrm{MS}}$} scheme, with $Q_0=2$~GeV).
\iteme{= 9 :} SaS 1D (in DIS scheme, with $Q_0=0.6$~GeV).
\iteme{= 10 :} SaS 1M (in {$\overline{\mrm{MS}}$} scheme, with $Q_0=0.6$~GeV).
\iteme{= 11 :} SaS 2D (in DIS scheme, with $Q_0=2$~GeV).
\iteme{= 12 :} SaS 2M (in {$\overline{\mrm{MS}}$} scheme, with $Q_0=2$~GeV).
\itemc{Note 1:} sets 5--8 use the parton distributions of the respective
set, and nothing else. These are appropriate for most applications, e.g.
jet production in $\gamma\mathrm{p}$ and $\gamma\gamma$ collisions. Sets 9--12
instead are appropriate for $\gamma^*\gamma$ processes, i.e. DIS
scattering on a photon, as measured in $F_2^{\gamma}$. Here the anomalous
contribution for $\c$ and $\b$ quarks are handled by the Bethe-Heitler
formulae, and the direct term is artificially lumped with the anomalous
one, so that the event simulation more closely agrees with what will be
experimentally observed in these processes. The agreement with the
$F_2^{\gamma}$ parametrization is still not perfect, e.g. in the treatment
of heavy flavours close to threshold.
\itemc{Note 2:} Sets 5--12 contain both VMD pieces and anomalous pieces,
separately parametrized. Therefore the respective piece is automatically
called, whatever \ttt{MSTP(14)} value is used to select only a part of the
allowed photon interactions. For other sets (set 1 above or \tsc{Pdflib}
sets), usually there is no corresponding subdivision. Then an option like
\ttt{MSTP(14)=2} (VMD part of photon only) is based on a rescaling of the
pion distributions, while \ttt{MSTP(14)=3} gives the SaS anomalous
parametrization.
\itemc{Note 3:} Formally speaking, the $k_0$ (or $p_0$) cut-off in
\ttt{PARP(15)} need not be set in any relation to the $Q_0$ cut-off
scales used by the various parametrizations. Indeed, due to the
familiar scale choice ambiguity problem, there could well be some offset
between the two. However, unless you know what you are doing, it is
strongly recommended that you let the two agree, i.e. set
\ttt{PARP(15)=0.6} for the SaS 1 sets and \ttt{=2.} for the SaS 2 sets.
\end{subentry}
\iteme{MSTP(56) :} (D=1) choice of photon parton-distribution-function
library.
\begin{subentry}
\iteme{= 1 :} the internal {\tsc{Pythia}} one, with parton distributions
according to the \ttt{MSTP(55)} above.
\iteme{= 2 :} the \tsc{Pdflib} one \cite{Plo93}, with the
\tsc{Pdflib} (version 4) \ttt{NGROUP} and \ttt{NSET} numbers to be
given as \ttt{MSTP(55) = 1000}$\times$\ttt{NGROUP + NSET}.
When the VMD and anomalous parts of the photon are split,
like for \ttt{MSTP(14)=10}, it is necessary to specify pion set to be
used for the VMD component, in \ttt{MSTP(53)} and \ttt{MSTP(54)},
while \ttt{MSTP(55)} here is irrelevant.
\iteme{= 3 :} when the parton distributions of the anomalous photon
are requested, the homogeneous solution is provided, evolved from a
starting value \ttt{PARP(15)} to the requested $Q$ scale. The homogeneous
solution is normalized so that the net momentum is unity,
i.e. any factors of $\alpha_{\mathrm{em}}/2\pi$ and charge have been left out.
The flavour of the original $\mathrm{q}$ is given in \ttt{MSTP(55)} (1, 2, 3, 4
or 5 for $\d$, $\u$, $\mathrm{s}$, $\c$ or $\b$); the value 0 gives a mixture
according to squared charge, with the exception that $\c$ and $\b$
are only allowed above the respective mass threshold ($Q > m_{\mathrm{q}}$).
The four-flavour $\Lambda$ value is assumed given in \ttt{PARP(1)};
it is automatically recalculated for 3 or 5 flavours at
thresholds. This option is not intended for standard event
generation, but is useful for some theoretical studies.
\itemc{Note:} to make use of option 2, it is necessary to link
\tsc{Pdflib}. Additionally, on most computers, the two dummy routines
\ttt{PDFSET} and \ttt{STRUCTM} at the end of the {\tsc{Pythia}} file should be
removed or commented out.
\itemc{Warning:} For external parton-distribution libraries, {\tsc{Pythia}}
does not check whether \ttt{MSTP(55)} corresponds to a valid code,
or if special $x$ and $Q^2$ restrictions exist for a given set,
such that crazy values could be returned. This puts an extra
responsibility on you.
\end{subentry}
\iteme{MSTP(57) :} (D=1) choice of $Q^2$ dependence in
parton-distribution functions.
\begin{subentry}
\iteme{= 0 :} parton distributions are evaluated at nominal lower
cut-off value $Q_0^2$, i.e. are made $Q^2$-independent.
\iteme{= 1 :} the parametrized $Q^2$ dependence is used.
\iteme{= 2 :} the parametrized parton-distribution behaviour is kept
at large $Q^2$ and $x$, but modified at small $Q^2$ and/or $x$,
so that parton distributions vanish in the limit $Q^2 \to 0$ and
have a theoretically motivated small-$x$ shape \cite{Sch93a}.
This option is only valid for the $\mathrm{p}$ and $\mathrm{n}$.
\iteme{= 3 :} as \ttt{=2}, except that also the $\pi^{\pm}$ and the
VMD component of a photon is modified in a corresponding manner.
\end{subentry}
\iteme{MSTP(58) :} (D=min(6, 2$\times$\ttt{MSTP(1)})) maximum number of
quark flavours used in parton distributions, and thus also for
initial-state space-like showers. If some distributions (notably $\t$)
are absent in the parametrization selected in \ttt{MSTP(51)}, these
are obviously automatically excluded.
\iteme{MSTP(61) :} (D=1) (C) master switch for initial-state QCD and
QED radiation.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\end{subentry}
\iteme{MSTP(62) - MSTP(67) :} (C) further switches for initial-state
radiation, see section \ref{ss:showrout}.
\iteme{MSTP(71) :} (D=1) (C) master switch for final-state QCD and
QED radiation.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\itemc{Note:} additional switches (e.g. for conventional/coherent
showers) are available in \ttt{MSTJ(40) - MSTJ(50)} and
\ttt{PARJ(81) - PARJ(89)}, see section \ref{ss:showrout}.
\end{subentry}
\iteme{MSTP(81) :} (D=1) master switch for multiple interactions.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\end{subentry}
\iteme{MSTP(82) - MSTP(83) :} further switches for multiple
interactions, see section \ref{ss:multintpar}.
\iteme{MSTP(85) :} possibility to dampen hard scattering matrix
elements in the limit $p_{\perp} \to 0$. It parellels some of the
multiple interactions options for QCD processes, but can be
used for any $2 \to 2$ process.
\begin{subentry}
\iteme{= 0 :} off, i.e. standard matrix elements are kept.
\iteme{= 1 :} on, i.e. matrix elements are multiplied by a
factor $p_{\perp}^4/(p_{\perp}^2 + p_{\perp 0}^2)^2$, where $p_{\perp 0}$
is given by \ttt{PARP(82)}. Additionally $\alpha_{\mathrm{s}}$ is evaluated
at a scale $p_{\perp}^2 + p_{\perp 0}^2$ rather than just $p_{\perp}^2$.
\end{subentry}
\iteme{MSTP(91) - MSTP(94) :} switches for beam remnant treatment,
see section \ref{ss:multintpar}.
\iteme{MSTP(101) :} (D=3) (C) structure of diffractive system.
\begin{subentry}
\iteme{= 1 :} forward moving diquark + interacting quark.
\iteme{= 2 :} forward moving diquark + quark joined via interacting
gluon (`hairpin' configuration).
\iteme{= 3 :} a mixture of the two options above, with a fraction
\ttt{PARP(101)} of the former type.
\end{subentry}
\iteme{MSTP(102) :} (D=1) (C) decay of a $\rho^0$ meson produced by
`elastic' scattering of an incoming $\gamma$, as in
$\gamma \mathrm{p} \to \rho^0 \mathrm{p}$, or the same with the hadron diffractively
excited.
\begin{subentry}
\iteme{= 0 :} the $\rho^0$ is allowed to decay isotropically, like
any other $\rho^0$.
\iteme{= 1 :} the decay $\rho^0 \to \pi^+ \pi^-$ is done with an
angular distribution proportional to $\sin^2 \theta$ in its rest frame,
where the $z$ axis is given by the direction of motion of the
$\rho^0$. The $\rho^0$ decay is then done as part of the hard process,
i.e. also when \ttt{MSTP(111)=0}.
\end{subentry}
\iteme{MSTP(111) :} (D=1) (C) master switch for fragmentation
and decay, as obtained with a \ttt{LUEXEC} call.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\iteme{= -1 :} only choose kinematical variables for hard scattering,
i.e. no jets are defined. This is useful, for instance, to calculate
cross sections (by Monte Carlo integration) without wanting
to simulate events; information obtained with \ttt{PYSTAT(1)}
will be correct.
\end{subentry}
\iteme{MSTP(112) :} (D=1) (C) cuts on partonic events; only affects
an exceedingly tiny fraction of events.
\begin{subentry}
\iteme{= 0 :} no cuts (can be used only with independent
fragmentation, at least in principle).
\iteme{= 1 :} string cuts (as normally required for fragmentation).
\end{subentry}
\iteme{MSTP(113) :} (D=1) (C) recalculation of energies of partons
from their momenta and masses, to be done immediately before
and after fragmentation, to partly compensate for some numerical
problems appearing at high energies.
\begin{subentry}
\iteme{= 0 :} not performed.
\iteme{= 1 :} performed.
\end{subentry}
\iteme{MSTP(121) :} (D=0) calculation of kinematics selection
coefficients and differential cross section maxima for
included (by user or default) subprocesses.
\begin{subentry}
\iteme{= 0 :} not known; to be calculated at initialization.
\iteme{= 1 :} not known; to be calculated at initialization;
however, the maximum value then obtained is to be multiplied by
\ttt{PARP(121)} (this may be useful if a violation factor has
been observed in a previous run of the same kind).
\iteme{= 2 :} known; kinematics selection coefficients stored
by user in \ttt{COEF(ISUB,J)} (\ttt{J} = 1--20) in common block
\ttt{PYINT2} and maximum of the corresponding differential
cross section times Jacobians in \ttt{XSEC(ISUB,1)} in
common block \ttt{PYINT5}. This is to be done for each included
subprocess ISUB before initialization, with the sum of all
\ttt{XSEC(ISUB,1)} values, except for ISUB = 95, stored in
\ttt{XSEC(0,1)}.
\end{subentry}
\iteme{MSTP(122) :} (D=1) initialization and differential
cross section maximization print-out.
\begin{subentry}
\iteme{= 0 :} none.
\iteme{= 1 :} short message.
\iteme{= 2 :} detailed message, including full maximization.
\end{subentry}
\iteme{MSTP(123) :} (D=2) reaction to violation of maximum
differential cross section.
\begin{subentry}
\iteme{= 0 :} stop generation, print message.
\iteme{= 1 :} continue generation, print message for each
subsequently larger violation.
\iteme{= 2 :} as \ttt{=1}, but also increase value of maximum.
\end{subentry}
\iteme{MSTP(124) :} (D=1) (C) frame for presentation of event.
\begin{subentry}
\iteme{= 1 :} as specified in \ttt{PYINIT}.
\iteme{= 2 :} c.m. frame of incoming particles.
\end{subentry}
\iteme{MSTP(125) :} (D=1) (C) documentation of partonic process,
see section \ref{sss:PYrecord} for details.
\begin{subentry}
\iteme{= 0 :} only list ultimate string/particle configuration.
\iteme{= 1 :} additionally list short summary of the hard process.
\iteme{= 2 :} list complete documentation of intermediate steps of
parton-shower evolution.
\end{subentry}
\iteme{MSTP(126) :} (D=20) number of lines at the beginning of event
record that are reserved for event-history information; see section
\ref{sss:PYrecord}. This value should never be reduced, but may be
increased at a later date if more complicated processes are included.
\iteme{MSTP(128) :} (D=0) storing of copy of resonance decay
products in the documentation section of the event record, and
mother pointer (\ttt{K(I,3)}) relation of the actual resonance
decay products (stored in the main section of the event record)
to the documentation copy.
\begin{subentry}
\iteme{= 0 :} products are stored also in the documentation section,
and each product stored in the main section points back
to the corresponding entry in the documentation section.
\iteme{= 1 :} products are stored also in the documentation section,
but the products stored in the main section point back to
the decaying resonance copy in the main section.
\iteme{= 2 :} products are not stored in the documentation section;
the products stored in the main section point back to the
decaying resonance copy in the main section.
\end{subentry}
\iteme{MSTP(129) :} (D=10) for the maximization of $2 \to 3$ processes
(\ttt{ISET(ISUB)=5}) each phase-space point in $\tau$, $y$ and $\tau'$
is tested \ttt{MSTP(129)} times in the other dimensions (at randomly
selected points) to determine the effective maximum in the
($\tau$, $y$, $\tau'$) point.
\iteme{MSTP(131) :} (D=0) master switch for pile-up events, i.e. several
independent hadron--hadron interactions generated in the same
bunch--bunch crossing, with the events following one after the
other in the event record.
\begin{subentry}
\iteme{= 0 :} off, i.e. only one event is generated at a time.
\iteme{= 1 :} on, i.e. several events are allowed in the same event
record. Information on the processes generated may be found in
\ttt{MSTI(41) - MSTI(50)}.
\end{subentry}
\iteme{MSTP(132) - MSTP(134) :} further switches for pile-up events,
see section \ref{ss:multintpar}.
\iteme{MSTP(141) :} (D=0) calling of \ttt{PYKCUT} in the
event-generation chain, for inclusion of user-specified cuts.
\begin{subentry}
\iteme{= 0 :} not called.
\iteme{= 1 :} called.
\end{subentry}
\iteme{MSTP(142) :} (D=0) calling of \ttt{PYEVWT} in the
event-generation chain, either to give weighted events or to modify
standard cross sections. See \ttt{PYEVWT} description in section
\ref{ss:PYTmainroutines} for further details.
\begin{subentry}
\iteme{= 0 :} not called.
\iteme{= 1 :} called; the distribution of events among subprocesses
and in kinematics variables is modified by the factor \ttt{WTXS},
set by the user in the \ttt{PYEVWT} call, but events come with a
compensating weight \ttt{PARI(10)=1./WTXS}, such that total
cross sections are unchanged.
\iteme{= 2 :} called; the cross section itself is modified by the
factor \ttt{WTXS}, set by the user in the \ttt{PYEVWT} call.
\end{subentry}
\iteme{MSTP(151) :} (D=0) introduce smeared position of primary vertex
of events.
\begin{subentry}
\iteme{= 0 :} no, i.e. the primary vertex of each event is at the
origin.
\iteme{= 1 :} yes, with Gaussian distributions separately in $x$, $y$,
$z$ and $t$. The respective widths of the Gaussians have to be given
in \ttt{PARP(151) - PARP(154)}. Also pile-up events obtain separate
primary vertices. No provisions are made for more complicated
beam-spot shapes, e.g. with a spread in $z$ that varies as a
function of $t$. Note that a large beam spot combined with some of the
\ttt{MSTJ(22)} options may lead to many particles not being allowed to
decay at all.
\end{subentry}
\iteme{MSTP(171) :} (D=0) possibility of variable energies from one
event to the next. For further details see subsection
\ref{ss:PYvaren}.
\begin{subentry}
\iteme{= 0 :} no; i.e. the energy is fixed at the initialization call.
\iteme{= 1 :} yes; i.e. a new energy has to be given for each new
event.
\end{subentry}
\iteme{MSTP(172) :} (D=2) options for generation of events with
variable energies, applicable when \ttt{MSTP(171)=1}.
\begin{subentry}
\iteme{= 1 :} an event is generated at the requested energy, i.e.
internally a loop is performed over possible event configurations
until one is accepted. If the requested c.m. energy of an event
is below \ttt{PARP(2)} the run is aborted. Cross-section information
can not be trusted with this option, since it depends on how you
decided to pick the requested energies.
\iteme{= 2 :} only one event configuration is tried. If that is
accepted, the event is generated in full. If not, no event is
generated, and the status code \ttt{MSTI(61)=1} is returned.
You are then expected to give a new energy, looping until an
acceptable event is found. No event is generated if the
requested c.m. energy is below \ttt{PARP(2)}, instead
\ttt{MSTI(61)=1} is set to signal the failure. In principle,
cross sections should come out correctly with this option.
\end{subentry}
\iteme{MSTP(173) :} (D=0) possibility for user to give in an event
weight to compensate for a biased choice of beam spectrum.
\begin{subentry}
\iteme{= 0 :} no, i.e. event weight is unity.
\iteme{= 1 :} yes; weight to be given for each event in
\ttt{PARP(173)}, with maximum weight given at initialization
in \ttt{PARP(174)}.
\end{subentry}
\iteme{MSTP(181) :} (R) {\tsc{Pythia}} version number.
\iteme{MSTP(182) :} (R) {\tsc{Pythia}} subversion number.
\iteme{MSTP(183) :} (R) last year of change for {\tsc{Pythia}}.
\iteme{MSTP(184) :} (R) last month of change for {\tsc{Pythia}}.
\iteme{MSTP(185) :} (R) last day of change for {\tsc{Pythia}}.
\iteme{MSTP(186) :} (R) earliest subversion of {\tsc{Jetset}} version 7
with which this {\tsc{Pythia}} subversion can be run.
\boxsep
\iteme{PARP(1) :}\label{p:PARP} (D=0.25 GeV) nominal
$\Lambda_{\mrm{QCD}}$ used in running $\alpha_{\mathrm{s}}$ for hard
scattering (see \ttt{MSTP(3)}).
\iteme{PARP(2) :} (D=10. GeV) lowest c.m. energy for the
event as a whole that the program will accept to simulate.
\iteme{PARP(13) :} (D=1. GeV$^2$) $Q_{\mathrm{max}}^2$ scale, to be set by
user for defining maximum scale allowed for photoproduction when
using the option \ttt{MSTP(13)=2}.
\iteme{PARP(14) :} (D=0.01) in the numerical integration of quark
and gluon parton distributions inside an electron, the successive
halvings of evaluation-point spacing is interrupted when two values
agree in relative size, $|$new$-$old$|$/(new$+$old), to better than
\ttt{PARP(14)}. There are hardwired lower and upper limits of 2 and
8 halvings, respectively.
\iteme{PARP(15) :} (D=0.5 GeV) lower cut-off $p_0$ used to define
minimum transverse momentum in branchings $\gamma \to \mathrm{q}\overline{\mathrm{q}}$ in
the anomalous event class of $\gamma\mathrm{p}$ interactions.
\iteme{PARP(16) :} (D=1.) the anomalous parton-distribution functions
of the photon are taken to have the charm and bottom flavour thresholds
at virtuality \ttt{PARP(16)}$\times m_{\mathrm{q}}^2$.
\iteme{PARP(17) :} (D=1.) rescaling factor used for the $Q$ argument
of the anomalous parton distributions of the photon, see
\ttt{MSTP(15)}.
\iteme{PARP(31) :} (D=1.5) common $K$ factor multiplying the
differential cross section for hard parton--parton processes
when \ttt{MSTP(33)=1} or \ttt{2}, with the exception of colour
annihilation graphs in the latter case.
\iteme{PARP(32) :} (D=2.0) special $K$ factor multiplying the
differential cross section in hard colour annihilation graphs,
including resonance production, when \ttt{MSTP(33)=2}.
\iteme{PARP(33) :} (D=0.075) this factor is used to multiply the
ordinary $Q^2$ scale in $\alpha_{\mathrm{s}}$ at the hard interaction for
\ttt{MSTP(33)=3}. The effective $K$ factor thus obtained is in
accordance with the results in \cite{Ell86}.
\iteme{PARP(34) :} (D=1.) the $Q^2$ scale defined by \ttt{MSTP(32)} is
multiplied by \ttt{PARP(34)} when it is used as argument for parton
distributions and $\alpha_{\mathrm{s}}$ at the hard interaction. It does not affect
$\alpha_{\mathrm{s}}$ when \ttt{MSTP(33)=3}, nor does it change the $Q^2$ argument
of parton showers.
\iteme{PARP(35) :} (D=0.20) fix $\alpha_{\mathrm{s}}$ value that is used in
the heavy-flavour threshold factor when \ttt{MSTP(35)=1}.
\iteme{PARP(36) :} (D=0. GeV) the width $\Gamma_{\mathrm{Q}}$ for the heavy
flavour studied in processes ISUB = 81 or 82; to be used for the
threshold factor when \ttt{MSTP(35)=2}.
\iteme{PARP(37) :} (D=2.) for \ttt{MSTP(37)=1} this regulates the
point at which the reference on-shell quark mass in Higgs couplings is
assumed defined; specifically the running quark mass is assumed
to coincide with the fix one at an energy scale \ttt{PARP(37)} times
the fix quark mass,
i.e. $m_{\mrm{running}}($\ttt{PARP(37)}$\times m_{\mrm{fix}}) =
m_{\mrm{fix}}$.
\iteme{PARP(38) :} (D=0.70 GeV$^3$) the squared wave function at the
origin, $|R(0)|^2$, of the $\mathrm{J}/\psi$ wave function. Used for process
86. See ref. \cite{Glo88}.
\iteme{PARP(39) :} (D=0.006 GeV$^3$) the squared derivative of the
wave function at the origin, $|R'(0)|^2/m^2$, of the $\chi_{\c}$
wave functions. Used for the processes 87, 88 and 89. See ref.
\cite{Glo88}.
\iteme{PARP(41) :} (D=0.020 GeV) in the process of generating mass
for resonances, and optionally to force that mass to be in a given
range, only resonances with a total width in excess of \ttt{PARP(41)}
are generated according to a Breit--Wigner shape (if allowed by
\ttt{MSTP(42)}), while narrower resonances are put on the mass
shell.
\iteme{PARP(42) :} (D=2. GeV) minimum mass of resonances assumed
to be allowed when evaluating total width of $\H^0$ to $\mathrm{Z}^0 \mathrm{Z}^0$ or
$\mathrm{W}^+ \mathrm{W}^-$ for cases when the $\H^0$ is so light that (at least)
one $\mathrm{Z}/\mathrm{W}$ is forced to be off the mass shell. Also generally used
as safety check on minimum mass of resonance. Note that some
\ttt{CKIN} values may provide additional constraints.
\iteme{PARP(43) :} (D=0.10) precision parameter used in numerical
integration of width into channel with at least one daughter off
the mass shell.
\iteme{PARP(44) :} (D=1000.) the $\nu$ parameter of the strongly
interacting $\mathrm{Z}/\mathrm{W}$ model of Dobado, Herrero and Terron \cite{Dob91}.
\iteme{PARP(45) :} (D=2054. GeV) the effective techni-$\rho$ mass
parameter of the strongly interacting model of Dobado, Herrero and
Terron \cite{Dob91}; see \ttt{MSTP(46)=5}. On physical grounds it
should not be chosen smaller than about 1 TeV or larger than
about the default value.
\iteme{PARP(46) :} (D=123. GeV) the $F_{\pi}$ decay constant that
appears inversely quadratically in all techni-$\eta$ partial decay
widths \cite{Eic84,App92}.
\iteme{PARP(47) :} (D=246. GeV) vacuum expectation value $v$ used
in the DHT scenario \cite{Dob91} to define the width of the
techni-$\rho$; this width is inversely proportional $v^2$.
\iteme{PARP(51) :} (D=1.) if parton distributions for light flavours
have to be extrapolated to $x$ values lower than covered by the
parametrizations, an $x^{-b}$ behaviour, with $b =$\ttt{PARP(51)},
is assumed in that region. This option only applies for the EHLQ
proton parton distributions that are internal to {\tsc{Pythia}}.
\iteme{PARP(61) - PARP(65) :} (C) parameters for initial-state
radiation, see section \ref{ss:showrout}.
\iteme{PARP(71) - PARP(72) :} (C) parameter for final-state
radiation, see section \ref{ss:showrout}.
\iteme{PARP(81) - PARP(88) :} parameters for multiple interactions,
see section \ref{ss:multintpar}.
\iteme{PARP(91) - PARP(100) :} parameters for beam remnant
treatment, see section \ref{ss:multintpar}.
\iteme{PARP(101) :} (D=0.50) fraction of diffractive systems in which
a quark is assumed kicked out by the pomeron rather than a gluon;
applicable for option \ttt{MSTP(101)=3}.
\iteme{PARP(102) :} (D=0.28 GeV) the mass spectrum of diffractive
states (in single and double diffractive scattering) is assumed to
start \ttt{PARP(102)} above the mass of the particle that is
diffractively excited. In this connection, an incoming $\gamma$
is taken to have the selected VMD meson mass, i.e. $m_{\rho}$,
$m_{\omega}$, $m_{\phi}$ or $m_{\mathrm{J}/\psi}$.
\iteme{PARP(103) :} (D=1.0 GeV) if the mass of a diffractive state
is less than \ttt{PARP(103)} above the mass of the particle that is
diffractively excited, the state is forced to decay isotropically
into a two-body channel. In this connection, an incoming $\gamma$
is taken to have the selected VMD meson mass, i.e. $m_{\rho}$,
$m_{\omega}$, $m_{\phi}$ or $m_{\mathrm{J}/\psi}$. If the mass is higher than
this threshold, the standard string fragmentation machinery is used.
The forced two-body decay is always carried out, also when
\ttt{MSTP(111)=0}.
\iteme{PARP(111) :} (D=2. GeV) used to define the minimum invariant
mass of the remnant hadronic system (i.e. when interacting partons
have been taken away), together with original hadron masses and
extra parton masses.
\iteme{PARP(121) :} (D=1.) the maxima obtained at initial
maximization are multiplied by this factor if \ttt{MSTP(121)=1};
typically \ttt{PARP(121)} would be given as the product of the
violation factors observed (i.e. the ratio of final maximum value
to initial maximum value) for the given process(es).
\iteme{PARP(122) :} (D=0.4) fraction of total probability that is
shared democratically between the \ttt{COEF} coefficients open for
the given variable, with the remaining fraction distributed according
to the optimization results of \ttt{PYMAXI}.
\iteme{PARP(131) :} parameter for pile-up events, see section
\ref{ss:multintpar}.
\iteme{PARP(151) - PARP(154) :} (D=4*0.) (C) regulate the assumed
beam-spot size. For \ttt{MSTP(151)=1} the $x$, $y$, $z$ and $t$
coordinates of the primary vertex of each event are selected
according to four independent Gaussians. The widths of these
Gaussians are given by the four parameters, where the first three
are in units of mm and the fourth in mm/$c$.
\iteme{PARP(161) - PARP(164) :} (D=2.20, 23.6, 18.4, 11.5) couplings
$f_V^2/4\pi$ of the photon to the $\rho^0$, $\omega$, $\phi$ and
$\mathrm{J}/\psi$ vector mesons.
\iteme{PARP(171) :} to be set, event-by-event, when variable
energies are allowed, i.e. when \ttt{MSTP(171)=1}. If \ttt{PYINIT} is
called with \ttt{FRAME='CMS'} (\ttt{='FIXT'}), \ttt{PARP(171)}
multiplies the c.m. energy (beam energy) used at initialization.
For the options \ttt{'USER'}, \ttt{'FOUR'} and \ttt{'FIVE'},
\ttt{PARP(171)} is dummy, since there the momenta are set in the
\ttt{P} array.
\iteme{PARP(173) :} event weight to be given by user when
\ttt{MSTP(173)=1}.
\iteme{PARP(174) :} (D=1.) maximum event weight that will be
encountered in \ttt{PARP(173)} during the course of a run with
\ttt{MSTP(173)=1}; to be used to optimize the efficiency of the
event generation. It is always allowed to use a larger bound than
the true one, but with a corresponding loss in efficiency.
\end{entry}
\subsection{General Event Information}
When an event is generated with \ttt{PYEVNT}, some information on
it is stored in the \ttt{MSTI} and \ttt{PARI} arrays of the
\ttt{PYPARS} common block (often copied directly from the internal
\ttt{MINT} and \ttt{VINT} variables). Further information is stored
in the complete event record; see section \ref{ss:evrec}.
Part of the information is only relevant for some subprocesses; by
default everything irrelevant is set to 0. Kindly note that, like the
\ttt{CKIN} constraints described in section \ref{ss:PYswitchkin},
kinematical variables normally (i.e. where it is not explicitly stated
otherwise) refer to the na\"{\i}ve hard scattering, before initial- and
final-state radiation effects have been included.
\drawbox{COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)}%
\label{p:PYPARS2}
\begin{entry}
\itemc{Purpose:} to provide information on latest event generated or,
in a few cases, on statistics accumulated during the run.
\iteme{MSTI(1) :}\label{p:MSTI} specifies the general type of
subprocess that has occurred, according to the ISUB code given
in section \ref{ss:ISUBcode}.
\iteme{MSTI(2) :} whenever \ttt{MSTI(1)} (together with \ttt{MSTI(15)}
and \ttt{MSTI(16)}) are not enough to specify the type of process
uniquely, \ttt{MSTI(2)} provides an ordering of the different
possibilities. This is particularly relevant for the different
colour-flow topologies possible in QCD $2 \to 2$ processes. With
$i = $\ttt{MSTI(15)}, $j = $\ttt{MSTI(16)} and $k = $\ttt{MSTI(2)},
the QCD possibilities are, in the classification scheme of
\cite{Ben84} (cf. section \ref{sss:QCDjetclass}):
\begin{subentry}
\itemn{ISUB = 11,} $i = j$, $\mathrm{q}_i \mathrm{q}_i \to \mathrm{q}_i \mathrm{q}_i$; \\
$k = 1$ : colour configuration $A$. \\
$k = 2$ : colour configuration $B$.
\itemn{ISUB = 11,} $i \neq j$, $\mathrm{q}_i \mathrm{q}_j \to \mathrm{q}_i \mathrm{q}_j$; \\
$k = 1$ : only possibility.
\itemn{ISUB = 12,} $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{q}_l \overline{\mathrm{q}}_l$; \\
$k = 1$ : only possibility.
\itemn{ISUB = 13,} $\mathrm{q}_i \overline{\mathrm{q}}_i \to \mathrm{g} \mathrm{g}$; \\
$k = 1$ : colour configuration $A$. \\
$k = 2$ : colour configuration $B$.
\itemn{ISUB = 28,} $\mathrm{q}_i \mathrm{g} \to \mathrm{q}_i \mathrm{g}$; \\
$k = 1$ : colour configuration $A$. \\
$k = 2$ : colour configuration $B$.
\itemn{ISUB = 53,} $\mathrm{g} \mathrm{g} \to \mathrm{q}_l \overline{\mathrm{q}}_l$; \\
$k = 1$ : colour configuration $A$. \\
$k = 2$ : colour configuration $B$.
\itemn{ISUB = 68,} $\mathrm{g} \mathrm{g} \to \mathrm{g} \mathrm{g}$; \\
$k = 1$ : colour configuration $A$. \\
$k = 2$ : colour configuration $B$. \\
$k = 3$ : colour configuration $C$.
\itemn{ISUB = 83,} $\mathrm{f} \mathrm{q} \to \mathrm{f}' \mathrm{Q}$ (by $t$-channel $\mathrm{W}$ exchange;
does not distinguish colour flows but result of user selection); \\
$k = 1$ : heavy flavour $\mathrm{Q}$ is produced on side 1. \\
$k = 2$ : heavy flavour $\mathrm{Q}$ is produced on side 2.
\end{subentry}
\iteme{MSTI(3) :} the number of partons produced in the hard
interactions, i.e. the number $n$ of the $2 \to n$ matrix elements
used; it is sometimes 3 or 4 when a basic $2 \to 1$ or $2 \to 2$
process has been folded with two $1 \to 2$ initial branchings (like
$\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \mathrm{q}''' \H^0$).
\iteme{MSTI(4) :} number of documentation lines at the beginning of
the common block \ttt{LUJETS} that are given with \ttt{K(I,1)=21};
0 for \ttt{MSTP(125)=0}.
\iteme{MSTI(5) :} number of events generated to date in current
run. In runs with the variable-energy option, \ttt{MSTP(171)=1}
and \ttt{MSTP(172)=2}, only those events that survive (i.e. that
do not have \ttt{MSTI(61)=1}) are counted in this number. That
is, \ttt{MSTI(5)} may be less than the total number of \ttt{PYEVNT}
calls.
\iteme{MSTI(6) :} current frame of event, cf. \ttt{MSTP(124)}.
\iteme{MSTI(7), MSTI(8) :} line number for documentation of
outgoing partons/particles from hard scattering for $2 \to 2$ or
$2 \to 1 \to 2$ processes (else = 0).
\iteme{MSTI(9) :} event class used in current event for $\gamma\mathrm{p}$ or
$\gamma\gamma$ events generated with the \ttt{MSTP(14)=10} option.
\begin{subentry}
\iteme{= 0 :} for other processes than the ones listed above.
\iteme{= 1 :} VMD (for $\gamma\mathrm{p}$) or VMD*VMD (for $\gamma\gamma$).
\iteme{= 2 :} direct (for $\gamma\mathrm{p}$) or VMD*direct (for
$\gamma\gamma$).
\iteme{= 3 :} anomalous (for $\gamma\mathrm{p}$) or VMD*anomalous (for
$\gamma\gamma$).
\iteme{= 4 :} direct*direct (for $\gamma\gamma$).
\iteme{= 5 :} direct*anomalous (for $\gamma\gamma$).
\iteme{= 6 :} anomalous*anomalous (for $\gamma\gamma$).
\end{subentry}
\iteme{MSTI(10) :} is 1 if cross section maximum was violated
in current event, and 0 if not.
\iteme{MSTI(11) :} KF flavour code for beam (side 1) particle.
\iteme{MSTI(12) :} KF flavour code for target (side 2) particle.
\iteme{MSTI(13), MSTI(14) :} KF flavour codes for side 1 and side 2
initial-state shower initiators.
\iteme{MSTI(15), MSTI(16) :} KF flavour codes for side 1 and side 2
incoming partons to the hard interaction.
\iteme{MSTI(17), MSTI(18) :} flag to signal if particle on side
1 or side 2 has been scattered diffractively; 0 if no, 1 if yes.
\iteme{MSTI(21) - MSTI(24) :} KF flavour codes for outgoing partons
from the hard interaction. The number of positions actually used is
process-dependent, see \ttt{MSTI(3)}; trailing positions not used
are set = 0.
\iteme{MSTI(25), MSTI(26) :} KF flavour codes of the products in the
decay of a single $s$-channel resonance formed in the hard interaction.
Are thus only used when \ttt{MSTI(3)=1} and the resonance is allowed
to decay.
\iteme{MSTI(31) :} number of hard or semi-hard scatterings that occurred
in the current event in the multiple-interaction scenario; is = 0 for a
low-$p_{\perp}$ event.
\iteme{MSTI(41) :} the number of pile-up events generated in the latest
\ttt{PYEVNT} call (including the first, `hard' event).
\iteme{MSTI(42) - MSTI(50) :} ISUB codes for the events 2--10
generated in the pile-up-events scenario. The first event ISUB code is
stored in \ttt{MSTI(1)}. If \ttt{MSTI(41)} is less than 10, only as
many positions are filled as there are pile-up events. If MSTI(41) is
above 10, some ISUB codes will not appear anywhere.
\iteme{MSTI(51) :} normally 0 but set to 1 if a \ttt{PYUPEV} call
did not return an event, such that \ttt{PYEVNT} could not generate
an event. For further details, see end of section \ref{ss:PYnewproc}.
\iteme{MSTI(52) :} counter for the number of times the current event
configuration failed in the generation machinery. For accepted events
this is always 0, but the counter can be used inside \ttt{PYUPEV}
to check on anomalous occurrences. For further
details, see end of section \ref{ss:PYnewproc}.
\iteme{MSTI(61) :} status flag set when events are generated. It is
only of interest for runs with variable energies, \ttt{MSTP(171)=1},
with the option \ttt{MSTP(172)=2}.
\begin{subentry}
\iteme{= 0 :} an event has been generated.
\iteme{= 1 :} no event was generated, either because the c.m. energy
was too low or because the Monte Carlo phase space point selection
machinery rejected the trial point. A new energy is to be picked by
the user.
\end{subentry}
\boxsep
\iteme{PARI(1) :}\label{p:PARI} total integrated cross section for
the processes
under study, in mb. This number is obtained as a by-product of the
selection of hard-process kinematics, and is thus known with better
accuracy when more events have been generated. The value stored here
is based on all events until the latest one generated.
\iteme{PARI(2) :} is the ratio \ttt{PARI(1)/MSTI(5)}, i.e. the ratio
of total integrated cross section and number of events generated.
Histograms filled with unit event weight have to be
multiplied by this factor, at the end of the run, to convert
results to mb. For \ttt{MSTP(142)=1}, \ttt{MSTI(5)} is replaced by
the sum of \ttt{PARI(10)} values. Histograms are then filled with
weight \ttt{PARI(10)} for each event and multiplied by \ttt{PARI(2)}
at the end. In runs with the variable-energy option, \ttt{MSTP(171)=1}
and \ttt{MSTP(172)=2}, only those events that survive (i.e. that
do not have \ttt{MSTI(61)=1}) are counted.
calls.
\iteme{PARI(9) :} is weight \ttt{WTXS} returned from \ttt{PYEVWT}
call when \ttt{MSTP(142)}$\geq 1$, otherwise is 1.
\iteme{PARI(10) :} is compensating weight \ttt{1./WTXS} that should
be associated to events when \ttt{MSTP(142)=1}, else is 1.
\iteme{PARI(11) :} $E_{\mrm{cm}}$, i.e. total c.m. energy.
\iteme{PARI(12) :} $s$, i.e. squared total c.m. energy.
\iteme{PARI(13) :} $\hat{m} = \sqrt{\hat{s}}$, i.e. mass of the
hard-scattering subsystem.
\iteme{PARI(14) :} $\hat{s}$ of the hard subprocess
($2 \to 2$ or $2 \to 1$).
\iteme{PARI(15) :} $\hat{t}$ of the hard subprocess
($2 \to 2$ or $2 \to 1 \to 2$).
\iteme{PARI(16) :} $\hat{u}$ of the hard subprocess
($2 \to 2$ or $2 \to 1 \to 2$).
\iteme{PARI(17) :} $\hat{p}_{\perp}$ of the hard subprocess
($2 \to 2$ or $2 \to 1 \to 2$),
evaluated in the rest frame of the hard interaction.
\iteme{PARI(18) :} $\hat{p}_{\perp}^2$ of the hard subprocess;
see \ttt{PARI(17)}.
\iteme{PARI(19) :} $\hat{m}'$, the mass of the complete three- or
four-body final state in $2 \to 3$ or $2 \to 4$ processes (while
$\hat{m}$, given in \ttt{PARI(13)}, here corresponds to the one-
or two-body central system).
Kinematically $\hat{m} \leq \hat{m}' \leq E_{\mrm{cm}}$.
\iteme{PARI(20) :} $\hat{s}' = \hat{m}'^2$; see \ttt{PARI(19)}.
\iteme{PARI(21) :} $Q$ of the hard-scattering subprocess. The exact
definition is process-dependent, see \ttt{MSTP(32)}.
\iteme{PARI(22) :} $Q^2$ of the hard-scattering subprocess; see
\ttt{PARI(21)}.
\iteme{PARI(23) :} $Q$ of the outer hard-scattering subprocess.
Agrees with \ttt{PARI(21)} for a $2 \to 1$ or $2 \to 2$ process.
For a $2 \to 3$ or $2 \to 4$ $\mathrm{W}/\mathrm{Z}$ fusion process, it is set by
the $\mathrm{W}/\mathrm{Z}$ mass scale, and for subprocesses 121 and 122 by the
heavy-quark mass.
\iteme{PARI(24) :} $Q^2$ of the outer hard-scattering subprocess;
see \ttt{PARI(23)}.
\iteme{PARI(25) :} $Q$ scale used as maximum virtuality in parton
showers. Is equal to \ttt{PARI(23)}, except for
deep-inelastic-scattering processes when \ttt{MSTP(22)}$\geq 1$.
\iteme{PARI(26) :} $Q^2$ scale in parton showers; see \ttt{PARI(25)}.
\iteme{PARI(31), PARI(32) :} the momentum fractions $x$ of the
initial-state parton-shower initiators on side 1 and 2, respectively.
\iteme{PARI(33), PARI(34) :} the momentum fractions $x$ taken by the
partons at the hard interaction, as used e.g. in the
parton-distribution functions.
\iteme{PARI(35) :} Feynman-$x$,
$x_{\mrm{F}} = x_1 - x_2 = $\ttt{PARI(33)}$-$\ttt{PARI(34)}.
\iteme{PARI(36) :}
$\tau = \hat{s}/s = x_1 \, x_2 = $\ttt{PARI(33)}$\times$\ttt{PARI(34)}.
\iteme{PARI(37) :} $y = (1/2) \ln(x_1/x_2)$, i.e. rapidity of the
hard-interaction subsystem in the c.m. frame of the event as a whole.
\iteme{PARI(38) :} $\tau' = \hat{s}'/s = $\ttt{PARI(20)/PARI(12)}.
\iteme{PARI(39), PARI(40) :} the primordial $k_{\perp}$ values
selected in the two beam remnants.
\iteme{PARI(41) :} $\cos\hat{\theta}$, where $\hat{\theta}$ is the
scattering angle of a $2 \to 2$ (or $2 \to 1 \to 2$) interaction,
defined in the rest frame of the hard-scattering subsystem.
\iteme{PARI(42) :} $x_{\perp}$, i.e. scaled transverse momentum of the
hard-scattering subprocess,
$x_{\perp} = 2 \hat{p}_{\perp}/E_{\mrm{cm}}$.
\iteme{PARI(43), PARI(44) :} $x_{L3}$ and $x_{L4}$, i.e. longitudinal
momentum fractions of the two scattered partons, in the range
$-1 < x_{\mrm{L}} < 1$, in the c.m. frame of the event as a whole.
\iteme{PARI(45), PARI(46) :} $x_3$ and $x_4$, i.e. scaled energy
fractions of the two scattered partons, in the c.m. frame of the
event as a whole.
\iteme{PARI(47), PARI(48) :} $y^*_3$ and $y^*_4$, i.e. rapidities
of the two scattered partons in the c.m. frame of the
event as a whole.
\iteme{PARI(49), PARI(50) :} $\eta^*_3$ and $\eta^*_4$, i.e.
pseudorapidities of the two scattered partons in the c.m. frame
of the event as a whole.
\iteme{PARI(51), PARI(52) :} $\cos\theta^*_3$ and $\cos\theta^*_4$,
i.e. cosines of the polar angles of the two scattered partons in
the c.m. frame of the event as a whole.
\iteme{PARI(53), PARI(54) :} $\theta^*_3$ and $\theta^*_4$, i.e.
polar angles of the two scattered partons, defined in the range
$0 < \theta^* < \pi$, in the c.m. frame of the event as a whole.
\iteme{PARI(55), PARI(56) :} azimuthal angles $\phi^*_3$ and
$\phi^*_4$ of the two scattered partons, defined in the range
$-\pi < \phi^* < \pi$, in the c.m. frame of the event as a whole.
\iteme{PARI(61) :} multiple interaction enhancement factor for
current event. A large value corresponds to a central collision
and a small value to a peripheral one.
\iteme{PARI(65) :} sum of the transverse momenta of partons
generated at the hardest interaction of the event, excluding
initial- and final-state radiation, i.e. $2 \times$\ttt{PARI(17)}.
\iteme{PARI(66) :} sum of the transverse momenta of all partons
generated at the hardest interaction, including initial- and
final-state radiation, resonance decay products, and primordial
$k_{\perp}$.
\iteme{PARI(67) :} sum of transverse momenta of partons generated
at hard interactions, excluding the hardest one (see \ttt{PARI(65)}),
and also excluding initial- and final-state radiation. Is non-vanishing
only in the multiple-interaction scenario.
\iteme{PARI(68) :} sum of transverse momenta of all partons generated
at hard interactions, excluding the hardest one (see \ttt{PARI(66)}),
but including initial- and final-state radiation. Is non-vanishing only
in the multiple-interaction scenario.
\iteme{PARI(69) :} sum of transverse momenta of all partons generated
in hard interactions (\ttt{PARI(66) + PARI(68)}) and, additionally,
of all beam remnant partons.
\iteme{PARI(71), PARI(72) :} sum of the momentum fractions $x$ taken
by initial-state parton-shower initiators on side 1 and and side 2,
excluding those of the hardest interaction. Is non-vanishing only in
the multiple-interaction scenario.
\iteme{PARI(73), PARI(74) :} sum of the momentum fractions $x$ taken
by the partons at the hard interaction on side 1 and side 2, excluding
those of the hardest interaction. Is non-vanishing only in the
multiple-interaction scenario.
\iteme{PARI(75), PARI(76) :} the $x$ value of a photon that branches
into quarks or gluons, i.e. $x$ at interface between initial-state QED
and QCD cascades.
\iteme{PARI(77), PARI(78) :} the $\chi$ values selected for beam
remnants that are split into two objects, describing how the energy
is shared (see \ttt{MSTP(92)} and \ttt{MSTP(94)}); is vanishing if
no splitting is needed.
\iteme{PARI(81) :} size of the threshold factor (enhancement or
suppression) in the latest event with heavy-flavour production;
see \ttt{MSTP(35)}.
\iteme{PARI(91) :} average multiplicity $\br{n}$ of pile-up events,
see \ttt{MSTP(133)}. Only relevant for \ttt{MSTP(133)=} 1 or 2.
\iteme{PARI(92) :} average multiplicity $\langle n \rangle$ of pile-up
events as actually simulated, i.e. with multiplicity = 0 events
removed and the high-end tail truncated. Only relevant for
\ttt{MSTP(133)=} 1 or 2.
\iteme{PARI(93) :} for \ttt{MSTP(133)=1} it is the probability that
a beam crossing will produce a pile-up event at all, i.e. that there
will be at least one hadron--hadron interaction; for
\ttt{MSTP(133)=2} the probability that a beam crossing will produce
a pile-up event with one hadron--hadron interaction of the desired rare
type.
\end{entry}
\subsection{How to include external processes in PYTHIA}
\label{ss:PYnewproc}
Despite a large repertory of processes in {\tsc{Pythia}}, the number
of missing ones clearly is even larger, and with time this discrepancy
is likely to increase. There are several reasons why it is not
practicable to imagine a {\tsc{Pythia}} which has `everything'. One is
the amount of time it takes to implement a process for the single
{\tsc{Pythia}} author, compared with the rate of new cross section results
produced by the rather larger matrix-element calculations community.
Another is the length of currently produced matrix-element expressions,
which would make the program very bulky. A third argument is that,
whereas the phase space of $2 \to 1$ and $2 \to 2$ processes can be set
up once and for all according to a reasonably flexible machinery,
processes with more final-state particles are less easy to generate.
To achieve a reasonable efficiency, it is necessary to tailor the
phase-space selection procedure to the dynamics of the given process,
and to the desired experimental cuts.
If the desired subprocess is missing, it can be included into {\tsc{Pythia}}
as an `external' subprocess. In this section we will describe how it
is possible to
specify the partonic state of some hard-scattering process in an
interface common block. {\tsc{Pythia}} will read this common block, and add
initial- and final-state showers, beam remnants and underlying events,
fragmentation and decays, to build up an event in as much detail as an
ordinary {\tsc{Pythia}} one. You may also use {\tsc{Pythia}} to mix events of different
kinds, and to keep track of cross section statistics. You have to
provide the matrix elements, the phase-space generator, and the storage
of event information in the common block.
First a minor comment, however. Some processes may be seen just as
trivial modifications of already existing ones. For instance, you might
want to add some extra term, corresponding to contact interactions,
to the matrix elements of a {\tsc{Pythia}} $2 \to 2$ process. In that case it is
not necessary to go through the machinery below, but instead you can use
the \ttt{PYEVWT} routine to introduce an additional weight for the
event, defined as the ratio of the modified to the unmodified
differential cross sections. If you use the option \ttt{MSTP(142)=2},
this weight is considered as part of the `true' cross section of the
process, and the generation is changed accordingly.
The more generic facility for including an external process is a bit
more complicated, and involves two routines and one common block. All
names contain \ttt{UP}, which is short for User Process.
If you want to include a new process, first you have to pick an unused
subprocess number ISUB (see tables in section \ref{ss:ISUBcode}). For
instance, the numbers 191--200 are currently unused, so this might be a
logical place to put a new process. This number and the `title' of the
process (plus \ttt{SIGMAX}, to be described below) have to be given in
to {\tsc{Pythia}} in a subroutine call
\drawbox{CALL PYUPIN(ISUB,TITLE,SIGMAX)}\label{p:PYUPIN}
\boxsep
\noindent
before the call to \ttt{PYINIT}. The \ttt{TITLE} can be any character
string up to 28 characters, e.g.
\begin{verbatim}
CALL PYUPIN(191,'g + g -> t + tbar + gamma',SIGMAX)
\end{verbatim}
The call to \ttt{PYUPIN} tells the program that a process ISUB exists,
but not that you want to generate it. This is done, as with normal
processes, by setting \ttt{MSUB(ISUB)=1} before the \ttt{PYINIT}
call.
Once the event generation chain has been started and \ttt{PYEVNT} is
called to generate an event, this routine may in its turn call the
routine \ttt{PYUPEV}, which is the routine you must supply, in which
the next event is selected. (A dummy copy of \ttt{PYUPEV} has been
included at the end of {\tsc{Pythia}}; depending on the machine you may have to
comment out this copy when you link your own.) The call arguments are
\drawbox{CALL PYUPEV(ISUB,SIGEV)}\label{p:PYUPEV}
\boxsep
\noindent
where \ttt{ISUB} is given by \ttt{PYEVNT}, while \ttt{SIGEV} is to be
calculated (see below) and returned to \ttt{PYEVNT}. If there is only
one user-defined process, then the \ttt{ISUB} input is superfluous;
otherwise it is necessary to branch to the relevant process.
The \ttt{SIGEV} variable is supposed to give the differential
cross section ot the current event, times the phase-space volume
within which events are generated, expressed in millibarns. This means
that, in the limit that many events are generated, the average value
of \ttt{SIGEV} gives the total cross section of the simulated process.
The \ttt{SIGMAX} value, handed to {\tsc{Pythia}} in the \ttt{PYUPIN} call, is
assumed to be the maximum value that \ttt{SIGEV} will reach. Events
will be accepted with a probability \ttt{SIGEV/SIGMAX}, i.e. the
acceptance/rejection of events according to
differential cross section is done by \ttt{PYEVNT}, not by the user.
This means that the events that come out in the end all have unit
weight, i.e. the user does not have to worry about events with
different weights. It also allows several subprocesses to be
generated together, in the proper mixture.
Of course, the tricky part is that the differential cross section
usually is strongly peaked in a few regions of the phase space, such
that the average probability to accept an event,
$\langle$\ttt{SIGEV}$\rangle /$\ttt{SIGMAX} is
small. It may then be necessary to find a suitable set of transformed
phase-space coordinates, for which the correspondingly transformed
differential cross section is better behaved.
To avoid unclarities, here is a more formal version of the two above
paragraphs. Call $\d X$ the differential phase space, e.g. for a
$2 \to 2$
process $\d X = \d x_1 \, \d x_2 \, \d \hat{t}$, where $x_1$ and $x_2$
are the momentum fractions carried by the two incoming partons and
$\hat{t}$ the Mandelstam variable of the scattering. Call
$\d \sigma / \d X$ the differential cross section
of the process, e.g. for $2 \to 2$: $\d \sigma / \d X = \sum_{ij}
f_i(x_1,Q^2) \, f_j(x_2,Q^2) \, \d \hat{\sigma}_{ij} / \d \hat{t}$,
i.e. the product of parton distributions and hard-scattering matrix
elements, summed over all allowed incoming flavours $i$ and $j$.
The physical cross section that one then wants to generate is
$\sigma = \int (\d \sigma / \d X) \, \d X$, where the integral is over
the allowed phase-space volume. The event generation procedure consists
of selecting an $X$ uniformly in $\d X$ and then evaluating the weight
$\d \sigma / \d X$ at this point. \ttt{SIGEV} is now simply
\ttt{SIGEV}$ = \d \sigma / \d X \, \int \d X$, i.e. the differential
cross section times the considered volume of phase space. Clearly,
when averaged over many events, \ttt{SIGEV} will correctly estimate
the desired cross section. If \ttt{SIGEV} fluctuates too much, one
may try to transform to new variables $X'$,
where events are now picked accordingly to $\d X'$ and
\ttt{SIGEV}$ = \d \sigma / \d X' \, \int \d X'$.
A warning. It is important that $X$ is indeed uniformly picked within
the allowed phase space, alternatively that any Jacobians are properly
taken into account. For instance, in the case above, one approach
would be to pick $x_1$, $x_2$ and $\hat{t}$ uniformly in the ranges
$0 < x_1 < 1$, $0 < x_2 < 1$, and $-s < \hat{t} < 0$, with full phase
space
volume $\int \d X = s$. The cross section would only be non-vanishing
inside the physical region given by $-s x_1 x_2 < \hat{t}$ (in the
massless case), i.e. Monte Carlo efficiency is likely to be low.
However, if one were to choose $\hat{t}$ values only in the range
$-\hat{s} < \hat{t} < 0$, small $\hat{s}$ values would be favoured,
since the density of selected $\hat{t}$ values would be larger there.
Without the use of a compensating Jacobian $\hat{s}/s$, an incorrect
answer would be obtained. Alternatively, one could start out with a
phase space like $\d X = \d x_1 \, \d x_2 \, \d (\cos\hat{\theta})$,
where the limits decouple. Of course, the $\cos\hat{\theta}$ variable
can be translated back into a $\hat{t}$, which will then always be in
the desired range $-\hat{s} < \hat{t} < 0$. The transformation itself
here gives the necessary Jacobian.
If you do not know how big \ttt{SIGMAX} is, you can put it to some
very small value (but larger than zero, however) and do an
exploratory run. When the program encounters events with
\ttt{SIGEV}$ > $\ttt{SIGMAX}, a warning message is printed, which
gives the new \ttt{SIGMAX} that the program will use from then
on. Hopefully such maximum violations only appear at the beginning of
the run, and later \ttt{SIGMAX} stabilizes to a level that can then
be used as input for a second, correct run.
If you want to do the event rejection yourself, simply put \ttt{SIGEV}
equal to \ttt{SIGMAX}. In that case events will not be rejected by
{\tsc{Pythia}} (except if there is something else wrong with them). If
\ttt{SIGMAX} is the correct total cross section of the process, event
mixing with other processes will still work fine. You could also decide
not to reject any events, but to use weighted ones. In that case you
can only have one ISUB switched on in a run, since the program will not
know how to mix different kinds of events, and you cannot use {\tsc{Pythia}} to
do cross section statistics for you. Therefore you could, for instance,
put \ttt{SIGMAX = SIGEV = 1}, and use a common block to transfer event
weight and other information from your \ttt{PYUPEV} routine to your
main program.
In addition to the \ttt{SIGEV} value returned for each event, it is also
necessary to return the event itself. This is done via the common block
\drawbox{COMMON/PYUPPR/NUP,KUP(20,7),PUP(20,5),NFUP,IFUP(10,2),%
Q2UP(0:10)}\label{p:PYUPPR}
\boxsep
The first part closely parallels the standard event record in the
\ttt{LUJETS} common block, see section \ref{ss:evrec}, although with
a few simplifications. The number \ttt{NUP}\label{p:NUP} gives the
number of particles involved in the process, where a particle may be
a quark, a lepton, a gauge boson, or anything else. The first two
are simply the two incoming particles that initiate the hard
scattering, while the remaining \ttt{NUP-2} are the outgoing
particles from the hard process. For each particle \ttt{I}, with
$1 \leq $\ttt{I}$ \leq $\ttt{NUP}, the following information is stored:
\begin{entry}
\iteme{KUP(I,1) :}\label{p:KUP} is = 1 normally. However, if you put
it = 2 that signifies intermediate states that are not to be treated
by {\tsc{Pythia}}, but are included only to make the event record easier to
read.
\iteme{KUP(I,2) :} is the flavour code of a particle, i.e. the two
incoming partons for \ttt{I} = 1 and 2, and the outgoing particles
for \ttt{I}$ \geq 3$. The flavour codes are the standard KF ones,
as used elsewhere in the program.
\iteme{KUP(I,3) :} may be used to indicate the position of a mother.
Such information may again make the record more readable, but is not
really needed, and so one may well put all \ttt{KUP(I,3)=0}.
\iteme{KUP(I,4) :} for a final-state parton which carries colour,
\ttt{KUP(I,4)} gives the position of the parton from which the colour
comes; otherwise it must be 0.
\iteme{KUP(I,5) :} for a final-state parton that carries anticolour,
\ttt{KUP(I,5)} gives the position of the parton from which the
anticolour comes; else it must be 0.
\iteme{KUP(I,6) :} for an initial-state parton that carries colour,
\ttt{KUP(I,6)} gives the position of the parton to which the colour
goes; else it must be 0.
\iteme{KUP(I,7) :} for an initial-state parton which carries
anticolour, \ttt{KUP(I,7)} gives the position of the parton to
which the anticolour goes; else it must be 0.
\iteme{PUP(I,1) :}\label{p:PUP} $p_x$, i.e. $x$ momentum.
\iteme{PUP(I,2) :} $p_y$, i.e. $y$ momentum.
\iteme{PUP(I,3) :} $p_z$, i.e. $z$ momentum.
\iteme{PUP(I,4) :} $E$, i.e. energy.
\iteme{PUP(I,5) :} $m$, i.e. mass.
\end{entry}
After this brief summary, we proceed with more details and examples.
To illustrate the issue of documentation in \ttt{KUP(I,1)} and
\ttt{KUP(I,3)}, consider the case of $\mathrm{W}^+$ production and decay to
$\u \overline{\mathrm{d}}$, maybe as part of a more complex process. The final-state
particles clearly are $\u$ and $\overline{\mathrm{d}}$, so the $\mathrm{W}^+$ need not be
given at all, but it would make the event listing easier to read.
Therefore one should add the $\mathrm{W}^+$, but with \ttt{KUP(I,1)=2}.
(If the $\mathrm{W}^+$ would have been added with \ttt{KUP(I,1)=1}, it would
later have been treated by {\tsc{Pythia}}/{\tsc{Jetset}}, which means it would have
been allowed to decay once more.) If the $\mathrm{W}^+$ is in line 3, the
$\u$ in 4 and the $\overline{\mathrm{d}}$ in 5, one could further put \ttt{KUP(4,3)=3}
and \ttt{KUP(5,3)=3} to indicate that the $\u$ and $\overline{\mathrm{d}}$ in lines
4 and 5 come from the $\mathrm{W}^+$ stored in line 3.
The switch \ttt{MSTP(128)} works in the same way for user-defined
processes as for ordinary ones, i.e. decay products of
resonances can optionally be omitted from the documentation section
of the event record, and history pointers can be set slightly
differently. The information the program has at its disposal
for this purpose is in \ttt{KUP(I,3)}; an entry with this
value non-zero is considered as a resonance decay product.
The colour-flow information for coloured particles (quarks, gluons,
leptoquarks, \ldots) is needed to set up parton showers and
fragmentation properly. Sometimes many different colour flows are
possible for one and the same process, as discussed in section
\ref{sss:QCDjetclass}.
It is up to you whether or not you will include all possible colour
flows in the appropriate mixture, but at least you must pick some
representative colour configuration. Consider e.g. the case of
$\mathrm{g}(1) + \mathrm{g}(2) \to \mathrm{q}(3) + \overline{\mathrm{q}}(4)$, where the numbers give the
position in the array. It is clear the $\mathrm{q}$ must get its colour from
either of the two gluons, which means there are (at least) two
possibilities. Picking the $\mathrm{q}$ colour
to come from gluon 1, one would thus write
\ttt{KUP(3,4)=1}, to be read `the colour of parton 3 comes from
parton 1'. By implication therefore also \ttt{KUP(1,6)=3}, i.e.
`the colour of parton 1 goes to parton 3', i.e. the colour flow is
bookkept doubly. The anticolour now must flow from parton 2 to parton
4, i.e. \ttt{KUP(2,7)=4} and \ttt{KUP(4,5)=2}. This completely
specifies the colours of the $\mathrm{q}$ and $\overline{\mathrm{q}}$, but not of the
two gluons. In fact, one colour in the initial state `annihilates'
between the $\mathrm{g}(1)$ and $\mathrm{g}(2)$, i.e. the anticolour of gluon 1 and the
colour of gluon 2 match, which may be expressed by \ttt{KUP(1,7)=2}
and \ttt{KUP(2,6)=1}. In other words colour/anticolour of an
initial-state parton may either go to a final-state parton or to
another
initial-state parton. Correspondingly, the colour/anticolour of a
final-state parton may come either from an initial-state parton or
from another final-state parton. An example of the latter possibility
is $\mathrm{W}$ decays, or generically the decay of any colour-singlet
particle. (Thus a third colour flow above is represented by
$\mathrm{g} \mathrm{g} \to \H^0 \to \mathrm{q} \overline{\mathrm{q}}$, where no colour passes through the
Higgs, and therefore colour flows between the two gluons and,
separately, between the $\mathrm{q}$ and $\overline{\mathrm{q}}$.)
Storing of momenta should be straightforward, but a few comments must
be made. Even if you ask, in the \ttt{PYINIT} call, to have events
generated in a fixed target or a user-specified frame, at
intermediate stages {\tsc{Pythia}} will still work in the c.m. frame of the
two incoming beam particles, with the first beam moving in the $+z$
direction and the second in the $-z$ one. This c.m. frame must also
be used when giving the momenta of the process. In addition, the two
incoming partons in lines 1 and 2 are assumed massless. Therefore the
initial-state partons are characterized only by the two energies
\ttt{P(1,4)} and \ttt{P(2,4)}, with \ttt{P(1,3) = P(1,4)},
\ttt{P(2,3) = -P(2,4)}, and everything else is zero. In the
final state, energies, momenta and masses are free, but must add up to
give the same four-momentum as that of the initial state. All momenta
are given in GeV, with the speed of light $c = 1$.
The second part of the \ttt{PYUPPR} common block is used to regulate the
initial- and final-state showering, as follows:
\begin{entry}
\iteme{Q2UP(0) :}\label{p:Q2UP} $Q^2$ scale of initial-state showers.
\iteme{NFUP :}\label{p:NFUP} number of parton pairs that undergo
final-state showers.
\iteme{IFUP(IF,1), IFUP(IF,2) :}\label{p:IFUP} positions of the two
partons of a final-state showering pair, where the index \ttt{IF}
runs between 1 and \ttt{NFUP}.
\iteme{Q2UP(IF) :} the $Q^2$ scale of the final-state shower between
parton pair \ttt{IF} above.
\end{entry}
If you do not want any showering at all, you can put \ttt{MSTP(61)=0}
and \ttt{MSTP(71)=0}, and then you do not have to give the above
quantities. In general the scale choices $Q^2$ are not unique, which
means that some guesswork is involved. Since the showers add extra
partonic activity at mass scales below the mentioned $Q^2$ choices, the
\ttt{Q2UP} should be of the order of the phase-space cut-offs, so as to
provide a reasonably smooth joining between partonic activity from
matrix elements and that from showers. There are a few cases where
choices are rather easy. In the decay of any $s$-channel colour
neutral state, such as a $\mathrm{W}^{\pm}$, the $Q^2$ scale of final-state
showers is just set by the squared mass of the resonance. For
initial-state radiation, \ttt{Q2UP(0)} should be about the same as
the $Q^2$
scale used for the evaluation of parton distributions for the hard
process, up to some factor of order unity. (One frequent choice for
this factor would be 4, if your parton-distribution scale is something
like the squared transverse momentum, simply because $m^2$ is of order
$4 p_{\perp}^2$.)
The `parton'-shower evolution actually also can include photon emission
off quarks and leptons, if the shower switches are properly set. It is
not possible to define only one particle in the above arrays, since
it would then not be possible to conserve energy and momentum in the
shower. You can very well have a pair where only one of the two can
branch, however. For instance, in a $\mathrm{g} \gamma$ final state, only
the gluon can shower, but the photon can lose energy to the gluon
in such a way that the gluon branchings becomes possible.
Currently, it is not possible to do showering where three or more
final-state particles are involved at the same time. This may be added
at a
later stage. It is therefore necessary to subdivide suitably into pairs,
and maybe leave some (especially colour-neutral) particles unshowered.
You are free to make use of whatever tools you want in your \ttt{PYUPEV}
routine, and normally there would be no contact with the rest of
{\tsc{Pythia}}, except as described above. However, you may want to use
some of the tools already available. One attractive possibility
is to use \ttt{PYSTFU} for parton-distribution-function evaluation.
Other possible tools could be \ttt{RLU} for random-number
generation, \ttt{ULALPS} for $\alpha_{\mathrm{s}}$ evaluation,
\ttt{ULALEM} for evaluation of a running $\alpha_{\mathrm{em}}$, and
maybe a few more.
We end with a few comments on anomalous situations. In some cases
one may want to decide, inside \ttt{PYUPEV}, when to stop
the event-generation loop. This is the case, for instance, if event
configurations are read in from a file, and the end of the file is
reached. One might be tempted just to put \ttt{SIGEV=0} when this
happens. Then \ttt{PYEVNT} will discard the event, as part of the
matrix-element-weighting procedure. However, next \ttt{PYEVNT} will
generate another event, which normally means a new request to
\ttt{PYUPEV}, so one does not really get out of the loop. Instead
you should put \ttt{NUP=0}. If the program encounters this value at a
return from \ttt{PYUPEV}, then it will also exit from \ttt{PYEVNT},
without incrementing the counters for the number of events generated.
It is then up to you to have a check on this condition in your main
event-generation loop. This you do either by looking at \ttt{NUP} or
at \ttt{MSTI(51)}; the latter is set to 1 if no event was generated.
It may also happen that a user-defined configuration fails elsewhere
in the \ttt{PYEVNT} call. For instance, the beam-remnant treatment
occasionally encounters situations it cannot handle, wherefore the
hard interaction is rejected and a new one generated. This happens also
with ordinary (not user-defined) events, and usually comes about as a
consequence of the initial-state radiation description leaving too
little energy for the remnant. If the same hard scattering were to
be used as input for a new initial-state radiation and beam-remnant
attempt, it could then work fine. There is a possibility to give events
that chance, as follows. \ttt{MSTI(52)} counts the number of times a
hard-scattering configuration has failed to date. If you come in to
\ttt{PYUPEV} with \ttt{MSTI(52)} non-vanishing, this means that the
latest configuration failed. So long as the contents of the \ttt{PYUPPR}
common block are not changed, such an event may be given another try.
For instance, a line
\begin{verbatim}
IF(MSTI(52).GE.1.AND.MSTI(52).LE.4) RETURN
\end{verbatim}
at the beginning of \ttt{PYUPEV} will give each event up to five
tries; thereafter a new one would be generated as usual. Note that the
counter for the number of events is updated at each new try. The
fraction of failed configurations is given in the bottom line of
the \ttt{PYSTAT(1)} table.
The above comment only refers to very rare occurrences (less than
one in a hundred), which are not errors in a strict sense; for instance,
they do not produce any error messages on output. If you get
warnings and error messages that the program does not understand the
flavour codes or cannot reconstruct the colour flows, it is due to
faults of yours, and giving such events more tries is not going to
help.
\subsection{How to run PYTHIA with varying energies}
\label{ss:PYvaren}
It is possible to use {\tsc{Pythia}} in a mode where the energy can be
varied from one event to the next, without the need to reinitialize
with a new \ttt{PYINIT} call. This allows a significant speed-up of
execution, although it is not as fast as running at a fixed
energy. It can not be used for everything --- we will come to
the fine print at the end --- but it should be applicable for most
tasks.
The master switch to access this possibility is in \ttt{MSTP(171)}.
By default it is off, so you must set \ttt{MSTP(171)=1} before
initialization. There are two submodes of running, with
\ttt{MSTP(172)} being 1 or 2. In the former mode, {\tsc{Pythia}} will generate
an event at the requested energy. This means that you have to know
which energy you want beforehand. In the latter mode, {\tsc{Pythia}}
will often return without having generated an event --- with flag
\ttt{MSTI(61)=1} to signal that --- and you are then requested to give
a new energy. The energy spectrum of accepted events will then,
in the end, be your naive input spectrum weighted with the
cross-section of the processes you study. We will come back to
this.
The energy can be varied, whichever frame is given in the \ttt{PYINIT}
call. When the frame is \ttt{'CMS'}, \ttt{PARP(171)} should be filled
with the fractional energy of each event, i.e.
$E_{\mrm{cm}} =$\ttt{PARP(171)}$\times$\ttt{WIN}, where \ttt{WIN} is
the nominal c.m. energy of the \ttt{PYINIT} call. Here \ttt{PARP(171)}
should normally be smaller than unity, i.e. initialization should
be done at the maximum energy to be encountered. For the \ttt{'FIXT'}
frame, \ttt{PARP(171)} should be filled by the fractional beam energy
of that one, i.e. $E_{\mrm{beam}} = $\ttt{PARP(171)}$\times$\ttt{WIN}.
For the \ttt{'USER'}, \ttt{'FOUR'} and \ttt{'FIVE'} options, the
two four-momenta are given in for each event in the same format as
used for the \ttt{PYINIT} call. Note that there is a minimum c.m.
energy allowed, \ttt{PARP(2)}. If you give in values below this,
the program will stop for \ttt{MSTP(172)=1}, and will return with
\ttt{MSTI(61)=1} for \ttt{MSTP(172)=1}.
To illustrate the use of the \ttt{MSTP(172)=2} facility, consider the
case of beamstrahlung in $\e^+\e^-$ linear colliders. This is just for
convenience; what is said here can be translated easily into other
situations. Assume that the beam spectrum is given by $D(z)$, where
$z$ is the fraction retained by the original $\mathrm{e}$ after beamstrahlung.
Therefore $0 \leq z \leq 1$ and the integral of $D(z)$ is unity.
This is not perfectly general; one could imagine branchings
$\mathrm{e}^- \to \mathrm{e}^- \gamma \to \mathrm{e}^-\mathrm{e}^+\mathrm{e}^-$, which gives a multiplication
in the number of beam particles. This could either be expressed in
terms of a $D(z)$ with integral larger than unity or in terms of an
increased luminosity. We will assume the latter, and use $D(z)$
properly normalized. Given a nominal $s = 4E_{\mrm{beam}}^2$,
the actual $s'$ after beamstrahlung is given by $s' = z_1 z_2 s$.
For a process with a cross section $\sigma(s)$ the total
cross section is then
\begin{equation}
\sigma_{\mrm{tot}} = \int_0^1 \int_0^1 D(z_1) \, D(z_2)
\sigma(z_1 z_2 s) \, \d z_1 \, \d z_2~.
\end{equation}
The cross section $\sigma$ may in itself be an integral over a number
of additional phase space variables. If the maximum of the differential
cross section is known, a correct procedure to generate events is
\begin{Enumerate}
\item pick $z_1$ and $z_2$ according to $D(z_1) \, \d z_1$ and
$D(z_2) \, \d z_2$, respectively;
\item pick a set of phase space variables of the process, for the
given $s'$ of the event;
\item evaluate $\sigma(s')$ and compare with $\sigma_{\mathrm{max}}$;
\item if event is rejected, then return to step 1 to generate new
variables;
\item else continue the generation to give a complete event.
\end{Enumerate}
You as a user are assumed to take care of step 1, and present the
resulting kinematics with incoming $\mathrm{e}^+$ and $\mathrm{e}^-$ of varying energy.
Thereafter {\tsc{Pythia}} will do steps 2--5, and either return an event or
put \ttt{MSTI(61)=1} to signal failure in step 4.
The maximization procedure does search in phase space to find
$\sigma_{\mathrm{max}}$, but it does not vary the $s'$ energy in this process.
Therefore the maximum search in the \ttt{PYINIT} call should be
performed where the cross section is largest. For processes with
increasing cross section as a function of energy this means at the
largest energy that will ever be encountered, i.e. $s' = s$ in the
case above. This is the `standard' case, but often one encounters
other behaviours, where more complicated procedures are needed. One
such case would be the process $\e^+\e^- \to \mathrm{Z}^{*0} \to \mathrm{Z}^0 \H^0$,
which is known to have a cross section that increases near the
threshold but is decreasing asymptotically. If one already knows
that the maximum, for a given Higgs mass, appears at 300 GeV, say,
then the \ttt{PYINIT} call should be made with that energy, even if
subsequently one will be generating events for a 500 GeV collider.
In general, it may be necessary to modify the selection of $z_1$ and
$z_2$ and assign a compensating event weight. For instance, consider
a process with a cross section behaving roughly like $1/s$. Then the
$\sigma_{\mrm{tot}}$ expression above may be rewritten as
\begin{equation}
\sigma_{\mrm{tot}} = \int_0^1 \int_0^1 \frac{D(z_1)}{z_1} \,
\frac{D(z_2)}{z_2} \, z_1 z_2 \sigma(z_1 z_2 s) \, \d z_1 \, \d z_2 ~.
\end{equation}
The expression $z_1 z_2 \sigma(s')$ is now essentially flat in $s'$,
i.e. not only can $\sigma_{\mathrm{max}}$ be found at a convenient energy
such as the maximum one, but additionally the {\tsc{Pythia}} generation
efficiency (the likelihood of surviving step 4) is greatly
enhanced. The price to be paid is that $z$ has to be selected
according to $D(z)/z$ rather than according to $D(z)$. Note that
$D(z)/z$ is not normalized to unity. One therefore needs to define
\begin{equation}
{\cal I}_D = \int_0^1 \frac{D(z)}{z} \, \d z ~,
\end{equation}
and a properly normalized
\begin{equation}
D'(z) = \frac{1}{{\cal I}_D} \, \frac{D(z)}{z} ~.
\end{equation}
Then
\begin{equation}
\sigma_{\mrm{tot}} = \int_0^1 \int_0^1 D'(z_1) \, D'(z_2) \,
{\cal I}_D^2 \, z_1 z_2 \sigma(z_1 z_2 s) \, \d z_1 \, \d z_2 ~.
\end{equation}
Therefore the proper event weight is ${\cal I}_D^2 \, z_1 z_2$.
This weight should be stored, for each event, in \ttt{PARP(173)}.
The maximum weight that will be encountered should be stored in
\ttt{PARP(174)} before the \ttt{PYINIT} call, and not changed
afterwards. It is not necessary to know the precise maximum;
any value larger than the true maximum will do, but the inefficiency
will be larger the cruder the approximation.
Additionally you must put \ttt{MSTP(173)=1} for the program to
make use of weights at all. Often $D(z)$ are not known analytically;
therefore ${\cal I}_D$ is also not known beforehand, but may have to
be evaluated (by you) during the course of the run. Then you should
just use the weight $z_1 z_2$ in \ttt{PARP(173)} and do the overall
normalization yourself in the end. Since \ttt{PARP(174)=1.} by
default, in this case you need not set this variable specially.
Only the cross sections are affected by the procedure selected for
overall normalization, the events themselves still are properly
distributed in $s'$ and internal phase space.
Above it has been assumed tacitly that $D(z) \to 0$ for $z \to 0$.
If not, $D(z)/z$ is divergent, and it is not possible to define a
properly normalized $D'(z) = D(z)/z$. If the cross section is truly
diverging like $1/s$, then a $D(z)$ which is nonvanishing for
$z \to 0$ does imply an infinite total cross section, whichever way
things are considered. In cases like that, it is necessary to impose
a lower cut on $z$, based on some physics or detector consideration.
Some such cut is anyway needed to keep away from the minimum c.m.
energy required for {\tsc{Pythia}} events, see above.
The most difficult cases are those with a very narrow and high
peak, such as the $\mathrm{Z}^0$. One could initialize at the energy of
maximum cross section and use $D(z)$ as is, but efficiency might
turn out to be very low. One might then be tempted to do more
complicated transforms of the kind illustrated above. As a rule
it is then convenient to work in the variables $\tau_z = z_1 z_2$
and $y_z = (1/2) \ln (z_1/z_2)$, cf. subsection \ref{ss:kinemtwo}.
Clearly, the better the behaviour of the cross section can be
modelled in the choice of $z_1$ and $z_2$, the better the overall
event generation efficieny. Even under the best of circumstances,
the efficiency will still be lower than for runs with fix energy.
There is also a non-negligible time overhead for using variable
energies in the first place, from kinematics reconstruction
and (in part) from the phase space selection. One should therefore
not use variable energies when not needed, and not use a large
range of energies $\sqrt{s'}$ if in the end only a smaller range
is of experimental interest.
This facility may be combined with most other aspects of the program.
For instance, it is possible to simulate beamstrahlung as above and
still include bremsstrahlung with \ttt{MSTP(11)=1}. Further, one may
multiply the overall event weight of \ttt{PARP(173)} with a
kinematics-dependent weight given by \ttt{PYEVWT}, although it is not
recommended (since the chances of making a mistake are also
multiplied). However, a few things do \textit{not} work.
\begin{Itemize}
\item It is not possible to use pile-up events, i.e. you must have
\ttt{MSTP(131)=0}.
\item The possibility of giving in your own cross-section optimization
coefficients, option \ttt{MSTP(121)=2}, would require more input
than with fixed energies, and this option should therefore not be
used. You can still use \ttt{MSTP(121)=1}, however.
\item The multiple interactions scenario with \ttt{MSTP(82)}$\geq 2$
only works approximately for energies different from the initialization
one. If the c.m. energy spread is smaller than a factor 2, say, the
approximation should be reasonable, but if the spread is larger
one may have to subdivide into subruns of different energy bins.
The initialization should be made at the largest energy to be
encountered --- whenever multiple interactions are possible (i.e. for
incoming hadrons and resolved photons) this is where the cross sections
are largest anyway, and so this is no further constraint. There is no
simple possibility to change \ttt{PARP(82)} during the course of the
run, i.e. an energy-independent $p_{\perp 0}$ must be assumed. The
default option \ttt{MSTP(82)=1} works fine, i.e. does not suffer
from the constraints above. If so desired,
$p_{\perp\mathrm{min}}=$\ttt{PARP(81)} can be set differently for each event,
as a function of c.m. energy. Initialization should then be done with
\ttt{PARP(81)} as low as it is ever supposed to become.
\end{Itemize}
\subsection{Other Routines and Common Blocks}
The subroutines and common blocks that you will come in direct
contact with have already been described. A number of other routines
and common blocks exist, and those not described elsewhere are here
briefly listed for the sake of completeness. The \ttt{PYG***}
routines are slightly modified versions of the \ttt{SAS***} ones
of the \tsc{SaSgam} library. The common block \ttt{SASCOM} is
renamed \ttt{PYINT8}. If you want to use the parton distributions
for standalone purposes, you are encouraged to use the original
\tsc{SaSgam} routines rather than going the way via the
{\tsc{Pythia}} adaptations.
\begin{entry}
\iteme{SUBROUTINE PYINRE :}\label{p:PYINRE}
to initialize the widths and effective widths of resonances.
\iteme{SUBROUTINE PYINBM(CHFRAM,CHBEAM,CHTARG,WIN) :}\label{p:PYINBM}
to read in and identify the beam (\ttt{CHBEAM}) and target
(\ttt{CTTARG}) particles and the frame (\ttt{CHFRAM}) as given in
the \ttt{PYINIT} call; also to save the original energy (\ttt{WIN}).
\iteme{SUBROUTINE PYINKI(MODKI) :}\label{p:PYINKI}
to set up the event kinematics, either at initialization
(\ttt{MODKI=0}) or for each separate event, the latter when the
program is run with varying kinematics (\ttt{MODKI=1}).
\iteme{SUBROUTINE PYINPR :}\label{p:PYINPR}
to set up the partonic subprocesses selected with \ttt{MSEL}.
For $\gamma\mathrm{p}$ and $\gamma\gamma$, also the \ttt{MSTP(14)} value
affects the choice of processes. In particular, the option
\ttt{MSTP(14)=10} sets up the three or six different processes that
need to be mixed, with separate cuts for each.
\iteme{SUBROUTINE PYXTOT :}\label{p:PYXTOT}
to give the parametrized total, double diffractive, single
diffractive and elastic cross sections for different energies and
colliding hadrons.
\iteme{SUBROUTINE PYMAXI :}\label{p:PYMAXI}
to find optimal coefficients \ttt{COEF} for the selection of
kinematical variables, and to find the related maxima for
the differential cross section times Jacobian factors,
for each of the subprocesses included.
\iteme{SUBROUTINE PYPILE(MPILE) :}\label{p:PYPILE}
to determine the number of pile-up events, i.e. events
appearing in the same beam--beam crossing.
\iteme{SUBROUTINE PYSAVE(ISAVE,IGA) :}\label{p:PYSAVE}
saves and restores parameters and cross section values between the
three $\gamma\mathrm{p}$ and the six $\gamma\gamma$ components
of \ttt{MSTP(14)=10}. The options for \ttt{ISAVE} are (1)
a complete save of all parameters specific to a given component,
(2) a partial save of cross-section information, (3) a restoration
of all parameters specific to a given component, (4) as 3 but
preceded by a random selection of component, and (5) a summation of
component cross sections (for \ttt{PYSTAT}). The subprocess code
in \ttt{IGA} is the one described for \ttt{MSTI(9)}; it is input
for options 1, 2 and 3 above, output for 4 and dummy for 5.
\iteme{SUBROUTINE PYRAND :}\label{p:PYRAND}
to generate the quantities characterizing a hard scattering
on the parton level, according to the relevant matrix elements.
\iteme{SUBROUTINE PYSCAT :}\label{p:PYSCAT}
to find outgoing flavours and to set up the kinematics and
colour flow of the hard scattering.
\iteme{SUBROUTINE PYRESD :}\label{p:PYRESD}
to allow resonances to decay,
including chains of successive decays and parton showers.
\iteme{SUBROUTINE PYMULT(MMUL) :}\label{p:PYMULT}
to generate semi-hard interactions according to the
multiple interaction formalism.
\iteme{SUBROUTINE PYREMN(IPU1,IPU2) :}\label{p:PYREMN}
to add on target remnants and include primordial $k_{\perp}$.
\iteme{SUBROUTINE PYDIFF :}\label{p:PYDIFF}
to handle diffractive and elastic scattering events.
\iteme{SUBROUTINE PYDOCU :}\label{p:PYDOCU}
to compute cross sections of processes, based on current
Monte Carlo statistics, and to store event information in the
\ttt{MSTI} and \ttt{PARI} arrays.
\iteme{SUBROUTINE PYWIDT(KFLR,SH,WDTP,WDTE) :}\label{p:PYWIDT}
to calculate widths and effective widths of resonances.
\iteme{SUBROUTINE PYOFSH(MOFSH,KFMO,KFD1,KFD2,PMMO,RET1,RET2) :}%
\label{p:PYOFSH}
to calculate partial widths into channels off the mass shell,
and to select correlated masses of resonance pairs.
\iteme{SUBROUTINE PYKLIM(ILIM) :}\label{p:PYKLIM}
to calculate allowed kinematical limits.
\iteme{SUBROUTINE PYKMAP(IVAR,MVAR,VVAR) :}\label{p:PYKMAP}
to calculate the value of a kinematical variable when this
is selected according to one of the simple pieces.
\iteme{SUBROUTINE PYSIGH(NCHN,SIGS) :}\label{p:PYSIGH}
to give the differential cross section (multiplied by the
relevant Jacobians) for a given subprocess and kinematical
setup.
\iteme{SUBROUTINE PYSTFL(KF,X,Q2,XPQ) :}\label{p:PYSTFL}
to give parton distributions for $\mathrm{p}$ and $\mathrm{n}$ in the option
with modified behaviour at small $Q^2$ and $x$, see
\ttt{MSTP(57)}.
\iteme{SUBROUTINE PYSTFU(KF,X,Q2,XPQ) :}\label{p:PYSTFU}
to give parton-distribution functions (multiplied by $x$, i.e.
$x f_i(x,Q^2)$) for an arbitrary particle (of those recognized by
{\tsc{Pythia}}). Generic driver routine for the following, specialized ones.
\begin{subentry}
\iteme{KF :} flavour of probed particle, according to KF code.
\iteme{X :} $x$ value at which to evaluate parton distributions.
\iteme{Q2 :} $Q^2$ scale at which to evaluate parton distributions.
\iteme{XPQ :} array of dimensions \ttt{XPQ(-25:25)}, which contains
the evaluated parton distributions $x f_i(x,Q^2)$. Components $i$
ordered according to standard KF code; additionally the gluon is
found in position 0 as well as 21 (for historical reasons).
\end{subentry}
\iteme{SUBROUTINE PYSTEL(X,Q2,XPEL) :}\label{p:PYSTEL}
to give electron parton distributions.
\iteme{SUBROUTINE PYSTGA(X,Q2,XPGA) :}\label{p:PYSTGA}
to give the photon parton distributions for sets other than the SaS
ones.
\iteme{SUBROUTINE PYGGAM(ISET,X,Q2,P2,F2GM,XPDFGM) :}\label{p:PYGGAM}
to construct the SaS $F_2$ and parton distributions of the photon
by summing homogeneous (VMD) and inhomogeneous (anomalous) terms.
For $F_2$, $\c$ and $\b$ are included by the Bethe-Heitler formula;
in the `{$\overline{\mrm{MS}}$}' scheme additionally a $C^{\gamma}$ term
is added. Calls \ttt{PYGVMD}, \ttt{PYGANO}, \ttt{PYGBEH},
and \ttt{PYGDIR}.
\iteme{SUBROUTINE PYGVMD(ISET,KF,X,Q2,P2,ALAM,XPGA) :}\label{p:PYGVMD}
to evaluate the VMD parton distributions of a photon,
evolved homogeneously from an initial scale $P^2$ to $Q^2$.
\iteme{SUBROUTINE PYGANO(KF,X,Q2,P2,ALAM,XPGA) :}\label{p:PYGANO}
to evaluate the parton distributions of the anomalous
photon, inhomogeneously evolved from a scale $P^2$ (where it vanishes)
to $Q^2$.
\iteme{SUBROUTINE PYGBEH(KF,X,Q2,P2,PM2,XPBH) :}\label{p:PYGBEH}
to evaluate the Bethe-Heitler cross section for
heavy flavour production.
\iteme{SUBROUTINE PYGDIR(X,Q2,P2,AK0,XPGA) :}\label{p:PYGDIR}
to evaluate the direct contribution, i.e. the $C^{\gamma}$ term,
as needed in {$\overline{\mrm{MS}}$} parametrizations.
\iteme{SUBROUTINE PYSTPI(X,Q2,XPPI) :}\label{p:PYSTPI}
to give pion parton distributions.
\iteme{SUBROUTINE PYSTPR(X,Q2,XPPR) :}\label{p:PYSTPR}
to give proton parton distributions.
\iteme{FUNCTION PYCTQ2(ISET,IPRT,X,Q) :}\label{p:PYCTQ2}
to give the CTEQ2 proton parton distributions.
\iteme{FUNCTION PYHFTH(SH,SQM,FRATT) :}\label{p:PYHFTH}
to give heavy-flavour threshold factor in matrix elements.
\iteme{SUBROUTINE PYSPLI(KF,KFLIN,KFLCH,KFLSP) :}\label{p:PYSPLI}
to give hadron remnant or remnants left behind when the reacting
parton is kicked out.
\iteme{FUNCTION PYGAMM(X) :}\label{p:PYGAMM}
to give the value of the ordinary $\Gamma (x)$ function
(used in some parton-distribution parametrizations).
\iteme{SUBROUTINE PYWAUX(IAUX,EPS,WRE,WIM) :}\label{p:PYWAUX}
to evaluate the two auxiliary functions $W_1$ and $W_2$ appearing
in the cross section expressions in \ttt{PYSIGH}.
\iteme{SUBROUTINE PYI3AU(EPS,RAT,Y3RE,Y3IM) :}\label{p:PYI3AU}
to evaluate the auxiliary function $I_3$ appearing
in the cross section expressions in \ttt{PYSIGH}.
\iteme{FUNCTION PYSPEN(XREIN,XIMIN,IREIM) :}\label{p:PYSPEN}
to calculate the real and imaginary part of the Spence
function.
\iteme{SUBROUTINE PYQQBH(WTQQBH) :}\label{p:PYQQBH}
to calculate matrix elements for the two processes
$\mathrm{g} \mathrm{g} \to \mathrm{Q} \overline{\mathrm{Q}} \H^0$ and $\mathrm{q} \overline{\mathrm{q}} \to \mathrm{Q} \overline{\mathrm{Q}} \H^0$.
\iteme{BLOCK DATA PYDATA :}\label{p:PYDATA}
to give sensible default values to all status codes and
parameters.
\end{entry}
\drawbox{COMMON/PYINT1/MINT(400),VINT(400)}\label{p:PYINT1}
\begin{entry}
\itemc{Purpose:} to collect a host of integer- and real-valued
variables used internally in the program during the initialization
and/or event generation stage. These variables must not be changed
by you.
\iteme{MINT(1) :}\label{p:MINT} specifies the general type of
subprocess that has occurred, according to the ISUB code given in
section \ref{ss:ISUBcode}.
\iteme{MINT(2) :} whenever \ttt{MINT(1)} (together with \ttt{MINT(15)}
and \ttt{MINT(16)}) are not sufficient to specify the type of process
uniquely, \ttt{MINT(2)} provides an ordering of the different
possibilities, see \ttt{MSTI(2)}.
\iteme{MINT(3) :} number of partons produced in the hard
interactions, i.e. the number $n$ of the $2 \to n$ matrix elements
used; is sometimes 3 or 4 when a basic $2 \to 1$ or $2 \to 2$ process
has been folded with two $1 \to 2$ initial branchings
(like $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \mathrm{q}''' \H^0$).
\iteme{MINT(4) :} number of documentation lines at the beginning of
the common block \ttt{LUJETS} that are given with \ttt{K(I,1)=21};
0 for \ttt{MSTP(125)=0}.
\iteme{MINT(5) :} number of events generated to date in current run.
In runs with the variable-energy option, \ttt{MSTP(171)=1}
and \ttt{MSTP(172)=2}, only those events that survive (i.e. that
do not have \ttt{MSTI(61)=1}) are counted in this number. That
is, \ttt{MINT(5)} may be less than the total number of \ttt{PYEVNT}
calls.
\iteme{MINT(6) :} current frame of event (see \ttt{MSTP(124)} for
possible values).
\iteme{MINT(7), MINT(8) :} line number for documentation of outgoing
partons/particles from hard scattering for $2 \to 2$ or
$2 \to 1 \to 2$ processes (else = 0).
\iteme{MINT(10) :} is 1 if cross section maximum was violated in
current event, and 0 if not.
\iteme{MINT(11) :} KF flavour code for beam (side 1) particle.
\iteme{MINT(12) :} KF flavour code for target (side 2) particle.
\iteme{MINT(13), MINT(14) :} KF flavour codes for side 1 and side 2
initial-state shower initiators.
\iteme{MINT(15), MINT(16) :} KF flavour codes for side 1 and side 2
incoming partons to the hard interaction.
\iteme{MINT(17), MINT(18) :} flag to signal if particle on side 1 or
side 2 has been scattered diffractively; 0 if no, 1 if yes.
\iteme{MINT(19), MINT(20) :} flag to signal initial-state structure
with parton inside photon inside electron on side 1 or side 2;
0 if no, 1 if yes.
\iteme{MINT(21) - MINT(24) :} KF flavour codes for outgoing partons
from the hard interaction. The number of positions actually used is
process-dependent, see \ttt{MINT(3)}; trailing positions not used are
set = 0.
\iteme{MINT(25), MINT(26) :} KF flavour codes of the products in the
decay of a single $s$-channel resonance formed in the hard interaction.
Are thus only used when \ttt{MINT(3)=1} and the resonance is allowed
to decay.
\iteme{MINT(31) :} number of hard or semi-hard scatterings that occurred
in the current event in the multiple-interaction scenario; is = 0 for a
low-$p_{\perp}$ event.
\iteme{MINT(35) :} in a true $2 \to 3$ process, where one particle is
a resonance with decay channel selected already before the
\ttt{PYRESD} call, the decay channel number (in the \ttt{/LUDAT3/}
numbering) is stored here.
\iteme{MINT(41), MINT(42) :} type of incoming beam or target particle;
1 for lepton and 2 for hadron. A photon counts as a lepton if it
is not resolved (\ttt{MSTP(14)=0}) and as a hadron if it is resolved
(\ttt{MSTP(14)}$\geq 1$).
\iteme{MINT(43) :} combination of incoming beam and target particles.
A photon counts as a hadron.
\begin{subentry}
\iteme{= 1 :} lepton on lepton.
\iteme{= 2 :} lepton on hadron.
\iteme{= 3 :} hadron on lepton.
\iteme{= 4 :} hadron on hadron.
\end{subentry}
\iteme{MINT(44) :} as \ttt{MINT(43)}, but a photon counts as a lepton.
\iteme{MINT(45), MINT(46) :} structure of incoming beam and target
particles.
\begin{subentry}
\iteme{= 1 :} no internal structure, i.e. an electron or photon
carrying the full beam energy.
\iteme{= 2 :} defined with parton distributions that are not peaked
at $x = 1$, i.e. a hadron or a resolved photon.
\iteme{= 3 :} defined with parton distributions that are peaked at
$x = 1$, i.e. a resolved electron.
\end{subentry}
\iteme{MINT(47) :} combination of incoming beam- and target-particle
parton-distribution function types.
\begin{subentry}
\iteme{= 1 :} no parton distribution either for beam or target.
\iteme{= 2 :} parton distributions for target but not for beam.
\iteme{= 3 :} parton distributions for beam but not for target.
\iteme{= 4 :} parton distributions for both beam and target, but not
both peaked at $x = 1$.
\iteme{= 5 :} parton distributions for both beam and target, with both
peaked at $x = 1$.
\end{subentry}
\iteme{MINT(48) :} total number of subprocesses switched on.
\iteme{MINT(49) :} number of subprocesses that are switched on, apart
from elastic scattering and single, double and central diffractive.
\iteme{MINT(50) :} combination of incoming particles from a multiple
interactions point of view.
\begin{subentry}
\iteme{= 0 :} the total cross section is not known; therefore no
multiple interactions are possible.
\iteme{= 1 :} the total cross section is known; therefore multiple
interactions are possible if switched on.
\end{subentry}
\iteme{MINT(51) :} internal flag that event failed cuts.
\begin{subentry}
\iteme{= 0 :} no problem.
\iteme{= 1 :} event failed; new one to be generated.
\iteme{= 2 :} event failed; no new event is to be generated but
instead control is to be given back to used. Is intended for
user-defined processes, when \ttt{NUP=0}.
\end{subentry}
\iteme{MINT(52) :} internal counter for number of lines used
(in \ttt{/LUJETS/}) before multiple interactions are considered.
\iteme{MINT(53) :} internal counter for number of lines used
(in \ttt{/LUJETS/}) before beam remnants are considered.
\iteme{MINT(55) :} the heaviest new flavour switched on for QCD
processes, specifically the flavour to be generated for ISUB = 81,
82, 83 or 84.
\iteme{MINT(56) :} the heaviest new flavour switched on for QED
processes, specifically for ISUB = 85. Note that, unlike
\ttt{MINT(55)}, the heaviest flavour may here be a lepton, and
that heavy means the one with largest KF code.
\iteme{MINT(57) :} number of times the beam remnant treatment has
failed, and the same basic kinematical setup is used to produce
a new parton shower evolution and beam remnant set. Mainly used
in leptoproduction, for the option when $x$ and $Q^2$ are to be
preserved.
\iteme{MINT(61) :} internal switch for the mode of operation of
resonance width calculations in \ttt{PYWIDT} for $\gamma^* / \Z^0$ or
$\gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0$.
\begin{subentry}
\iteme{= 0 :} without reference to initial-state flavours.
\iteme{= 1 :} with reference to given initial-state flavours.
\iteme{= 2 :} for given final-state flavours.
\end{subentry}
\iteme{MINT(62) :} internal switch for use at initialization of
$\H^0$ width.
\begin{subentry}
\iteme{= 0 :} use widths into $\mathrm{Z} \mathrm{Z}^*$ or $\mathrm{W} \mathrm{W}^*$ calculated
before.
\iteme{= 1 :} evaluate widths into $\mathrm{Z} \mathrm{Z}^*$ or $\mathrm{W} \mathrm{W}^*$ for
current Higgs mass.
\end{subentry}
\iteme{MINT(65) :} internal switch to indicate initialization
without specified reaction.
\begin{subentry}
\iteme{= 0 :} normal initialization.
\iteme{= 1 :} initialization with argument \ttt{'none'} in
\ttt{PYINIT} call.
\end{subentry}
\iteme{MINT(71) :} switch to tell whether current process is singular
for $p_{\perp} \to 0$ or not.
\begin{subentry}
\iteme{= 0 :} non-singular process, i.e. proceeding via an
$s$-channel resonance or with both products having a mass above
\ttt{CKIN(6)}.
\iteme{= 1 :} singular process.
\end{subentry}
\iteme{MINT(72) :} number of $s$-channel resonances that may
contribute to the cross section.
\iteme{MINT(73) :} KF code of first $s$-channel resonance;
0 if there is none.
\iteme{MINT(74) :} KF code of second $s$-channel resonance;
0 if there is none.
\iteme{MINT(81) :} number of selected pile-up events.
\iteme{MINT(82) :} sequence number of currently considered pile-up
event.
\iteme{MINT(83) :} number of lines in the event record already
filled by previously considered pile-up events.
\iteme{MINT(84) :} \ttt{MINT(83) + MSTP(126)}, i.e. number of lines
already filled by previously considered events plus number of lines
to be kept free for event documentation.
\iteme{MINT(91) :} is 1 for a lepton--hadron event and 0
else. Used to determine whether a \ttt{PYFRAM(3)} call is possible.
\iteme{MINT(92) :} is used to denote region in $(x,Q^2)$ plane when
\ttt{MSTP(57)=2}, according to numbering in \cite{Sch93a}. Simply put,
0 means that the modified proton parton distributions were not used,
1 large $x$ and $Q^2$, 2 small $Q^2$ but large $x$,
3 small $x$ but large $Q^2$ and 4 small $x$ and $Q^2$.
\iteme{MINT(93) :} is used to keep track of parton distribution set
used in the latest \ttt{STRUCTM} call to \tsc{Pdflib}. The code for this
set is stored in the form
\ttt{MINT(93) = 1000000}$\times$\ttt{NPTYPE + 1000}$\times$%
\ttt{NGROUP + NSET}.
The stored previous value is compared with the current new value
to decide whether a \ttt{PDFSET} call is needed to switch to another
set.
\iteme{MINT(101), MINT(102) :} is normally 1, but is 4 when a resolved
photon (appearing on side 1 or 2) can be represented by either of the
four vector mesons $\rho^0$, $\omega$, $\phi$ and $\mathrm{J}/\psi$.
\iteme{MINT(103), MINT(104) :} KF flavour code for the two incoming
particles, i.e. the same as \ttt{MINT(11)} and \ttt{MINT(12)}. The
exception is when a resolved photon is represented by a vector meson
(a $\rho^0$, $\omega$, $\phi$ or $\mathrm{J}/\psi$). Then the code of the vector
meson is given.
\iteme{MINT(105) :} is either \ttt{MINT(103)} or \ttt{MINT(104)},
depending on which side of the event currently is being studied.
\iteme{MINT(107), MINT(108) :} if either or both of the two incoming
particles is a photon, then the respective value gives the nature
assumed for that photon. The code follows the one used for
\ttt{MSTP(14)}:
\begin{subentry}
\iteme{= 0 :} direct photon.
\iteme{= 1 :} resolved photon.
\iteme{= 2 :} VMD-like photon.
\iteme{= 3 :} anomalous photon.
\end{subentry}
\iteme{MINT(109) :} is either \ttt{MINT(107)} or \ttt{MINT(108)},
depending on which side of the event currently is being studied.
\iteme{MINT(111) :} the frame given in \ttt{PYINIT} call, 0--5 for
\ttt{'NONE'}, \ttt{'CMS'}, \ttt{'FIXT'}, \ttt{'USER'}, \ttt{'FOUR'}
and \ttt{'FIVE'}, respectively.
\iteme{MINT(121) :} number of separate event classes to initialize
and mix.
\begin{subentry}
\iteme{= 1 :} the normal value.
\iteme{= 3 :} for a $\gamma\mathrm{p}$ interaction when \ttt{MSTP(14)=10}.
\iteme{= 6 :} for a $\gamma\gamma$ interaction when
\ttt{MSTP(14)=10}.
\end{subentry}
\iteme{MINT(122) :} event class used in current event for $\gamma\mathrm{p}$ or
$\gamma\gamma$ events generated with the \ttt{MSTP(14)=10} option;
code as described for \ttt{MSTI(9)}.
\iteme{MINT(123) :} event class used in the current event, with the
same list of possibilities as for \ttt{MSTP(14)}, except that
options 1, 4 or 10 do not appear. Apart from a different coding, this
is exactly the same information as is available in \ttt{MINT(122)}.
\boxsep
\iteme{VINT(1) :}\label{p:VINT} $E_{\mrm{cm}}$, c.m. energy.
\iteme{VINT(2) :} $s$ ($=E_{\mrm{cm}}^2$) squared mass of
complete system.
\iteme{VINT(3) :} mass of beam particle.
\iteme{VINT(4) :} mass of target particle.
\iteme{VINT(5) :} momentum of beam (and target) particle in c.m.
frame.
\iteme{VINT(6) - VINT(10) :} $\theta$, $\varphi$ and
\mbox{{\boldmath $\beta$}} for rotation and boost
from c.m. frame to user-specified frame.
\iteme{VINT(11) :} $\tau_{\mathrm{min}}$.
\iteme{VINT(12) :} $y_{\mathrm{min}}$.
\iteme{VINT(13) :} $\cos\hat{\theta}_{\mathrm{min}}$ for
$\cos\hat{\theta} \leq 0$.
\iteme{VINT(14) :} $\cos\hat{\theta}_{\mathrm{min}}$ for
$\cos\hat{\theta} \geq 0$.
\iteme{VINT(15) :} $x^2_{\perp \mathrm{min}}$.
\iteme{VINT(16) :} $\tau'_{\mathrm{min}}$.
\iteme{VINT(21) :} $\tau$.
\iteme{VINT(22) :} $y$.
\iteme{VINT(23) :} $\cos\hat{\theta}$.
\iteme{VINT(24) :} $\varphi$ (azimuthal angle).
\iteme{VINT(25) :} $x_{\perp}^2$.
\iteme{VINT(26) :} $\tau'$.
\iteme{VINT(31) :} $\tau_{\mathrm{max}}$.
\iteme{VINT(32) :} $y_{\mathrm{max}}$.
\iteme{VINT(33) :} $\cos\hat{\theta}_{\mathrm{max}}$ for
$\cos\hat{\theta} \leq 0$.
\iteme{VINT(34) :} $\cos\hat{\theta}_{\mathrm{max}}$ for
$\cos\hat{\theta} \geq 0$.
\iteme{VINT(35) :} $x^2_{\perp \mathrm{max}}$.
\iteme{VINT(36) :} $\tau'_{\mathrm{max}}$.
\iteme{VINT(41), VINT(42) :} the momentum fractions $x$ taken by the
partons at the hard interaction, as used e.g. in the
parton-distribution functions.
\iteme{VINT(43) :} $\hat{m} = \sqrt{\hat{s}}$, mass of
hard-scattering subsystem.
\iteme{VINT(44) :} $\hat{s}$ of the hard subprocess ($2 \to 2$ or
$2 \to 1$).
\iteme{VINT(45) :} $\hat{t}$ of the hard subprocess ($2 \to 2$ or
$2 \to 1 \to 2$).
\iteme{VINT(46) :} $\hat{u}$ of the hard subprocess ($2 \to 2$ or
$2 \to 1 \to 2$).
\iteme{VINT(47) :} $\hat{p}_{\perp}$ of the hard subprocess
($2 \to 2$ or $2 \to 1 \to 2$), i.e. transverse momentum evaluated
in the rest frame of the scattering.
\iteme{VINT(48) :} $\hat{p}_{\perp}^2$ of the hard subprocess;
see \ttt{VINT(47)}.
\iteme{VINT(49) :} $\hat{m}'$, the mass of the complete three- or
four-body final state in $2 \to 3$ or $2 \to 4$ processes.
\iteme{VINT(50) :} $\hat{s}' = \hat{m}'^2$; see \ttt{VINT(49)}.
\iteme{VINT(51) :} $Q$ of the hard subprocess. The exact definition is
process-dependent, see \ttt{MSTP(32)}.
\iteme{VINT(52) :} $Q^2$ of the hard subprocess; see \ttt{VINT(51)}.
\iteme{VINT(53) :} $Q$ of the outer hard-scattering subprocess.
Agrees with \ttt{VINT(51)} for a $2 \to 1$ or $2 \to 2$ process.
For a $2 \to 3$ or $2 \to 4$ $\mathrm{W}/\mathrm{Z}$ fusion process, it is set by
the $\mathrm{W}/\mathrm{Z}$ mass scale, and for
subprocesses 121 and 122 by the heavy-quark mass.
\iteme{VINT(54) :} $Q^2$ of the outer hard-scattering subprocess;
see \ttt{VINT(53)}.
\iteme{VINT(55) :} $Q$ scale used as maximum virtuality in parton
showers. Is equal to \ttt{VINT(53)}, except for
deep-inelastic-scattering processes when \ttt{MSTP(22)}$ > 0$.
\iteme{VINT(56) :} $Q^2$ scale in parton showers; see \ttt{VINT(55)}.
\iteme{VINT(57) :} $\alpha_{\mathrm{em}}$ value of hard process.
\iteme{VINT(58) :} $\alpha_{\mathrm{s}}$ value of hard process.
\iteme{VINT(59) :} $\sin\hat{\theta}$ (cf. \ttt{VINT(23)}); used for
improved numerical precision in elastic and diffractive scattering.
\iteme{VINT(61), VINT(62) :} nominal $m^2$ values, i.e. without
initial-state radiation effects, for the two partons entering
the hard interaction.
\iteme{VINT(63), VINT(64) :} nominal $m^2$ values, i.e. without
final-state radiation effects, for the two (or one) partons/particles
leaving the hard interaction.
\iteme{VINT(65) :} $\hat{p}_{\mrm{init}}$, i.e. common nominal absolute
momentum of the two partons entering the hard interaction, in their
rest frame.
\iteme{VINT(66) :} $\hat{p}_{\mrm{fin}}$, i.e. common nominal absolute
momentum of the two partons leaving the hard interaction, in their
rest frame.
\iteme{VINT(67), VINT(68) :} mass of beam and target particle, as
\ttt{VINT(3)} and \ttt{VINT(4)}, except that an incoming $\gamma$ is
assigned the $\rho^0$, $\omega$ or $\phi$ mass. Used for elastic
scattering $\gamma \mathrm{p} \to \rho^0 \mathrm{p}$ and other similar processes.
\iteme{VINT(71) :} $p_{\perp\mathrm{min}}$ of process, i.e. \ttt{CKIN(3)} or
\ttt{CKIN(5)}, depending on which is larger, and whether the process
is singular in $p_{\perp} \to 0$ or not.
\iteme{VINT(73) :} $\tau = m^2/s$ value of first resonance, if any;
see \ttt{MINT(73)}.
\iteme{VINT(74) :} $m \Gamma/s$ value of first resonance, if any;
see \ttt{MINT(73)}.
\iteme{VINT(75) :} $\tau = m^2/s$ value of second resonance, if any;
see \ttt{MINT(74)}.
\iteme{VINT(76) :} $m \Gamma/s$ value of second resonance, if any;
see \ttt{MINT(74)}.
\iteme{VINT(80) :} correction factor (evaluated in \ttt{PYOFSH}) for
the cross section of resonances produced in $2 \to 2$ processes, if
only some mass range of the full Breit--Wigner shape is allowed by
user-set mass cuts (\ttt{CKIN(2)}, \ttt{CKIN(45) - CKIN(48)}).
\iteme{VINT(81) - VINT(84) :} the $\cos\theta$ and $\varphi$ variables
of a true $2 \to 3$ process, where one product is a resonance,
effectively giving $2 \to 4$. The first two are $\cos\theta$ and
$\varphi$ for the resonance decay, the other two ditto for the
effective system formed by the other two particles.
\iteme{VINT(85), VINT(86) :} transverse momenta in a true $2 \to 3$
process; one is stored in \ttt{VINT(47)} (that of the
$\mathrm{Z}^0$ in $\mathrm{g} \mathrm{g} \to \mathrm{Z}^0 \mathrm{Q} \overline{\mathrm{Q}}$), while the smaller of the other
two is stored in \ttt{VINT(85)} and the larger in \ttt{VINT(86)}.
\iteme{VINT(91), VINT(92) :} gives a dimensionless suppression factor,
to take into account reduction in cross section due to the allowed
channels for a $\mathrm{W}^+\mathrm{W}^+$ or $\mathrm{W}^-\mathrm{W}^-$ pair, respectively, in the
same sense as \ttt{WIDS(24,1)} gives it for a $\mathrm{W}^+\mathrm{W}^-$ pair.
\iteme{VINT(98) :} is sum of \ttt{VINT(100)} values for current run.
\iteme{VINT(99) :} is weight \ttt{WTXS} returned from \ttt{PYEVWT} call
when \ttt{MSTP(142)}$\geq 1$, otherwise is 1.
\iteme{VINT(100) :} is compensating weight \ttt{1./WTXS} that should be
associated with events when \ttt{MSTP(142)=1}, otherwise is 1.
\iteme{VINT(108) :} ratio of maximum differential cross section
observed to maximum differential cross section assumed for the
generation; cf. \ttt{MSTP(123)}.
\iteme{VINT(109) :} ratio of minimal (negative!) cross section
observed to maximum differential cross section assumed for the
generation; could only become negative if cross sections are
incorrectly included.
\iteme{VINT(111) - VINT(116) :} for \ttt{MINT(61)=1} gives
kinematical factors for the different pieces contributing to
$\gamma^* / \Z^0$ or $\gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0$ production, for \ttt{MINT(61)=2}
gives sum of final-state weights for the same; coefficients are
given in the order pure $\gamma^*$, $\gamma^*$--$\mathrm{Z}^0$ interference,
$\gamma^*$--$\mathrm{Z}'^0$ interference, pure $\mathrm{Z}^0$, $\mathrm{Z}^0$--$\mathrm{Z}'^0$
interference and pure $\mathrm{Z}'^0$.
\iteme{VINT(117) :} width of $\mathrm{Z}^0$; needed in
$\gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0$ production.
\iteme{VINT(131) :} total cross section (in mb) for subprocesses
allowed in the pile-up events scenario according to the
\ttt{MSTP(132)} value.
\iteme{VINT(132) :} $\br{n} = $\ttt{VINT(131)}$\times$\ttt{PARP(131)},
cf. \ttt{PARI(91)}.
\iteme{VINT(133) :} $\langle n \rangle = \sum_i i \, {\cal P}_i /
\sum_i {\cal P}_i$ as actually simulated, i.e. $1 \leq i \leq 200$
(or smaller), see \ttt{PARI(92)}.
\iteme{VINT(134) :} number related to probability to have event in
beam--beam crossing; is $\exp(-\br{n}) \sum_i \br{n}^i/i!$ for
\ttt{MSTP(133)=1} and $\exp(-\br{n}) \sum_i \br{n}^i/(i-1)!$
for \ttt{MSTP(133)=2}, cf. \ttt{PARI(93)}.
\iteme{VINT(138) :} size of the threshold factor (enhancement or
suppression) in the latest event with heavy-flavour production;
see \ttt{MSTP(35)}.
\iteme{VINT(141), VINT(142) :} $x$ values for the parton-shower
initiators of the hardest interaction; used to find what is left
for multiple interactions.
\iteme{VINT(143), VINT(144) :} $1 - \sum_i x_i$ for all scatterings;
used for rescaling each new $x$-value in the multiple-interaction
parton-distribution-function evaluation.
\iteme{VINT(145) :} estimate of total parton--parton cross section for
multiple interactions; used for \ttt{MSTP(82)}$\geq 2$.
\iteme{VINT(146) :} common correction factor $f_c$ in the
multiple-interaction probability; used for \ttt{MSTP(82)}$\geq 2$
(part of $e(b)$, see eq.~(\ref{mi:ebenh})).
\iteme{VINT(147) :} average hadronic matter overlap; used for
\ttt{MSTP(82)}$\geq 2$ (needed in evaluation of $e(b)$, see
eq.~(\ref{mi:ebenh})).
\iteme{VINT(148) :} enhancement factor for current event in the
multiple-interaction probability, defined as the actual overlap
divided by the average one; used for \ttt{MSTP(82)}$\geq 2$
(is $e(b)$ of eq.~(\ref{mi:ebenh})).
\iteme{VINT(149) :} $x_{\perp}^2$ cut-off or turn-off for multiple
interactions. For \ttt{MSTP(82)}$\leq 1$ it is $4 p_{\perp\mathrm{min}}^2/s$,
for \ttt{MSTP(82)}$\geq 2$ it is $4 p_{\perp 0}^2/s$.
\iteme{VINT(150) :} probability to keep the given event in the
multiple-interaction scenario, as given by the `Sudakov' form factor.
\iteme{VINT(151), VINT(152) :} sum of $x$ values for all the
multiple-interaction partons.
\iteme{VINT(153) :} current differential cross section value
obtained from \ttt{PYSIGH}; used in multiple interactions only.
\iteme{VINT(155), VINT(156) :} the $x$ value of a photon that branches
into quarks or gluons, i.e. $x$ at interface between initial-state QED
and QCD cascades.
\iteme{VINT(157), VINT(158) :} the primordial $k_{\perp}$ values
selected in the two beam remnants.
\iteme{VINT(159), VINT(160) :} the $\chi$ values selected for beam
remnants that are split into two objects, describing how the energy
is shared (see \ttt{MSTP(92)} and \ttt{MSTP(94)}); is 0 if no
splitting is needed.
\iteme{VINT(161) - VINT(200) :} sum of Cabibbo--Kobayashi--Maskawa
squared matrix elements that a given flavour is allowed to couple to.
Results are stored in format \ttt{VINT(180+KF)} for quark and lepton
flavours and antiflavours (which need not be the same; see
\ttt{MDME(IDC,2)}). For leptons, these factors are normally unity.
\iteme{VINT(201) - VINT(220) :} additional variables needed in
phase-space selection for $2 \to 3$ processes with \ttt{ISET(ISUB)=5}.
Below indices 1, 2 and 3 refer to scattered partons 1, 2 and 3, except
that the $q$ four-momentum variables are $q_1 + q_2 \to q_1' q_2' q_3'$.
All kinematical variables refer to the internal kinematics of
the 3-body final state --- the kinematics of the system as a whole
is described by $\tau'$ and $y$, and the mass distribution of
particle 3 (a resonance) by $\tau$.
\begin{subentry}
\iteme{VINT(201) :} $m_1$.
\iteme{VINT(202) :} $p_{\perp 1}^2$.
\iteme{VINT(203) :} $\varphi_1$.
\iteme{VINT(204) :} $M_1$ (mass of propagator particle).
\iteme{VINT(205) :} weight for the $p_{\perp 1}^2$ choice.
\iteme{VINT(206) :} $m_2$.
\iteme{VINT(207) :} $p_{\perp 2}^2$.
\iteme{VINT(208) :} $\varphi_2$.
\iteme{VINT(209) :} $M_2$ (mass of propagator particle).
\iteme{VINT(210) :} weight for the $p_{\perp 2}^2$ choice.
\iteme{VINT(211) :} $y_3$.
\iteme{VINT(212) :} $y_{3 \mathrm{max}}$.
\iteme{VINT(213) :} $\epsilon = \pm 1$; choice between two mirror
solutions $1 \leftrightarrow 2$.
\iteme{VINT(214) :} weight associated to $\epsilon$-choice.
\iteme{VINT(215) :} $t_1 = (q_1 - q_1')^2$.
\iteme{VINT(216) :} $t_2 = (q_2 - q_2')^2$.
\iteme{VINT(217) :} $q_1 q_2'$ four-product.
\iteme{VINT(218) :} $q_2 q_1'$ four-product.
\iteme{VINT(219) :} $q_1' q_2'$ four-product.
\iteme{VINT(220) :} $\sqrt{(m_{\perp 12}^2 - m_{\perp 1}^2 -
m_{\perp 2}^2)^2 - 4 m_{\perp 1}^2 m_{\perp 2}^2}$, where
$m_{\perp 12}$ is the transverse mass of the $q'_1 q'_2$ system.
\end{subentry}
\iteme{VINT(221) - VINT(225) :} $\theta$, $\varphi$ and
\mbox{{\boldmath $\beta$}} for rotation and boost
from c.m. frame to hadronic c.m. frame of a lepton--hadron
event.
\iteme{VINT(231) :} $Q^2_{\mathrm{min}}$ scale for current
parton-distribution function set.
\iteme{VINT(281) :} for resolved photon events, it gives the
ratio between the total $\gamma X$ cross section and the total
$\pi^0 X$ cross section, where $X$ represents the target particle.
\iteme{VINT(283), VINT(284) :} virtuality scale at which an anomalous
photon on the beam or target side of the event is being resolved.
More precisely, it gives the $p_{\perp}^2$ of the $\gamma \to \mathrm{q}\overline{\mathrm{q}}$
vertex.
\iteme{VINT(285) :} the \ttt{CKIN(3)} value provided by the user
at initialization; subsequently \ttt{CKIN(3)} may be overwritten
(for \ttt{MSTP(14)=10}) but \ttt{VINT(285)} stays.
\iteme{VINT(289) :} squared c.m. energy found in \ttt{PYINIT} call.
\iteme{VINT(290) :} the \ttt{WIN} argument of a \ttt{PYINIT} call.
\iteme{VINT(291) - VINT(300) :} the two five-vectors of the two
incoming particles, as reconstructed in \ttt{PYINKI}.
These may vary from one event to the next.
\end{entry}
\drawbox{COMMON/PYINT2/ISET(200),KFPR(200,2),COEF(200,20),ICOL(40,4,2)}%
\label{p:PYINT2}
\begin{entry}
\itemc{Purpose:} to store information necessary for efficient
generation of the different subprocesses, specifically type of
generation scheme and coefficients of the Jacobian. Also to store
allowed colour-flow configurations. These variables must not be
changed by you.
\iteme{ISET(ISUB) :}\label{p:ISET} gives the type of
kinematical-variable selection scheme used for subprocess ISUB.
\begin{subentry}
\iteme{= 0 :} elastic, diffractive and low-$p_{\perp}$ processes.
\iteme{= 1 :} $2 \to 1$ processes (irrespective of subsequent decays).
\iteme{= 2 :} $2 \to 2$ processes (i.e. the bulk of processes).
\iteme{= 3 :} $2 \to 3$ processes
(like $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \mathrm{q}''' \H^0$).
\iteme{= 4 :} $2 \to 4$ processes
(like $\mathrm{q} \mathrm{q}' \to \mathrm{q}'' \mathrm{q}''' \mathrm{W}^+ \mathrm{W}^-$).
\iteme{= 5 :} `true' $2 \to 3$ processes, one method.
\iteme{= 6 :} `true' $2 \to 3$ processes, another method;
currently only $\mathrm{g} \mathrm{g} \to \mathrm{Z}^0 \mathrm{Q} \overline{\mathrm{Q}}$.
\iteme{= 9 :} $2 \to 2$ in multiple interactions ($p_{\perp}$ as kinematics
variable).
\iteme{= 11 :} a user-defined process.
\iteme{= -1 :} legitimate process which has not yet been implemented.
\iteme{= -2 :} ISUB is an undefined process code.
\end{subentry}
\iteme{KFPR(ISUB,J) :}\label{p:KFPR} give the KF flavour codes for the
products produced in subprocess ISUB. If there is only one product, the
\ttt{J=2} position is left blank. Also, quarks and leptons assumed
massless in the matrix elements are denoted by 0. The main
application is thus to identify resonances produced
($\mathrm{Z}^0$, $\mathrm{W}^{\pm}$, $\H^0$, etc.).
\iteme{COEF(ISUB,J) :}\label{p:COEF} factors used in the Jacobians in
order to speed
up the selection of kinematical variables. More precisely, the shape
of the cross section is given as the sum of terms with different
behaviour, where the integral over the allowed phase space is
unity for each term. \ttt{COEF} gives the relative strength of these
terms, normalized so that the sum of coefficients for each variable
used is unity. Note that which coefficients are indeed used is
process-dependent.
\begin{subentry}
\iteme{ISUB :} standard subprocess code.
\iteme{J = 1 :} $\tau$ selected according $1/\tau$.
\iteme{J = 2 :} $\tau$ selected according to $1/\tau^2$.
\iteme{J = 3 :} $\tau$ selected according to $1/(\tau(\tau+\tau_R))$,
where $\tau_R = m_R^2/s$ is $\tau$ value of resonance; only used for
resonance production.
\iteme{J = 4 :} $\tau$ selected according to Breit--Wigner of form
$1/((\tau-\tau_R)^2+\gamma_R^2)$, where $\tau_R = m_R^2/s$ is $\tau$
value of resonance and $\gamma_R = m_R \Gamma_R/s$ is its scaled mass
times width; only used for resonance production.
\iteme{J = 5 :} $\tau$ selected according to
$1/(\tau(\tau+\tau_{R'}))$, where $\tau_{R'} = m_{R'}^2/s$ is $\tau$
value of second resonance; only used for simultaneous production of
two resonances.
\iteme{J = 6 :} $\tau$ selected according to second Breit--Wigner of
form $1/((\tau-\tau_{R'})^2+\gamma_{R'}^2)$, where
$\tau_{R'} = m_{R'}^2/s$ is $\tau$ value of second resonance and
$\gamma_{R'} = m_{R'} \Gamma_{R'}/s$ is its scaled
mass times width; is used only for simultaneous production
of two resonances, like $\gamma^*/\mathrm{Z}^0/\mathrm{Z}'^0$.
\iteme{J = 7 :} $\tau$ selected according to $1/(1-\tau)$; only used
when both parton distributions are peaked at $x = 1$.
\iteme{J = 8 :} $y$ selected according to $y - y_{\mathrm{min}}$.
\iteme{J = 9 :} $y$ selected according to $y_{\mathrm{max}} - y$.
\iteme{J = 10 :} $y$ selected according to $1/\cosh(y)$.
\iteme{J = 11 :} $y$ selected according to $1/(1-\exp(y-y_{\mathrm{max}}))$;
only used when beam parton distribution is peaked close to $x = 1$.
\iteme{J = 12 :} $y$ selected according to $1/(1-\exp(y_{\mathrm{min}}-y))$;
only used when target parton distribution is peaked close to $x = 1$.
\iteme{J = 13 :} $z = \cos\hat{\theta}$ selected evenly between limits.
\iteme{J = 14 :} $z = \cos\hat{\theta}$ selected according to $1/(a-z)$,
where $a = 1 + 2 m_3^2 m_4^2/\hat{s}^2$, $m_3$ and $m_4$ being the
masses of the two final-state particles.
\iteme{J = 15 :} $z = \cos\hat{\theta}$ selected according to $1/(a+z)$,
with $a$ as above.
\iteme{J = 16 :} $z = \cos\hat{\theta}$ selected according to
$1/(a-z)^2$, with $a$ as above.
\iteme{J = 17 :} $z = \cos\hat{\theta}$ selected according to
$1/(a+z)^2$, with $a$ as above.
\iteme{J = 18 :} $\tau'$ selected according to $1/\tau'$.
\iteme{J = 19 :} $\tau'$ selected according to
$(1 - \tau/\tau')^3/\tau'^2$.
\iteme{J = 20 :} $\tau'$ selected according to $1/(1-\tau')$; only
used when both parton distributions are peaked close to $x = 1$.
\end{subentry}
\iteme{ICOL :}\label{p:ICOL} contains information on different
colour-flow topologies in hard $2 \to 2$ processes.
\end{entry}
\drawbox{COMMON/PYINT3/XSFX(2,-40:40),ISIG(1000,3),SIGH(1000)}%
\label{p:PYINT3}
\begin{entry}
\itemc{Purpose:} to store information on parton distributions,
subprocess cross sections and different final-state relative
weights. These variables must not be changed by you.
\iteme{XSFX :}\label{p:XSFX} current values of parton-distribution
functions $xf(x)$ on beam and target side.
\iteme{ISIG(ICHN,1) :}\label{p:ISIG} incoming parton/particle on the
beam side to
the hard interaction for allowed channel number \ttt{ICHN}. The number
of channels filled with relevant information is given by \ttt{NCHN},
one of the arguments returned in a \ttt{PYSIGH} call. Thus only
$1 \leq $\ttt{ICHN}$ \leq $\ttt{NCHN} is filled with relevant
information.
\iteme{ISIG(ICHN,2) :} incoming parton/particle on the target side
to the hard interaction for allowed channel number \ttt{ICHN}. See also
comment above.
\iteme{ISIG(ICHN,3) :} colour-flow type for allowed channel number
\ttt{ICHN}; see \ttt{MSTI(2)} list. See also above comment. For
`subprocess' 96 uniquely, \ttt{ISIG(ICHN,3)} is also used to
translate information on what is the correct subprocess number
(11, 12, 13, 28, 53 or 68); this is used for reassigning subprocess
96 to either of these.
\iteme{SIGH(ICHN) :}\label{p:SIGH} evaluated differential
cross section for allowed
channel number \ttt{ICHN}, i.e. matrix-element value times
parton distributions, for current kinematical setup (in addition,
Jacobian factors are included in the figures, as used to speed up
generation). See also comment for \ttt{ISIG(ICHN,1)}.
\end{entry}
\drawbox{COMMON/PYINT4/WIDP(21:40,0:40),WIDE(21:40,0:40),WIDS(21:40,3)}%
\label{p:PYINT4}
\begin{entry}
\itemc{Purpose:} to store partial and effective decay widths for the
different resonances. These variables must not be changed by you.
\iteme{WIDP(KF,J) :}\label{p:WIDP} gives partial decay widths of
resonances into different channels (in GeV), given that all physically
allowed final states are included.
\begin{subentry}
\iteme{KF :} standard KF code for resonance considered. When top is
treated like a resonance (see \ttt{MSTP(48)}) it is stored in position
26. When the fourth generation fermions $\mathrm{l}$, $\mathrm{h}$, $\chi^-$ and
$\nu_{\chi}$ are treated like resonances (see \ttt{MSTP(49)}) they
are stored in positions 27, 28, 29 and 30, respectively.
\iteme{J :} enumerates the different decay channels possible for
resonance KF, as stored in the {\tsc{Jetset}} \ttt{LUDAT3} common block,
with the first channel in \ttt{J=1}, etc.
\end{subentry}
\iteme{WIDE(KF,J) :}\label{p:WIDE} gives effective decay widths of
resonances into
different channels (in GeV), given the decay modes actually left open
in the current run. The on/off status of decay modes is set by the
\ttt{MDME} switches in {\tsc{Jetset}}; see section \ref{ss:parapartdat}.
\begin{subentry}
\iteme{KF :} standard KF code for resonance considered. For comment
about top and fourth generation see \ttt{WIDP} above.
\iteme{J :} enumerates the different decay channels possible for
resonance KF, as stored in the {\tsc{Jetset}} \ttt{LUDAT3} common block,
with the first channel in \ttt{J=1}, etc.
\end{subentry}
\iteme{WIDS(KF,J) :}\label{p:WIDS} gives a dimensionless suppression
factor, which
is defined as the ratio of the total width of channels switched on to
the total width of all possible channels (replace width by squared
width for a pair of resonances). The on/off status of channels is set
by the \ttt{MDME} switches in {\tsc{Jetset}}; see section \ref{ss:parapartdat}.
The information in \ttt{WIDS} is used e.g. in cross-section
calculations.
\begin{subentry}
\iteme{KF :} standard KF code for resonance considered. For comment
about top and fourth generation see \ttt{WIDP} above.
\iteme{J = 1 :} suppression when a pair of resonances of type KF are
produced together. When an antiparticle exists, the
particle--antiparticle pair (such as $\mathrm{W}^+ \mathrm{W}^-$) is the relevant
combination, else the particle--particle one (such as $\mathrm{Z}^0 \mathrm{Z}^0$).
\iteme{J = 2 :} suppression for a particle of type KF when produced
on its own, or together with a particle of another type.
\iteme{J = 3 :} suppression for an antiparticle of type KF when produced
on its own, or together with a particle of another type.
\end{subentry}
\end{entry}
\drawbox{COMMON/PYINT5/NGEN(0:200,3),XSEC(0:200,3)}\label{p:PYINT5}
\begin{entry}
\itemc{Purpose:} to store information necessary for cross-section
calculation and differential cross-section maximum violation.
These variables must not be changed by you.
\iteme{NGEN(ISUB,1) :}\label{p:NGEN} gives the number of times that
the differential cross section (times Jacobian factors) has been
evaluated for subprocess ISUB, with \ttt{NGEN(0,1)} the sum of
these.
\iteme{NGEN(ISUB,2) :} gives the number of times that a kinematical
setup for subprocess ISUB is accepted in the generation procedure,
with \ttt{NGEN(0,2)} the sum of these.
\iteme{NGEN(ISUB,3) :} gives the number of times an event of
subprocess type ISUB is generated, with \ttt{NGEN(0,3)} the sum of
these. Usually \ttt{NGEN(ISUB,3) = NGEN(ISUB,2)}, i.e. an accepted
kinematical configuration can normally be used to produce an event.
\iteme{XSEC(ISUB,1) :}\label{p:XSEC} estimated maximum differential
cross section
(times the Jacobian factors used to speed up the generation process)
for the different subprocesses in use, with \ttt{XSEC(0,1)} the sum
of these (except low-$p_{\perp}$, i.e. ISUB = 95).
\iteme{XSEC(ISUB,2) :} gives the sum of differential cross sections
(times Jacobian factors) for the \ttt{NGEN(ISUB,1)} phase-space
points evaluated so far.
\iteme{XSEC(ISUB,3) :} gives the estimated integrated cross section
for subprocess ISUB, based on the statistics accumulated so far,
with \ttt{XSEC(0,3)} the estimated total cross section for all
subprocesses included (all in mb). This is exactly the
information obtainable by a \ttt{PYSTAT(1)} call.
\end{entry}
\drawboxtwo{COMMON/PYINT6/PROC(0:200)}{CHARACTER PROC*28}%
\label{p:PYINT6}
\begin{entry}
\itemc{Purpose:} to store character strings for the different
possible subprocesses; used when printing tables.
\iteme{PROC(ISUB) :}\label{p:PROC} name for the different
subprocesses, according to ISUB code. \ttt{PROC(0)} denotes
all processes.
\end{entry}
\drawbox{COMMON/PYINT7/SIGT(0:6,0:6,0:5)}\label{p:PYINT7}
\begin{entry}
\itemc{Purpose:} to store information on total, elastic and
diffractive cross sections. These variables should only be set
by you for the option \ttt{MSTP(31)=0}; else they should not be
touched. All numbers are given in mb.
\iteme{SIGT(I1,I2,J) :}\label{p:SIGT} the cross section, both total
and subdivided by class (elastic, diffractive etc.). For a photon
to be considered as a VMD meson the cross sections are additionally
split into the contributions from the various meson states.
\begin{subentry}
\iteme{I1, I2 :} allowed states for the incoming particle on side
1 and 2, respectively.
\begin{subentry}
\iteme{= 0 :} sum of all allowed states. Except for a photon to
be considered as a VMD meson this is the only nonvanishing entry.
\iteme{= 1 :} the contribution from the $\rho^0$ VMD state.
\iteme{= 2 :} the contribution from the $\omega$ VMD state.
\iteme{= 3 :} the contribution from the $\phi$ VMD state.
\iteme{= 4 :} the contribution from the $\mathrm{J}/\psi$ VMD state.
\iteme{= 5, 6 :} reserved for future use.
\end{subentry}
\iteme{J :} the total and partial cross sections.
\begin{subentry}
\iteme{= 0 :} the total cross section.
\iteme{= 1 :} the elastic cross section.
\iteme{= 2 :} the single diffractive cross section $AB \to XB$.
\iteme{= 3 :} the single diffractive cross section $AB \to AX$.
\iteme{= 4 :} the double diffractive cross section.
\iteme{= 5 :} the inelastic, non-diffractive cross section.
\end{subentry}
\itemc{Warning:} If you set these values yourself, it is important
that they are internally consistent, since this is not explicitly
checked by the program. Thus the contributions \ttt{J=}1--5 should
add up to the \ttt{J=}0 one and, for VMD photons, the contributions
\ttt{I=}1--4 should add up to the \ttt{I=}0 one.
\end{subentry}
\end{entry}
\drawboxtwo{~COMMON/PYINT8/XPVMD(-6:6),XPANL(-6:6),XPANH(-6:6),%
XPBEH(-6:6),}{\&XPDIR(-6:6)}\label{p:PYINT8}
\begin{entry}
\itemc{Purpose:} to store the various components of the photon
parton distributions when the \ttt{PYGGAM} routine is called.
\iteme{XPVMD(KFL) :}\label{p:XPVMD} gives distributions of the VMD part
($\rho^0$, $\omega$ and $\phi$).
\iteme{XPANL(KFL) :}\label{p:XPANL} gives distributions of the anomalous
part of light quarks ($\d$, $\u$ and $\mathrm{s}$).
\iteme{XPANH(KFL) :}\label{p:XPANH} gives distributions of the anomalous
part of heavy quarks ($\c$ and $\b$).
\iteme{XPBEH(KFL) :}\label{p:XPBEH} gives Bethe-Heitler distributions of
heavy quarks ($\c$ and $\b$). This provides an alternative to \ttt{XPANH},
i.e. both should not be used at the same time.
\iteme{XPDIR(KFL) :}\label{p:XPDIR} gives direct correction to the
production of light quarks ($\d$, $\u$ and $\mathrm{s}$). This term is
nonvanishing only in the {$\overline{\mrm{MS}}$} scheme, and is applicable for
$F_2^{\gamma}$ rather than for the parton distributions themselves.
\end{entry}
\boxsep
Finally, in addition a number of routines and common blocks with names
beginning with \ttt{RK} come with the program. These contain the
matrix-element evaluation for the process
$\mathrm{g} \mathrm{g} \to \mathrm{Z} \mathrm{q} \overline{\mathrm{q}}$, based on a
program of Ronald Kleiss, with only minor modifications.
\subsection{Examples}
The program is built as a slave system, i.e. you supply the main
program, which calls on the {\tsc{Pythia}} and {\tsc{Jetset}} routines to perform
specific tasks and then resumes control.
A typical program for the analysis of collider events at 630 GeV c.m.
energy with a minimum $p_{\perp}$ of 10 GeV/c at the hard scattering
(because of initial-state radiation, fragmentation effects, etc.,
the actual $p_{\perp}$ cut-off will be smeared around this value) might look
like
\begin{verbatim}
COMMON/LUJETS/N,K(4000,5),P(4000,5),V(4000,5)
COMMON/PYSUBS/MSEL,MSUB(200),KFIN(2,-40:40),CKIN(200)
COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)
... ! set all common block variables that
... ! did not have desired default values
CKIN(3)=10. ! lower p_T cut-off
CALL PYINIT('CMS','p','pbar',630.) ! initialize
... ! initialize analysis statistics
DO 100 IEVENT=1,1000 ! loop over events
CALL PYEVNT ! generate event
IF(IEVENT.EQ.1) CALL LULIST(1) ! list first event
... ! insert desired analysis chain for
... ! each event
100 CONTINUE
CALL PYSTAT(1) ! print cross sections
... ! user output
END
\end{verbatim}
\clearpage
\section{Initial- and Final-State Radiation}
\label{s:showinfi}
Starting from the hard interaction, initial- and final-state
radiation corrections may be added. This is normally done by
making use of the parton-shower language --- only for the
$\e^+\e^- \to \mathrm{q} \overline{\mathrm{q}}$ process does {\tsc{Jetset}} offer a matrix-element
option (described in section \ref{ss:eematrix}).
The algorithms used to generate initial- and final-state showers
are rather different, and are therefore described separately
below, starting with the conceptually easier final-state one.
Before that, some common elements are introduced.
The main reference for final-state showers is ref. \cite{Ben87a} and
for initial-state ones ref. \cite{Sjo85}.
\subsection{Shower Evolution}
In the leading log picture, a shower may be viewed as a sequence
of $1 \to 2$ branchings $a \to bc$. Here $a$ is called the mother
and $b$ and $c$ the two daughters. Each daughter is free to branch
in its turn, so that a tree-like stucture can evolve. We will use
the work `parton' for all the objects $a$, $b$ and $c$ involved
in the branching process, i.e. not only for quarks and gluons
but also for leptons and photons. The branchings included in the
program are $\mathrm{q} \to \mathrm{q} \mathrm{g}$, $\mathrm{g} \to \mathrm{g} \mathrm{g}$, $\mathrm{g} \to \mathrm{q} \overline{\mathrm{q}}$,
$\mathrm{q} \to \mathrm{q} \gamma$ and $\ell \to \ell \gamma$. Photon branchings,
i.e. $\gamma \to \mathrm{q} \overline{\mathrm{q}}$ and $\gamma \to \ell \br{\ell}$, have not
been included so far, since they are reasonably rare and since no
urgent need for them has been perceived.
\subsubsection{The evolution equations}
In the shower formulation, the kinematics of each branching is
given in terms of two variables, $Q^2$ and $z$. Slightly different
interpretations may be given to these variables, and indeed this is
one main area where the various programs on the market differ.
$Q^2$ has dimensions of squared mass, and is related to the
mass or transverse momentum scale of the branching. $z$ gives the
sharing of the $a$ energy and momentum between the two daughters,
with parton $b$ taking a fraction $z$ and parton $c$ a fraction
$1-z$. To specify the kinematics, an azimuthal angle
$\varphi$ of the $b$ around the $a$ direction is needed in addition;
normally $\varphi$ is chosen to be isotropically distributed,
although options for non-isotropic distributions exist.
The probability for a parton to branch is given by the evolution
equations (also called DGLAP or Altarelli--Parisi \cite{Gri72,Alt77}).
It is convenient to introduce
\begin{equation}
t = \ln(Q^2/\Lambda^2) ~~~ \Rightarrow ~~~
\d t = \d \ln(Q^2) = \frac{\d Q^2}{Q^2} ~,
\end{equation}
where $\Lambda$ is the QCD $\Lambda$ scale in $\alpha_{\mathrm{s}}$. Of course,
this choice is more directed towards the QCD parts of the shower,
but it can be used just as well for the QED ones. In terms of the two
variables $t$ and $z$, the differential probability $\d {\cal P}$ for
parton $a$ to branch is now
\begin{equation}
\d {\cal P}_a = \sum_{b,c} \frac{\alpha_{abc}}{2 \pi} \,
P_{a \to bc}(z) \, \d t \, \d z ~.
\label{sh:Patobc}
\end{equation}
Here the sum is supposed to run over all allowed branchings,
for a quark $\mathrm{q} \to \mathrm{q} \mathrm{g}$ and $\mathrm{q} \to \mathrm{q} \gamma$, and so on. The
$\alpha_{abc}$ factor is $\alpha_{\mathrm{em}}$ for QED branchings and
$\alpha_{\mathrm{s}}$ for QCD ones (to be evaluated at some suitable scale,
see below).
The splitting kernels $P_{a \to bc}(z)$ are
\begin{eqnarray}
P_{\mathrm{q} \to \mathrm{q}\mathrm{g}}(z)&=&C_F \, \frac{1+z^2}{1-z} ~,
\nonumber \\
P_{\mathrm{g} \to \mathrm{g}\g}(z)&=&N_C \, \frac{(1-z(1-z))^2}{z(1-z)} ~,
\nonumber \\
P_{\mathrm{g} \to \mathrm{q}\overline{\mathrm{q}}}(z)&=&T_R \, (z^2 + (1-z)^2) ~,
\nonumber \\
P_{\mathrm{q} \to \mathrm{q}\gamma}(z)&=&e_{\mathrm{q}}^2 \, \frac{1+z^2}{1-z} ~,
\nonumber \\
P_{\ell \to \ell\gamma}(z)&=&e_{\ell}^2 \, \frac{1+z^2}{1-z} ~,
\label{sh:APker}
\end{eqnarray}
with $C_F = 4/3$, $N_C = 3$, $T_R = n_f/2$ (i.e. $T_R$ receives
a contribution of $1/2$ for each allowed $\mathrm{q}\overline{\mathrm{q}}$ flavour),
and $e_{\mathrm{q}}^2$ and $e_{\ell}^2$ the squared electric charge
($4/9$ for $\u$-type quarks, $1/9$ for $\d$-type ones, and 1 for
leptons).
Persons familiar with analytical calculations may wonder why the
`+ prescriptions' and $\delta(1-z)$ terms of the splitting kernels
in eq.~(\ref{sh:APker}) are missing. These complications fulfil
the task of ensuring flavour and energy conservation in the analytical
equations. The corresponding problem is solved trivially in
Monte Carlo programs, where the shower evolution is traced in detail,
and flavour and four-momentum are conserved at each branching.
The legacy left is the need to introduce a cut-off on the allowed range
of $z$ in splittings, so as to avoid the singular regions corresponding
to excessive production of very soft gluons.
Also note that $P_{\mathrm{g} \to \mathrm{g}\g}(z)$ is given here with a factor
$N_C$ in front, while it is sometimes shown with $2 N_C$. The
confusion arises because the final state contains two identical
partons. With the normalization above, $P_{a \to bc}(z)$
is interpreted as the branching probability for the original parton
$a$. On the other hand, one could also write down the probability
that a parton $b$ is produced with a fractional energy $z$. Almost
all the above kernels can be used unchanged also for this purpose,
with the obvious symmetry $P_{a \to bc}(z) = P_{a \to cb}(1-z)$.
For $\mathrm{g} \to \mathrm{g}\g$, however, the total probability to find a gluon
with energy fraction $z$ is the sum of the probability to find either
the first or the second daughter there, and that gives the factor of
2 enhancement.
\subsubsection{The Sudakov form factor}
\label{ss:sudakov}
The $t$ variable fills the function of a kind of time for the shower
evolution. In final-state showers, $t$ is constrained to be gradually
decreasing away from the hard scattering, in initial-state ones to
be gradually increasing towards the hard scattering. This does not
mean that an individual parton runs through a range of $t$ values:
in the end, each parton is associated with a fixed $t$ value, and the
evolution procedure is just a way of picking that value. It is only
the ensemble of partons in many events that evolves continuously with
$t$, cf. the concept of parton distributions.
For a given $t$ value we define the integral of the branching
probability over all allowed $z$ values,
\begin{equation}
{\cal I}_{a \to bc}(t) = \int_{z_{-}(t)}^{z_{+}(t)} \d z \,
\frac{\alpha_{abc}}{2 \pi} \, P_{a \to bc}(z) ~.
\end{equation}
The na\"{\i}ve probability that a branching occurs during a small range
of $t$ values, $\delta t$, is given by
$\sum_{b,c} {\cal I}_{a \to bc}(t) \, \delta t$, and thus the
probability for no emission by
$1 - \sum_{b,c} {\cal I}_{a \to bc}(t) \, \delta t$.
If the evolution of parton $a$ starts at a `time' $t_0$, the
probability that the parton has not yet branched at a `later time'
$t > t_0$ is given by the product of the probabilities that it did not
branch in any of the small intervals $\delta t$ between $t_0$ and $t$.
In other words, letting $\delta t \to 0$, the no-branching probability
exponentiates:
\begin{equation}
{\cal P}_{\mrm{no-branching}}(t_0,t) =
\exp \left\{ - \int_{t_0}^t \d t' \, \sum_{b,c}
{\cal I}_{a \to bc}(t') \right\} = S_a(t) ~.
\label{sh:Pnobranch}
\end{equation}
Thus the actual probability that a branching of $a$ occurs at $t$
is given by
\begin{equation}
\frac{\d {\cal P}_a}{\d t} =
- \frac{\d {\cal P}_{\mrm{no-branching}}(t_0,t)}{\d t} =
\left( \sum_{b,c} {\cal I}_{a \to bc}(t) \right)
\exp \left\{ - \int_{t_0}^t \d t' \, \sum_{b,c}
{\cal I}_{a \to bc}(t') \right\} ~.
\label{sh:Pbranch}
\end{equation}
The first factor is the na\"{\i}ve branching probability, the second
the suppression due to the conservation of total
probability: if a parton has already branched at a `time' $t' < t$,
it can no longer branch at $t$. This is nothing but the exponential
factor that is familiar from radioactive decay. In parton-shower
language the exponential factor
$S_a(t) = {\cal P}_{\mrm{no-branching}}(t_0,t)$ is referred to as
the Sudakov form factor \cite{Sud56}.
The ordering in terms of increasing $t$ above
is the appropriate one for initial-state showers. In
final-state showers the evolution is from an initial $t_{\mathrm{max}}$
(set by the hard scattering) and towards smaller $t$. In that case
the integral from $t_0$ to $t$ in eqs.~(\ref{sh:Pnobranch})
and (\ref{sh:Pbranch}) is replaced by an integral from $t$
to $t_{\mathrm{max}}$. Since, by convention,
the Sudakov factor is still defined from the lower cut-off $t_0$,
i.e. gives the probability that a parton starting at scale $t$ will
not have branched by the lower cut-off scale $t_0$, the no-branching
factor is actually
${\cal P}_{\mrm{no-branching}}(t_{\mathrm{max}},t) =
S_a(t_{\mathrm{max}})/S_a(t)$.
We note that the above structure is exactly of the kind discussed
in section \ref{ss:vetoalg}. The veto algorithm is therefore
extensively used in the Monte Carlo simulation of parton showers.
\subsubsection{Matching to the hard scattering}
\label{sss:showermatching}
The evolution in $Q^2$ is begun from some maximum scale
$Q_{\mathrm{max}}^2$ for final-state parton showers, and is terminated
at (a possibly different) $Q_{\mathrm{max}}^2$ for initial-state showers.
In general $Q_{\mathrm{max}}^2$ is not known. Indeed, since
the parton-shower language does not guarantee agreement with
higher-order matrix-element results, neither in absolute
shape nor normalization, there is no unique prescription for
a `best' choice. Generically $Q_{\mathrm{max}}$ should be of the order
of the hard-scattering scale, i.e. the largest virtuality
should be associated with the hard scattering, and initial- and
final-state parton showers should only involve virtualities
smaller than that. This may be viewed just as a matter of sound
bookkeeping: in a $2 \to n$ graph, a $2 \to 2$ hard-scattering
subgraph could be chosen several different ways, but if all
the possibilities were to be generated then the cross section
would be double-counted. Therefore one should define the $2 \to 2$
`hard' piece of a $2 \to n$ graph as the one that involves the
largest virtuality.
Of course, the issue of double-counting depends a bit on what
processes are actually generated in the program. If one
considers a $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ final state in hadron colliders,
this could come either as final-state radiation off a
$\mathrm{q} \overline{\mathrm{q}}$ pair, or by a gluon splitting in a $\mathrm{q} \overline{\mathrm{q}}$ pair,
or many other ways, so that the danger of double-counting is
very real. On the other hand, consider the production of a
low-$p_{\perp}$, low-mass Drell--Yan pair of leptons, together with two
quark jets. Such a process in principle could proceed by having
a $\gamma^*$ emitted off a quark leg, with a quark--quark
scattering as hard interaction. However, since this process is
not included in the program, there is no actual danger of
(this particular) double-counting, and so the scale of evolution
could be picked larger than the mass of the Drell--Yan pair, at
least by some amount.
For most $2 \to 2$ scattering processes in {\tsc{Pythia}}, the
$Q^2$ scale of the hard scattering is chosen to be
$Q_{\mrm{hard}}^2 = p_{\perp}^2$ (when the final-state particles are
massless, otherwise masses are added). In final-state showers, where
$Q$ is associated with the mass of the branching parton,
transverse momenta generated in the shower are constrained by
$p_{\perp} < Q/2$. An ordering that shower $p_{\perp}$ be smaller than
the hard-scattering $p_{\perp}$ therefore corresponds roughly
to $Q_{\mathrm{max}}^2 = 4 Q_{\mrm{hard}}^2$, which is the default
assumption. In principle, the constraints are slightly different
for initial-state showers, but not enough to warrant a separate
$Q_{\mathrm{max}}$ choice.
The situation is rather better for the final-state showers in the
decay of any colour-singlet particles, such as the $\mathrm{Z}^0$ or the
$\H^0$, either as part of a hard $2 \to 1 \to 2$ process, or
anywhere else in the final state. Then we know that $Q_{\mathrm{max}}$
has to be put equal to the particle mass. It is also possible
to match the parton-shower evolution to the first-order
matrix-element results. In the program this is done under the
assumption that the resonance has spin one, and this approach is known
to work very well for $\gamma^* / \Z^0$. The machinery is not fully correct
for the spin-zero $\H^0$, but should also there provide a rather good
description.
QCD processes such as $\mathrm{q}\mathrm{g} \to \mathrm{q}\mathrm{g}$ pose a special problem when
the scattering angle is small. Coherence effects (see below) may
then restrict the emission further than what is just given by
the $Q_{\mathrm{max}}$ scale introduced above. This is most easily viewed
in the rest frame of the $2 \to 2$ hard scattering subprocess.
Some colours flow from the initial to the final state. The radiation
associated with such a colour flow should be restricted to a cone
with opening angle given by the difference between the original and
the final colour directions; there is one such cone around the
incoming parton for initial state radiation and one around the
outgoing parton for final state radiation. Colours that are
annihilated or created in the process effectively correspond to an
opening angle of 180$^{\circ}$ and therefore the emission is not
constrained for these. For a gluon, which have two colours and
therefore two different cones, a random choice is made between the
two for the first branching. Further, coherence effects also imply
azimuthal anisotropies of the emission inside the allowed cones.
\subsection{Final-State Showers}
Final-state showers are time-like, i.e. all virtualities
$m^2 = E^2 - \mbf{p}^2 \geq 0$. The maximum allowed virtuality
scale $Q^2_{\mathrm{max}}$ is set by the hard-scattering process, and
thereafter the virtuality is decreased in each subsequent
branching, down to the cut-off scale $Q_0^2$. This cut-off scale
is used to regulate both soft and collinear divergences in the
emission probabilities.
The main points of the {\tsc{Jetset}} showering algorithm are as follows.
\begin{Itemize}
\item It is a leading-log algorithm, of the improved, coherent kind,
i.e. with angular ordering.
\item It can be used for an arbitrary initial pair of partons
or, in fact, for any one, two or three given entities (including
hadrons and gauge bosons) although only quarks, gluons and
leptons can initiate a shower.
\item The pair of showering partons may be given in any frame, but
the evolution is carried out in the c.m. frame of the showering
partons.
\item Energy and momentum are conserved exactly at each step of the
showering process.
\item If the initial pair is $\mathrm{q}\overline{\mathrm{q}}$ or $\ell^+ \ell^-$ (coming
from a resonance decay) an additional rejection technique is used
in the first branching of each of the two original partons, so as
to reproduce the lowest-order differential 3-jet cross section.
\item In subsequent branchings, angular
ordering (coherence effects) is imposed.
\item Gluon helicity effects, i.e. correlations between the
production plane and the decay plane of a gluon, can be included.
\item The first-order
$\alpha_{\mathrm{s}}$ expression is used, with the $Q^2$ scale given by (an
approximation to) the squared transverse momentum of a branching.
The default $\Lambda_{\mrm{QCD}}$,
which should not be regarded as a proper
$\Lambda_{\br{\mrm{MS}}}$, is 0.4 GeV.
\item The parton shower is by default
cut off at a mass scale of 1 GeV.
\end{Itemize}
Let us now proceed with a more detailed description.
\subsubsection{The choice of evolution variable}
In the {\tsc{Jetset}} shower algorithm, the evolution variable $Q^2$ is
associated with the squared mass of the branching parton,
$Q^2 = m_a^2$ for a branching $a \to bc$. As a consequence,
$t = \ln(Q^2/\Lambda^2) = \ln(m_a^2/\Lambda^2)$.
This $Q^2$ choice is not unique, and indeed other programs have
other definitions: \tsc{Herwig} uses $Q^2 \approx m^2/(2z(1-z))$
\cite{Mar88} and \tsc{Ariadne}
$Q^2 = p_{\perp}^2 \approx z(1-z)m^2$ \cite{Pet88}.
With $Q$ a mass scale, the lower cut-off $Q_0$ is one in mass.
To be more precise, in a QCD shower, the $Q_0$ parameter is used
to derive effective masses
\begin{eqnarray}
m_{\mrm{eff},\mathrm{g}} & = & \frac{1}{2} Q_0 ~, \nonumber \\
m_{\mrm{eff},\mathrm{q}} & = & \sqrt{ m_{\mathrm{q}}^2 + \frac{1}{4} Q_0^2 } ~,
\end{eqnarray}
where the $m_{\mathrm{q}}$ have been chosen as typical current-algebra quark
masses. A parton cannot branch unless its mass is at least the sum
of the lightest pair of allowed decay products, i.e. the minimum
mass scale at which a branching is possible is
\begin{eqnarray}
m_{\mathrm{min},\mathrm{g}} & = & 2 \, m_{\mrm{eff},\mathrm{g}} = Q_0 ~, \nonumber \\
m_{\mathrm{min},\mathrm{q}} & = & m_{\mrm{eff},\mathrm{q}} + m_{\mrm{eff},\mathrm{g}} \geq Q_0 ~.
\end{eqnarray}
The above masses are used to constrain the allowed range of
$Q^2$ and $z$ values. However, once it has been decided that a
parton cannot branch any further, that parton is put on the
mass shell, i.e. `final-state' gluons are massless.
When also photon emission is included, a separate $Q_0$ scale is
introduced for the QED part of the shower, exactly reproducing
the QCD one above \cite{Sjo92c}. By default the two $Q_0$ scales
are chosen equal, and have the value 1 GeV. If anything, one would
be inclined to allow a cut-off lower for photon emission than for
gluon one. In that case the allowed $z$ range of photon emission
would be larger than that of gluon emission, and at the end of
the shower evolution only photon emission would be allowed.
Photon and gluon emission differ fundamentally in that photons appear
as physical particles in the final state, while gluons are confined.
For photon emission off quarks, however, the confinement forces
acting on the quark may provide an effective photon emission
cut-off at larger scales than the bare quark mass. Soft and collinear
photons could also be emitted by the final-state charged hadrons;
the matching between emission off quarks and off hadrons is a
delicate issue, and we therefore do not attempt to address the
soft-photon region.
For photon emission off leptons, there is no need to introduce any
collinear emission cut-off beyond what is given by the lepton mass,
but we keep the same cut-off approach as for quarks: firstly,
the program is not aimed at high-precision studies of lepton
pairs (where interference terms between initial- and final-state
radiation also would have to be included); secondly, most experimental
procedures would include the energy of collinear photons into the
effective energy of a final-state lepton.
\subsubsection{The choice of energy splitting variable}
The final-state radiation
machinery is always applied in the c.m. frame of the hard scattering,
from which normally emerges a pair of evolving partons.
Occasionally there may be one evolving parton recoiling against a
non-evolving one, as in $\mathrm{q} \overline{\mathrm{q}} \to \mathrm{g} \gamma$, where only the
gluon evolves in the final state, but where the energy of the photon
is modifed by the branching activity of the gluon. (With only one
evolving parton and nothing else, it would not be possible to
conserve energy and momentum when the parton is assigned a mass.)
Thus, before the evolution is performed, the parton pair is boosted
to their common c.m. frame, and rotated to sit along the $z$ axis.
After the evolution, the full parton shower is rotated and boosted
back to the original frame of the parton pair.
The interpretation of the energy and momentum splitting variable
$z$ is not unique, and in fact the program allows the possibility
to switch between four different alternatives \cite{Ben87a},
`local' and `global' $z$ definition combined with `constrained'
or `unconstrained' evolution. In all
four of them, the $z$ variable is interpreted as an energy fraction,
i.e. $E_b = z E_a$ and $E_c = (1-z) E_a$. In the `local' choice of
$z$ definition, energy fractions are defined in the rest frame
of the grandmother, i.e. the mother of parton $a$. The preferred
choice is the `global' one, in which energies are always evaluated
in the c.m. frame of the hard scattering. The two definitions agree
for the branchings of the partons that emerge directly from the hard
scattering, since the hard scattering itself is considered to be the
`mother' of the first generation of partons. For instance, in
$\mathrm{Z}^0 \to \mathrm{q}\overline{\mathrm{q}}$ the $\mathrm{Z}^0$ is considered the mother of the $\mathrm{q}$
and $\overline{\mathrm{q}}$, even though the branching is not handled by the
parton-showering machinery. The `local' and `global' definitions
diverge for subsequent branchings, where the `global' tends to allow
more shower evolution.
In a branching $a \to bc$ the kinematically allowed range of
$z = z_a$ values, $z_{-} < z < z_{+}$, is given by
\begin{equation}
z_{\pm} = \frac{1}{2} \left\{ 1 + \frac{m_b^2 - m_c^2}{m_a^2} \pm
\frac{|\mbf{p}_a|}{E_a} \, \frac{\sqrt{ (m_a^2 - m_b^2 - m_c^2)^2
- 4 m_b^2 m_c^2}}{m_a^2} \right\} ~.
\label{sh:zbounds}
\end{equation}
With `constrained' evolution, these bounds are respected in the
evolution. The cut-off masses $m_{\mrm{eff},b}$ and $m_{\mrm{eff},c}$
are used to
define the maximum allowed $z$ range, within which $z_a$ is chosen,
together with the $m_a$ value. In the subsequent evolution of $b$ and
$c$, only pairs of $m_b$ and $m_c$ are allowed for which the
already selected $z_a$ fulfils the constraints in
eq.~(\ref{sh:zbounds}).
For `unconstrained' evolution, which is the preferred alternative,
one may start off by assuming the daughters to be massless, so that the
allowed $z$ range is
\begin{equation}
z_{\pm} = \frac{1}{2} \left\{ 1 \pm \frac{|\mbf{p}_a|}{E_a}
\theta(m_a - m_{\mathrm{min},a}) \right\} ~,
\label{sh:zrange}
\end{equation}
where $\theta(x)$ is the step function, $\theta(x) = 1$ for $x > 0$
and $\theta(x) = 0$ for $x < 0$. The decay kinematics into two
massless four-vectors $p_b^{(0)}$ and $p_c^{(0)}$ is now
straightforward. Once $m_b$ and $m_c$ have been found from the
subsequent evolution, subject only to the constraints
$m_b < z_a E_a$, $m_c < (1-z_a) E_a$ and $m_b + m_c < m_a$,
the actual massive four-vectors may be defined as
\begin{equation}
p_{b,c} = p_{b,c}^{(0)} \pm (r_c p_c^{(0)} - r_b p_b^{(0)}) ~,
\label{sh:pshowershift}
\end{equation}
where
\begin{equation}
r_{b,c} = \frac{m_a^2 \pm (m_c^2 -m_b^2) -
\sqrt{ (m_a^2 - m_b^2 - m_c^2)^2 - 4 m_b^2 m_c^2}}{2 m_a^2} ~.
\end{equation}
In other words, the meaning of $z_a$ is somewhat reinterpreted
{\it post facto}. Needless to say, the `unconstrained' option allows
more branchings to take place than the `constrained' one.
In the following discussion we will only refer to the
`global, unconstrained' $z$ choice.
\subsubsection{First branchings and matrix-element matching}
The final-state evolution is normally started from some initial
parton pair $1 + 2$, at a $Q_{\mathrm{max}}^2$ scale determined by
deliberations already discussed. When the evolution of parton 1
is considered, it is assumed that parton 2 is massless,
so that the parton 1 energy and momentum are simple
functions of its mass (and of the c.m. energy of the pair, which is
fixed), and hence also the allowed $z_1$ range for splittings is a
function of this mass,
eq.~(\ref{sh:zrange}). Correspondingly, parton 2 is evolved under the
assumption that parton 1 is massless. After both partons have been
assigned masses, their correct energies may be found, which are
smaller than originally assumed. Therefore the allowed $z$ ranges
have shrunk, and it may happen that a branching has been assigned
a $z$ value outside this range. If so, the parton is evolved
downwards in mass from the rejected mass value; if both $z$ values
are rejected, the parton with largest mass is evolved further.
It may also happen that the sum of $m_1$ and $m_2$ is larger
than the c.m. energy, in which case the one with the larger mass
is evolved downwards. The checking and evolution steps are
iterated until an acceptable set of $m_1$, $m_2$, $z_1$
and $z_2$ has been found.
The procedure is an extension of the veto
algorithm, where an initial overestimation of the allowed $z$
range is compensated by rejection of some branchings. One should
note, however, that the veto algorithm is not strictly
applicable for the coupled evolution in two variables ($m_1$
and $m_2$), and that therefore some arbitrariness is involved.
This is manifest in the choice of which parton will be evolved
further if both $z$ values are unacceptable, or if the mass sum
is too large.
For quark and lepton pairs which come from the decay of a
colour-singlet particle, the first branchings are matched to
the explicit first-order matrix elements for gauge boson decays.
This is also done, e.g. in $\H^0$ decays, which has spin 0 rather
than 1, and for which in principle therefore the matrix elements
are slightly different.
The matching is based on a mapping of the parton-shower variables
on to the 3-jet phase space. To produce a 3-jet event,
$\gamma^* / \Z^0 \to \mathrm{q}(p_1) \overline{\mathrm{q}}(p_2) \mathrm{g}(p_3)$, in the shower language,
one will pass through an intermediate
state, where either the $\mathrm{q}$ or the $\overline{\mathrm{q}}$ is off the mass shell.
If the former is the case then
\begin{eqnarray}
m^2 & = & (p_1 + p_3)^2 = E_{\mrm{cm}}^2 (1 - x_2) ~, \nonumber \\
z & = & \frac{E_1}{E_1 + E_3} = \frac{x_1}{x_1 + x_3} =
\frac{x_1}{2-x_2} ~,
\label{sh:MEfrPS}
\end{eqnarray}
where $x_i = 2 E_i/E_{\mrm{cm}}$. The $\overline{\mathrm{q}}$ emission case is obtained
with $1 \leftrightarrow 2$. The parton-shower splitting expression
in terms of $m^2$ and $z$, eq.~(\ref{sh:Patobc}), can therefore be
translated into the following differential 3-jet rate:
\begin{eqnarray}
\frac{1}{\sigma} \, \frac{\d \sigma_{\mrm{PS}}}{\d x_1 \, \d x_2}
& = & \frac{\alpha_{\mathrm{s}}}{2 \pi} \, C_F \, \frac{1}{(1-x_1)(1-x_2)}
\times
\nonumber \\
& \times &
\left\{ \frac{1-x_1}{x_3} \left( 1 + \left( \frac{x_1}{2-x_2}
\right)^2 \right) + \frac{1-x_2}{x_3} \left( 1 +
\left( \frac{x_2}{2-x_1} \right)^2 \right) \right\} ~,
\label{sh:PSwt}
\end{eqnarray}
where the first term inside the curly bracket comes from emission
off the quark and the second term from emission off the antiquark.
The corresponding expression in matrix-element language is
\begin{equation}
\frac{1}{\sigma} \, \frac{\d \sigma_{\mrm{ME}}}{\d x_1 \, \d x_2} =
\frac{\alpha_{\mathrm{s}}}{2 \pi} \, C_F \, \frac{1}{(1-x_1)(1-x_2)}
\left\{ x_1^2 +x_2^2 \right\} ~.
\label{sh:MEwt}
\end{equation}
With the kinematics choice of \tsc{Jetset},
the matrix-element expression is always smaller than the parton-shower
one. It is therefore possible to run the shower as usual,
but to impose an extra weight factor
$\d \sigma_{\mrm{ME}} / \d \sigma_{\mrm{PS}}$,
which is just the ratio of the expressions in curly brackets.
If a branching is rejected, the evolution is continued from the
rejected $Q^2$ value onwards (the veto algorithm). The weighting
procedure is applied to the first branching of both the $\mathrm{q}$ and the
$\overline{\mathrm{q}}$, in each case with the (nominal) assumption that none of
the other partons branch (neither the sister nor the daughters),
so that the relations of eq.~(\ref{sh:MEfrPS}) are applicable.
If a photon is emitted instead of a gluon, the emission rate in
parton showers is given by
\begin{eqnarray}
\frac{1}{\sigma} \, \frac{\d \sigma_{\mrm{PS}}}{\d x_1 \, \d x_2}
& = & \frac{\alpha_{\mathrm{em}}}{2 \pi} \, \frac{1}{(1-x_1)(1-x_2)}
\times \nonumber \\
& \times &
\left\{ e_{\mathrm{q}}^2 \, \frac{1-x_1}{x_3} \left( 1 + \left(
\frac{x_1}{2-x_2} \right)^2 \right) + e_{\overline{\mathrm{q}}}^2 \, \frac{1-x_2}{x_3}
\left( 1 + \left( \frac{x_2}{2-x_1} \right)^2 \right) \right\} ~,
\end{eqnarray}
and in matrix elements by \cite{Gro81}
\begin{equation}
\frac{1}{\sigma} \, \frac{\d \sigma_{\mrm{ME}}}{\d x_1 \, \d x_2} =
\frac{\alpha_{\mathrm{em}}}{2 \pi} \, \frac{1}{(1-x_1)(1-x_2)}
\left\{ \left( e_{\mathrm{q}} \, \frac{1-x_1}{x_3} - e_{\overline{\mathrm{q}}} \,
\frac{1-x_2}{x_3} \right)^2 \left( x_1^2 +x_2^2 \right) \right\} ~.
\end{equation}
As in the gluon emission case, a weighting factor
$\d \sigma_{\mrm{ME}} / \d \sigma_{\mrm{PS}}$ can therefore be applied
when either the original $\mathrm{q}$ ($\ell$) or the original $\overline{\mathrm{q}}$
($\br{\ell}$) emits a photon. For a
neutral resonance, such as $\mathrm{Z}^0$, where $e_{\overline{\mathrm{q}}} = - e_{\mathrm{q}}$,
the above expressions simplify and one recovers
exactly the same ratio $\d \sigma_{\mrm{ME}} / \d \sigma_{\mrm{PS}}$
as for gluon emission.
Compared with the standard matrix-element treatment, a few
differences remain. The shower one automatically contains the
Sudakov form factor and an $\alpha_{\mathrm{s}}$ running as a function
of the $p_{\perp}^2$ scale of the branching.
The shower also allows all partons to evolve
further, which means that the na\"{\i}ve kinematics assumed for a
comparison with matrix elements is modified by subsequent
branchings, e.g. that the energy of parton 1 is reduced when
parton 2 is assigned a mass. All these effects are formally of
higher order, and so do not affect a first-order comparison.
This does not mean that the corrections need be small, but experimental
results are encouraging: the approach outlined does every bit as
good as explicit second-order matrix elements for the description
of 4-jet production.
\subsubsection{Subsequent branches and angular ordering}
The shower evolution is (almost) always done on a pair of partons,
so that energy and momentum can be conserved. In the first
step of the evolution, the two original partons thus undergo
branchings $1 \to 3 + 4$ and $2 \to 5 + 6$. As described
above, the allowed $m_1$, $m_2$, $z_1$ and $z_2$ ranges
are coupled by kinematical constraints. In the second step,
the pair $3 + 4$ is evolved and,
separately, the pair $5 + 6$. Considering only the former (the
latter is trivially obtained by symmetry), the partons thus have
nominal initial energies $E_3^{(0)} = z_1 E_1$ and
$E_4^{(0)} = (1-z_1) E_1$, and maximum allowed virtualities
$m_{\mathrm{max},3} = \min(m_1,E_3^{(0)})$ and
$m_{\mathrm{max},4} = \min(m_1,E_4^{(0)})$. Initially partons 3 and 4 are
evolved separately, giving masses $m_3$ and $m_4$ and splitting
variables $z_3$ and $z_4$. If $m_3 + m_4 > m_1$,
the parton of 3 and 4 that has the largest ratio of
$m_i/m_{\mathrm{max},i}$ is
evolved further. Thereafter eq.~(\ref{sh:pshowershift}) is used to
construct corrected energies $E_3$ and $E_4$, and the $z$
values are checked for consistency. If a branching has to be
rejected because the change of parton energy puts $z$ outside the
allowed range, the parton is evolved further.
This procedure can then be iterated for the evolution of the two
daughters of parton 3 and for the two of parton 4, etc., until each
parton reaches the cut-off mass $m_{\mathrm{min}}$. Then the parton is put on
the mass shell.
The model, as described so far, produces so-called conventional
showers, wherein masses are strictly decreasing in the shower
evolution. Emission angles are decreasing only in an average sense,
however, which means that also fairly `late' branchings can give
partons at large angles. Theoretical studies beyond the leading-log
level show that this is not correct \cite{Mue81}, but that destructive
interference effects are large in the region of non-ordered
emission angles. To a very good first approximation, these so-called
coherence effects can be taken into account in parton shower
programs by requiring a strict ordering in terms of decreasing
emission angles.
The coherence phenomenon is known already from QED. One
manifestation is the Chudakov effect \cite{Chu55}, discovered
in the study of high-energy cosmic $\gamma$ rays impinging on a
nuclear target. If a $\gamma$ is
converted into a highly collinear $\e^+\e^-$ pair inside the
emulsion, the $\mathrm{e}^+$ and $\mathrm{e}^-$ in their travel through the
emulsion ionize atoms and thereby produce blackening.
However, near the conversion point the blackening is small:
the $\mathrm{e}^+$ and $\mathrm{e}^-$ then are still close together, so that
an atom traversed by the pair does not resolve the individual
charges of the $\mathrm{e}^+$ and the $\mathrm{e}^-$, but only feels a net
charge close to zero. Only later,
when the $\mathrm{e}^+$ and $\mathrm{e}^-$ are separated by more than a typical
atomic radius, are the two able to ionize independently of
each other.
The situation is similar in QCD, but is further extended, since
now also gluons carry colour. For example, in a branching
$\mathrm{q}_0 \to \mathrm{q}\mathrm{g}$ the $\mathrm{q}$ and $\mathrm{g}$ share a colour--anticolour
pair, and therefore the $\mathrm{q}$ and $\mathrm{g}$ cannot emit subsequent gluons
incoherently. Again the net effect is to reduce the amount of soft
gluon emission: since a soft gluon (emitted at large angles)
corresponds to a large (transverse) wavelength,
the soft gluon is unable to resolve the separate colour charges
of the $\mathrm{q}$ and the $\mathrm{g}$, and only feels the net charge carried by
the $\mathrm{q}_0$. Such a soft gluon $\mathrm{g}'$ (in the region
$\theta_{\mathrm{q}_0 \mathrm{g}'} > \theta_{\mathrm{q} \mathrm{g}}$)
could therefore be thought of as being
emitted by the $\mathrm{q}_0$ rather than by the $\mathrm{q}$--$\mathrm{g}$ system.
If one considers only emission that should be associated with the
$\mathrm{q}$ or the $\mathrm{g}$, to a good approximation, there is
a complete destructive interference in the
regions of non-decreasing opening angles, while partons
radiate independently of each other inside the regions
of decreasing opening angles ($\theta_{q g'} < \theta_{q g}$ and
$\theta_{g g'} < \theta_{q g}$), once azimuthal angles are averaged
over. The details of the colour interference pattern are
reflected in non-uniform azimuthal emission probabilities.
The first branchings of the shower are not affected by the
angular-ordering requirement --- since the evolution is performed
in the c.m. frame of the original parton pair, where the original
opening angle is 180$^{\circ}$, any angle would anyway be smaller
than this --- but here instead the matrix-element matching procedure
is used, where applicable. Subsequently, each opening angle is
compared with that of the preceding branching in the shower.
For a branching $a \to bc$ the kinematical approximation
\begin{equation}
\theta_a \approx \frac{p_{\perp b}}{E_b} + \frac{p_{\perp c}}{E_c}
\approx \sqrt{z_a (1-z_a)} m_a \left( \frac{1}{z_a E_a} +
\frac{1}{(1-z_a) E_a} \right) = \frac{1}{\sqrt{z_a(1-z_a)}}
\frac{m_a}{E_a}
\end{equation}
is used to derive the opening angle (this is anyway to the same level
of approximation as the one in which angular ordering is derived).
With $\theta_b$ of the $b$ branching calculated similarly, the
requirement $\theta_b < \theta_a$ can be reduced to
\begin{equation}
\frac{z_b (1-z_b)}{m_b^2} > \frac{1-z_a}{z_a m_a^2} ~.
\end{equation}
Since photons do not obey angular ordering, the check on angular
ordering is not performed when a photon is emitted.
When a gluon is emitted in the branching after a photon, its emission
angle is restricted by that of the preceding QCD branching in the
shower, i.e. the photon emission angle does not enter.
\subsubsection{Other final-state shower aspects}
The electromagnetic coupling constant for the emission of photons
on the mass shell is
$\alpha_{\mathrm{em}} = \alpha_{\mathrm{em}}(Q^2 = 0) \approx 1/137$. For
the strong coupling constant several alternatives are available, the
default being the first-order expression $\alpha_{\mathrm{s}}(p_{\perp}^2)$, where
$p_{\perp}^2$ is defined by the approximate expression
$p_{\perp}^2 \approx z(1-z) m^2$. Studies of next-to-leading-order
corrections favour this choice \cite{Ama80}. The other alternatives
are a fixed $\alpha_{\mathrm{s}}$ and an $\alpha_{\mathrm{s}}(m^2)$.
With the default choice of $p_{\perp}^2$ as scale in $\alpha_{\mathrm{s}}$,
a further cut-off is introduced on the allowed phase space
of gluon emission, not present in the options with fixed
$\alpha_{\mathrm{s}}$ or with $\alpha_{\mathrm{s}}(m^2)$, nor in the QED shower.
A minimum requirement, to ensure a well-defined $\alpha_{\mathrm{s}}$,
is that $p_{\perp} / \Lambda > 1.1$, but additionally
{\tsc{Jetset}} requires that $p_{\perp} > Q_0/2$. This latter
requirement is not a necessity, but it makes sense when
$p_{\perp}$ is taken to be the preferred scale of the branching
process, rather than e.g. $m$. It reduces the allowed $z$ range,
compared with the purely kinematical constraints.
Since the $p_{\perp}$ cut is not
present for photon emission, the relative ratio of photon to gluon
emission off a quark is enhanced at small virtualities compared with
na\"{\i}ve expectations; in actual fact this enhancement is largely
compensated by the running of $\alpha_{\mathrm{s}}$, which acts in the
opposite direction. The main consequence, however, is that the
gluon energy spectrum is peaked at around $Q_0$ and rapidly
vanishes for energies below that, whilst the photon spectum
extends all the way to zero energy.
Previously it was said that azimuthal angles in branchings are
chosen isotropically. In fact, as an option, it is possible to
include some effects of gluon polarization, which correlate the
production and the decay planes of a gluon, such that a
$\mathrm{g} \to \mathrm{g}\g$ branching tends to take place in the production
plane of the gluon, while a decay out of the plane is favoured
for $\mathrm{g} \to \mathrm{q}\overline{\mathrm{q}}$. The formulae are given e.g. in ref.
\cite{Web86}, as simple functions of the $z$ value at the
vertex where the gluon is produced and of the $z$ value when
it branches. Also coherence phenomena lead to non-isotropic
azimuthal distributions \cite{Web86}, which are included as a
further option. In either case the $\varphi$ azimuthal variable
is first chosen isotropically, then the weight factor due to
polarization times coherence is evaluated, and the $\varphi$ value
is accepted or rejected. In case of rejection, a new $\varphi$ is
generated, and so on.
While the rule is to have an initial pair of partons, there are
a few examples where one or three partons have to be allowed to
shower. If only one parton is given, it is not possible to
conserve both energy and momentum. The choice has been made
to conserve energy and jet direction, but the momentum vector is
scaled down when the radiating parton acquires a mass. The `rest
frame of the system', used e.g. in the $z$ definition, is taken to
be whatever frame the jet is given in.
In $\Upsilon \to \mathrm{g}\g\mathrm{g}$ decays and other primary three-parton
configurations, one is left with the issue how the energy
sharing variables $x_1$ and $x_2$ from the massless matrix elements
should be reinterpreted for a massive three-parton configuration.
We have made the arbitrary choice of preserving the energy of each
parton, which means that relative angles between the original partons
is changed. Mass triplets outside the allowed phase space are
rejected and the evolution continued.
Finally, it should be noted that two toy shower models are
included as options. One is a scalar gluon model, in which the
$\mathrm{q} \to \mathrm{q}\mathrm{g}$ branching kernel is replaced by
$P_{\mathrm{q} \to \mathrm{q}\mathrm{g}}(z) = \frac{2}{3} (1-z)$. The couplings of the
gluon, $\mathrm{g} \to \mathrm{g}\g$ and $\mathrm{g} \to \mathrm{q}\overline{\mathrm{q}}$, have been left as free
parameters, since they depend on the colour structure assumed in the
model. The spectra are flat in $z$ for a spin 0 gluon.
Higher-order couplings of the type $\mathrm{g} \to \mathrm{g}\g\mathrm{g}$ could well
contribute significantly, but are not included.
The second toy model is an Abelian vector one. In this
option $\mathrm{g} \to \mathrm{g} \mathrm{g}$ branchings are absent, and $\mathrm{g} \to \mathrm{q} \overline{\mathrm{q}}$
ones enhanced. More precisely, in the splitting kernels, eq.
(\ref{sh:APker}), the Casimir factors are changed as follows:
$C_F = 4/3 \to 1$, $N_C = 3 \to 0$, $T_R = n_f/2 \to 3n_f$.
When using either of these options, one should be aware that also
a number of other components in principle should be changed, from
the running of $\alpha_{\mathrm{s}}$ to the whole concept of fragmentation.
One should therefore not take them too seriously.
\subsection{Initial-State Showers}
The initial-state showe algorithm in {\tsc{Pythia}} is not quite as
sophisticated as the final-state one. This is partly because
initial-state radiation is less well understood theoretically,
partly because the programming task is
more complicated and ambiguous. Still, the program at disposal
is known to do a reasonably good job of describing existing data,
such as $\mathrm{Z}^0$ production properties at hadron colliders
\cite{Sjo85}.
\subsubsection{The shower structure}
\label{sss:initshowstruc}
A fast hadron may be viewed as a cloud of quasireal partons.
Similarly a fast lepton may be viewed as surrounded by a cloud
of photons and partons; in the program the two situations are on
an equal footing, but here we choose the hadron as example. At
each instant, an individual parton can initiate a virtual cascade,
branching into a number of partons. This cascade can be described
in terms of a tree-like structure, composed of many subsequent
branchings $a \to bc$. Each branching involves some relative
transverse momentum between the two daughters. In a language where
four-momentum is conserved at each vertex, this implies that at
least one of the $b$ and $c$ partons must have a space-like
virtuality, $m^2 < 0$. Since the partons are not on the mass
shell, the cascade only lives a finite time before reassembling, with
those parts of the cascade that are most off the mass shell living the
shortest time.
A hard scattering, e.g. in deep inelastic leptoproduction, will
probe the hadron at a given instant. The probe, i.e. the virtual
photon in the leptoproduction case, is able to resolve fluctuations
in the hadron up to the $Q^2$ scale of the hard scattering. Thus
probes at different $Q^2$ values will seem to see different parton
compositions in the hadron. The change in parton composition with
$t = \ln(Q^2/\Lambda^2)$ is given by the evolution equations
\begin{equation}
\frac{\d f_b(x,t)}{\d t} = \sum_{a,c} \int \frac{\d x'}{x'} \,
f_a(x',t) \, \frac{\alpha_{abc}}{2 \pi} \,
P_{a \to bc} \left( \frac{x}{x'} \right) ~.
\label{sh:sfevol}
\end{equation}
Here the $f_i(x,t)$ are the parton-distribution functions, expressing
the probability of finding a parton $i$ carrying a fraction $x$
of the total momentum if the hadron is probed at virtuality $Q^2$.
The $P_{a \to bc}(z)$ are given in eq.~(\ref{sh:APker}). As before,
$\alpha_{abc}$ is $\alpha_{\mathrm{s}}$ for QCD shower and $\alpha_{\mathrm{em}}$
for QED ones.
Eq.~(\ref{sh:sfevol}) is closely related to eq.~(\ref{sh:Patobc}):
$\d {\cal P}_a$ describes the probability that a given parton $a$ will
branch (into partons $b$ and $c$), $\d f_b$ the influx of partons
$b$ from the branchings of partons $a$. (The expression $\d f_b$ in
principle also should contain a loss term for partons $b$ that
branch; this term is important for parton-distribution evolution,
but does not appear explicitly in what we shall be using eq.
(\ref{sh:sfevol}) for.) The absolute form of
hadron parton distributions cannot be predicted in perturbative
QCD, but rather have to be parametrized at some $Q_0$ scale, with
the $Q^2$ dependence thereafter given by eq.~(\ref{sh:sfevol}).
Available parametrizations are discussed in section \ref{ss:structfun}.
The lepton and photon parton distributions inside a lepton can be
fully predicted, but here for simplicity are treated on equal footing
with hadron parton distributions.
If a hard interaction scatters a parton out of the incoming hadron,
the `coherence' \cite{Gri83} of the cascade is broken: the partons can
no longer reassemble completely back to the cascade-initiating parton.
In this semiclassical picture, the partons on the `main chain' of
consecutive branchings that lead directly from the initiating parton
to the scattered parton can no longer reassemble, whereas fluctuations
on the `side branches' to this chain may still disappear. A
convenient description is obtained by assigning a space-like
virtuality to the partons on the main chain, in such a way that
the partons on the side branches may still be on the mass shell.
Since the momentum transfer of the hard process can put the scattered
parton on the mass shell (or even give it a time-like virtuality, so
that it can initiate a final-state shower), one is then guaranteed
that no partons have a space-like virtuality in the final state. (In
real life, confinement effects obviously imply that partons need not
be quite on the mass shell.) If no hard scattering had taken place,
the virtuality of the space-like parton line would still force the
complete cascade to reassemble. Since the virtuality of the cascade
probed is carried by one single parton, it is possible to equate the
space-like virtuality of this parton with the $Q^2$ scale of the
cascade, to be used e.g. in the evolution equations. Further,
coherence effects \cite{Gri83,Bas83} guarantee that the $Q^2$ vaules
of the partons along the main chain are strictly ordered, with the
largest $Q^2$ values close to the hard scattering.
In recent years, further coherence effects have been studied
\cite{Cia87}, with particular implications for the
structure of parton showers at small $x$. None of these additional
complications are implemented in the current algorithm, with the
exception of a few rather primitive options that do not address
the full complexity of the problem.
Instead of having a tree-like structure, where all legs are treated
democratically, the cascade is reduced to a single sequence of
branchings $a \to bc$, where the $a$ and $b$ partons are on the
main chain of space-like virtuality, $m_{a,b}^2 < 0$, while the $c$
partons are on the mass shell and do not branch. (Later we will
include the possibility that the $c$ partons may have positive
virtualities, $m_c^2 > 0$, which leads to the appearance of time-like
`final-state' parton showers on the side branches.) This truncation
of the cascade is only possible when it is known which parton
actually partakes in the hard scattering: of all the possible
cascades that exist virtually in the incoming hadron, the hard
scattering will select one.
To obtain the correct $Q^2$ evolution of parton distributions,
e.g., it is essential that all branches of the cascade be treated
democratically. In Monte Carlo simulation of space-like showers
this is a major problem. If indeed the evolution of the complete
cascade is to be followed from some small $Q_0^2$ up to the
$Q^2$ scale of the hard scattering, it is no possible at the
same time to handle kinematics exactly, since the virtuality of the
various partons cannot be found until after the hard scattering
has been selected. This kind of `forward evolution' scheme therefore
requires a number of extra tricks to be made to work. Further, in
this approach it is not known e.g. what the $\hat{s}$ of the
hard scattering subsystem will be until the evolution has been
carried out, which means that the initial-state evolution
and the hard scattering have to be selected jointly, a not so
trivial task.
Instead we use the `backwards evolution' approach \cite{Sjo85},
in which the hard scattering is first selected, and the parton
shower that preceded it is subsequently reconstructed. This
reconstruction is started at the hard interaction, at the
$Q_{\mathrm{max}}^2$ scale, and thereafter step by step one moves
`backwards' in `time', towards smaller $Q^2$, all the way back
to the parton-shower initiator at the cut-off scale $Q_0^2$.
This procedure is possible if evolved parton distributions are
used to select the hard scattering, since the $f_i(x,Q^2)$
contain the inclusive summation of all initial-state parton-shower
histories that can lead to the appearance of an interacting
parton $i$ at the hard scale. What remains is thus to select an
exclusive history from the set of inclusive ones.
\subsubsection{Longitudinal evolution}
The evolution equations, eq.~(\ref{sh:sfevol}), express that, during
a small increase $\d t$ there is a probability for parton $a$ with
momentum fraction $x'$ to become resolved into parton $b$ at
$x = z x'$ and another parton $c$ at $x' - x = (1-z) x'$.
Correspondingly, in backwards evolution, during a decrease $\d t$
a parton $b$ may be `unresolved' into parton $a$. The relative
probability $\d {\cal P}_b$ for this to happen is given by
the ratio $\d f_b / f_b$. Using eq.~(\ref{sh:sfevol}) one obtains
\begin{equation}
\d {\cal P}_b = \frac{\d f_b(x,t)}{f_b(x,t)} = |\d t| \, \sum_{a,c}
\int \frac{\d x'}{x'} \, \frac{f_a(x',t)}{f_b(x,t)} \,
\frac{\alpha_{abc}}{2 \pi} \,
P_{a \to bc} \left( \frac{x}{x'} \right) ~.
\end{equation}
Summing up the cumulative effect of many small changes $\d t$, the
probability for no radiation exponentiates. Therefore one may
define a form factor
\begin{eqnarray}
S_b(x,t_{\mathrm{max}},t) & = &
\exp \left\{ - \int_t^{t_{\mathrm{max}}} \d t' \, \sum_{a,c} \int
\frac{\d x'}{x'} \, \frac{f_a(x',t')}{f_b(x,t')} \,
\frac{\alpha_{abc}(t')}{2\pi} \, P_{a \to bc} \left(
\frac{x}{x'} \right) \right\} \nonumber \\
& = & \exp \left\{ - \int_t^{t_{\mathrm{max}}} \d t' \, \sum_{a,c}
\int \d z \, \frac{\alpha_{abc}(t')}{2\pi} \, P_{a \to bc}(z) \,
\frac{x'f_a(x',t')}{xf_b(x,t')} \right\} ~,
\end{eqnarray}
giving the probability that a parton $b$ remains at $x$ from
$t_{\mathrm{max}}$ to a $t < t_{\mathrm{max}}$.
It may be useful to compare this with the corresponding expression
for forward evolution, i.e. with $S_a(t)$ in eq.~(\ref{sh:Pnobranch}).
The most obvious difference is the appearance of parton distributions
in $S_b$. Parton distributions are absent in $S_a$: the probability
for a given parton $a$ to branch, once it exists, is independent of
the density of partons $a$ or $b$. The parton distributions in $S_b$,
on the other hand, express the fact that the probability for a parton
$b$ to come from the branching of a parton $a$ is proportional to
the number of partons $a$ there are in the hadron, and inversely
proportional to the number of partons $b$. Thus the numerator
$f_a$ in the exponential of $S_b$ ensures that the parton composition
of the hadron
is properly reflected. As an example, when a gluon is chosen at the
hard scattering and evolved backwards, this gluon is more likely to
have been emitted by a $\u$ than by a $\d$ if the incoming hadron is
a proton. Similarly, if a heavy flavour is chosen at the hard
scattering, the denominator $f_b$ will vanish at the $Q^2$ threshold
of the heavy-flavour production, which means that the integrand
diverges and $S_b$ itself vanishes, so that no heavy flavour remain
below threshold.
Another difference between $S_b$ and $S_a$, already touched upon, is
that the $P_{\mathrm{g} \to \mathrm{g}\g}(z)$ splitting kernel appears with a
normalization $2 N_C$ in $S_b$ but only with $N_C$ in $S_a$, since
two gluons are produced but only one decays in a branching.
A knowledge of $S_b$ is enough to reconstruct the parton shower
backwards. At each branching $a \to bc$, three quantities have to
be found: the $t$ value of the branching (which defines the
space-like virtuality $Q_b^2$ of parton $b$), the parton flavour $a$
and the splitting variable $z$. This information may be extracted as
follows:
\begin{Enumerate}
\item If parton $b$ partook in the hard scattering or branched into
other partons at a scale $t_{\mathrm{max}}$, the probability that $b$ was
produced in a branching $a \to bc$ at a lower scale $t$ is
\begin{equation}
\frac{\d {\cal P}_b}{\d t} = - \frac{\d S_b(x,t_{\mathrm{max}},t)}{\d t} =
\left( \sum_{a,c} \int dz \, \frac{\alpha_{abc}(t')}{2\pi} \,
P_{a \to bc}(z) \, \frac{x'f_a(x',t')}{xf_b(x,t')} \right)
S_b(x,t_{\mathrm{max}},t) ~.
\end{equation}
If no branching is found above the cut-off scale $t_0$ the
iteration is stopped and parton $b$ is assumed to be massless.
\item Given the $t$ of a branching, the relative probabilities
for the different allowed branchings $a \to bc$ are given by the
$z$ integrals above, i.e. by
\begin{equation}
\int \d z \, \frac{\alpha_{abc}(t)}{2\pi} \, P_{a \to bc}(z) \,
\frac{x'f_a(x',t)}{xf_b(x,t)} ~.
\label{sh:Pbspace}
\end{equation}
\item Finally, with $t$ and $a$ known, the probability distribution
in the splitting variable $z = x/x' = x_b/x_a$ is given by the
integrand in eq.~(\ref{sh:Pbspace}).
\end{Enumerate}
In addition, the azimuthal angle $\varphi$ of the branching is
selected isotropically, i.e. no spin or coherence effects are
included in this distribution.
The selection of $t$, $a$ and $z$ is then a standard task of the
kind than can be performed with the help of the veto algorithm.
Specifically, upper and lower bounds for parton distributions are
used to find simple functions that are everywhere larger than the
integrands in eq.~(\ref{sh:Pbspace}). Based on these simple expressions,
the integration over $z$ may be carried out, and $t$, $a$ and $z$
values selected. This set is then accepted with a weight given
by a ratio of the correct integrand in eq.~(\ref{sh:Pbspace}) to
the simple approximation used, both evaluated for the given set.
Since parton distributions, as a rule, are not in a simple
analytical form, it may be tricky to find reasonably good bounds to
parton distributions. It is necessary to make different assumptions
for valence and sea quarks, and be especially attentive close to a
flavour threshold (\cite{Sjo85}). An electron distribution
inside an electron behaves differently from parton distributions
encountered in hadrons, and has to be considered separately.
A comment on soft gluon emission. Nominally the range of the $z$
integral in $S_b$ is $x \leq z \leq 1$. The lower limit corresponds
to $x' = x/z = 1$, and parton distributions vanish in this limit,
wherefore no problems are encountered here. At the upper cut-off
$z=1$ the splitting kernels $P_{\mathrm{q} \to \mathrm{q}\mathrm{g}}(z)$ and
$P_{\mathrm{g} \to \mathrm{g}\g}$ diverge. This is the soft gluon singularity:
the energy carried by the emitted gluon is vanishing,
$x_{\mathrm{g}} = x' - x = (1-z) x' = (1-z) x/z \to 0$ for $z \to 1$.
In order to calculate the integral over $z$ in $S_b$, an upper
cut-off $z_{\mathrm{max}} = x/(x + x_{\epsilon})$ is introduced, i.e. only
branchings with $z \leq z_{\mathrm{max}}$ are included in $S_b$. Here
$x_{\epsilon}$ is a small number, typically chosen so that the
gluon energy $x_{\mathrm{g}} \sqrt{s}/2 \geq x_{\epsilon} \sqrt{s}/2 = 2$
GeV. The average amount of energy carried away by gluons in the
range $x_{g} < x_{\epsilon}$, over the given range of $t$ values
from $t_a$ to $t_b$, may be estimated \cite{Sjo85}. The finally
selected $z$ value may thus be picked as
$z = z_{\mrm{hard}} \langle z_{\mrm{soft}}(t_a, t_b) \rangle$,
where $z_{\mrm{hard}}$
is the originally selected $z$ value and $z_{\mrm{soft}}$ is the
correction factor for soft gluon emission.
In QED showers, the smallness of $\alpha_{\mathrm{em}}$ means that one can
use rather smaller cut-off values without obtaining large amounts of
emission. A fixed small cut-off $x_{\gamma} > 10^{-6}$ is therefore
used to avoid the region of very soft photons. As has been discussed
in section \ref{sss:estructfun}, the electron distribution
inside the electron is cut off at $x_{\mathrm{e}} < 1 - 10^{-6}$, for
numerical reasons, so the two cuts are closely matched.
The cut-off scale $Q_0$ may be chosen separately for QCD and QED
showers, just as in final-state radiation. The defaults are
1 GeV and 0.001 GeV, respectively. The former is the typical hadronic
mass scale, below which radiation is not expected resolvable; the
latter is of the order of the electron mass.
Normally QED and QCD showers do not appear mixed. The most notable
exception is resolved photoproduction (in $\e\p$) and resolved
2$\gamma$ events
(in $\e^+\e^-$), i.e. shower histories of the type $\mathrm{e} \to \gamma \to \mathrm{q}$.
Here the $Q^2$ scales need not be ordered at the interface, i.e.
the last $\mathrm{e} \to \mathrm{e}\gamma$ branching may well have a larger $Q^2$
than the first $\mathrm{q} \to \mathrm{q} \mathrm{g}$ one, and the branching $\gamma \to \mathrm{q}$
does not even have a strict parton-shower interpretation for the
vector dominance model part of the photon parton distribution.
These issues are currently
not addressed in full. Rather, based on the $x$ selected for the
parton (quark or gluon) at the hard scattering, the $x_{\gamma}$
is selected once and for all in the range $x < x_{\gamma} <1$,
according to the distribution implied by eq.~(\ref{pg:foldqgine}).
The QCD parton shower is then traced backwards from the hard
scattering to the QCD shower initiator at $t_0$. No attempt is
made to perform the full QED shower, but rather the beam remnant
treatment (see section \ref{ss:beamrem}) is used to find the
$\overline{\mathrm{q}}$ (or $\mathrm{g}$) remnant that matches the $\mathrm{q}$ (or $\mathrm{g}$) QCD shower
initiator, with the electron itself considered as a second beam
remnant.
\subsubsection{Transverse evolution}
\label{sss:initshowtrans}
We have above seen that two parton lines may be defined, stretching
back from the hard scattering to the initial incoming hadron
wavefunctions at small $Q^2$. Specifically, all parton flavours
$i$, virtualities $Q^2$ and energy fractions $x$ may be found.
The exact kinematical interpretation of the $x$ variable is not
unique, however. For partons with small virtualities and transverse
momenta, essentially all definitions agree, but differences may
appear for branchings close to the hard scattering.
In first-order QED \cite{Ber85} and in some simple QCD toy models
\cite{Got86}, one may show that the `correct' choice is the
`$\hat{s}$ approach'. Here one requires that $\hat{s} = x_1 x_2 s$,
both at the hard scattering scale and at any lower scale, i.e.
$\hat{s}(Q^2) = x_1(Q^2) \, x_2(Q^2) \, s$, where $x_1$ and $x_2$ are
the $x$ values of the two resolved partons (one from each incoming
beam particle) at the given $Q^2$ scale. In practice this means
that, at a branching with the splitting variable $z$, the total
$\hat{s}$ has to be increased by a factor $1/z$ in the backwards
evolution. It also means that branchings on the two incoming legs
have to be interleaved in a single monotonic sequence of $Q^2$
values of branchings.
For a reconstruction of the complete kinematics in this approach,
one should start with the hard scattering, for which $\hat{s}$
has been chosen according to the hard scattering matrix element.
By backwards evolution, the virtualities $Q_1^2 = -m_1^2$ and
$Q_2^2 = -m_2^2$ of
the two interacting partons are reconstructed. Initially the two
partons are considered in their common c.m. frame, coming in along
the $\pm z$ directions. Then the four-momentum vectors have the
non-vanishing components
\begin{eqnarray}
E_{1,2} & = & \frac{ \hat{s} \pm (Q_2^2 - Q_1^2)}{2 \sqrt{\hat{s}}} ~,
\nonumber \\
p_{z1} = - p_{z2} & = & \sqrt{ \frac{ (\hat{s} + Q_1^2 + Q_2^2 )^2
- 4 Q_1^2 Q_2^2 }{ 4 \hat{s} } } ~,
\end{eqnarray}
with $(p_1 + p_2)^2 = \hat{s}$.
If, say, $Q_1^2 > Q_2^2$, then the branching $3 \to 1 + 4$, which
produced parton 1, is the one that took place closest to the hard
scattering, and the one to be reconstructed first. With the
four-momentum $p_3$ known, $p_4 = p_3 - p_1$ is automatically
known, so there are four degrees of freedom. One corresponds to
a trivial azimuthal angle around the $z$ axis. The $z$ splitting
variable for the $3 \to 1 + 4$ vertex is found as the same time as
$Q_1^2$, and provides the constraint $(p_3 + p_2)^2 = \hat{s}/z$.
The virtuality $Q_3^2$ is given by backwards evolution of parton 3.
One degree of freedom remains to be specified, and this is related
to the possibility that parton 4 initiates a time-like parton shower,
i.e. may have a non-zero mass. The maximum allowed squared mass
$m_{\mathrm{max},4}^2$ is found for a collinear branching $3 \to 1 + 4$.
In terms of the combinations
\begin{eqnarray}
s_1 & = & \hat{s} + Q_2^2 + Q_1^2 ~, \nonumber \\
s_3 & = & \frac{\hat{s}}{z} + Q_2^2 + Q_3^2 ~, \nonumber \\
r_1 & = & \sqrt{s_1^2 - 4 Q_2^2 Q_1^2} ~, \nonumber \\
r_3 & = & \sqrt{s_3^2 - 4 Q_2^2 Q_3^2} ~,
\end{eqnarray}
one obtains
\begin{equation}
m_{\mathrm{max},4}^2 = \frac{s_1 s_3 - r_1 r_3}{2 Q_2^2} - Q_1^2 - Q_3^2 ~,
\end{equation}
which, for the special case of $Q_2^2 = 0$, reduces to
\begin{equation}
m_{\mathrm{max},4}^2 = \left\{ \frac{Q_1^2}{z} - Q_3^2 \right\}
\left\{ \frac{\hat{s}}{\hat{s} + Q_1^2} -
\frac{\hat{s}}{\hat{s}/z + Q_3^2} \right\} ~.
\label{sh:zrangespace}
\end{equation}
These constraints on $m_4$ are only the kinematical ones, in
addition coherence phenomena could constrain the $m_{\mathrm{max},4}$
values further. Some options of this kind are available; the
default one is to require additionally that $m_4^2 \leq Q_1^2$,
i.e. lesser than the space-like virtuality of the sister parton.
With the maximum virtuality given, the final-state showering
machinery may be used to give the development of the subsequent
cascade, including the actual mass $m_4^2$, with
$0 \leq m_4^2 \leq m_{\mathrm{max},4}^2$. The evolution is performed in
the c.m. frame of the two `resolved' partons, i.e. that of
partons 1 and 2 for the
branching $3 \to 1 + 4$, and parton 4 is assumed to have a nominal
energy $E_{\mrm{nom},4} = (1/z - 1) \sqrt{\hat{s}}/2$. (Slight
modifications appear if parton 4 has a non-vanishing mass
$m_{\mathrm{q}}$ or $m_{\ell}$.)
Using the relation $m_4^2 = (p_3 - p_1)^2$, the momentum of parton
3 may now be found as
\begin{eqnarray}
E_3 & = & \frac{1}{2 \sqrt{\hat{s}} } \left\{ \frac{\hat{s}}{z}
+ Q_2^2 - Q_1^2 - m_4^2 \right\} ~, \nonumber \\
p_{z3} & = & \frac{1}{2 p_{z1}} \left\{ s_3 - 2 E_2 E_3 \right\} ~,
\nonumber \\
p_{\perp ,3}^2 & = & \left\{ m_{\mathrm{max},4}^2 - m_4^2 \right\} \,
\frac{ (s_1 s_3 + r_1 r_3)/2 - Q_2^2 (Q_1^2 + Q_3^2 + m_4^2)}{r_1^2} ~.
\end{eqnarray}
The requirement that $m_4^2 \geq 0$ (or $\geq m_f^2$ for heavy
flavours) imposes a constraint on allowed $z$ values. This constraint
cannot be included in the choice of $Q_1^2$, where it logically
belongs, since it also depends on $Q_2^2$ and $Q_3^2$, which are
unknown at this point. It is fairly rare (in the order of 10\% of all
events) that an unallowed $z$ value is generated, and when it happens
it is almost always for one of the two branchings closest to the
hard interaction: for $Q_2^2 = 0$ eq.~(\ref{sh:zrangespace}) may be
solved to yield $z \leq \hat{s}/(\hat{s} + Q_1^2 - Q_3^2)$, which is
a more severe cut for $\hat{s}$ small and $Q_1^2$ large. Therefore
an essentially bias-free way of coping is to redo completely any
initial-state cascade for which this problem appears.
This completes the reconstruction of the $3 \to 1 + 4$ vertex.
The subsystem made out of partons 3 and 2 may now be boosted to its
rest frame and rotated to bring partons 3 and 2 along the $\pm z$
directions. The partons 1 and 4 now have opposite and compensating
transverse momenta with respect to the event axis. When the next
vertex is considered, either the one that produces parton 3 or the
one that produces parton 2, the 3--2 subsystem will fill the function
the 1--2 system did above, e.g. the r\^ole of
$\hat{s} = \hat{s}_{12}$ in the formulae above is now played by
$\hat{s}_{32} = \hat{s}_{12}/z$. The internal structure of the
3--2 system, i.e. the branching $3 \to 1 + 4$, appears nowhere
in the continued description, but has become
`unresolved'. It is only reflected in the successive rotations and
boosts performed to bring back the new endpoints to their common
rest frame. Thereby the hard scattering subsystem 1--2 builds up
a net transverse momentum and also an overall rotation of the
hard scattering subsystem.
After a number of steps, the two outermost partons have virtualities
$Q^2 < Q_0^2$ and then the shower is terminated and the endpoints
assigned $Q^2 = 0$. Up to small corrections from primordial
$k_{\perp}$, discussed in section \ref{ss:beamrem}, a final boost
will bring the partons from their c.m. frame to the overall c.m.
frame, where the $x$ values of the outermost partons agree also
with the light-cone definition.
\subsubsection{Other initial-state shower aspects}
In the formulae above, $Q^2$ has been used as argument for
$\alpha_{\mathrm{s}}$, and not only as the space-like virtuality of partons.
This is one possibility, but in fact loop calculations tend to
indicate that the proper argument for $\alpha_{\mathrm{s}}$ is not $Q^2$
but $p_{\perp}^2 = (1-z) Q^2$ \cite{Bas83}. The variable $p_{\perp}$ does
have the interpretation of transverse momentum, although it
is only exactly so for a branching $a \to bc$
with $a$ and $c$ massless and $Q^2 = - m_b^2$, and with $z$
interpreted as light-cone fraction of energy and momentum.
The use of $\alpha_{\mathrm{s}}((1-z)Q^2)$ is default in the program.
Indeed, if one wanted to, the complete shower might be
interpreted as an evolution in $p_{\perp}^2$ rather than in $Q^2$.
As we see, the initial-state showering algorithm leads to a
net boost and rotation of the hard scattering subsystems. The
overall final state is made even more complex by the additional
final-state radiation. In principle, the complexity is very
physical, but it may still have undesirable side effects. One
such, discussed further in section \ref{ss:PYswitchkin}, is that
it is very difficult to generate events that fulfill specific
kinematics conditions, since kinematics is smeared and even,
at times, ambiguous.
A special case is encountered in deep inelastic scattering in
$\e\p$ collisions. Here the DIS $x$ and $Q^2$ values are defined
in terms of the scattered electron direction and energy, and
therefore are unambiguous (except for issues of final-state
photon radiation close to the electron direction).
Neither initial- nor final-state showers preserve the kinematics
of the scattered electron, however, and hence the DIS $x$ and
$Q^2$ are changed. In principle, this is perfectly legitimate,
with the caveat that one then also should use different sets
of parton distributions than ones derived from DIS, since these
are based on the kinematics of the scattered lepton and nothing
else. Alternatively, one might consider showering schemes that
leave $x$ and $Q^2$ unchanged. In \cite{Ben88} detailed
modifications are presented that make a preservation
possible when radiation off the incoming and outgoing electron is
neglected, but these are not included in the current version of
{\tsc{Pythia}}.
What is available, as an option, is a simple machinery
which preserves $x$ and $Q^2$ from the effects of QCD radiation,
and also from those of primordial $k_{\perp}$ and the beam remnant
treatment, as follows. After the showers have been generated,
the four-momentum of the scattered lepton is changed to the
expected one, based on the nominal $x$ and $Q^2$ values.
The azimuthal angle of the lepton is maintained when the transverse
momentum is adjusted. Photon radiation off the lepton leg is not fully
accounted for, i.e. it is assumed that the energy of final-state
photons is added to that of the scattered electron for the definition
of $x$ and $Q^2$ (this is the normal procedure for parton-distribution
definitions).
The change of three-momentum on the lepton side of the event is
balanced by the final state partons on the hadron side, excluding
the beam remnant but including all the partons both from initial-
and final-state showering. The fraction of three-momentum shift taken
by each parton is proportional to its original light-cone momentum
in the direction of the incoming lepton, i.e. to $E \mp p_z$ for a
hadron moving in the $\pm$ direction. This procedure guarantees
momentum but not energy conservation. For the latter, one additional
degree of freedom is needed, which is taken to be the longitudinal
momentum of the initial state shower initiator. As this momentum is
modified, the change is shared by the final state partons on the
hadron side, according to the same light-cone fractions as before
(based on the original momenta). Energy conservation
requires that the total change in final state parton energies plus
the change in lepton side energy equals the change in initiator
energy. This condition can be turned into an iterative procedure to
find the initiator momentum shift.
Sometimes the procedure may break down. For instance, an initiator
with $x > 1$ may be reconstructed. If this should happen, the $x$ and
$Q^2$ values of the event are preserved, but new initial and final
state showers are generated. After five such failures, the event is
completely discared in favour of a new kinematical setup.
Kindly note that the four-momentum of intermediate partons in the
shower history are not being adjusted. In a listing of the complete
event history, energy and momentum need then not be conserved in
shower branchings. This mismatch could be fixed up, if need be.
The scheme presented above should not be taken too
literally, but is rather intended as a contrast to the more
sophisticated schemes already on the market, if one would like to
understand whether the kind of conservation scheme chosen does affect
the observable physics.
\subsection{Routines and Common Block Variables}
\label{ss:showrout}
In this section we collect information on how to use the initial-
and final-state showering routines. Of these \ttt{LUSHOW} for
final-state radiation is the more generally interesting, since it
can be called to let a user-defined parton configuration shower.
\ttt{PYSSPA}, on the other hand, is so intertwined with the general
structure of a {\tsc{Pythia}} event that it is of little use as a
stand-alone product.
\drawbox{CALL LUSHOW(IP1,IP2,QMAX)}\label{p:LUSHOW}
\begin{entry}
\itemc{Purpose:} to generate time-like parton showers, conventional
or coherent. The performance of the program is regulated by the
switches \ttt{MSTJ(40) - MSTJ(50)} and parameters
\ttt{PARJ(81) - PARJ(89)}. In order to keep track of the colour
flow information, the positions \ttt{K(I,4)} and \ttt{K(I,5)} have
to be organized properly for showering partons. Inside the {\tsc{Jetset/Pythia}}
programs, this is done automatically, but for external use
proper care must be taken.
\iteme{IP1 > 0, IP2 = 0 :} generate a time-like parton shower for the
parton in line \ttt{IP1} in common block \ttt{LUJETS}, with maximum
allowed mass \ttt{QMAX}. With only one parton at hand, one cannot
simultaneously conserve both energy and momentum: we here choose to
conserve energy and jet direction, while longitudinal momentum (along
the jet axis) is not conserved.
\iteme{IP1 > 0, IP2 > 0 :} generate time-like parton showers for the
two partons in lines \ttt{IP1} and \ttt{IP2} in the common block
\ttt{LUJETS}, with maximum allowed mass for each parton \ttt{QMAX}.
For shower evolution, the two partons are boosted to their c.m. frame.
Energy and momentum is conserved for the pair of partons, although
not for each individually. One of the two partons may be
replaced by a nonradiating particle, such as a photon or a
diquark; the energy and momentum of this particle will then be
modified to conserve the total energy and momentum.
\iteme{IP1 > 0, IP2 < 0 :} generate time-like parton showers for the
\ttt{-IP2} (at most 3) partons in lines \ttt{IP1}, \ttt{IP1+1}, \ldots
\ttt{IPI-IP2-1} in the common block \ttt{LUJETS}, with maximum allowed
mass for each parton \ttt{QMAX}. The actions for \ttt{IP2=-1} and
\ttt{IP2=-2} correspond to what is described above, but additionally
\ttt{IP2=-3} may be used to generate the evolution starting from
three given partons (e.g. in $\Upsilon \to \mathrm{g}\g\mathrm{g}$). Then the three
partons are boosted to their c.m. frame, energy is conserved for each
parton individually and momentum for the system as a whole.
\iteme{QMAX :} the maximum allowed mass of a radiating parton, i.e.
the starting value for the subsequent evolution. (In addition, the
mass of a single parton may not exceed its energy, the mass of a
parton in a system may not exceed the invariant mass of the
system.)
\end{entry}
\boxsep
\begin{entry}
\iteme{SUBROUTINE PYSSPA(IPU1,IPU2) :}\label{p:PYSSPA} to generate
the space-like showers of the initial-state radiation.
\end{entry}
\drawbox{COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)}
\begin{entry}
\itemc{Purpose:} to give access to a number of status codes and
parameters which regulate the performance of {\tsc{Jetset}}.
Most parameters are described in section \ref{ss:JETswitch};
here only those related to \ttt{LUSHOW} are described.
\boxsep
\iteme{MSTJ(40) :}\label{p:MSTJ40} (D=0) possibility to suppress the
branching probability for a branching $\mathrm{q} \to \mathrm{q}\mathrm{g}$ (or
$\mathrm{q} \to \mathrm{q}\gamma$) of a quark produced in the decay of an unstable
particle with width $\Gamma$, where this width has to
be specified by the user in \ttt{PARJ(89)}. The algorithm used is not
exact, but still gives some impression of potential effects. This
switch ought to have appeared at the end of the current list of shower
switches (after \ttt{MSTJ(50)}), but because of lack of
space it appears immediately before.
\begin{subentry}
\iteme{= 0 :} no suppression, i.e. the standard parton-shower machinery.
\iteme{= 1 :} suppress radiation by a factor
$\chi(\omega) = \Gamma^2 / (\Gamma^2 + \omega^2)$, where $\omega$ is
the energy of the gluon (or photon) in the rest frame of the radiating
dipole. Essentially this means that hard radiation with
$\omega > \Gamma$ is removed.
\iteme{= 2 :} suppress radiation by a factor
$1 - \chi(\omega) = \omega^2 / (\Gamma^2 + \omega^2)$, where $\omega$
is the energy of the gluon (or photon) in the rest frame of the
radiating dipole. Essentially this means that soft radiation with
$\omega < \Gamma$ is removed.
\end{subentry}
\iteme{MSTJ(41) :} (D=2) type of branchings allowed in shower.
\begin{subentry}
\iteme{= 0 :} no branchings at all, i.e. shower is switched off.
\iteme{= 1 :} QCD type branchings of quarks and gluons.
\iteme{= 2 :} also emission of photons off quarks and leptons; the
photons are assumed on the mass shell.
\iteme{= 10 :} as \ttt{=2}, but enhance photon emission by a factor
\ttt{PARJ(84)}. This option is unphysical, but for moderate values,
\ttt{PARJ(84)}$\leq 10$, it may be used to enhance the prompt photon
signal in $\mathrm{q}\overline{\mathrm{q}}$ events. The normalization of the prompt photon
rate should then be scaled down by the same factor. The dangers
of an improper use are significant, so do not use this option if you
do not know what you are doing.
\end{subentry}
\iteme{MSTJ(42) :} (D=2) branching mode for time-like showers.
\begin{subentry}
\iteme{= 1 :} conventional branching, i.e. without angular ordering.
\iteme{= 2 :} coherent branching, i.e. with angular ordering.
\end{subentry}
\iteme{MSTJ(43) :} (D=4) choice of $z$ definition in branching.
\begin{subentry}
\iteme{= 1 :} energy fraction in grandmother's rest frame (`local,
constrained').
\iteme{= 2 :} energy fraction in grandmother's rest frame assuming
massless daughters, with energy and momentum reshuffled for massive
ones (`local, unconstrained').
\iteme{= 3 :} energy fraction in c.m. frame of the showering partons
(`global, constrained').
\iteme{= 4 :} energy fraction in c.m. frame of the showering partons
assuming massless daughters, with energy and momentum reshuffled for
massive ones (`global, unconstrained').
\end{subentry}
\iteme{MSTJ(44) :} (D=2) choice of $\alpha_{\mathrm{s}}$ scale for shower.
\begin{subentry}
\iteme{= 0 :} fixed at \ttt{PARU(111)} value.
\iteme{= 1 :} running with $Q^2 = m^2/4$, $m$ mass of decaying
parton, $\Lambda$ as stored in \ttt{PARJ(81)} (natural choice for
conventional showers).
\iteme{= 2 :} running with $Q^2 = z(1-z)m^2$, i.e. roughly $p_{\perp}^2$
of branching, $\Lambda$ as stored in \ttt{PARJ(81)} (natural choice
for coherent showers).
\end{subentry}
\iteme{MSTJ(45) :} (D=5) maximum flavour that can be produced in
shower by $\mathrm{g} \to \mathrm{q} \overline{\mathrm{q}}$; also used to determine the maximum
number of active flavours in the $\alpha_{\mathrm{s}}$ factor in parton showers
(here with a minimum of 3).
\iteme{MSTJ(46) :} (D=3) nonhomogeneous azimuthal distributions in
a shower branching.
\begin{subentry}
\iteme{= 0 :} azimuthal angle is chosen uniformly.
\iteme{= 1 :} nonhomogeneous azimuthal angle in gluon decays due to
a kinematics-dependent effective gluon polarization.
Not meaningful for scalar model, i.e. then same as \ttt{=0}.
\iteme{= 2 :} nonhomogeneous azimuthal angle in gluon decay due to
interference with nearest neighbour (in colour).
Not meaningful for Abelian model, i.e. then same as \ttt{=0}.
\iteme{= 3 :} nonhomogeneous azimuthal angle in gluon decay due to
both polarization (\ttt{=1}) and interference (\ttt{=2}).
Not meaningful for Abelian model, i.e. then same as \ttt{=1}.
Not meaningful for scalar model, i.e. then same as \ttt{=2}.
\end{subentry}
\iteme{MSTJ(47) :} (D=3) corrections to the lowest-order $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}$,
$\mathrm{q}\overline{\mathrm{q}}\gamma$, $\ell^+\ell^-\gamma$ or $\ell\nu_{\ell}\gamma$
3-parton matrix element at the first branching of either initial
parton in a shower.
\begin{subentry}
\iteme{= 0 :} no corrections.
\iteme{= 1 :} included whenever scattered partons are $\mathrm{q}\overline{\mathrm{q}}$,
$\ell^+\ell^-$ or $\ell\nu_{\ell}$.
\iteme{= 2 :} always included when shower starts from two partons.
\iteme{= 3 :} as \ttt{=1} except that for massive quarks also the
massive matrix element expression is used, eq.~(\ref{ee:threejMEmass}),
while \ttt{=1} is always based on massless matrix elements.
\iteme{= 4 :} as \ttt{=2} except that for massive quarks also the
massive matrix element expression is used, while \ttt{=2} is always
based on massless matrix elements.
\end{subentry}
\iteme{MSTJ(48) :} (D=0) possibility to impose maximum angle for the
first branching in a shower.
\begin{subentry}
\iteme{= 0 :} no explicit maximum angle.
\iteme{= 1 :} maximum angle given by \ttt{PARJ(85)} for single
showering parton, by \ttt{PARJ(85)} and \ttt{PARJ(86)} for pair
of showering partons.
\end{subentry}
\iteme{MSTJ(49) :} (D=0) possibility to change the branching
probabilities according to some alternative toy models (note that
the $Q^2$ evolution of $\alpha_{\mathrm{s}}$ may well be different in these
models, but that only the \ttt{MSTJ(44)} options are at the
disposal of the user).
\begin{subentry}
\iteme{= 0 :} standard QCD branchings.
\iteme{= 1 :} branchings according to a scalar gluon theory, i.e. the
splitting kernels in the evolution equations are,
with a common factor $\alpha_{\mathrm{s}}/(2\pi)$ omitted,
$P_{\mathrm{q} \to \mathrm{q}\mathrm{g}} = (2/3) (1-z)$, $P_{\mathrm{g} \to \mathrm{g}\g} =$ \ttt{PARJ(87)},
$P_{\mathrm{g} \to \mathrm{q}\overline{\mathrm{q}}} =$ \ttt{PARJ(88)} (for each separate flavour).
The couplings of the gluon have been left as free parameters,
since they depend on the colour structure assumed. Note that,
since a spin 0 object decays isotropically, the gluon splitting
kernels contain no $z$ dependence.
\iteme{= 2 :} branchings according to an Abelian vector gluon theory,
i.e. the colour factors are changed (compared with QCD) according to
$C_F = 4/3 \to 1$, $N_C = 3 \to 0$, $T_R = 1/2 \to 3$. Note that an
Abelian model is not expected to contain any coherence effects
between gluons, so that one should normally use \ttt{MSTJ(42)=1} and
\ttt{MSTJ(46)=} 0 or 1. Also, $\alpha_{\mathrm{s}}$ is expected to increase
with increasing $Q^2$ scale, rather than decrease. No such
$\alpha_{\mathrm{s}}$ option is available; the one that comes closest
is \ttt{MSTJ(44)=0}, i.e. a fix value.
\end{subentry}
\iteme{MSTJ(50) :} (D=3) possibility to introduce colour coherence
effects in the first branching of a final state shower; mainly
of relevance for QCD parton--parton scattering processes.
\begin{subentry}
\iteme{= 0 :} none.
\iteme{= 1 :} impose an azimuthal anisotropy.
\iteme{= 2 :} restrict the polar ange of a branching to be smaller
than the scattering angle of the relevant colour flow.
\iteme{= 3 :} both azimuthal anisotropy and restricted polar angles.
\itemc{Note:} for subsequent branchings the (polar) angular ordering
is automatic (\ttt{MSTP(42)=2}) and \ttt{MSTJ(46)=3}).
\end{subentry}
\boxsep
\iteme{PARJ(81) :}\label{p:PARJ81} (D=0.29 GeV) $\Lambda$ value
in running $\alpha_{\mathrm{s}}$ for parton showers (see \ttt{MSTJ(44)}). This is
used in all user calls to \ttt{LUSHOW}, in the $\e^+\e^-$ routines of {\tsc{Jetset}},
and in a {\tsc{Pythia}} (or {\tsc{Jetset}}) resonance decay. It is not intended for
other timelike showers in {\tsc{Pythia}}, however, for which \ttt{PARP(72)}
is used.
\iteme{PARJ(82) :} (D=1.0 GeV) invariant mass cut-off $m_{\mathrm{min}}$ of
parton showers, below which partons are not assumed to radiate.
For $Q^2 = p_{\perp}^2$ (\ttt{MSTJ(44)=2}) \ttt{PARJ(82)}/2
additionally gives the minimum $p_{\perp}$ of a branching. To avoid
infinite $\alpha_{\mathrm{s}}$ values, one must have
\ttt{PARJ(82)}$ > 2 \times$\ttt{PARJ(81)} for \ttt{MSTJ(44)}$\geq 1$
(this is automatically checked in the program, with
$2.2 \times$\ttt{PARJ(81)} as the lowest value attainable).
\iteme{PARJ(83) :} (D=1.0 GeV) invariant mass cut-off $m_{\mathrm{min}}$ used
for photon emission in parton showers, below which quarks and leptons
are not assumed to radiate. The function of \ttt{PARJ(83)} closely
parallels that of \ttt{PARJ(82)} for QCD branchings, but there is a
priori no requirement that the two be equal.
\iteme{PARJ(84) :} (D=1.) used for option \ttt{MSTJ(41)=10} as a
multiplicative factor in the promt photon emission rate in final
state parton showers. Unphysical but useful technical trick, so
beware!
\iteme{PARJ(85), PARJ(86) :} (D=10.,10.) maximum opening angles
allowed in the first branching of parton showers; see \ttt{MSTJ(48)}.
\iteme{PARJ(87) :} (D=0.) coupling of $\mathrm{g} \to \mathrm{g}\g$ in scalar gluon
shower, see \ttt{MSTJ(49)=1}.
\iteme{PARJ(88) :} (D=0.) coupling of $\mathrm{g} \to \mathrm{q}\overline{\mathrm{q}}$ in scalar
gluon shower (per quark species), see \ttt{MSTJ(49)=1}.
\iteme{PARJ(89) :} (D=0. GeV) the width of the unstable particle studied
for the \ttt{MSTJ(40) > 0} options; to be set by the user (separately
for each \ttt{LUSHOW} call, if need be).
\end{entry}
\drawbox{COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)}
\begin{entry}
\itemc{Purpose:} to give access to status code and parameters which
regulate the performance of {\tsc{Pythia}}.
Most parameters are described in section \ref{ss:PYswitchpar};
here only those related to \ttt{PYSSPA} and \ttt{LUSHOW} are
described.
\iteme{MSTP(22) :}\label{p:MSTP22} (D=0) special override of normal
$Q^2$ definition used for maximum of parton-shower evolution. This
option only affects processes 10 and 83 (deep inelastic scattering)
and only in lepton--hadron events.
\begin{subentry}
\iteme{= 0 :} use the scale as given in \ttt{MSTP(32)}.
\iteme{= 1 :} use the DIS $Q^2$ scale, i.e. $-\hat{t}$.
\iteme{= 2 :} use the DIS $W^2$ scale, i.e. $(-\hat{t})(1-x)/x$.
\iteme{= 3 :} use the DIS $Q \times W$ scale, i.e.
$(-\hat{t}) \sqrt{(1-x)/x}$.
\iteme{= 4 :} use the scale $Q^2 (1-x) \max(1, \ln(1/x))$, as
motivated by first order matrix elements \cite{Ing80,Alt78}.
\itemc{Note:} in all of these alternatives, a multiplicative factor is
introduced by \ttt{PARP(67)} and \ttt{PARP(71)}, as usual.
\end{subentry}
\iteme{MSTP(61) :}\label{p:MSTP61} (D=1) master switch for
initial-state QCD and QED radiation.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\end{subentry}
\iteme{MSTP(62) :} (D=3) level of coherence imposed on the
space-like parton-shower evolution.
\begin{subentry}
\iteme{= 1 :} none, i.e. neither $Q^2$ values nor angles need be
ordered.
\iteme{= 2 :} $Q^2$ values at branches are strictly ordered,
increasing towards the hard interaction.
\iteme{= 3 :} $Q^2$ values and opening angles of emitted
(on-mass-shell or time-like) partons are both strictly ordered,
increasing towards the hard interaction.
\end{subentry}
\iteme{MSTP(63) :} (D=2) structure of associated time-like
showers, i.e. showers initiated by emission off the incoming
space-like partons.
\begin{subentry}
\iteme{= 0 :} no associated showers are allowed, i.e. emitted
partons are put on the mass shell.
\iteme{= 1 :} a shower may evolve, with maximum allowed time-like
virtuality set by the phase space only.
\iteme{= 2 :} a shower may evolve, with maximum allowed time-like
virtuality set by phase space or by \ttt{PARP(71)} times the $Q^2$
value of the space-like parton created in the same vertex, whichever
is the stronger constraint.
\end{subentry}
\iteme{MSTP(64) :} (D=2) choice of $\alpha_{\mathrm{s}}$ and $Q^2$ scale
in space-like parton showers.
\begin{subentry}
\iteme{= 0 :} $\alpha_{\mathrm{s}}$ is taken to be fix at the value
\ttt{PARU(111)}.
\iteme{= 1 :} first-order running $\alpha_{\mathrm{s}}$ with argument
\ttt{PARP(63)}$Q^2$.
\iteme{= 2 :} first-order running $\alpha_{\mathrm{s}}$ with argument
\ttt{PARP(64)}$k_{\perp}^2 = $\ttt{PARP(64)}$(1-z)Q^2$.
\end{subentry}
\iteme{MSTP(65) :} (D=1) treatment of soft gluon emission in
space-like parton-shower evolution.
\begin{subentry}
\iteme{= 0 :} soft gluons are entirely neglected.
\iteme{= 1 :} soft gluon emission is resummed and included
together with the hard radiation as an effective $z$ shift.
\end{subentry}
\iteme{MSTP(66) :} (D=1) choice of lower cut-off for initial-state
QCD radiation in anomalous photoproduction events
(see \ttt{MSTP(14)=3}).
\begin{subentry}
\iteme{= 0 :} the lower $Q^2$ cutoff is the standard one in
\ttt{PARP(62)}$^2$.
\iteme{= 1 :} the lower cutoff is the larger of \ttt{PARP(62)}$^2$
and \ttt{VINT(283)} or \ttt{VINT(284)}, where the latter is the
virtuality scale of the $\gamma \to \mathrm{q}\overline{\mathrm{q}}$ vertex on the appropriate
side of the event.
\end{subentry}
\iteme{MSTP(67) :} (D=2) possibility to introduce colour coherence
effects in the first branching of the backwards evolution of an
initial state shower; mainly of relevance for QCD parton--parton
scattering processes.
\begin{subentry}
\iteme{= 0 :} none.
\iteme{= 2 :} restrict the polar angle of a branching to be smaller
than the scattering angle of the relevant colour flow.
\itemc{Note 1:} azimuthal anisotropies have not yet been included.
\itemc{Note 2:} for subsequent branchings, \ttt{MSTP(62)=3} is
used to restrict the (polar) angular range of branchings.
\end{subentry}
\iteme{MSTP(71) :} (D=1) master switch for final-state QCD and
QED radiation.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\end{subentry}
\boxsep
\iteme{PARP(61) :}\label{p:PARP61} (D=0.25 GeV) $\Lambda$ value
used in space-like parton shower (see \ttt{MSTP(64)}). This value
may be overwritten, see \ttt{MSTP(3)}.
\iteme{PARP(62) :} (D=1. GeV) effective cut-off $Q$ or
$k_{\perp}$ value (see \ttt{MSTP(64)}), below which space-like
parton showers are not evolved.
\iteme{PARP(63) :} (D=0.25) in space-like shower evolution the
virtuality $Q^2$ of a parton is multiplied by \ttt{PARP(63)} for use
as a scale in $\alpha_{\mathrm{s}}$ and parton distributions when
\ttt{MSTP(64)=1}.
\iteme{PARP(64) :} (D=1.) in space-like parton-shower evolution
the squared transverse momentum evolution scale $k_{\perp}^2$ is
multiplied by \ttt{PARP(64)} for use as a scale in $\alpha_{\mathrm{s}}$ and
parton distributions when \ttt{MSTP(64)=2}.
\iteme{PARP(65) :} (D=2. GeV) effective minimum energy (in c.m.
frame) of time-like or on-shell parton emitted in space-like shower;
see also \ttt{PARP(66)}.
\iteme{PARP(66) :} (D=0.001) effective lower cut-off on $1-z$ in
space-like showers, in addition to the cut implied by \ttt{PARP(65)}.
\iteme{PARP(67) :} (D=4.) the $Q^2$ scale of the hard scattering
(see \ttt{MSTP(32)}) is multiplied by \ttt{PARP(67)} to define the
maximum parton virtuality allowed in space-like showers. This does not
apply to $s$-channel resonances, where the maximum virtuality is set
by $m^2$.
\iteme{PARP(68) :} (D=1E-3) lower $Q$ cut-off for QED space-like
showers.
\iteme{PARP(71) :} (D=4.) the $Q^2$ scale of the hard scattering
(see \ttt{MSTP(32)}) is multiplied by \ttt{PARP(71)} to define the
maximum parton virtuality allowed in time-like showers. This does not
apply to $s$-channel resonances, where the maximum virtuality is set
by $m^2$.
\iteme{PARP(72) :} (D=0.25 GeV) $\Lambda$ value used in running
$\alpha_{\mathrm{s}}$ for timelike parton showers, except for showers in the
decay of a resonance. (Resonance decay, e.g. $\gamma^*/Z^0$ decay,
is instead set by \ttt{PARJ(81)}.)
\end{entry}
\clearpage
\section{Beam Remnants and Underlying Events}
Each incoming beam particle may leave behind a beam remnant, which
does not take part in the initial-state radiation or hard scattering
process. If nothing else, the remnants need be reconstructed and
connected to the rest of the event. In hadron--hadron collisions,
the composite nature of the two incoming beam particles implies the
additional possibility that several parton pairs undergo separate
hard or semi-hard scatterings, `multiple interactions'.
This may give a non-negligible
contribution to the `underlying event' structure, and thus to the
total multiplicity. Finally, in high-luminosity colliders, it is
possible to have several collisions between beam particles in one
and the same beam crossing, i.e. pile-up events, which further act
to build up the
general particle production activity that is to be observed by
detectors. These three aspects are described in turn, with emphasis
on the middle one, that of multiple interactions within a single
hadron--hadron collision.
The main reference on the multiple interactions model is
\cite{Sjo87a}.
\subsection{Beam Remnants}
\label{ss:beamrem}
The initial-state radiation algorithm reconstructs one shower
initiator in each beam. (If initial-state radiation is not included,
the initiator is nothing but the incoming parton to the hard
interaction.) Together the two initiators delineate an
interaction subsystem, which contains all the partons that
participate in the initial-state showers, in the hard interaction,
and in the final-state showers. Left behind are two beam remnants
which, to first approximation, just sail through, unaffected by the
hard process. (The issue of additional interactions is covered in
the next section.)
A description of the beam remnant structure contains a few components.
First, given the flavour content of a (colour-singlet) beam particle,
and the flavour and colour of the initiator parton, it is possible
to reconstruct the flavour and colour of the
beam remnant. Sometimes the remnant may be
represented by just a single parton or diquark, but often the
remnant has to be subdivided into two separate objects. In the latter
case it is necessary to share the remnant energy and momentum between
the two. Due to Fermi motion inside hadron beams, the initiator parton
may have a `primordial $k_{\perp}$' transverse momentum motion,
which has to be compensated by the beam remnant. If the remnant is
subdivided, there may also be a relative transverse momentum.
In the end, total energy and momentum has to be conserved.
To first approximation, this is ensured within each remnant
separately, but some final global adjustments are necessary to
compensate for the
primordial $k_{\perp}$ and any effective beam remnant mass.
Consider first a proton (or, with trivial modifications, any other
baryon or antibaryon).
\begin{Itemize}
\item If the initiator parton is a $\u$ or $\d$ quark, it is
assumed to be a valence quark, and therefore leaves behind a
diquark beam remnant, i.e. either a $\u\d$ or a $\u\u$ diquark,
in a colour antitriplet state.
Relative probabilities for different diquark spins are derived within
the context of the non-relativistic {\bf SU(6)} model, i.e. flavour
{\bf SU(3)} times spin {\bf SU(2)}. Thus a $\u\d$ is $3/4$ $\u\d_0$
and $1/4$ $\u\d_1$, while a $\u\u$ is always $\u\u_1$.
\item An initiator gluon leaves behind a colour octet $\u\u\d$
state, which is subdivided into a colour triplet quark and a colour
antitriplet diquark. {\bf SU(6)} gives the appropriate subdivision,
$1/2$ of the time into $\u + \u\d_0$, $1/6$ into $\u + \u\d_1$ and
$1/3$ into $\d + \u\u_1$.
\item A sea quark initiator, such as an $\mathrm{s}$, leaves behind a
$\u\u\d\overline{\mathrm{s}}$ four-quark state. The PDG flavour coding scheme and
the fragmentation routines do not foresee such a state, so therefore
it is subdivided into a meson plus a diquark, i.e. $1/2$ into
$\u\overline{\mathrm{s}} + \u\d_0$, $1/6$ into $\u\overline{\mathrm{s}} + \u\d_1$ and $1/3$ into
$\d\overline{\mathrm{s}} + \u\u_1$. Once the flavours of the meson are determined,
the choice of meson multiplet is performed as in the standard
fragmentation description.
\item Finally, an antiquark initiator, such as an $\overline{\mathrm{s}}$, leaves
behind a $\u\u\d\mathrm{s}$ four-quark state, which is subdivided into a
baryon plus a quark. Since, to first approximation, the $\mathrm{s}\overline{\mathrm{s}}$
pair comes from the branching $\mathrm{g} \to \mathrm{s}\overline{\mathrm{s}}$ of a colour octet
gluon, the subdivision $\u\u\d + \mathrm{s}$ is not allowed, since it would
correspond to a colour-singlet $\mathrm{s}\overline{\mathrm{s}}$. Therefore the subdivision
is $1/2$ into $\u\d_0\mathrm{s} + \u$, $1/6$ into $\u\d_1\mathrm{s} + \u$ and $1/3$
into $\u\u_1\mathrm{s} + \d$. A baryon is formed among the ones possible for
the given flavour content and diquark spin, according to the relative
probabilities used in the fragmentation. One could argue for an
additional weighting to count the number of baryon states available
for a given diquark plus quark combination, but this has not been
included.
\end{Itemize}
One may note that any $\u$ or $\d$ quark taken out of the proton is
automatically assumed to be a valence quark. Clearly this is
unrealistic,
but not quite as bad as it might seem. In particular, one should
remember that the beam remnant scenario is applied to the
initial-state shower initiators at a scale of $Q_0 \approx 1$ GeV
and at an
$x$ value usually much larger than the $x$ at the hard scattering.
The sea quark contribution therefore normally is negligible.
For a meson beam remnant, the rules are in the same spirit, but
somewhat easier, since no diquark or baryons need be taken into
account. Thus a valence quark (antiquark) initiator leaves
behind a valence antiquark (quark), a gluon initiator leaves
behind a valence quark plus a valence antiquark, and a sea quark
(antiquark) leaves behind a meson (which contains the partner to
the sea parton) plus a valence antiquark (quark).
A resolved photon is even simpler than a meson,
since one does not have to make
the distinction between valence and sea flavour. Thus any quark
(antiquark) initiator leaves behind the matching antiquark (quark),
and a gluon leaves behind a quark + antiquark pair. The relative
quark flavour composition in the latter case is assumed proportional
to $e_{\mathrm{q}}^2$ among light flavours, i.e. $2/3$ into $\u + \overline{\mathrm{u}}$,
$1/6$ into $\d + \overline{\mathrm{d}}$, and $1/6$ into $\mathrm{s} + \overline{\mathrm{s}}$. If one wanted
to, one could also have chosen to represent the remnant by a single
gluon.
If no initial-state radiation is assumed, an electron (or, in general,
a lepton or a neutrino) leaves behind no beam remnant. Also when
radiation is included, one would expect to recover a single electron
with the full beam energy when the shower initiator is reconstructed.
This does not have to happen, e.g. if the initial-state shower is cut
off at a non-vanishing scale, such that some of the emission at low
$Q^2$ values is not simulated. Further, for purely technical reasons,
the distribution of an electron inside an electron,
$f_{\mathrm{e}}^{\mathrm{e}}(x,Q^2)$, is cut off at $x = 1 - 10^{-6}$. This means that
always, when initial-state radiation is included, a fraction
of at least $10^{-6}$ of the beam energy has to be put into one single
photon along the beam direction, to represent this not simulated
radiation. The physics is here slightly different from the standard
beam remnant concept, but it is handled with the same machinery.
Beam remnants can also apper when the electron is resolved with the
use of parton distributions, but initial-state radiation is switched
off. Conceptually, this is a contradiction, since it is the
initial-state radiation that builds up the parton distributions,
but sometimes
the combination is still useful. Finally, since QED radiation has not
yet been included in events with resolved photons inside electrons,
also in this case effective beam remnants have to be assigned by the
program.
The beam remnant assignments inside an electron, in either of the cases
above, is as follows.
\begin{Itemize}
\item An $\mathrm{e}^-$ initiator leaves behind a $\gamma$ remnant.
\item A $\gamma$ initiator leaves behind an $\mathrm{e}^-$ remnant.
\item An $\mathrm{e}^+$ initiator leaves behind an $\mathrm{e}^- + \mathrm{e}^-$ remnant.
\item A $\mathrm{q}$ ($\overline{\mathrm{q}}$) initiator leaves behind a
$\overline{\mathrm{q}} + \mathrm{e}^-$ ($\mathrm{q} + \mathrm{e}^-$) remnant.
\item A $\mathrm{g}$ initiator leaves behind a $\mathrm{g} + \mathrm{e}^-$ remnant.
One could argue that, in agreement with the treatment of photon
beams above, the remnant should be $\mathrm{q} + \overline{\mathrm{q}} + \mathrm{e}^-$. The program
currently does not allow for three beam remnant objects, however.
\end{Itemize}
By the hard scattering and initial-state radiation machinery,
the shower initiator has been assigned some fraction $x$ of the
four-momentum of the beam particle, leaving behind $1-x$ to the
remnant. If the remnant consists of two objects, this energy
and momentum has to be shared, somehow. For an electron, the
sharing is given from first principles: if, e.g., the initiator
is a $\mathrm{q}$, then that $\mathrm{q}$ was produced in the sequence of branchings
$\mathrm{e} \to \gamma \to \mathrm{q}$, where $x_{\gamma}$ is distributed according
to the convolution in eq.~(\ref{pg:foldqgine}). Therefore the $\overline{\mathrm{q}}$
remnant takes a fraction $\chi = (x_{\gamma} -x)/(1-x)$ of the total
remnant energy, and the $\mathrm{e}$ takes $1 - \chi$.
For the other beam remnants, the relative energy-sharing variable
$\chi$ is not known from first principles, but picked according to
some suitable parametrization. Normally several different options are
available, that can be set separately for baryon and meson beams, and
for hadron + quark and quark + diquark (or antiquark) remnants. In one
extreme are shapes in agreement with na\"{\i}ve counting rules, i.e.
where energy is shared evenly between `valence' partons. For instance,
${\cal P}(\chi) = 2 \, (1-\chi)$ for the energy fraction taken by the
$\mathrm{q}$ in a $\mathrm{q} + \mathrm{q}\q$ remnant. In the other extreme, an uneven
distribution could be used, like in parton distributions, where the
quark only takes a small fraction and most is retained by the diquark.
The default for a $\mathrm{q} + \mathrm{q}\q$ remnant is of this type,
\begin{equation}
{\cal P}(\chi) \propto \frac{(1 - \chi)^3}
{\sqrt{\chi^2 + c_{\mathrm{min}}^2}} ~,
\end{equation}
with $c_{\mathrm{min}} = 2 \langle m_{\mathrm{q}} \rangle / E_{\mrm{cm}} = (0.6$
GeV)$/ E_{\mrm{cm}}$ providing a lower cut-off. In general, the more
uneven the sharing of the energy, the less the total multiplicity in the
beam remnant fragmentation. If no multiple interactions are allowed,
a rather even sharing is needed to come close to the experimental
multiplicity (and yet one does not quite make it). With an
uneven sharing there is room to generate more of the total multiplicity
by multiple interactions \cite{Sjo87a}.
In a photon beam, with a remnant $\mathrm{q} + \overline{\mathrm{q}}$, the $\chi$ variable is
chosen the same way it would have been in a corresponding meson
remnant.
Before the $\chi$ variable is used to assign remnant momenta, it
is also necessary to consider the issue of primordial $k_{\perp}$.
The initiator partons are thus assigned each a $k_{\perp}$ value,
vanishing for an electron or photon inside an electron, distributed
either according to a Gaussian or an exponential shape for a hadron,
and according to either of these shapes or a power-like shape
for a quark or gluon inside a photon (which may in its turn be inside
an electron). The interaction subsystem is boosted and rotated to bring
it from the frame assumed so far, with each initiator along the $\pm z$
axis, to one where the initiators have the required primordial
$k_{\perp}$ values.
The $p_{\perp}$ recoil is taken by the remnant. If the remnant is composite,
the recoil is all taken by one of the two, namely the one that,
in some imagined perturbative splitting language, is the sister of
the initiator parton. For instance, when a gluon is taken out of a
proton, the recoil is all taken by the lone quark
(i.e. nothing by the diquark), since one could
have imagined an earlier branching $\mathrm{q}_0 \to \mathrm{q} \mathrm{g}$, below the shower
cut-off scale $Q_0$, with $p_{\perp q_0} = 0$.
In addition, however, two beam remnants may be
given a relative $p_{\perp}$, which is then always chosen as for
$\mathrm{q}_i \overline{\mathrm{q}}_i$ pairs in the fragmentation description.
The $\chi$ variable is interpreted as a sharing of light-cone energy and
momentum, i.e. $E + p_z$ for the beam moving in the $+z$ direction and
$E - p_z$ for the other one. When the two transverse masses
$m_{\perp 1}$ and $m_{\perp 2}$ of a composite remnant have been
constructed, the total transverse mass can therefore be found as
\begin{equation}
m_{\perp}^2 = \frac{m_{\perp 1}^2}{\chi} +
\frac{m_{\perp 2}^2}{1 - \chi} ~,
\end{equation}
if remnant 1 is the one that takes the fraction $\chi$. The choice of
a light-cone interpretation to $\chi$ means the definition is invariant
under longitudinal boosts, and therefore does not depend on the beam
energy itself. A $\chi$ value close to the na\"{\i}ve borders 0 or 1 can
lead to an unreasonably large remnant $m_{\perp}$.
Therefore an additional check is introduced, that the remnant
$m_{\perp}$ be smaller than the na\"{\i}ve c.m. frame remnant energy,
$(1-x) E_{\mrm{cm}}/2$. If this is not the case, a new $\chi$ and a
new relative transverse momentum is selected.
Whether there is one remnant parton or two, the transverse mass of
the remnant is not likely to agree with $1-x$ times the mass of the
beam particle, i.e. it is not going to be possible to preserve
the energy and momentum
in each remnant separately. One therefore allows a
shuffling of energy and momentum between the beam remnants from
each of the two incoming beams. This may be achieved by performing a
(small) longitudinal boost of each remnant system. Since there are
two boost degrees of freedom, one for each remnant, and two constraints,
one for energy and one for longitudinal momentum, a solution may be
found.
Under some circumstances, one beam remnant may be absent or of very
low energy, while the other one is more complicated. One example is
deep inelastic scattering in $\e\p$ collisions, where the electron
leaves no remnant, or maybe only a low-energy photon.
It is clearly then not possible to balance the two beam remnants
against each other. Therefore, if one beam remnant has an energy
below 0.2 of the beam energy, i.e. if the initiator parton has
$x > 0.8$, then the two boosts needed to ensure energy and momentum
conservation are instead performed on the other remnant and on the
interaction subsystem. If there is a low-energy remnant at all then,
before that, energy and momentum are assigned to the remnant
constituent(s) so that the appropriate light-cone combination
$E \pm p_z$ is conserved, but not energy or momentum separately.
If both beam remnants have low energy, but both still exist, then
the one with lower $m_{\perp} / E$ is the one that will not be
boosted.
\subsection{Multiple Interactions}
\label{ss:multint}
In this section we present the model used in {\tsc{Pythia}} to describe the
possibility that several parton pairs undergo hard interactions in a
hadron--hadron collision, and thereby contribute to the overall event
activity, in particular at low $p_{\perp}$. The same model is also used
to describe the VMD $\gamma \mathrm{p}$ events, where the photon interacts
like a hadron. It should from the onset be made
clear that this is not an easy topic. In fact, in the full event
generation process, probably no other area is as poorly understood
as this one. The whole concept of multiple interactions is very
controversial, with contraditory experimental conclusions
\cite{AFS87}.
The multiple interactions scenario presented here \cite{Sjo87a} was
the first detailed model for this kind of physics, and is still one
of the very few available. We will present two related but separate
scenarios, one `simple' model and one somewhat more sophisticated.
In fact, neither of them are all that simple, which may make the
models look unattractive. However, the world of hadron physics
{\it is} complicated, and if we err, it is most likely in being
too unsophisticated. The experience gained with the model(s), in
failures as well as successes, could be used as a guideline in
the evolution of yet more detailed models.
Our basic philosophy will be as follows. The total rate of
parton--parton interactions, as a function of the transverse
momentum scale $p_{\perp}$, is assumed to be given by perturbative
QCD. This is certainly true for reasonably large $p_{\perp}$ values, but
here we shall also extend the perturbative parton--parton
scattering framework into the low-$p_{\perp}$ region. A regularization of
the divergence in the cross section for $p_{\perp} \to 0$ has to be
introduced, however, which will provide us with the main free
parameter of the model. Since each incoming hadron is a composite
object, consisting of many partons, there should exist the possibility
of several parton pairs interacting when two hadrons collide. It is
not unreasonable to assume that the different pairwise interactions
take place essentially independently of each other, and that therefore
the number of interactions in an event is given by a Poissonian
distribution. This is the strategy of the `simple' scenario.
Furthermore, hadrons are not only composite but also extended
objects, meaning that collisions range from very central to rather
peripheral ones. Reasonably, the average number of interactions
should be larger in the former than in the latter case. Whereas
the assumption of a Poissonian distribution should hold for each
impact parameter separately, the distribution
in number of interactions should be widened by the spread of impact
parameters. The amount of widening depends on the assumed
matter distribution inside the colliding hadrons. In the `complex'
scenario, different matter distributions are therefore introduced.
\subsubsection{The basic cross sections}
The QCD cross section for hard $2 \to 2$ processes, as a function
of the $p_{\perp}^2$ scale, is given by
\begin{equation}
\frac{\d \sigma}{\d p_{\perp}^2} = \sum_{i,j,k} \int \d x_1 \int \d x_2
\int \d \hat{t} \, f_i(x_1,Q^2) \, f_j(x_2,Q^2)
\, \frac{\d \hat{\sigma}_{ij}^k}{\d \hat{t}} \,
\delta \! \left( p_{\perp}^2 - \frac{\hat{t}\hat{u}}{\hat{s}} \right) ~,
\label{mi:sigmapt}
\end{equation}
cf. section \ref{ss:kinemtwo}. Implicitly, from now on we are assuming
that the `hardness' of processes is given by the $p_{\perp}$ scale of the
scattering. For an application of the formula above to small $p_{\perp}$
values, a number of caveats could be made. At low $p_{\perp}$, the
integrals receive major contributions from the small-$x$ region,
where parton distributions are poorly understood theoretically
(Regge limit behaviour, dense packing problems etc. \cite{Lev90})
and not yet measured. Different sets of parton distributions can
therefore give numerically rather different results for the
phenomenology of interest. One may also worry about higher-order
corrections to the jet rates ($K$ factors), beyond what is given
by parton-shower corrections --- one simple option we allow here
is to evaluate $\alpha_{\mathrm{s}}$ of the hard scattering process at an
optimized scale, $\alpha_{\mathrm{s}}(0.075p_{\perp}^2)$ \cite{Ell86}.
The hard scattering cross section above some given $p_{\perp\mathrm{min}}$
is given by
\begin{equation}
\sigma_{\mrm{hard}} (p_{\perp\mathrm{min}}) = \int_{p_{\perp\mathrm{min}}^2}^{s/4}
\frac{\d \sigma}{\d p_{\perp}^2} \, \d p_{\perp}^2 ~.
\label{mi:sigmahard}
\end{equation}
Since the differential cross section diverges roughly like
$\d p_{\perp}^2 / p_{\perp}^4$, $\sigma_{\mrm{hard}}$ is also divergent for
$p_{\perp\mathrm{min}} \to 0$. We may compare this with the total
inelastic, non-diffractive cross section $\sigma_{\mrm{nd}}(s)$
--- elastic and diffractive events are not the topic of this
section. At current collider energies
$\sigma_{\mrm{hard}} (p_{\perp\mathrm{min}})$
becomes comparable with $\sigma_{\mrm{nd}}$
for $p_{\perp\mathrm{min}} \approx$1.5--2 GeV. This need not lead to
contradictions: $\sigma_{\mrm{hard}}$ does not give the hadron--hadron
cross section but the parton--parton one. Each of the incoming
hadrons may be viewed as a beam of partons, with the possibility of
having several parton--parton interactions when the hadrons pass
through each other. In this language,
$\sigma_{\mrm{hard}} (p_{\perp\mathrm{min}}) / \sigma_{\mrm{nd}}(s)$
is simply the
average number of parton--parton scatterings above $p_{\perp\mathrm{min}}$
in an event, and this number may well be larger than unity.
While the introduction of several interactions per event is the
natural consequence of allowing small $p_{\perp\mathrm{min}}$ values
and hence large $\sigma_{\mrm{hard}}$ ones, it is not the solution of
$\sigma_{\mrm{hard}} (p_{\perp\mathrm{min}})$ being divergent for
$p_{\perp\mathrm{min}} \to 0$: the average $\hat{s}$ of a scattering
decreases slower with $p_{\perp\mathrm{min}}$ than the number of
interactions increases, so na\"{\i}vely the total amount of scattered
partonic energy becomes infinite. One cut-off is therefore
obtained via the need to introduce proper multi-parton correlated
parton distributions inside a hadron. This is not a part of the
standard perturbative QCD formalism and is therefore not built into
eq.~(\ref{mi:sigmahard}). In practice, even correlated
parton-distribution functions seems to provide too weak a cut,
i.e. one is lead to a
picture with too little of the incoming energy remaining in the
small-angle beam jet region.
A more credible reason for an effective cut-off is that the incoming
hadrons are colour neutral objects. Therefore, when the $p_{\perp}$ of an
exchanged gluon is made small and the transverse wavelength
correspondingly large, the gluon can no longer resolve the individual
colour charges, and the effective coupling is decreased. This
mechanism is not in contradiction to perturbative
QCD calculations, which are always performed assuming scattering
of free partons (rather than partons inside hadrons), but neither does
present knowledge of QCD provide an understanding of how such a
decoupling mechanism would work in detail. In the simple model one
makes use of a sharp cut-off at some scale $p_{\perp\mathrm{min}}$, while
a more smooth dampening is assumed for the complex scenario.
\subsubsection{The simple model}
In an event with several interactions, it is convenient to impose an
ordering. The logical choice is to arrange the scatterings in falling
sequence of $x_{\perp} = 2 p_{\perp} / E_{\mrm{cm}}$. The `first' scattering
is thus the hardest one, with the `subsequent' (`second', `third', etc.)
successively softer. It is important to remember that this terminology
is in no way related to any picture in physical time; we do not know
anything about the latter. In principle, all the scatterings that occur
in an event must be correlated somehow, na\"{\i}vely by momentum
and flavour conservation for the partons from each incoming hadron,
less na\"{\i}vely by
various quantum mechanical effects. When averaging over all
configurations of soft partons, however, one should effectively obtain
the standard QCD phenomenology for a hard scattering, e.g. in terms of
parton distributions. Correlation effects, known or estimated, can be
introduced in the choice of subsequent scatterings, given that the
`preceding' (harder) ones are already known.
With a total cross section of hard interactions
$\sigma_{\mrm{hard}} (p_{\perp\mathrm{min}})$ to be distributed among
$\sigma_{\mrm{nd}}(s)$ (non-diffractive, inelastic) events, the average
number of interactions per event is just the ratio
$\br{n} = \sigma_{\mrm{hard}} (p_{\perp\mathrm{min}}) / \sigma_{\mrm{nd}}(s)$.
As a starting point we will assume that all hadron collisions are
equivalent (no impact parameter dependence), and that the different
parton--parton interactions take place completely independently of
each other. The number of scatterings per event is then distributed
according to a Poissonian with mean $\br{n}$. A fit to collider
multiplicity data gives $p_{\perp\mathrm{min}} \approx 1.6$ GeV, which
corresponds to $\br{n} \approx 1$. For Monte Carlo
generation of these interactions it is useful to define
\begin{equation}
f(x_{\perp}) = \frac{1}{\sigma_{\mrm{nd}}(s)} \,
\frac{\d \sigma}{\d x_{\perp}} ~,
\end{equation}
with $\d \sigma / \d x_{\perp}$ obtained by analogy with eq.
(\ref{mi:sigmapt}). Then $f(x_{\perp})$ is simply the probability to
have a parton--parton interaction at $x_{\perp}$, given that the two
hadrons undergo a non-diffractive, inelastic collision.
The probability that the hardest interaction, i.e. the one with
highest $x_{\perp}$, is at $x_{\perp 1}$, is now given by
\begin{equation}
f(x_{\perp 1}) \exp \left\{ - \int_{x_{\perp 1}}^1
f(x'_{\perp}) \, \d x'_{\perp} \right\} ~,
\label{mi:hardest}
\end{equation}
i.e. the na\"{\i}ve probability to have a scattering at $x_{\perp 1}$
multiplied by the probability that there was no scattering with
$x_{\perp}$ larger than $x_{\perp 1}$. This is the familiar
exponential dampening in radioactive decays, encountered e.g. in
parton showers in section \ref{ss:sudakov}. Using the same technique
as in the proof of the veto algorithm, section \ref{ss:vetoalg},
the probability to have an $i$:th scattering at an
$x_{\perp i} < x_{\perp i-1} < \cdots < x_{\perp 1} < 1$ is found
to be
\begin{equation}
f(x_{\perp i}) \, \frac{1}{(i-1)!} \left( \int_{x_{\perp i}}^1
f(x'_{\perp}) \, \d x'_{\perp} \right)^{i-1} \exp \left\{
- \int_{x_{\perp i}}^1 f(x'_{\perp}) \, \d x'_{\perp} \right\} ~.
\end{equation}
The total probability to have a scattering at a given $x_{\perp}$,
irrespectively of it being the first, the second or whatever,
obviously adds up to give back $f(x_{\perp})$. The multiple
interaction formalism thus retains the correct perturbative
QCD expression for the scattering probability at any given
$x_{\perp}$.
With the help of the integral
\begin{equation}
F(x_{\perp}) = \int_{x_{\perp}}^1 f(x'_{\perp}) \, \d x'_{\perp}
= \frac{1}{\sigma_{\mrm{nd}}(s)} \, \int_{s x_{\perp}^2/4}^{s/4}
\frac{\d \sigma}{\d p_{\perp}^2} \, \d p_{\perp}^2
\end{equation}
(where we assume $F(x_{\perp}) \to \infty$ for $x_{\perp} \to 0$)
and its inverse $F^{-1}$, the iterative procedure to generate
a chain of scatterings
$1 > x_{\perp 1} > x_{\perp 2} > \cdots > x_{\perp i}$
is given by
\begin{equation}
x_{\perp i} = F^{-1}(F(x_{\perp i-1}) - \ln R_i) ~.
\label{mi:iter}
\end{equation}
Here the $R_i$ are random numbers evenly distributed between 0 and 1.
The iterative chain is started with a fictitious $x_{\perp 0} = 1 $
and is terminated when $x_{\perp i}$ is smaller than
$x_{\perp \mathrm{min}} = 2 p_{\perp\mathrm{min}} / E_{\mrm{cm}}$. Since $F$ and
$F^{-1}$ are not known analytically, the standard veto algorithm is
used to generate a much denser set of $x_{\perp}$ values, whereof
only some are retained in the end. In addition to the $p_{\perp}^2$ of an
interaction, it is also necessary to generate the other flavour
and kinematics variables according to the relevant matrix elements.
Whereas the ordinary parton distributions should be used for the
hardest scattering, in order to reproduce standard QCD
phenomenology, the parton distributions to be used for subsequent
scatterings must depend on all preceding $x$ values and flavours
chosen. We do not know enough about the hadron wave function to
write down such joint probability distributions. To take
into account the energy `already' used in harder scatterings, a
conservative approach is to evaluate the parton distributions, not at
$x_i$ for the $i$:th scattered parton from hadron, but at
the rescaled value
\begin{equation}
x'_i = \frac{x_i}{\sum_{j=1}^{i-1} x_j} ~.
\end{equation}
This is our standard procedure in the program; we have
tried a few alternatives without finding any significantly
different behaviour in the final physics.
In a fraction $\exp(-F(x_{\perp \mathrm{min}}))$ of the events studied, there
will be no hard scattering above $x_{\perp \mathrm{min}}$ when the iterative
procedure in eq.~(\ref{mi:iter}) is applied. It is therefore also
necessary to have a model for what happens in events with no
(semi)hard interactions. The simplest possible way to produce an event
is to have an exchange of a very soft gluon between the two colliding
hadrons. Without (initially) affecting the momentum distribution of
partons, the `hadrons' become colour octet objects rather than colour
singlet ones. If only valence quarks are considered, the colour
octet state of a baryon can be decomposed into a colour triplet quark
and an antitriplet diquark. In a baryon-baryon collision, one would
then obtain a two-string picture, with each string stretched from the
quark of one baryon to the diquark of the other. A baryon-antibaryon
collision would give one string between a quark and an antiquark and
another one between a diquark and an antidiquark.
In a hard interaction, the number of possible string drawings are many
more, and the overall situation can become quite complex
when several hard scatterings are present in an event.
Specifically, the string drawing now depends on the relative colour
arrangement, in each hadron individually, of the partons that are about
to scatter. This is a subject about which nothing is known.
To make matters worse, the standard string fragmentation description
would have to be extended, to handle events where two or more valence
quarks have been kicked out of an incoming hadron by separate
interactions. In particular, the position of the baryon number would
be unclear. We therefore here assume that, following the hardest
interaction, all subsequent interactions belong to one of three
classes.
\begin{Itemize}
\item Scatterings of the $\mathrm{g}\g \to \mathrm{g}\g $ type, with
the two gluons in a colour-singlet state, such that a double string is
stretched directly between the two outgoing gluons, decoupled from the
rest of the system.
\item Scatterings $\mathrm{g}\g \to \mathrm{g}\g$, but colour correlations
assumed to be such that each of the gluons is connected to one
of the strings `already' present. Among the different possibilities of
connecting the colours of the gluons, the one which minimizes the total
increase in string length is chosen. This is in contrast to the
previous alternative, which roughly corresponds to a maximization of
the extra string length.
\item Scatterings $\mathrm{g}\g \to \mathrm{q}\overline{\mathrm{q}}$, with the final pair again
in a colour-singlet state, such that a single string is stretched
between the outgoing $\mathrm{q}$ and $\overline{\mathrm{q}}$.
\end{Itemize}
By default, the three possibilities are assumed equally probable.
Note that the total jet rate is maintained at its nominal value, i.e.
scatterings such as $\mathrm{q}\mathrm{g} \to \mathrm{q}\mathrm{g}$ are included in the cross section,
but are replaced by a mixture of $\mathrm{g}\g$ and $\mathrm{q}\overline{\mathrm{q}}$ events for string
drawing issues. Only the hardest interaction is guaranteed to give
strings coupled to the beam remnants. One should not take this approach
to colour flow too seriously --- clearly
it is a simplification --- but the overall picture does not tend to be
very dependent on the particular choice you make.
Since a $\mathrm{g}\g \to \mathrm{g}\g$ or $\mathrm{q}\overline{\mathrm{q}}$ scattering need not remain that if
initial- and final-state showers were to be included, radiation is only
included for the hardest interaction. In practice, there is no
problem: except for the hardest interaction, which can be hard because
of experimental trigger conditions, it is unlikely for a parton
scattering to be so hard that radiation plays a significant r\^ole.
In events with multiple interactions, the beam remnant treatment is
slightly modified. First the hard scattering is generated, with its
associated initial- and final-state radiation, and next any additional
multiple interactions. Only thereafter are beam remnants attached to
the initator partons of the hardest scattering, using the same
machinery as before, except that the energy and momentum already
taken away from the beam remnants also include that of the
subsequent interactions.
\subsubsection{A model with varying impact parameters}
Up to this point, it has been assumed that the initial state is the
same for all hadron collisions, whereas in fact each collision also
is characterized by a varying impact parameter $b$. Within the
classical framework of this paper, $b$ is to be thought of as a
distance of closest approach, not as the Fourier transform of the
momentum transfer. A small $b$ value corresponds to a large
overlap between the two colliding hadrons, and hence an enhanced
probability for multiple interactions. A large $b$, on the other
hand, corresponds to a grazing collision, with a large probability
that no parton--parton interactions at all take place.
In order to quantify the concept of hadronic matter overlap, one may
assume a spherically symmetric distribution of matter inside the
hadron, $\rho(\mbf{x}) \, \d^3 x = \rho(r) \, \d^3 x$.
For simplicity, the same spatial distribution is taken to apply
for all parton species and momenta. Several different matter
distributions have been tried, and are available. We will here
concentrate on the most extreme one, a double Gaussian
\begin{equation}
\rho(r) \propto \frac{1 - \beta}{a_1^3} \, \exp \left\{
- \frac{r^2}{a_1^2} \right\} + \frac{\beta}{a_2^3} \,
\exp \left\{ - \frac{r^2}{a_2^2} \right\} ~.
\label{mi:doubleGauss}
\end{equation}
This corresponds to a distribution with a small core region, of radius
$a_2$ and containing a fraction $\beta$ of the total hadronic matter,
embedded in a larger hadron of radius $a_1$. While it is mathematically
convenient to have the origin of the two Gaussians coinciding, the
physics could well correspond to having three disjoint core regions,
reflecting the presence of three valence quarks, together carrying
the fraction $\beta$ of the proton momentum. One could alternatively
imagine a hard hadronic core surrounded by a pion cloud.
Such details would affect e.g. the predictions for the $t$ distribution
in elastic scattering, but are not of any consequence for the
current topics. To be
specific, the values $\beta = 0.5$ and $a_2/a_1 = 0.2$ have been
picked as default values. It should be noted that the overall distance
scale $a_1$ never enters in the subsequent calculations, since the
inelastic, non-diffractive cross section $\sigma_{\mrm{nd}}(s)$ is taken
from literature rather than calculated from the $\rho(r)$.
Compared to other shapes, like a simple Gaussian, the double Gaussian
tends to give larger fluctuations, e.g. in the multiplicity
distribution of minimum bias events: a collision in which the two
cores overlap tends to have a strongly increased activity, while
ones where they do not are rather less active. One also has a
biasing effect: hard processes are more likely
when the cores overlap, thus hard scatterings are associated with
an enhanced multiple interaction rate. This provides one possible
explanation for the experimental `pedestal effect'.
For a collision with impact parameter $b$, the time-integrated
overlap ${\cal O}(b)$ between the matter distributions of the
colliding hadrons is given by
\begin{eqnarray}
& & {\cal O}(b) \propto \int \d t \int \d^3 x \, \rho(x,y,z) \,
\rho(x+b,y,z+t) \nonumber \\
& & \propto \frac{(1 - \beta)^2}{2a_1^2} \exp \left\{
- \frac{b^2}{2a_1^2} \right\} +
\frac{2 \beta (1-\beta)}{a_1^2+a_2^2} \exp \left\{
- \frac{b^2}{a_1^2+ a_2^2} \right\} +
\frac{\beta^2}{2a_2^2} \exp \left\{ - \frac{b^2}{2a_2^2} \right\} ~.
\end{eqnarray}
The necessity to use boosted $\rho(\mbf{x})$ distributions
has been circumvented by a suitable scale transformation of the $z$
and $t$ coordinates.
The overlap ${\cal O}(b)$ is obviously strongly related to the
eikonal $\Omega(b)$ of optical models. We have kept a separate
notation, since the physics context of the two is slightly
different: $\Omega(b)$ is based on the quantum mechanical
scattering of waves in a potential, and is normally used to
describe the elastic scattering of a hadron-as-a-whole, while
${\cal O}(b)$ comes from a purely classical picture of point-like
partons distributed inside the two colliding hadrons. Furthermore,
the normalization and energy dependence is differently realized
in the two formalisms.
The larger the overlap ${\cal O}(b)$ is, the more likely it is to
have interactions between partons in the two colliding hadrons.
In fact, there should be a linear relationship
\begin{equation}
\langle \tilde{n}(b) \rangle = k {\cal O}(b) ~,
\end{equation}
where $\tilde{n} = 0, 1, 2, \ldots$ counts the number of interactions
when two hadrons pass each other with an impact parameter $b$.
The constant of proportionality, $k$, is related to the
parton--parton cross section and hence increases with c.m. energy.
For each given impact parameter, the number of interactions is
assumed to be distributed according to a Poissonian. If the matter
distribution has a tail to infinity (as the double Gaussian does),
events may be obtained with arbitrarily large $b$ values. In order
to obtain finite total cross sections, it is necessary to assume
that each event contains at least one semi-hard interaction. The
probability that two hadrons, passing each other with an impact
parameter $b$, will actually undergo a collision is then given by
\begin{equation}
{\cal P}_{\mrm{int}}(b) = 1 - \exp ( - \langle \tilde{n}(b) \rangle )
= 1 - \exp (- k {\cal O}(b) ) ~,
\end{equation}
according to Poissonian statistics. The average number of
interactions per event at impact parameter $b$ is now
\begin{equation}
\langle n(b) \rangle =
\frac{ \langle \tilde{n}(b) \rangle }{{\cal P}_{\mrm{int}}(b)} =
\frac{ k {\cal O}(b) }{ 1 - \exp (- k {\cal O}(b) ) } ~,
\label{mi:nofb}
\end{equation}
where the denominator comes from the removal of hadron pairs which
pass without colliding, i.e. with $\tilde{n} = 0$.
The relationship
$\langle n \rangle = \sigma_{\mrm{hard}}/\sigma_{\mrm{nd}}$
was earlier introduced for the average number of interactions per
non-diffractive, inelastic event. When averaged over all
impact parameters, this relation must still hold true: the
introduction of variable impact parameters may give more interactions
in some events and less in others, but it does not affect either
$\sigma_{\mrm{hard}}$ or $\sigma_{\mrm{nd}}$.
For the former this is because the
perturbative QCD calculations only depend on the total parton flux,
for the latter by construction. Integrating eq.~(\ref{mi:nofb}) over
$b$, one then obtains
\begin{equation}
\langle n \rangle =
\frac{ \int \langle n(b) \rangle \, {\cal P}_{\mrm{int}}(b) \, \d^2 b}
{ \int {\cal P}_{\mrm{int}}(b) \, \d^2 b} =
\frac{ \int k {\cal O}(b) \, \d^2 b}
{ \int \left( 1 - \exp (- k {\cal O}(b) ) \right) \, \d^2 b} =
\frac{\sigma_{\mrm{hard}}}{\sigma_{\mrm{nd}}} ~.
\label{mi:kfromsigma}
\end{equation}
For ${\cal O}(b)$, $\sigma_{\mrm{hard}}$ and $\sigma_{\mrm{nd}}$ given,
with $\sigma_{\mrm{hard}} / \sigma_{\mrm{nd}} > 1$, $k$ can thus always
be found (numerically) by solving the last equality.
The absolute normalization of ${\cal O}(b)$ is not interesting
in itself, but only the relative variation with impact parameter.
It is therefore useful to introduce an `enhancement factor'
$e(b)$, which gauges how the interaction probability for a passage
with impact parameter $b$ compares with the average, i.e.
\begin{equation}
\langle \tilde{n}(b) \rangle = k{\cal O}(b) =
e(b) \, \langle k{\cal O}(b) \rangle ~.
\label{mi:ebenh}
\end{equation}
The definition of the average $\langle k{\cal O}(b) \rangle$ is a
bit delicate, since the average number of interactions per event
is pushed up by the requirement that each event contain at least
one interaction. However, an exact meaning can be given \cite{Sjo87a}.
With the knowledge of $e(b)$, the $f(x_{\perp})$ function of the
simple model generalizes to
\begin{equation}
f(x_{\perp},b) = e(b) \, f(x_{\perp}) ~.
\end{equation}
The na\"{\i}ve generation procedure is thus to pick a $b$ according
to the phase space $\d^2 b$, find the relevant $e(b)$ and plug in the
resulting $f(x_{\perp},b)$ in the formalism of the simple model.
If at least one hard interaction is generated, the event is retained,
else a new $b$ is to be found. This algorithm would work fine for
hadronic matter distributions which vanish outside some radius, so
that the $\d^2 b$ phase space which needs to be probed is finite.
Since this is not true for the distributions under study, it is
necessary to do better.
By analogy with eq.~(\ref{mi:hardest}), it is possible to ask what
the probability is to find the hardest scattering of an event at
$x_{\perp 1}$. For each impact parameter separately, the probability
to have an interaction at $x_{\perp 1}$ is given by $f(x_{\perp},b)$,
and this should be multiplied by the probability that the event
contains no interactions at a scale $x'_{\perp} > x_{\perp 1}$,
to yield the total probability distribution
\begin{eqnarray}
\frac{\d {\cal P}_{\mrm{hardest}}}{\d^2 b \, \d x_{\perp 1}} & = &
f(x_{\perp 1},b) \, \exp \left\{ - \int_{x_{\perp 1}}^1
f(x'_{\perp},b) \, \d x'_{\perp} \right\} \nonumber \\
& = & e(b) \, f(x_{\perp 1}) \, \exp \left\{ - e(b)
\int_{x_{\perp 1}}^1 f(x'_{\perp}) \, \d x'_{\perp} \right\} ~.
\end{eqnarray}
If the treatment of the exponential is deferred for a moment,
the distribution in $b$ and $x_{\perp 1}$ appears in factorized form,
so that the two can be chosen independently of each other. In
particular, a high-$p_{\perp}$ QCD scattering or any other hard scattering
can be selected with whatever kinematics desired for that process,
and thereafter assigned some suitable `hardness' $x_{\perp 1}$.
With the $b$ chosen according to $e(b) \, \d^2 b$, the neglected
exponential can now be evaluated, and the event retained with a
probability proportional to it. From the $x_{\perp 1}$ scale of the
selected interaction, a sequence of softer $x_{\perp i}$ values may
again be generated as in the simple model, using the known
$f(x_{\perp},b)$. This sequence may be empty, i.e. the
event need not contain any further interactions.
It is interesting to understand how the algorithm above works.
By selecting $b$ according to $e(b) \, \d^2 b$, i.e.
${\cal O}(b) \, \d^2 b$, the primary $b$ distribution is
maximally biased towards small impact parameters. If the first
interaction is hard, by choice or by chance, the integral of
the cross section above $x_{\perp 1}$ is small, and the exponential
close to unity. The rejection procedure is therefore very efficient
for all standard hard processes in the program --- one may even
safely drop the weighting with the exponential completely. The large
$e(b)$ value is also likely to lead to the generation of many further,
softer interactions. If, on the other hand, the first interaction is
not hard, the exponential is no longer close to unity, and many
events are rejected. This pulls down the efficiency for `minimum bias'
event generation. Since the exponent is proportional to $e(b)$,
a large $e(b)$ leads to an enhanced probability for rejection,
whereas the chance of acceptance is larger with a small $e(b)$.
Among events where the hardest interaction is soft, the $b$
distribution is therefore biased towards larger values
(smaller $e(b)$), and there is a small probability for yet softer
interactions.
To evaluate the exponential factor, the program pretabulates the
integral of $f(x_{\perp})$ at the initialization stage, and further
increases the Monte Carlo statistics of this tabulation as the run
proceeds. The $x_{\perp}$ grid is concentrated towards small
$x_{\perp}$, where the integral is large. For a selected
$x_{\perp 1}$ value, the $f(x_{\perp})$ integral is obtained by
interpolation. After multiplication by the known $e(b)$ factor,
the exponential factor may be found.
In this section, nothing has yet been assumed about the form of the
$\d \sigma / \d p_{\perp}$ spectrum. Like in the impact parameter independent
case, it is possible to use a sharp cut-off at some given
$p_{\perp\mathrm{min}}$ value. However, now each event is required to have
at least one interaction, whereas before events without interactions
were retained and put at $p_{\perp} = 0$. It is therefore aesthetically
more appealing to assume a gradual turn-off, so that a (semi)hard
interaction can be rather soft part of the time. The matrix elements
roughly diverge like $\alpha_{\mathrm{s}}(p_{\perp}^2) \, \d p_{\perp}^2 / p_{\perp}^4$ for
$p_{\perp} \to 0$. They could therefore be regularized as follows. Firstly,
to remove the $1/p_{\perp}^4$ behaviour, multiply by a factor
$p_{\perp}^4 / (p_{\perp}^2 + p_{\perp 0}^2)^2$. Secondly, replace the $p_{\perp}^2$
argument in $\alpha_{\mathrm{s}}$ by $p_{\perp}^2 + p_{\perp 0}^2$ or, with the
inclusion of the $K$ factor procedure introduced earlier, replace
$0.075 \, p_{\perp}^2$ by $0.075 \, (p_{\perp}^2 + p_{\perp 0}^2)$.
With these substitutions, a continuous $p_{\perp}$ spectrum is obtained,
stretching from $p_{\perp} = 0$ to $E_{\mrm{cm}}/2$. For
$p_{\perp} \gg p_{\perp 0}$
the standard perturbative QCD cross section is recovered, while
values $p_{\perp} \ll p_{\perp 0}$ are strongly damped. The $p_{\perp 0}$
scale, which now is the main free parameter of the model, in
practice comes out to be of the same order of magnitude as the sharp
cut-off $p_{\perp\mathrm{min}}$ did, i.e. 1.5--2 GeV.
If gluons with large transverse wavelength decouple because of the
colour-singlet nature of hadrons, and if the transverse structure of
hadrons is assumed to be energy-independent, it is natural to assume
that also $p_{\perp\mathrm{min}}$ and $p_{\perp 0}$ are independent of the
c.m. energy of the hadron collision. For the impact parameter
independent picture this works out fine, with all events being reduced
to low-$p_{\perp}$ two-string ones when the c.m. energy is reduced. In the
variable impact parameter picture, the whole formalism only makes sense
if $\sigma_{\mrm{hard}} > \sigma_{\mrm{nd}}$, see e.g.
eq.~(\ref{mi:kfromsigma}). Since $\sigma_{\mrm{nd}}$ does not vanish
with decreasing energy, but $\sigma_{\mrm{hard}}$
would do that for a fixed $p_{\perp 0}$, this means
that $p_{\perp 0}$ has to be reduced when the energy is decreased
below some given threshold. The more `sophisticated' model of this
section therefore makes sense at collider energies, whereas it is not
well suited for applications at lower energies.
\subsection{Pile-up Events}
\label{ss:pileup}
In high luminosity colliders, there is a non-negligible probability
that one single bunch crossing may produce several separate events,
so-called pile-up events.
This in particular applies to future $\mathrm{p}\p$ colliders like LHC,
but one could also consider e.g. $\e^+\e^-$ colliders with high rates of
$\gamma\gamma$ collisions. The program therefore contains an option,
currently only applicable to hadron--hadron collisions,
wherein several events may be generated and put one after the other
in the event record, to simulate the full amount of particle
production a detector might be facing.
The program needs to know the assumed luminosity per bunch--bunch
crossing, expressed in mb$^{-1}$. Multiplied by the cross section
for pile-up processes studied, $\sigma_{\mrm{pile}}$, this gives the
average number of collisions per beam crossing, $\br{n}$. These pile-up
events are taken to be of the minimum bias type, with diffractive
and elastic events included or not (and a further subdivision of
diffractive events into single and double). This means that
$\sigma_{\mrm{pile}}$ may be either $\sigma_{\mrm{tot}}$,
$\sigma_{\mrm{tot}} - \sigma_{\mrm{el}}$ or
$\sigma_{\mrm{tot}} - \sigma_{\mrm{el}} - \sigma_{\mrm{diffr}}$.
Which option to choose depends on the detector: most detectors
would not be able to observe elastic $\mathrm{p}\p$ scattering, and therefore
it would be superfluous to generate that kind of events.
In addition, we allow for the possibility that one interaction may
be of a rare kind, selected freely by the user.
There is no option to generate two `rare' events in the same crossing;
normally the likelihood for that kind of occurences should be small.
If only minimum bias type events are generated, i.e. if only one
cross section is involved in the problem, then the number of
events in a crossing is distributed according to a Poissonian
with the average number $\br{n}$ as calculated above.
The program actually will simulate only those beam crossings
where at least one event occurs, i.e. not consider the fraction
$\exp(-\br{n})$ of zero-event crossings. Therefore the actually
generated average number of pile-up events is
$\langle n \rangle = \br{n}/(1-\exp(-\br{n}))$.
Now instead consider the other extreme, where one event is supposed
be rare, with a cross section $\sigma_{\mrm{rare}}$ much smaller
than $\sigma_{\mrm{pile}}$, i.e.
$f \equiv \sigma_{\mrm{rare}}/\sigma_{\mrm{pile}} \ll 1$.
The probability that a bunch crossing will give $i$ events, whereof
one of the rare kind, now is
\begin{equation}
{\cal P}_i = f \, i \, \exp(-\br{n}) \, \frac{\br{n}^i}{i!} =
f \, \br{n} \exp(-\br{n}) \, \frac{\br{n}^{i-1}}{(i-1)!} ~.
\end{equation}
The na\"{\i}ve Poissonian is suppressed by a factor $f$, since one of
the events is rare rather than of the normal kind, but enhanced by a
factor $i$, since any one of the $i$ events may be the rare one.
As the equality shows, the probability distribution is now a
Poissonian in $i-1$:
in a beam crossing which produces one rare event, the multiplicity of
additional pile-up events is distributed according to a Poissonian
with average number $\br{n}$. The total average number of events
thus is $\langle n \rangle = \br{n} + 1$.
Clearly, for processes with intermediate cross sections,
$\br{n} \, \sigma_{\mrm{rare}}/\sigma_{\mrm{pile}} \simeq 1$,
also the average number of events will be intermediate, and it is not
allowed to assume only one event to be of the `rare' type. We do not
consider that kind of situations.
Kindly note that, in the current implementation, all events are
supposed to be produced at the same vertex (the origin). To simulate
the spatial extent of the colliding beams, you would have to assign
interaction points yourself, and then shift each event separately
by the required amount in space and time.
When the pile-up option is used, one main limitation is that event
records may become very large when several events are put one after
the other, so that the space limit in the \ttt{LUJETS} common block
is reached. It is possible to expand the dimension of the common block,
see \ttt{MSTU(4)} and \ttt{MSTU(5)}, but only up to about
20\,000 entries, which might not always be enough.
For practical reasons, the program will only allow a $\br{n}$ up to
120. The multiplicity distribution is truncated above 200,
or when the probability for a multiplicity has fallen below
$10^{-6}$, whichever occurs sooner. Also low multiplicities with
probabilities below $10^{-6}$ are truncated.
\subsection{Common Block Variables}
\label{ss:multintpar}
Of the routines used to generate beam remnants, multiple interactions
and pile-up events, none are intended to be used in standalone mode.
The only way to regulate these aspects is therefore via the
variables in the \ttt{PYPARS} common block.
\drawbox{COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)}
\begin{entry}
\itemc{Purpose:} to give access to a number of status codes and
parameters which regulate the performance of {\tsc{Pythia}}.
Most parameters are described in section \ref{ss:PYswitchpar};
here only those related to beam remnants, multiple interactions
and pile-up events are described. If the default values,
below denoted by (D=\ldots), are not satisfactory, they must in
general be changed before the \ttt{PYINIT} call. Exceptions, i.e.
variables which can be changed for each new event, are denoted by
(C).
\iteme{MSTP(81) :}\label{p:MSTP81} (D=1) master switch for
multiple interactions.
\begin{subentry}
\iteme{= 0 :} off.
\iteme{= 1 :} on.
\end{subentry}
\iteme{MSTP(82) :} (D=1) structure of multiple interactions. For QCD
processes, used down to $p_{\perp}$ values below $p_{\perp\mathrm{min}}$, it also
affects the choice of structure for the one hard/semi-hard interaction.
\begin{subentry}
\iteme{= 0 :} simple two-string model without any hard interactions.
\iteme{= 1 :} multiple interactions assuming the same probability in
all events, with an abrupt $p_{\perp\mathrm{min}}$ cut-off at \ttt{PARP(81)}.
\iteme{= 2 :} multiple interactions assuming the same probability in
all events, with a continuous turn-off of the cross section at
$p_{\perp 0} = $\ttt{PARP(82)}.
\iteme{= 3 :} multiple interactions assuming a varying impact
parameter and a hadronic matter overlap consistent with a Gaussian
matter distribution, with a continuous turn-off of the cross section
at $p_{\perp 0} = $\ttt{PARP(82)}.
\iteme{= 4 :} multiple interactions assuming a varying impact
parameter and a hadronic matter overlap consistent with a double
Gaussian matter distribution given by \ttt{PARP(83)} and
\ttt{PARP(84)}, with a continuous turn-off of the cross section at
$p_{\perp 0} = $\ttt{PARP(82)}.
\itemc{Note 1:} For \ttt{MSTP(82)}$\geq 2$ and
\ttt{CKIN(3)}$ > $\ttt{PARP(82)}, cross sections
given with \ttt{PYSTAT(1)} may be somewhat too large, since (for
reasons of efficiency) the probability factor that the hard
interaction is indeed the hardest in the event is not
included in the cross sections. It is included in the event
selection, however, so the events generated are correctly
distributed. For \ttt{CKIN(3)} values a couple of times larger than
\ttt{PARP(82)} this ceases to be a problem.
\itemc{Note 2:} The \ttt{PARP(81)} and, in particular, \ttt{PARP(82)}
values are sensitive to the choice of parton distributions,
$\Lambda_{\mrm{QCD}}$,
etc., in the sense that a change in the latter variables
leads to a net change in the multiple interaction rate, which has
to be compensated by a retuning of \ttt{PARP(81)} or \ttt{PARP(82)}
if one wants to keep the net multiple interaction structure the
same. The default \ttt{PARP(81)} value is consistent with the other
default values give, i.e. CTEQ 2L parton distributions etc.
When options \ttt{MSTP(82)=} 2--4 are used, the default
\ttt{PARP(82)} value is to be used in conjunction with
\ttt{MSTP(2)=2} and \ttt{MSTP(33)=3}. These switches
should be set by you.
\end{subentry}
\iteme{MSTP(83) :} (D=100) number of Monte Carlo generated phase-space
points per bin (whereof there are 20) in the initialization
(in \ttt{PYMULT}) of multiple interactions for
\ttt{MSTP(82)}$\geq 2$.
\iteme{MSTP(91) :} (D=1) (C) primordial $k_{\perp}$ distribution
in hadron. See \ttt{MSTP(93)} for photon.
\begin{subentry}
\iteme{= 0 :} no primordial $k_{\perp}$.
\iteme{= 1 :} Gaussian, width given in \ttt{PARP(91)}, upper cut-off
in \ttt{PARP(93)}.
\iteme{= 2 :} exponential, width given in \ttt{PARP(92)}, upper
cut-off in \ttt{PARP(93)}.
\end{subentry}
\iteme{MSTP(92) :} (D=4) (C) energy partitioning in hadron or
resolved photon remnant, when this remnant is split into two jets.
(For a splitting into a hadron plus a jet, see \ttt{MSTP(94)}.)
The energy fraction $\chi$ taken by one of the two objects, with
conventions as described for \ttt{PARP(94)} and \ttt{PARP(96)},
is chosen according to the different distributions below. Here
$c_{\mathrm{min}} = 2 \langle m_{\mathrm{q}} \rangle/E_{\mrm{cm}} = 0.6$
GeV$/E_{\mrm{cm}}$.
\begin{subentry}
\iteme{= 1 :} 1 for meson or resolved photon, $2(1-\chi)$ for
baryon, i.e. simple counting rules.
\iteme{= 2 :} $(k+1)(1-\chi)^k$, with $k$ given by
\ttt{PARP(94)} or \ttt{PARP(96)}.
\iteme{= 3 :} proportional to
$(1-\chi)^k/\sqrt[4]{\chi^2+c_{\mathrm{min}}^2}$, with $k$ given by
\ttt{PARP(94)} or \ttt{PARP(96)}.
\iteme{= 4 :} proportional to $(1-\chi)^k/\sqrt{\chi^2+c_{\mathrm{min}}^2}$,
with $k$ given by \ttt{PARP(94)} or \ttt{PARP(96)}.
\iteme{= 5 :} proportional to $(1-\chi)^k/(\chi^2+c_{\mathrm{min}}^2)^{b/2}$,
with $k$ given by \ttt{PARP(94)} or \ttt{PARP(96)}, and
$b$ by \ttt{PARP(98)}.
\end{subentry}
\iteme{MSTP(93) :} (D=1) (C) primordial $k_{\perp}$ distribution
in photon, either it is one of the incoming particles or inside
an electron.
\begin{subentry}
\iteme{= 0 :} no primordial $k_{\perp}$.
\iteme{= 1 :} Gaussian, width given in \ttt{PARP(99)}, upper
cut-off in \ttt{PARP(100)}.
\iteme{= 2 :} exponential, width given in \ttt{PARP(99)}, upper
cut-off in \ttt{PARP(100)}.
\iteme{= 3 :} power-like of the type
$\d k_{\perp}^2/(k_{\perp 0}^2 + k_{\perp}^2)^2$, with
$k_{\perp 0}$ in \ttt{PARP(99)} and upper $k_{\perp}$ cut-off in
\ttt{PARP(100)}.
\iteme{= 4 :} power-like of the type
$\d k_{\perp}^2/(k_{\perp 0}^2 + k_{\perp}^2)$, with
$k_{\perp 0}$ in \ttt{PARP(99)} and upper $k_{\perp}$ cut-off in
\ttt{PARP(100)}.
\iteme{= 5 :} power-like of the type
$\d k_{\perp}^2/(k_{\perp 0}^2 + k_{\perp}^2)$, with
$k_{\perp 0}$ in \ttt{PARP(99)} and upper $k_{\perp}$ cut-off
given by the $p_{\perp}$ of the hard process or by
\ttt{PARP(100)}, whichever is smaller.
\itemc{Note:} for options 1 and 2 the \ttt{PARP(100)} value is of
minor importance, once \ttt{PARP(100)}$\gg$\ttt{PARP(99)}. However,
options 3 and 4 correspond to distributions with infinite
$\langle k_{\perp}^2 \rangle$ if the $k_{\perp}$ spectrum is not
cut off, and therefore the \ttt{PARP(100)} value is as important
for the overall distribution as is \ttt{PARP(99)}.
\end{subentry}
\iteme{MSTP(94) :} (D=2) (C) energy partitioning in hadron or
resolved photon remnant, when this remnant is split into a hadron
plus a remainder-jet. The energy fraction chi is taken by one of the
two objects, with conventions as described below or for \ttt{PARP(95)}
and \ttt{PARP(97)}.
\begin{subentry}
\iteme{= 1 :} 1 for meson or resolved photon, $2(1-\chi)$ for
baryon, i.e. simple counting rules.
\iteme{= 2 :} $(k+1)(1-\chi)^k$, with $k$ given by
\ttt{PARP(95)} or \ttt{PARP(97)}.
\iteme{= 3 :} the $\chi$ of the hadron is selected according to the
normal fragmentation function used for the hadron in jet
fragmentation, see \ttt{MSTJ(11)}. The possibility of a changed
fragmentation function shape in diquark fragmentation
(see \ttt{PARJ(45)}) is not included.
\iteme{= 4 :} as \ttt{=3}, but the shape is changed as allowed in
diquark fragmentation (see \ttt{PARJ(45)}); this change is here also
allowed for meson production. (This option is not so natural
for mesons, but has been added to provide the same amount of freedom
as for baryons).
\end{subentry}
\iteme{MSTP(131) :}\label{p:MSTP131} (D=0) master switch for pile-up
events, i.e. several
independent hadron--hadron interactions generated in the same
bunch--bunch crossing, with the events following one after the
other in the event record.
\begin{subentry}
\iteme{= 0 :} off, i.e. only one event is generated at a time.
\iteme{= 1 :} on, i.e. several events are allowed in the same event
record. Information on the processes generated may be found in
\ttt{MSTI(41) - MSTI(50)}.
\end{subentry}
\iteme{MSTP(132) :} (D=4) the processes that are switched on for
pile-up events. The first event may be set up completely arbitrarily,
using the switches in the \ttt{PYSUBS} common block, while all the
subsequent events have to be of one of the `inclusive' processes
which dominate the cross section, according to the options below.
It is thus not possible to generate two rare events in the pile-up
option.
\begin{subentry}
\iteme{= 1 :} low-$p_{\perp}$ processes (ISUB = 95) only. The
low-$p_{\perp}$ model actually used, both in the hard event and in
the pile-up events, is the one set by \ttt{MSTP(81)} etc. This means
that implicitly also high-$p_{\perp}$ jets can be generated in the pile-up
events.
\iteme{= 2 :} low-$p_{\perp}$ + double diffractive processes
(ISUB = 95 and 94).
\iteme{= 3 :} low-$p_{\perp}$ + double diffractive + single diffractive
processes (ISUB = 95, 94, 93 and 92).
\iteme{= 4 :} low-$p_{\perp}$ + double diffractive + single diffractive
+ elastic processes, together corresponding to the full
hadron--hadron cross section (ISUB = 95, 94, 93, 92 and 91).
\end{subentry}
\iteme{MSTP(133) :} (D=0) multiplicity distribution of pile-up events.
\begin{subentry}
\iteme{= 0 :} selected by user, before each \ttt{PYEVNT} call, by
giving the \ttt{MSTP(134)} value.
\iteme{= 1 :} a Poissonian multiplicity distribution in the total
number of pile-up events. This is the relevant distribution if the
switches set for the first event in \ttt{PYSUBS} give the same
subprocesses as are implied by \ttt{MSTP(132)}. In that case the
mean number of events per beam crossing is
$\br{n} = \sigma_{\mrm{pile}} \times$\ttt{PARP(31)}, where
$\sigma_{\mrm{pile}}$ is the sum of the cross section for allowed
processes. Since bunch crossing which do not give any events at all
(probability $\exp(-\br{n})$) are not simulated, the actual average
number per \ttt{PYEVNT} call is
$\langle n \rangle = \br{n}/(1-\exp(-\br{n}))$.
\iteme{= 2 :} a biased distribution, as is relevant when one of the
events to be generated is assumed to belong to an event class
with a cross section much smaller than the total hadronic
cross section. If $\sigma_{\mrm{rare}}$ is the cross section for this
rare process (or the sum of the cross sections of several rare
processes) and $\sigma_{\mrm{pile}}$ the cross section for the
processes allowed by \ttt{MSTP(132)}, then define
$\br{n} = \sigma_{\mrm{pile}} \times$\ttt{PARP(131)}
and $f = \sigma_{\mrm{rare}}/\sigma_{\mrm{pile}}$. The probability
that a bunch crossing will give $i$ events is then
${\cal P}_i = f \, i \, \exp(-\br{n}) \, \br{n}^i/i!$,
i.e. the na\"{\i}ve Poissonian is suppressed by a factor $f$ since
one of the events will be rare rather than frequent, but
enhanced by a factor $i$ since any of the $i$ events may be the
rare one. Only beam crossings which give at least one event
of the required rare type are simulated, and the distribution
above normalized accordingly.
\itemc{Note:} for practical reasons, it is required that
$\br{n} < 120$, i.e. that an average beam crossing does not contain
more than 120 pile-up events. The multiplicity distribution is
truncated above 200, or when the probability for a multiplicity
has fallen below $10^{-6}$, whichever occurs sooner. Also low
multiplicities with probabilities below $10^{-6}$ are truncated.
See also \ttt{PARI(91) - PARI(93)}.
\end{subentry}
\iteme{MSTP(134) :} (D=1) a user selected multiplicity, i.e. total
number of pile-up events, to be generated in the next \ttt{PYEVNT}
call. May be reset for each new event, but must be in the range
$1 \leq$\ttt{MSTP(134)}$\leq 200$.
\boxsep
\iteme{PARP(81) :}\label{p:PARP81} (D=1.40 GeV/$c$) effective
minimum transverse momentum $p_{\perp\mathrm{min}}$ for multiple
interactions with \ttt{MSTP(82)=1}.
\iteme{PARP(82) :} (D=1.55 GeV/$c$) regularization scale $p_{\perp 0}$
of the transverse momentum spectrum for multiple interactions with
\ttt{MSTP(82)}$\geq 2$.
\iteme{PARP(83), PARP(84) :} (D=0.5, 0.2) parameters of an assumed
double Gaussian matter distribution inside the colliding hadrons for
\ttt{MSTP(82)=4}, of the form given in eq.~(\ref{mi:doubleGauss}),
i.e. with a core of radius \ttt{PARP(84)} of the main radius and
containing a fraction \ttt{PARP(83)} of the total hadronic matter.
\iteme{PARP(85) :} (D=0.33) probability that an additional
interaction in the multiple interaction formalism gives two gluons,
with colour connections to `nearest neighbours' in momentum space.
\iteme{PARP(86) :} (D=0.66) probability that an additional
interaction in the multiple interaction formalism gives two gluons,
either as described in \ttt{PARP(85)} or as a closed gluon loop.
Remaining fraction is supposed to consist of quark--antiquark pairs.
\iteme{PARP(87), PARP(88) :} (D=0.7, 0.5) in order to account for an
assumed dominance of valence quarks at low transverse momentum scales,
a probability is introduced that a $\mathrm{g}\g$-scattering according to
na\"{\i}ve cross section is replaced by a $\mathrm{q}\overline{\mathrm{q}}$ one; this is used
only for \ttt{MSTP(82)}$\geq 2$. The probability is parametrized as
${\cal P} = a (1 - (p_{\perp}^2/(p_{\perp}^2 + b^2)^2)$, where
$a =$\ttt{PARP(87)} and $b =$\ttt{PARP(88)}$\times$\ttt{PARP(82)}.
\iteme{PARP(91) :} (D=0.44 GeV/$c$) (C) width of Gaussian primordial
$k_{\perp}$ distribution inside hadron for \ttt{MSTP(91)=1}, i.e.
$\exp(-k_{\perp}^2/\sigma^2) \, k_{\perp} \, \d k_{\perp}$ with
$\sigma =$\ttt{PARP(91)} and
$\langle k_{\perp}^2 \rangle = $\ttt{PARP(91)}$^2$.
\iteme{PARP(92) :} (D=0.20 GeV/$c$) (C) width parameter of
exponential primordial $k_{\perp}$ distribution inside hadron for
\ttt{MSTP(91)=2}, i.e.
$\exp(-k_{\perp}/\sigma) \, k_{\perp} \, \d k_{\perp}$ with
$\sigma =$\ttt{PARP(92)} and
$\langle k_{\perp}^2 \rangle = 6 \times$\ttt{PARP(92)}$^2$.
Thus one should put \ttt{PARP(92)}$\approx$\ttt{PARP(91)}$/\sqrt{6}$
to have continuity with the option above.
\iteme{PARP(93) :} (D=2. GeV/$c$) (C) upper cut-off for primordial
$k_{\perp}$ distribution inside hadron.
\iteme{PARP(94) :} (D=1.) (C) for \ttt{MSTP(92)}$\geq 2$ this gives
the value of the parameter $k$ for the case when a meson or
resolved photon remnant is split into two fragments (which is which
is chosen at random).
\iteme{PARP(95) :} (D=0.) (C) for \ttt{MSTP(94)=2} this gives
the value of the parameter $k$ for the case when a meson or
resolved photon remnant is split into a meson and a spectator
fragment jet, with $\chi$ giving the energy fraction taken by the
meson.
\iteme{PARP(96) :} (D=3.) (C) for \ttt{MSTP(92)}$\geq 2$ this gives
the value of the parameter $k$ for the case when a nucleon remnant
is split into a diquark and a quark fragment, with $\chi$ giving
the energy fraction taken by the quark jet.
\iteme{PARP(97) :} (D=1.) (C) for \ttt{MSTP(94)=2} this gives
the value of the parameter $k$ for the case when a nucleon remnant
is split into a baryon and a quark jet or a meson and a diquark jet,
with $\chi$ giving the energy fraction taken by the quark jet or
meson, respectively.
\iteme{PARP(98) :} (D=0.75) (C) for \ttt{MSTP(92)=5} this gives
the power of an assumed basic $1/\chi^b$ behaviour in the splitting
distribution, with $b =$\ttt{PARP(98)}.
\iteme{PARP(99) :} (D=0.44 GeV/$c$) (C) width parameter of primordial
$k_{\perp}$ distribution inside photon; exact meaning depends on
\ttt{MSTP(93)} value chosen (cf. \ttt{PARP(91)} and \ttt{PARP(92)}
above).
\iteme{PARP(100) :} (D=2. GeV/$c$) (C) upper cut-off for primordial
$k_{\perp}$ distribution inside photon.
\iteme{PARP(131) :}\label{p:PARP131} (D=0.01 mb$^{-1}$) in the
pile-up events scenario, \ttt{PARP(131)}
gives the assumed luminosity per bunch--bunch crossing, i.e.
if a subprocess has a cross section $\sigma$, the average number
of events of this type per bunch--bunch crossing is
$\br{n} = \sigma \times$\ttt{PARP(131)}. \ttt{PARP(131)} may be
obtained by dividing the integrated luminosity over a given time
(1 s, say) by the number of bunch--bunch crossings that this
corresponds to. Since the program will not generate more than
200 pile-up events, the initialization procedure will crash if
$\br{n}$ is above 120.
\end{entry}
\clearpage
\section{Fragmentation}
The main fragmentation option in {\tsc{Jetset/Pythia}}
is the Lund string scheme, but independent fragmentation
options are also available. These latter options should not be
taken too seriously, since we know that independent fragmentation
does not provide a consistent alternative, but occasionally one
may like to compare string fragmentation with something else.
The subsequent four subsections give further details;
the first one on flavour selection, which is common to the two
approaches, the second on string fragmentation, the third on
independent fragmentation, while the fourth and final contains
information on a few other minor issues.
The Lund fragmentation model is described in \cite{And83}, where
all the basic ideas are presented and earlier papers
\cite{And79,And80,And82,And82a} summarized.
The details given there on how a multiparton jet system is allowed
to fragment are out of date, however, and for this one should turn
to \cite{Sjo84}. Also the `popcorn' baryon production mechanism
is not covered, see \cite{And85}. Reviews of fragmentation models
in general may be found in \cite{Sjo88,Sjo89}.
\subsection{Flavour Selection}
\label{ss:flavoursel}
In either string or independent fragmentation, an iterative
approach is used to describe the fragmentation process.
Given an initial quark $\mathrm{q} = \mathrm{q}_0$, it is assumed that a new
$\mathrm{q}_1 \overline{\mathrm{q}}_1$
pair may be created, such that a meson $\mathrm{q}_0 \overline{\mathrm{q}}_1$ is
formed, and a $\mathrm{q}_1$ is left behind. This $\mathrm{q}_1$ may at a later
stage pair off with a $\overline{\mathrm{q}}_2$, and so on. What need be
given is thus the relative probabilities to produce the various
possible $\mathrm{q}_i \overline{\mathrm{q}}_i$ pairs, $\u \overline{\mathrm{u}}$, $\d \overline{\mathrm{d}}$,
$\mathrm{s} \overline{\mathrm{s}}$, etc., and the relative probilities that a given
$\mathrm{q}_{i-1}\overline{\mathrm{q}}_i$ quark pair combination forms a specific meson,
e.g. for $\u \overline{\mathrm{u}}$ either $\pi^+$, $\rho^+$ or some higher state.
In {\tsc{Jetset}}, it is assumed that the two aspects can be factorized,
i.e. that it is possible first to select a $\mathrm{q}_i \overline{\mathrm{q}}_i$ pair,
without any reference to allowed physical meson states, and
that, once the $\mathrm{q}_{i-1} \overline{\mathrm{q}}_i$ flavour combination is given,
it can be assigned to a given meson state with total probability
unity.
\subsubsection{Quark flavours and transverse momenta}
In order to generate the quark--antiquark pairs $\mathrm{q}_i \overline{\mathrm{q}}_i$ which
lead to string breakups, the Lund model invokes the idea of
quantum mechanical tunnelling, as follows. If the $\mathrm{q}_i$ and
$\overline{\mathrm{q}}_i$ have no (common) mass or transverse momentum, the pair can
classically be created at one point and then be pulled apart
by the field. If the quarks have mass and/or transverse momentum,
however, the $\mathrm{q}_i$ and $\overline{\mathrm{q}}_i$ must classically be produced at a
certain distance so that the field energy between them can be
transformed into the sum of the two transverse masses $m_{\perp}$.
Quantum mechanically, the quarks may be created in one point
(so as to keep the concept of local flavour conservation) and then
tunnel out to the classically allowed region. In terms of a
common transverse mass $m_{\perp}$ of the $\mathrm{q}_i$ and the $\overline{\mathrm{q}}_i$,
the tunnelling probability is given by
\begin{equation}
\exp \left( -\frac{\pi m_{\perp}^2}{\kappa} \right) =
\exp \left( -\frac{\pi m^2}{\kappa} \right)
\exp \left( -\frac{\pi p_{\perp}^2}{\kappa} \right) ~.
\end{equation}
The factorization of the transverse momentum and the mass terms leads
to a flavour-independent Gaussian spectrum for the $p_x$ and $p_y$
components of $\mathrm{q}_i \overline{\mathrm{q}}_i$ pairs. Since the string is assumed to
have no transverse excitations, this $p_{\perp}$ is locally compensated
between the quark and the antiquark of the pair. The $p_{\perp}$ of a
meson $\mathrm{q}_{i-1} \overline{\mathrm{q}}_i$ is given by the vector sum of the $p_{\perp}$:s
of the $\mathrm{q}_{i-1}$ and $\overline{\mathrm{q}}_i$ constituents, which implies
Gaussians in $p_x$ and $p_y$ with a width $\sqrt{2}$ that of the
quarks themselves. The assumption of a Gaussian shape may be a good
first approximation, but there remains the possibility of non-Gaussian
tails, that can be important in some situations.
In a perturbative QCD framework, a hard scattering is
associated with gluon radiation, and further contributions to what is
na\"{\i}vely called fragmentation $p_{\perp}$ comes from unresolved radiation.
This is used as an explanation
why the experimental $\left\langle p_{\perp} \right\rangle$ is somewhat
higher than obtained with the formula above.
The formula also implies a suppression of heavy quark production
$u : d : s : c \approx$ \mbox{$1 : 1 : 0.3 : 10^{-11}$}. Charm and
heavier quarks
are hence not expected to be produced in the soft fragmentation.
Since the predicted flavour suppressions are in terms of quark masses,
which are notoriously difficult to assign (should it be current algebra,
or constituent, or maybe something in between?), the
suppression of $\mathrm{s} \overline{\mathrm{s}}$ production is left as a free parameter in
the program: $\u \overline{\mathrm{u}}$ : $\d \overline{\mathrm{d}}$ : $\mathrm{s} \overline{\mathrm{s}}$ = 1 : 1 : $\gamma_s$,
where by default $\gamma_s = 0.3$. At least qualitatively, the
experimental value agrees with theoretical prejudice.
There is no production at all of heavier flavours in the fragmentation
process, but only as part of the shower evolution.
\subsubsection{Meson production}
\label{sss:mesonprod}
Once the flavours $\mathrm{q}_{i-1}$ and $\overline{\mathrm{q}}_i$ have been selected, a choice
is made between the possible multiplets. The relative composition of
different multiplets is not given from first principles, but must
depend on the details of the fragmentation process. To some
approximation one would expect a negligible fraction of states with
radial excitations or non-vanishing orbital angular momentum. Spin
counting arguments would then suggest a 3:1 mixture between the
lowest lying vector and pseudoscalar multiplets. Wave function
overlap arguments lead to a relative enhancement of the lighter
pseudoscalar states, which is more pronounced the larger the mass
splitting is \cite{And82a}.
In the program, six meson multiplets are
included. If the nonrelativistic classification scheme is used, i.e.
mesons are assigned a valence quark spin $S$ and an internal orbital
angular momentum $L$, with the physical spin $s$ denoted $J$,
$\mbf{J} = \mbf{L} + \mbf{S}$, then the multiplets are:
\begin{Itemize}
\item $L = 0$, $S = 0$, $J = 0$: the ordinary pseudoscalar meson
multiplet;
\item $L = 0$, $S = 1$, $J = 1$: the ordinary vector meson multiplet;
\item $L = 1$, $S = 0$, $J = 1$: an axial vector meson multiplet;
\item $L = 1$, $S = 1$, $J = 0$: the scalar meson multiplet;
\item $L = 1$, $S = 1$, $J = 1$: another axial vector meson multiplet;
and
\item $L = 1$, $S = 1$, $J = 2$: the tensor meson multiplet.
\end{Itemize}
Each multiplet has the full four-generation setup of $8 \times 8$
states included in the program, although many could never actually be
produced. Some simplifications have been made; thus
there is no mixing included between the two axial vector multiplets.
In the program, the spin $S$ is first chosen to be either 0 or 1.
This is done according to parametrized relative probabilities,
where the probability for spin 1 by default is taken to be 0.5 for
a meson consisting only of $\u$ and $\d$ quark, 0.6 for one which
contains $\mathrm{s}$ as well, and $0.75$ for quarks with $\c$ or heavier
quark, in accordance with the deliberations above.
By default, it is assumed that $L = 0$, such that only pseudoscalar
and vector mesons are produced. For inclusion of $L = 1$ production,
four parameters can be used, one to give the probability that a $S = 0$
state also has $L =1$, the other three for the probability that a
$S = 1$ state has $L = 1$ and $J$ either 0, 1, or 2.
For the
flavour-diagonal meson states $\u \overline{\mathrm{u}}$, $\d \overline{\mathrm{d}}$ and $\mathrm{s} \overline{\mathrm{s}}$,
it is also necessary to include mixing into the physical mesons.
This is done according to a parametrization, based on the mixing angles
given in the Review of Particle Properties \cite{PDG88}. In particular,
the default choices correspond to
\begin{eqnarray}
\eta & = & \frac{1}{2} (\u\overline{\mathrm{u}} + \d\overline{\mathrm{d}}) - \frac{1}{\sqrt{2}}
\mathrm{s}\overline{\mathrm{s}} ~; \nonumber \\
\eta' & = & \frac{1}{2} (\u\overline{\mathrm{u}} + \d\overline{\mathrm{d}}) + \frac{1}{\sqrt{2}}
\mathrm{s}\overline{\mathrm{s}} ~; \nonumber \\
\omega & = & \frac{1}{\sqrt{2}} (\u\overline{\mathrm{u}} + \d\overline{\mathrm{d}}) \nonumber \\
\phi & = & \mathrm{s}\overline{\mathrm{s}} ~.
\end{eqnarray}
In the $\pi^0 - \eta - \eta'$ system, no account is therefore
taken of the difference in masses, an approximation which seems to
lead to an overestimate of $\eta'$ rates \cite{ALE92}. Recently,
parameters have been introduced to allow an additional `brute force'
suppression of $\eta$ and $\eta'$ states.
\subsubsection{Baryon production}
Baryon production may, in its simplest form, be obtained by assuming
that any flavour $\mathrm{q}_i$ given above could represent either a quark or
an antidiquark in a colour triplet state. Then the same basic
machinery can be run through as above, supplemented with the probability
to produce various diquark pairs. In principle, there is one
parameter for each diquark, but if tunnelling is still assumed to give
an effective description, mass relations can be used to reduce the
effective number of parameters. There are three main ones
appearing in the program:
\begin{Itemize}
\item the relative probability to pick a $\overline{\mathrm{q}}\qbar$ diquark
rather than a $\mathrm{q}$;
\item the extra suppression associated with a diquark
containing a strange quark (over and above the ordinary $\mathrm{s} / \u$
suppression factor $\gamma_s$); and
\item the suppression of spin 1 diquarks relative to spin 0 ones
(apart from the factor of 3 enhancement of the former based on
counting the number of spin states).
\end{Itemize}
The extra strange diquark suppression factor comes about since what
appears in the exponent of the tunnelling formula is $m^2$ and not
$m$, so that the diquark and the strange quark suppressions do not
factorize.
Only two baryon multiplets are included, i.e. there are no $L=1$
excited states. The two multiplets are:
\begin{Itemize}
\item $S = J = 1/2$: the `octet' multiplet of SU(3) (in the full
four-generation scenario in the program 168 states are available);
\item $S = J = 3/2$: the `decuplet' multiplet of SU(3) (120 states
in the program).
\end{Itemize}
In contrast to the meson case, different flavour combinations have
different numbers of states available: for $\u \u \u$ only
$\Delta^{++}$, whereas $\u \d \mathrm{s}$ may become either $\Lambda$,
$\Sigma^0$ or $\Sigma^{*0}$.
An important constraint is that a baryon is a
symmetric state of three quarks, neglecting the colour degree of
freedom. When a diquark and a quark are joined to form a baryon,
the combination is therefore weighted with the probability that
they form a symmetric three-quark state. The program implementation
of this principle is to first select a diquark at random, with
the strangeness and spin 1 suppression factors above included,
but then to accept the selected diquark with a weight proportional to
the number of states available for the quark-diquark combination. This
means that, were it not for the tunnelling suppression factors, all
states in the {\bf SU(6)} (flavour {\bf SU(3)} times spin {\bf SU(2)})
56-multiplet would become equally populated. Of course also heavier
baryons may come from the fragmentation of e.g. $\c$ quark jets, but
although the particle classification scheme used in the program is
{\bf SU(16)}, i.e. with eight flavours, all possible quark-diquark
combinations can be related to {\bf SU(6)} by symmetry arguments.
As in the case for mesons, one could imagine an explicit further
suppression of the heavier spin 3/2 baryons. We do not expect it to
be an important effect, since baryon mass splittings are much smaller
than in the meson case.
In case of rejection, a new diquark is selected and tested, etc.
A corresponding procedure is used for the quark selection when
a diquark has already been formed in the previous step. Properly
speaking both the quark and the diquark flavour should be chosen anew.
This would become a tedious process, since also the hadron produced
in the step before would have to be rejected. In practice only the
last produced pair, be that the quark or diquark one, is rejected.
The error introduced by this is small.
A more general framework for baryon production is the `popcorn' one
\cite{And85}, in which diquarks as such are never produced, but
rather baryons appear from the successive production of several
$\mathrm{q}_i \overline{\mathrm{q}}_i$ pairs. The picture is the following. Assume that the
original $\mathrm{q}$ is red $r$ and the $\overline{\mathrm{q}}$ is $\br{r}$. Normally a
new $\mathrm{q}_1 \overline{\mathrm{q}}_1$ pair produced in the field would also be
$r \br{r}$, so that the $\overline{\mathrm{q}}_1$ is pulled towards the $\mathrm{q}$
end and vice versa, and two separate colour-singlet systems
$\mathrm{q} \overline{\mathrm{q}}_1$ and $\mathrm{q}_1 \overline{\mathrm{q}}$ are formed. Occasionally, the
$\mathrm{q}_1 \overline{\mathrm{q}}_1$ pair may be e.g. $g \br{g}$ ($g$ = green), in which
case there is no net colour charge acting on either $\mathrm{q}_1$ or
$\overline{\mathrm{q}}_1$. Therefore, the pair cannot gain energy from the field,
and normally would exist only as a fluctuation. If $\mathrm{q}_1$ moves
towards $\mathrm{q}$ and $\overline{\mathrm{q}}_1$ towards $\overline{\mathrm{q}}$, the net field remaining
between $\mathrm{q}_1$ and $\overline{\mathrm{q}}_1$ is $\br{b} b$ ($b$ = blue;
$g + r = \br{b}$ if only colour triplets are assumed). In this central
field, an additional $\mathrm{q}_2 \overline{\mathrm{q}}_2$ pair can be created, where
$\mathrm{q}_2$ now is pulled towards $\mathrm{q} \mathrm{q}_1$ and $\overline{\mathrm{q}}_2$
towards $\overline{\mathrm{q}} \overline{\mathrm{q}}_1$, with no net colour field between $\mathrm{q}_2$
and $\overline{\mathrm{q}}_2$. If this is all that happens, the baryon $B$ will be
made up out of $\mathrm{q}_1$, $\mathrm{q}_2$ and some $\mathrm{q}_4$ produced between
$\mathrm{q}$ and $\mathrm{q}_1$, and $\br{B}$ of $\overline{\mathrm{q}}_1$, $\overline{\mathrm{q}}_2$ and some
$\overline{\mathrm{q}}_5$, i.e. the $B$ and $\br{B}$ will be nearest neighbours in
rank and share two quark pairs. Specifically, $\mathrm{q}_1$ will gain
energy from $\mathrm{q}_2$ in order to end up on mass shell, and the
tunnelling formula for an effective $\mathrm{q}_1 \mathrm{q}_2$ diquark is recovered.
Part of the time, several $b \br{b}$ colour pair productions
may take place between the $\mathrm{q}_1$ and $\overline{\mathrm{q}}_1$, however. With two
production vertices $\mathrm{q}_2 \overline{\mathrm{q}}_2$ and $\mathrm{q}_3 \overline{\mathrm{q}}_3$, a central
meson $\overline{\mathrm{q}}_2 \mathrm{q}_3$ may be formed, surrounded by a baryon
$\mathrm{q}_4 \mathrm{q}_1 \mathrm{q}_2$ and an antibaryon $\overline{\mathrm{q}}_3 \overline{\mathrm{q}}_1 \overline{\mathrm{q}}_5$.
We call this a $BM\br{B}$ configuration to distinguish it from the
$\mathrm{q}_4 \mathrm{q}_1 \mathrm{q}_2$ + $\overline{\mathrm{q}}_2 \overline{\mathrm{q}}_1 \overline{\mathrm{q}}_5$ $B\br{B}$ configuration
above. For $BM\br{B}$ the $B$ and $\br{B}$ only share one
quark--antiquark pair, as opposed to two for $B\br{B}$
configurations. The relative probability for a $BM\br{B}$
configuration is given by the uncertainty relation suppression for
having the $\mathrm{q}_1$ and $\overline{\mathrm{q}}_1$ sufficiently far apart that a meson
may be formed in between. Strictly speaking, also configurations
like $BMM\br{B}$, $BMMM\br{B}$, etc. should be possible, but the
probability for this is small in our model. Further, since larger
masses corresponds to longer string pieces, the production of
pseudoscalar mesons is favoured over that of vector ones. If only
$B\br{B}$ and $BM\br{B}$ states are included, and if the probability
for having a vector meson $M$ is not suppressed extra, two partly
compensating errors are made (since a vector meson typically
decays into two or more pseudoscalar ones).
In total, the flavour iteration procedure therefore contains the
following possible subprocesses (plus, of course, their charge
conjugates):
\begin{Itemize}
\item $\mathrm{q}_1 \to \mathrm{q}_2 + (\mathrm{q}_1 \overline{\mathrm{q}}_2)$ meson;
\item $\mathrm{q}_1 \to \overline{\mathrm{q}}_2\overline{\mathrm{q}}_3 + (\mathrm{q}_1 \mathrm{q}_2 \mathrm{q}_3)$ baryon;
\item $\mathrm{q}_1 \mathrm{q}_2 \to \overline{\mathrm{q}}_3 + (\mathrm{q}_1 \mathrm{q}_2 \mathrm{q}_3)$ baryon;
\item $\mathrm{q}_1 \mathrm{q}_2 \to \mathrm{q}_1 \mathrm{q}_3 + (\mathrm{q}_2 \overline{\mathrm{q}}_3)$ meson;
\end{Itemize}
with the constraint that the last process cannot be iterated to
obtain several mesons in between the baryon and the antibaryon.
Unfortunately, the resulting baryon production model has a fair
number of parameters, which would be given by the model only if
quark and diquark masses were known unambiguously.
We have already mentioned the $\mathrm{s} / \u$ ratio and the $\mathrm{q}\q / \mathrm{q}$
one; the latter has to be increased from 0.09 to 0.10 for the
popcorn model, since the total number of possible baryon
production configurations is lower in this case (the particle
produced between the $B$ and $\br{B}$ is constrained to be a
meson). For the popcorn model, exactly the same parameters as
already found in the diquark model are needed to describe the
$B\br{B}$ configurations. For $BM\br{B}$ configurations, the square
root of a suppression factor should be applied if the factor is
relevant only for one of the $B$ and $\br{B}$, e.g. if the $B$
is formed with a spin 1 `diquark' $\mathrm{q}_1 \mathrm{q}_2$ but the $\br{B}$
with a spin 0 diquark $\overline{\mathrm{q}}_1 \overline{\mathrm{q}}_3$. Additional parameters
include the relative probability for $BM\br{B}$ configurations,
which is assumed to be roughly 0.5 (with the remaining 0.5 being
$B\br{B}$), a suppression factor for having a strange meson $M$
between the $B$ and $\br{B}$ (as opposed to having a lighter
nonstrange one) and a suppression factor for having a $\mathrm{s} \overline{\mathrm{s}}$
pair (rather than a $\u \overline{\mathrm{u}}$ one) shared between the $B$ and
$\br{B}$ of a $BM\br{B}$ configuration. The default parameter
values are based on a combination of experimental observation
and internal model predictions.
In the diquark model, a diquark is expected to have exactly the
same transverse momentum distribution as a quark. For $BM\br{B}$
configurations the situation is somewhat more unclear, but we
have checked that various possibilities give very similar results.
The option implemented in the program is to assume no transverse
momentum at all for the $\mathrm{q}_1 \overline{\mathrm{q}}_1$ pair shared by the $B$ and
$\br{B}$, with all other pairs having the standard Gaussian
spectrum with local momentum conservation. This means that the
$B$ and $\br{B}$ $p_{\perp}$:s are uncorrelated in a $BM\br{B}$
configuration and (partially) anticorrelated in the $B\br{B}$
configurations, with the same mean transverse momentum for primary
baryons as for primary mesons.
Occasionally, the endpoint of a string is not a single parton,
but a diquark or antidiquark, e.g. when a quark has been kicked
out of a proton beam particle. One could consider fairly complex
schemes for the resulting fragmentation. One such \cite{And81}
was available in {\tsc{Jetset}} version 6 but is no longer found in version
7. Instead the same basic scheme is used as for diquark pair
production above. Thus a $\mathrm{q}\q$ diquark endpoint is let to fragment
just as would a $\mathrm{q}\q$ produced in the field behind a matching
$\overline{\mathrm{q}}\qbar$ flavour, i.e. either the two quarks of the diquark enter
into the same leading baryon, or else a meson is first produced,
containing one of the quarks, while the other is contained in the
baryon produced in the next step.
\subsection{String Fragmentation}
An iterative procedure can also be used for other aspects of the
fragmentation. This is possible because, in the string picture,
the various points where the string break by the production of
$\mathrm{q} \overline{\mathrm{q}}$ pairs are causally disconnected. Whereas the space--time
picture in the c.m. frame is such that slow particles
(in the middle of the system) are formed first, this ordering is
Lorentz frame dependent and hence irrelevant. One may therefore
make the convenient choice of starting an iteration process at
the ends of the string and proceeding towards the middle.
The string fragmentation scheme is rather complicated for a generic
multiparton state. In order to simplify the discussion, we will
therefore start with the simple $\mathrm{q} \overline{\mathrm{q}}$ process, and only later
survey the complications that appear when additional gluons
are present. (This distinction is made for pedagogical reasons,
in the program there is only one general-purpose algorithm).
\subsubsection{Fragmentation functions}
Assume a $\mathrm{q} \overline{\mathrm{q}}$ jet system, in its c.m. frame, with the quark moving
out in the $+z$ direction and the antiquark in the $-z$ one. We have
discussed how it is possible to start the flavour iteration from the
$\mathrm{q}$ end, i.e. pick a $\mathrm{q}_1 \overline{\mathrm{q}}_1$ pair, form a hadron $\mathrm{q} \overline{\mathrm{q}}_1$,
etc. It has also been noted that the tunnelling mechanism
is assumed to give a transverse momentum $p_{\perp}$ for each new
$\mathrm{q}_i\overline{\mathrm{q}}_i$ pair created, with the $p_{\perp}$ locally compensated between
the $\mathrm{q}_i$ and the $\overline{\mathrm{q}}_i$ member of the pair, and with a Gaussian
distribution in $p_x$ and $p_y$ separately. In the program, this is
regulated by one parameter, which gives the root-mean-square $p_{\perp}$
of a quark. Hadron transverse momenta are obtained as the sum of
$p_{\perp}$:s of the constituent $\mathrm{q}_i$ and $\overline{\mathrm{q}}_{i+1}$, where a diquark
is considered just as a single quark.
What remains to be determined is the energy and longitudinal
momentum of the hadron. In fact, only one variable can be selected
independently, since the momentum of the hadron is constrained by
the already determined hadron transverse mass $m_{\perp}$,
\begin{equation}
(E+p_z)(E-p_z) = E^2 - p_z^2 = m_{\perp}^2 = m^2 + p_x^2 + p_y^2 ~.
\label{fr:massconstr}
\end{equation}
In an iteration from the quark end, one is led (by the desire for
longitudinal boost invariance and other considerations)
to select the $z$ variable as the fraction of
$E+p_z$ taken by the hadron, out of the available $E+p_z$.
As hadrons are split off, the $E+p_z$ (and $E-p_z$) left for
subsequent steps is reduced accordingly:
\begin{eqnarray}
(E+p_z)_{\mrm{new}} & = & (1-z) (E+p_z)_{\mrm{old}} ~, \nonumber \\
(E-p_z)_{\mrm{new}} & = & (E-p_z)_{\mrm{old}} -
\frac{m_{\perp}^2}{z (E+p_z)_{\mrm{old}}} ~.
\end{eqnarray}
The fragmentation function $f(z)$, which expresses the probability
that a given $z$ is picked, could in principle be arbitrary --- indeed,
several such choices can be used inside the program, see below.
If one, in addition, requires that the fragmentation
process as a whole should look the same,
irrespectively of whether the iterative procedure is performed
from the $\mathrm{q}$ end or the $\overline{\mathrm{q}}$ one, `left--right symmetry',
the choice is essentially unique \cite{And83a}: the `Lund symmetric
fragmentation function',
\begin{equation}
f(z) \propto \frac{1}{z} z^{a_{\alpha}} \left( \frac{1-z}{z}
\right)^{a_{\beta}} \exp \left( - \frac{bm_{\perp}^2}{z}
\right) ~.
\label{fr:LSFFlong}
\end{equation}
There is one separate parameter $a$ for each flavour,
with the index $\alpha$ corresponding to the `old' flavour in the
iteration process, and $\beta$ to the `new' flavour.
It is customary to put all $a_{\alpha,\beta}$ the same, and thus
arrive at the simplified expression
\begin{equation}
f(z) \propto z^{-1} (1-z)^a \exp (-bm_{\perp}^2/z) ~.
\label{fr:LSFF}
\end{equation}
In the program, only two separate $a$ values can be given, that
for quark pair production and that for diquark one; by default the
two are taken to be the same. In addition, there is the $b$ parameter,
which is universal.
It should be noted that the explicit mass dependence in $f(z)$ implies
a harder fragmentation function for heavier hadrons; the
asymptotic behaviour of the mean $z$ value for heavy hadrons is
\begin{equation}
\langle z \rangle \approx 1 - \frac{1+a}{bm_{\perp}^2} ~.
\end{equation}
Unfortunately it seems this predicts a somewhat harder spectrum
for $\mathrm{B}$ mesons than observed in data.
For future reference we note that the derivation of $f(z)$ as a
by-product also gives the probability distribution in proper
time $\tau$ of $\mathrm{q}_i \overline{\mathrm{q}}_i$ breakup vertices. In terms of
$\Gamma = (\kappa \tau)^2$, this distribution is
\begin{equation}
{\cal P}(\Gamma) \, \d \Gamma \propto \Gamma^a \, \exp(-b \Gamma) \,
\d \Gamma ~,
\end{equation}
with the same $a$ and $b$ as above.
Many different other fragmentation functions have been proposed,
and a few are available as options in the program.
\begin{Itemize}
\item The Field-Feynman parametrization \cite{Fie78},
\begin{equation}
f(z) = 1 - a + 3a(1-z)^2 ~,
\end{equation}
with default value $a = 0.77$, is intended to be used only for
ordinary hadrons made out of $\u$, $\d$ and $\mathrm{s}$ quarks.
\item Since there are indications that the shape above is too
strongly peaked at $z = 0$, instead a shape like
\begin{equation}
f(z) = (1+c) (1-z)^c
\end{equation}
may be used.
\item Charm and bottom data clearly indicate the need for a
harder fragmentation function for heavy flavours.
The best known of these is the Peterson et al.~formula \cite{Pet83}
\begin{equation}
f(z) \propto \frac{1}{ z \left( 1 -
\frac{\displaystyle 1}{\displaystyle z} -
\frac{\displaystyle \epsilon_Q}{\displaystyle 1-z}
\right)^2 } ~,
\label{fr:PetHF}
\end{equation}
where $\epsilon_Q$ is a free parameter, expected to scale between
flavours like $\epsilon_Q \propto 1/m_Q^2$.
\item As a crude alternative, that is also peaked at $z=1$, one may
use
\begin{equation}
f(z) = (1+c) z^c ~.
\end{equation}
\item Bowler \cite{Bow81} has shown, within the framework of the
Artru--Mennessier model \cite{Art74}, that a massive endpoint quark
with mass $m_Q$ leads to a modification of the symmetric
fragmentation function, due to
the fact that the string area swept out is reduced for massive endpoint
quarks, compared with massless ditto. The Artru--Mennessier model in
principle only applies for clusters with a continuous mass spectrum,
and does not allow an $a$ term (i.e. $a \equiv 0$); however, it has
been shown \cite{Mor89} that, for a discrete mass spectrum, one may
still retain an effective $a$ term. In the program an approximate
form with an $a$ term has therefore been used:
\begin{equation}
f(z) \propto \frac{1}{z^{1 + r_Q b m_Q^2}}
z^{a_{\alpha}} \left( \frac{1-z}{z}
\right)^{a_{\beta}} \exp \left( - \frac{b m_{\perp}^2}{z} \right) ~.
\label{fr:LSFFBowler}
\end{equation}
In principle the prediction is that $r_Q \equiv 1$, but so as to
be able to extrapolate smoothly between this form and the Lund
symmetric one, it is possible to pick $r_Q$ separately for $\c$, $\b$
and $\t$ hadrons.
\end{Itemize}
\subsubsection{Joining the jets}
The $f(z)$ formula above is only valid, for the breakup of a jet
system into a hadron plus a remainder-system, when the remainder
mass is large. If the fragmentation algorithm were to be used all the
way from the $\mathrm{q}$ end to the $\overline{\mathrm{q}}$ one, the mass of the last hadron to
be formed at the $\overline{\mathrm{q}}$ end would be completely constrained by energy
and momentum conservation, and could not be on its mass shell. In theory
it is known how to take such effects into account, but the resulting
formulae are wholly unsuitable for Monte Carlo implementation.
The practical solution to this problem is to carry out the
fragmentation both from the $\mathrm{q}$ and the $\overline{\mathrm{q}}$ end, such that for
each new step in the fragmentation process, a random
choice is made as to from what side the step is to be taken.
If the step is on the $\mathrm{q}$ side, then $z$ is interpreted as fraction
of the remaining $E+p_z$ of the system, while $z$ is interpreted as
$E-p_z$ fraction for a step from the $\overline{\mathrm{q}}$ end. At some point, when
the remaining mass of the system has dropped below a given value,
it is decided that the next breakup will produce two final hadrons,
rather than a hadron and a remainder-system.
Since the momenta of two hadrons are to be selected, rather than
that of one only, there are enough degrees of freedom to have both
total energy and total momentum completely conserved.
The mass at which the normal fragmentation
process is stopped and the final two hadrons formed is not actually
a free parameter of the model: it is given by the requirement that
the string everywhere looks the same, i.e. that the rapidity spacing
of the final two hadrons, internally and with respect to surrounding
hadrons, is the same as elsewhere in the fragmentation process.
The stopping mass, for a given setup of fragmentation parameters,
has therefore been determined in separate runs. If the fragmentation
parameters are changed, some retuning should be done but, in practice,
reasonable changes can be made without any special arrangements.
Consider a fragmentation process which has already split off a number
of hadrons from the $\mathrm{q}$ and $\overline{\mathrm{q}}$ sides, leaving behind a
a $\mathrm{q}_i \overline{\mathrm{q}}_j$ remainder system. When this system breaks by the
production of a $\mathrm{q}_n \overline{\mathrm{q}}_n$ pair, it is decided to make this pair
the final one, and produce the last two hadrons $\mathrm{q}_i\overline{\mathrm{q}}_n$ and
$\mathrm{q}_n\overline{\mathrm{q}}_j$, if
\begin{equation}
( (E+p_z)(E-p_z) )_{\mrm{remaining}} = W_{\mrm{rem}}^2 <
W_{\mathrm{min}}^2 ~.
\end{equation}
The $W_{\mathrm{min}}$ is calculated according to
\begin{equation}
W_{\mathrm{min}} = ( W_{\mathrm{min} 0} + m_{\mathrm{q} i} + m_{\mathrm{q} j} + k \, m_{\mathrm{q} n} )
\, (1 \pm \delta) ~.
\end{equation}
Here $W_{\mathrm{min} 0}$ is the main free parameter, typically around
1 GeV, determined to give a flat
rapidity plateau (separately for each particle species), while the
default $k = 2$ corresponds to the mass of the final pair being taken
fully into account. Smaller values may also be considered, depending
on what criteria are used to define the `best' joining of the $\mathrm{q}$
and the $\overline{\mathrm{q}}$ chain. The factor $1 \pm \delta$, by default evenly
distributed between 0.8 and 1.2, signifies a smearing of the
$W_{\mathrm{min}}$ value, to avoid an abrupt and unphysical cut-off in
the invariant mass distribution of the final two hadrons. Still,
this distribution will be somewhat different from that of any two
adjacent hadrons elsewhere. Due to the cut there will be no tail up to
very high masses; there are also fewer events close to the lower limit,
where the two hadrons are formed at rest with respect to each other.
This procedure does not work all that well for heavy flavours, since it
does not fully take into account the harder fragmentation function
encountered. Therefore, in addition to the check above, one further
test is performed for charm and heavier flavours, as follows. If the
check above allows more particle production,
a heavy hadron $\mathrm{q}_i \overline{\mathrm{q}}_n$ is formed, leaving a
remainder $\mathrm{q}_n \overline{\mathrm{q}}_j$. The range of allowed $z$ values, i.e. the
fraction of remaining $E+p_z$ that may be taken by the
$\mathrm{q}_i \overline{\mathrm{q}}_n$ hadron, is constrained away from 0 and 1 by the
$\mathrm{q}_i \overline{\mathrm{q}}_n$ mass and minimal mass of the $\mathrm{q}_n \overline{\mathrm{q}}_j$ system.
The limits of the physical $z$ range is obtained when the
$\mathrm{q}_n \overline{\mathrm{q}}_j$ system only consists of one single particle, which then
has a well-determined
transverse mass $m_{\perp}^{(0)}$. From the $z$ value obtained with the
infinite-energy fragmentation function formulae, a rescaled $z'$ value
between these limits is given by
\begin{equation}
z' = \frac{1}{2} \left\{ 1 + \frac{m_{\perp i n}^2}{W_{\mrm{rem}}^2} -
\frac{m_{\perp n j}^{(0)2}}{W_{\mrm{rem}}^2} +
\sqrt{ \left( 1 - \frac{m_{\perp i n}^2}{W_{\mrm{rem}}^2} -
\frac{m_{\perp n j}^{(0)2}}{W_{\mrm{rem}}^2} \right)^2
- 4 \frac{m_{\perp i n}^2}{W_{\mrm{rem}}^2}
\frac{m_{\perp n j}^{(0)2}}{W_{\mrm{rem}}^2} }
\, (2 z -1) \right\} ~.
\end{equation}
From the $z'$ value, the actual transverse mass
$m_{\perp n j} \geq m_{\perp n j}^{(0)}$ of the $\mathrm{q}_n \overline{\mathrm{q}}_j$
system may be calculated. For more than one particle
to be produced out of this system, the requirement
\begin{equation}
m_{\perp n j}^2 = (1-z') \, \left( W_{\mrm{rem}}^2 -
\frac{m_{\perp i n}^2}{z'} \right) > (m_{q j} + W_{\mathrm{min} 0})^2
+ p_{\perp}^2
\end{equation}
has to be fulfilled. If not, the $\mathrm{q}_n \overline{\mathrm{q}}_j$ system is assumed to
collapse to one single particle.
The consequence of the procedure above is that, the more the infinite
energy fragmentation function $f(z)$ is peaked close to $z=1$, the more
likely it is that only two particles are produced. In particular, for
$\t \overline{\mathrm{t}}$ systems, where very large $\langle z \rangle$ values are
predicted, the
expectation is that two particle final states will dominate far above
the threshold region. The procedure above has been constructed so that
the two particle fraction can be calculated directly from the shape of
$f(z)$ and the (approximate) mass spectrum, but it is not unique.
For the symmetric Lund fragmentation function, a number of
alternatives tried all give essentially the same result, whereas other
fragmentation functions may be more sensitive to details.
Assume now that two final hadrons have been picked. If the
transverse mass of the remainder-system is smaller than the sum of
transverse masses of the final two hadrons, the whole fragmentation
chain is rejected, and started over from the $\mathrm{q}$ and $\overline{\mathrm{q}}$
endpoints. This does not introduce any significant bias, since the
decision to reject a fragmentation chain only depends on what happens
in the very last step, specifically that the next-to-last step took
away too much energy, and not on what happened in the steps before
that.
If, on the other hand, the remainder-mass is large enough, there
are two kinematically allowed solutions for the final two hadrons:
the two mirror images in the rest frame of the remainder-system. Also
the choice between these two solutions is given by the consistency
requirements, and can be derived from studies of infinite energy jets.
The probability for the reverse ordering, i.e. where
the rapidity and the flavour orderings disagree, is parametrized by
\begin{equation}
{\cal P}_{\mrm{reverse}} = \frac{1}{2} \left(
\frac{m_{\perp i n} + m_{\perp n j}}{W_{\mrm{rem}}} \right)^d ~.
\label{fr:revord}
\end{equation}
For symmetric fragmentation, the ordering is expected to be
increasingly strict when the particles involved are more massive.
In the program it is therefore assumed that $d$ is a function of the
masses, $d = d_0 (m_{\perp i n} + m_{\perp n j})^2$, where $d_0$ is a
free parameter.
When baryon production is included, some particular problems arise.
First consider $B\br{B}$ situations. In the na\"{\i}ve iterative scheme,
away from the middle of the event, one
already has a quark and is to chose a matching diquark flavour or the
other way around. In either case the choice of the new flavour can
be done taking into account the number of {\bf SU(6)} states available
for the quark-diquark combination. For a
case where the final $\mathrm{q}_n \overline{\mathrm{q}}_n$ breakup is an
antidiquark-diquark one, the weights for forming $\mathrm{q}_i \overline{\mathrm{q}}_n$ and
$\mathrm{q}_n \overline{\mathrm{q}}_i$ enter at the same time, however. We do not know how to
handle this problem; what is done is to use weights as usual for the
$\mathrm{q}_i \overline{\mathrm{q}}_n$ baryon to select $\mathrm{q}_n$, but then consider
$\mathrm{q}_n \overline{\mathrm{q}}_i$ as given (or the other way around with equal
probability). If $\mathrm{q}_n \overline{\mathrm{q}}_i$ turns out to be
an antidiquark-diquark combination, the whole fragmentation chain is
rejected, since we do not know how to form corresponding hadrons.
A similar problem arises, and is solved in the same spirit, for a
$BM\br{B}$ configuration in which the $B$ (or $\br{B}$) was chosen
as third-last particle. When only two particles remain to be
generated, it is obviously too late to consider having a $BM\br{B}$
configuration. This is as it should, however, as can be found by
looking at all possible ways a hadron of given rank can be a baryon.
While some practical compromises have to be accepted in the
joining procedure, the fact that the joining takes place in
different parts of the string in different events means that,
in the end, essentially no visible effects remain.
\subsubsection{String motion and infrared stability}
We have now discussed the SF scheme for the fragmentation of a simple
$\mathrm{q} \overline{\mathrm{q}}$ jet system. In order to understand how these results
generalize
to arbitrary jet systems, it is first necessary to understand the
string motion for the case when no fragmentation takes place. In the
following we will assume that quarks as well as gluons are massless,
but all arguments can be generalized to massive quarks without too
much problem.
For a $\mathrm{q} \overline{\mathrm{q}}$ event viewed in the c.m. frame, with total energy $W$,
the partons start moving out back-to-back, carrying half the energy
each. As they move apart, energy and momentum is lost to the string.
When the partons are a distance $W / \kappa$ apart, all the energy
is stored in the string. The partons now turn around and come
together again with the original momentum vectors reversed. This
corresponds to half a period of the full string motion; the second
half the process is repeated, mirror-imaged. For further
generalizations to multiparton systems, a convenient description of
the energy and momentum flow is given in terms of `genes'
\cite{Art83}, infinitesimal packets of the four-momentum given up
by the partons to the string. Genes with $p_z = E$, emitted from the
$\mathrm{q}$ end in the initial stages of the string motion above,
will move in the $\overline{\mathrm{q}}$ direction with the speed of light, whereas
genes with $p_z = -E$ given up by the $\overline{\mathrm{q}}$ will move in the $\mathrm{q}$
direction. Thus, in this simple case, the direction of motion for a
gene is just opposite to that of a free particle with the same
four-momentum. This is due to the string tension. If the system is
not viewed in the c.m. frame, the rules are that any parton gives up
genes with four-momentum proportional to its own four-momentum, but
the direction of motion of any gene is given by the momentum direction
of the genes it meets, i.e. that were emitted by the parton at the
other end of that particular string piece. When the $\mathrm{q}$ has lost
all its energy, the $\overline{\mathrm{q}}$ genes, which before could not catch up
with $\mathrm{q}$, start impinging on it, and the $\mathrm{q}$ is pulled back,
accreting $\overline{\mathrm{q}}$ genes in the process. When the $\mathrm{q}$ and $\overline{\mathrm{q}}$
meet in the origin again, they have completely traded genes with
respect to the initial situation.
A 3-jet $\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ event initially corresponds to having a
string piece stretched between $\mathrm{q}$ and $\mathrm{g}$ and another between
$\mathrm{g}$ and $\overline{\mathrm{q}}$. Gluon four-momentum genes are thus flowing towards
the $\mathrm{q}$ and $\overline{\mathrm{q}}$. Correspondingly, $\mathrm{q}$ and $\overline{\mathrm{q}}$ genes are
flowing towards the $\mathrm{g}$. When the gluon has lost all its energy,
the $\mathrm{g}$ genes continue moving apart, and instead a third
string region is formed in the `middle' of the total string,
consisting of overlapping $\mathrm{q}$ and $\overline{\mathrm{q}}$ genes. The two `corners'
on the string, separating the three string regions, are not of the
gluon-kink type: they do not carry any momentum.
If this third region would only appear at a time later than the typical
time scale for fragmentation, it could not affect the sharing of energy
between different particles. This is true in the limit of high energy,
well separated partons. For a small gluon energy, on the other hand, the
third string region appears early, and the overall drawing of the string
becomes fairly 2-jet-like, since the third string region consists of
$\mathrm{q}$ and $\overline{\mathrm{q}}$ genes and therefore behaves exactly as a sting pulled
out directly between the $\mathrm{q}$ and $\overline{\mathrm{q}}$.
In the limit of vanishing gluon energy,
the two initial string regions collapse to naught, and the ordinary
2-jet event is recovered. Also for a collinear gluon, i.e.
$\theta_{\mathrm{q} \mathrm{g}}$ (or $\theta_{\overline{\mathrm{q}} \mathrm{g}}$) small, the stretching
becomes 2-jet-like. In particular, the $\mathrm{q}$ string endpoint first
moves out a distance $\mbf{p}_{\mathrm{q}} / \kappa$ losing genes to the
string, and then a further distance $\mbf{p}_{\mathrm{g}} / \kappa$, a first
half accreting genes from the $\mathrm{g}$ and the second half re-emitting
them. (This latter half actually includes yet another
string piece; a corresponding piece appears at the $\overline{\mathrm{q}}$ end, such
that half a period of the system involves five different string
regions.) The end result is, approximately, that a string is drawn out
as if there had only been a single parton with energy
$|\mbf{p}_{\mathrm{q}} + \mbf{p}_{\mathrm{g}}|$, such that the simple 2-jet event
again is recovered in the limit $\theta_{\mathrm{q} \mathrm{g}} \to 0$. These
properties of the string motion are the reason why
the string fragmentation scheme is `infrared safe' with respect to
soft or collinear gluon emission.
The discussions for the 3-jet case can be generalized to the motion
of a string with $\mathrm{q}$ and $\overline{\mathrm{q}}$ endpoints and an arbitrary number of
intermediate gluons. For $n$ partons, whereof $n-2$
gluons, the original string contains $n-1$ pieces. Anytime one of the
original gluons has lost its energy, a new string region is formed,
delineated by a pair of `corners'.
As the extra `corners' meet each other, old string regions vanish
and new are created, so that half a period of the string contains
$2n^2 - 6n + 5$ different string regions. Each of these regions can
be understood simply as built up from the overlap of (opposite-moving)
genes from two of the original partons, according to well specified
rules.
\subsubsection{Fragmentation of multiparton systems}
The full machinery needed for a multiparton system is very
complicated, and is described in detail in \cite{Sjo84}. The
following outline is far from complete, and is complicated
nonetheless. The main message to be conveyed is that a
Lorentz covariant algorithm exists for handling an arbitrary parton
configuration, but that the necessary machinery is more complex
than in either cluster or independent fragmentation.
Assume $n$ partons, with ordering along the string, and related
four-momenta, given by
$\mathrm{q}(p_1) \mathrm{g}(p_2) \mathrm{g}(p_3) \cdots \mathrm{g}(p_{n-1}) \overline{\mathrm{q}}(p_n)$.
The initial string then contains $n-1$ separate pieces.
The string piece between the quark and its neigbouring gluon is, in
four-momentum space, spanned by one side with four-momentum
$p_+^{(1)} = p_1$ and another with $p_-^{(1)} = p_2/2$. The factor of
1/2 in the second expression
comes from the fact that the gluon shares its energy between two
string pieces. The indices `$+$' and `$-$' denotes direction towards
the $\mathrm{q}$ and $\overline{\mathrm{q}}$ end, respectively. The next string piece,
counted from the quark end, is spanned by $p_+^{(2)} = p_2/2$ and
$p_-^{(2)} = p_3/2$, and so on, with the last one being
$p_+^{(n-1)} = p_{n-1}/2$ and $p_-^{(n-1)} = p_n$.
For the algorithm to work, it is important
that all $p_{\pm}^{(i)}$ be light-cone-like, i.e. $p_{\pm}^{(i)2} = 0$.
Since gluons are massless, it is only the two endpoint quarks which
can cause problems. The procedure here is to create new $p_{\pm}$
vectors for each of the two endpoint regions, defined to be linear
combinations of the old $p_{\pm}$ ones for the same region, with
coefficients determined so that the
new vectors are light-cone-like. De facto, this corresponds to
replacing a massive quark at the end of a string piece with a massless
quark at the end of a somewhat longer string piece. With the exception
of the added fictitious piece, which anyway ends up entirely within
the heavy hadron produced from the heavy quark, the string motion
remains unchanged by this.
In the continued string motion, when new string regions appear as
time goes by, the first such string regions that appear can be
represented as being spanned by one $p_+^{(j)}$ and another
$p_-^{(k)}$ four-vector, with $j$ and $k$ not necessarily
adjacent. For instance, in the $\mathrm{q} \mathrm{g} \overline{\mathrm{q}}$
case, the `third' string region is spanned by $p_+^{(1)}$ and
$p_-^{(3)}$. Later on in the string evolution history,
it is also possible to have regions made up of two $p_+$ or two
$p_-$ momenta. These appear when an endpoint quark has
lost all its original momentum, has accreted the momentum of an
gluon, and is now re-emitting this momentum. In practice, these
regions may be neglected. Therefore only pieces made up by a
$(p_+^{(j)},p_-^{(k)})$ pair of momenta are considered in the
program.
The allowes string regions may be ordered in an abstract parameter
plane, where the $(j,k)$ indices of the four-momentum pairs define
the position of each region along the two (parameter plane)
coordinate axes. In this plane the fragmentation procedure can be
described as a sequence of steps, starting at the quark end,
where $(j,k) = (1,1)$, and ending at the antiquark one,
$(j,k) = (n-1,n-1)$. Each step is taken from an `old'
$\mathrm{q}_{i-1} \overline{\mathrm{q}}_{i-1}$ pair production vertex, to the
production vertex of a `new' $\mathrm{q}_i \overline{\mathrm{q}}_i$
pair, and the string piece between these two string
breaks represent a hadron. Some steps may be taken within one and
the same region, while others may have one vertex in one region
and the other vertex in another region. Consistency requirements,
like energy-momentum conservation, dictates that vertex $j$ and
$k$ region values be ordered in a monotonic sequence, and that
the vertex positions are monotonically ordered inside each region.
The four-momentum of each hadron can be read off, for $p_+$ ($p_-$)
momenta, by projecting the separation between the old and the new
vertex on to the $j$ ($k$) axis. If the four-momentum fraction of
$p_{\pm}^{(i)}$ taken by a hadron is denoted $x_{\pm}^{(i)}$,
then the total hadron four-momentum is given by
\begin{equation}
p = \sum_{j=j_1}^{j_2} x_+^{(j)} p_+^{(j)} +
\sum_{k=k_1}^{k_2} x_-^{(k)} p_-^{(k)} +
p_{x1} \hat{e}_x^{(j_1 k_1)} + p_{y1} \hat{e}_y^{(j_1 k_1)} +
p_{x2} \hat{e}_x^{(j_2 k_2)} + p_{y2} \hat{e}_y^{(j_2 k_2)} ~,
\label{fr:fourmom}
\end{equation}
for a step from region $(j_1,k_1)$ to region $(j_2,k_2)$.
By necessity, $x_+^{(j)}$ is unity for a $j_1 < j < j_2$,
and correspondingly for $x_-^{(k)}$.
The $(p_x,p_y)$ pairs are the transverse momenta produced at
the two string breaks, and the $(\hat{e}_x,\hat{e}_y)$ pairs
four-vectors transverse to the string directions in the regions
of the respective string breaks:
\begin{eqnarray}
& & \hat{e}_x^{(jk)2} = \hat{e}_y^{(jk)2} = -1 ~, \nonumber \\
& & \hat{e}_x^{(jk)} \hat{e}_y^{(jk)} = \hat{e}_{x,y}^{(jk)}
p_+^{(j)} = \hat{e}_{x,y}^{(jk)} p_-^{(k)} = 0 ~.
\end{eqnarray}
The fact that the hadron should be on mass shell, $p^2 = m^2$, puts
one constraint on where a new breakup may be, given that the old
one is already known, just as eq.~(\ref{fr:massconstr}) did in the
simple 2-jet case. The remaining degree of freedom is, as before,
to be given by
the fragmentation function $f(z)$. The interpretation of the $z$
is only well-defined for a step entirely constrained to one of the
initial string regions, however, which is not enough. In the
2-jet case, the $z$ values can be related to the proper times
of string breaks, as follows. The variable
$\Gamma = (\kappa \tau)^2$, with $\kappa$ the string tension and
$\tau$ the proper time between the production vertex of the
partons and the breakup point, obeys an iterative relation of the
kind
\begin{eqnarray}
\Gamma_0 & = & 0 ~, \nonumber \\
\Gamma_i & = & (1-z_i) \left( \Gamma_{i-1} + \frac{m_{\perp i}^2}{z_i}
\right) ~.
\end{eqnarray}
Here $\Gamma_0$ represents the value at the $\mathrm{q}$ and $\overline{\mathrm{q}}$ endpoints,
and $\Gamma_{i-1}$ and $\Gamma_i$ the values
at the old and new breakup vertices needed to produce
a hadron with transverse mass $m_{\perp i}$, and with the $z_i$ of the
step chosen according to $f(z_i)$. The proper time can be
defined in an unambiguous way, also over boundaries between the
different string regions, so for multijet events the $z$ variable may
be interpreted just as
an auxiliary variable needed to determine the next $\Gamma$ value.
(In the Lund symmetric fragmentation function derivation, the
$\Gamma$ variable actually does appear naturally, so the choice
is not as arbitrary as it may seem here.)
The mass and $\Gamma$ constraints together are sufficient to determine
where the next string breakup is to be chosen, given the preceding
one in the iteration scheme. Actually, several ambiguities remain,
but are of no importance for the overall picture.
The algorithm for finding the next breakup then works something
like follows. Pick a hadron, $p_{\perp}$, and $z$, and calculate the next
$\Gamma$. If the old breakup is in the region $(j,k)$, and if the
new breakup is also assumed to be in the same region, then the
$m^2$ and $\Gamma$ constraints can be reformulated in terms of
the fractions $x_+^{(j)}$ and $x_-^{(k)}$ the hadron must take of the
total four-vectors $p_+^{(j)}$ and $p_-^{(k)}$:
\begin{eqnarray}
m^2 &=&
c_1 + c_2 x_+^{(j)} + c_3 x_-^{(k)} + c_4 x_+^{(j)} x_-^{(k)} ~,
\nonumber \\
\Gamma &=&
d_1 + d_2 x_+^{(j)} + d_3 x_-^{(k)} + d_4 x_+^{(j)} x_-^{(k)} ~.
\label{fr:mGa}
\end{eqnarray}
Here the coefficients
$c_n$ are fairly simple expressions, obtainable by squaring
eq.~(\ref{fr:fourmom}), while $d_n$ are slightly more complicated
in that they depend on the position of the old string break,
but both the $c_n$ and the $d_n$ are explicitly calculable. What
remains is an equation system with two unknowns, $x_+^{(j)}$
and $x_-^{(k)}$.
The absence of any quadratic terms is due to
the fact that all $p_{\pm}^{(i)2} = 0$, i.e. to the choice of a
formulation based on light-cone-like longitudinal vectors.
Of the two possible solutions to the equation system (elimination of
one variable gives a second degree equation in the other), one is
unphysical and can be discarded outright. The other solution is checked
for
whether the $x_{\pm}$ values are actually inside the physically allowed
region, i.e. whether the $x_{\pm}$ values of the current step, plus
whatever has already been used up in previous steps, are less than
unity. If yes, a solution has been found. If no, it is because the
breakup could not take place inside the region studied, i.e. because
the equation system was solved for the wrong region. One therefore
has to change either index $j$ or index $k$ above by one step, i.e.
go to the next nearest string region. In this new region, a new equation
system of the type in eq.~(\ref{fr:mGa}) may be written down, with new
coefficients. A new solution is found and tested, and so on until a
physically
acceptable solution is found. The hadron four-momentum is now
given by an expression of the type (\ref{fr:fourmom}). The breakup
found forms the starting point for the new step in the fragmentation
chain, and so on. The final joining in the middle is done as in the
2-jet case, with minor extensions.
\subsection{Independent Fragmentation}
The independent fragmentation (IF) approach dates back to the early
seventies \cite{Krz72}, and gained widespread popularity with
the Field-Feynman paper \cite{Fie78}. Subsequently, IF was the basis
for two programs widely used in the early PETRA/PEP days, the
Hoyer et al.~\cite{Hoy79} and the Ali et al.~\cite{Ali80} programs.
JETSET has as (non-default) options a wide selection of independent
fragmentation algorithms.
\subsubsection{Fragmentation of a single jet}
In the IF approach, it is assumed that the fragmentation of any
system of partons can be described as an incoherent sum of
independent fragmentation procedures for each parton separately.
The process is to be carried out in the overall c.m. frame of the
jet system, with each jet fragmentation axis given by the direction
of motion of the corresponding parton in that frame.
Exactly as in string fragmentation, an iterative ansatz can be used
to describe the sucessive production of one hadron after the next.
Assume that a quark is kicked out by some hard interaction, carrying a
well-defined amount of energy and momentum. This quark jet $\mathrm{q}$ is
split into a hadron $\mathrm{q} \overline{\mathrm{q}}_1$ and a remainder-jet $\mathrm{q}_1$, essentially
collinear with each other. New quark and hadron flavours are picked
as already described. The sharing of energy and momentum is given
by some probability distribution $f(z)$, where $z$ is the fraction
taken by the hadron, leaving $1-z$ for the remainder-jet. The
remainder-jet is assumed to be just a scaled-down version of the
original jet, in an average sense. The process of splitting off a
hadron can therefore be iterated, to yield a sequence of hadrons.
In particular, the function $f(z)$ is assumed to be the
same at each step, i.e. independent of remaining energy. If $z$ is
interpreted as the fraction of the jet $E+p_{\mrm{L}}$, i.e. energy plus
longitudinal momentum with respect to the jet axis, this leads to a
flat central rapidity plateau $dn/dy$ for a large initial energy.
Fragmentation functions can be chosen among those listed above for
string fragmentation, but also here the default is the Lund symmetric
fragmentation function.
The normal $z$ interpretation means that
a choice of a $z$ value close to $0$ corresponds to a particle
moving backwards, i.e. with $p_{\mrm{L}} < 0$.
It makes sense to allow only the production of particles with
$p_{\mrm{L}} > 0$, but to explicitly constrain $z$ accordingly
would destroy longitudinal invariance. The most
straightforward way out is to allow all $z$ values but discard
hadrons with $p_{\mrm{L}} < 0$. Flavour, transverse momentum and
$E + p_{\mrm{L}}$
carried by these hadrons are `lost' for the forward jet. The average
energy of the final jet comes out roughly right this way, with a spread
of 1--2 GeV around the mean. The jet longitudinal
momentum is decreased, however, since the jet acquires an effective
mass during the fragmentation procedure. For a 2-jet event this is
as it should be, at least on average, because also the momentum of the
compensating opposite-side parton is decreased.
In addition to local flavour conservation in $\mathrm{q}_i \overline{\mathrm{q}}_i$ splittings,
it is also assumed that transverse momentum is locally conserved, i.e.
the net $p_{\perp}$ of the $\mathrm{q}_i \overline{\mathrm{q}}_i$ pair as a whole is assumed to be
vanishing. The $p_{\perp}$ of the $\mathrm{q}$ is taken to be a Gaussian in the two
transverse degrees of freedom separately, with the transverse momentum
of a hadron obtained by the sum of constituent quark transverse momenta.
Within the IF framework, there is no unique recipe for how gluon jet
fragmentation should be handled. One possibility is to treat it exactly
like a quark jet, with the initial quark flavour chosen
at random among $\u$, $\overline{\mathrm{u}}$, $\d$, $\overline{\mathrm{d}}$, $\mathrm{s}$ and
$\overline{\mathrm{s}}$, including the ordinary $\mathrm{s}$ quark suppression factor.
Since the gluon is supposed to fragment more softly
than a quark jet, the fragmentation fuction may be chosen
independently. Another common option is to split the $\mathrm{g}$ jet into
a pair of parallel $\mathrm{q}$ and $\overline{\mathrm{q}}$ ones, sharing the energy,
e.g. as in a perturbative branching $\mathrm{g} \to \mathrm{q} \overline{\mathrm{q}}$, i.e.
$f(z) \propto z^2 + (1-z)^2$.
The fragmentation function could still be chosen
independently, if so desired. Further, in either case the fragmentation
$p_{\perp}$ could be chosen to have a different mean.
\subsubsection{Fragmentation of a jet system}
In a system of many jets, each jet is fragmented independently. Since
each jet by itself does not conserves the flavour, energy and momentum,
then neither does a system of jets.
At the end of the generation, special algorithms are
therefore used to patch this up. The choice of approach has major
consequences, e.g. for event shapes and $\alpha_{\mathrm{s}}$ determinations
\cite{Sjo84a}.
Little attention is usually given to flavour conservation, and we only
offer one scheme. When the fragmentation of all jets has been performed,
independently of each other, the net initial flavour composition, i.e.
number of $\u$ quarks minus number of $\overline{\mathrm{u}}$ quarks etc., is compared
with the net final flavour composition. In case of an imbalance,
the flavours of the hadron with lowest three-momentum are removed, and
the imbalance is re-evaluated. If the remaining imbalance could be
compensated by a suitable choice of new flavours for this hadron,
flavours are so chosen, a new mass is found and the new energy can be
evaluated, keeping the three-momentum of the original hadron. If
the removal of flavours from the hadron with lowest momentum is not
enough, flavours are removed from the one with next-lowest momentum,
and so on until enough freedom is obtained, whereafter the necessary
flavours are recombined at random to form the new hadrons. Occasionally
one extra $\mathrm{q}_i \overline{\mathrm{q}}_i$ pair must be created, which is then done
according to the customary probabilities.
Several different schemes for energy and momentum conservation have
been devised. One \cite{Hoy79} is to conserve transverse momentum
locally within each jet, so that the final momentum vector of a jet
is always parallel with that of the corresponding parton. Then
longitudinal momenta
may be rescaled separately for particles within each jet, such that
the ratio of rescaled jet momentum to initial parton momentum is the
same in all jets. Since the initial partons had net vanishing
three-momentum, so do now the hadrons. The rescaling factors may
be chosen such that also energy comes out right. Another common
approach \cite{Ali80} is to boost the event to the frame
where the total hadronic momentum is vanishing. After that, energy
conservation can be obtained by rescaling all particle three-momenta
by a common factor.
The number of possible schemes is infinite.
Two further options are available in the program. One is to
shift all particle three-momenta by a common amount to give
net vanishing momentum, and then rescale as before. Another is
to shift all particle three-momenta, for each particle by an
amount proportional to the longitudinal mass with respect to the
imbalance direction, and with overall magnitude selected to give
momentum conservation, and then rescale as before.
In addition, there is a choice of whether to treat separate
colour singlets (like $\mathrm{q} \overline{\mathrm{q}}'$ and $\mathrm{q}' \overline{\mathrm{q}}$ in a
$\mathrm{q} \overline{\mathrm{q}} \mathrm{q}' \overline{\mathrm{q}}'$ event) separately or as one single big system.
A serious conceptual weakness of the IF framework is the issue of
Lorentz invariance. The outcome of the fragmentation procedure
depends on the coordinate frame chosen, a problem circumvented by
requiring fragmentation always to be carried out in the c.m. frame.
This is a consistent procedure for 2-jet events, but only a
technical trick for multijets.
It should be noted, however, that a Lorentz covariant generalization
of the independent fragmentation model exists, in which separate
`gluon-type' and `quark-type' strings are used,
the Montvay scheme \cite{Mon79}.
The `quark string' is characterized by the ordinary string constant
$\kappa$, whereas a `gluon string' is taken to have a string constant
$\kappa_{\mathrm{g}}$. If $\kappa_{\mathrm{g}} > 2 \kappa$ it is always energetically
favourable to split a gluon string into two quark ones, and the
ordinary Lund string model is recovered. Otherwise, for a 3-jet
$\mathrm{q} \overline{\mathrm{q}} \mathrm{g}$ event the three different string pieces are joined at
a junction. The motion of this junction is given by the composant of
string tensions acting on it. In particular, it is always possible
to boost an event to a frame where this junction is at rest. In this
frame, much of the standard na\"{\i}ve IF picture holds for the
fragmentation of the three jets; additionally, a correct treatment
would automatically give flavour, momentum and energy conservation.
Unfortunately, the simplicity is lost when studying events with
several gluon jets. In general, each event will contain a number of
different junctions, resulting in a polypod shape with a number of
quark and gluons strings sticking out from a skeleton of gluon
strings. With the shift of emphasis from three-parton to
multi-parton configurations, the simple option existing in {\tsc{Jetset}}~6.3
therefore is no longer included.
A second conceptual weakness of IF is the issue of collinear
divergences. In a parton-shower picture, where a quark or gluon is
expected to branch into several reasonably collimated partons,
the independent fragmentation of one single parton or of a bunch of
collinear ones gives quite different outcomes, e.g. with a much
larger hadron multiplicity in the latter case. It is conceivable that
a different set of fragmentation functions could be constructed in the
shower case in order to circumvent this problem
(local parton--hadron duality \cite{Dok89} would correspond to having
$f(z) = \delta(z-1)$).
\subsection{Other Fragmentation Aspects}
Here two aspects are considered, which are applicable
regardless of whether string or
independent fragmentation is used.
\subsubsection{Small mass systems}
Occasionally, a jet system may have too small an invariant mass for the
ordinary jet fragmentation schemes. This is particularly a problem when
showers are used, since two nearby $\mathrm{g} \to \mathrm{q}' \overline{\mathrm{q}}'$ branchings
may give rise to an intermediate low-mass colour-singlet system. Before
the ordinary fragmentation, one includes an optional additional step,
to catch situations of this kind.
First the jet system with lowest invariant mass, minus endpoint quark
masses, is found. If this is too low for jet fragmentation, an attempt
is made to split the system into two hadrons by producing a new
$\mathrm{q}_n \overline{\mathrm{q}}_n$ pair (with $\mathrm{q}_n$ chosen according to the standard
fragmentation scheme, so that e.g. also diquarks are allowed)
to go with the existing endpoint flavours. If the sum of the two
thus constructed
hadron masses is smaller than the total invariant mass, a simple
isotropic two-particle decay is performed. If not, the endpoint flavours
are combined to give one single hadron. Next, the parton (or hadron) is
found which, when taken together with the jet system, has the largest
invariant mass. A minimal transfer of four-momentum is then performed,
which puts the hadron on mass shell while keeping the mass of the
parton unchanged. With this done, one may again search for a low-mass
jet system, and iterate the procedure above, if need be.
The procedure may be seen as a `poor man's cluster fragmentation',
i.e. a cluster and a low-mass string are considered to be more or
less the same thing.
\subsubsection{Bose--Einstein effects}
A crude option for the simulation of Bose--Einstein effects is
included, but is turned off by default. Here the detailed physics is
not that well understood, see e.g. \cite{Lor89}. What is
offered is an algorithm, more than just a parametrization (since very
specific assumptions and choices have been made), and yet less than a
true model (since the underlying physics picture is rather fuzzy).
In this scheme, the
fragmentation is allowed to proceed as usual, and so is the decay of
short-lived particles like $\rho$. Then pairs of identical particles,
$\pi^+$ say, are considered one by one. The $Q_{ij}$ value of a pair
$i$ and $j$ is evaluated,
\begin{equation}
Q_{ij} = \sqrt{ (p_i + p_j)^2 - 4m^2} ~,
\end{equation}
where $m$ is the common particle mass. A shifted (smaller) $Q'_{ij}$
is then to be found such that the (infinite statistics) ratio
$C_2(Q)$ of shifted to unshifted $Q$ distributions is given by the
requested parametrization. The shape may be chosen either
exponential or Gaussian,
\begin{equation}
C_2(Q) = 1 + \lambda \exp \left( - (Q/d)^r \right),
~~~~r = 1~\mrm{or}~2 ~.
\end{equation}
(In fact, the distribution has to dip slightly below unity at $Q$
values outside the Bose enhancement region, from conservation of
total multiplicity.) If the inclusive distribution of $Q_{ij}$
values is assumed given
just by phase space, at least at small relative momentum then,
with $\d^3 p / E \propto Q^2 \, \d Q / \sqrt{Q^2 + 4m^2}$,
then $Q'_{ij}$ is found as the solution to the equation
\begin{equation}
\int_0^{Q_{ij}} \frac{Q^2 \, \d Q}{\sqrt{Q^2 + 4 m^2}} =
\int_0^{Q'_{ij}} C_2(Q) \, \frac{Q^2 \, \d Q}{\sqrt{Q^2 + 4 m^2}} ~.
\end{equation}
The change of $Q_{ij}$ can be translated into an effective shift of
the three-momenta of the two particles, if one uses
as extra constraint that the total three-momentum of each pair be
conserved in the c.m. frame of the event. Only after all pairwise
momentum shifts have been evaluated, with respect to the original
momenta, are these momenta actually shifted, for each particle by the
sum of evaluated shifts. The total energy of the event is slightly
reduced in the process, which is compensated by an overall rescaling of
all c.m. frame momentum vectors. It can be discussed which are the
particles to involve in this rescaling. Currently the only
exceptions to using everything are leptons and neutrinos coming from
resonance decays (such as $\mathrm{W}$'s) and photons radiated by leptons
(also in initial state radiation). Finally, the decay chain is resumed
with more long-lived particles like $\pi^0$.
Two comments can be made.
The Bose--Einstein effect is here interpreted almost as a
classical force acting on the `final state', rather than
as a quantum mechanical phenomenon on the production amplitude. This
is not a credo, but just an ansatz to make things manageable.
Also, since only pairwise interactions
are considered, the effects associated with three or more nearby
particles tend to get overestimated. (More exact, but also more
time-consuming methods may be found in
\cite{Zaj87}.) Thus the input $\lambda$
may have to be chosen smaller than what one wants to get out.
(On the other hand, many of the pairs of an event contains at least
one particle produced in some secondary vertex, like a $\mathrm{D}$ decay.
This reduces the fraction of pairs which may contribute to the
Bose--Einstein effects, and thus reduces the potential signal.)
This option should therefore be used with caution, and only as a
first approximation to what Bose--Einstein effects can mean.
\clearpage
\section{Particles and Their Decays}
Particles are the building blocks from which events are constructed.
We here use the word `particle' in its broadest sense, i.e. including
partons, resonances, hadrons, and so on, subgroups we will describe
in the following. Each particle is characterized by some quantities,
such as charge and mass. In addition, many of the particles are
unstable and subsequently decay. This section contains a survey of
the particle content of the programs, and the particle properties
assumed. In particular, the decay treatment is discussed. Some
particle and decay properties form part already of the hard
subprocess description, and are therefore described in sections
\ref{s:JETSETproc}, \ref{s:PYTprocgen} and \ref{s:pytproc}.
\subsection{The Particle Content}
\label{ss:decpartcont}
In order to describe both current and potential future physics,
a number of different particles are needed. A list of some
particles, along with their codes, is given in section \ref{ss:codes}.
Here we therefore emphasize the generality rather than the details.
Four full generations of quarks and leptons are included in the
program, although indications from LEP strongly suggest that
only three exist in Nature. There is no standard terminology for the
fourth generation; we use $\mathrm{l}$ for the down type quark
($\mathrm{l}$ for low), $\mathrm{h}$ for the up type quark
($\mathrm{h}$ for high), $\chi$ for the lepton
and $\nu_{\chi}$ for the neutrino. Quarks may appear either singly
or in pairs; the latter are called diquarks and are characterized
by their flavour content and their spin. A diquark is always assumed
to be in a colour antitriplet state.
From the coloured quarks (and diquarks), the colour neutral hadrons
may be build up. Six full meson multiplets are included and two
baryon ones, see section \ref{ss:flavoursel}. In addition,
$\mathrm{K}_{\mrm{S}}^0$ and $\mathrm{K}_{\mrm{L}}^0$ are considered as separate
particles coming from the `decay' of $\mathrm{K}^0$ and $\br{\mathrm{K}}^0$ (or,
occasionally, produced directly).
Other particles from the Standard Model include the gluon $\mathrm{g}$, the
photon $\gamma$, the intermediate gauge bosons $\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$,
and the standard Higgs $\H^0$. Non-standard particles include
additional gauge bosons, $\mathrm{Z}'^0$ and $\mathrm{W}'^{\pm}$, additional
Higgs bosons $\H'^0$, $\mathrm{A}^0$ and $\H^{\pm}$, a leptoquark $\L_{\mathrm{Q}}$
and a horizontal gauge boson $\mathrm{R}^0$. It is also possible
to use the particle codes of the current fourth generation
fermions to represent excited quarks and leptons, $\mathrm{q}^*$ and
$\ell^*$.
From the point of view of usage inside the programs, particles
may be subdivided into three classes, partly overlapping.
\begin{Enumerate}
\item A parton is generically any object which may be found in the
wave function of the incoming beams, and may participate in initial-
or final-state showers. This includes what is normally meant by
partons, i.e. quarks and gluons, but here also leptons
and photons. In a few cases other particles may be
classified as partons in this sense.
\item A resonance is an unstable particle produced as part of the
hard process, and where the decay treatment normally is also part
of the hard process. Resonance partial widths are perturbatively
calculable, and therefore it is possible to dynamically recalculate
branching ratios as a function of the mass assigned to a resonance.
Resonances includes particles like the $\mathrm{Z}^0$ and other massive
gauge bosons and Higgs particles. It does not include hadrons
with non-vanishing width, like the $\rho$, which are just called
`unstable hadrons'.
\item Hadrons, i.e. mesons and baryons produced either in the
fragmentation process, in secondary decays or as part of the beam
remnant treatment, but not directly as part of the hard process
(except in a few special cases). Hadrons may be stable or unstable.
Branching ratios are not assumed perturbatively calculable, and can
therefore be set freely. Also leptons and photons produced in decays
belong to this class.
\end{Enumerate}
Usually the subdivision above is easy to understand and gives you
the control you would expect. However, the classification of top
and the fourth generation fermions may lead to some confusion, as
already mentioned, section \ref{sss:heavflavclass}. The problem is
that the top did not use to be
treated as a resonance, but was rather allowed to fragment
to hadrons, which subsequently decayed. This approach was a reasonable
choice in the days when the top mass was assumed quite light by
today's standards. However, given current top limits, the fragmentation
and the decay of the top quark is being played out on comparable time
scales, and the treatment becomes much more difficult (see e.g. ref.
\cite{Sjo92a} for a toy model description). Starting at masses of
around 120 GeV the top decay time is so short that no top hadrons at
all are formed, and then a true resonance description is
appropriate, but still with some complications due to the net colour
charge of the top quark. Such an option is now default, wherein the top
quark is assumed to decay immediately, but intermediate scenarios can
not be modelled. The appearance of intermediate top hadrons
in the description has little influence on event shapes, even for a
very heavy top.
\subsection{Masses, Widths and Lifetimes}
\subsubsection{Masses}
Quark masses are not particularly well defined. In the program it is
necessary to make use of two kinds of masses, current algebra ones
and constituent ones. The former are relevant for the kinematics in
hard processes (e.g. in $\mathrm{g}\g \to \c\overline{\mathrm{c}}$) and for couplings to
Higgs particles, and therefore directly affect cross sections.
These values are the ones stored in the standard mass array
\ttt{PMAS}. Constituent masses are used to derive the masses of hadrons,
and are stored separately in the \ttt{PARF} array. We maintain
this distinction for the five first flavours, using the following
values by default: \\
\begin{tabular}{ccc@{\protect\rule{0mm}{\tablinsep}}}
quark & current algebra mass & constituent mass \\
d & 0.0099 GeV & 0.325 GeV \\
u & 0.0056 GeV & 0.325 GeV \\
s & 0.199 GeV & 0.5 GeV \\
c & 1.35 GeV & 1.6 GeV \\
b & 5.0 GeV & 5.0 GeV. \\
\end{tabular} \\
For top and fourth generation quarks the distinction is not as
important, so only one set of mass values is used, namely the one in
\ttt{PMAS}. The default top mass is 160 GeV. Constituent masses for
diquarks are defined as the sum of the respective quark masses. The
gluon is always assumed massless.
Particle masses, when known, are taken from ref. \cite{PDG92}.
Hypothesized particles, such as fourth generation fermions and
Higgs bosons, are assigned some not unreasonable set of default
values, in the sense of where you want to search for them in the
not too distant future. Here it is understood that you will go in
and change the default values according to your own opinions at
the beginning of a run.
The total number of hadrons in the program is very large, whereof
many are not yet discovered (like charm and bottom baryons) and
other may or may not exist (top and fourth generation hadrons).
In particular for the latter, it would be messy for the user to have
to recalculate the masses of hadron each time the assumed quark mass
is changed. Therefore the masses of yet undiscovered mesons and
baryons are built up, when needed, from the constituent masses.
For this purpose one uses formulae of the type
\cite{DeR75}
\begin{equation}
m = m_0 + \sum_i m_i + k \, m_{\d}^2 \sum_{i<j} \frac{\langle
\mbox{\boldmath $\sigma$}_i \cdot
\mbox{\boldmath $\sigma$}_j \rangle}{m_i \, m_j} ~,
\end{equation}
i.e. one constant term, a sum over constituent masses
and a spin-spin interaction term for each quark pair in the hadron.
The constants $m_0$ and $k$ are fitted from known masses,
treating mesons and baryons separately. For mesons with orbital
angular momentum $L=1$ the spin-spin coupling is assumed vanishing,
and only $m_0$ is fitted.
One may also define `constituent diquarks masses' using the formula
above, with a $k$ value $2/3$ that of baryons. The default values
are: \\
\begin{tabular}{ccc@{\protect\rule{0mm}{\tablinsep}}}
multiplet & $m_0$ & $k$ \\
pseudoscalars and vectors & 0. & 0.16 GeV \\
axial vectors ($S=0$) & 0.50 GeV & 0. \\
scalars & 0.45 GeV & 0. \\
axial vectors ($S=1$) & 0.55 GeV & 0. \\
tensors & 0.60 GeV & 0. \\
baryons & 0.11 GeV & 0.048 GeV \\
diquarks & 0.077 GeV & 0.048 GeV.\\
\end{tabular} \\
There is one exception to the rule above, and that is flavour neutral
mesons, i.e. the onia states of a heavy quark--antiquark pair. These
are defined individually, to allow more flexibility.
\subsubsection{Widths}
A width is calculated perturbatively for those resonances which
appear in the {\tsc{Pythia}} hard process generation machinery. The width is
used to select masses in hard processes according to a relativistic
Breit--Wigner shape. In many processes the width is allowed to be
$\hat{s}$-dependent, see section \ref{ss:kinemreson}.
Other particle masses, as discussed so far, have been fixed at their
nominal value, i.e. with no mass broadening for short-lived particles
such as $\rho$, $\mathrm{K}^*$ or $\Delta$. Compared to the $\mathrm{Z}^0$, it is
much more difficult to describe the $\rho$ resonance shape, since
nonperturbative and threshold effects act to distort the na\"{\i}ve
shape. Thus the $\rho$ mass is limited from below by its decay
$\rho \to \pi\pi$, but also from above, e.g. in the decay
$\phi \to \rho \pi$. In some decay chains, several mass choices are
coupled, like in $\a_2 \to \rho \pi$, where also the $\a_2$ has a
non-negligible width. Finally, there are some extreme cases, like
the $\mathrm{f}_0$, which has a nominal mass below the $\mathrm{K}\K$ threshold, but
a tail extending beyond that threshold, and therefore a
non-negligible branching ratio to the $\mathrm{K}\K$ channel.
In view of examples like these, no attempt is made to provide a
full description. Instead a simplified description is used, which
should be enough to give the general smearing of events due to
mass broadening, but maybe not sufficient for detailed studies of
a specific resonance. By default, hadrons are therefore given a
mass distribution according to a non-relativistic Breit--Wigner
\begin{equation}
{\cal P}(m) \, \d m \propto \frac{1}{(m - m_0)^2 + \Gamma^2/4}
\, \d m ~.
\label{dec:BWlin}
\end{equation}
Leptons and resonances not taken care of by the hard
process machinery are distributed according to a relativistic
Breit--Wigner
\begin{equation}
{\cal P}(m^2) \, \d m^2 \propto \frac{1}{(m^2 - m_0^2)^2 +
m_0^2 \Gamma^2} \, \d m^2 ~.
\label{dec:BWtwo}
\end{equation}
Here $m_0$ and $\Gamma$ are the nominal mass and width of the
particle. The Breit--Wigner shape is truncated symmetrically,
$|m - m_0| < \delta$, with $\delta$ arbitrarily chosen for each
particle so that no problems are encountered in the decay chains.
It is possible to switch off the mass broadening, or to use either
a non-relativistic or a relativistic Breit--Wigners everywhere.
The $\mathrm{f}_0$ problem has been `solved' by shifting the $\mathrm{f}_0$ mass to
be slightly above the $\mathrm{K}\K$ threshold and have vanishing width.
Then kinematics in decays $\mathrm{f}_0 \to \mathrm{K}\K$ is reasonably well
modelled. The $\mathrm{f}_0$ mass is too large in the
$\mathrm{f}_0 \to \pi\pi$ channel, but this does not really matter, since
one anyway is far above threshold here.
\subsubsection{Lifetimes}
Clearly the lifetime and the width of a particle are inversely
related. For practical applications, however, any particle with
a non-negligible width decays too close to its production vertex
for the lifetime to be of any interest. In the program, the two
aspects are therefore considered separately. Particles with a
non-vanishing nominal proper lifetime
$\tau_0 = \langle \tau \rangle$ are
assigned an actual lifetime according to
\begin{equation}
{\cal P}(\tau) \, \d \tau \propto \exp(- \tau / \tau_0 ) \, \d \tau ~,
\end{equation}
i.e. a simple exponential decay is assumed. Since the program
uses dimensions where the speed of light $c \equiv 1$, and
space dimensions are in mm, then actually the unit of $c \tau_0$ is
mm and of $\tau_0$ itself mm$/c \approx 3.33\times10^{-12}$ s.
If a particle is produced at a vertex $v = (\mbf{x}, t)$ with a
momentum $p = (\mbf{p}, E)$ and a lifetime $\tau$, the
decay vertex position is assumed to be
\begin{equation}
v' = v + \tau \, \frac{p}{m} ~,
\label{dec:newvertex}
\end{equation}
where $m$ is the mass of the particle. With the primary
interaction (normally) in the origin, it is therefore possible to
construct all secondary vertices in parallel with the ordinary
decay treatment.
The formula above does not take into account any detector effects,
such as a magnetic field. It is therefore possible to stop the
decay chains at some suitable point, and leave any subsequent
decay treatment to the detector simulation program. One may
select that particles are only allowed to decay if they have a
nominal lifetime $\tau_0$ shorter than some given value or,
alternatively, if their decay vertices $\mbf{x}'$ are inside some
spherical or cylindrical volume around the origin.
\subsection{Decays}
\label{ss:partdecays}
Several different kinds of decay treatment are used in the program,
depending on the nature of the decay. Not discussed here are the
decays of resonances which are handled as part of the hard process.
\subsubsection{Strong and electromagnetic decays}
The decays of hadrons containing the `ordinary' $\u$, $\d$ and $\mathrm{s}$
quarks into two or three particles are known, and branching ratios
may be found in \cite{PDG92}. We normally assume that the momentum
distributions are given by phase space. There are a few exceptions,
where the phase space is weighted by a matrix-element expression,
as follows.
In $\omega$ and $\phi$ decays to $\pi^+ \pi^- \pi^0$, a matrix element
of the form
\begin{equation}
|{\cal M}|^2 \propto | \mbf{p}_{\pi^+} \times \mbf{p}_{\pi^-} |^2
\label{dec:omegphi}
\end{equation}
is used, with the $\mbf{p}_{\pi}$ the pion momenta in the rest frame
of the decay. (Actually, what is coded is the somewhat more lengthy
Lorentz invariant form of the expression above.)
Consider the decay chain $P_0 \to P_1 + V \to P_1 + P_2 + P_3$,
with $P$ representing pseudoscalar mesons and $V$ a vector one. Here
the decay angular distribution of the $V$ in its rest frame is
\begin{equation}
|{\cal M}|^2 \propto \cos^2 \theta_{02} ~,
\label{dec:psvpsps}
\end{equation}
where $\theta_{02}$ is the angle between $P_0$ and $P_2$.
The classical example is $\mathrm{D} \to \mathrm{K}^* \pi \to \mathrm{K} \pi \pi$.
If the $P_1$ is replaced by a $\gamma$, the angular distribution in
the $V$ decay is instead $\propto \sin^2 \theta_{02}$.
In Dalitz decays, $\pi^0$ or $\eta \to \mathrm{e}^+\mathrm{e}^- \gamma$, the mass $m^*$
of the $\e^+\e^-$ pair is selected according to
\begin{equation}
{\cal P}(m^{*2}) \, \d m^{*2} \propto \frac{\d m^{*2}}{m^{*2}} \,
\left( 1 + \frac{2m_{\mathrm{e}}^2}{m^{*2}} \right) \,
\sqrt{ 1 - \frac{4m_{\mathrm{e}}^2}{m^{*2}} } \,
\left( 1 - \frac{m^{*2}}{m_{\pi,\eta}^2} \right)^3 \,
\frac{1}{ (m_{\rho}^2 - m^{*2})^2 + m_{\rho}^2 \Gamma_{\rho}^2 } ~.
\label{dec:Dalitz}
\end{equation}
The last factor, the VMD-inspired $\rho^0$ propagator, is negligible
for $\pi^0$ decay. Once the $m^*$ has been selected, the angular
distribution of the $\e^+\e^-$ pair is given by
\begin{equation}
|{\cal M}|^2 \propto (m^{*2} - 2 m_{\mathrm{e}}^2) \left\{
(p_{\gamma} p_{\mathrm{e}^+})^2 + (p_{\gamma} p_{\mathrm{e}^-})^2 \right\} +
4m_{\mathrm{e}}^2 \left\{ (p_{\gamma} p_{\mathrm{e}^+}) (p_{\gamma} p_{\mathrm{e}^-}) +
(p_{\gamma} p_{\mathrm{e}^+})^2 + (p_{\gamma} p_{\mathrm{e}^-})^2 \right\} ~.
\end{equation}
Also a number of simple decays involving resonances of heavier
hadrons, e.g. $\Sigma_{\c}^0 \to \Lambda_{\c}^+ \pi^-$ or
$\mathrm{B}^{*-} \to \mathrm{B}^- \gamma$ are treated in the same way as the
other two-particle decays.
\subsubsection{Weak decays of charm hadrons}
The charm hadrons have a mass in an intermediate range, where the
effects of the na\"{\i}ve $V-A$ weak decay matrix element is partly but
not fully reflected in the kinematics of final-state particles.
Therefore different decay strategies ar combined. We start with
hadronic decays, and subseqently consider semileptonic ones.
For the four `main' charm hadrons, $\mathrm{D}^+$, $\mathrm{D}^0$, $\mathrm{D}_{\mathrm{s}}^+$ and
$\Lambda_{\c}^+$, a number of branching ratios are already known.
The known braching ratios have been combined with reasonable
guesses, to construct more or less complete tables of all channels.
For hadronic decays of $\mathrm{D}^0$ and $\mathrm{D}^+$, where rather much is known,
all channels have an explicitly listed particle content.
However, only for the two-body decays is resonance production
properly taken into account. It means that the experimentally measured
branching ratio for a $\mathrm{K} \pi \pi$ decay channel, say, is represented
by contributions from a direct $\mathrm{K} \pi \pi$ channel as well as from
indirect ones, such as $\mathrm{K}^* \pi$ and $\mathrm{K} \rho$. For a channel like
$\mathrm{K} \pi \pi \pi$, on the other hand, only the $\mathrm{K}^* \rho$ appears
separately, while the rest is lumped into one entry in the decay
tables. This is more or less in agreement with the philosophy adopted
in the PDG tables \cite{PDG92}. For $\mathrm{D}_{\mathrm{s}}^+$ and $\Lambda_{\c}^+$
knowledge is rather incomplete, and only two-body decay channels are
listed. Final states with three or more hadron are only listed in
terms of a flavour content.
The way the program works, it is important to include all the
allowed decay channels up to a given multiplicity. Channels with
multiplicity higher than this may then be generated according to
a simple flavour combination scheme. For instance, in a $\mathrm{D}_{\mathrm{s}}^+$
decay, the normal quark content is $\mathrm{s}\overline{\mathrm{s}}\u\overline{\mathrm{d}}$, where one
$\overline{\mathrm{s}}$ is the spectator quark and the others come from the weak
decay of the $\c$ quark. The spectator quark may also be annihilated,
like in $\mathrm{D}_{\mathrm{s}}^+ \to \u\overline{\mathrm{d}}$. The flavour content to make up
one or two hadrons is therefore present from the onset.
If one decides to generate more hadrons,
this means new flavour-antiflavour pairs have to be generated
and combined with the existing flavours. This is done using the
same flavour approach as in fragmentation.
In more detail, the following scheme is used.
\begin{Enumerate}
\item The multiplicity is first selected. The $\mathrm{D}_{\mathrm{s}}^+$ and
$\Lambda_{\c}^+$ multiplicity is selected according to a distribution
described further below. The program can also be asked to
generate events of a predetermined multiplicity.
\item One of the non-spectator flavours is selected at random.
This flavour is allowed to `fragment' into a hadron plus a new
remaining flavour, using exactly the same flavour generation
algorithm as in the standard jet fragmentation, section
\ref{ss:flavoursel}.
\item Step 2 is iterated until only one or two hadrons remain to
be generated, depending on whether the original number of flavours
is two or four. In each step one `unpaired' flavour is replaced by
another one as a hadron is `peeled off', so the number of unpaired
flavours is preserved.
\item If there are two flavours, these are combined to form the last
hadron. If there are four, then one of the two possible pairings
into two final hadrons is selected at random. To find the hadron
species, the same flavour rules are used as when final flavours are
combined in the joining of two jets.
\item If the sum of decay product masses is larger than the mass of
the decaying particle, the flavour selection is rejected and the
process is started over at step 1. Normally a new multiplicity is
picked, but for $\mathrm{D}^0$ and $\mathrm{D}^+$ the old multiplicity is retained.
\item Once an acceptable set of hadrons has been found, these are
distributed according to phase space.
\end{Enumerate}
The picture then is one of a number of partons moving apart,
fragmenting almost like jets, but with momenta so low that phase-space
considerations are enough to give the average behaviour of
the momentum distribution. Like in jet fragmentation, endpoint
flavours are not likely to recombine with each other. Instead
new flavour pairs are created in between them. One should also note
that, while vector and pseudoscalar mesons are produced at their
ordinary relative rates, events with many vectors are likely to
fail in step 5. Effectively, there is therefore a shift towards
lighter particles, especially at large multiplicities.
When a multiplicity is to be picked, this is done according to a
Gaussian distribution, centered at $c + n_{\mathrm{q}}/4$ and with a
width $\sqrt{c}$, with the final number rounded off to the nearest
integer. The value for the number of quarks $n_{\mathrm{q}}$ is 2 or 4,
as described above, and
\begin{equation}
c = c_1 \, \ln \left( \frac{m - \sum m_{\mathrm{q}}}{c_2} \right) ~,
\label{dec:multsel}
\end{equation}
where $m$ is the hadron mass and $c_1$ and $c_2$ have been tuned
to give a reasonable description of multiplicities. There is always
some lower limit for the allowed multiplicity; if a number
smaller than this is picked the choice is repeated. Since two-body
decays are explicitly enumerated for $\mathrm{D}_{\mathrm{s}}^+$ and
$\Lambda_{\c}^+$, there the mimimum multiplicity is three.
Semileptonic branching ratios are explicitly given in the program
for all the four particles discussed here, i.e. it is never
necessary to generate the flavour content using the fragmentation
description. This does not mean that all branching ratios are known;
a fair amount of guesswork is involved for
the channels with higher multiplicities, based
on a knowledge of the inclusive semileptonic branching ratio and
the exclusive branching ratios for low multiplicities.
In semileptonic decays it is not appropriate to distribute the
lepton and neutrino momenta according to phase space. Instead the
simple $V-A$
matrix element is used, in the limit that decay product masses may
be neglected and that quark momenta can be replaced by hadron
momenta. Specifically, in the decay $H \to \ell^+ \nu_{\ell} h$,
where $H$ is a charm hadron and $h$ and ordinary hadron, the matrix
element
\begin{equation}
|{\cal M}|^2 = (p_H p_{\ell}) (p_{\nu} p_h)
\end{equation}
is used to distribute the products. It is not clear how to
generalize this formula when several hadrons are present in the final
state. In the program, the same matrix element is used as above,
with $p_h$ replaced by the total four-momentum of all the hadrons.
This tends to favour a low invariant mass for the hadronic system
compared with na\"{\i}ve phase space.
There are a few charm hadrons, such as $\Xi_c$ and $\Omega_c$, which
decay weakly but are so rare that little is known about them. For
these a simplified generic charm decay treatment is used. For
hadronic decays only the quark content is given, and then a
multiplicity and a flavour composition is picked at random, as
already described. Semileptonic decays are assumed to produce only
one hadron, so that $V-A$ matrix element can be simply applied.
\subsubsection{Weak decays of the $\tau$ lepton}
For the $\tau$ lepton, an explicit list of decay channels has been
put together, which includes channels with up to five final-state
particles, some of which may be unstable and subsequently decay to
produce even larger total multiplicities. Because of the well-known
`$\tau$ puzzle', i.e. that experimentally the sum of branching
ratios for exclusive one-prong decays is lower than the inclusive
one-prong branching ratio, such a table cannot be constructed in
full agreement with the PDG data. (The problem is nowadays less severe
than it used to be, but still not fully resolved.)
The leptonic decays $\tau^- \to \nu_{\tau} \ell^- \br{\nu}_{\ell}$,
where $\ell$ is $\mathrm{e}$ or $\mu$, are distributed according to the
standard $V-A$ matrix element
\begin{equation}
|{\cal M}|^2 = (p_{\tau} p_{\br{\nu}_{\ell}})
(p_{\ell} p_{\nu_{\tau}}) ~.
\end{equation}
(The corresponding matrix element is also used in $\mu$ decays, but
normally the $\mu$ is assumed stable.)
In $\tau$ decays to hadrons, the hadrons and the $\nu_{\tau}$ are
distributed according to phase space times the factor
$x_{\nu} \, (3 - x_{\nu})$, where $x_{\nu} = 2E_{\nu}/m_{\tau}$
in the rest frame of the $\tau$. The latter factor is the
$\nu_{\tau}$ spectrum predicted by the parton level $V-A$ matrix
element, and therefore represents an attempt to take into account
that the $\nu_{\tau}$ should take a larger momentum fraction than
given by phase space alone.
The probably largest shortcoming of the $\tau$ decay treatment is
that no polarization effects are included, i.e. the $\tau$ is
always assumed to decay isotropically. Usually this is not correct,
since a $\tau$ is produced polarized in $\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$ decays.
The \ttt{LUTAUD} routine provides a generic interface to an external
$\tau$ decay library, where such effects could be handled (see also
\ttt{MSTJ(28)}).
\subsubsection{Weak decays of bottom hadrons}
Some exclusive branching ratios now start to be known for $\mathrm{B}$
decays. In this version, the $\mathrm{B}^0$, $\mathrm{B}^+$, $\mathrm{B}_{\mathrm{s}}^0$ and
$\Lambda_{\b}^0$ therefore appear in a similar vein to the one
outlined above for $\mathrm{D}_{\mathrm{s}}^+$ and $\Lambda_{\c}^+$ above. That
is, all leptonic channels and all hadronic two-body decay channels
are explicitly listed, while hadronic channels with three or more
particles are only given in terms of a quark content. The $\mathrm{B}_{\c}$
is exceptional, in that either the bottom or the charm quark may
decay first, and in that annihilation graphs may be non-negligible.
Leptonic and semileptonic channels are here given in full, while
hadronic channels are only listed in terms of a quark content,
with a relative composition as given in \cite{Lus91}. No separate
branching ratios are set for any of the other weakly decaying
bottom hadrons, but instead a pure `spectator quark' model is assumed,
where the decay of the $\b$ quark is the same in all hadrons and the
only difference in final flavour content comes from the spectator quark.
Compared to the charm decays, the weak decay matrix elements are given
somewhat larger importance in the hadronic decay channels.
In semileptonic decays $\b \to \c \ell^- \br{\nu}_{\ell}$ the $\c$
quark is combined with the spectator antiquark or diquark to form
one single hadron. This hadron may be either a pseudoscalar, a vector
or a higher resonance (tensor etc.). The relative fraction of the
higher resonances has been picked to be about 30\%, in order to give
a leptonic spectrum in reasonable experiment with data. (This only
applies to the main particles $\mathrm{B}^0$, $\mathrm{B}^+$, $\mathrm{B}_{\mathrm{s}}^0$ and
$\Lambda_{\b}^0$; for the rest the choice is according to the standard
composition in the fragmentation.) The overall process is therefore
$H \to h \ell^- \br{\nu}_{\ell}$, where $H$ is a bottom antimeson
or a bottom baryon (remember that $\br{\mathrm{B}}$ is the one that contains
a $\b$ quark), and the matrix element used to distribute momenta is
\begin{equation}
|{\cal M}|^2 = (p_H p_{\nu}) (p_{\ell} p_h) ~.
\end{equation}
Again decay product masses have been neglected in the matrix element,
but in the branching ratios the $\tau^- \br{\nu}_{\tau}$ channel has
been reduced in rate, compared with $\mathrm{e}^- \br{\nu}_{\mathrm{e}}$ and
$\mu^- \br{\nu}_{\mu}$ ones, according to the expected mass effects.
No CKM-suppressed decays $\b \to \u \ell^- \br{\nu}_{\ell}$ are
currently included.
In most multibody hadronic decays, e.g.
$\b \to \c \d \overline{\mathrm{u}}$, the $\c$ quark is
again combined with the spectator flavour to form one single hadron,
and thereafter the hadron and the two quark momenta are distributed
according to the same matrix element as above, with
$\ell^- \leftrightarrow \d$ and
$\br{\nu}_{\ell} \leftrightarrow \overline{\mathrm{u}}$.
The invariant mass of the two quarks is calculated next. If this mass
is so low that two hadrons cannot be formed from the system, the
two quarks are combined into one single hadron. Else the same kind of
approach as in hadronic charm decays is adopted, wherein a
multiplicity is selected, a number of hadrons are formed and
thereafter momenta are distributed according to phase space. The
difference is that here the charm decay product is distributed according
to the $V-A$ matrix element, and only the rest of the system is
assumed isotropic in its rest frame, while in charm decays all hadrons
are distributed isotropically.
Note that the $\c$ quark and the spectator are assumed to form
one colour singlet and the $\d \overline{\mathrm{u}}$ another, separate one.
It is thus assumed that the original colour assignments of the
basic hard process are better retained than in charm decays.
However, sometimes this will not be true, and with about 20\%
probability the colour assignment is flipped around so that
$\c \overline{\mathrm{u}}$ forms one singlet. (In the program, this is achieved by
changing the order in which decay products are given.) In particular,
the decay $\b \to \c \mathrm{s} \overline{\mathrm{c}}$ is allowed to give a $\c\overline{\mathrm{c}}$
colour-singlet state part of the time, and this state may collapse
to a single $\mathrm{J}/\psi$. Two-body decays of this type are explicitly
listed for $\mathrm{B}^0$, $\mathrm{B}^+$, $\mathrm{B}_{\mathrm{s}}^0$ and $\Lambda_{\b}^0$;
while other $\mathrm{J}/\psi$ production channels appear from the flavour
content specification.
The $\mathrm{B}^0$--$\br{\mathrm{B}}^0$ and $\mathrm{B}_{\mathrm{s}}^0$--$\br{\mathrm{B}}_{\mathrm{s}}^0$ systems
mix before decay. This is optionally included. With a probability
\begin{equation}
{\cal P}_{\mrm{flip}} = \sin^2 \left( \frac{x \, \tau}
{ 2\, \langle \tau \rangle} \right)
\end{equation}
a $\mathrm{B}$ is therefore allowed to decay like a $\br{\mathrm{B}}$, and vice versa.
The mixing parameters are by default $x_{\d} = 0.7$ in the
$\mathrm{B}^0$--$\br{\mathrm{B}}^0$ system and $x_{\mathrm{s}} = 10$ in the
$\mathrm{B}_{\mathrm{s}}^0$--$\br{\mathrm{B}}_{\mathrm{s}}^0$ one.
The generic $\mathrm{B}$ meson and baryon decay properties are stored for
`particle' 85. This particle contains a description of the free $\b$
quark decay, with an instruction to find the spectator flavour
according to the particle code of the actual decaying hadron.
Currently baryons other than $\Lambda_{\b}^0$ are treated this
way. If so desired, each hadron could be given a separate decay
channel list, or all $\mathrm{B}$ hadrons could be mapped to particle 85,
as used to be the case..
\subsubsection{Weak decays of top and fourth generation}
As already explained in section \ref{ss:decpartcont},
heavy quarks are normally assumed to decay before they fragment.
Optionally, they may be allowed to fragment before they decay.
In either case, the decay itself is handled as
if the heavy flavour is free.
For a hadron, some of the hadron energy is reserved for the spectator
quark. The decay matrix element used for $\mathrm{Q} \to \mathrm{q} \overline{\mathrm{f}} \mathrm{f}$ is
\begin{equation}
|{\cal M}|^2 \propto \frac{ (p_{\mathrm{Q}} p_{\overline{\mathrm{f}}}) (p_{\mathrm{f}} p_{\mathrm{q}})}
{\left( (p_{\mathrm{f}} + p_{\overline{\mathrm{f}}})^2 - m_{\mathrm{W}}^2 \right)^2 +
m_{\mathrm{W}}^2 \Gamma_{\mathrm{W}}^2} ~.
\end{equation}
Here $\mathrm{Q}$ may represent the $\t$ or any of the fourth generation
quarks, $\mathrm{l}$ and $\mathrm{h}$. With trivial change of
notation, the lepton $\chi$ obeys the same formula. The $\mathrm{f}\overline{\mathrm{f}}$
pair are the fermions from the $\mathrm{W}$ decay, either quarks or leptons.
The program takes care of the effects of the $\mathrm{W}$ propagator,
whatever the mass difference $m_{\mathrm{Q}} - m_{\mathrm{q}}$, with one proviso:
the selection of the $\mathrm{q}$ flavour is done according to fixed branching
ratios, and does thus not take into account the relative enhancement
of a CKM-suppressed $\mathrm{q}$ due to mass effects. This would play a
r\^ole around thresholds, e.g., with $m_{\t} \approx m_{\mathrm{W}}$, the
$\t \to \mathrm{s}$ would be enhanced compared with $\t \to \b$.
On the other hand, threshold factors are included for the
choice of the $\mathrm{f}\overline{\mathrm{f}}$ fermion pair from the $\mathrm{W}$ decay.
For the alternative with a rapidly decaying top quark, so that no
hadron is formed, one is not close to threshold. The composition of
the light flavour produced in the decay is then calculated according
to the respective phase space times CKM weight. By default the $\mathrm{W}$
decays with the spin information implicit in the matrix element above,
but isotropic $\mathrm{W}$ decay is an option.
The $\b$ quark produced in the decay $\t \to \b\mathrm{W}^+$ may be allowed
to radiate. It thereby acquires an effective mass, which means that
the kinematics of the decay is changed, with energy shuffled from the
$\mathrm{W}$ to the $\b$.
The system containing the spectator quark will often have a mass
too small to allow it to fragment like a jet system. In these cases
a single particle is formed from the flavour content, with a momentum
vector given by the sum of the two quark momenta. Since the energy of
this particle then will come out wrong, the momenta of the other jets
or leptons in the decay are modified slightly to obtain total energy
conservation. (Of course, for $\chi$ decay, there is no spectator and
thus no treatment of this kind.)
The $\mathrm{f} \overline{\mathrm{f}}$ pair from the $\mathrm{W}$ decay is allowed to shower, i.e.
emit gluons and photons according to the standard final-state
radiation algorithm, including matching to first-order matrix
elements. The resulting jet system is fragmented with
ordinary string fragmentation --- the mass is here so high that
a fragmentation description is quite appropriate. Only very rarely
would the $\mathrm{W}$ mass be below the threshold for the production of a
pair of particles; such kinematical configurations are rejected.
\subsubsection{Other decays}
For onia spin 1 resonances, decay channels into a pair of leptons
are explicitly given. Hadronic decays of the $\mathrm{J}/\psi$ are simulated
using the flavour generation model introduced for charm. For
$\Upsilon$ a fraction of the hadronic decays is into $\mathrm{q}\overline{\mathrm{q}}$
pairs, while the rest is into $\mathrm{g}\g\mathrm{g}$ or $\mathrm{g}\g\gamma$, using the
matrix elements of eq.~(\ref{ee:Upsilondec}). The $\eta_c$ and
$\eta_b$ are both allowed to decay into a $\mathrm{g}\g$ pair, which then
subsequently fragments. In $\Upsilon$ and $\eta_b$ decays the partons
are allowed to shower before fragmentation, but energies are too low
for showering to have any impact.
With current bounds on the top mass, one does not expect the
formation of well-defined toponium states. A complete description of
the resonance structure in the threshold region is beyond the scope
of the program. The approach taken for the toponium states that
have been defined is to let either the $\t$ or the $\overline{\mathrm{t}}$ decay
weakly first, then do the fragmentation, and subsequently let the
produced antitop or top hadron decay. A better description is provided
by the {\tsc{Pythia}} machinery for resonance decays.
Default branching ratios are given for resonances like the $\mathrm{Z}^0$,
the $\mathrm{W}^{\pm}$ or the $\H^0$. When {\tsc{Pythia}} is initialized, these
numbers are replaced by branching ratios evaluated from the given
masses. For $\mathrm{Z}^0$ and $\mathrm{W}^{\pm}$ the branching ratios depend only
marginally on the masses assumed, while effects are large e.g. for
the $\H^0$. In fact, branching ratios may vary over the
Breit--Wigner resonance shape, something which is also taken into
account in {\tsc{Pythia}}. Therefore the default resonance treatment of {\tsc{Jetset}}
is normally not so useful, and should be avoided (except, of course,
the standard $\e^+\e^- \to \gamma^* / \Z^0 \to \mathrm{q}\overline{\mathrm{q}}$ description). When it is
used, a channel is selected according to the given fixed branching
ratios. If the decay is into a $\mathrm{q}\overline{\mathrm{q}}$ pair, the quarks are allowed
to shower and subsequently the parton system is fragmented.
\clearpage
\section{The JETSET Program Elements}
In this section we collect information on most of the routines and
common block variables found in {\tsc{Jetset}}. A few parts are discussed
elsewhere; this includes the $\e^+\e^-$ routines, parton showers
and event-analysis routines. In this section
the emphasis is on the fragmentation and decay package, and on
generic utilities for things like event listing.
\subsection{Definition of Initial Configuration or Variables}
With the use of the conventions described for the event record, it
is possible to specify any initial jet/particle configuration. This
task is simplified for a number of often occuring situations by
the existence of the filling routines below. It should be noted that
many users do not come in direct contact with these routines, since
that is taken care of by higher-level routines for specific
processes, particularly \ttt{LUEEVT} and \ttt{PYEVNT}.
Several calls to the routines can be combined in the specification.
In case one call is enough, the complete
fragmentation/decay chain may be simulated at the same time. At each
call, the value of \ttt{N} is updated to the last line used for
information in the call, so if several calls are used, they should be
made with increasing \ttt{IP} number, or else \ttt{N} should be
redefined by hand afterwards.
The routine \ttt{LUJOIN} is very useful
to define the colour flow in more complicated parton configurations;
thereby one can bypass the not so trivial rules for how to set the
\ttt{K(I,4)} and \ttt{K(I,5)} colour-flow information.
As an experiment, the routine \ttt{LUGIVE} contains a facility to set
various comonblock variables in a controlled and documented fashion.
\drawbox{CALL LU1ENT(IP,KF,PE,THE,PHI)}\label{p:LU1ENT}
\begin{entry}
\itemc{Purpose:} to add one entry to the event record, i.e. either
a jet or a particle.
\iteme{IP :} normally line number for the jet/particle. There are two
exceptions. \\
If \ttt{IP=0}, line number 1 is used and \ttt{LUEXEC} is called. \\
If \ttt{IP<0}, line \ttt{-IP} is used, with status code
\ttt{K(-IP,2)=2} rather than 1; thus a jet system may be built up by
filling all but the last jet of the system with \ttt{IP<0}.
\iteme{KF :} jet/particle flavour code.
\iteme{PE :} jet/particle energy. If \ttt{PE} is smaller than the mass,
the jet/particle is taken to be at rest.
\iteme{THE, PHI :} polar and azimuthal angle for the momentum vector
of the jet/particle.
\end{entry}
\drawbox{CALL LU2ENT(IP,KF1,KF2,PECM)}\label{p:LU2ENT}
\begin{entry}
\itemc{Purpose:} to add two entries to the event record, i.e. either a
2-jet system or two separate particles.
\iteme{IP :} normally line number for the first jet/particle, with
second in line \ttt{IP+1}. There are two exceptions. \\
If \ttt{IP=0}, lines 1 and 2 are used and \ttt{LUEXEC} is called. \\
If \ttt{IP<0}, lines \ttt{-IP} and \ttt{-IP+1} are used, with status
code \ttt{K(I,1)=3}, i.e. with special colour connection information,
so that a parton shower can be generated by a \ttt{LUSHOW} call,
followed by a \ttt{LUEXEC} call, if so desired (only relevant for jets).
\iteme{KF1, KF2 :} flavour codes for the two jets/particles.
\iteme{PECM :} ($=E_{\mrm{cm}}$) the total energy of the system.
\itemc{Remark:} the system is given in the c.m. frame, with the
first jet/particle going out in the $+z$ direction.
\end{entry}
\drawbox{CALL LU3ENT(IP,KF1,KF2,KF3,PECM,X1,X3)}\label{p:LU3ENT}
\begin{entry}
\itemc{Purpose:} to add three entries to the event record, i.e. either
a 3-jet system or three separate particles.
\iteme{IP :} normally line number for the first jet/particle, with
other two in \ttt{IP+1} and \ttt{IP+2}. There are two exceptions. \\
If \ttt{IP=0}, lines 1, 2 and 3 are used and \ttt{LUEXEC} is
called. \\
If \ttt{IP<0}, lines \ttt{-IP} through \ttt{-IP+2} are used, with
status code \ttt{K(I,1)=3}, i.e. with special colour connection
information, so that a parton shower can be generated by a \ttt{LUSHOW}
call, followed by a \ttt{LUEXEC} call, if so desired (only relevant
for jets).
\iteme{KF1, KF2, KF3:} flavour codes for the three jets/particles.
\iteme{PECM :} ($E_{\mrm{cm}}$) the total energy of the system.
\iteme{X1, X3 :} $x_i = 2E_i/E_{\mrm{cm}}$, i.e. twice the energy
fraction taken by the $i$'th jet. Thus $x_2 = 2 - x_1 - x_3$, and
need not be given. Note that not all combinations of $x_i$ are
inside the physically allowed region.
\itemc{Remark :} the system is given in the c.m. frame, in the
$xz$-plane, with the first jet going out in the $+z$ direction and the
third one having $p_x > 0$.
\end{entry}
\drawbox{CALL LU4ENT(IP,KF1,KF2,KF3,KF4,PECM,X1,X2,X4,X12,X14)}%
\label{p:LU4ENT}
\begin{entry}
\itemc{Purpose:} to add four entries to the event record, i.e. either a
4-jet system or four separate particles (or, for $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}'\overline{\mathrm{q}}'$
events, two 2-jet systems).
\iteme{IP :} normally line number for the first jet/particle, with
other three in lines \ttt{IP+1}, \ttt{IP+2} and \ttt{IP+3}.
There are two exceptions. \\
If \ttt{IP=0}, lines 1, 2, 3 and 4 are used and \ttt{LUEXEC} is
called. \\
If \ttt{IP<0}, lines \ttt{-IP} through \ttt{-IP+3} are used, with
status code \ttt{K(I,1)=3}, i.e. with special colour connection
information, so that a parton shower can be generated by a
\ttt{LUSHOW} call, followed by a \ttt{LUEXEC} call, if so desired
(only relevant for jets).
\iteme{KF1,KF2,KF3,KF4 :} flavour codes for the four jets/particles.
\iteme{PECM :} ($=E_{\mrm{cm}}$) the total energy of the system.
\iteme{X1,X2,X4 :} $x_i = 2E_i/E_{\mrm{cm}}$, i.e. twice the energy
fraction taken by the $i$'th jet. Thus $x_3 = 2 - x_1 - x_2 - x_4$,
and need not be given.
\iteme{X12,X14 :} $x_{ij} = 2 p_i p_j/E_{\mrm{cm}}^2$, i.e. twice the
four-vector product of the momenta for jets $i$ and $j$, properly
normalized. With the masses known, other $x_{ij}$ may be constructed
from the $x_i$ and $x_{ij}$ given. Note that not all combinations of
$x_i$ and $x_{ij}$ are inside the physically allowed region.
\itemc{Remark:} the system is given in the c.m. frame, with the first
jet going out in the $+z$ direction and the fourth jet lying in the
$xz$-plane with $p_x > 0$. The second jet will have $p_y > 0$ and
$p_y < 0$ with equal probability, with the third jet balancing this
$p_y$ (this corresponds to a random choice between the two possible
stereoisomers).
\end{entry}
\drawbox{CALL LUJOIN(NJOIN,IJOIN)}\label{p:LUJOIN}
\begin{entry}
\itemc{Purpose:} to connect a number of previously defined partons
into a string configuration. Initially the partons must be given with
status codes \ttt{K(I,1)=} 1, 2 or 3. Afterwards the partons
all have status code 3, i.e. are given with full colour-flow
information. Compared to the normal way of defining a parton
system, the partons need therefore not appear in the same sequence
in the event record as they are assumed to do along the string. It
is also possible to call \ttt{LUSHOW} for all or some of the
entries making up the string formed by \ttt{LUJOIN}.
\iteme{NJOIN:} the number of entries that are to be joined by one
string.
\iteme{IJOIN:} an one-dimensional array, of size at least \ttt{NJOIN}.
The \ttt{NJOIN} first numbers are the positions of the partons that
are to be joined, given in the order the partons are assumed to appear
along the string. If the system consists entirely of gluons,
the string is closed by connecting back the last to the first
entry.
\itemc{Remarks:} only one string (i.e. one colour singlet) may be
defined per call, but one is at liberty to use any number of
\ttt{LUJOIN} calls for a given event. The program will check that the
parton configuration specified makes sense, and not take any action
unless it does. Note, however, that an initially sensible parton
configuration may become nonsensical, if only some of the partons
are reconnected, while the others are left unchanged.
\end{entry}
\drawbox{CALL LUGIVE(CHIN)}\label{p:LUGIVE}
\begin{entry}
\itemc{Purpose:} to set the value of any variable residing in the
commmonblocks \ttt{LUJETS}, \ttt{LUDAT1}, \ttt{LUDAT2}, \ttt{LUDAT3},
\ttt{LUDAT4}, \ttt{LUDATR}, \ttt{PYSUBS}, \ttt{PYPARS}, \ttt{PYINT1},
\ttt{PYINT2}, \ttt{PYINT3}, \ttt{PYINT4}, \ttt{PYINT5}, \ttt{PYINT6},
or \ttt{PYINT7}. This is done in a more controlled fashion than by
directly including the common blocks in the user program, in that
array bounds are checked and the old and new values for the variable
changed are written to the output for reference.
\iteme{CHIN :} character expression of length at most 100 characters,
with requests for variables to be changed, stored in the form \\
\ttt{variable1=value1;variable2=value2;variable3=value3}\ldots~. \\
Note that an arbitrary number of instructions can be stored in
one call if separated by semicolons, and that blanks may be included
anyplace. The variable$_i$ may be any single variable in the {\tsc{Jetset/Pythia}}
common blocks, and the value$_i$ must be of the correct integer, real
or character (without extra quotes) type. Array indices and values
must be given explicitly, i.e. cannot be variables in their own
right. The exception is that the first index can be preceded by
a \ttt{C}, signifying that the index should be translated from normal
KF to compressed KC code with a \ttt{LUCOMP} call; this is allowed for
the \ttt{KCHG}, \ttt{PMAS}, \ttt{MDCY} and \ttt{CHAF} arrays. If a
value$_i$ is omitted, i.e. with the construction \ttt{variable=},
the current value is written to the output, but the variable itself
is not changed.
\itemc{Remark :} The checks on array bounds are hardwired into this
routine. Therefore, if some user changes array dimensions and
\ttt{MSTU(3)}, \ttt{MSTU(6)} and/or \ttt{MSTU(7)}, as allowed by
other considerations, these changes will not be known to \ttt{LUGIVE}.
Normally this should not be a problem, however.
\end{entry}
\subsection{The JETSET Physics Routines}
\label{ss:JETphysrout}
The physics routines form the major part of {\tsc{Jetset}}, but once
the initial jet/particle configuration has been specified and
default parameter values changed, if so desired, only a \ttt{LUEXEC}
call is necessary to simulate the whole fragmentation and decay chain.
Therefore a normal user will not directly see any of the other routines
in this section. Some of them could be called directly, but the danger
of faulty usage is then non-negligible.
The \ttt{LUTAUD} routine provides an optional interface to an external
$\tau$ decay library, where polarization effects could be included.
It is up to the user to write the appropriate calls, as explained at
the end of this section.
\drawbox{CALL LUEXEC}\label{p:LUEXEC}
\begin{entry}
\itemc{Purpose:} to administrate the fragmentation and decay chain.
\ttt{LUEXEC} may be called several times, but only entries which have
not yet been treated (more precisely, which have
$1 \leq$\ttt{K(I,1)}$\leq 10$)
can be affected by further calls. This may apply if more jets/particles
have been added by the user, or if particles previously considered
stable are now allowed to decay. The actions that will be taken during
a \ttt{LUEXEC} call can be tailored extensively via the
\ttt{LUDAT1}--\ttt{LUDAT3} common blocks, in particular by setting the
\ttt{MSTJ} values suitably.
\end{entry}
\boxsep
\begin{entry}
\iteme{SUBROUTINE LUPREP(IP) :}\label{p:LUPREP}
to rearrange parton shower end products (marked with \ttt{K(I,1)=3})
sequentially along strings; also to (optionally) allow small jet
systems to collapse into two particles or one only, in the latter
case with energy and momentum to be shuffled elsewhere in the event;
also to perform checks that e.g. flavours of colour-singlet systems
make sense.
\iteme{SUBROUTINE LUSTRF(IP) :}\label{p:LUSTRF}
to generate the fragmentation of an arbitrary colour-singlet
jet system according to the Lund string fragmentation model. In many
respects, this routine is the very heart and soul of {\tsc{Jetset}}.
\iteme{SUBROUTINE LUINDF(IP) :}\label{p:LUINDF}
to handle the fragmentation of a jet system according to
independent fragmentation models, and implement energy, momentum
and flavour conservation, if so desired. Also the fragmentation of
a single jet, not belonging to a jet system, is considered here
(this is of course physical nonsense, but may sometimes be
convenient for specific tasks).
\iteme{SUBROUTINE LUDECY(IP) :}\label{p:LUDECY}
to perform a particle decay, according to known
branching ratios or different kinds of models, depending on our level
of knowledge. Various matrix elements are included for specific
processes.
\iteme{SUBROUTINE LUKFDI(KFL1,KFL2,KFL3,KF) :}\label{p:LUKFDI}
to generate a new quark or diquark flavour and to
combine it with an existing flavour to give a hadron.
\begin{subentry}
\iteme{KFL1:} incoming flavour.
\iteme{KFL2:} extra incoming flavour, e.g. for formation of final
particle, where the flavours are completely specified. Is normally 0.
\iteme{KFL3:} newly created flavour; is 0 if \ttt{KFL2} is non-zero.
\iteme{KF:} produced hadron. Is 0 if something went wrong (e.g.
inconsistent combination of incoming flavours).
\end{subentry}
\iteme{SUBROUTINE LUPTDI(KFL,PX,PY) :}\label{p:LUPTDI}
to give transverse momentum, e.g. for a $\mathrm{q}\overline{\mathrm{q}}$ pair created in the
colour field, according to independent Gaussian distributions in
$p_x$ and $p_y$.
\iteme{SUBROUTINE LUZDIS(KFL1,KFL3,PR,Z) :}\label{p:LUZDIS}
to generate the longitudinal scaling variable $z$ in jet
fragmentation, either according to the Lund symmetric fragmentation
function, or according to a choice of other shapes.
\iteme{SUBROUTINE LUBOEI :}\label{p:LUBOEI}
to include Bose--Einstein effects according to a simple
para\-metri\-zation. By default, this routine is not called.
If called from \ttt{LUEXEC}, this is done after the decay of
short-lived resonances, but before the decay of long-lived ones.
This means the routine should never be called directly by you,
nor would effects be correctly simulated if decays are switched off.
See \ttt{MSTJ(51) - MSTJ(52)} for switching on the routine.
\iteme{FUNCTION ULMASS(KF) :}\label{p:ULMASS}
to give the mass for a parton/particle.
\iteme{SUBROUTINE LUNAME(KF,CHAU) :}\label{p:LUNAME}
to give the parton/particle name (as a string of type
\ttt{CHARACTER CHAU*16}).
\iteme{FUNCTION LUCHGE(KF) :}\label{p:LUCHGE}
to give three times the charge for a parton/particle.
\iteme{FUNCTION LUCOMP(KF) :}\label{p:LUCOMP}
to give the compressed parton/particle code KC for a given
KF code, as required to find entry into mass and decay data tables.
Also checks whether the given KF code is actually an allowed one
(i.e. known by the program), and returns 0 if not. Note that KF
may be positive or negative, while the resulting KC code is never
negative.
\iteme{SUBROUTINE LUERRM(MERR,MESSAG) :}\label{p:LUERRM}
to keep track of the number of errors and warnings encountered,
write out information on them, and abort the program in case of too
many errors.
\iteme{FUNCTION ULANGL(X,Y) :}\label{p:ULANGL}
to calculate the angle from the $x$ and $y$ coordinates.
\iteme{SUBROUTINE LULOGO :}\label{p:LULOGO}
to write a titlepage for the {\tsc{Jetset/Pythia}} programs. Called by LULIST(0).
\iteme{BLOCK DATA LUDATA :}\label{p:LUDATA}
to give default values for variables in the \ttt{LUDAT1}, \ttt{LUDAT2},
\ttt{LUDAT3}, \ttt{LUDAT4} and \ttt{LUDATR} common blocks.
\end{entry}
\drawbox{CALL LUTAUD(ITAU,IORIG,KFORIG,NDECAY)}\label{p:LUTAUD}
\begin{entry}
\itemc{Purpose:} to act as an interface between the standard decay
routine \ttt{LUDECY} and a user-supplied $\tau$ lepton decay library.
The latter library would normally know how to handle polarized $\tau$'s,
given the $\tau$ polarization as input, so one task of the interface
routine is to construct the $\tau$ polarization/helicity from the
information available. Input to the routine (from \ttt{LUDECY}) is
provided in the first three arguments, while the last argument and
some event record information have to be set before return.
To use this facility you have to set the switch \ttt{MSTJ(28)},
include your own interface routine \ttt{LUTAUD} and see to it that
the dummy routine \ttt{LUTAUD} in {\tsc{Jetset}} is not linked. The dummy
routine is there only to avoid unresolved external references when
no user-supplied interface is linked.
\iteme{ITAU :} line number in the event record where the $\tau$ is
stored. The four-momentum of this $\tau$ has first been boosted back
to the rest frame of the decaying mother and thereafter rotated to move
out along the $+z$ axis. It would have been possible to also perform
a final boost to the rest frame of the $\tau$ itself, but this has been
avoided so as not to suppress the kinematics aspect of
close-to-threshold production (e.g. in $\mathrm{B}$ decays) vs.
high-energy production (e.g. in real $\mathrm{W}$ decays). The choice of frame
should help the calculation of the helicity configuration. After the
\ttt{LUTAUD} call the $\tau$ and its decay products will automatically
be rotated and boosted back. However, seemingly, the event record does
not conserve momentum at this intermediate stage.
\iteme{IORIG :} line number where the mother particle to the $\tau$
is stored. Is 0 if the mother is not stored. This does not have
to mean the mother is unknown. For instance, in semileptonic $\mathrm{B}$ decays
the mother is a $\mathrm{W}^{\pm}$ with known four-momentum
$p_{\mathrm{W}} = p_{\tau} + p_{\nu_{\tau}}$, but there is no $\mathrm{W}$ line in the
event record. When several copies of the mother is stored (e.g. one in
the documentation section of the event record and one in the main
section), \ttt{IORIG} points to the last. If a branchings like
$\tau \to \tau\gamma$ occurs, the `grandmother' is given, i.e. the
mother of the direct $\tau$ before branching.
\iteme{KFORIG :} flavour code for the mother particle. Is 0 if the
mother is unknown. The mother would typically be a resonance such as
$\gamma^* / \Z^0$ (23), $\mathrm{W}^{\pm}$ ($\pm 24$), $\H^0$ (25), or $\H^{\pm}$
($\pm 37$). Often the helicity choice would be clear just by the
knowledge of this mother species, e.g., $\mathrm{W}^{\pm}$ vs. $\H^{\pm}$.
However, sometimes further complications may exist. For instance,
the KF code 23 represents a mixture of $\gamma^*$ and $\mathrm{Z}^0$; a
knowledge of the mother mass (in \ttt{P(IORIG,5)}) would here be
required to make the choice of helicities. Further, a $\mathrm{W}^{\pm}$ or
$\mathrm{Z}^0$ may either be (predominantly) transverse or longitudinal,
depending on the production process under study.
\iteme{NDECAY :} the number of decay products of the $\tau$; to be
given by the user. You must also store the KF flavour codes of those
decay products in the positions \ttt{K(I,2)},
\ttt{N+1}$\leq$\ttt{I}$\leq$\ttt{N+NDECAY}, of the event record.
The corresponding five-momentum (momentum, energy and mass) should be
stored in the associated \ttt{P(I,J)} positions, 1$\leq$\ttt{J}$\leq$5.
The four-momenta are expected to add up to the four-momentum of the
$\tau$ in position \ttt{ITAU}. You should not change the \ttt{N} value
or any of the other \ttt{K} or \ttt{V} values (neither for the
$\tau$ nor for its decay products) since this is automatically done
in \ttt{LUDECY}.
\end{entry}
\subsection{Event Study and Data Listing Routines}
After an \ttt{LUEXEC} call, the event generated is stored in the
\ttt{LUJETS} common block, and whatever physical variable is desired
may be constructed from this record. An event may be rotated, boosted
or listed, and particle data may be listed or modified. Via the
functions \ttt{KLU} and \ttt{PLU} the values of some frequently
appearing variables may be obtained more easily. As described in
section \ref{ss:evanrout}, also more detailed event shape analyses
may be performed simply.
\drawbox{CALL LUROBO(THE,PHI,BEX,BEY,BEZ)}\label{p:LUROBO}
\begin{entry}
\itemc{Purpose:} to perform rotations and Lorentz boosts (in that
order, if both in the same call) of jet/particle momenta and vertex
position variables.
\iteme{THE, PHI :} standard polar coordinates $\theta, \varphi$,
giving the rotated direction of a momentum vector initially along
the $+z$ axis.
\iteme{BEX, BEY, BEZ :} gives the direction and size
\mbox{\boldmath $\beta$} of a Lorentz boost,
such that a particle initially at rest will have
$\mbf{p}/E = $\mbox{\boldmath $\beta$} afterwards.
\itemc{Remark:} all entries 1 through \ttt{N} are affected by the
transformation, unless lower and upper bounds are explicitly given
by \ttt{MSTU(1)} and \ttt{MSTU(2)}, or if status code
\ttt{K(I,1)}$\leq 0$.
\end{entry}
\drawbox{ENTRY LUDBRB(IMI,IMA,THE,PHI,DBEX,DBEY,DBEZ)}\label{p:LUDBRB}
\begin{entry}
\itemc{Purpose:} to perform rotations and Lorentz boosts (in that
order, if both in the same call) of jet/particle momenta and vertex
position variables, for a specific range of entries, and with the
boost vector given in double precision. Is entry to \ttt{LUROBO},
mainly intended for internal use.
\iteme{IMI, IMA :} range of entries affected by transformation,
\ttt{IMI}$\leq$\ttt{I}$\leq$\ttt{IMA}.
\iteme{THE, PHI :} standard polar coordinates $\theta, \varphi$,
giving the rotated direction of a momentum vector initially along
the $+z$ axis.
\iteme{DBEX, DBEY, DBEZ :} gives the direction and size
\mbox{\boldmath $\beta$} of a Lorentz boost,
such that a particle initially at rest will have
$\mbf{p}/E = $\mbox{\boldmath $\beta$} afterwards.
Is to be given in double precision.
\itemc{Remark:} all entries with status codes \ttt{K(I,1)>0} in
the requested range are affected by the transformation.
\end{entry}
\drawbox{CALL LUEDIT(MEDIT)}\label{p:LUEDIT}
\begin{entry}
\itemc{Purpose:} to exclude unstable or undetectable jets/particles
from the event record. One may also use \ttt{LUEDIT} to store spare
copies of events (specifically initial parton configuration) that can
be recalled to allow e.g. different fragmentation schemes to be run
through with one and the same parton configuration. Finally, an
event which has been analyzed with \ttt{LUSPHE}, \ttt{LUTHRU} or
\ttt{LUCLUS} (see section \ref{ss:evanrout}) may be rotated to align
the event axis with the $z$ direction.
\iteme{MEDIT :} tells which action is to be taken.
\begin{subentry}
\iteme{= 0 :} empty (\ttt{K(I,1)=0}) and documentation
(\ttt{K(I,1)>20}) lines
are removed. The jets/particles remaining are compressed in the
beginning of the \ttt{LUJETS} common block and the \ttt{N} value is
updated accordingly. The event history is lost, so that information
stored in \ttt{K(I,3)}, \ttt{K(I,4)} and \ttt{K(I,5)} is no longer
relevant.
\iteme{= 1 :} as \ttt{=0}, but in addition all jets/particles that have
fragmented/decayed (\ttt{K(I,1)>10}) are removed.
\iteme{= 2 :} as \ttt{=1}, but also all neutrinos and unknown particles
(i.e. compressed code KC$ = 0$) are removed.
\iteme{= 3 :} as \ttt{=2}, but also all uncharged, colour neutral
particles are removed, leaving only charged, stable particles (and
unfragmented partons, if fragmentation has not been performed).
\iteme{= 5 :} as \ttt{=0}, but also all partons which have branched or
been rearranged in a parton shower and all particles which have
decayed are removed, leaving only the fragmenting parton
configuration and the final-state particles.
\iteme{= 11 :} remove lines with \ttt{K(I,1)<0}. Update event
history information (in \ttt{K(I,3) - K(I,5)}) to refer to remaining
entries.
\iteme{= 12 :} remove lines with \ttt{K(I,1)=0}. Update event
history information (in \ttt{K(I,3) - K(I,5)}) to refer to remaining
entries.
\iteme{= 13 :} remove lines with \ttt{K(I,1)}$ = 11$, 12 or 15, except
for any line with \ttt{K(I,2)=94}. Update event history information
(in \ttt{K(I,3) - K(I,5)}) to refer to remaining entries. In
particular, try to trace origin of daughters, for which the mother
is decayed, back to entries not deleted.
\iteme{= 14 :} remove lines with \ttt{K(I,1)}$ = 13$ or 14, and also
any line with \ttt{K(I,2)=94}. Update event history information
(in \ttt{K(I,3) - K(I,5)}) to refer to remaining entries. In
particular, try to trace origin of rearranged jets back through
the parton-shower history to the shower initiator.
\iteme{= 15 :} remove lines with \ttt{K(I,1)>20}. Update event
history information (in \ttt{K(I,3) - K(I,5)}) to refer to remaining
entries.
\iteme{= 16 :} try to reconstruct missing daughter pointers of
decayed particles from the mother pointers of decay products.
These missing pointers typically come from the need to use
\ttt{K(I,4)} and \ttt{K(I,5)} also for colour flow information.
\iteme{= 21 :} all partons/particles in current event record are
stored (as a spare copy) in bottom of common block \ttt{LUJETS}
(is e.g. done to save original partons before calling \ttt{LUEXEC}).
\iteme{= 22 :} partons/particles stored in bottom of event record with
\ttt{=21} are placed in beginning of record again, overwriting previous
information there (so that e.g. a different fragmentation scheme
can be used on the same partons). Since the copy at bottom is
unaffected, repeated calls with \ttt{=22} can be made.
\iteme{= 23 :} primary partons/particles in the beginning of event
record are marked as not fragmented or decayed, and number of entries
\ttt{N} is updated accordingly. Is simpe substitute for \ttt{=21} plus
\ttt{=22} when no fragmentation/decay products precede any of the
original partons/particles.
\iteme{= 31 :} rotate largest axis, determined by \ttt{LUSPHE},
\ttt{LUTHRU} or \ttt{LUCLUS}, to sit along the $z$ direction, and the
second largest axis into the $xz$ plane. For \ttt{LUCLUS} it can be
further specified to $+z$ axis and $xz$ plane with $x > 0$,
respectively. Requires that one of these routines has been called
before.
\iteme{= 32 :} mainly intended for \ttt{LUSPHE} and \ttt{LUTHRU}, this
gives a further alignment of the event, in addition to the one implied
by \ttt{=31}. The `slim' jet, defined as the side ($z > 0$ or $z < 0$)
with the smallest summed $p_{\perp}$ over square root of number of particles,
is rotated into the $+z$ hemisphere. In the opposite hemisphere
(now $z < 0$), the side of $x > 0$ and $x < 0$ which has the largest
summed $|p_z|$ is rotated into the $z < 0, x > 0$ quadrant.
Requires that \ttt{LUSPHE} or \ttt{LUTHRU} has been called before.
\end{subentry}
\itemc{Remark:} all entries 1 through \ttt{N} are affected by the
editing. For options 0--5 lower and upper bounds can be explicitly
given by \ttt{MSTU(1)} and \ttt{MSTU(2)}.
\end{entry}
\drawbox{CALL LULIST(MLIST)}\label{p:LULIST}
\begin{entry}
\itemc{Purpose:} to list an event, jet or particle data, or current
parameter values.
\iteme{MLIST :} determines what is to be listed.
\begin{subentry}
\iteme{= 0 :} writes a title page, common for {\tsc{Jetset}} and {\tsc{Pythia}}, with
program version numbers and last dates of change; is mostly for
internal use.
\iteme{= 1 :} gives a simple list of current event record, in an 80
column format suitable for viewing directly on the computer terminal.
For each entry, the following information is given: the entry
number \ttt{I}, the parton/particle name (see below), the status code
(\ttt{K(I,1)}), the flavour code KF (\ttt{K(I,2)}), the line number
of the mother (\ttt{K(I,3)}), and the three-momentum, energy and mass
(\ttt{P(I,1) - P(I,5)}). If \ttt{MSTU(3)} is non-zero, lines immediately
after the event record proper are also listed. A final line
contains information on total charge, momentum, energy and
invariant mass. \\
The particle name is given by a call to the routine \ttt{LUNAME}.
For an entry which has decayed/fragmented (\ttt{K(I,1)=} 11--20),
this particle name is given within parentheses. Similarly, a
documentation line (\ttt{K(I,1)=} 21--30) has the name enclosed in
expression signs (!\ldots!) and an event/jet axis information line
the name within inequality signs ($<$\ldots$>$). If the last
character of the name is a `?', it is a signal that the complete name
has been truncated to fit in, and can therefore not be trusted;
this is very rare. For partons which have been arranged along
strings (\ttt{K(I,1)=} 1, 2, 11 or 12), the end of the parton name
column contains information about the colour string arrangement:
an \ttt{A} for the first entry of a string, an \ttt{I} for all
intermediate ones, and a \ttt{V} for the final one (a poor man's
rendering of a vertical doublesided arrow, $\updownarrow$). \\
It is possible to insert lines just consisting of sequences
of \ttt{======} to separate different sections of the event record,
see \ttt{MSTU(70) - MSTU(80)}.
\iteme{= 2 :} gives a more extensive list of the current event record,
in a 132 column format, suitable for printers or workstations.
For each entry, the following information is given: the entry
number \ttt{I}, the parton/particle name (with padding as described
for \ttt{=1}), the status code (\ttt{K(I,1)}), the flavour code
KF (\ttt{K(I,2)}), the line number of the mother (\ttt{K(I,3)}), the
decay product/colour-flow pointers (\ttt{K(I,4), K(I,5)}), and the
three-momentum, energy and mass (\ttt{P(I,1) - P(I,5)}). If
\ttt{MSTU(3)} is non-zero, lines immediately after the event record
proper are also listed. A final line contains information on total
charge, momentum, energy and invariant mass. Lines with only
\ttt{======} may be inserted as for \ttt{=1}.
\iteme{= 3 :} gives the same basic listing as \ttt{=2}, but with an
additional line for each entry containing information on production
vertex position and time (\ttt{V(I,1) - V(I,4)}) and, for unstable
particles, proper lifetime (\ttt{V(I,5)}).
\iteme{= 11 :} provides a simple list of all parton/particle codes
defined in the program, with KF code and corresponding particle name.
The list is grouped by particle kind, and only within each group
in ascending order.
\iteme{= 12 :} provides a list of all parton/particle and decay data
used in the program. Each parton/particle code is represented by one
line containing KF flavour code, KC compressed code, particle
name, antiparticle name (where appropriate), electrical and
colour charge (stored in \ttt{KCHG}), mass, resonance width and
maximum broadening, average proper lifetime (in \ttt{PMAS}) and
whether the particle is considered stable or not (in \ttt{MDCY}).
Immediately after a particle, each decay channel gets one line,
containing decay channel number (\ttt{IDC} read from \ttt{MDCY}),
on/off switch for the channel, matrix element type (\ttt{MDME}),
branching ratio (\ttt{BRAT}), and decay products (\ttt{KFDP}).
The \ttt{MSTU(14)} flag can be used to set the maximum flavour
for which particles are listed, with the default (= 0)
corresponding to separately defined ones (KC$ > 100$ if KF$ > 0$).
In order to keep the size down, decay modes of heavy hadrons
collectively defined are never listed; these have KC codes
84--88, where the relevant information may be found.
\iteme{= 13 :} gives a list of current parameter values for
\ttt{MSTU}, \ttt{PARU}, \ttt{MSTJ} and \ttt{PARJ}, and the first
200 entries of \ttt{PARF}. This is useful to keep check of which
default values were changed in a given run.
\end{subentry}
\itemc{Remark:} for options 1--3 and 12 lower and upper bounds of the
listing can be explicitly given by \ttt{MSTU(1)} and \ttt{MSTU(2)}.
\end{entry}
\drawbox{CALL LUUPDA(MUPDA,LFN)}\label{p:LUUPDA}
\begin{entry}
\itemc{Purpose:} to give you the ability to update particle data,
or to keep several versions of modified particle data for special
purposes (e.g. charm studies).
\iteme{MUPDA :} gives the type of action to be taken.
\begin{subentry}
\iteme{= 1 :} write a table of particle data, that you then can
edit at leisure. For ordinary listing of decay data, \ttt{LULIST(12)}
should be used, but that listing could not be read back in
by the program. \\
For each compressed flavour code KC = 1--500, one line is
written containing KC (\ttt{I5}), the basic particle name (i.e.
excluding charge etc.) (\ttt{2X,A8}) in \ttt{CHAF}, the electric
(\ttt{I3}), colour charge (\ttt{I3}) and particle/antiparticle
distinction (\ttt{I3}) codes in \ttt{KCHG}, the mass (\ttt{F12.5}),
the mass width (\ttt{F12.5}), maximum broadening (\ttt{F12.5}) and
average proper lifetime (\ttt{2X,F12.5}) in \ttt{PMAS}, and the
on/off decay switch (I3) in \ttt{MDCY(KC,1).} \\
After a KC line follows one line for each possible decay
channel, containing the \ttt{MDME} codes (\ttt{5X,2I5}), the branching
ratio (\ttt{5X,F12.5}) in \ttt{BRAT}, and the \ttt{KFDP} codes for the
decay products (\ttt{5I8}), with trailing 0's if the number of decay
products is smaller than 5.
\iteme{= 2 :} read in particle data, as written with \ttt{=1} and
thereafter edited by you, and use this data subsequently in
the current run. Reading is done with fixed format,
which means that you have to preserve the format codes
described for \ttt{=1} during the editing. A number of checks will
be made to see if input looks reasonable, with warnings if not.
If some decay channel is said not to conserve charge, it should
be taken seriously. Warnings that decay is kinematically
unallowed need not be as serious, since that particular decay
mode may not be switched on unless the particle mass is
increased.
\iteme{= 3 :} write current particle data as data lines, which
can be edited into \ttt{BLOCK DATA LUDATA} for a permanent
replacement of the particle data. This option is intended for the
program author only, not for you.
\end{subentry}
\iteme{LFN :} the file number which the data should be written to or
read from. You must see to it that this file is properly
opened for read or write (since the definition of file names
is machine dependent).
\end{entry}
\drawbox{KK = KLU(I,J)}\label{p:KLU}
\begin{entry}
\itemc{Purpose:} to provide various integer-valued event data. Note
that many of the options available (in particular \ttt{I}$ > 0$,
\ttt{J}$\geq 14$) which refer to event history will not work after a
\ttt{LUEDIT} call. Further, the options 14--18 depend on the way the
event history has been set up, so with the explosion of different
allowed formats these options are no longer as safe as they may have
been. For instance, option 16 can only work if \ttt{MSTU(16)=2}.
\iteme{I=0, J= :} properties referring to the complete event.
\begin{subentry}
\iteme{= 1 :} \ttt{N}, total number of lines in event record.
\iteme{= 2 :} total number of partons/particles remaining after
fragmentation and decay.
\iteme{= 6 :} three times the total charge of remaining (stable)
partons and particles.
\end{subentry}
\iteme{I>0, J= :} properties referring to the entry in line no.
\ttt{I} of the event record.
\begin{subentry}
\iteme{= 1 - 5 :} \ttt{K(I,1) - K(I,5)}, i.e. parton/particle status
KS, flavour code KF and origin/decay product/colour-flow information.
\iteme{= 6 :} three times parton/particle charge.
\iteme{= 7 :} 1 for a remaining entry, 0 for a decayed, fragmented or
documentation entry.
\iteme{= 8 :} KF code (\ttt{K(I,2)}) for a remaining entry, 0 for a
decayed, fragmented or documentation entry.
\iteme{= 9 :} KF code (\ttt{K(I,2)}) for a parton (i.e. not colour
neutral entry), 0 for a particle.
\iteme{= 10 :} KF code (\ttt{K(I,2)}) for a particle (i.e. colour
neutral entry), 0 for a parton.
\iteme{= 11 :} compressed flavour code KC.
\iteme{= 12 :} colour information code, i.e. 0 for colour neutral,
1 for colour triplet, -1 for antitriplet and 2 for octet.
\iteme{= 13 :} flavour of `heaviest' quark or antiquark (i.e. with
largest code) in hadron or diquark (including sign for antiquark),
0 else.
\iteme{= 14 :} generation number. Beam particles or virtual exchange
particles are generation 0, original jets/particles generation
1 and then 1 is added for each step in the fragmentation/decay
chain.
\iteme{= 15 :} line number of ancestor, i.e. predecessor in first
generation (generation 0 entries are disregarded).
\iteme{= 16 :} rank of a hadron in the jet it belongs to. Rank denotes
the ordering in flavour space, with hadrons containing the original
flavour of the jet having rank 1, increasing by 1 for each step
away in flavour ordering. All decay products inherit the rank
of their parent. Whereas the meaning of a first-rank hadron
in a quark jet is always well-defined, the definition of higher
ranks is only meaningful for independently fragmenting quark
jets. In other cases, rank refers to the ordering in the actual
simulation, which may be of little interest.
\iteme{= 17 :} generation number after a collapse of a jet system into
one particle, with 0 for an entry not coming from a collapse, and
-1 for entry with unknown history. A particle formed in a
collapse is generation 1, and then one is added in each decay
step.
\iteme{= 18 :} number of decay/fragmentation products (only defined
in a collective sense for fragmentation).
\iteme{= 19 :} origin of colour for showering parton, 0 else.
\iteme{= 20 :} origin of anticolour for showering parton, 0 else.
\iteme{= 21 :} position of colour daughter for showering parton,
0 else.
\iteme{= 22 :} position of anticolour daughter for showering parton,
0 else.
\end{subentry}
\end{entry}
\drawbox{PP = PLU(I,J)}\label{p:PLU}
\begin{entry}
\itemc{Purpose:} to provide various real-valued event data. Note that
some of the options available (\ttt{I}$ > 0$, \ttt{J}$ = 20$--25),
which are primarily intended for studies of systems in their
respective c.m. frame, requires that a \ttt{LUEXEC} call has been
made for the current initial parton/particle configuration, but that
the latest \ttt{LUEXEC} call has not been followed by a \ttt{LUROBO}
one.
\iteme{I=0, J= :} properties referring to the complete event.
\begin{subentry}
\iteme{= 1 - 4 :} sum of $p_x$, $p_y$, $p_z$ and $E$, respectively,
for the stable remaining entries.
\iteme{= 5 :} invariant mass of the stable remaining entries.
\iteme{= 6 :} sum of electric charge of the stable remaining entries.
\end{subentry}
\iteme{I>0, J= :} properties referring to the entry in line no.
\ttt{I} of the event record.
\begin{subentry}
\iteme{= 1 - 5 :} \ttt{P(I,1) - P(I,5)}, i.e. normally $p_x$, $p_y$,
$p_z$, $E$ and $m$ for jet/particle.
\iteme{= 6 :} electric charge $e$.
\iteme{= 7 :} squared momentum $|\mbf{p}|^2 = p_x^2 + p_y^2 + p_z^2$.
\iteme{= 8 :} absolute momentum $|\mbf{p}|$.
\iteme{= 9 :} squared transverse momentum $p_{\perp}^2 = p_x^2 + p_y^2$.
\iteme{= 10 :} transverse momentum $p_{\perp}$.
\iteme{= 11 :} squared transverse mass
$m_{\perp}^2 = m^2 + p_x^2 + p_y^2$.
\iteme{= 12 :} transverse mass $m_{\perp}$.
\iteme{= 13 - 14 :} polar angle $\theta$ in radians (between 0 and
$\pi$) or degrees, respectively.
\iteme{= 15 - 16 :} azimuthal angle $\varphi$ in radians (between
$-\pi$ and $\pi$) or degrees, respectively.
\iteme{= 17 :} true rapidity $y = (1/2) \, \ln((E+p_z)/(E-p_z))$.
\iteme{= 18 :} rapidity $y_{\pi}$ obtained by assuming that the
particle is a pion when calculating the energy $E$, to be used in the
formula above, from the (assumed known) momentum $\mbf{p}$.
\iteme{= 19 :} pseudorapidity $\eta = (1/2) \, \ln((p+p_z)/(p-p_z))$.
\iteme{= 20 :} momentum fraction $x_p = 2|\mbf{p}|/W$, where $W$
is the total energy of initial jet/particle configuration.
\iteme{= 21 :} $x_{\mrm{F}} = 2p_z/W$ (Feynman-$x$ if system is
studied in the c.m. frame).
\iteme{= 22 :} $x_{\perp} = 2p_{\perp}/W$.
\iteme{= 23 :} $x_E = 2E/W$.
\iteme{= 24 :} $z_+ = (E+p_z)/W$.
\iteme{= 25 :} $z_- = (E-p_z)/W$.
\end{subentry}
\end{entry}
\subsection{The General Switches and Parameters}
\label{ss:JETswitch}
The common block \ttt{LUDAT1} is, next to \ttt{LUJETS}, the one a {\tsc{Jetset}}
user is most likely to access. Here one may control in detail what the
program is to do, if the default mode of operation is not satisfactory.
\drawbox{COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)}%
\label{p:LUDAT1}
\begin{entry}
\itemc{Purpose:} to give access to a number of status codes and
parameters which regulate the performance of the program as a whole.
Here \ttt{MSTU} and \ttt{PARU} are related to utility functions, as
well as a few parameters of the Standard Model, while \ttt{MSTJ} and
\ttt{PARJ} affect the underlying physics assumptions. Some of the
variables in \ttt{LUDAT1} are described elsewhere, and are therefore
here only reproduced as references to the relevant sections. This
in particular applies to many coupling constants mainly used by
{\tsc{Pythia}}, which are found just after this, in section
\ref{ss:coupcons}.
\boxsep
\iteme{MSTU(1),MSTU(2) :}\label{p:MSTU} (D=0,0) can be used to replace
the ordinary
lower and upper limits (normally 1 and \ttt{N}) for the action of
\ttt{LUROBO}, and most \ttt{LUEDIT} and \ttt{LULIST} calls. Are reset
to 0 in a \ttt{LUEXEC} call.
\iteme{MSTU(3) :} (D=0) number of lines with extra information added
after line \ttt{N}. Is reset to 0 in a \ttt{LUEXEC} call, or in an
\ttt{LUEDIT} call when particles are removed.
\iteme{MSTU(4) :} (D=4000) number of lines available in the
common block \ttt{LUJETS}. Should always be changed if the
dimensions of the \ttt{K} and \ttt{P} arrays are changed by the user,
but should otherwise never be touched. Maximum allowed value is 10000,
unless \ttt{MSTU(5)} is also changed.
\iteme{MSTU(5) :} (D=10000) is used in building up the special
colour-flow information stored in \ttt{K(I,4)} and \ttt{K(I,5)}
for \ttt{K(I,3)=} 3, 13 or 14. The generic form for \ttt{j=} 4 or 5
is \\
\ttt{K(I,j)}$ = 2 \times$\ttt{MSTU(5)}$^2 \times$MCFR$ +
$\ttt{MSTU(5)}$^2 \times$MCTO$ + $\ttt{MSTU(5)}$\times$ICFR$ + $ICTO, \\
with notation as in section \ref{ss:evrec}.
One should always have \ttt{MSTU(5)}$\geq$\ttt{MSTU(4)}. On a 32 bit
machine, values \ttt{MSTU(5)}$> 20000$ may lead to overflow problems,
and should be avoided.
\iteme{MSTU(6) :} (D=500) number of KC codes available in the
\ttt{KCHG}, \ttt{PMAS}, \ttt{MDCY}, and \ttt{CHAF} arrays; should be
changed if these dimensions are changed.
\iteme{MSTU(7) :} (D=2000) number of decay channels available in the
\ttt{MDME}, \ttt{BRAT} and \ttt{KFDP} arrays; should be changed if
these dimensions are changed.
\iteme{MSTU(10) :} (D=2) use of parton/particle masses in filling
routines (\ttt{LU1ENT}, \ttt{LU2ENT}, \ttt{LU3ENT}, \ttt{LU4ENT}).
\begin{subentry}
\iteme{= 0 :} assume the mass to be zero.
\iteme{= 1 :} keep the mass value stored in \ttt{P(I,5)}, whatever it
is. (This may be used e.g. to describe kinematics with
off-mass-shell partons).
\iteme{= 2 :} find masses according to mass tables as usual.
\end{subentry}
\iteme{MSTU(11) :} (D=6) file number to which all program output is
directed. It is your responsibility to see to it that the
corresponding file is also opened for output.
\iteme{MSTU(12) :} (D=1) writing of title page (version number and last
date of change for {\tsc{Pythia}} and {\tsc{Jetset}}) on output file.
\begin{subentry}
\iteme{= 0 :} not done.
\iteme{= 1 :} title page is written at first occasion, at which time
\ttt{MSTU(12)} is set =0.
\end{subentry}
\iteme{MSTU(13) :} (D=1) writing of information on variable values
changed by a \ttt{LUGIVE} call.
\begin{subentry}
\iteme{= 0 :} no information is provided.
\iteme{= 1 :} information is written to standard output.
\end{subentry}
\iteme{MSTU(14) :} (D=0) if non-zero, this gives the maximum flavour
for which a \ttt{LULIST(12)} call will give particle data on possible
hadrons. With \ttt{MSTU(14)=5} only known hadrons, i.e. up to bottom,
are listed. If \ttt{=0}, only separately specified particles are
listed (i.e. either KF$\leq 100$ or else both KF$> 100$ and
KC$> 100$).
\iteme{MSTU(15) :} (D=1) selection for characters used in particle
names to denote an antiparticle; appear in \ttt{LULIST} listings
or other \ttt{LUNAME} applications.
\begin{subentry}
\iteme{= 1 :} the tilde character `$\sim$'.
\iteme{= 2 :} the characters `bar'.
\end{subentry}
\iteme{MSTU(16) :} (D=1) choice of mother pointers for the particles
produced by a fragmenting parton system.
\begin{subentry}
\iteme{= 1 :} all primary particles of a system point to a line with
KF = 92 or 93, for string or independent fragmentation, respectively,
or to a line with KF = 91 if a jet system has so small a mass
that it is forced to decay into one or two particles. The two
(or more) shower initiators of a showering parton system point
to a line with KF = 94. The entries with KF = 91--94 in their
turn point back to the predecessor partons, so that the
KF = 91--94 entries form a part of the event history proper.
\iteme{= 2 :} although the lines with KF = 91--94 are present, and
contain the correct mother and daughter pointers, they are not part
of the event history proper, in that particles produced in string
fragmentation point directly to either of the two endpoint
partons of the string (depending on the side they were generated
from), particles produced in independent fragmentation point
to the respective parton they were generated from, particles in
small mass systems point to either endpoint parton, and
shower initiators point to the original on-mass-shell
counterparts. Also the daugher pointers bypass the KF = 91--94
entries. In independent fragmentation, a parton need not produce
any particles at all, and then have daughter pointers 0.
\itemc{Note :} \ttt{MSTU(16)} should not be changed between the
generation of an event and the translation of this event record with
a \ttt{LUHEPC} call, since this may give an erroneous translation of
the event history.
\end{subentry}
\iteme{MSTU(17) :} (D=0) storage option for \ttt{MSTU(90)} and
associated information on $z$ values for heavy-flavour production.
\begin{subentry}
\iteme{= 0 :} \ttt{MSTU(90)} is reset to zero at each \ttt{LUEXEC}
call. This is the appropriate course if \ttt{LUEXEC} is only called
once per event, as is normally the case when you do not
yourself call \ttt{LUEXEC}.
\iteme{= 1 :} you have to reset \ttt{MSTU(90)} to zero yourself
before each new event. This is the appropriate course if several
\ttt{LUEXEC} calls may appear for one event, i.e. if you
call \ttt{LUEXEC} directly.
\end{subentry}
\iteme{MSTU(19) :} (D=0) advisory warning for unphysical flavour
setups in \ttt{LU2ENT}, \ttt{LU3ENT} or \ttt{LU4ENT} calls.
\begin{subentry}
\iteme{= 0 :} yes.
\iteme{= 1 :} no; \ttt{MSTU(19)} is reset to 0 in such a call.
\end{subentry}
\iteme{MSTU(21) :} (D=2) check on possible errors during program
execution. Obviously no guarantee is given that all errors will be
caught, but some of the most trivial user-caused errors may be found.
\begin{subentry}
\iteme{= 0 :} errors do not cause any immediate action, rather the
program will try to cope, which may mean e.g. that it
runs into an infinite loop.
\iteme{= 1 :} parton/particle configurations are checked for possible
errors. In case of problem, an exit is made from the
misbehaving subprogram, but the generation of the event is
continued from there on. For the first \ttt{MSTU(22)} errors a
a message is printed; after that no messages appear.
\iteme{= 2 :} parton/particle configurations are checked for possible
errors. In case of problem, an exit is made from the
misbehaving subprogram, and subsequently from \ttt{LUEXEC}.
You may then choose to correct the error, and continue the
execution by another \ttt{LUEXEC} call. For the first \ttt{MSTU(22)}
errors a message is printed, after that the last event is
printed and execution is stopped.
\end{subentry}
\iteme{MSTU(22) :} (D=10) maximum number of errors that are printed.
\iteme{MSTU(23) :} (I) count of number of errors experienced to date.
\iteme{MSTU(24) :} (R) type of latest error experienced; reason that
event was not generated in full. Is reset at each \ttt{LUEXEC} call.
\begin{subentry}
\iteme{= 0 :} no error experienced.
\iteme{= 1 :} have reached end of or are writing outside
\ttt{LUJETS} memory.
\iteme{= 2 :} unknown flavour code or unphysical combination of codes;
may also be caused by erroneous string connection information.
\iteme{= 3 :} energy or mass too small or unphysical kinematical
variable setup.
\iteme{= 4 :} program is caught in an infinite loop.
\iteme{= 5 :} momentum, energy or charge was not conserved (even
allowing for machine precision errors, see \ttt{PARU(11)}); is
evaluated only after event has been generated in full, and does not
apply when independent fragmentation without momentum conservation
was used.
\iteme{= 6 :} error call from outside the fragmentation/decay package
(e.g. the $\e^+\e^-$ routines).
\iteme{= 7 :} inconsistent particle data input in \ttt{LUUPDA}
(\ttt{MUPDA = 2}) or other \ttt{LUUPDA}-related problem.
\iteme{= 8 :} problems in more peripheral service routines.
\iteme{= 9 :} various other problems.
\end{subentry}
\iteme{MSTU(25) :} (D=1) printing of warning messages.
\begin{subentry}
\iteme{= 0 :} no warnings are written.
\iteme{= 1 :} first \ttt{MSTU(26)} warnings are printed, thereafter
no warnings appear.
\end{subentry}
\iteme{MSTU(26) :} (D=10) maximum number of warnings that are printed.
\iteme{MSTU(27) :} (I) count of number of warnings experienced to
date.
\iteme{MSTU(28) :} (R) type of latest warning given, with codes
paralleling those for \ttt{MSTU(24)}, but of a less serious nature.
\iteme{MSTU(31) :} (I) number of \ttt{LUEXEC} calls in present run.
\iteme{MSTU(32) :} (I) number of entries stored with
\ttt{LUEDIT(-1)} call.
\iteme{MSTU(33) :} (I) if set 1 before a \ttt{LUDBRB} call, the
\ttt{V} vectors (in the particle range to be rotated/boosted) are
set 0 before the rotation/boost. \ttt{MSTU(33)} is set back to 0 in
the \ttt{LUDBRB} call. Is inactive in a \ttt{LUROBO} call.
\iteme{MSTU(41) - MSTU(63) :} switches for event-analysis routines,
see section \ref{ss:evanrout}.
\iteme{MSTU(70) :} (D=0) the number of lines consisting only of equal
signs (\ttt{======}) that are inserted in the event listing obtained
with \ttt{LULIST(1)}, \ttt{LULIST(2)} or \ttt{LULIST(3)}, so as to
distinguish different sections of the event record on output. At most
10 such lines can be inserted; see \ttt{MSTU(71) - MSTU(80)}. Is reset
at \ttt{LUEDIT} calls with arguments 0--5.
\iteme{MSTU(71) - MSTU(80) :} line numbers below which lines
consisting only of equal signs (\ttt{======}) are inserted in event
listings. Only the first \ttt{MSTU(70)} of the 10 allowed positions
are enabled.
\iteme{MSTU(90) :} number of heavy-flavour hadrons (i.e. hadrons
containing charm or heavier flavours) produced in current event, for
which the positions in the event record are stored in
\ttt{MSTU(91) - MSTU(98)} and the $z$ values in the fragmentation in
\ttt{PARU(91) - PARU(98)}. At most eight values will be stored
(normally this is no problem). No $z$ values can be stored for those
heavy hadrons produced when a string has so small mass that it
collapses to one or two particles, nor for those produced as one of
the final two particles in the fragmentation of a string. If
\ttt{MSTU(17)=1}, \ttt{MSTU(90)} should be reset to zero by you
before each new event, else this is done automatically.
\iteme{MSTU(91) - MSTU(98) :} the first \ttt{MSTU(90)} positions will
be filled with the line numbers of the heavy-flavour hadrons produced
in the current event. See \ttt{MSTU(90)} for additional comments. Note
that the information is corrupted by calls to \ttt{LUEDIT} with options
0--5 and 21--23; calls with options 11--15 work, however.
\iteme{MSTU(101) - MSTU(118) :} switches related to couplings, see
section \ref{ss:coupcons}.
\iteme{MSTU(161), MSTU(162) :} information used by event-analysis
routines, see section \ref{ss:evanrout}.
\iteme{MSTU(181) :} (R) {\tsc{Jetset}} version number.
\iteme{MSTU(182) :} (R) {\tsc{Jetset}} subversion number.
\iteme{MSTU(183) :} (R) last year of change for {\tsc{Jetset}}.
\iteme{MSTU(184) :} (R) last month of change for {\tsc{Jetset}}.
\iteme{MSTU(185) :} (R) last day of change for {\tsc{Jetset}}.
\iteme{MSTU(186) :} (R) earliest subversion of {\tsc{Pythia}} version 5
with which this {\tsc{Jetset}} subversion can be run.
\boxsep
\iteme{PARU(1) :}\label{p:PARU} (R) $\pi \approx 3.1415927$.
\iteme{PARU(2) :} (R) $2\pi \approx 6.2831854$.
\iteme{PARU(3) :} (D=0.1973) conversion factor for GeV$^{-1} \to$ fm
or fm$^{-1} \to$ GeV.
\iteme{PARU(4) :} (D=5.068) conversion factor for fm $\to$ GeV$^{-1}$
or GeV $\to$ fm$^{-1}$.
\iteme{PARU(5) :} (D=0.3894) conversion factor for GeV$^{-2} \to$ mb
or mb$^{-1} \to$ GeV$^2$.
\iteme{PARU(6) :} (D=2.568) conversion factor for mb $\to$ GeV$^{-2}$
or GeV$^2 \to$ mb$^{-1}$.
\iteme{PARU(11) :} (D=0.001) relative error, i.e. nonconservation of
momentum and energy divided by total energy, that may be attributable
to machine precision problems before a physics error is suspected
(see \ttt{MSTU(24)=5}).
\iteme{PARU(12) :} (D=0.09 GeV$^2$) effective cut-off in squared mass,
below which partons may be recombined to simplify (machine precision
limited) kinematics of string fragmentation.
\iteme{PARU(13) :} (D=0.01) effective angular cut-off in radians for
recombination of partons, used in conjunction with \ttt{PARU(12)}.
\iteme{PARU(21) :} (I) contains the total energy $W$ of all first
generation jets/particles after a \ttt{LUEXEC} call; to be used by the
\ttt{PLU} function for \ttt{I>0}, \ttt{J=} 20--25.
\iteme{PARU(41) - PARU(63) :} parameters for event-analysis routines,
see section \ref{ss:evanrout}.
\iteme{PARU(91) - PARU(98) :} the first \ttt{MSTU(90)} positions will
be filled with the fragmentation $z$ values used internally in the
generation of heavy-flavour hadrons --- how these are translated into
the actual energies and momenta of the observed hadrons is a
complicated function of the string configuration. The particle with
$z$ value stored in \ttt{PARU(i)} is to be found in line \ttt{MSTU(i)}
of the event record. See \ttt{MSTU(90)} and \ttt{MSTU(91) - MSTU(98)}
for additional comments.
\iteme{PARU(101) - PARU(195) :} various coupling constants and
parameters related to couplings, see section \ref{ss:coupcons}.
\boxsep
\iteme{MSTJ(1) :}\label{p:MSTJ} (D=1) choice of fragmentation scheme.
\begin{subentry}
\iteme{= 0 :} no jet fragmentation at all.
\iteme{= 1 :} string fragmentation according to the Lund model.
\iteme{= 2 :} independent fragmentation, according to specification
in \ttt{MSTJ(2)} and \ttt{MSTJ(3)}.
\end{subentry}
\iteme{MSTJ(2) :} (D=3) gluon jet fragmentation scheme in independent
fragmentation.
\begin{subentry}
\iteme{= 1 :} a gluon is assumed to fragment like a random $\d$, $\u$
or $\mathrm{s}$ quark or antiquark.
\iteme{= 2 :} as \ttt{=1}, but longitudinal (see \ttt{PARJ(43)},
\ttt{PARJ(44)} and \ttt{PARJ(59)}) and transverse (see \ttt{PARJ(22)})
momentum properties of quark or antiquark substituting for gluon may
be separately specified.
\iteme{= 3 :} a gluon is assumed to fragment like a pair of a
$\d$, $\u$ or $\mathrm{s}$ quark and its antiquark, sharing the gluon energy
according to the Altarelli-Parisi splitting function.
\iteme{= 4 :} as \ttt{=3}, but longitudinal (see \ttt{PARJ(43)},
\ttt{PARJ(44)} and \ttt{PARJ(59)}) and transverse (see \ttt{PARJ(22)})
momentum properties of quark and antiquark substituting for gluon may
be separately specified.
\end{subentry}
\iteme{MSTJ(3) :} (D=0) energy, momentum and flavour conservation
options in independent fragmentation. Whenever momentum conservation
is described below, energy and flavour conservation is also
implicitly assumed.
\begin{subentry}
\iteme{= 0 :} no explicit conservation of any kind.
\iteme{= 1 :} particles share momentum imbalance compensation according
to their energy (roughly equivalent to boosting event to c.m.
frame). This is similar to the approach in the Ali et al.
program \cite{Ali80}.
\iteme{= 2 :} particles share momentum imbalance compensation according
to their longitudinal mass with respect to the imbalance
direction.
\iteme{= 3 :} particles share momentum imbalance compensation equally.
\iteme{= 4 :} transverse momenta are compensated separately within
each jet, longitudinal momenta are rescaled so that ratio
of final jet to initial parton momentum is the same for
all the jets of the event. This is similar to the approach in
the Hoyer et al. program \cite{Hoy79}.
\iteme{= 5 :} only flavour is explicitly conserved.
\iteme{= 6 - 10 :} as \ttt{=1 - 5}, except that above several colour
singlet systems that followed immediately after each other
in the event listing (e.g. $\mathrm{q}\overline{\mathrm{q}}\mathrm{q}\overline{\mathrm{q}}$) were treated as
one single system, whereas here they are treated as
separate systems.
\iteme{= -1 :} independent fragmentation, where also particles moving
backwards with respect to the jet direction are kept, and
thus the amount of energy and momentum mismatch may be large.
\end{subentry}
\iteme{MSTJ(11) :} (D=4) choice of longitudinal fragmentation
function, i.e. how large a fraction of the energy available a
newly-created hadron takes.
\begin{subentry}
\iteme{= 1 :} the Lund symmetric fragmentation function, see
\ttt{PARJ(41) - PARJ(45)}.
\iteme{= 2 :} choice of some different forms for each flavour
separately, see \ttt{PARJ(51) - PARJ(59)}.
\iteme{= 3 :} hybrid scheme, where light flavours are treated with
symmetric Lund (\ttt{=1}), but charm and heavier can be separately
chosen, e.g. according to the SLAC function (\ttt{=2}).
\iteme{= 4 :} the Lund symmetric fragmentation function (\ttt{=1}),
for heavy endpoint quarks modified according to the Bowler
(Artru--Mennessier, Morris) space--time picture of string evolution,
see \ttt{PARJ(46)}.
\iteme{= 5 :} as \ttt{=4}, but with possibility to interpolate
between Bowler and Lund separately for $\c$, $\b$ and $\t$; see
\ttt{PARJ(46) - PARJ(48)}.
\end{subentry}
\iteme{MSTJ(12) :} (D=2) choice of baryon production model.
\begin{subentry}
\iteme{= 0 :} no baryon-antibaryon pair production at all; initial
diquark treated as a unit.
\iteme{= 1 :} diquark-antidiquark pair production allowed; diquark
treated as a unit.
\iteme{= 2 :} diquark-antidiquark pair production allowed, with
possibility for diquark to be split according to the `popcorn'
scheme.
\iteme{= 3 :} as \ttt{=2}, but additionally the production of first
rank baryons may be suppressed by a factor \ttt{PARJ(19)}.
\end{subentry}
\iteme{MSTJ(13) :} (D=0) generation of transverse momentum for
endpoint quark(s) of single quark jet or $\mathrm{q}\overline{\mathrm{q}}$ jet system (in
multijet events no endpoint transverse momentum is ever allowed for).
\begin{subentry}
\iteme{= 0 :} no transverse momentum for endpoint quarks.
\iteme{= 1 :} endpoint quarks obtain transverse momenta like ordinary
$\mathrm{q}\overline{\mathrm{q}}$ pairs produced in the field (see \ttt{PARJ(21)}); for
2-jet systems the endpoints obtain balancing transverse momenta.
\end{subentry}
\iteme{MSTJ(14) :} (D=1) treatment of a colour-singlet jet system
with a low invariant mass.
\begin{subentry}
\iteme{= 0 :} no precautions are taken, meaning that problems may
occur in \ttt{LUSTRF} (or \ttt{LUINDF}) later on.
\iteme{= 1 :} small jet systems are allowed to collapse into two
particles or, failing that, one single particle. Normally
all small systems are treated this way, starting with the
smallest one, but some systems would require more work and
are left untreated; they include diquark-antidiquark pairs
below the two-particle threshold.
\iteme{= -1 :} special option for \ttt{LUPREP} calls, where no
precautions are taken (as for \ttt{=0}), but, in addition, no checks
are made on the presence of small-mass systems; i.e. \ttt{LUPREP}
only rearranges colour strings.
\end{subentry}
\iteme{MSTJ(15) :} (D=0) production probability for new flavours.
\begin{subentry}
\iteme{= 0 :} according to standard Lund parametrization, as given
by \ttt{PARJ(1) - PARJ(20)}.
\iteme{= 1 :} according to probabilities stored in
\ttt{PARF(201) - PARF(1960)}; note that no default values exist here,
i.e. \ttt{PARF} must be set by you. The \ttt{MSTJ(12)} switch
can still be used to set baryon production mode, with the
modification that \ttt{MSTJ(12)=2} here allows an arbitrary number
of mesons to be produced between a baryon and an antibaryon (since
the probability for diquark $\to$ meson $+$ new diquark is assumed
independent of prehistory).
\end{subentry}
\iteme{MSTJ(21) :} (D=2) form of particle decays.
\begin{subentry}
\iteme{= 0 :} all particle decays are inhibited.
\iteme{= 1 :} a particle declared unstable in the \ttt{MDCY} vector,
and with decay channels defined, may decay within the region given
by \ttt{MSTJ(22)}. A particle may decay into jets, which then fragment
further according to the \ttt{MSTJ(1)} value.
\iteme{= 2 :} as \ttt{=1}, except that a $\mathrm{q}\overline{\mathrm{q}}$ jet system produced
in a decay (e.g. of a $\mathrm{B}$ meson) is always allowed to fragment
according to string fragmentation, rather than according to the
\ttt{MSTJ(1)} value (this means that momentum, energy and charge
are conserved in the decay).
\end{subentry}
\iteme{MSTJ(22) :} (D=1) cut-off on decay length for a particle that is
allowed to decay according to \ttt{MSTJ(21)} and the \ttt{MDCY} value.
\begin{subentry}
\iteme{= 1 :} a particle declared unstable is also forced to decay.
\iteme{= 2 :} a particle is decayed only if its average proper
lifetime is larger than \ttt{PARJ(71)}.
\iteme{= 3 :} a particle is decayed only if the decay vertex is
within a distance \ttt{PARJ(72)} of the origin.
\iteme{= 4 :} a particle is decayed only if the decay vertex is
within a cylindrical volume with radius \ttt{PARJ(73)} in the
$xy$-plane and extent to $\pm$\ttt{PARJ(74)} in the $z$ direction.
\end{subentry}
\iteme{MSTJ(23) :} (D=1) possibility of having a shower evolving from
a $\mathrm{q}\overline{\mathrm{q}}$ pair created as decay products.
\begin{subentry}
\iteme{= 0 :} never.
\iteme{= 1 :} whenever the decay channel matrix-element code is
\ttt{MDME(IDC,2)=} 4, 32, 33, 44 or 46, the two first decay products
(if they are partons) are allowed to shower, like a colour-singlet
subsystem, with maximum virtuality given by the invariant mass
of the pair.
\end{subentry}
\iteme{MSTJ(24) :} (D=2) particle masses.
\begin{subentry}
\iteme{= 0 :} discrete mass values are used.
\iteme{= 1 :} particles registered as having a mass width in the
\ttt{PMAS} vector are given a mass according to a truncated
Breit--Wigner shape, linear in $m$, eq.~(\ref{dec:BWlin}).
\iteme{= 2 :} as \ttt{=1}, but gauge bosons (actually all particles
with $|$KF$| \leq 100$) are distributed according to a Breit--Wigner
quadratic in $m$, as obtained from propagators.
\iteme{= 3 :} as \ttt{=1}, but Breit--Wigner shape is always
quadratic in $m$, eq.~(\ref{dec:BWtwo}).
\end{subentry}
\iteme{MSTJ(25) :} (D=1) inclusion of the $\mathrm{W}^{\pm}$ propagator, in
addition to the standard, `infinitely heavy' weak $V-A$ matrix
element, in the decay of a $\t$, $\mathrm{l}$ or $\mathrm{h}$ quark, or
$\chi$ lepton.
\begin{subentry}
\iteme{= 0 :} not included.
\iteme{= 1 :} included.
\end{subentry}
\iteme{MSTJ(26) :} (D=2) inclusion of $\mathrm{B}$--$\br{\mathrm{B}}$ mixing in
decays.
\begin{subentry}
\iteme{= 0 :} no.
\iteme{= 1 :} yes, with mixing parameters given by \ttt{PARJ(76)}
and \ttt{PARJ(77)}. Mixing decays are not specially marked.
\iteme{= 2 :} yes, as \ttt{=1}, but a $\mathrm{B}$ ($\br{\mathrm{B}}$) that decays
as a $\br{\mathrm{B}}$ ($\mathrm{B}$) is marked as \ttt{K(I,1)=12} rather than the
normal \ttt{K(I,1)=11}.
\end{subentry}
\iteme{MSTJ(27) :} (D=2) possibility for the $\b$ quark to develop
a shower in the decay of a top hadron, i.e. $\mathrm{T} \to \mathrm{W}^+ \b \overline{\mathrm{q}}$,
where $\overline{\mathrm{q}}$ is a spectator quark.
\begin{subentry}
\iteme{= 0 :} no, i.e. $\b$ jet is narrow, low-multiplicity.
\iteme{= 1 :} the $\b$ is allowed to shower and the $\mathrm{W}$ momentum
(in the rest frame of the $\mathrm{T}$) is reduced acccordingly. The $\mathrm{W}$
is therafter assumed to decay isotropically.
\iteme{= 2 :} the $\b$ is allowed to shower, as in \ttt{=1}, but
the $\mathrm{W}$ decays anisotropically, with the same polarization as in
the standard weak decay of option \ttt{=0}. In principle this is
better than option \ttt{=1}, but in practice there is no big
difference.
\end{subentry}
\iteme{MSTJ(28) :} (D=0) call to an external $\tau$ decay library.
For this option to be meaningful, it is up to you to write the
appropriate interface and include that in the routine \ttt{LUTAUD},
as explained in section \ref{ss:JETphysrout}.
\begin{subentry}
\iteme{= 0 :} not done, i.e. the internal \ttt{LUDECY} treatment is
used.
\iteme{= 1 :} done whenever the $\tau$ mother particle species can
be identified, else the internal \ttt{LUDECY} treatment is used.
Normally the mother particle should always be identified, but it is
possible for a user to remove event history information or to add
extra $\tau$'s directly to the event record, and then the mother is
not known.
\iteme{= 2 :} always done.
\end{subentry}
\iteme{MSTJ(40) - MSTJ(50) :} switches for time-like parton showers,
see section \ref{ss:showrout}.
\iteme{MSTJ(51) :} (D=0) inclusion of Bose--Einstein effects.
\begin{subentry}
\iteme{= 0 :} no effects included.
\iteme{= 1 :} effects included according to an exponential
parametrization
$C_2(Q) = 1 + $\ttt{PARJ(92)}$\times \exp(-Q/$\ttt{PARJ(93)}$)$,
where $C_2(Q)$ represents the ratio of particle production at
$Q$ with Bose--Einstein effects to that without, and the relative
momentum $Q$ is defined by
$Q^2(p_1,p_2) = -(p_1 - p_2)^2 = (p_1 + p_2)^2 - 4m^2$. Particles
with width broader than \ttt{PARJ(91)} are assumed to have time to
decay before Bose--Einstein effects are to be considered.
\iteme{= 2 :} effects included according to a Gaussian
parametrization
$C_2(Q) = 1 + $\ttt{PARJ(92)}$\times \exp(- (Q/$\ttt{PARJ(93)}$)^2 )$,
with notation and comments as above.
\end{subentry}
\iteme{MSTJ(52) :} (D=3) number of particle species for which
Bose--Einstein correlations are to be included, ranged along the
chain $\pi^+$, $\pi^-$, $\pi^0$, $\mathrm{K}^+$, $\mathrm{K}^-$, $\mathrm{K}^0_{\mrm{L}}$,
$\mathrm{K}^0_{\mrm{S}}$, $\eta$ and $\eta'$. Default corresponds to
including all pions ($\pi^+$, $\pi^-$, $\pi^0$), 7 to including all
Kaons as well, and 9 is maximum.
\iteme{MSTJ(91) :} (I) flag when generating gluon jet with options
\ttt{MSTJ(2)=} 2 or 4 (then \ttt{=1}, else \ttt{=0}).
\iteme{MSTJ(92) :} (I) flag that a $\mathrm{q}\overline{\mathrm{q}}$ or $\mathrm{g}\g$ pair or a
$\mathrm{g}\g\mathrm{g}$ triplet created in \ttt{LUDECY} should be allowed to shower,
is 0 if no pair or triplet, is the entry number of the first parton
if a pair indeed exists, is the entry number of the first parton,
with a $-$ sign, if a triplet indeed exists.
\iteme{MSTJ(93) :} (I) switch for \ttt{ULMASS} action. Is reset to 0
in \ttt{ULMASS} call.
\begin{subentry}
\iteme{= 0 :} ordinary action.
\iteme{= 1 :} light ($\d$, $\u$, $\mathrm{s}$, $\c$, $\b$) quark masses are
taken from \ttt{PARF(101) - PARF(105)} rather than
\ttt{PMAS(1,1) - PMAS(5,1)}. Diquark masses are
given as sum of quark masses, without spin splitting term.
\iteme{= 2 :} as \ttt{=1}. Additionally the constant terms
\ttt{PARF(121)} and \ttt{PARF(122)} are subtracted from quark and
diquark masses, respectively.
\end{subentry}
\iteme{MSTJ(101) - MSTJ(121) :} switches for $\e^+\e^-$ event generation,
see section \ref{ss:eeroutines}.
\boxsep
\iteme{PARJ(1) :}\label{p:PARJ} (D=0.10) is
${\cal P}(\mathrm{q}\q)/{\cal P}(\mathrm{q})$, the
suppression of diquark-antidiquark pair production in the colour
field, compared with quark--antiquark production.
\iteme{PARJ(2) :} (D=0.30) is ${\cal P}(\mathrm{s})/{\cal P}(u)$, the
suppression of $\mathrm{s}$ quark pair production in the field compared with
$\u$ or $\d$ pair production.
\iteme{PARJ(3) :} (D=0.4) is
$({\cal P}(\u\mathrm{s})/{\cal P}(\u\d))/({\cal P}(\mathrm{s})/{\cal P}(\d))$,
the extra suppression of strange diquark production compared with
the normal suppression of strange quarks.
\iteme{PARJ(4) :} (D=0.05) is $(1/3){\cal P}(\u\d_1)/{\cal P}(\u\d_0)$,
the suppression of spin 1 diquarks compared with spin 0 ones
(excluding the factor 3 coming from spin counting).
\iteme{PARJ(5) :} (D=0.5) parameter determining relative occurence of
baryon production by $BM\br{B}$ and by $B\br{B}$ configurations in the
popcorn baryon production model, roughly
${\cal P}(BM\br{B})/({\cal P}(B\br{B})+{\cal P}(BM\br{B})) =$
\ttt{PARJ(5)}$/(0.5+$\ttt{PARJ(5)}$)$.
\iteme{PARJ(6) :} (D=0.5) extra suppression for having a $\mathrm{s}\overline{\mathrm{s}}$
pair shared by the $B$ and $\br{B}$ of a $BM\br{B}$ situation.
\iteme{PARJ(7) :} (D=0.5) extra suppression for having a strange
meson $M$ in a $BM\br{B}$ configuration.
\iteme{PARJ(11) - PARJ(17) :} parameters that determine the spin of
mesons.
\begin{subentry}
\iteme{PARJ(11) :} (D=0.5) is the probability that a light meson
(containing $\u$ and $\d$ quarks only) has spin 1 (with
\ttt{1-PARJ(11)} the probability for spin 0) when formed in
fragmentation.
\iteme{PARJ(12) :} (D=0.6) is the probability that a strange meson
has spin 1.
\iteme{PARJ(13) :} (D=0.75) is the probability that a charm or
heavier meson has spin 1.
\iteme{PARJ(14) :} (D=0.) is the probability that a spin = 0
meson is produced with an orbital angular momentum 1, for a
total spin = 1.
\iteme{PARJ(15) :} (D=0.) is the probability that a spin = 1
meson is produced with an orbital angular momentum 1, for a
total spin = 0.
\iteme{PARJ(16) :} (D=0.) is the probability that a spin = 1
meson is produced with an orbital angular momentum 1, for a
total spin = 1.
\iteme{PARJ(17) :} (D=0.) is the probability that a spin = 1
meson is produced with an orbital angular momentum 1, for a
total spin = 2.
\itemc{Note :} the end result of the numbers above is
that, with \ttt{i} = 11, 12 or 13, depending on flavour content, \\
${\cal P}(S = 0, L = 0, J = 0) = (1 - \mtt{PARJ(i)}) \times
(1 - \mtt{PARJ(14)})$, \\
${\cal P}(S = 0, L = 1, J = 1) = (1 - \mtt{PARJ(i)}) \times
\mtt{PARJ(14)}$, \\
${\cal P}(S = 1, L = 0, J = 1) = $ \\
\mbox{~}\hfill$\mtt{PARJ(i)} \times (1 - \mtt{PARJ(15)} -
\mtt{PARJ(16)} - \mtt{PARJ(17)})$, \\
${\cal P}(S = 1, L = 1, J = 0) = \mtt{PARJ(i)} \times
\mtt{PARJ(15)}$, \\
${\cal P}(S = 1, L = 1, J = 1) = \mtt{PARJ(i)} \times
\mtt{PARJ(16)}$, \\
${\cal P}(S = 1, L = 1, J = 2) = \mtt{PARJ(i)} \times
\mtt{PARJ(17)}$, \\
where $S$ is the quark `true' spin and $J$ is the total spin, usually
called the spin $s$ of the meson.
\end{subentry}
\iteme{PARJ(18) :} (D=1.) is an extra suppression factor multiplying
the ordinary {\bf SU(6)} weight for spin $3$/2 baryons, and hence a
means to break {\bf SU(6)} in addition to the dynamic breaking implied
by \ttt{PARJ(2)}, \ttt{PARJ(3)}, \ttt{PARJ(4)}, \ttt{PARJ(6)} and
\ttt{PARJ(7)}.
\iteme{PARJ(19) :} (D=1.) extra baryon suppression factor, which
multiplies the ordinary diquark-antidiquark production probability
for the breakup closest to the endpoint of a string, but leaves other
breaks unaffected. Is only used for \ttt{MSTJ(12)=3}.
\iteme{PARJ(21) :} (D=0.36 GeV) corresponds to the width $\sigma$ in
the Gaussian $p_x$ and $p_y$ transverse momentum distributions for
primary hadrons. See also \ttt{PARJ(22) - PARJ(24)}.
\iteme{PARJ(22) :} (D=1.) relative increase in transverse momentum in
a gluon jet generated with \ttt{MSTJ(2)=} 2 or 4.
\iteme{PARJ(23), PARJ(24) :} (D=0.01, 2.) a fraction \ttt{PARJ(23)}
of the Gaussian transverse momentum distribution is taken to be a
factor \ttt{PARJ(24)} larger than input in \ttt{PARJ(21)}. This
gives a simple parametrization of non-Gaussian tails to the Gaussian
shape assumed above.
\iteme{PARJ(25) :} (D=1.) extra suppression factor for $\eta$
production in fragmentation; if an $\eta$ is rejected a new flavour
pair is generated and a new hadron formed.
\iteme{PARJ(26) :} (D=0.4) extra suppression factor for $\eta'$
production in fragmentation; if an $\eta'$ is rejected a new flavour
pair is generated and a new hadron formed.
\iteme{PARJ(31) :} (D=0.1 GeV) gives the remaining $W_+$ below which
the generation of a single jet is stopped (it is chosen smaller than a
pion mass, so that no hadrons moving in the forward direction are
missed).
\iteme{PARJ(32) :} (D=1. GeV) is, with quark masses added, used to
define the minimum allowable energy of a colour-singlet jet system.
\iteme{PARJ(33) - PARJ(34) :} (D=0.8 GeV, 1.5 GeV) are, together with
quark masses, used to define the remaining energy below which the
fragmentation of a jet system is stopped and two final hadrons
formed. \ttt{PARJ(33)} is normally used, except for \ttt{MSTJ(11)=2},
when \ttt{PARJ(34)} is used.
\iteme{PARJ(36) :} (D=2.) represents the dependence on the mass of the
final quark pair for defining the stopping point of the fragmentation.
Is strongly correlated to the choice of \ttt{PARJ(33) - PARJ(35)}.
\iteme{PARJ(37) :} (D=0.2) relative width of the smearing of the
stopping point energy.
\iteme{PARJ(38) - PARJ(39) :} (D=2.5, 0.6) refers to the probability
for reverse rapidity ordering of the final two hadrons,
according to eq.~(\ref{fr:revord}), where
$d_0 = $\ttt{PARJ(39)} for \ttt{MSTJ(11)}$\neq 2$, and
$d = $\ttt{PARJ(39)} for \ttt{MSTJ(11)=2}.
\iteme{PARJ(41), PARJ(42) :} (D=0.3, 0.58 GeV$^{-2}$) give the $a$ and
$b$ parameters of the symmetric Lund fragmentation function for
\ttt{MSTJ(11)=}1, 4 and 5 (and \ttt{MSTJ(11)=3} for ordinary hadrons).
\iteme{PARJ(43), PARJ(44) :} (D=0.5, 0.9 GeV$^{-2}$) give the $a$ and
$b$ parameters as above for the special case of a gluon jet generated
with IF and \ttt{MSTJ(2)=} 2 or 4.
\iteme{PARJ(45) :} (D=0.5) the amount by which the effective $a$
parameter in the Lund flavour dependent symmetric fragmentation
function is assumed to be larger than the normal $a$ when diquarks
are produced. More specifically, referring to eq.~(\ref{fr:LSFFlong}),
$a_{\alpha} = $\ttt{PARJ(41)} when considering the
fragmentation of a quark and = \ttt{PARJ(41) + PARJ(45)} for
the fragmentation of a diquark, with corresponding expression for
$a_{\beta}$ depending on whether the newly created object is a quark
or diquark (for an independent gluon jet generated with
\ttt{MSTJ(2)=} 2 or 4, replace \ttt{PARJ(41)} by \ttt{PARJ(43)}).
In the popcorn model, a meson created in between the baryon and
antibaryon has $a_{\alpha} = a_{\beta} = $\ttt{PARJ(41) + PARJ(45)}.
\iteme{PARJ(46) - PARJ(48) :} (D=3*1.) modification of the Lund
symmetric fragmentation for heavy endpoint quarks according to the
recipe by Bowler, available when
\ttt{MSTJ(11)=} 4 or 5 is selected. The shape is given
by eq.~(\ref{fr:LSFFBowler}). If \ttt{MSTJ(11)=4} then
$r_{\mathrm{Q}} = $\ttt{PARJ(46)} for all flavours, while if
\ttt{MSTJ(11)=5} then $r_{\c} = $\ttt{PARJ(46)},
$r_{\b} = $\ttt{PARJ(47)} and $r_{\mathrm{Q}} = $\ttt{PARJ(48)} for
$\t$ and heavier. \ttt{PARJ(46) - PARJ(48)} thus provide a
possibility to interpolate between the `pure' Bowler shape,
$r = 1$, and the normal Lund one, $r = 0$. The additional
modifications made in \ttt{PARJ(43) - PARJ(45)}
are automatically taken into account, if necessary.
\iteme{PARJ(51) - PARJ(58) :} (D=3*0.77, $-0.05$, $-0.005$,
3*$-0.00001$) give a choice of four possible ways to parametrize the
fragmentation function for \ttt{MSTJ(11)=2} (and \ttt{MSTJ(11)=3}
for charm and heavier). The fragmentation of each flavour KF may
be chosen separately; for a diquark the flavour of the heaviest
quark is used. With $c = $\ttt{PARJ(50+KF)}, the parametrizations
are: \\
$0 \leq c \leq 1$ : Field-Feynman, $f(z) = 1 - c + 3c(1-z)^2$; \\
$-1 \leq c < 0$ : SLAC, $f(z) = 1/(z(1-1/z-(-c)/(1-z))^2)$; \\
$c > 1$ : power peaked at $z=0$, $f(z) = (1-z)^{c-1}$; \\
$c < -1$ : power peaked at $z=1$, $f(z) = z^{-c-1}$.
\iteme{PARJ(59) :} (D=1.) replaces \ttt{PARJ(51) - PARJ(53)} for
gluon jet generated with \ttt{MSTJ(2)=} 2 or 4.
\iteme{PARJ(61) - PARJ(63) :} (D=4.5, 0.7, 0.) parametrizes the
energy dependence of the primary multiplicity distribution in
phase-space decays. The former two correspond to $c_1$ and $c_2$
of eq.~(\ref{dec:multsel}), while the latter allows a further
additive term in the multiplicity specifically for onium decays.
\iteme{PARJ(64) :} (0.003 GeV) minimum kinetic energy in decays
(safety margin for numerical precision errors).
\iteme{PARJ(65) :} (D=0.5 GeV) mass which, in addition to the
spectator quark ordiquark mass, is not assumed to partake in the
weak decay of a heavy quark in a hadron.
\iteme{PARJ(66) :} (D=0.5) relative probability that colour is
rearranged when two singlets are to be formed from decay products.
Only applies for \ttt{MDME(IDC,2)=} 11--30, i.e. low-mass
phase-space decays.
\iteme{PARJ(71) :} (D=10 mm) maximum average proper lifetime for
particles allowed to decay in the \ttt{MSTJ(22)=2} option. With
the default value, $\mathrm{K}_{\mrm{S}}^0$, $\Lambda$, $\Sigma^-$, $\Sigma^+$,
$\Xi^-$, $\Xi^0$ and $\Omega^-$ are stable (in addition to those
normally taken to be stable), but charm and bottom do still decay.
\iteme{PARJ(72) :} (D=1000 mm) maximum distance from the origin at
which a decay is allowed to take place in the \ttt{MSTJ(22)=3}
option.
\iteme{PARJ(73) :} (D=100 mm) maximum cylindrical distance
$\rho = \sqrt{x^2 + y^2}$ from the origin at which a decay is allowed
to take place in the \ttt{MSTJ(22)=4} option.
\iteme{PARJ(74) :} (D=1000 mm) maximum z distance from the origin
at which a decay is allowed to take place in the \ttt{MSTJ(22)=4}
option.
\iteme{PARJ(76) :} (D=0.7) mixing parameter $x_d = \Delta M/\Gamma$
in $\mathrm{B}^0$--$\br{\mathrm{B}}^0$ system.
\iteme{PARJ(77) :} (D=10.) mixing parameter $x_s = \Delta M/\Gamma$
in $\mathrm{B}_s^0$--$\br{\mathrm{B}}_s^0$ system.
\iteme{PARJ(81) - PARJ(89) :} parameters for time-like parton showers,
see section \ref{ss:showrout}.
\iteme{PARJ(91) :} (D=0.020 GeV) minimum particle width in
\ttt{PMAS(KC,2)}, above which particle decays are assumed to take
place before the stage where Bose--Einstein effects are introduced.
\iteme{PARJ(92) :} (D=1.) nominal strength of Bose--Einstein effects
for $Q = 0$, see \ttt{MSTJ(51)}. This parameter, often denoted
$\lambda$, expresses the amount of incoherence in particle
production. Due to the simplified picture used for the
Bose--Einstein effects, in particular for effects from three
nearby identical particles, the actual $\lambda$
of the simulated events may be larger than the input value.
\iteme{PARJ(93) :} (D=0.20 GeV) size of the Bose--Einstein effect
region in terms of the $Q$ variable, see \ttt{MSTJ(51)}. The more
conventional measure, in terms of the radius $R$ of the production
volume, is given by
$R = \hbar/$\ttt{PARJ(93)}$ \approx 0.2$
fm$\times$GeV/\ttt{PARJ(93)}$ = $\ttt{PARU(3)}/\ttt{PARJ(93)}.
\iteme{PARJ(121) - PARJ(171) :} parameters for $\e^+\e^-$ event generation,
see section \ref{ss:eeroutines}.
\end{entry}
\subsection{Couplings}
\label{ss:coupcons}
In this section we collect information on the two routines for
running $\alpha_{\mathrm{s}}$ and $\alpha_{\mathrm{em}}$, and on other couplings
of standard and non-standard particles. Although originally begun
for {\tsc{Jetset}} applications, this section has rapidly expanded towards
the non-standard aspects, and is thus more of interest for {\tsc{Pythia}}
applications than for {\tsc{Jetset}} itself. It could therefore equally
well have been put somewhere else in this manual. A few couplings
indeed appear in the \ttt{PARP} array, see section
\ref{ss:PYswitchpar}.
\drawbox{ALEM = ULALEM(Q2)}\label{p:ULALEM}
\begin{entry}
\itemc{Purpose:} to calculate the running electromagnetic coupling
constant $\alpha_{\mathrm{em}}$. Expressions used are described in
ref. \cite{Kle89}. See \ttt{MSTU(101)}, \ttt{PARU(101)},
\ttt{PARU(103)} and \ttt{PARU(104)}.
\iteme{Q2 :} the momentum transfer scale $Q^2$ at which to evaluate
$\alpha_{\mathrm{em}}$.
\end{entry}
\drawbox{ALPS = ULALPS(Q2)}\label{p:ULALPS}
\begin{entry}
\itemc{Purpose:} to calculate the running strong coupling constant
$\alpha_{\mathrm{s}}$. The first- and second-order expressions are given by
eqs.~(\ref{ee:aS3j}) and (\ref{ee:aS4j}). See
\ttt{MSTU(111) - MSTU(118)}
and \ttt{PARU(111) - PARU(118)} for options.
\iteme{Q2 :} the momentum transfer scale $Q^2$ at which to evaluate
$\alpha_{\mathrm{s}}$.
\end{entry}
\drawbox{COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)}
\begin{entry}
\itemc{Purpose:} to give access to a number of status codes and
parameters which regulate the performance of the program as a whole.
Here only those related to couplings are described; the main
description is found in section \ref{ss:JETswitch}.
\boxsep
\iteme{MSTU(101) :}\label{p:MSTU101} (D=1) procedure for
$\alpha_{\mathrm{em}}$ evaluation in the \ttt{ULALEM} function.
\begin{subentry}
\iteme{= 0 :} $\alpha_{\mathrm{em}}$ is taken fixed at the value
\ttt{PARU(101)}.
\iteme{= 1 :} $\alpha_{\mathrm{em}}$ is running with the $Q^2$ scale,
taking into account corrections from fermion loops ($\mathrm{e}$, $\mu$,
$\tau$, $\d$, $\u$, $\mathrm{s}$, $\c$, $\b$).
\iteme{= 2 :} $\alpha_{\mathrm{em}}$ is fixed, but with separate values at low
and high $Q^2$. For $Q^2$ below (above) \ttt{PARU(104)} the value
\ttt{PARU(101)} (\ttt{PARU(103)}) is used. The former value is
then intended for real photon emission, the latter for
electroweak physics, e.g. of the $\mathrm{W}/\mathrm{Z}$ gauge bosons.
\end{subentry}
\iteme{MSTU(111) :} (D=1) order of $\alpha_{\mathrm{s}}$ evaluation in the
\ttt{ULALPS} function. Is overwritten in \ttt{LUEEVT}, \ttt{LUONIA} or
\ttt{PYINIT} calls with the value desired for the process under study.
\begin{subentry}
\iteme{= 0 :} $\alpha_{\mathrm{s}}$ is fixed at the value \ttt{PARU(111)}.
\iteme{= 1 :} first-order running $\alpha_{\mathrm{s}}$ is used.
\iteme{= 2 :} second-order running $\alpha_{\mathrm{s}}$ is used.
\end{subentry}
\iteme{MSTU(112) :} (D=5) the nominal number of flavours assumed in
the $\alpha_{\mathrm{s}}$ expression, with respect to which $\Lambda$ is defined.
\iteme{MSTU(113) :} (D=3) minimum number of flavours that may be
assumed in $\alpha_{\mathrm{s}}$ expression, see \ttt{MSTU(112)}.
\iteme{MSTU(114) :} (D=5) maximum number of flavours that may be
assumed in $\alpha_{\mathrm{s}}$ expression, see \ttt{MSTU(112)}.
\iteme{MSTU(115) :} (D=0) treatment of $\alpha_{\mathrm{s}}$ singularity for
$Q^2 \to 0$.
\begin{subentry}
\iteme{= 0 :} allow it to diverge like $1/\ln(Q^2/\Lambda^2)$.
\iteme{= 1 :} soften the divergence to $1/\ln(1 + Q^2/\Lambda^2)$.
\iteme{= 2 :} freeze $Q^2$ evolution below \ttt{PARU(114)}, i.e.
the effective argument is $\max(Q^2, $\ttt{PARU(114)}$)$.
\end{subentry}
\iteme{MSTU(118) :} (I) number of flavours $n_f$ found and used in
latest \ttt{ULALPS} call.
\boxsep
\iteme{PARU(101) :}\label{p:PARU101} (D=0.00729735=1/137.04)
$\alpha_{\mathrm{em}}$, the electromagnetic fine structure constant at
vanishing momentum transfer.
\iteme{PARU(102) :} (D=0.232) $\sin^2 \! \theta_W$, the weak mixing angle of the
standard electroweak model.
\iteme{PARU(103) :} (D=0.007764=1/128.8) typical $\alpha_{\mathrm{em}}$ in
electroweak processes; used for $Q^2 >$\ttt{PARU(104)} in the
option \ttt{MSTU(101)=2} of \ttt{ULALEM}.
\iteme{PARU(104) :} (D=1 GeV$^2$) dividing line between `low' and
`high' $Q^2$ values in the option \ttt{MSTU(101)=2} of \ttt{ULALEM}.
\iteme{PARU(105) :} (D=1.16639E-5 GeV$^{-2}$) $G_{\mathrm{F}}$, the Fermi
constant of weak interactions.
\iteme{PARU(108) :} (I) the $\alpha_{\mathrm{em}}$ value obtained in the
latest call to the \ttt{ULALEM} function.
\iteme{PARU(111) :} (D=0.20) fix $\alpha_{\mathrm{s}}$ value assumed in
\ttt{ULALPS} when \ttt{MSTU(111)=0} (and also in parton showers
when $\alpha_{\mathrm{s}}$ is assumed fix there).
\iteme{PARU(112) :} (D=0.25 GeV) $\Lambda$ used in running $\alpha_{\mathrm{s}}$
expression in \ttt{ULALPS}. Like \ttt{MSTU(111)}, this value is
overwritten by the calling physics routines, and is therefore purely
nominal.
\iteme{PARU(113) :} (D=1.) the flavour thresholds, for the effective
number of flavours $n_f$ to use in the $\alpha_{\mathrm{s}}$ expression, are
assumed to sit at $Q^2 = $\ttt{PARU(113)}$\times m_{\mathrm{q}}^2$, where
$m_{\mathrm{q}}$ is the quark mass. May be overwritten from the calling
physics routine.
\iteme{PARU(114) :} (D=4 GeV$^2$) $Q^2$ value below which the
$\alpha_{\mathrm{s}}$ value is assumed constant for \ttt{MSTU(115)=2}.
\iteme{PARU(115) :} (D=10.) maximum $\alpha_{\mathrm{s}}$ value that \ttt{ULALPS}
will ever return; is used as a last resort to avoid singularities.
\iteme{PARU(117) :} (I) $\Lambda$ value (associated with
\ttt{MSTU(118)} effective flavours) obtained in latest \ttt{ULALPS}
call.
\iteme{PARU(118) :} (I) $\alpha_{\mathrm{s}}$ value obtained in latest
\ttt{ULALPS} call.
\iteme{PARU(121) - PARU(130) :} couplings of a new $\mathrm{Z}'^0$; for
fermion default values are given by the Standard Model $\mathrm{Z}^0$ values,
assuming $\sin^2 \! \theta_W = 0.23$. Note that e.g. the $\mathrm{Z}'^0$ width contains
squared couplings, and thus depends quadratically on the values below.
\begin{subentry}
\iteme{PARU(121), PARU(122) :} (D=-0.693,-1.) vector and axial
couplings of down type quarks to $\mathrm{Z}'^0$.
\iteme{PARU(123), PARU(124) :} (D=0.387,1.) vector and axial
couplings of up type quarks to $\mathrm{Z}'^0$.
\iteme{PARU(125), PARU(126) :} (D=-0.08,-1.) vector and axial
couplings of leptons to $\mathrm{Z}'^0$.
\iteme{PARU(127), PARU(128) :} (D=1.,1.) vector and axial
couplings of neutrinos to $\mathrm{Z}'^0$.
\iteme{PARU(129) :} (D=1.) the coupling $Z'^0 \to \mathrm{W}^+ \mathrm{W}^-$ is
taken to be \ttt{PARU(129)}$\times$(the Standard Model
$\mathrm{Z}^0 \to \mathrm{W}^+ \mathrm{W}^-$ coupling)$\times (m_{\mathrm{W}}/m_{\mathrm{Z}'})^2$.
This gives a $\mathrm{Z}'^0 \to \mathrm{W}^+ \mathrm{W}^-$ partial width that
increases proportionately to the $\mathrm{Z}'^0$ mass.
\iteme{PARU(130) :} (D=0.) in the decay chain
$\mathrm{Z}'^0 \to \mathrm{W}^+ \mathrm{W}^- \to 4$ fermions, the angular distribution in
the $\mathrm{W}$ decays is supposed to be a mixture, with fraction
\ttt{1-PARU(130)} corresponding to the same angular distribution
between the four final fermions as in $\mathrm{Z}^0 \to \mathrm{W}^+ \mathrm{W}^-$ (mixture
of transverse and longitudinal $\mathrm{W}$'s), and fraction \ttt{PARU(130)}
corresponding to $\H^0 \to \mathrm{W}^+ \mathrm{W}^-$ the same way (longitudinal
$\mathrm{W}$'s).
\end{subentry}
\iteme{PARU(131) - PARU(136) :} couplings of a new $\mathrm{W}'^{\pm}$;
for fermions default values are given by the Standard Model
$\mathrm{W}^{\pm}$ values (i.e. $V-A$). Note that e.g. the $\mathrm{W}'^{\pm}$
width contains squared couplings, and
thus depends quadratically on the values below.
\begin{subentry}
\iteme{PARU(131), PARU(132) :} (D=1.,-1.) vector and axial couplings
of a quark--antiquark pair to $\mathrm{W}'^{\pm}$; is further multiplied by the
ordinary CKM factors.
\iteme{PARU(133), PARU(134) :} (D=1.,-1.) vector and axial couplings
of a lepton-neutrino pair to $\mathrm{W}'^{\pm}$.
\iteme{PARU(135) :} (D=1.) the coupling $\mathrm{W}'^{\pm} \to \mathrm{Z}^0 \mathrm{W}^{\pm}$
is taken to be \ttt{PARU(135)}$\times$(the Standard Model
$\mathrm{W}^{\pm} \to \mathrm{Z}^0 \mathrm{W}^{\pm}$ coupling)$\times (m_{\mathrm{W}}/m_{W'})^2$.
This gives a $\mathrm{W}'^{\pm} \to \mathrm{Z}^0 \mathrm{W}^{\pm}$ partial width that
increases proportionately to the $\mathrm{W}'$ mass.
\iteme{PARU(136) :} (D=0.) in the decay chain
$\mathrm{W}'^{\pm} \to \mathrm{Z}^0 \mathrm{W}^{\pm} \to 4$ fermions,
the angular distribution in the $\mathrm{W}/\mathrm{Z}$ decays is supposed to be a
mixture, with fraction \ttt{1-PARU(130)} corresponding to the same
angular distribution between the four final fermions as in
$\mathrm{W}^{\pm} \to \mathrm{Z}^0 \mathrm{W}^{\pm}$ (mixture of transverse and longitudinal
$\mathrm{W}/\mathrm{Z}$'s), and fraction \ttt{PARU(130)} corresponding to
$\H^{\pm} \to \mathrm{Z}^0 \mathrm{W}^{\pm}$ the same way (longitudinal $\mathrm{W}/\mathrm{Z}$'s).
\end{subentry}
\iteme{PARU(141) :} (D=5.) $\tan\beta$ parameter of a two Higgs
doublet scenario, i.e. the ratio of vacuum expectation values.
This affects mass relations and couplings in the Higgs sector.
\iteme{PARU(142) :} (D=1.) the $\mathrm{Z}^0 \to \H^+ \H^-$ coupling is
taken to be \ttt{PARU(142)}$\times$(the MSSM $\mathrm{Z}^0 \to \H^+ \H^-$
coupling).
\iteme{PARU(143) :} (D=1.) the $\mathrm{Z}'^0 \to \H^+ \H^-$ coupling is
taken to be \ttt{PARU(143)}$\times$(the MSSM $\mathrm{Z}^0 \to \H^+ \H^-$
coupling).
\iteme{PARU(145) :} (D=1.) quadratically multiplicative factor in the
$\mathrm{Z}'^0 \to \mathrm{Z}^0 \H^0$ partial width in left--right-symmetric models,
expected to be unity (see \cite{Coc91}).
\iteme{PARU(146) :} (D=1.) $\sin(2\alpha)$ parameter, enters
quadratically as multiplicative factor in the
$\mathrm{W}'^{\pm} \to \mathrm{W}^{\pm} \H^0$ partial width in
left--right-symmetric models (see \cite{Coc91}).
\iteme{PARU(151) :} (D=1.) multiplicative factor in the
$\L_{\mathrm{Q}} \to \mathrm{q} \ell$ squared Yukawa coupling, and thereby in the
$\L_{\mathrm{Q}}$ partial width and the $\mathrm{q} \ell \to \L_{\mathrm{Q}}$ and other
cross sections. Specifically,
$\lambda^2/(4\pi) = $\ttt{PARU(151)}$\times \alpha_{\mathrm{em}}$, i.e. it
corresponds to the $k$ factor of \cite{Hew88}.
\iteme{PARU(153) :} (D=0.) anomalous magnetic moment of the
$\mathrm{W}^{\pm}$; $\eta = \kappa - 1$, where $\eta = 0$ ($\kappa = 1$)
is the Standard Model value.
\iteme{PARU(155) :} (D=1000. GeV) compositeness scale $\Lambda$.
\iteme{PARU(156) :} (D=1.) sign of interference term between
standard cross section and composite term ($\eta$ parameter);
should be $\pm 1$.
\iteme{PARU(157) - PARU(159) :} (D=3*1.) strength of {\bf SU(2)},
{\bf U(1)} and {\bf SU(3)} couplings, respectively, in an excited
fermion scenario; cf. $f$, $f'$ and $f_s$ of \cite{Bau90}.
\iteme{PARU(161) - PARU(168) :} (D=5*1.,3*0.) multiplicative factors
that can be used to modify the default couplings of the $\H^0$
particle in {\tsc{Pythia}}. Note that the factors enter quadratically in the
partial widths. The default values correspond to the couplings given
in the minimal one-Higgs-doublet Standard Model.
\begin{subentry}
\iteme{PARU(161) :} $\H^0$ coupling to down type quarks.
\iteme{PARU(162) :} $\H^0$ coupling to up type quarks.
\iteme{PARU(163) :} $\H^0$ coupling to leptons.
\iteme{PARU(164) :} $\H^0$ coupling to $\mathrm{Z}^0$.
\iteme{PARU(165) :} $\H^0$ coupling to $\mathrm{W}^{\pm}$.
\iteme{PARU(168) :} $\H^0$ coupling to $\H^{\pm}$ in
$\gamma\gamma \to \H^0$ loops, in MSSM
$\sin(\beta-\alpha)+\cos(2\beta)\sin(\beta+\alpha) /
(2\cos^2 \! \theta_W)$.
\end{subentry}
\iteme{PARU(171) - PARU(178) :} (D=7*1.,0.) multiplicative factors
that can be used to modify the default couplings of the $\H'^0$
particle in {\tsc{Pythia}}. Note that the factors enter quadratically in
partial widths. The default values for \ttt{PARU(171) - PARU(175)}
correspond to the couplings given to $\H^0$ in the minimal
one-Higgs-doublet Standard Model, and are therefore not realistic
in a two-Higgs-doublet scenario. The default values should
be changed appropriately by you. Also the last two default
values should be changed; for these the expressions of the
minimal supersymmetric Standard Model (MSSM) are given to show
parameter normalization.
\begin{subentry}
\iteme{PARU(171) :} $\H'^0$ coupling to down type quarks.
\iteme{PARU(172) :} $\H'^0$ coupling to up type quarks.
\iteme{PARU(173) :} $\H'^0$ coupling to leptons.
\iteme{PARU(174) :} $\H'^0$ coupling to $\mathrm{Z}^0$.
\iteme{PARU(175) :} $\H'^0$ coupling to $W^{\pm}$.
\iteme{PARU(176) :} $\H'^0$ coupling to $\H^0 \H^0$, in MSSM
$\cos(2\alpha) \cos(\beta+\alpha) - 2 \sin(2\alpha)
\sin(\beta+\alpha)$.
\iteme{PARU(177) :} $\H'^0$ coupling to $\mathrm{A}^0 \mathrm{A}^0$, in MSSM
$\cos(2\beta) \cos(\beta+\alpha)$.
\iteme{PARU(178) :} $\H'^0$ coupling to $\H^{\pm}$ in
$\gamma \gamma \to \H'^0$ loops, in MSSM
$\cos(\beta-\alpha) - \cos(2\beta)\cos(\beta+\alpha) /
(2\cos^2 \! \theta_W)$.
\end{subentry}
\iteme{PARU(181) - PARU(190) :} (D=3*1.,2*0.,2*1.,3*0.)
multiplicative factors that can be used to modify the default
couplings of the $\mathrm{A}^0$ particle in PYTHIA. Note that the factors
enter quadratically in partial widths. The default values for
\ttt{PARU(181) - PARU(183)} correspond
to the couplings given to $\H^0$ in the minimal one-Higgs-doublet
Standard Model, and are therefore not realistic in a
two-Higgs-doublet scenario. The default values should
be changed appropriately by you. \ttt{PARU(184)} and \ttt{PARU(185)}
should be vanishing at the tree level, and are so set;
normalization of these couplings agrees with what is used for
H and $\H'^0$. Also the other default values should be changed; for
these the expressions of the Minimal Supersymmetric Standard
Model (MSSM) are given to show parameter normalization.
\begin{subentry}
\iteme{PARU(181) :} $\mathrm{A}^0$ coupling to down type quarks.
\iteme{PARU(182) :} $\mathrm{A}^0$ coupling to up type quarks.
\iteme{PARU(183) :} $\mathrm{A}^0$ coupling to leptons.
\iteme{PARU(184) :} $\mathrm{A}^0$ coupling to $\mathrm{Z}^0$.
\iteme{PARU(185) :} $\mathrm{A}^0$ coupling to $\mathrm{W}^{\pm}$.
\iteme{PARU(186) :} $\mathrm{A}^0$ coupling to $\mathrm{Z}^0 \H^0$ (or
$\mathrm{Z}^*$ to $\mathrm{A}^0 \H^0$), in MSSM $\cos(\beta-\alpha)$.
\iteme{PARU(187) :} $\mathrm{A}^0$ coupling to $\mathrm{Z}^0 \H'^0$ (or $\mathrm{Z}^*$ to
$\mathrm{A}^0 \H'^0$), in MSSM $\sin(\beta-\alpha)$.
\iteme{PARU(188) :} As \ttt{PARU(186)}, but coupling to $\mathrm{Z}'^0$
rather than $\mathrm{Z}^0$.
\iteme{PARU(189) :} As \ttt{PARU(187)}, but coupling to $\mathrm{Z}'^0$
rather than $\mathrm{Z}^0$.
\iteme{PARU(190) :} $\mathrm{A}^0$ coupling to $\H^{\pm}$ in
$\gamma \gamma \to \mathrm{A}^0$ loops, 0 in MSSM.
\end{subentry}
\iteme{PARU(191) - PARU(195) :} (D=4*0.,1.) multiplicative factors
that can be used to modify the couplings of the $\H^{\pm}$ particle
in {\tsc{Pythia}}. Currently only \ttt{PARU(195)} is in use. See above for
related comments.
\begin{subentry}
\iteme{PARU(195) :} $\H^{\pm}$ coupling to $\mathrm{W}^{\pm} \H^0$ (or
$\mathrm{W}^{* \pm}$ to $\H^{\pm} \H^0$), in MSSM $\cos(\beta-\alpha)$.
\end{subentry}
\end{entry}
\subsection{Further Parameters and Particle Data}
\label{ss:parapartdat}
The following common blocks are maybe of a more peripheral interest,
with the exception of the \ttt{MDCY} array, which allows a selective
inhibiting of particle decays, and masses of not yet discovered
particles, such as \ttt{PMAS(6,1)}, the top quark mass.
\drawbox{COMMON/LUDAT2/KCHG(500,3),PMAS(500,4),PARF(2000),VCKM(4,4)}%
\label{p:LUDAT2}
\begin{entry}
\itemc{Purpose:} to give access to a number of flavour treatment
constants or parameters and particle/parton data. Particle data is
stored by compressed code KC rather than by the full KF code.
You are reminded that the way to know the KC value is to use the
\ttt{LUCOMP} function, i.e. \ttt{KC = LUCOMP(KF)}.
\boxsep
\iteme{KCHG(KC,1) :}\label{p:KCHG} three times particle/parton charge
for compressed code KC.
\iteme{KCHG(KC,2) :} colour information for compressed code KC.
\begin{subentry}
\iteme{= 0 :} colour-singlet particle.
\iteme{= 1 :} quark or antidiquark.
\iteme{= -1 :} antiquark or diquark.
\iteme{= 2 :} gluon.
\end{subentry}
\iteme{KCHG(KC,3) :} particle/antiparticle distinction for
compressed code KC.
\begin{subentry}
\iteme{= 0 :} the particle is its own antiparticle.
\iteme{= 1 :} a nonidentical antiparticle exists.
\end{subentry}
\boxsep
\iteme{PMAS(KC,1) :}\label{p:PMAS} particle/parton mass $m$ (in GeV)
for compressed code KC.
\iteme{PMAS(KC,2) :} the total width $\Gamma$ (in GeV) of an assumed
symmetric Breit--Wigner mass shape for compressed particle code KC.
\iteme{PMAS(KC,3) :} the maximum deviation (in GeV) from the
\ttt{PMAS(KC,1)} value at which the Breit--Wigner shape above is
truncated. (Is used in particle decays, but not in the {\tsc{Pythia}} resonance
treatment; cf. the \ttt{CKIN} variables.)
\iteme{PMAS(KC,4) :} the average lifetime $\tau$ for compressed
particle code KC, with $c \tau$ in mm, i.e. $\tau$ in units of
about $3.33 \times 10^{-12}$ s.
\boxsep
\iteme{PARF(1) - PARF(60) :}\label{p:PARF} give a parametrization of
the $\d\overline{\mathrm{d}}$--$\u\overline{\mathrm{u}}$--$\mathrm{s}\overline{\mathrm{s}}$ flavour mixing in production of
flavour-diagonal mesons. Numbers are stored in groups of 10, for
the six multiplets pseudoscalar, vector, axial vector ($S=0$),
scalar, axial vector ($S=1$) and tensor, in this order; see
section \ref{sss:mesonprod}. Within each group, the first two
numbers determine the fate of a $\d\overline{\mathrm{d}}$ flavour state, the
second two that of a $\u\overline{\mathrm{u}}$ one, the next two that of an
$\mathrm{s}\overline{\mathrm{s}}$ one, while the last four are unused. Call the numbers
of a pair $p_1$ and $p_2$. Then the probability to produce the
state with smallest KF code is $1-p_1$, the probability for the
middle one is $p_1 - p_2$ and the probability for the one with
largest code is $p_2$, i.e. $p_1$ is the probability to produce
either of the two `heavier' ones.
\iteme{PARF(61) - PARF(80) :} give flavour {\bf SU(6)} weights for
the production of a spin 1/2 or spin 3/2 baryon from a given
diquark--quark combination. Should not be changed.
\iteme{PARF(101) - PARF(108) :} first five contain $\d$, $\u$, $\mathrm{s}$,
$\c$ and $\b$ constituent masses, as to be used in mass formulae, and
should not be changed. For $\t$, $\mathrm{l}$ and $\mathrm{h}$ masses the current
values stored in \ttt{PMAS(6,1) - PMAS(8,1)} are copied in.
\iteme{PARF(111), PARF(112) :} (D=0.0, 0.11 GeV) constant terms in the
mass formulae for heavy mesons and baryons, respectively (with diquark
getting 2/3 of baryon).
\iteme{PARF(113), PARF(114) :} (D=0.16,0.048 GeV) factors which,
together with Clebsch-Gordan coefficients and quark constituent
masses, determine the mass splitting due to spin-spin interactions
for heavy mesons and baryons, respectively. The latter factor is
also used for the splitting between spin 0 and spin 1 diquarks.
\iteme{PARF(115) - PARF(118) :} (D=0.50, 0.45, 0.55, 0.60 GeV),
constant mass terms, added to the constituent masses, to get the
mass of heavy mesons with orbital angular momentum $L = 1$. The four
numbers are for pseudovector mesons with quark spin 0, and for scalar,
pseudovector and tensor mesons with quark spin 1, respectively.
\iteme{PARF(121), PARF(122) :} (D=0.1, 0.2 GeV) constant terms, which
are subtracted for quark and diquark masses, respectively, in defining
the allowed phase space in particle decays into partons.
\iteme{PARF(201) - PARF(1960) :} (D=1760*0) relative probabilities
for flavour production in the \ttt{MSTJ(15)=1} option; to be
defined by you before any {\tsc{Jetset}} calls. \\
The index in \ttt{PARF} is of the compressed form \\
$120 + 80 \times$KTAB1$ + 25 \times$KTABS$ + $KTAB3. \\
Here KTAB1 is the
old flavour, fixed by preceding fragmentation history, while KTAB3
is the new flavour, to be selected according to the relevant relative
probabilities (except for the very last particle, produced when
joining two jets, where both KTAB1 and KTAB3 are known).
Only the most frequently appearing quarks/diquarks are defined,
according to the code $1 = \d$, $2 = \u$, $3 = \mathrm{s}$, $4 = \c$,
$5 = \b$, $6 = \t$, $7 = \d\d_1$, $8 = \u\d_0$, $9 = \u\d_1$,
$10 = \u\u_1$, $11 = \mathrm{s}\d_0$, $12 = \mathrm{s}\d_1$, $13 = \mathrm{s}\u_0$,
$14 = \mathrm{s}\u_1$, $15 = \mathrm{s}\s_1$, $16 = \c\d_0$, $17 = \c\d_1$,
$18 = \c\u_0$, $19 = \c\u_1$, $20 = \c\mathrm{s}_0$, $21 = \c\mathrm{s}_1$,
$22 = \c\c_1$.
These are thus the only possibilities for the new flavour to be
produced; for an occasional old flavour not on this list, the
ordinary relative flavour production probabilities will be used. \\
Given the initial and final flavour, the intermediate hadron that
is produced is almost fixed. (Initial and final diquark here
corresponds to `popcorn' production of mesons intermediate between
a baryon and an antibaryon). The additional index KTABS gives the
spin type of this hadron, with \\
0 = pseudoscalar meson or $\Lambda$-like spin 1/2 baryon, \\
1 = vector meson or $\Sigma$-like spin 1/2 baryon, \\
2 = tensor meson or spin 3/2 baryon. \\
(Some meson multiplets, not frequently produced, are not accessible
by this parametrization.) \\
Note that some combinations of KTAB1, KTAB3 and KTABS do not
correspond to a physical particle (a $\Lambda$-like baryon must
contain three different quark flavours, a $\Sigma$-like one at least
two), and that you must see to it that the corresponding
\ttt{PARF} entries are vanishing. One additional complication exist
when KTAB3 and KTAB1 denote the same flavour content (normally
KTAB3$ = $KTAB1, but for diquarks the spin freedom may give
KTAB3$ = $KTAB1$\pm 1$): then a flavour neutral meson is to be
produced, and here $\d\overline{\mathrm{d}}$, $\u\overline{\mathrm{u}}$ and $\mathrm{s}\overline{\mathrm{s}}$ states mix
(heavier flavour states do not, and these are therefore no problem).
For these cases the ordinary KTAB3 value gives the total probability
to produce either of the mesons possible, while KTAB3$ = $23 gives the
relative probability to produce the lightest meson state ($\pi^0$,
$\rho^0$, $\a_2^0$), KTAB3$ = $24 relative probability for the middle
meson ($\eta$, $\omega$, $\mathrm{f}_2^0$), and KTAB3 = 25 relative
probability for the heaviest one ($\eta'$, $\phi$, $f'^0_2$). Note
that, for simplicity, these relative probabilities are assumed the
same whether initial and final diquark have the same spin or not; the
total probability may well be assumed different, however. \\
As a general comment, the sum of \ttt{PARF} values for a given KTAB1
need not be normalized to unity, but rather the program will find the
sum of relevant weights and normalize to that. The same goes for
the KTAB3$ = $23--25 weights. This makes it straightforward to use
one common setup of \ttt{PARF} values and still switch between
different \ttt{MSTJ(12)} baryon production modes.
\boxsep
\iteme{VCKM(I,J) :}\label{p:VCKM} squared matrix elements of the
Cabibbo-Kobayashi-Maskawa flavour mixing matrix.
\begin{subentry}
\iteme{I :} up type generation index, i.e. $1 = \u$, $2 = \c$,
$3 = \t$ and $4 = \mathrm{h}$.
\iteme{J :} down type generation index, i.e. $1 = \d$, $2 = \mathrm{s}$,
$3 = \b$ and $4 = \mathrm{l}$.
\end{subentry}
\end{entry}
\drawbox{COMMON/LUDAT3/MDCY(500,3),MDME(2000,2),BRAT(2000),%
KFDP(2000,5)}\label{p:LUDAT3}
\begin{entry}
\itemc{Purpose:} to give access to particle decay data and
parameters. In particular, the \ttt{MDCY(KC,1)} variables may be
used to switch on or off the decay of a given particle species,
and the \ttt{MDME(IDC,1)} ones
to switch on or off an individual decay channel of a particle.
For quarks, leptons and gauge bosons, a number of decay channels
are included that are not allowed for on-mass-shell particles, see
\ttt{MDME(IDC,2)=102}. These channels are not currently used in
{\tsc{Jetset}}, but instead find applications in {\tsc{Pythia}}. Particle data is
stored by compressed code KC rather than by the full KF code.
You are reminded that the way to know the KC value is to use the
\ttt{LUCOMP} function, i.e. \ttt{KC = LUCOMP(KF)}.
\boxsep
\iteme{MDCY(KC,1) :}\label{p:MDCY} switch to tell whether a particle
with compressed code KC may be allowed to decay or not.
\begin{subentry}
\iteme{= 0 :} the particle is not allowed to decay.
\iteme{= 1 :} the particle is allowed to decay (if decay information
is defined below for the particle).
\end{subentry}
\iteme{MDCY(KC,2) :} gives the entry point into the decay channel
table for compressed particle code KC. Is 0 if no decay channels
have been defined.
\iteme{MDCY(KC,3) :} gives the total number of decay channels defined
for compressed particle code KC, independently of whether they have
been assigned a non-vanishing branching ratio or not. Thus the
decay channels are found in positions \ttt{MDCY(KC,2)} to
\ttt{MDCY(KC,2)+MDCY(KC,3)-1}.
\boxsep
\iteme{MDME(IDC,1) :}\label{p:MDME} on/off switch for individual
decay channel IDC. In addition, a channel may be left selectively
open; this has some special applications in {\tsc{Pythia}} which are not
currently used in {\tsc{Jetset}}. Effective branching ratios are
automatically recalculated for the decay channels left open.
Also process cross sections are affected; see section
\ref{sss:resdecaycross}. If a particle is allowed to decay by the
\ttt{MDCY(KC,1)} value, at least one channel must be left open
by you. A list of decay channels with current IDC numbers
may be obtained with \ttt{LULIST(12)}.
\begin{subentry}
\iteme{= -1 :} this is a non-Standard Model decay mode, which by
default is assumed not to exist. Normally, this option is used for
decays involving fourth generation or $\H^{\pm}$ particles.
\iteme{= 0 :} channel is switched off.
\iteme{= 1 :} channel is switched on.
\iteme{= 2 :} channel is switched on for a particle but off for an
antiparticle. It is also on for a particle its own antiparticle,
i.e. here it means the same as \ttt{=1}.
\iteme{= 3 :} channel is switched on for an antiparticle but off for
a particle. It is off for a particle its own antiparticle.
\iteme{= 4 :} in the production of a pair of equal or charge
conjugate resonances in {\tsc{Pythia}}, say $\H^0 \to \mathrm{W}^+ \mathrm{W}^-$, either one
of the resonances is allowed to decay according to this group of
channels, but not both. If the two particles of the pair
are different, the channel is on.
Within {\tsc{Jetset}}, this option only means that the channel is
switched off.
\iteme{= 5 :} as \ttt{=4}, but an independent group of channels, such
that in a pair of equal or charge conjugate resonances the decay of
either resonance may be specified independently. If the two
particles in the pair are different, the channel is off.
Within {\tsc{Jetset}}, this option only means that the channel is
switched off.
\itemc{Warning:} the two values -1 and 0 may look similar, but in fact
are quite different. In neither case the channel so set is
generated, but in the latter case the channel still contributes
to the total width of a resonance, and thus affects both
simulated line shape and the generated cross section when
{\tsc{Pythia}} is run. The value 0 is appropriate to a channel
we assume exists, even if we are not currently simulating it,
while -1 should be used for channels we believe do not exist.
In particular, you are warned unwittingly to set fourth
generation channels 0 (rather than -1), since by now the
support for a fourth generation is small.
\itemc{Remark:} all the options above may be freely mixed. The
difference, for those cases where both make sense, between using
values 2 and 3 and using 4 and 5 is that the latter automatically
include charge conjugate states, e.g.
$\H^0 \to \mathrm{W}^+ \mathrm{W}^- \to \mathrm{e}^+ \nu_e \d \overline{\mathrm{u}}$ or
$\overline{\mathrm{d}} \u \mathrm{e}^- \br{\nu}_e$, but the former only one
of them. In calculations of the joint branching ratio, this
makes a factor 2 difference.
\itemc{Example:} to illustrate the above options, consider the case
of a $\mathrm{W}^+ \mathrm{W}^-$ pair. One might then set the following combination
of switches for the $\mathrm{W}$:\\
\begin{tabular}{ccl@{\protect\rule{0mm}{\tablinsep}}}
channel & value & comment \\
$\u\overline{\mathrm{d}}$ & 1 & allowed for $\mathrm{W}^+$ and $\mathrm{W}^-$ in any combination, \\
$\u\overline{\mathrm{s}}$ & 0 & never produced but contributes to $\mathrm{W}$ width, \\
$\c\overline{\mathrm{d}}$ & 2 & allowed for $\mathrm{W}^+$ only, \\
$\c\overline{\mathrm{s}}$ & 3 & allowed for $\mathrm{W}^-$ only, i.e. properly
$\mathrm{W}^- \to \overline{\mathrm{c}}\mathrm{s}$, \\
$\t\overline{\mathrm{b}}$ & 0 & never produced but contributes to $\mathrm{W}$ width \\
& & if the channel is kinematically allowed, \\
$\nu_{\mathrm{e}}\mathrm{e}^+$ & 4 & allowed for one of $\mathrm{W}^+$ or $\mathrm{W}^-$, but not
both, \\
$\nu_{\mu}\mu^+$ & 4 & allowed for one of $\mathrm{W}^+$ or $\mathrm{W}^-$, but not
both, \\
& & and not in combination with $\nu_{\mathrm{e}}\mathrm{e}^+$, \\
$\nu_{\tau}\tau^+$ & 5 & allowed for the other $\mathrm{W}$, but not both, \\
$\nu_{\chi}\chi^-$ & $-1$ & not produced and does not contribute to
$\mathrm{W}$ width.\\
\end{tabular}\\
A $\mathrm{W}^+\mathrm{W}^-$ final state $\u\overline{\mathrm{d}} + \overline{\mathrm{c}}\mathrm{s}$ is allowed, but not its
charge conjugate $\overline{\mathrm{u}}\d + \c\overline{\mathrm{s}}$, since the latter decay mode is
not allowed for a $\mathrm{W}^+$. The combination
$\nu_{\mathrm{e}}\mathrm{e}^+ + \bar{\nu}_{\tau}\tau^-$ is allowed, since the two
channels belong to different groups, but not
$\nu_{\mathrm{e}}\mathrm{e}^+ + \bar{\nu}_{\mu}\mu^-$, where both belong to the same.
Both $\u\overline{\mathrm{d}} + \bar{\nu}_{\tau}\tau^-$ and
$\overline{\mathrm{u}}\d + \nu_{\tau}\tau^+$ are allowed, since there is no clash.
The full rulebook, for this case, is given by
eq.~(\ref{eq:WWallchancombgen}). A term $r_i^2$ means channel
$i$ is allowed
for $\mathrm{W}^+$ and $\mathrm{W}^-$ simultaneously, a term $r_i r_j$ that channels
$i$ and $j$ may be combined, and a term $2 r_i r_j$ that channels
$i$ and $j$ may be combined two ways, i.e. that also a charge
conjugate combination is allowed.
\end{subentry}
\iteme{MDME(IDC,2) :} information on special matrix-element treatment
for decay channel IDC. In addition to the outline below, special rules
apply for the order in which decay products should be given, so that
matrix elements and colour flow is properly treated. One such example
is the weak matrix elements, which only will be correct if decay
products are given in the right order. The program does not police
this, so if you introduce channels of your own and use these codes,
you should be guided by the existing particle data.
\begin{subentry}
\iteme{= 0 :} no special matrix-element treatment; partons and
particles are copied directly to the event record, with momentum
distributed according to phase space.
\iteme{= 1 :} $\omega$ and $\phi$ decays into three pions, eq.
(\ref{dec:omegphi}).
\iteme{= 2 :} $\pi^0$ or $\eta$ Dalitz decay to $\gamma \mathrm{e}^+ \mathrm{e}^-$,
eq.~(\ref{dec:Dalitz}).
\iteme{= 3 :} used for vector meson decays into two pseudoscalars,
to signal non-isotropic decay angle according to eq.
(\ref{dec:psvpsps}), where relevant.
\iteme{= 4 :} decay of a spin 1 onium resonance to three gluons or
to a photon and two gluons, eq.~(\ref{ee:Upsilondec}). The gluons may
subsequently develop a shower if \ttt{MSTJ(23)=1}.
\iteme{= 11 :} phase-space production of hadrons from the quarks
available.
\iteme{= 12 :} as \ttt{=11}, but for onia resonances, with the option
of modifying the multiplicity distribution separately.
\iteme{= 13 :} as \ttt{=11}, but at least three hadrons to be produced
(useful when the two-body decays are given explicitly).
\iteme{= 14 :} as \ttt{=11}, but at least four hadrons to be produced.
\iteme{= 15 :} as \ttt{=11}, but at least five hadrons to be produced.
\iteme{= 22 - 30 :} phase-space production of hadrons from the quarks
available, with the multiplicity fixed to be \ttt{MDME(IDC,2)-20},
i.e. 2--10.
\iteme{= 31 :} two or more quarks and particles are distributed
according to phase space. If three or more products, the last product
is a spectator quark, i.e. sitting at rest with respect to the
decaying hadron.
\iteme{= 32 :} a $\mathrm{q}\overline{\mathrm{q}}$ or $\mathrm{g}\g$ pair, distributed according to
phase space (in angle), and allowed to develop a shower if
\ttt{MSTJ(23)=1}.
\iteme{= 33 :} a triplet $\mathrm{q} X \overline{\mathrm{q}}$, where $X$ is either a gluon or
a colour-singlet particle; the final particle ($\overline{\mathrm{q}}$) is assumed to
sit at rest with respect to the decaying hadron, and the
two first particles ($\mathrm{q}$ and $X$) are allowed to develop a shower if
\ttt{MSTJ(23)=1}.
\iteme{= 41 :} weak decay, where particles are distributed according
to phase space, multiplied by a factor from the expected shape of
the momentum spectrum of the direct product of the weak decay
(the $\nu_{\tau}$ in $\tau$ decay).
\iteme{= 42 :} weak decay matrix element for quarks and leptons.
Products may be given either in terms of quarks or hadrons, or
leptons for some channels. If the spectator system is given in
terms of quarks, it is assumed to collapse into one particle from
the onset. If the virtual $\mathrm{W}$ decays into quarks, these quarks
are converted to particles, according to phase space in the
$\mathrm{W}$ rest frame, as in \ttt{=11}. Is intended for $\tau$, charm and
bottom.
\iteme{= 43 :} as \ttt{=42}, but if the $\mathrm{W}$ decays into quarks, these
will either appear as jets or, for small masses, collapse into a
one- or two-body system.
\iteme{= 44 :} weak decay matrix element for quarks and leptons, where
the spectator system may collapse into one particle for a small
invariant mass. If the first two decay products are a $\mathrm{q}\overline{\mathrm{q}}'$
pair, they may develop a parton shower if \ttt{MSTJ(23)=1}.
Is intended for top and beyond, but largely superseded by
the following option.
\iteme{= 45 :} weak decay $\mathrm{q} \to \mathrm{W} \mathrm{q}'$ or $\ell \to \mathrm{W} \nu_{\ell}$,
where the $\mathrm{W}$ is registered as a decay product and subsequently
treated with \ttt{MDME=46}. To distinguish from ordinary $\mathrm{W}$'s
on the mass shell, code KF$ = \pm 89$ is used. The virtual $\mathrm{W}$ mass
is selected according to the standard weak decay matrix element,
times the $\mathrm{W}$ propagator (for \ttt{MSTJ(25)=1}). There may be two
or three decay products; if a third this is a spectator taken to sit
at rest. The spectator system may collapse into one particle. Is
intended for top and beyond.
\iteme{= 46 :} $\mathrm{W}$ (KF = 89) decay into $\mathrm{q}\overline{\mathrm{q}}'$ or
$\ell \nu_{\ell}$ according to relative probabilities given
by couplings (as stored in the \ttt{BRAT} vector)
times a dynamical phase-space factor given by the current $\mathrm{W}$
mass. In the decay, the correct $V-A$ angular distribution is
generated if the $\mathrm{W}$ origin is known (heavy quark or lepton).
This is therefore the second step of a decay with \ttt{MDME=45}.
A $\mathrm{q}\overline{\mathrm{q}}'$ pair may subsequently develop a shower if
\ttt{MSTJ(23)=1}.
\iteme{= 84 - 88 :} map the decay of this particle onto the generic
$\c$, $\b$, $\t$, $\mathrm{l}$ or $\mathrm{h}$ decay modes defined for
KC$ = $84--88.
\iteme{= 48 :} as \ttt{=42}, but require at least three decay
products.
\iteme{= 101 :} this is not a proper decay channel, but only to be
considered as a continuation line for the decay product listing
of the immediately preceding channel. Since the \ttt{KFDP} array can
contain five decay products per channel, with this code it is
possible to define channels with up to ten decay products. It is
not allowed to have several continuation lines after each other.
\iteme{= 102 :} this is not a proper decay channel for a decaying
particle on the mass shell (or nearly so), and is therefore assigned
branching ratio 0. For a particle off the mass shell, this decay
mode is allowed, however. By including this channel among the
others, the switches \ttt{MDME(IDC,1)} may be used to allow or
forbid these channels in hard processes, with cross sections
to be calculated separately. As an example, $\gamma \to \u \overline{\mathrm{u}}$
is not possible for a massless photon, but is an allowed
channel in $\e^+\e^-$ annihilation.
\end{subentry}
\boxsep
\iteme{BRAT(IDC) :}\label{p:BRAT} give branching ratios for the
different decay
channels. In principle, the sum of branching ratios for a given
particle should be unity. Since the program anyway has to calculate
the sum of branching ratios left open by the \ttt{MDME(IDC,1)} values
and normalize to that, you need not explicitly ensure this
normalization, however. (Warnings are printed in \ttt{LUUPDA(2)} calls
if the sum is not unity, but this is entirely intended as a help
for finding user mistypings.) For decay channels with
\ttt{MDME(IDC,2)}$> 80$ the \ttt{BRAT} values are dummy.
\boxsep
\iteme{KFDP(IDC,J) :}\label{p:KFDP} contain the decay products in the
different channels, with five positions \ttt{J=} 1--5 reserved for each
channel IDC. The decay products are given following the standard KF
code for jets and particles, with 0 for trailing empty positions. Note
that the \ttt{MDME(IDC+1,2)=101} option allows you to double the
maximum number of decay product in a given channel from 5 to 10,
with the five latter products stored \ttt{KFDP(IDC+1,J)}.
\end{entry}
\drawboxtwo{COMMON/LUDAT4/CHAF(500)}{CHARACTER CHAF*8}\label{p:LUDAT4}
\begin{entry}
\itemc{Purpose:} to give access to character type variables.
\boxsep
\iteme{CHAF :}\label{p:CHAF} particle names (excluding charge)
according to KC code.
\end{entry}
\subsection{Miscellaneous Comments}
The previous sections have dealt with the subroutine options and
variables one at a time. This is certainly important, but for a full
use of the capabilities of the program, it is also necessary
to understand how to make different pieces work together. This is
something that cannot be explained fully in a manual, but must also
be learnt by trial and error. This section contains some examples of
relationships between subroutines, common blocks and parameters.
It also contains comments on issues that did not fit in naturally
anywhere else, but still might be useful to have on record.
\subsubsection{Interfacing to detector simulation}
Very often, the output of the program is to be fed into a subsequent
detector simulation program. It therefore becomes necessary to set up
an interface between the \ttt{LUJETS} common block and the detector
model. Preferrably this should be done via the \ttt{HEPEVT} standard
common block, see section \ref{ss:HEPEVT}, but sometimes this may not
be convenient. If a \ttt{LUEDIT(2)} call is made, the remaining
entries exactly correspond to those an ideal detector could see: all
non-decayed particles, with the exception of neutrinos. The translation
of momenta should be trivial (if need be, a \ttt{LUROBO} call can be
made to rotate the `preferred' $z$ direction to whatever is the
longitudinal direction of the detector), and so should the translation
of particle codes. In particular, if the
detector simulation program also uses the standard Particle Data Group
codes, no conversion at all is needed. The problem then is to select
which particles are allowed to decay, and how decay vertex information
should be used.
Several switches regulate which particles are allowed to decay. First,
the master switch \ttt{MSTJ(21)} can be used to switch on/off all
decays (and it also contains a choice of how fragmentation should
be interfaced). Second, a particle must have decay modes defined for
it, i.e. the corresponding \ttt{MDCY(KC,2)} and \ttt{MDCY(KC,3)}
entries must be non-zero for compressed code \ttt{KC = LUCOMP(KF)}.
This is true for all colour neutral particles except the neutrinos,
the photon, the proton and the neutron. (This statement is actually
not fully correct, since irrelevant `decay modes' with
\ttt{MDME(IDC,2)=102} exist in some cases.) Third, the
individual switch in \ttt{MDCY(KC,1)} must be on. Of all the particles
with decay modes defined, only $\mu^{\pm}$, $\pi^{\pm}$, $\mathrm{K}^{\pm}$
and $\mathrm{K}_{\mrm{L}}^0$ are by default considered stable.
Finally, if \ttt{MSTJ(22)} does not have its default value 1, checks
are also made on the lifetime of a particle before it is allowed to
decay. In the simplest alternative, \ttt{MSTJ(22)=2}, the comparison
is based on the average lifetime, or rather $c \tau$, measured in mm.
Thus if the limit \ttt{PARJ(71)} is (the default) 10 mm, then decays
of $\mathrm{K}_{\mrm{S}}^0$, $\Lambda$, $\Sigma^-$, $\Sigma^+$, $\Xi^-$,
$\Xi^0$ and $\Omega^-$ are all switched off, but charm and bottom
still decay. No $c \tau$ values below 1 micron are defined. With the
two options \ttt{MSTJ(22)=} 3 or 4, a spherical or cylindrical volume
is defined around the origin, and all decays taking place inside this
volume are ignored.
Whenever a particle is in principle allowed to decay, i.e.
\ttt{MSTJ(21)} and \ttt{MDCY} on, an proper lifetime is selected
once and for all and stored in \ttt{V(I,5)}. The \ttt{K(I,1)} is then
also changed to 4. For \ttt{MSTJ(22)=1}, such a particle will also
decay, but else it could remain in the event record. It is then
possible, at a later stage, to expand the volume inside which decays
are allowed, and do a new \ttt{LUEXEC} call to have particles
fulfilling the new conditions (but not the old) decay. As a further
option, the \ttt{K(I,1)} code may be put to 5, signalling that the
particle will definitely decay in the next \ttt{LUEXEC} call, at the
vertex position given (by the user) in the \ttt{V} vector.
This then allows the {\tsc{Jetset}} decay routines to be used inside a
detector simulation program, as follows. For a particle which did not
decay before entering the detector, its point of decay is still well
defined (in the absence of deflections by electric or magnetic fields),
eq.~(\ref{dec:newvertex}). If it interacts before that point, the
detector simulation program is left to handle things. If not, the
\ttt{V} vector is updated according to the formula above, \ttt{K(I,1)}
is set to 5, and \ttt{LUEXEC} is called, to give a set of decay
products, that can again be tracked.
A further possibility is to force particles to decay into specific
decay channels; this may be particularly interesting for charm or
bottom physics. The choice of channels left open is determined by the
values of the switches \ttt{MDME(IDC,1)} for decay channel IDC
(use \ttt{LULIST(12)} to obtain the full listing). One or several
channels may be left open;
in the latter case effective branching ratios are automatically
recalculated without the need for your intervention. It is also
possible to differentiate between which channels are left open for
particles and which for antiparticles. Lifetimes are not affected by
the exclusion of some decay channels. Note that, whereas forced decays
can enhance the efficiency for several kinds of studies, it can
also introduce unexpected biases, in particular when events may contain
several particles with forced decays, cf. section
\ref{sss:resdecaycross}.
\subsubsection{Parameter values}
A non-trivial question is to know which parameter values to use. The
default values stored in the program are based on comparisons with
LEP $\e^+\e^- \to \mathrm{Z}^0$ data at around 91 GeV \cite{LEP90}, using a
parton-shower picture followed by string fragmentation. If
fragmentation is indeed an universal phenomenon, as
we would like to think, then the same parameters should also apply at
other energies and in other processes. The former aspect, at least,
seems to be borne out by comparisons with lower-energy PETRA/PEP
data. Note, however, that the choice of parameters is intertwined with
the choice of perturbative QCD description. If instead matrix elements
are used, a best fit to 30 GeV data would require the values
\ttt{PARJ(21)=0.40}, \ttt{PARJ(41)=1.0} and \ttt{PARJ(42)=0.7}.
With matrix elements one does not expect an energy
independence of the parameters, since the effective minimum invariant
mass cut-off is then energy dependent, i.e. so is the amount of soft
gluon emission effects lumped together with the fragmentation
parameters. This is indeed confirmed by the LEP data.
A mismatch in the perturbative QCD treatment could also
lead to small differences between different processes.
It is often said that the string fragmentation model contains a wealth
of parameters. This is certainly true, but it must be remembered that
most of these deal with flavour properties, and to a large extent
factorize from the treatment of the general event shape. In a fit to
the latter it is therefore usually enough to consider the parameters of
the perturbative QCD treatment, like $\Lambda$ in $\alpha_{\mathrm{s}}$ and a
shower cut-off $Q_0$ (or $\alpha_{\mathrm{s}}$ itself and $y_{\mathrm{min}}$, if matrix
elements are used), the $a$ and $b$ parameter of the Lund symmetric
fragmentation function (\ttt{PARJ(41)} and \ttt{PARJ(42)}) and the
width of the transverse momentum distribution
($\sigma = $\ttt{PARJ(21)}). In addition, the $a$ and $b$ parameters
are very strongly correlated by the requirement of having the correct
average multiplicity, such that in a typical $\chi^2$ plot, the
allowed region corresponds to a very narrow but very long valley,
stretched diagonally from small ($a$,$b$) pairs to large ones.
As to the flavour parameters, these are certainly many more, but most
of them are understood qualitiatively within one single framework, that
of tunnelling pair production of flavours.
Since the use of independent fragmentation has fallen in disrespect, it
should be pointed out that the default parameters here are not
particularly well tuned to the data. This especially applies if one,
in addition to asking for independent fragmentation, also asks for
another setup of fragmentation functions, i.e. other than the standard
Lund symmetric one. In particular, note that most fits to the popular
Peterson et al. (SLAC) heavy-flavour fragmentation function are based
on the actual observed spectrum. In a Monte Carlo simulation, one must
then start out with something harder, to compensate for the energy lost
by initial-state photon radiation and gluon bremsstrahlung. Since
independent fragmentation is not collinear safe (i.e, the emission of
a collinear gluon changes the properties of the final event), the tuning
is strongly dependent on the perturbative QCD treatment chosen. All the
parameters needed for a tuning of independent fragmentation are
available, however.
\subsubsection{Particle properties}
The masses of most frequently used particles are taken from tables.
For some rare charm and bottom hadrons, and for heavier flavour
hadrons, this would be unwieldy, and instead mass formulae are used,
based on the quark content. For the known quarks $\d$, $\u$, $\mathrm{s}$,
$\c$ and $\b$, the masses used for this purpose are actually the ones
stored in positions 101--105 in the \ttt{PARF} vector, rather than
the ones found in \ttt{PMAS}. This means that the \ttt{PMAS} masses
can be freely changed by you, to modify the masses that appear
in the event record, without courting disaster elsewhere (since mass
formulae typically contain $1/m$ terms from spin-spin splittings, it
is necessary to have the non-zero `constituent' masses here). Thus you
should never touch the mass values stored in \ttt{PARF}. For the
heavier flavours top, low and high, the current \ttt{PMAS}
values are always used. For these flavours, the only individually
defined hadrons are the flavour neutral $\eta$, $\Theta$,
$\mathrm{h}_1$, $\chi_0$, $\chi_1$ and $\chi_2$ states. A complete change
of top mass in the program thus requires changing \ttt{PMAS(6,1)},
\ttt{PMAS(LUCOMP(661),1)}, \ttt{PMAS(LUCOMP(663),1)},
\ttt{PMAS(LUCOMP(665),1)}, \ttt{PMAS(LUCOMP(10661),1)},
\ttt{PMAS(LUCOMP(10663),1)} and \ttt{PMAS(LUCOMP(20663),1)}.
Since the latter heavy-flavour-diagonal states
are not normally produced in fragmentation, it would be no
disaster to forget changing their masses.
Most particles have separately defined decay channels. However,
there are so many heavy-flavour hadrons with common decay desciptions,
that a few `pseudoparticles' have been introduced for generic
decays. The most frequently used ones are 84 for charm decays,
85 for bottom decays and 86 for top decays. Instead of a long list
of decay channels, several bottom and
charm baryons and all top hadrons therefore only have one
`decay channel', which is the instruction to make use of the decay
data for particle 84/85/86. The spectator quark of the generic
decay channels is found as the light quark/diquark of the hadron
considered. All the mesons in the bottom and charm sectors are
individually defined, as are the $\Lambda_{\c}$ and $\Lambda_{\b}$
states. For top and heavier hadrons, the decay is likely to be so
fast that no hadrons at all are produced, but if they are, the generic
pseudoparticle approach is a good approximation.
The program contains space so that additional new particles may be
introduced. Although not completely trivial, this should not be
beyond the ability of an ordinary user. Basically, three steps are
involved. First, a mechanism of production has to be introduced.
This production may well take place in another program, like
{\tsc{Pythia}} or some user-written correspondence, where matrix elements are
used to select the hard process. In this case the new particle already
exists in the \ttt{LUJETS} common block when {\tsc{Jetset}} is called. A new
particle, meson, baryon or glueball, may also be a part of the
fragmentation process, in which case \ttt{LUKFDI} would have to be
suitably modified. The particle might also appear as a decay product
from some already existing particle, and then the decay data in
\ttt{/LUDAT3/} would have to be expanded; conceivably also
\ttt{LUDECY} would be affected.
The second step is to teach to program to recognize the new particle.
If a KF code in the range 41 to 80 is used, this is automatically taken
care of, and in particular the compressed code KC coincides with KF.
If a whole sequence of particles is to be introduced, with KF codes
paralleling that of ordinary mesons/baryons (a supersymmetric `meson'
multiplet, made of a squark plus an antiquark, say), then \ttt{LUCOMP}
must be modified to include a mapping from these KF values to currently
unused KC ones, like the range 401 - 500. It is the presence of such a
mapping that the program uses to accept a given KF code as bona fide.
The third and final step is to define the properties of this new
particle. Thus particle charge information must be given in \ttt{KCHG},
mass, width and lifetime in \ttt{PMAS}, particle name in \ttt{CHAF},
and decay data in the \ttt{MDCY}, \ttt{MDME}, \ttt{BRAT} and \ttt{KFDP}
arrays. This process is most conveniently carried out by using
\ttt{LUUPDA(1)} to produce a table of particle data, which can then be
modified, and afterwards read back in with \ttt{LUUPDA(2)}. Note that
the particle data is to be introduced for the compressed code KC, not
for KF proper.
\subsection{Examples}
A 10 GeV $\u$ quark jet going out along the $+z$ axis is generated
with
\begin{verbatim}
CALL LU1ENT(0,2,10.,0.,0.)
\end{verbatim}
Note that such a single jet is not required to conserve energy,
momentum or flavour. In the generation scheme, particles with
negative $p_z$ are produced as well, but these are automatically
rejected unless \ttt{MSTJ(3)=-1}. While frequently used in former
days, the one-jet generation option is not of much current interest.
In e.g. a leptoproduction event a typical situation could be a $\u$
quark going out in the $+z$ direction and a $\u\d_0$ target remnant
essentially at rest. (Such a process can be simulated by {\tsc{Pythia}}, but
here we illustrate how to do part of it yourself.) The simplest
procedure is probably to treat the process in the c.m. frame
and boost it to the lab frame afterwards. Hence, if the c.m. energy is
20 GeV and the boost $\beta_z = 0.996$ (corresponding to
$x_B = 0.045$), then
\begin{verbatim}
CALL LU2ENT(0,2,2101,20.)
CALL LUROBO(0.,0.,0.,0.,0.996)
\end{verbatim}
The jets could of course also be defined and allowed to fragment in the
lab frame with
\begin{verbatim}
CALL LU1ENT(-1,2,223.15,0.,0.)
CALL LU1ENT(2,12,0.6837,3.1416,0.)
CALL LUEXEC
\end{verbatim}
Note here that the target diquark is required to move in the backwards
direction with $E-p_z = m_{\mathrm{p}}(1-x_B)$ to obtain the correct
invariant mass for the system. This is, however, only an artefact of
using a fixed diquark mass to represent a varying target remnant mass,
and is of no importance for the fragmentation. If one wants a
nicer-looking event record, it is possible to use the following
\begin{verbatim}
CALL LU1ENT(-1,2,223.15,0.,0.)
MSTU(10)=1
P(2,5)=0.938*(1.-0.045)
CALL LU1ENT(2,2101,0.,0.,0.)
MSTU(10)=2
CALL LUEXEC
\end{verbatim}
A 30 GeV $\u\overline{\mathrm{u}}\mathrm{g}$ event with $E_{\u} = 8$ GeV and
$E_{\overline{\mathrm{u}}} = 14$ GeV is simulated with
\begin{verbatim}
CALL LU3ENT(0,2,21,-2,30.,2.*8./30.,2.*14./30.)
\end{verbatim}
The event will be given in a standard orientation with the $\u$
quark along the $+z$ axis and the $\overline{\mathrm{u}}$ in the $-z, +x$ quadrant.
Note that the flavours of the three partons have to be given in the
order they are found along a string, if string fragmentation options
are to work. Also note that, for 3-jet events, and particularly
4-jet ones, not all setups of kinematical variables $x$ lie within
the kinematically allowed regions of phase space.
All common block variables can obviously be changed by including the
corresponding common block in the user-written main program.
Alternatively, the routine \ttt{LUGIVE} can be used to feed in
values, with some additional checks on array bounds then performed.
A call
\begin{verbatim}
CALL LUGIVE('MSTJ(21)=3;PMAS(C663,1)=210.;CHAF(401)=funnyino;'//
&'PMAS(21,4)=')
\end{verbatim}
will thus change the value of \ttt{MSTJ(21)} to 3, the value of
\ttt{PMAS(LUCOMP(663),1) = PMAS(136,1)} to 210., the value of
\ttt{CHAF(401)} to 'funnyino', and print the current value of
\ttt{PMAS(21,4)}. Since old and new values of parameters changed are
written to output, this may offer a convenient way of documenting
non-default values used in a given run. On the other hand, if a
variable is changed back and forth frequently, the resulting
voluminous output may be undesirable, and a direct usage
of the common blocks is then to be recommended (the output can also be
switched off, see \ttt{MSTU(13)}).
A general rule of thumb is that none of the physics routines
(\ttt{LUSTRF}, \ttt{LUINDF}, \ttt{LUDECY}, etc.) should ever be
called directly, but only via \ttt{LUEXEC}. This routine may be
called repeatedly for one single event. At each call only those
entries that are allowed to fragment or decay, and have not yet
done so, are treated. Thus
\begin{verbatim}
CALL LU2ENT(1,1,-1,20.) ! fill 2 jets without fragmenting
MSTJ(1)=0 ! inhibit jet fragmentation
MSTJ(21)=0 ! inhibit particle decay
MDCY(LUCOMP(111),1)=0 ! inhibit pi0 decay
CALL LUEXEC ! will not do anything
MSTJ(1)=1 !
CALL LUEXEC ! jets will fragment, but no decays
MSTJ(21)=2 !
CALL LUEXEC ! particles decay, except pi0
CALL LUEXEC ! nothing new can happen
MDCY(LUCOMP(111),1)=1 !
CALL LUEXEC ! pi0:s decay
\end{verbatim}
A partial exception to the rule above is \ttt{LUSHOW}. Its main
application is for internal use by \ttt{LUEEVT}, \ttt{LUDECY}, and
\ttt{PYEVNT}, but it can also be directly called by you. Note that a
special format for storing colour-flow information in \ttt{K(I,4)}
and \ttt{K(I,5)} must then be used. For simple cases, the \ttt{LU2ENT}
can be made to take care of that automatically, by calling with the
first argument negative.
\begin{verbatim}
CALL LU2ENT(-1,1,-2,80.) ! store d ubar with colour flow
CALL LUSHOW(1,2,80.) ! shower partons
CALL LUEXEC ! subsequent fragmentation/decay
\end{verbatim}
For more complicated configurations, \ttt{LUJOIN} should be used.
It is always good practice to list one or a few events during a run to
check that the program is working as intended. With
\begin{verbatim}
CALL LULIST(1)
\end{verbatim}
all particles will be listed and in addition total charge, momentum and
energy of stable entries will be given. For string fragmentation these
quantities should be conserved exactly (up to machine precision errors),
and the same goes when running independent fragmentation with one of
the momentum conservation options. \ttt{LULIST(1)} gives a format that
comfortably fits on an 80 column screen, at the price of not giving
the complete story. With \ttt{LULIST(2)} a more extensive listing is
obtained, and \ttt{LULIST(3)} also gives vertex information. Further
options are available, like \ttt{LULIST(12)}, which gives a list of
particle data.
An event, as stored in the \ttt{LUJETS} common block, will contain the
original jets and the whole decay chain, i.e. also particles which
subsequently decayed. If parton showers are used, the amount of parton
information is also considerable: first the on-shell partons before
showers have been considered, then a \ttt{K(I,1)=22} line with total
energy of the showering
subsystem, after that the complete shower history tree-like structure,
starting off with the same initial partons (now off-shell), and finally
the end products of the shower rearranged along the string directions.
This detailed record is useful in many connections, but if one only
wants to retain the final particles, superfluous information may be
removed with \ttt{LUEDIT}. Thus e.g.
\begin{verbatim}
CALL LUEDIT(2)
\end{verbatim}
will leave you with the final charged and neutral particles, except
for neutrinos.
The information in \ttt{LUJETS} may be used directly to study an event.
Some useful additional quantities derived from these, such as charge
and rapidity, may easily be found via the \ttt{KLU} and \ttt{PLU}
functions. Thus electric charge \ttt{=PLU(I,6)} (as integer,
three times charge \ttt{=KLU(I,6)}) and true rapidity $y$ with respect
to the $z$ axis \ttt{= PLU(I,17)}.
A number of utility (\ttt{MSTU}, \ttt{PARU}) and physics (\ttt{MSTJ},
\ttt{PARJ}) switches and parameters are available in common block
\ttt{LUDAT1}. All of these have sensible default values. Particle data
is stored in common blocks \ttt{LUDAT2}, \ttt{LUDAT3} and \ttt{LUDAT4}.
Note that the data in the arrays \ttt{KCHG}, \ttt{PMAS}, \ttt{MDCY}
and \ttt{CHAF} is not stored by KF code, but by the compressed
code KC. This code is not to be learnt by heart, but instead accessed
via the conversion function \ttt{LUCOMP}, \ttt{KC = LUCOMP(KF)}.
In the particle tables, the following particles are considered stable:
the photon, $\mathrm{e}^{\pm}$, $\mu^{\pm}$, $\pi^{\pm}$, $\mathrm{K}^{\pm}$,
$\mathrm{K}_{\mrm{L}}^0$,
$\mathrm{p}$, $\overline{\mathrm{p}}$, $\mathrm{n}$, $\br{\mathrm{n}}$ and all the neutrinos. It is, however,
always possible to inhibit the decay of any given particle by putting
the corresponding \ttt{MDCY} value zero or negative, e.g.
\ttt{MDCY(LUCOMP(310),1)=0} makes $\mathrm{K}_{\mrm{S}}^0$ and
\ttt{MDCY(LUCOMP(3122),1)=0} $\Lambda$ stable. It is also possible
to select stability based on the average lifetime (see \ttt{MSTJ(22)}),
or based on whether the decay takes place within a given spherical
or cylindrical volume around the origin.
The Field-Feynman jet model \cite{Fie78} is available in the program
by changing the following values: \ttt{MSTJ(1)=2} (independent
fragmentation), \ttt{MSTJ(3)=-1} (retain particles with $p_z < 0$;
is not mandatory), \ttt{MSTJ(11)=2} (choice of longitudinal
fragmentation function, with the $a$ parameter stored in
\ttt{PARJ(51) - PARJ(53)}), \ttt{MSTJ(12)=0} (no baryon production),
\ttt{MSTJ(13)=1} (give endpoint quarks $p_{\perp}$ as quarks created
in the field), \ttt{MSTJ(24)=0} (no mass broadening of resonances),
\ttt{PARJ(2)=0.5} ($\mathrm{s}/\u$ ratio for the production of new $\mathrm{q}\overline{\mathrm{q}}$
pairs), \ttt{PARJ(11)=PARJ(12)=0.5} (probability for mesons to
have spin 1) and \ttt{PARJ(21)=0.35} (width of Gaussian transverse
momentum distribution). In addition only $\d$, $\u$ and $\mathrm{s}$ single
quark jets may be generated following the FF recipe. Today the FF
`standard jet' concept is probably dead and buried, so the numbers
above should more be taken as an example of the flexibility of the
program, than as something to apply in practice.
A wide range of independent fragmentation options are implemented,
to be accessed with the master switch \ttt{MSTJ(1)=2}. In particular,
with \ttt{MSTJ(2)=1} a gluon jet is assumed to fragment like a
random $\d$, $\overline{\mathrm{d}}$, $\u$, $\overline{\mathrm{u}}$, $\mathrm{s}$ or $\overline{\mathrm{s}}$ jet, while with
\ttt{MSTJ(2)=3} the gluon is split into a $\d\overline{\mathrm{d}}$, $\u\overline{\mathrm{u}}$ or
$\mathrm{s}\overline{\mathrm{s}}$ pair of jets sharing the energy according to the
Altarelli-Parisi splitting function. Whereas energy, momentum and
flavour is not explicitly conserved in independent fragmentation, a
number of options are available in \ttt{MSTJ(3)} to ensure this
`post facto', e.g. \ttt{MSTJ(3)=1} will boost the event to ensure
momentum conservation and then (in the c.m. frame) rescale momenta by
a common factor to obtain energy conservation, whereas
\ttt{MSTJ(3)=4} rather uses a method of stretching the jets in
longitudinal momentum along the respective jet axis to keep angles
between jets fixed.
\clearpage
\section{Event Analysis Routines}
To describe the complicated geometries encountered in multihadronic
events, a number of event measures have been introduced. These
measures are intended to provide a global view of the properties
of a given event, wherein the full information content of the event
is condensed into one or a few numbers. A steady stream of such
measures are proposed for different purposes. Many are rather
specialized or never catch on, but a few become standards, and are
useful to have easy access to. \tsc{Jetset} therefore contains a
number of routines that can be called for any event, and that will
directly access the event record to extract the required information.
In the presentation below, measures have been grouped in three kinds.
The first contains simple event shape quantities, such as sphericity
and thrust. The second is jet finding algorithms.
The third is a mixed bag of particle multiplicities and compositions,
factorial moments and energy--energy correlations, put together
in a small statistics package.
None of the measures presented here are Lorentz invariant. The
analysis will be performed in whatever frame the event happens to
be given in. It it therefore up to you to decide whether the
frame in which events were generated is the right one, or whether
events beforehand should be boosted, e.g. to the c.m. frame. You
can also decide which particles you want to have affected by the
analysis.
\subsection{Event Shapes}
\label{ss:evshape}
In this section we study general event shape variables: sphericity,
thrust, Fox-Wolfram moments, and jet masses. These measures are
implemented in the routines \ttt{LUSPHE}, \ttt{LUTHRU}, \ttt{LUFOWO}
and \ttt{LUJMAS}, respectively.
Each event is assumed characterized by the particle four-momentum
vectors $p_i = (\mbf{p}_i, E_i)$, with $i = 1, 2, \cdots , n$ an
index running over the particles of the event.
\subsubsection{Sphericity}
The sphericity tensor is defined as \cite{Bjo70}
\begin{equation}
S^{\alpha \beta} = \frac{\displaystyle \sum_i p^{\alpha}_{i} \,
p^{\beta}_{i} } {\displaystyle \sum_i |\mbf{p}_i|^2 } ~,
\end{equation}
where $\alpha, \beta = 1, 2, 3$ corresponds to the $x$, $y$ and $z$
components.
By standard diagonalization of $S^{\alpha \beta}$ one may find three
eigenvalues $\lambda_1 \geq \lambda_2 \geq \lambda_3$, with
$\lambda_1 + \lambda_2 + \lambda_3 = 1$. The sphericity of the event
is then defined as
\begin{equation}
S = \frac{3}{2} \, (\lambda_2 + \lambda_3) ~,
\end{equation}
so that $0 \leq S \leq 1$. Sphericity is essentially a measure of the
summed $p_{\perp}^2$ with respect to the event axis; a 2-jet event
corresponds to $S \approx 0$ and an isotropic event to $S \approx 1$.
The aplanarity $A$, with definition $A = \frac{3}{2} \lambda_3$, is
constrained to the range $0 \leq A \leq \frac{1}{2}$. It measures the
transverse momentum component out of the event plane: a planar event
has $A \approx 0$ and an isotropic one $A \approx \frac{1}{2}$.
Eigenvectors $\mbf{v}_j$ can be found that correspond to the three
eigenvalues $\lambda_j$ of the sphericity tensor. The $\mbf{v}_1$
one is called the sphericity axis (or event axis, if it is clear
from the context that sphericity has been used), while the sphericity
event plane is spanned by $\mbf{v}_1$ and $\mbf{v}_2$.
The sphericity tensor is quadratic in particle momenta. This means
that the sphericity
value is changed if one particle is split up into two collinear ones
which share the original momentum. Thus sphericity is
not an infrared safe quantity in QCD perturbation theory. A useful
generalization of the sphericity tensor is
\begin{equation}
S^{(r) \alpha \beta} = \frac{\displaystyle \sum_i |\mbf{p}_i|^{r-2} \,
p^{\alpha}_{i} \, p^{\beta}_{i} }{\displaystyle
\sum_i |\mbf{p}_i|^r } ~,
\end{equation}
where $r$ is the power of the momentum dependence. While $r = 2$
thus corresponds to sphericity, $r = 1$ corresponds to linear
measures calculable in perturbation theory \cite{Par78}:
\begin{equation}
S^{(1) \alpha \beta} = \frac{\displaystyle \sum_i \frac{\displaystyle
p^{\alpha}_{i} \, p^{\beta}_{i}}{\displaystyle |\mbf{p}_i|} }
{\displaystyle \sum_i |\mbf{p}_i| } ~.
\end{equation}
Eigenvalues and eigenvectors may be defined exactly as before, and
therefore also equivalents of $S$ and $A$. These have no standard
names; I tend to call them linearized sphericity $S_{\mrm{lin}}$ and
linearized aplanarity $A_{\mrm{lin}}$. Quantities that are standard in
the literature are instead the combinations \cite{Ell81}
\begin{eqnarray}
C & = & 3 ( \lambda_1 \lambda_2 + \lambda_1 \lambda_3 +
\lambda_2 \lambda_3 ) ~, \\
D & = & 27 \lambda_1 \lambda_2 \lambda_3 ~.
\end{eqnarray}
Each of these is constrained to be in the range between 0 and 1.
Typically, $C$ is used to measure the 3-jet structure and $D$
the 4-jet one, since $C$ is vanishing for a perfect 2-jet event
and $D$ is vanishing for a planar event. The $C$ measure is related
to the second Fow-Wolfram moment (see below), $C = 1 - H_2$.
Noninteger $r$ values may also be used, and corresponding generalized
sphericity and aplanarity measures calculated. While perturbative
arguments favour $r = 1$, we know that the fragmentation `noise', e.g.
from transverse momentum fluctuations, is
proportionately larger for low momentum particles, and so $r > 1$
should be better for experimental event axis determinations. The use
of too large an $r$ value, on the other hand, puts all the emphasis
on a few high-momentum particles, and therefore involves a loss of
information. It should then come as no surprise that intermediate
$r$ values, of around 1.5, gives the best performance for event
axis determinations in 2-jet events, where the theoretical
meaning of the event axis is well-defined. The gain in accuracy
compared with the more conventional choices $r=2$ or $r=1$ is
rather modest, however.
\subsubsection{Thrust}
The quantity thrust $T$ is defined by \cite{Bra64}
\begin{equation}
T = \max_{|\mbf{n}| = 1} \,
\frac{\displaystyle \sum_i |\mbf{n} \cdot \mbf{p}_i|}
{\displaystyle \sum_i |\mbf{p}_i|} ~,
\label{em:thrust}
\end{equation}
and the thrust axis $\mbf{v}_1$ is given by the $\mbf{n}$ vector
for which maximum is attained. The allowed range is
$1/2 \leq T \leq 1$, with a 2-jet event corresponding
to $T \approx 1$ and an isotropic event to $T \approx 1/2$.
In passing, we note that this is not the only definition found in
the literature. The definitions agree for events studied in the c.m.
frame and where all particles are detected. However, a definition
like
\begin{equation}
T = 2 \, \max_{|\mbf{n}| = 1} \,
\frac{\displaystyle \left| \sum_i \theta(\mbf{n} \cdot \mbf{p}_i) \,
\mbf{p}_i \right| }{\displaystyle \sum_i |\mbf{p}_i|} =
2 \, \max_{\theta_i = 0,1} \,
\frac{\displaystyle \left| \sum_i \theta_i \, \mbf{p}_i \right| }
{\displaystyle \sum_i |\mbf{p}_i|}
\label{em:thrusttwo}
\end{equation}
(where $\theta(x)$ is the step function, $\theta(x) = 1 $ if $x > 0$,
else $\theta(x) = 0$) gives different results than the one above
if e.g. only charged particles are detected. It would even be possible
to have $T > 1$; to avoid such problems, often an extra fictitious
particle is introduced to balance the total momentum \cite{Bra79}.
Eq.~(\ref{em:thrust}) may be rewritten as
\begin{equation}
T = \max_{\epsilon_i = \pm 1} \,
\frac{\displaystyle \left| \sum_i \epsilon_i \, \mbf{p}_i \right| }
{\displaystyle \sum_i |\mbf{p}_i|} ~.
\label{em:thrustthree}
\end{equation}
(This may also be viewed as applying eq.~(\ref{em:thrusttwo}) to an
event with $2 n$ particles, $n$ carrying the momenta $\mbf{p}_i$
and $n$ the momenta $- \mbf{p}_i$, thus automatically balancing
the momentum.) To find the thrust value and axis this way, $2^{n-1}$
different possibilities would have to be tested. The reduction by a
factor of 2 comes from $T$ being unchanged when all
$\epsilon_i \to - \epsilon_i$. Therefore this approach rapidly becomes
prohibitive. Other exact methods exist, which `only' require about
$4n^2$ combinations to be tried.
In the implementation in {\tsc{Jetset}}, a faster alternative method is
used, in which the thrust axis is iterated from a starting direction
$\mbf{n}^{(0)}$ according to
\begin{equation}
\mbf{n}^{(j+1)} = \frac{\displaystyle \sum_i \epsilon(
\mbf{n}^{(j)} \cdot \mbf{p}_i) \, \mbf{p}_i }
{\displaystyle \left| \sum_i \epsilon(
\mbf{n}^{(j)} \cdot \mbf{p}_i) \, \mbf{p}_i \right| }
\end{equation}
(where $\epsilon(x) = 1$ for $x > 0$ and $\epsilon(x) = -1$ for
$x < 0$). It is easy to show that the related thrust value will
never decrease, $T^{(j+1)} \geq T^{(j)}$. In fact, the method
normally converges in 2--4 iterations. Unfortunately, this
convergence need not be towards the correct thrust axis but is
occasionally only towards a local maximum of the thrust function
\cite{Bra79}. We know of no foolproof way around this complication,
but the danger of an error may be lowered if several different
starting axes $\mbf{n}^{(0)}$ are tried and found to agree. These
$\mbf{n}^{(0)}$ are suitably constructed from the $n'$ (by default
4) particles with the largest momenta in the event, and the
$2^{n' -1}$ starting directions $\sum_i \epsilon_i \, \mbf{p}_i$
constructed from these are tried in falling order of the
corresponding absolute momentum values. When a predetermined number
of the starting axes have given convergence towards the same
(best) thrust axis this one is accepted.
In the plane perpendicular to the thrust axis, a major \cite{MAR79}
axis and value may be defined in just the same fashion as thrust, i.e.
\begin{equation}
M_a = \max_{|\mbf{n}| = 1, \, \mbf{n} \cdot \mbf{v}_1 = 0} \,
\frac{\displaystyle \sum_i |\mbf{n} \cdot \mbf{p}_i|}
{\displaystyle \sum_i |\mbf{p}_i| } ~.
\end{equation}
In a plane more efficient methods can be used to find an axis than in
three dimensions \cite{Wu79}, but for simplicity we use the same
method as above. Finally, a third axis, the minor axis, is defined
perpendicular to the thrust and major ones, and a minor value $M_i$
is calculated just as thrust and major.
The difference between major and minor is called
oblateness, $O = M_a -M_i$. The upper limit on oblateness depends
on the thrust value in a not so simple way. In general
$O \approx 0$ corresponds to an event symmetrical around the thrust
axis and high $O$ to a planar event.
As in the case of sphericity, a generalization to arbitrary momentum
dependence may easily be obtained, here by replacing the $\mbf{p}_i$
in the formulae above by $|\mbf{p}_i|^{r-1} \, \mbf{p}_i$. This
possibility is included, although so far it has not found any
experimental use.
\subsubsection{Fox-Wolfram moments}
The Fox-Wolfram moments $H_l$, $l = 0, 1, 2, \ldots$, are defined by
\cite{Fox79}
\begin{equation}
H_l = \sum_{i,j} \frac{ |\mbf{p}_i| \, |\mbf{p}_j| }
{E_{\mrm{vis}}^2} \, P_l (\cos \theta_{ij}) ~,
\end{equation}
where $\theta_{ij}$ is the opening angle between hadrons $i$ and
$j$ and $E_{\mrm{vis}}$ the total visible energy of the event. Note
that also autocorrelations, $i = j$, are included. The $P_l(x)$ are
the Legendre polynomials,
\begin{eqnarray}
P_0(x) & = & 1 ~, \nonumber \\
P_1(x) & = & x ~, \nonumber \\
P_2(x) & = & \frac{1}{2} \, (3x^2 -1) ~, \nonumber \\
P_3(x) & = & \frac{1}{2} \, (5x^3 - 3x) ~, \nonumber \\
P_4(x) & = & \frac{1}{8} \, (35x^4 - 30x^2 + 3) ~.
\end{eqnarray}
To the extent that particle masses may be neglected, $H_0 \equiv 1$.
It is customary to normalize the results to $H_0$, i.e. to give
$H_{l0} = H_l / H_0$. If momentum is balanced then $H_1 \equiv 0$.
2-jet events tend to give $H_l \approx 1$ for $l$ even and
$\approx 0$ for $l$ odd.
\subsubsection{Jet masses}
The particles of an event may be divided into two classes.
For each class a squared invariant mass may be calculated, $M_1^2$
and $M_2^2$. If the assignment of particles is adjusted such that
the sum $M_1^2 + M_2^2$ is minimized, the two masses thus obtained
are called heavy and light jet mass, $M_{\mrm{H}}$ and $M_{\mrm{L}}$.
It has been shown that these quantities are well behaved in perturbation
theory \cite{Cla79}. In $\e^+\e^-$ annihilation, the heavy jet mass
obtains a contribution from $\mathrm{q}\overline{\mathrm{q}}\mathrm{g}$ 3-jet events, whereas
the light mass is non-vanishing only when 4-jet events also
are included. In the c.m. frame of an event one has the limits
$0 \leq M_{\mrm{H}}^2 \leq E_{\mrm{cm}}^2/3$.
In general, the subdivision of particles tends to be into two
hemispheres, separated by a plane perpendicular to an event axis.
As with thrust, it is time-consuming to find the exact solution.
Different approximate strategies may therefore be used. In the
program, the sphericity axis is used to perform a fast subdivision
into two hemispheres, and thus into two preliminary jets. Thereafter
one particle at a time is tested to determine whether the sum
$M_1^2 + M_2^2$ would be decreased if that particle were to be
assigned to the other jet. The precedure is stopped when no
further significant change is obtained. Often the original
assignment is retained as it is, i.e. the sphericity axis gives
a good separation. This is not a full guarantee, since the program
might get stuck in a local mimimum which is not the global one.
\subsection{Cluster Finding}
\label{ss:clustfind}
Global event measures, like sphericity or thrust, can only be used to
determine the jet axes for back-to-back 2-jet events. To determine
the individual jet axes in events with three or more jets,
or with two (main) jets which are not back-to-back, cluster
algorithms are customarily used. In these, nearby particles are
grouped together into a variable number of clusters. Each cluster
has a well-defined direction, given by a suitably weighted average
of the constituent particle directions.
The cluster algorithms traditionally used in $\e^+\e^-$ and in $\p\p$
physics differ in several respects. The former tend to be
spherically symmetric, i.e. have no preferred axis in space,
and normally all particles have to be assigned to some jet. The
latter pick the beam axis as preferred direction, and make use
of variables related to this choice, such as rapidity and transverse
momentum; additionally only a fraction of all particles are assigned
to jets.
This reflects a difference in the underlying physics: in
$\p\p$ collisions, the beam remnants found at low transverse momenta
are not related to any hard processes, and therefore only provide an
unwanted noise to many studies. (Of course, also hard processes may
produce particles at low transverse momenta, but at a rate much
less than that from soft or semi-hard processes.) Further, the
kinematics of hard processes is, to a good approximation, factorized
into the hard subprocess itself, which is boost invariant in rapidity,
and parton-distribution effects, which determine the overall position
of a hard scattering in rapidity. Hence rapidity, azimuthal angle and
transverse momentum is a suitable coordinate frame to describe hard
processes in.
In standard $\e^+\e^-$ annihilation events, on the other hand, the hard
process c.m. frame tends to be almost at rest, and the event
axis is just about randomly distributed in space, i.e. with no
preferred r\^ole for the axis defined by the incoming $\mathrm{e}^{\pm}$.
All particle production is initiated by and related to the hard
subprocess. Some of the particles may be less easy to associate to
a specific jet, but there is no compelling reason to remove any
of them from consideration.
This does not mean that the separation above is always required.
$2\gamma$ events in $\e^+\e^-$ may have a structure with `beam jets' and
`hard scattering' jets, for which the $\p\p$ type algorithms might
be well suited. Conversely, a heavy particle produced in $\p\p$
collisions could profitably be studied, in its own rest frame,
with $\e^+\e^-$ techniques.
In the following, particles are only characterized by their
three-momenta or, alternatively, their energy and direction of
motion. No knowledge is therefore assumed of particle types, or
even of mass and charge. Clearly, the more is known, the more
sophisticated clustering algorithms can be used. The procedure
then also becomes more detector-dependent, and therefore less
suitable for general usage.
{\tsc{Jetset}} contains two cluster finding routines. \ttt{LUCLUS} is of
the $\e^+\e^-$ type and \ttt{LUCELL} of the $\p\p$ one. Each of them
allows some variations of the basic scheme.
\subsubsection{Cluster finding in an $\e^+\e^-$ type of environment}
The usage of cluster algorithms for $\e^+\e^-$ applications started in
the late 1970's. A number of different approaches were proposed
\cite{Dor81}. Of these, we will here only discuss those based on
binary joining. In this kind of
approach, initially each final-state particle is considered to be a
cluster. Using some distance measure, the two nearest clusters are
found. If their distance is smaller than some cut-off value,
the two clusters are joined into one. In this new configuration, the
two clusters that are now nearest are found and joined, and so on until
all clusters are separated by a distance larger than the cut-off.
The clusters remaining at the end are often also called jets. Note that,
in this approach, each single particle belongs to exactly one
cluster. Also note that the resulting jet picture explicitly depends
on the cut-off value used. Normally the number of clusters is allowed to
vary from event to event, but occasionally it is more useful to have
the cluster algorithm find a predetermined number of jets (like 3).
The obvious choice for a distance measure is to use squared invariant
mass, i.e. for two clusters $i$ and $j$ to define the
distance to be
\begin{equation}
m^2_{ij} = (E_i + E_j)^2 - (\mbf{p}_i + \mbf{p}_j)^2 ~.
\end{equation}
(Equivalently, one could have used the invariant mass as measure
rather than its square; this is just a matter of convenience.)
In fact, a number of people (including the author)
tried this measure long ago and gave up on it, since
it turns out to have severe instability problems. The reason is well
understood: in general, particles tend to cluster closer in
invariant mass in the region of small momenta. The clustering process
therefore tends to start in the center of the event, and only
subsequently spread outwards to encompass also the fast particles.
Rather than clustering slow particles around the fast ones (where the
latter na\"{\i}vely should best represent the jet directions), the
invariant mass measure will tend to cluster fast particles around
the slow ones.
Another instability may be seen by considering the clustering in a
simple 2-jet event. By the time that clustering has reached the
level of three clusters, the `best' the clustering algorithm can
possibly have achieved, in terms of finding three low-mass clusters,
is to have one fast cluster around each jet, plus a third slow cluster
in the middle. In the last step this third cluster would be joined with
one of the fast ones, to produce two final asymmetric clusters:
one cluster would contain all the slow particles, also those that
visually look like belonging to the opposite jet. A simple binary
joining process, with no possiblity to reassign particles between
clusters, is therefore not likely to be optimal.
The solution adopted by the author \cite{Sjo83} is to reject invariant
mass as distance measure. Instead a jet is defined as a collection
of particles which have a limited transverse momentum with respect to
a common jet axis, and hence also with respect to each other. This
picture is clearly inspired by the standard fragmentation picture,
e.g. in string fragmentation. A distance measure $d_{ij}$ between two
particles (or clusters) with momenta $\mbf{p}_i$ and $\mbf{p}_j$ should
thus not depend critically on the longitudinal momenta but only on the
relative transverse momentum. A number of such measures were tried,
and the one eventually selected is
\begin{equation}
d_{ij}^2 = \frac{1}{2} \, (|\mbf{p}_i| \, |\mbf{p}_j| -
\mbf{p}_i \cdot \mbf{p}_j) \, \frac{4 \, |\mbf{p}_i| \, |\mbf{p}_j|}
{(|\mbf{p}_i| + |\mbf{p}_j|)^2} =
\frac{4 \, |\mbf{p}_i|^2 \, |\mbf{p}_j|^2 \, \sin^2(\theta_{ij}/2)}
{(|\mbf{p}_i| + |\mbf{p}_j|)^2} ~.
\end{equation}
For small relative angle $\theta_{ij}$, where
$2 \sin(\theta_{ij}/2) \approx \sin\theta_{ij}$ and
$\cos \theta_{ij} \approx 1$, this measure reduces to
\begin{equation}
d_{ij} \approx \frac{|\mbf{p}_i \times \mbf{p}_j|}
{|\mbf{p}_i + \mbf{p}_j|} ~,
\end{equation}
where `$\times$' represents the cross product. We therefore see that
$d_{ij}$ in this limit has the simple physical interpretation as the
transverse momentum of either particle with respect to the direction
given by the sum of the two particle momenta. Unlike the approximate
expression, however, $d_{ij}$ does not vanish for two back-to-back
particles, but is here more related to the invariant mass
between them.
The basic scheme is of the binary joining type, i.e. initially each
particle is assumed to be a cluster by itself. Then the two clusters
with smallest relative distance $d_{ij}$ are found and, if
$d_{ij} < d_{\mrm{join}}$, with $d_{\mrm{join}}$
some predetermined distance, the two clusters are joined
to one, i.e. their momenta are added vectorially to give the
momentum of the new cluster. This is repeated until the distance between
any two clusters is $> d_{\mrm{join}}$. The number and momenta of these
final clusters then represent our reconstruction of the initial jet
configuration, and each particle is assigned to one of the clusters.
To make this scheme workable, two further ingredients are necessary,
however. Firstly, after two clusters have been joined, some particles
belonging to the new cluster may actually be closer to
another cluster. Hence, after each joining, all particles in the event
are reassigned to the closest of the clusters. For particle $i$, this
means that the distance $d_{ij}$ to all clusters $j$ in the event has
to be evaluated and compared. After all particles
have been considered, and only then, are cluster momenta recalculated
to take into account any reassignments. To save time, the assignment
procedure is not iterated until a stable configuration is reached,
but, since all particles are reassigned at each step, such an
iteration is effectively taking place in parallel with the cluster
joining. Only at the very end, when all $d_{ij} > d_{\mrm{join}}$, is
the reassignment procedure iterated to convergence --- still with the
possibility to continue the cluster joining if some $d_{ij}$ should
drop below $d_{\mrm{join}}$ due to the reassignment.
Occasionally, it may occur that the reassignment step leads to an empty
cluster, i.e. one to which no particles are assigned. Since such a
cluster has a distance $d_{ij} = 0$ to any other cluster, it is
automatically removed in the next cluster joining. However, it is
possible to run the program in a mode where a minimum number of jets
is to be reconstructed. If this minimum is reached with one cluster
empty, the particle is found which has largest distance to the cluster
it belongs to. That cluster is then split into two, namely the
large-distance particle and a remainder. Thereafter the reassignment
procedure is continued as before.
Secondly, the large multiplicities normally encountered means that,
if each particle initially is to be treated as a separate cluster, the
program will become very slow. Therefore a smaller number of clusters,
for a normal $\e^+\e^-$ event typically 8--12, is constructed as a starting
point for the iteration above, as follows. The particle with the
highest momentum is found, and thereafter all particles within a
distance $d_{ij} < d_{\mrm{init}}$ from it, where
$d_{\mrm{init}} \ll d_{\mrm{join}}$.
Together these are allowed to form a single cluster. For the
remaining particles, not assigned to this cluster, the procedure is
iterated, until all particles have been used up. Particles in the
central momentum region, $|\mbf{p}| < 2d_{\mrm{init}}$ are treated
separately; if their vectorial momentum sum is above $2d_{\mrm{init}}$
they are allowed to form one cluster, otherwise they are left
unassigned in the initial configuration. The value of $d_{\mrm{init}}$,
as long as reasonably small, has no physical importance, in that the
same final cluster configuration will be found as if each particle
initially is assumed to be a cluster by itself: the particles
clustered at this step are so nearby anyway that they almost
inevitably must enter the same jet; additionally the reassignment
procedure allows any possible `mistake' to be corrected in later
steps of the iteration.
Thus the jet reconstruction depends on one single parameter,
$d_{\mrm{join}}$, with a clearcut physical meaning of a transverse
momentum `jet-resolution power'. Neglecting smearing from
fragmentation, $d_{ij}$ between two clusters of equal energy
corresponds to half the invariant mass of the two original partons.
If one only wishes to reconstruct well separated jets, a large
$d_{\mrm{join}}$ should be chosen, while a small $d_{\mrm{join}}$
would allow the separation of close jets, at the
cost of sometimes artificially dividing a single jet into two. In
particular, $\b$ quark jets may here be a nuisance. The value of
$d_{\mrm{join}}$ to use for a fixed jet-resolution power in principle
should be independent of the c.m. energy of events, although
fragmentation effects may give a contamination of spurious extra jets
that increases slowly with $E_{\mrm{cm}}$ for fixed $d_{\mrm{join}}$.
Therefore a $d_{\mrm{join}} = 2.5$ GeV was acceptable at PETRA/PEP,
while 3--4 GeV may be better for applications at LEP and beyond.
This completes the description of the main option of the \ttt{LUCLUS}
routine. Variations are possible. One such is to skip the reassignment
step, i.e. to make use only of the simple binary joining procedure,
without any possibility to reassign particles between jets.
(This option is included mainly as a reference, to check how important
reassignment really is.) The other main alternative is
to replace the distance measure used above with the one used in the
JADE algorithm \cite{JAD86}.
The JADE cluster algorithm is an attempt to save the invariant mass
measure. The distance measure is defined to be
\begin{equation}
y_{ij} = \frac{2 E_i E_j (1-\cos \theta_{ij})}{E^2_{\mrm{vis}}} ~.
\end{equation}
Here $E_{\mrm{vis}}$ is the total visible energy of the event. The
usage of $E^2_{\mrm{vis}}$ in the denominator rather than
$E_{\mrm{cm}}^2$ tends to make the measure less
sensitive to detector acceptance corrections; in addition the
dimensionless nature of $y_{ij}$ makes it well suited for a comparison
of results at different c.m. energies. For the subsequent discussions,
this normalization will be irrelevant, however.
The $y_{ij}$ measure is very closely related to the squared mass
distance measure: the two coincide (up to the difference in
normalization) if $m_i = m_j = 0$. However, consider a pair of
particles or clusters with non-vanishing individual masses and a fixed
pair mass. Then, the larger the net momentum of the pair, the
smaller the $y_{ij}$ measure. This somewhat tends to favour clustering
of fast particles, and makes the algorithm less unstable than the one
based on true invariant mass.
The successes of the JADE
algorithm are well known: one obtains a very good
agreement between the number of partons generated on the matrix-element
(or parton-shower) level and the number of clusters reconstructed from
the hadrons, such that QCD aspects like the running of $\alpha_{\mathrm{s}}$ can
be studied with a minimal dependence on fragmentation effects. Of
course, the insensitivity to fragmentation effects depends on the choice
of fragmentation model. Fragmentation effects are small in the string
model, but not in independent fragmentation scenarios.
Although independent fragmentation in itself is not credible,
this may be seen as a signal for caution.
One should note that the JADE measure still suffers from some of the
diseases of the simple mass measure (without reassignments), namely
that particles which go in opposite directions may well be joined
into the same cluster. Therefore, while the JADE algorithm is a good
way to find the number of jets, it is inferior to the standard
$d_{ij}$ measure for a determination of jet directions and energies
\cite{Bet92}. The $d_{ij}$ measure also gives narrower jets, which
agree better with the visual impression of jet structure.
Recently, the `Durham algorithm' has been introduced \cite{Cat91},
which works as the JADE one but with a distance measure
\begin{equation}
\tilde{y}_{ij} = \frac{2 \min(E_i^2,E_j^2) (1-\cos \theta_{ij})}
{E^2_{cm}} ~.
\end{equation}
Like the $d_{ij}$ measure, this is a transverse momentum, but
$\tilde{y}_{ij}$ has the geometrical interpretation as the
transverse momentum of the softer particle with respect to the
direction of the harder one, while $d_{ij}$ is the transverse
momentum of either particle with respect to the common direction
given by the momentum vector sum. The two definitions agree when
one cluster is much softer than the other, so the soft gluon
exponentiation proven for the Durham measure also holds for the
$d_{ij}$ one.
The main difference therefore is that the standard \ttt{LUCLUS}
option allows reassignments, while the Durham algorithm does not.
The latter is therefore more easily calculable on the perturbative
parton level. This point is sometimes overstressed, and one could
give counterexamples why reassignments in fact may bring better
agreement with the underlying perturbative level. In particular,
without reassignments, one will make the recombination that seems
the `best' in the current step, even when that forces you to make
`worse' choices in subsequent steps. With reassignments, it is
possible to correct for mistakes due to the too local sensitivity of
a simple binary joining scheme.
\subsubsection{Cluster finding in a $\p\p$ type of environment}
The \ttt{LUCELL} cluster finding routines is of the kind pioneered
by UA1 \cite{UA183}, and commonly used in $\p\p$ physics. It is based
on a choice of pseudorapidity $\eta$, azimuthal angle $\varphi$ and
transverse momentum $p_{\perp}$ as the fundamental coordinates.
This choice is discussed in the introduction to cluster finding
above, with the proviso that the theoretically preferred true
rapidity has to be replaced by pseudorapidity, to make contact with
the real-life detector coordinate system.
A fix detector grid is assumed, with the pseudorapidity range
$|\eta| < \eta_{\mathrm{max}}$ and the full azimuthal range each divided
into a number of equally large bins, giving a rectangular grid.
The particles of an event impinge on this detector grid. For each
cell in ($\eta$,$\varphi$) space, the transverse momentum which enters
that cell is summed up to give a total cell $E_{\perp}$ flow.
Clearly the model remains very primitive in a number of
respects, compared with a real detector. There is no magnetic field
allowed for, i.e. also charged particles move in straight tracks.
The dimensions of the detector are not specified; hence the positions
of the primary vertex and any secondary vertices are neglected when
determining which cell a particle belongs to. The rest mass of
particles is not taken into account, i.e. what is used is really
$p_{\perp} = \sqrt{p_x^2 + p_y^2}$, while in a real detector some
particles would decay or annihilate, and then deposit additional
amounts of energy.
To take into account the energy resolution of the detector, it is
possible to smear the $E_{\perp}$ contents, bin by bin. This is done
according to a Gaussian, with a width assumed proportional to the
$\sqrt{E_{\perp}}$ of the bin. The Gaussian is cut off at zero and
at some predetermined multiple of the unsmeared $E_{\perp}$, by default
twice it. Alternatively, the smearing may be performed in $E$
rather than in $E_{\perp}$. To find the $E$, it is assumed
that the full energy of a cell is situated at its center, so
that one can translate back and forth with
$E = E_{\perp} \cosh\eta_{\mrm{center}}$.
The cell with largest $E_{\perp}$ is taken as a jet initiator if its
$E_{\perp}$ is above some threshold. A candidate jet is defined to
consist of all cells which are within some given radius $R$ in the
($\eta$,$\varphi$) plane, i.e. which have
$(\eta - \eta_{\mrm{initiator}})^2
+ (\varphi - \varphi_{\mrm{initiator}})^2 < R^2$.
Coordinates are always given with respect to the center of the
cell. If the summed $E_{\perp}$ of the jet is above the
required minimum jet energy, the candidate jet is accepted, and all
its cells are removed from further consideration. If not, the
candidate is rejected. The sequence is now repeated with the
remaining cell of highest $E_{\perp}$, and so on until no single
cell fulfills the jet initiator condition.
The number of jets reconstruced can thus vary from none to a maximum
given by purely geometrical considerations, i.e. how many circles of
radius $R$ are needed to cover the allowed ($\eta$,$\varphi$) plane.
Normally only a fraction of the particles are assigned to jets.
One could consider to iterate the jet assignment process, using the
$E_{\perp}$-weighted center of a jet to draw a new cirle of radius
$R$. In the current algorithm there is no such iteration step.
For an ideal jet assignment it would also be necessary to improve
the treatment when two jet circles partially overlap.
A final technical note. A natural implementation of a cell finding
algorithm is based on having a two-dimensional array of $E_{\perp}$
values, with dimensions to match the detector grid.
Very often most of the cells would then
be empty, in particular for low-multiplicity events in fine-grained
calorimeters. Our implementation is somewhat atypical, since cells are
only reserved space (contents and position) when they are shown to be
non-empty. This means that all non-empty cells have to be looped over
to find which are within the required distance $R$ of a potential jet
initiator. The algorithm is therefore faster than the ordinary kind
if the average cell occupancy is low, but slower if it is high.
\subsection{Event Statistics}
All the event-analysis routines above are defined on an
event-by-event basis. Once found, the quantities are about equally
often used to define inclusive distributions as to select specific
classes of events for continued study. For instance, the thrust
routine might be used either to find the inclusive $T$ distribution
or to select events with $T < 0.9$. Other measures, although still
defined for the individual event, only make sense to discuss in
terms of averages over many events. A small set of such measures
is found in \ttt{LUTABU}. This routine
has to be called once after each event to accumulate statistics,
and once in the end to print the final tables. Of course, among the
wealth of possibilities imaginable, the ones collected here are only
a small sample, selected because the author at some point has found
a use for them himself.
\subsubsection{Multiplicities}
Three options are available to collect information on multiplicities
in events. One gives the flavour content of the final state in hard
interaction processes, e.g. the relative composition of
$\d\overline{\mathrm{d}} / \u\overline{\mathrm{u}} / \mathrm{s}\overline{\mathrm{s}} / \c\overline{\mathrm{c}} / \b\overline{\mathrm{b}}$ in $\e^+\e^-$
annihilation events. Additionally it gives the total parton multiplicity
distribution at the end of parton showering. Another gives the
inclusive rate of all the different particles produced in events,
either as intermediate resonances or as final-state particles.
The number is subdivided into particles produced from fragmentation
(primary particles) and those produced in decays (secondary
particles).
The third option tabulates the rate of exclusive final states, after
all allowed decays have occurred. Since only events with up to 8
final-state particles are analyzed, this is clearly not intended for the
study of complete high-energy events. Rather the main application is
for an analysis of the decay modes of a single particle. For instance,
the decay data for $\mathrm{D}$ mesons is given in terms of channels that
also contain unstable particles, such as $\rho$ and $\eta$, which
decay further. Therefore a given final state may receive
contributions from several tabulated decay channels; e.g.
$\mathrm{K} \pi \pi$ from $\mathrm{K}^* \pi$ and $\mathrm{K} \rho$, and so on.
\subsubsection{Energy-Energy Correlation}
The Energy-Energy Correlation is defined by \cite{Bas78}
\begin{equation}
\mrm{EEC}(\theta) = \sum_{i < j}
\frac{2 E_i E_j}{E_{\mrm{vis}}^2} \,
\delta(\theta - \theta_{ij}) ~,
\end{equation}
and its Asymmetry by
\begin{equation}
\mrm{EECA}(\theta) = \mrm{EEC}(\pi - \theta) - \mrm{EEC}(\theta) ~.
\end{equation}
Here $\theta_{ij}$ is the opening angle between the two particles
$i$ and $j$, with energies $E_i$ and $E_j$. In principle,
normalization should be to $E_{\mrm{cm}}$, but if not all particles
are detected it is convenient to normalize to the total visible
energy $E_{\mrm{vis}}$. Taking into account the autocorrelation term
$i = j$, the total $\mrm{EEC}$ in an event then is unity.
The $\delta$ function peak is smeared out by the
finite bin width $\Delta \theta$ in the histogram, i.e.,
it is replaced by a contribution $1 / \Delta \theta$ to the bin
which contains $\theta_{ij}$.
The formulae above refer to an individual event, and are to be
averaged over all events to suppress statistical fluctuations, and
obtain smooth functions of $\theta$.
\subsubsection{Factorial moments}
Factorial moments may be used to search for intermittency in events
\cite{Bia86}. The whole field has been much studied in recent years,
and a host of different measures have been proposed. We only implement
one of the original prescriptions.
To calculate the factorial moments, the full rapidity (or
pseudorapidity) and
azimuthal ranges are subdivided into bins of successively smaller
size, and the multiplicity distributions in bins is studied. The
program calculates pseudorapidity with respect to the $z$ axis;
if desired, one could first find
an event axis, e.g. the sphericity or thrust axis, and subsequently
rotate the event to align this axis with the $z$ direction.
The full rapidity range $|y| < y_{\mathrm{max}}$ (or pseudorapidity range
$|\eta| < \eta_{\mathrm{max}}$) and
azimuthal range $0 < \varphi < 2\pi$ are subdivided into $m_y$ and
$m_{\varphi}$ equally large bins. In fact, the whole analysis is
performed thrice: once with $m_{\varphi}=1$ and the $y$ (or $\eta$)
range gradually
divided into 1, 2, 4, 8, 16, 32, 64, 128, 256 and 512 bins, once
with $m_y = 1$ and the $\varphi$ range subdivided as above, and
finally once with $m_y = m_{\varphi}$ according to the same binary
sequence. Given the multiplicity $n_j$ in bin $j$, the $i$:th
factorial moment is defined by
\begin{equation}
F_i = (m_y m_{\varphi})^{i-1} \, \sum_j
\frac{n_j(n_j-1)\cdots(n_j-i+1)}{n(n-1)\cdots(n-i+1)} ~.
\end{equation}
Here $n = \sum_j n_j$ is the total multiplicity of the event within
the allowed $y$ (or $\eta$) limits. The calculation is performed for
the second through the fifth moments, i.e. $F_2$ through $F_5$.
The $F_i$ as given here are defined for the individual event,
and have to be averaged over many events to give a reasonably smooth
behaviour. If particle production is uniform and uncorrelated
according to Poissonian statistics, one expects
$\langle F_i \rangle \equiv 1$ for all moments and all bin sizes.
If, on the other hand, particles are locally clustered, factorial
moments should increase when bins are made smaller, down to the
characteristic dimensions of the clustering.
\subsection{Routines and Common Block Variables}
\label{ss:evanrout}
The six routines \ttt{LUSPHE}, \ttt{LUTHRU}, \ttt{LUCLUS},
\ttt{LUCELL}, \ttt{LUJMAS} and \ttt{LUFOWO} give you the possibility
to find some global event shape properties. The routine \ttt{LUTABU}
performs a statistical analysis of a number of different quantities
like particle content, factorial moments and the energy--energy
correlation.
Note that, by default, all remaining partons/particles except
neutrinos are used in the analysis. Neutrinos may be included with
\ttt{MSTU(41)=1}. Also note that axes determined are stored in
\ttt{LUJETS}, but are not proper four-vectors and, as a general rule
(with some exceptions), should therefore not be rotated or boosted.
\drawbox{CALL LUSPHE(SPH,APL)}\label{p:LUSPHE}
\begin{entry}
\itemc{Purpose:} to diagonalize the momentum tensor, i.e. find the
eigenvalues $\lambda_1 > \lambda_2 > \lambda_3$, with sum unity,
and the corresponding eigenvectors. \\
Momentum power dependence is given by \ttt{PARU(41)}; default
corresponds to sphericity, \ttt{PARU(41)=1.} gives measures linear
in momenta. Which particles (or partons) are used in the analysis
is determined by the \ttt{MSTU(41)} value.
\iteme{SPH :} $\frac{3}{2} (\lambda_2 + \lambda_3)$, i.e. sphericity
(for \ttt{PARU(41)=2.}).
\begin{subentry}
\iteme{= -1. :} analysis not performed because event contained less
than two particles (or two exactly back-to-back particles, in
which case the two transverse directions would be undefined).
\end{subentry}
\iteme{APL :} $\frac{3}{2} \lambda_3$, i.e. aplanarity (for
\ttt{PARU(41)=2.}).
\begin{subentry}
\iteme{= -1. :} as \ttt{SPH=-1.}.
\end{subentry}
\itemc{Remark:} the lines \ttt{N+1} through \ttt{N+3} (\ttt{N-2}
through \ttt{N} for \ttt{MSTU(43)=2}) in \ttt{LUJETS} will, after
a call, contain the following information: \\
\ttt{K(N+i,1) =} 31; \\
\ttt{K(N+i,2) =} 95; \\
\ttt{K(N+i,3) :} $i$, the axis number, $i=1,2,3$; \\
\ttt{K(N+i,4), K(N+i,5) =} 0; \\
\ttt{P(N+i,1) - P(N+i,3) :} the $i$'th eigenvector, $x$, $y$ and $z$
components; \\
\ttt{P(N+i,4) :} $\lambda_i$, the $i$'th eigenvalue; \\
\ttt{P(N+i,5) =} 0; \\
\ttt{V(N+i,1) - V(N+i,5) =} 0. \\
Also, the number of particles used in the analysis is given in
\ttt{MSTU(62)}.
\end{entry}
\drawbox{CALL LUTHRU(THR,OBL)}\label{p:LUTHRU}
\begin{entry}
\itemc{Purpose:} to find the thrust, major and minor axes and
corresponding projected momentum quantities, in particular thrust
and oblateness. The performance of the program is affected by
\ttt{MSTU(44)}, \ttt{MSTU(45)}, \ttt{PARU(42)} and \ttt{PARU(48)}.
In particular, \ttt{PARU(42)} gives the momentum dependence, with
the default value \ttt{=1.} corresponding to linear dependence.
Which particles (or partons) are used in the analysis
is determined by the \ttt{MSTU(41)} value.
\iteme{THR :} thrust (for \ttt{PARU(42)=1.}).
\begin{subentry}
\iteme{= -1. :} analysis not performed because event contained
less than two particles.
\iteme{= -2. :} remaining space in \ttt{LUJETS} (partly used as
working area) not large enough to allow analysis.
\end{subentry}
\iteme{OBL :} oblateness (for \ttt{PARU(42)=1.}).
\begin{subentry}
\iteme{= -1., -2. :} as for \ttt{THR}.
\end{subentry}
\itemc{Remark:} the lines \ttt{N+1} through \ttt{N+3} (\ttt{N-2}
through \ttt{N} for \ttt{MSTU(43)=2}) in \ttt{LUJETS} will, after
a call, contain the following information: \\
\ttt{K(N+i,1) =} 31; \\
\ttt{K(N+i,2) =} 96; \\
\ttt{K(N+i,3) :} $i$, the axis number, $i=1,2,3$; \\
\ttt{K(N+i,4), K(N+i,5) =} 0; \\
\ttt{P(N+i,1) - P(N+i,3) :} the thrust, major and minor axis,
respectively, for $i = 1, 2$ and 3; \\
\ttt{P(N+i,4) :} corresponding thrust, major and minor value; \\
\ttt{P(N+i,5) =} 0; \\
\ttt{V(N+i,1) - V(N+i,5) =} 0. \\
Also, the number of particles used in the analysis is given in
\ttt{MSTU(62)}.
\end{entry}
\drawbox{CALL LUCLUS(NJET)}\label{p:LUCLUS}
\begin{entry}
\itemc{Purpose:} to reconstruct an arbitrary number of jets using a
cluster analysis method based on particle momenta. \\
Two different distance measures are available, see section
\ref{ss:clustfind}. The choice is controlled by \ttt{MSTU(46)}. The
distance scale $d_{\mrm{join}}$, above which two clusters may not be
joined, is normally given by \ttt{PARU(44)}. In general,
$d_{\mrm{join}}$
may be varied to describe different `jet-resolution powers';
the default value, 2.5 GeV, is fairly well suited for $\e^+\e^-$ physics
at 30--40 GeV. With the alternative mass distance measure,
\ttt{PARU(44)} can be used to set the absolute maximum cluster mass,
or \ttt{PARU(45)} to set the scaled one, i.e. in
$y = m^2/E_{\mrm{cm}}^2$, where $E_{\mrm{cm}}$ is the total
invariant mass of the particles being considered. \\
It is possible to continue the cluster search from the configuration
already found, with a new higher $d_{\mrm{join}}$ scale, by selecting
\ttt{MSTU(48)} properly. In \ttt{MSTU(47)} one can also require a
minimum number of jets to be reconstructed; combined with an
artificially large $d_{\mrm{join}}$ this can be used to reconstruct a
predetermined number of jets. \\
Which particles (or partons) are used in the analysis is determined
by the \ttt{MSTU(41)} value, whereas assumptions about particle masses
is given by \ttt{MSTU(42)}. The parameters \ttt{PARU(43)} and
\ttt{PARU(48)} regulate more technical details (for events at high
energies and large multiplicities, however, the choice of a larger
\ttt{PARU(43)} may be necessary to obtain reasonable reconstruction
times).
\iteme{NJET :} the number of clusters reconstructed.
\begin{subentry}
\iteme{= -1 :} analysis not performed because event contained less than
\ttt{MSTU(47)} (normally 1) particles, or analysis failed to
reconstruct the requested number of jets.
\iteme{= -2 :} remaining space in \ttt{LUJETS} (partly used as working
area) not large enough to allow analysis.
\end{subentry}
\itemc{Remark:} if the analysis does not fail, further information is
found in \ttt{MSTU(61) - MSTU(63)} and \ttt{PARU(61) - PARU(63)}.
In particular, \ttt{PARU(61)} contains the invariant mass for the
system analyzed, i.e. the number used in determining the denominator
of $y = m^2/E_{\mrm{cm}}^2$. \ttt{PARU(62)} gives the generalized
thrust, i.e. the sum of (absolute values of) cluster momenta divided
by the sum of particle momenta (roughly the same as multicity).
\ttt{PARU(63)} gives the minimum distance $d$ (in $p_{\perp}$ or $m$)
between two clusters in the final cluster configuration, 0 in case
of only one cluster. \\
Further, the lines \ttt{N+1} through \ttt{N+NJET} (\ttt{N-NJET+1}
through \ttt{N} for \ttt{MSTU(43)=2}) in \ttt{LUJETS}
will, after a call, contain the following information: \\
\ttt{K(N+i,1) =} 31; \\
\ttt{K(N+i,2) =} 97; \\
\ttt{K(N+i,3) :} $i$, the jet number, with the jets arranged in
falling order of absolute momentum; \\
\ttt{K(N+i,4) :} the number of particles assigned to jet $i$; \\
\ttt{K(N+i,5) =} 0; \\
\ttt{P(N+i,1) - P(N+i,5) :} momentum, energy and invariant mass of
jet $i$; \\
\ttt{V(N+i,1) - V(N+i,5) =} 0. \\
Also, for a particle which was used in the analysis,
\ttt{K(I,4)}$ = i$, where \ttt{I} is the particle number and $i$
the number of the jet it has ben assigned to. Undecayed particles
not used then have \ttt{K(I,4)=0}. An exception is made for lines
with \ttt{K(I,1)=3} (which anyhow are not normally interesting for
cluster search), where the colour-flow information stored in
\ttt{K(I,4)} is left intact.
\end{entry}
\drawbox{CALL LUCELL(NJET)}\label{p:LUCELL}
\begin{entry}
\itemc{Purpose:} to provide a simpler cluster routine more in line
with what is currently used in the study of high-$p_{\perp}$ collider
events. \\
A detector is assumed to stretch in pseudorapidity between
\ttt{-PARU(51)} and \ttt{+PARU(51)} and be segmented in
\ttt{MSTU(51)} equally large $\eta$ (pseudorapidity) bins and
\ttt{MSTU(52)} $\varphi$ (azimuthal) bins. Transverse
energy $E_{\perp}$ for undecayed entries are summed up in each bin.
For \ttt{MSTU(53)} non-zero, the energy is smeared by calorimetric
resolution effects, cell by cell. This is done according to a Gaussian
distribution; if \ttt{MSTU(53)=1} the standard deviation for the
$E_{\perp}$ is \ttt{PARU(55)}$\times \sqrt{E_{\perp}}$, if
\ttt{MSTU(53)=2} the standard deviation for the $E$ is
\ttt{PARU(55)}$\times \sqrt{E}$, $E_{\perp}$ and $E$ expressed in GeV.
The Gaussian is cut off at 0 and at a factor \ttt{PARU(56)} times the
correct $E_{\perp}$ or $E$. Cells with an $E_{\perp}$ below a given
threshold \ttt{PARU(58)} are removed from further consideration;
by default \ttt{PARU(58)=0.} and thus all cells are kept.
\\
All bins with $E_{\perp} > $\ttt{PARU(52)} are taken to be possible
initiators of jets, and are tried in falling $E_{\perp}$ sequence to
check whether the total $E_{\perp}$ summed over cells no more distant
than \ttt{PARU(54)} in $\sqrt{(\Delta\eta)^2 + (\Delta\varphi)^2}$
exceeds \ttt{PARU(53)}. If so, these cells define one jet, and are
removed from further consideration. Contrary to \ttt{LUCLUS}, not all
particles need be assigned to jets. Which particles (or partons) are
used in the analysis is determined by the \ttt{MSTU(41)} value.
\iteme{NJET :} the number of jets reconstructed (may be 0).
\begin{subentry}
\iteme{= -2 :} remaining space in \ttt{LUJETS} (partly used as
working area) not large enough to allow analysis.
\end{subentry}
\itemc{Remark:} the lines \ttt{N+1} through \ttt{N+NJET}
(\ttt{N-NJET+1} through \ttt{N} for \ttt{MSTU(43)=2}) in
\ttt{LUJETS} will, after a call, contain the following information: \\
\ttt{K(N+i,1) =} 31; \\
\ttt{K(N+i,2) =} 98; \\
\ttt{K(N+i,3) :} $i$, the jet number, with the jets arranged in
falling order in $E_{\perp}$; \\
\ttt{K(N+i,4) :} the number of particles assigned to jet $i$; \\
\ttt{K(N+i,5) =} 0; \\
\ttt{V(N+i,1) - V(N+i,5) =} 0. \\
Further, for \ttt{MSTU(54)=1} \\
\ttt{P(N+i,1), P(N+i,2) =} position in $\eta$ and $\varphi$ of the
center of the jet initiator cell, i.e. geometrical center of jet; \\
\ttt{P(N+i,3), P(N+i,4) =} position in $\eta$ and $\varphi$ of the
$E_{\perp}$-weighted center of the jet, i.e. the center of gravity
of the jet; \\
\ttt{P(N+i,5) =} sum $E_{\perp}$ of the jet; \\
while for \ttt{MSTU(54)=2} \\
\ttt{P(N+i,1) - P(N+i,5) :} the jet momentum vector, constructed
from the summed $E_{\perp}$ and the $\eta$ and $\varphi$ of the
$E_{\perp}$-weighted center of the jet as \\
$(p_x, p_y, p_z, E, m)=E_{\perp} (\cos\varphi, \sin\varphi,
\sinh\eta, \cosh\eta, 0)$; \\
and for \ttt{MSTU(54)=3} \\
\ttt{P(N+i,1) - P(N+i,5) :} the jet momentum vector, constructed by
adding vectorially the momentum of each cell assigned to the jet,
assuming that all the $E_{\perp}$ was deposited at the center of the
cell, and with the jet mass in \ttt{P(N+i,5)} calculated from the
summed $E$ and $\mbf{p}$ as $m^2 = E^2 - p_x^2 - p_y^2 - p_z^2$. \\
Also, the number of particles used in the analysis is given in
\ttt{MSTU(62)}, and the number of cells hit in \ttt{MSTU(63)}.
\end{entry}
\drawbox{CALL LUJMAS(PMH,PML)}\label{p:LUJMAS}
\begin{entry}
\itemc{Purpose:} to reconstruct high and low jet mass of an event.
A simplified algorithm is used, wherein a preliminary division of
the event into two hemispheres is done transversely to the sphericity
axis. Then one particle at a time is reassigned to the other
hemisphere if that reduces the sum of squares of the two jet masses,
$m_{\mrm{H}}^2 + m_{\mrm{L}}^2$. The procedure is stopped when no
further significant change (see \ttt{PARU(48)}) is obtained. Often, the
original assignment is retained as it is. Which particles (or partons)
used in the analysis is determined by the \ttt{MSTU(41)} value,
whereas assumptions about particle masses is given by \ttt{MSTU(42)}.
\iteme{PMH :} heavy jet mass (in GeV).
\begin{subentry}
\iteme{= -2. :} remaining space in \ttt{LUJETS} (partly used as
working area) not large enough to allow analysis.
\end{subentry}
\iteme{PML :} light jet mass (in GeV).
\begin{subentry}
\iteme{= -2. :} as for \ttt{PMH=-2.}.
\end{subentry}
\itemc{Remark:} After a successful call, \ttt{MSTU(62)} contains the
number of particles used in the analysis, and \ttt{PARU(61)} the
invariant mass of the system analyzed. The latter number is helpful
in constructing scaled jet masses.
\end{entry}
\drawbox{CALL LUFOWO(H10,H20,H30,H40)}\label{p:LUFOWO}
\begin{entry}
\itemc{Purpose:} to do an event analysis in terms of the Fox-Wolfram
moments. The moments $H_i$ are normalized to the lowest one, $H_0$.
Which particles (or partons) are used in the analysis is determined
by the \ttt{MSTU(41)} value.
\iteme{H10 :} $H_1/H_0$. Is $=0$ if momentum is balanced.
\iteme{H20 :} $H_2/H_0$.
\iteme{H30 :} $H_3/H_0$.
\iteme{H40 :} $H_4/H_0$.
\itemc{Remark:} the number of particles used in the analysis is given
in \ttt{MSTU(62)}.
\end{entry}
\drawbox{CALL LUTABU(MTABU)}\label{p:LUTABU}
\begin{entry}
\itemc{Purpose:} to provide a number of event-analysis options which
can be be used on each new event, with accumulated statistics to be
written out on request. When errors are quoted, these refer to
the uncertainty in the average value for the event sample as a
whole, rather than to the spread of the individual events, i.e.
errors decrease like one over the square root of the number of
events analyzed. For a correct use of \ttt{LUTABU}, it is not
permissible to freely mix generation and analysis of different
classes of events, since only one set of statistics counters exists.
A single run may still contain sequential `subruns', between
which statistics is reset. Whenever an event is analyzed, the
number of particles/partons used is given in \ttt{MSTU(62)}.
\iteme{MTABU :} determines which action is to be taken. Generally, a
last digit equal to 0 indicates that the statistics counters for this
option is to be reset; since the counters are reset (by \ttt{DATA}
statements) at the beginning of a run, this is not used normally. Last
digit 1 leads to an analysis of current event with respect to the
desired properties. Note that the resulting action may depend on how
the event generated has been rotated, boosted or edited before this
call. The statistics accumulated is output in tabular form with
last digit 2, while it is dumped in the \ttt{LUJETS} common block for
last digit 3. The latter option may be useful for interfacing to
graphics output.
\begin{subentry}
\iteme{= 10 :} statistics on parton multiplicity is reset.
\iteme{= 11 :} the parton content of the current event is analyzed,
classified according to the flavour content of the hard
interaction and the total number of partons. The flavour
content is assumed given in \ttt{MSTU(161)} and \ttt{MSTU(162)};
these are automatically set e.g. in \ttt{LUEEVT} and \ttt{PYEVNT}
calls.
\iteme{= 12 :} gives a table on parton multiplicity distribution.
\iteme{= 13 :} stores the parton multiplicity distribution of events
in \ttt{/LUJETS/}, using the following format: \\
\ttt{N =} total number of different channels found; \\
\ttt{K(I,1) =} 32; \\
\ttt{K(I,2) =} 99; \\
\ttt{K(I,3), K(I,4) =} the two flavours of the flavour content; \\
\ttt{K(I,5) =} total number of events found with flavour content of
\ttt{K(I,3)} and \ttt{K(I,4)}; \\
\ttt{P(I,1) - P(I,5) =} relative probability to find given flavour
content and a total of 1, 2, 3, 4 or 5 partons, respectively; \\
\ttt{V(I,1) - V(I,5) =} relative probability to find given flavour
content and a total of 6--7, 8--10, 11--15, 16--25 or above 25
partons, respectively. \\
In addition, \ttt{MSTU(3)=1} and \\
\ttt{K(N+1,1) =} 32; \\
\ttt{K(N+1,2) =} 99; \\
\ttt{K(N+1,5) =} number of events analyzed.
\iteme{= 20 :} statistics on particle content is reset.
\iteme{= 21 :} the particle/parton content of the current event is
analyzed, also for particles which have subsequently decayed and
partons which have fragmented (unless this has been made impossible
by a preceding \ttt{LUEDIT} call). Particles are subdivided into
primary and secondary ones, the main principle being that primary
particles are those produced in the fragmentation of a string,
while secondary come from decay of other particles. Since
particles (top, say), may decay into partons, the distinction
is not always unique.
\iteme{= 22 :} gives a table of particle content in events.
\iteme{= 23 :} stores particle content in events in \ttt{/LUJETS/},
using the following format: \\
\ttt{N =} number of different particle species found; \\
\ttt{K(I,1) =} 32; \\
\ttt{K(I,2) =} 99; \\
\ttt{K(I,3) =} particle KF code; \\
\ttt{K(I,5) =} total number of particles and antiparticles of this
species; \\
\ttt{P(I,1) =} average number of primary particles per event; \\
\ttt{P(I,2) =} average number of secondary particles per event; \\
\ttt{P(I,3) =} average number of primary antiparticles per event; \\
\ttt{P(I,4) =} average number of secondary antiparticles per event; \\
\ttt{P(I,5) =} average total number of particles or antiparticles
per event. \\
In addition, \ttt{MSTU(3)=1} and \\
\ttt{K(N+1,1) =} 32; \\
\ttt{K(N+1,2) =} 99; \\
\ttt{K(N+1,5) =} number of events analyzed; \\
\ttt{P(N+1,1) =} average primary multiplicity per event; \\
\ttt{P(N+1,2) =} average final multiplicity per event; \\
\ttt{P(N+1,3) =} average charged multiplicity per event.
\iteme{= 30 :} statistics on factorial moments is reset.
\iteme{= 31 :} analyzes the factorial moments of the multiplicity
distribution in different bins of rapidity and azimuth.
Which particles (or partons) are used in the analysis is
determined by the \ttt{MSTU(41)} value. The selection between usage
of true rapidity, pion rapidity or pseudorapidity is regulated
by \ttt{MSTU(42)}. The $z$ axis is assumed to be event axis; if this
is not desirable find an event axis e.g. with \ttt{LUSPHE} or
\ttt{LUTHRU} and use \ttt{LUEDIT(31)}. Maximum (pion-, pseudo-)
rapidity, which sets the limit for the rapidity plateau or the
experimental acceptance, is given by \ttt{PARU(57)}.
\iteme{= 32 :} prints a table of the first four factorial moments
for various bins of pseudorapidity and azimuth. The moments are
properly normalized so that they would be unity (up to
statistical fluctuations) for uniform and uncorrelated particle
production according to Poissonian statistics, but increasing
for decreasing bin size in case of `intermittent' behaviour.
The error on the average value is based on the actual
statistical sample (i.e. does not use any assumptions on the
distribution to relate errors to the average values of higher
moments). Note that for small bin sizes, where the average
multiplicity is small and the factorial moment therefore only
very rarely is non-vanishing, moment values may fluctuate wildly
and the errors given may be too low.
\iteme{= 33 :} stores the factorial moments in \ttt{/LUJETS/},
using the format: \\
\ttt{N =} 30, with \ttt{I = }$i=1$--10 corresponding to results for
slicing the rapidity range in $2^{i-1}$ bins, \ttt{I = }$i = 11$--20
to slicing the azimuth in $2^{i-11}$ bins, and \ttt{I = }$i = 21$--30
to slicing both rapidity and azimuth, each in $2^{i-21}$ bins; \\
\ttt{K(I,1) =} 32; \\
\ttt{K(I,2) =} 99; \\
\ttt{K(I,3) =} number of bins in rapidity; \\
\ttt{K(I,4) =} number of bins in azimuth; \\
\ttt{P(I,1) =} rapidity bin size; \\
\ttt{P(I,2) - P(I,5) =} $\langle F_2 \rangle$--$\langle F_5 \rangle$,
i.e. mean of second, third, fourth and fifth factorial moment; \\
\ttt{V(I,1) =} azimuthal bin size; \\
\ttt{V(I,2) - V(I,5) =} statistical errors on
$\langle F_2 \rangle$--$\langle F_5 \rangle$. \\
In addition, \ttt{MSTU(3) =} 1 and \\
\ttt{K(31,1) =} 32; \\
\ttt{K(31,2) =} 99; \\
\ttt{K(31,5) =} number of events analyzed.
\iteme{= 40 :} statistics on energy--energy correlation is reset.
\iteme{= 41 :} the energy--energy correlation $\mrm{EEC}$ of the
current
event is analyzed. Which particles (or partons) are used in the
analysis is determined by the \ttt{MSTU(41)} value. Events are
assumed given in their c.m. frame. The weight assigned to a pair
$i$ and $j$ is $2 E_i E_j/E_{\mrm{vis}}^2$, where $E_{\mrm{vis}}$
is the sum of energies of
all analyzed particles in the event. Energies are determined from
the momenta of particles, with mass determined according to the
\ttt{MSTU(42)} value. Statistics is accumulated for the relative
angle $\theta_{ij}$, ranging between 0 and 180 degrees, subdivided
into 50 bins.
\iteme{= 42 :} prints a table of the energy--energy correlation
$\mrm{EEC}$ and its asymmetry $\mrm{EECA}$, with errors.
The definition of errors is not unique. In our approach each event
is viewed as one observation, i.e. an $\mrm{EEC}$ and
$\mrm{EECA}$ distribution is obtained by
summing over all particle pairs of an event, and then the
average and spread of this event-distribution is calculated
in the standard fashion. The quoted error is therefore inversely
proportional to the square root of the number of events. It could
have been possible to view each single particle pair as one
observation, which would have given somewhat lower errors, but then
one would also be forced to do a complicated correction procedure
to account for the pairs in an event not being uncorrelated
(two hard jets separated by a given angle typically corresponds
to several pairs at about that angle). Note, however, that
in our approach the squared error on an $\mrm{EECA}$ bin is smaller
than the sum of the squares of the errors on the corresponding
$\mrm{EEC}$ bins (as it should be). Also note that it is not possible
to combine the errors of two nearby bins by hand from the
information given, since nearby bins are correlated (again a
trivial consequence of the presence of jets).
\iteme{= 43 :} stores the $\mrm{EEC}$ and $\mrm{EECA}$ in
\ttt{/LUJETS/}, using the format: \\
\ttt{N =} 25; \\
\ttt{K(I,1) =} 32; \\
\ttt{K(I,2) =} 99; \\
\ttt{P(I,1) =} $\mrm{EEC}$ for angles between \ttt{I-1} and \ttt{I},
in units of $3.6^{\circ}$; \\
\ttt{P(I,2) =} $\mrm{EEC}$ for angles between \ttt{50-I} and
\ttt{51-I}, in units of $3.6^{\circ}$; \\
\ttt{P(I,3) =} $\mrm{EECA}$ for angles between \ttt{I-1} and
\ttt{I}, in units of $3.6^{\circ}$; \\
\ttt{P(I,4), P(I,5) :} lower and upper edge of angular range of bin
\ttt{I}, expressed in radians; \\
\ttt{V(I,1) - V(I,3) :} errors on the $\mrm{EEC}$ and $\mrm{EECA}$
values stored in \ttt{P(I,1) - P(I,3)} (see \ttt{=42} for comments); \\
\ttt{V(I,4), V(I,5) :} lower and upper edge of angular range of bin
\ttt{I}, expressed in degrees. \\
In addition, \ttt{ MSTU(3)=1} and \\
\ttt{K(26,1) =} 32; \\
\ttt{K(26,2) =} 99; \\
\ttt{K(26,5) =} number of events analyzed.
\iteme{= 50 :} statistics on complete final states is reset.
\iteme{= 51 :} analyzes the particle content of the final state of
the current event record. During the course of the run, statistics
is thus accumulated on how often different final states appear.
Only final states with up to 8 particles are analyzed, and there
is only reserved space for up to 200 different final states.
Most high energy events have multiplicities far above 8, so the
main use for this tool is to study the effective branching
ratios obtained with a given decay model for e.g. charm or bottom
hadrons. Then \ttt{LU1ENT} may be used to generate one decaying
particle at a time, with a subsequent analysis by \ttt{LUTABU}.
Depending on at what level this studied is to be carried out,
some particle decays may be switched off, like $\pi^0$.
\iteme{= 52 :} gives a list of the (at most 200) channels with up
to 8 particles in the final state, with their relative branching
ratio. The ordering is according to multiplicity, and within
each multiplicity according to an ascending order of KF codes.
The KF codes of the particles belonging to a given channel are
given in descending order.
\iteme{= 53 :} stores the final states and branching ratios found in
\ttt{/LUJETS/}, using the format: \\
\ttt{N =} number of different explicit final states found (at most
200); \\
\ttt{K(I,1) =} 32; \\
\ttt{K(I,2) =} 99; \\
\ttt{K(I,5) =} multiplicity of given final state, a number between
1 and 8; \\
\ttt{P(I,1) - P(I,5), V(I,1) - V(I,3) :} the KF codes of the up to 8
particles of the given final state, converted to real
numbers, with trailing zeroes for positions not used; \\
\ttt{V(I,5) :} effective branching ratio for the given final state. \\
In addition, \ttt{MSTU(3)=1} and \\
\ttt{K(N+1,1) =} 32; \\
\ttt{K(N+1,2) =} 99; \\
\ttt{K(N+1,5) =} number of events analyzed; \\
\ttt{V(N+1,5) =} summed branching ratio for finals states not given
above, either because they contained more than 8 particles
or because all 200 channels have been used up.
\end{subentry}
\end{entry}
\drawbox{COMMON/LUDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)}
\begin{entry}
\itemc{Purpose:} to give access to a number of status codes and
parameters which regulate the performance of {\tsc{Jetset}}.
Most parameters are described in section \ref{ss:JETswitch};
here only those related to the event-analysis routines are described.
\boxsep
\iteme{MSTU(41) :}\label{p:MSTU41} (D=2) partons/particles used in
the event-analysis routines \ttt{LUSPHE}, \ttt{LUTHRU}, \ttt{LUCLUS},
\ttt{LUCELL}, \ttt{LUJMAS}, \ttt{LUFOWO} and \ttt{LUTABU}
(\ttt{LUTABU(11)} excepted).
\begin{subentry}
\iteme{= 1 :} all partons/particles that have not fragmented/decayed.
\iteme{= 2 :} ditto, with the exception of neutrinos and unknown
particles.
\iteme{= 3 :} only charged, stable particles, plus any partons
still not fragmented.
\end{subentry}
\iteme{MSTU(42) :} (D=2) assumed particle masses, used in
calculating energies $E^2 = \mbf{p}^2 + m^2$, as subsequently
used in \ttt{LUCLUS}, \ttt{LUJMAS} and \ttt{LUTABU}
(in the latter also for pseudorapidity, pion rapidity or true
rapidity selection).
\begin{subentry}
\iteme{= 0 :} all particles are assumed massless.
\iteme{= 1 :} all particles, except the photon, are assumed to have
the charged pion mass.
\iteme{= 2 :} the true masses are used.
\end{subentry}
\iteme{MSTU(43) :} (D=1) storing of event-analysis information (mainly
jet axes), in \ttt{LUSPHE}, \ttt{LUTHRU}, \ttt{LUCLUS} and
\ttt{LUCELL}.
\begin{subentry}
\iteme{= 1 :} stored after the event proper, in positions \ttt{N+1}
through \ttt{N+MSTU(3)}. If several of the routines are used in
succession, all but the latest information is overwritten.
\iteme{= 2 :} stored with the event proper, i.e. at the end of the
event listing, with \ttt{N} updated accordingly. If several of the
routines are used in succession, all the axes determined are
available.
\end{subentry}
\iteme{MSTU(44) :} (D=4) is the number of the fastest (i.e. with
largest momentum) particles used to construct the (at most) 10 most
promising starting configurations for the thrust axis determination.
\iteme{MSTU(45) :} (D=2) is the number of different starting
configurations above, which have to converge to the same (best) value
before this is accepted as the correct thrust axis.
\iteme{MSTU(46) :} (D=1) distance measure used for the joining of
clusters in \ttt{LUCLUS}.
\begin{subentry}
\iteme{= 1 :} $d_{ij}$, i.e. approximately relative transverse
momentum. Anytime two clusters have been joined, particles are
reassigned to the cluster they now are closest to. The distance
cut-off $d_{\mrm{join}}$ is stored in \ttt{PARU(44)}.
\iteme{= 2 :} distance measure as in \ttt{=1}, but particles are
never reassigned to new jets.
\iteme{= 3 :} JADE distance measyre $y_{ij}$, but with dimensions
to correspond approximately to total invariant mass. Particles may
never be reassigned between clusters. The distance cut-off
$m_{\mathrm{min}}$ is stored in \ttt{PARU(44)}.
\iteme{= 4 :} as \ttt{=3}, but a scaled JADE distance $y_{ij}$ is
used instead of $m_{ij}$. The distance cut-off $y_{\mathrm{min}}$ is stored
in \ttt{PARU(45)}.
\end{subentry}
\iteme{MSTU(47) :} (D=1) the minimum number of clusters to be
reconstructed by \ttt{LUCLUS}.
\iteme{MSTU(48) :} (D=0) mode of operation of the \ttt{LUCLUS}
routine.
\begin{subentry}
\iteme{= 0 :} the cluster search is started from scratch.
\iteme{= 1 :} the clusters obtained in a previous cluster search on the
same event (with \ttt{MSTU(48)=0}) are to be taken as the starting
point for subsequent cluster joining. For this call to have any effect,
the joining scale in \ttt{PARU(44)} or \ttt{PARU(45)} must have been
changed. If the event record has been modified after the last
\ttt{LUCLUS} call, or if any other cluster search parameter setting
has been changed, the subsequent result is unpredictable.
\end{subentry}
\iteme{MSTU(51) :} (D=25) number of pseudorapidity bins that the range
between \ttt{-PARU(51)} and \ttt{+PARU(51)} is divided into to define
cell size for \ttt{LUCELL}.
\iteme{MSTU(52) :} (D=24) number of azimuthal bins, used to define the
cell size for \ttt{LUCELL}.
\iteme{MSTU(53) :} (D=0) smearing of correct energy, imposed
cell-by-cell in \ttt{LUCELL}, to simulate calorimeter resolution
effects.
\begin{subentry}
\iteme{= 0 :} no smearing.
\iteme{= 1 :} the transverse energy in a cell, $E_{\perp}$, is smeared
according to a Gaussian distribution with standard deviation
\ttt{PARU(55)}$\times \sqrt{E_{\perp}}$, where $E_{\perp}$ is given in
GeV. The Gaussian is cut off so that $0 < E_{\perp \mrm{smeared}}
< $\ttt{PARU(56)}$\times E_{\perp \mrm{true}}$.
\iteme{= 2 :} as \ttt{=1}, but it is the energy $E$ rather than the
transverse energy $E_{\perp}$ that is smeared.
\end{subentry}
\iteme{MSTU(54) :} (D=1) form for presentation of information about
reconstructed clusters in \ttt{LUCELL}, as stored in \ttt{LUJETS}
according to the \ttt{MSTU(43)} value.
\begin{subentry}
\iteme{= 1 :} the \ttt{P} vector in each line contains $\eta$ and
$\varphi$ for the geometric origin of the jet, $\eta$ and $\varphi$
for the weighted center of the jet, and jet $E_{\perp}$, respectively.
\iteme{= 2 :} the \ttt{P} vector in each line contains a massless
four-vector giving the direction of the jet, obtained as \\
$(p_x,p_y,p_z,E,m) = E_{\perp}
(\cos\varphi,\sin\varphi,\sinh\eta,\cosh\eta,0)$, \\
where $\eta$ and $\varphi$ give the weighted center of a jet and
$E_{\perp}$ its transverse energy.
\iteme{= 3 :} the \ttt{P} vector in each line contains a massive
four-vector, obtained by adding the massless four-vectors of all cells
that form part of the jet, and calculating the jet mass from
$m^2 = E^2-p_x^2-p_y^2-p_z^2$. For each cell, the total $E_{\perp}$ is
summed up, and then translated into a massless four-vector
assuming that all the $E_{\perp}$ was deposited in the center of the
cell.
\end{subentry}
\iteme{MSTU(61) :} (I) first entry for storage of event-analysis
information in last event analyzed with \ttt{LUSPHE}, \ttt{LUTHRU},
\ttt{LUCLUS} or \ttt{LUCELL}.
\iteme{MSTU(62) :} (R) number of particles/partons used in the last
event analysis with \ttt{LUSPHE}, \ttt{LUTHRU}, \ttt{LUCLUS},
\ttt{LUCELL}, \ttt{LUJMAS}, \ttt{LUFOWO} or \ttt{LUTABU}.
\iteme{MSTU(63) :} (R) in a \ttt{LUCLUS} call, the number of
preclusters constructed in order to speed up analysis (should be
equal to \ttt{MSTU(62)} if \ttt{PARU(43)=0.}). In a \ttt{LUCELL}
call, the number of cells hit.
\iteme{MSTU(161), MSTU(162) :}\label{p:MSTU161} hard flavours involved
in current event,
as used in an analysis with \ttt{LUTABU(11)}. Either or both may be set
0, to indicate the presence of one or none hard flavours in event.
Is normally set by high-level routines, like \ttt{LUEEVT} or
\ttt{PYEVNT}, but can also be set by you.
\boxsep
\iteme{PARU(41) :}\label{p:PARU41} (D=2.) power of momentum-dependence
in \ttt{LUSPHE}, default corresponds to sphericity, \ttt{=1.} to linear
event measures.
\iteme{PARU(42) :} (D=1.) power of momentum-dependence in \ttt{LUTHRU},
default corresponds to thrust.
\iteme{PARU(43) :} (D=0.25 GeV) maximum distance $d_{\mrm{init}}$
allowed in
\ttt{LUCLUS} when forming starting clusters used to speed up
reconstruction. The meaning of the parameter is in $p_{\perp}$ for
\ttt{MSTU(46)}$\leq 2$ and in $m$ for \ttt{MSTU(46)}$\geq 3$. If
\ttt{=0.}, no preclustering is obtained.
If chosen too large, more joining may be generated at this stage
than is desirable. The main application is at high energies,
where some speedup is imperative, and the small details are not
so important anyway.
\iteme{PARU(44) :} (D=2.5 GeV) maximum distance $d_{\mrm{join}}$,
below
which it is allowed to join two clusters into one in \ttt{LUCLUS}.
Is used for \ttt{MSTU(46)}$\leq 3$, i.e. both for $p_{\perp}$ and mass
distance measure.
\iteme{PARU(45) :} (D=0.05) maximum distance
$y_{\mrm{join}} = m^2/E_{\mrm{vis}}^2$,
below which it is allowed to join two clusters into one in
\ttt{LUCLUS} for \ttt{MSTU(46)=4}.
\iteme{PARU(48) :} (D=0.0001) convergence criterion for thrust (in
\ttt{LUTHRU}) or generalized thrust (in \ttt{LUCLUS}), or relative
change of $m_{\mrm{H}}^2 + m_{\mrm{L}}^2$ (in \ttt{LUJMAS}),
i.e. when the value changes by less than this amount between two
iterations the process is stopped.
\iteme{PARU(51) :} (D=2.5) defines maximum absolute pseudorapidity
used for detector assumed in \ttt{LUCELL}.
\iteme{PARU(52) :} (D=1.5 GeV) gives minimum $E_{\perp}$ for a cell
to be considered as a potential jet initiator by \ttt{LUCELL}.
\iteme{PARU(53) :} (D=7.0 GeV) gives minimum summed $E_{\perp}$ for
a collection of cells to be accepted as a jet.
\iteme{PARU(54) :} (D=1.) gives the maximum distance in
$R = \sqrt{(\Delta\eta)^2 + (\Delta\varphi)^2}$ from cell initiator
when grouping cells to check whether they qualify as a jet.
\iteme{PARU(55) :} (D=0.5) when smearing the transverse energy
(or energy, see \ttt{MSTU(53)}) in \ttt{LUCELL}, the calorimeter
cell resolution is taken to be
\ttt{PARU(55)}$\times \sqrt{E_{\perp}}$ (or
\ttt{PARU(55)}$\times \sqrt{E}$) for $E_{\perp}$ (or $E$) in GeV.
\iteme{PARU(56) :} (D=2.) maximum factor of upward fluctuation in
transverse energy or energy in a given cell when calorimeter
resolution is included in \ttt{LUCELL} (see \ttt{MSTU(53)}).
\iteme{PARU(57) :} (D=3.2) maximum rapidity (or pseudorapidity or
pion rapidity, depending on \ttt{MSTU(42)}) used in the factorial
moments analysis in \ttt{LUTABU}.
\iteme{PARU(58) :} (D=0. GeV) in a \ttt{LUCELL} call, cells with a
transverse energy $E_{\perp}$ below \ttt{PARP(58)} are removed from
further consideration. This may be used to represent a threshold
in an actual calorimeter, or may be chosen just to speed up the
algorithm in a high-multiplicity environment..
\iteme{PARU(61) :} (I) invariant mass $W$ of a system analyzed with
\ttt{LUCLUS} or \ttt{LUJMAS}, with energies calculated according to
the \ttt{MSTU(42)} value.
\iteme{PARU(62) :} (R) the generalized thrust obtained after a
successful \ttt{LUCLUS} call, i.e. ratio of summed cluster momenta
and summed particle momenta.
\iteme{PARU(63) :} (R) the minimum distance $d$ between two clusters
in the final cluster configuration after a successful \ttt{LUCLUS}
call; is 0 if only one cluster left.
\end{entry}
\clearpage
\section{Summary and Outlook}
A complete description of the {\tsc{Pythia/Jetset}} programs would have to cover
four aspects:
\begin{Enumerate}
\item the basic philosophy and principles underlying the programs;
\item the detailed physics scenarios implemented, with all the
necessary compromises and approximations;
\item the structure of the implementation, including program flow,
internal variable names and programming tricks; and
\item the manual, which describes how to use the programs.
\end{Enumerate}
Of these aspects, the first has been dealt with in reasonable detail.
The second is unevenly covered: in depth for aspects which are not
discussed anywhere else, more summarily for areas where separate
up-to-date papers already exist. The third is not included at all,
but `left as an exercise' for the reader, to figure out from the code
itself. The fourth, finally, should be largely covered, although
many further comments could have been made, in particular about the
interplay between different parts of the programs. Still, in the
end, no manual, however complete, can substitute for `hands on'
experience.
The {\tsc{Pythia/Jetset}} programs are continuously being developed. We are aware
of many shortcomings, which hopefully will be addressed in the
future, such as:
\begin{Itemize}
\item polarization effects should be included in more places,
in particular for $\tau$ production and decay;
\item the photoproduction and $\gamma\gamma$-physics scenarios
should be expanded;
\item many processes of interest are missing; and
\item mass relations and couplings need to be included beyond the
Born level in the MSSM two Higgs doublet scenario.
\end{Itemize}
This list could have been made much longer (I almost certainly missed
your top priority). One other aspect would be to provide more and
longer examples of working main programs for a number of standard
applications.
Apart from these physics aspects, one may also worry about the
programming ones. For instance, for historical reasons,
single precision real is used almost everywhere. With the
push to higher energies, this is becoming more and more of a
problem, so it would be logical to move to double precision
throughout.
One should also note that the {\tsc{Jetset}} and {\tsc{Pythia}} programs these days
are becoming so intertwined, that it would make sense to join
them into one single program. This would e.g. mean that the current
$\e^+\e^-$ generation routines of {\tsc{Jetset}} are made part of the generic
{\tsc{Pythia}} process generation machinery --- this is particular affects
the matrix-element options, since $\e^+\e^-$ events with parton showers
already are available in {\tsc{Pythia}}. A joint product would likely adopt
the name {\tsc{Pythia}}: although {\tsc{Jetset}} is the older of the two programs, it
has a less well developed identity of its own. (It is also often
referred to as `Lund', which today is more confusing than it was in
the early days.) In the process of joining the programs, one would
probably also remove a number of options that are no longer used.
Another possible change on longer time scales would be an introduction
of Fortran 90 programming elements. In particular, derived data types
could be used to define the event record as a one-dimensional array,
where each element represents a particle, with integer and real
components to give flavour, history, momentum and production vertex.
No timetable is set up for future changes. After all, this is not
a professionally maintained software product, but part of a one-man
physics research project. Very often, developments
of the programs have come about as a direct response to the evolution
of the physics stage, i.e. experimental results and studies for
future accelerators. Hopefully, the program will keep on evolving
in step with the new challenges opening up.
\clearpage
|
\section{Introduction}
Photon-proton processes
are traditionally classified according to the virtuality ($Q^2$)
of the photon.
For quasi-real photoproduction interactions
$Q^2$ is close to zero, and
correspondingly the photon is nearly
on mass shell. Interactions which involve
a $Q^2$ larger than a few GeV$^2$ are usually
termed
deep-inelastic scattering processes (DIS).
This distinction results mainly from the
different theoretical
descriptions adopted for these processes.
Photoproduction has turned out to be very
similar to hadron-hadron collisions and is described in a VMD-like
(Vector Meson Dominance) picture~\cite{sakurai},
where the photon is assumed to fluctuate into a vector meson before interacting
with the proton. Additional diagrams such as the direct coupling of the
photon to quarks in the proton and a pointlike contribution where the
photon splits into a $q\overline{q}$ pair are needed to accommodate the
large tail
observed
in the transverse momentum
distribution of produced particles~\cite{H1:incl}, but the
majority of photoproduction collisions show
features of soft low-$p_T$ processes as seen in hadronic collisions. For DIS
on the other hand,
the observation of scaling of the structure function in
early experiments suggested that this interaction could, in the infinite
momentum frame of the proton, be interpreted as a hard scattering process
in which a
point-like virtual photon ``probes'' the structure of the hadronic
target.
The subsequently measured scaling violations are well
described by perturbative QCD.
However,
recent DIS measurements at the high energy $ep$
collider HERA and high precision measurements from fixed target
leptoproduction experiments have shown that
the data in the low Bjorken-$x$ region reveal
properties from soft interactions as well. Here $x = Q^2/2P\cdot q$ where
$P$ and $q$ are the
four momenta of the incoming proton and photon, respectively.
Elastic and other diffractive hadronic final states have been
produced~\cite{h1rapgap,disdif}
and shadowing
on nuclear targets has been observed~\cite{shadowing}, phenomena
which are well known from
hadron-hadron and real photon collisions~\cite{gvmd}.
These observations suggest that the different treatment of
low and high $Q^2$ interactions is somewhat artificial, and it is
therefore worthwhile testing a prescription which provides a
natural transition between these two classes.
A qualitative approach for a smooth transition of the high
energy $\gamma^*p$~\footnote{In this paper
the generic symbol $\gamma^*$ is used to denote a
colliding photon irrespective of the virtuality.}
interactions over a large range of $Q^2$ results
from viewing the photon-proton collision in the proton rest frame and
the hypothesis that the photon can develop
as a hadronic object before
interacting with the nuclear target.
The time in which a real photon can fluctuate into e.g. a $\rho$ meson is given
by the Heisenberg uncertainty principle and amounts to
$\tau \approx 2\nu/M^2_{\rho}$ where $\nu$ is the photon energy in the
proton rest frame and $M_{\rho}$ the mass of the $\rho$ meson.
The lifetime
of a high energy real photon to fluctuate
into a hadronic state is much longer than the time of the
strong interaction itself, and therefore
this picture is
generally applied to
describe photoproduction interactions.
The same argument can be used for DIS
interactions at low $x$. In the proton rest frame the time in which the
virtual photon can fluctuate and stay in a hadronic state, e.g. a
quark-antiquark pair,
is given by
$\tau \approx 1/(xM_p)$ ~\cite{tioffe,brodsky}, where
$M_p$ is the mass of the nucleon target.
Thus, for small $x$, the virtual photon
can convert into a quark-antiquark
pair and cover a distance which is long compared to the interaction length.
For an $x$ in the range of
$10^{-2}-10^{-4}$ this distance is in the range of 10 to 1000 fm,
much larger than
the typical radius of the target. In the HERA kinematic range such $x$ values
can be reached for $Q^2 \sim 10 $ GeV$^2$.
Therefore, at low $x$, a virtual photon
can stay
in a hadronic state for a long time and interact
with the target strongly, leading to a final state
similar to the one in a hadron-hadron collision.
This picture of DIS in the frame where the incoming photon fluctuates
into a hadronic system before interacting with the proton has already been
advocated in various papers~\cite{nikolaev,alligned,kwiecinski}.
Note that this
is not in
contradiction with the traditional treatment of DIS as a point-like process,
where a virtual photon ``probes" the hadronic structure of the nuclear target.
In fact both pictures are taken to be complementary, as discussed
in~\cite{brodsky}.
The
HERA collider
provides a unique opportunity to study the final state
of both photoproduction and DIS interactions at high energy and small
Bjorken-$x$: 26.7 GeV electrons collide with 820 GeV protons, yielding a
centre of mass energy $\sqrt{s}$ of 296 GeV. This allows for a study of
DIS interactions for $x$ values down to $\sim 10^{-4}$~\cite{h1f2},
and photoproduction
collisions at a centre of mass energy of $\sim 200$ GeV~\cite{H1:incl}.
The data recorded with the H1 experiment are used to
study the transverse energy behaviour in photoproduction and
DIS interactions, and the analogy with hadronic collisions is checked.
Within this analogy $\mbox{$\gamma^{*}p$}$ events can be sub-divided
in rapidity-space into
three regions, which differ in the mechanism by which the hadronic final
state is produced~\cite{bj73}: the proton fragmentation region,
the photon fragmentation region and a hadron plateau
spanning the rapidity interval between the
two. The height of the hadronic plateau depends logarithmically
on the centre of mass energy of the $\mbox{$\gamma^{*}p$}$ collision.
The region of the hadronic plateau is expected to be independent of the
nature
of the two incoming ``hadrons",
as was verified in
hadron-hadron collisions~\cite{hadver}.
Particle production in the proton fragmentation region is also expected
to be very similar compared to hadron-proton collisions,
but the limited experimental
acceptance in this region does not allow this to be verified. The ``hadronic"
nature of the photon is assumed to change
with increasing $Q^2$, thus
the photon
fragmentation region is expected to change with $Q^2$.
Also studied is
the fraction of diffractive events in
DIS and photoproduction.
In hadronic collisions
approximate factorization of the cross section for high mass
diffractive dissociation has been observed, in accordance with
Regge theory predictions~\cite{triple}:
the ratio of the hadron diffractive dissociation to the total
cross section is found to be approximately independent of the type of
dissociating
hadron and is thus the
same for pions, kaons, protons and real
photons~\cite{triple2}.
Applying this factorization property
to real and virtual photons leads to the expectation that the
ratio of the
photon diffractive dissociation cross section to the total cross
section
is independent of the $Q^2$ of the photon, in the high dissociative mass
region.
\section{Detector Description}
A detailed description of the H1 apparatus can be found elsewhere~\cite{h1}.
In the following the components of the detector relevant for
this analysis are briefly described.
The hadronic energy flow and the scattered electrons in DIS processes
are measured with a
liquid argon~(LAr) calorimeter and a
backward electromagnetic lead-scintillator calorimeter (BEMC).
The LAr calorimeter~\cite{larc} covers the polar angular range
$4^\circ < \theta < 153^\circ$ with full azimuthal coverage, where
$\theta$ is defined with respect to the proton
beam direction ($+z$ axis).
It consists of an electromagnetic section with lead absorbers
and a hadronic section with steel absorbers.
Both sections are highly segmented in the transverse and
longitudinal direction
with about 44000 cells in total.
The total depth of both sections
varies between 4.5 and 8 interaction lengths.
Test beam measurements of the LAr~calorimeter modules show an
energy resolution of $\sigma_{E}/E\approx 0.12/\sqrt{E\;[\rm GeV]} \oplus 0.01$
for electrons ~\cite{elcern} and
$\sigma_{E}/E\approx 0.50/\sqrt{E\;[\rm GeV]} \oplus 0.02$ for
charged pions~\cite{h1calo3}.
The uncertainty in the absolute energy scale for electrons is
$3\%$.
The absolute scale of the hadronic energy is presently known to $5\%$, as
determined from studies of the
transverse momentum balance in DIS events.
The BEMC (depth of 22.5 radiation lengths or 1
interaction length) covers the backward region of the detector,
$151^\circ < \theta < 177^\circ$.
A major task of the BEMC is to trigger on and measure scattered electrons
in DIS processes with $Q^2$ values ranging from 5 to 100 GeV$^{2}$.
The BEMC energy scale for electrons is known to an accuracy of
$1.7\%$~\cite{bemc}.
Its resolution is given by
$\sigma_{E}/E = 0.10/\sqrt{E\;[\rm GeV]} \oplus 0.42/E[\rm GeV] \oplus 0.03$
\cite{f2pap}.
The calorimeters are
surrounded by a superconducting solenoid providing a uniform
magnetic field of $1.15$ T parallel to the beam axis in the tracking region.
Charged particle
tracks are measured in a central drift chamber and the forward
tracking system,
covering the polar angular range $ 7^\circ < \theta < 165^\circ$.
The central chamber is interleaved with an inner and an outer double
layer of multi-wire proportional chambers (MWPC),
which were used for the trigger to select
events with charged tracks pointing to the interaction region.
A backward proportional chamber (BPC) in front of the BEMC with an angular
acceptance of $155.5^\circ < \theta < 174.5^\circ$ serves to
support electron identification
and to precisely measure electron direction.
Using information from the BPC, the BEMC and the reconstructed event vertex the
polar angle of the scattered electron is known to better than 2 mrad.
A small angle detector (electron tagger),
which is part of the luminosity system, is a TlCl/TlBr
crystal calorimeter with an energy resolution
$\sigma_{E}/E = 0.1/\sqrt{E\;[\rm GeV]}$. It is located at $z$ = --33 m
and accepts electrons from photoproduction processes
with an energy fraction between 0.2 and 0.8 with
respect to the beam energy and scattering angles $\theta'< 5$~mrad
($\theta' = \pi-\theta$).
\section{Event Selection and Correction for Detector Effects}
The data used in this analysis were collected in 1993
and correspond to
an integrated luminosity of about $0.3$ pb$^{-1}$.
The kinematic variables $Q^2, x$ and $y$
of the $ep$ collision are determined from
the scattered electron:
$ Q^2 = 4\,E_e \, E'_e\cos^2(\theta_{e}/2)$ and
$ y = 1-(E'_e/E_e)\cdot \sin^2(\theta_{e}/2)$,
where $E_e$ is the energy of the incident electron
and $E_e'$ and $\theta_e$ are
the energy and the polar angle of the scattered electron respectively.
The variable $y$ represents the fraction of the energy of the electron
transferred to the proton, in the proton rest frame.
The scaling variable $x$
is then derived via $x=Q^2/(ys)$, and the total hadronic invariant mass
is given by
$W^2=sy-Q^2$.
\begin{table}[thb]
\centering
\begin{tabular}{|c|c|c|} \hline
photoproduction & low $Q^2$ DIS & high $Q^2$ DIS \\
sample & sample & sample \\ \hline
$Q^2<10^{-2}$~GeV$^2$ & $5<Q^2<100$~GeV$^2$ & $Q^2>100$~GeV$^2$ \\ \hline
$\theta_e-\pi < 5$~mrad & $157^\circ<\theta_e < 173^\circ$ &
$10^\circ < \theta_e < 148^\circ$
\\ \hline
$0.3 < y < 0.5$ & $E'_e > 12$~GeV & $0.05< y < 0.7$ \\
& $y > 0.05$ & \\ \hline
\end{tabular}
\caption[~]
{\small
Accepted kinematic regions for the three sub-samples used in this analysis.
}
\label{seltable}
\end{table}
The data are classified into three event sub-samples
depending on the $Q^2$ range. A different detector component of H1
is used to detect
the scattered electron in each of these ranges. In the studies below
these subsamples
get further sub-divided into samples with different $y$ or $Q^2$ values.
\begin{itemize}
\item
The photoproduction sub-sample ($Q^2<10^{-2}$~GeV$^2$)
consists of events where the scattered electron is detected in the
small angle electron detector.
To
avoid regions of low acceptance and to
facilitate the data correction procedure,
the kinematic range was further restricted to $0.3<y<0.5$.
The photoproduction events
are triggered by a coincidence of an energy deposit in the small
angle electron detector
and at least one track pointing to the vertex region.
The track condition is derived from the cylindrical MWPC and
requires a $p_T\;\raisebox{-0.5mm}{$\stackrel{>}{\scriptstyle{\sim}}$}\;200$ MeV/c.
\item
The low $Q^2$ DIS sub-sample ($5<Q^2<100$~GeV$^2$) consists of events
where the scattered electron is detected in the BEMC. The events
are triggered by requiring
a cluster of more than 4~GeV in the BEMC.
The most energetic BEMC cluster is taken to be
the scattered electron.
\item
For the high $Q^2$ DIS sub-sample ($Q^2>100$~GeV$^2$)
the scattered electron is detected in the LAr calorimeter.
The events
are triggered by requiring
a cluster of more than 5~GeV in the electromagnetic part of the LAr
calorimeter,
and no associated hadronic energy.
The electromagnetic cluster in the LAr calorimeter with the highest transverse
energy is considered to be the scattered electron.
\end{itemize}
Further details on the scattered electron identification procedure
in the BEMC and LAr calorimeters and in the small angle electron detector
can be found in~\cite{h1f2,h1alphas,h1gamp},
respectively.
The selected kinematic regions for each sub-sample, as given
in Table~\ref{seltable}, are
chosen to ensure a large acceptance,
high trigger efficiency and,
for the DIS sub-samples, a small photoproduction background (less than 3\%).
For all three sub-samples the $z$ position of the
event vertex reconstructed from charged
tracks was required to be within $\pm30$~cm
of the nominal interaction point.
A minimum $y$ cut, $y > 0.05$ was imposed for the DIS sub-samples.
Events suffering from QED radiation or from a
badly reconstructed electron
are strongly reduced by requiring that they also fulfill
this cut if $y$ is calculated from the measured hadrons.
The final event samples contain 82850 photoproduction events,
15324 low $Q^2$ DIS events with
$5 < Q^2 < 100$~GeV$^2$ ($10^{-4} < x < 10^{-2}$),
692 high $Q^2$ DIS events with
$Q^2 > 100$~GeV$^2$ ($10^{-3} < x < 10^{-1}$).
The data are corrected for detector effects using
samples of Monte Carlo generated events which were fully
simulated in the H1 detector.
The PHOJET~\cite{phojet} generator for photoproduction
and the CDM~\cite{ariadne}
(Colour Dipole Model) generator for DIS processes were used.
The program PHOJET generates $\gamma p$ interactions, treating the
photon as a hadron-like object.
The model used
to simulate the hadronic final states is similar to that used
in the Monte Carlo program
DTUJET~\cite{dtujet} which simulates particle production
in $pp$ and $\bar{p}p$ collisions up to very high energies.
The CDM Monte Carlo program generates DIS events, and uses
the colour dipole
model for QCD radiation in the hadronic final state.
Here the final state is
assumed to be a chain of independently radiating dipoles formed by
emitted gluons~\cite{dipole}. Since all radiation is
assumed to originate from the dipole
formed by the struck quark and the remnant, photon-gluon fusion events
have to be added and are taken from the QCD matrix elements~\cite{lepto}.
Version 4.03 of the ARIADNE program was used for the CDM
studies in this paper.
To estimate systematic uncertainties of the correction procedure
for DIS events
another model for the hadronic final state was used as well:
the MEPS (Matrix Elements plus Parton Showers) model~\cite{lepto}.
This model incorporates
QCD matrix elements up to first order, with additional
soft emissions generated by adding leading log parton showers.
Divergences of the matrix element are avoided
by imposing a lower limit
on the parton-parton invariant masses,
which was parametrized as a function of $W$ such that it is always
2 GeV above the region in phase space where the
order $\alpha_s$ contributions
would exceed the total cross section. More details on this implementation are
given in~\cite{h1gljet}.
\section{Results}
For comparisons of event properties at different $Q^2$ values,
the $\gamma^*p$ centre of mass system (CMS) is chosen as frame of
reference.
The orientation of the CMS is such that the
direction of the proton defines the positive $z'$ axis.
Transverse quantities are defined with respect to the proton direction
in this frame.
The flow of transverse energy, $E_T$,
as a function of pseudorapidity $\eta^*=-\ln(\tan \frac{\theta^*}{2})$
in the CMS is shown in Fig.~\ref{ETA}.
Here $\theta^*$ is the polar angle of a particle in the CMS frame with
respect to the proton direction.
The transverse energy is calculated from the energy deposits in the
calorimeter cells.
For this study the DIS data are restricted to the same kinematic region
$0.3<y<0.5$ as the photoproduction events.
The $y$ range corresponds
to a mean $\mbox{$\gamma^{*}p$}$ collision energy of
$\sqrt{s_{\gamma p}} = W \approx 185$~GeV.
The corrections which have been applied to correct for detector
effects never exceed 30$\%$.
\begin{figure}[htb] \centering
\Large \boldmath
\begin{picture}(170,110)(-8,0)
\put( 26,98){\small $\average{Q^2}\sim 0 $ GeV$^2$}
\put( 26,92){\small $\average{Q^2}\sim 11 $ GeV$^2$}
\put( 26,86){\small $\average{Q^2}\sim 38 $ GeV$^2$}
\put( 26,80){\small $\average{Q^2}\sim 520 $ GeV$^2$}
\put( 115,5){$\eta^* $}
\put( 5,+42){\begin{sideways}
$ 1/N \cdot dE_T/d\eta^*~~[$GeV$]$ \end{sideways}}
\epsfig
{file=fig1.ps,height=130mm,width=130mm,angle=90}
\end{picture}
\normalsize \unboldmath
\caption[~]
{\small
The flow of transverse energy $E_T$ in the hadronic CMS as a function of
pseudorapidity $\eta^*$ normalized to the number of events $N$.
Photoproduction data (open circles) are compared with
DIS data (full symbols: circles -- $\average{Q^2}\approx 11$~GeV$^2$,
triangles -- $\average{Q^2}\approx 38$~GeV$^2$,
squares -- $\average{Q^2}\approx 520$~GeV$^2$)
in the same $y$ range $0.3 < y < 0.5$
($W\approx 185$ GeV).
}
\label{ETA}
\end{figure}
\begin{figure}[htb] \centering
\Large \boldmath
\begin{picture}(170,110)(7,0)
\put( 7,+45){\begin{sideways}
$1/N\cdot dE_T/d\eta^* ~[$GeV$]$ \end{sideways}}
\put( 38,89){\small $-3<\eta^*<-2$}
\put( 38,96){\small $-0.5<\eta^*<0.5$}
\put( 33,82){\small -------- CDM}
\put( 33,76){\small -- -- -- MEPS}
\put(125,5){$Q^2~~[$GeV$^2]$ }
\put( 36,18){\large $\wr$}
\put( 37,18){\large $\wr$}
\put( 36,107){\large $\wr$}
\put( 37,107){\large $\wr$}
\put( 26,15){\small \bf 0}
\epsfig
{file=fig2.ps,height=130mm,width=170mm,angle=90}
\end{picture}
\normalsize \unboldmath
\caption[~]
{\small
The transverse energy per unit of pseudorapidity in
the CMS central region ($-0.5<\eta^*<0.5$, full circles) and in the photon
fragmentation region ($-3<\eta^*<-2$, open circles) as a function of $Q^2$
for $0.3 < y < 0.5$.
For comparison, the CDM (full line) and
MEPS (dashed line) models are shown. The lower curves correspond to
the CMS central region and the upper ones to the photon
fragmentation region.
}
\label{Q2}
\end{figure}
\begin{figure}[htb] \centering
\Large \boldmath
\begin{picture}(170,110)(7,0)
\put( 40,97){\large H1, $Q^2\approx~8~$GeV$^2$}
\put( 40,92){\large H1, $Q^2\approx~14~$GeV$^2$}
\put( 40,87){\large H1, $Q^2\approx~30~$GeV$^2$}
\put( 40,82){\large H1, photoproduction}
\put( 40,77){\large UA1}
\put( 40,72){\large AFS}
\put( 40,67){\large NA22}
\put( 7,20){\begin{sideways}
$1/N\cdot dE_T/d\eta^*$ at $\eta^*=0~[$GeV$]$ \end{sideways}}
\put(125,5){$ W ~~[$GeV$]$ }
\epsfig
{file=fig3.ps,height=130mm,width=170mm,angle=90}
\end{picture}
\normalsize \unboldmath
\caption[~]
{\small
The transverse energy per unit of pseudorapidity
in the CMS central region
as a function of the hadronic CMS energy.
The DIS data are compared with photoproduction data and with
data from hadron-hadron collisions ($p\overline{p}$ for UA1;
$pp$ for NA22 and AFS).
Systematic point-to-point errors of $6\%$ and an overall scale
error of $9\%$ for the H1 data are not shown,
neither are the global scale uncertainties for the other experiments.
}
\label{W}
\end{figure}
\begin{figure}[htb] \centering
\Large \boldmath
\begin{picture}(170,110)(7,0)
\put( 25,100){\Large $a)$}
\put(105,100){\Large $b)$}
\put( 50,5){$E_T~~[$GeV$]$}
\put( 5,+30){\begin{sideways}
$ 1/N \cdot dN/dE_T~~[$GeV$^{-1}]$ \end{sideways}}
\put( 140,5){$\eta^*_{max} $}
\put(85,+45){\begin{sideways}
$ 1/N \cdot dN/d\eta^*_{max}$ \end{sideways}}
\epsfig
{file=fig4.ps,height=130mm,width=170mm,angle=90}
\end{picture}
\normalsize \unboldmath
\caption[~]
{\small
$(a)$ distribution of the uncorrected transverse energy per unit
of pseudorapidity in the central region $(-0.5<\eta^*<0.5)~$,
$(b)$ $\eta^*_{max}$ distribution normalized to the total number of events.
Open circles - photoproduction data, full circles - DIS data
with~$10<Q^2<100$~GeV$^2$ and $0.3<y<0.5$.}
\label{ET-MAX}
\end{figure}
Only statistical errors are shown in Fig.~\ref{ETA}.
For the comparison of the $E_T$ spectra at
the four different $Q^2$ values it
is important to identify the relative systematic uncertainty, so
the systematic errors on each ``$Q^2$ point"
at a given $\eta^*$ value are considered
in two parts.
Firstly there is a point-to-point systematic error, which has a different
effect on each of the four measurements,
accounting
for kinematically dependent systematic effects. Secondly there is
an overall scale error which affects all four points
at a given $\eta^*$ value in the same way.
The main source of the point-to-point systematic error
is the model dependence of the correction for detector effects.
This dependence was investigated with the
CDM and MEPS models and leads to a point-to-point error of
$6\%$\footnote{In the central region the correction for photoproduction
events determined with PHOJET is close to the one for DIS events determined
with MEPS at $<Q^2>\sim 11$ GeV$^2$}.
Additionally, the model dependence contributes
$6\%$
to the overall scale error at each $\eta^*$ value.
Further contributions to the overall scale error
arise from the LAr and BEMC calorimeter calibration
($5\%$ and $20\%$ respectively), and
from details of the analysis method, such as the calorimeter noise
treatment,
the clustering scheme for the calorimeter cells,
and the accuracy of the simulation of the calorimeter
response,
affecting the results by $5\%$
in the region close to the proton direction ($\eta^* > -1$).
All these contributions
give an overall scale error varying
from $8\%$ in the photon fragmentation region ($-3.5 < \eta^* < -1$) to
$9\%$ in the central region ($\eta^* > -1$), and
up to $20\%$ in the region $\eta^* < -3.5$, which does not affect the
results of
the present analysis.
This discussion on the systematic errors is also valid
for data shown below.
Fig.\ref{ETA} demonstrates that
the transverse energy flow
exhibits a strong increase in the photon fragmentation region
($\eta^* < -1$),
from $Q^2=0$ (photoproduction) to $Q^2 \approx 500$~GeV$^2$ (DIS),
while in the central region the level of $E_T$ remains
almost the same.
This can be taken as evidence that the influence of the $Q^2$
of the photon on the $E_T$ flow diffuses
away quite quickly towards the central region.
Such behaviour is expected if the central region corresponds
to the hadronic plateau, discussed in the introduction.
In order to quantify the change of $E_T$ with $Q^2$,
the value of
$E_T$ per unit of pseudorapidity ($E_T$ density) for an $\eta^*$ slice
in the central
pseudorapidity region ($-0.5<\eta^*<0.5$) and
an $\eta^*$ slice in the photon fragmentation
region ($-3<\eta^*<-2$) is shown in Fig.~\ref{Q2}.
The data clearly demonstrate
that the transverse energy in the central region is
essentially independent of the
$Q^2$ of the photon, but increases significantly
in the photon fragmentation region.
However, in the photon fragmentation region
the difference in the $E_T$ density for DIS and
photoproduction
data is significant only for large $Q^2$ values, $Q^2 > 20 $ GeV$^2$.
The level of transverse energy
in the photon fragmentation region for interactions with $Q^2 < 10 $ GeV$^2$
is quite similar, roughly independent of the transverse size,
$Q^2$, of the photon.
Since the former can be described by perturbative calculations, this
observation hints that the result of some
predicted features of the final state may be
transportable to the hitherto
assumed non-perturbative ``VDM region" at $Q^2 = 0$.
In Fig.~\ref{Q2}
the results are compared with predictions from the CDM and
MEPS models, as described in the previous section.
The $E_T$ density
predicted by the CDM model is found to be independent of
$Q^2$, which is in accordance with the data in the central region, but
does not
reproduce the data in the photon fragmentation region.
In this version of the CDM model the scale for QCD radiation is given by
the $p_T^2$ of the radiated gluons, which is limited by $W^{4/3}$.
Since data corresponding to
the same $W^2$ region are selected,
the predictions for the
$E_T$ density are similar for the different $Q^2$ data samples.
A new version of the CDM model (version 4.06)\cite{leif2} partially
corrects for this defect in the photon fragmentation region, but still does
not yield a good description of the data.
The MEPS model on the other hand describes the behaviour
of the transverse energy flow in the photon
fragmentation region but fails in the central region.
Hence, in this model, the influence of the
photon virtuality extends too far in rapidity away from
the photon-$q\overline{q}$ vertex.
The comparison of absolute values of the $E_T$ density
between the data and
models should be viewed with some caution since diffractive events are not
included in these models. These events, which amount to $\approx 6\%$
of all events,
give rise to a
rapidity gap~\cite{h1rapgap},
i.e. no energy deposition
in the central region.
The exclusion of diffractive events however does not change the shape of
the distribution of the $E_T$ density versus $Q^2$.
Comparisons of the energy flows for models and data with rapidity gap
events removed can be found in~\cite{h1bfkl}.
Within the hadronic picture,
the observation of a rise of the $E_T$ density
with $Q^2$ in the photon
fragmentation region implies that, on
arrival at the target, the virtual photon has a hadronic structure
(parton configuration) which depends on~$Q^2$.
With increasing $Q^2$ the transverse momentum of the constituent
partons in this hadronic structure increases.
A formal explanation of this effect based on QCD
calculations is given in~\cite{khoze}.
The central region however shows a more universal behaviour, independent
of the partonic nature of the colliding particles.
It is therefore interesting to compare our measurement with
high energy hadron-hadron collisions. In
Fig.~\ref{W} the transverse energy density,
$dE_T/d\eta^*$, at $\eta^* = 0$ is shown for DIS, photoproduction,
$pp$ and $\overline{p}p$ interactions as a function of $W$
(= $\sqrt{s}$ for hadron-hadron collisions).
The results for
DIS include only the data with $Q^2 < 50$~GeV$^2$, where enough
statistics is available
to cover a large $W$ (or equivalently
$y$) range, as given in Table 1.
The $W$ dependence of $dE_T/d\eta^*$ observed in DIS processes at low $Q^2$
(and low $x$) agrees with
the $W$ interpolation between measurements
from $p\bar{p}$~\cite{ua1mb} and
$pp$~\mbox{\cite{ua1mb,na22}}
collisions.
Also the photoproduction data show the same level of $E_T$ for the
$W$ value of this data sample.
This observation is consistent with the ansatz of the analogy of
$\gamma^*p$ interactions with real photon-hadron and
hadron-hadron collisions.
In Fig.~\ref{ET-MAX}a
the distribution of (uncorrected)
transverse energy summed over a
unit of pseudorapidity in the central region ($-0.5<\eta^*<0.5$)
is shown for photoproduction and for DIS data
($10<Q^2<100$~GeV$^2$, $0.3<y<0.5$).
The comparison shows that in the central region not
only the mean $E_T$,
but also the energy spectra themselves, are quite similar.
This apparently holds down to very low $E_T$ values,
where diffractive dissociation processes are expected to dominate
(see peak at $E_T=0$).
Since
for $\eta^* = 0$ essentially the
same detector regions are used in
DIS
and photoproduction interactions
the agreement is not influenced by detector effects.
Following this observation, it is interesting to compare
the fraction of diffractive events in
photoproduction and DIS.
An effective way to detect diffractive processes at HERA is
via the $\eta^*_{max}$ variable, which
is the maximum pseudorapidity in the $\mbox{$\gamma^{*}p$}$ CMS frame,
of a reconstructed track or calorimetric
cluster with an energy larger than 400 MeV observed in the detector.
In diffractive events $\eta^*_{max}$ indicates the maximum pseudorapidity of
secondary hadrons from photon fragmentation and is related
to the fraction $x_p$
of the initial proton momentum carried by the diffractively scattered
proton via $\eta^*_{max} \sim \ln(1-x_p) + C$~\cite{kaidalov}.
The comparison of the $\eta^*_{max}$ distributions for DIS and photoproduction
interactions is shown in Fig.~\ref{ET-MAX}b.
The figure shows spectra which fall off rapidly
from $\eta^*_{max} = 1$ with
decreasing $\eta^*_{max}$. This part of the spectra
can be described by non-diffractive DIS and
photoproduction interactions as demonstrated in~\cite{h1rapgap,h1diffphoto}.
For $\eta^*_{max} \sim -1$ the
spectra level off to a constant plateau. The events with this and
lower $\eta^*_{max}$ values have been shown to originate predominantly from
a diffractive mechanism~\cite{h1rapgap2,h1diffphoto}. Below
$\eta^*_{max}<-3 $ the photoproduction and DIS data are
affected differently by the detector
acceptance and therefore not included in this analysis.
As a consequence of this cut the low mass resonance region is avoided as well.
For diffractive events
the distribution $1/N\cdot dN/d\eta^*_{max} \sim
(1/\sigma_{tot})(d\sigma/d(1-x_p))(1-x_p)$ is expected to be roughly
independent of $\eta^*_{max}$ if the
diffractive cross section $d\sigma/d(1-x_p)$ is approximately
inversely proportional to $(1-x_p)$, see ref.~\cite{h1rapgap2} \footnote
{It was shown that for DIS events $d\sigma/d(1-x_p)$
is proportional to $(1-x_p)^n$ with $n=1.19\pm0.06\pm0.07$}.
This lack of dependence of $1/N\cdot dN/d\eta^*_{max}$ on $\eta^*_{max}$
is itself often taken as the signature of diffraction~\cite{bjdiff}.
Fig.~\ref{ET-MAX}b shows clearly that indeed at low $\eta^*_{max}$ both for
DIS and photoproduction the differential distribution
$1/N\cdot dN/d\eta^*_{max}$ is roughly independent of $\eta^*_{max}$.
Furthermore the
relative contribution of the photon high mass diffractive dissociation
for photoproduction and DIS interactions
is found to be the same to within 15-20\%.
This agreement is not affected by detector effects since
for a given $\eta^*_{max}$ value largely the
same detector regions are explored in
DIS
and photoproduction interactions.
This observation is at the same level of agreement as measurements made for
hadron-hadron interactions, and is
in accord with the expectation of
approximate factorization of high mass
diffractive cross sections, as observed in hadron collisions and
explained with triple Regge phenomenology~\cite{triple}:
the ratio of the differential diffractive cross section,
$d\sigma/d(1-x_p)$, to the total
cross section is approximately independent of the type of
dissociating
hadron.
The fact that this factorization rule also seems
to hold for virtual photon
interactions gives an additional
argument in favour of the validity of a universal
hadron-like description of low-$x$ DIS, real photoproduction, and
hadron-hadron collisions.
\section{Conclusions}
The comparison of photoproduction and low-$x$ DIS data
at the $ep$ collider HERA reveals striking similarities
in the energy flow of the hadronic final state
and relative rate of high mass diffractive
dissociation of the photon.
The $W$ dependence of the
transverse energy density in the central rapidity region is found to
be similar to that seen in high energy hadron-hadron collisions.
These findings are consistent with the
hadronic picture of the photon which can therefore be considered to be
complementary to the conventional deep-inelastic scattering picture, even
at large $Q^2$ values.
This picture gives a description of
the transition region from the high $Q^2$
perturbative region to the low $Q^2$ non-perturbative region, and provides
a basis for a universal description of hadron-hadron, real photon-hadron
and virtual photon-hadron high energy interactions.
\section*{Acknowledgements}
We are grateful to the HERA machine group whose outstanding efforts
made this experiment possible. We appreciate the immense
effort of the
engineers and technicians who constructed and maintained the detector.
We thank the funding agencies for their financial support of the
experiment. We wish to thank the DESY directorate for the support
and hospitality extended to the non-DESY members of the collaboration.
We thank further J. Bartels, A. Kaidalov, J. Kwiecinski and
M. Ryskin for useful
discussions.
{\Large\normalsize}
\begin{footnotesize}
|
\section{INTRODUCTION}
The visual search for absorption lines in the gamma ray burst spectra
accumulated by the Burst and Transient Source Experiment (BATSE) on the {\it
Compton Gamma Ray Observatory (CGRO)} has resulted in a number of line
candidates, but no definitive detections (Palmer et al. 1994, hereafter
Paper~I). Here we present the results of a systematic evaluation of the line
candidates identified thus far, and discuss the two most significant
candidates. The existence of lines in the BATSE spectra is one of the major
unresolved observational issues confronting BATSE since the absorption lines
reported by the KONUS (Mazets et al. 1981), {\it HEAO~1} (Hueter 1987) and {\it
Ginga} (Murakami et al. 1988) detectors were attributed to cyclotron absorption
in $\sim$10$^{12}$ gauss magnetic fields (\cite{wang89}), which suggests that
bursts originate on neutron stars, the only known anchors for teragauss fields.
This paper is part of a series describing and analyzing our search for lines in
the BATSE data. Paper~I reported the absence of line detections by the visual
search of the BATSE spectra. Band et al. (1994, hereafter Paper~II) developed
the methodology for comparing the absence of a BATSE detection with the
detections reported by previous missions; a preliminary calculation shows that
the apparent discrepancy between BATSE and {\it Ginga} is not yet compelling.
Band et al. (1995, hereafter Paper~III) demonstrated that BATSE would have
detected the lines that {\it Ginga} observed; the simulations in Paper~III also
provide the probabilities for detecting lines in the BATSE spectra which are
necessary for a comprehensive comparison between BATSE and previous detectors.
The visual search provides neither the statistics necessary for this
comprehensive search, nor measures of its completeness, as is evident from the
candidate evaluation presented here. Therefore, we have begun a more
sophisticated computerized search (Schaefer et al. 1994).
The evaluation of line candidates involves many instrumental and data analysis
details which are summarized in \S 2. We first present the results of the
candidate evaluation (\S 3), and then discuss two particular candidates in
greater detail (\S 3.1-2).
\section{METHODOLOGY}
We search for line features in the spectra accumulated by BATSE's Spectroscopy
Detectors (SDs); the SDs have been described in detail elsewhere (Fishman et
al. 1989; Band et al. 1992). In brief, the SDs are simple 5" diameter by 3"
thick cylindrical NaI(Tl) scintillation detectors with an energy resolution of
19\% at 60~keV. An SD is found in each of BATSE's eight modules located at the
corners of the {\it CGRO} spacecraft. Each module also contains a Large Area
Detector (LAD) for detecting transient events, determining their positions, and
recording their time histories. For 4-10 minutes after a burst trigger, BATSE
accumulates 192 SD spectra with 256 channels spanning two energy decades with
a
low energy cutoff between 15 and 30~keV (the SHERB data type). These spectra
are gathered from four detectors, with a greater number from the most brightly
illuminated detectors. The accumulation times are based on the count rate.
Background spectra are accumulated every $\sim5$ minutes when BATSE is not in
burst mode (the SHER data type).
Unfortunately, a few channels just above each spectrum's low energy cutoff are
distorted by the SLED ({\bf S}D {\bf L}ow {\bf E}nergy {\bf D}istortion), an
electronic artifact discovered after launch (Band et al. 1992). For a typical
SD in its highest gain setting the upper end of the SLED is between 15 and
20~keV. Although the SLED can be partially mitigated by the calibration
software, the channels affected by the SLED will probably never be trusted for
more than determining the continuum below a line candidate. However, for the
purpose of establishing consistency among all detectors we do search for
features in the SLED-affected channels when lines have been found at the same
energy in other detectors.
For an observed spectral feature to be considered intrinsic to the incident
spectrum (i.e., a ``real'' feature of astrophysical interest), the probability
that it is a statistical fluctuation must be small, and the spectra accumulated
by all detectors must be consistent. Thus the first detection criterion is
that the feature is significant in at least one detector, while the second
criterion is consistency among all detectors which could have observed the
feature. Note that we do not require multiple independent detections (although
it is very rare that a feature would be observable by only one detector), nor
must a feature be significant in all detectors where it is observed. As we
showed in Paper~III, fluctuations can vary the apparent significance of an
absorption line. In addition, the probability that a line will be significant
enough to be considered a detection varies with the burst angle (the angle
between the burst and the detector normal). Thus we cannot expect significant
detections in all detectors.
We use the $F$-test to calculate the probability that an observed feature is a
fluctuation. As applied here, the $F$-test compares a spectral fit with a
continuum-only model to the fit with a continuum+line model. Assume the
continuum-only fit gives $\chi_1^2$ with $\nu_1$ degrees-of-freedom, while the
continuum+line fit results in $\chi_2^2$ with $\nu_2$. We define the $F$
statistic as (Martin 1971)
\begin{equation}
F = \left({{\chi_1^2-\chi_2^2}\over{\nu_1-\nu_2}}\right)
\left({{\nu_2}\over{\chi_2^2}}\right) \quad .
\end{equation}
This $F$ statistic is characterized by a normalized distribution which is a
function of $\nu_1-\nu_2$ and $\nu_2$; the integral over this distribution from
$F$ to infinity is the $F$-test probability $P(\ge F)$, which we define as the
significance of the feature, with a small probability indicating a more
significant feature. We choose a maximum probability $P(\ge F)$ as the
threshold for a candidate to be considered a detection. Note that $P(\ge F)$
is only the {\it a priori} probability of a fluctuation with the observed
parameters;
it does not consider all the possible ``trials'', the range of parameters
(e.g., time intervals, line centroids, widths, etc.) in which line-like
fluctuations could have occurred. If there are a large number of trials then
it is likely that in some trial there will be a rare fluctuation. The
probability that a fluctuation will produce a spurious line feature with a
given significance somewhere in our dataset is the product of $P(\ge F)$ and
the number of trials; therefore we require a very small $P(\ge F)$ to conclude
the line is real. The calculation of the number of trials in our data is a
difficult issue in standard ``frequentist'' statistics which we will address
later in this series of papers. We estimate there are of order $\sim 3000$
trials in our spectra, and therefore use a detection threshold of $P(\ge F)\le
10^{-4}$. While this threshold may not appear stringent enough since the
probability of a false positive is then $\sim30\%$, we also require consistency
among detectors (Paper~I). The expected distribution of fluctuations will be
analyzed in greater detail later in this series of papers.
Only $\sim 20\%$ of the bursts detected by BATSE have a sufficiently high count
rate to warrant detailed spectral analysis, which roughly corresponds to a LAD
peak count rate over 10,000 counts-s$^{-1}$ in the 50-300~keV band (i.e., LAD
discriminators 2 and 3). Thus of the over 1000 bursts detected, only $\sim$250
have been searched, and of these bursts only $\sim$50 were intense enough for
lines comparable to those observed by \hbox{\it Ginga}\ to be detectable.
Although the spacecraft telemetry provides SHERB spectra for the four most
brightly illuminated detectors, not all the data are appropriate for line
searches. First, the detectors must be run at high gain to extend the
accumulated spectra low enough to cover the energy range where previous
missions reported spectral features. For most of the mission, two of the eight
SDs have been operated at low gain (to satisfy other mission objectives), and
the gain of one of the nominally high gain detectors cannot be pushed high
enough to cover energies below $\sim30$~keV. Second, the burst angle may be
greater than $\sim 80^\circ$, the angle at which the spacecraft and the rest of
the BATSE module affects the incident flux and our detector response model
becomes unreliable. The SHERB data are from the SDs in the modules of the four
most brightly-illuminated LADs; these LADs normally have burst angles less than
$90^\circ$. However, the detector normals of the SDs and LADs are offset by
$18.5^\circ$, and therefore a module whose LAD has a burst angle less than
$90^\circ$ may have an SD with an angle greater than $90^\circ$. Earth-scatter
of GRB photons off the atmosphere only complicates matters by occasionally
increasing the count rate in detectors with large burst angles. Thus usable
spectra may be available from only two or three SDs.
The visual search inspects background-subtracted spectra on different time
scales. By adding together the basic SHERB spectra, we construct spectra for
the entire burst and for the major temporal substructures (e.g., each major
emission spike). In addition, spectra are produced on the shortest time scale
over which all the detectors are accumulated, that of the third and fourth rank
detectors (i.e., the SDs in the third and fourth most brightly illuminated
modules). Thus spectra are available from all relevant detectors over the same
time ranges, allowing comparisons on the same time scale. The resulting
background-subtracted spectra are inspected visually for both emission and
absorption lines over the entire energy range. Features which appear to be
very significant are studied further by inspecting spectra averaged over
different time ranges. Thus a feature might be identified in a spectrum
averaged over the third rank accumulation time (e.g., from 3.328 to 6.592~s
after the trigger) but might be most significant in a spectrum averaged over an
overlapping time range (e.g., from 4.160 to 7.360~s). This search is sensitive
to features which are significant on the searched time scales, or which are
apparent to the eye in the searched time intervals but become more significant
by varying the time range; thus this search is not sensitive to short lived
features which are washed out in spectra accumulated over longer periods, nor
to features which persist a long time but are not evident in the searched
spectra over shorter time periods. Our simulations indicate that the eye is
sensitive to features with an $F$-test significance of $10^{-2}$ or smaller
(i.e., $P(\ge F)\le 10^{-2}$). Since in this study we find that we identified
many candidates with significances of order 0.1, we should have detected
features which persisted over many of the searched spectra.
We evaluate the line candidates identified by the visual search. Because the
gain differs between detectors and varies with time, each burst's spectra must
be calibrated for each detector (Band et al. 1992). Our response matrices
include Earth-scatter. Background spectra are created channel-by-channel by
fitting 2nd, 3rd, or 4th order polynomials to spectra before and after the
burst. As always, background creation is something of an art; we evaluate the
calculated background by the quality of the fits to each channel. These fits
provide the background as a function of time; for background subtraction these
functions are integrated over the time interval of interest.
We fit the spectra with the standard Marquardt-Levenberg algorithm minimizing
$\chi^2$ (Bevington 1969, pp.~232-241; Press et al. 1992, pp.~678-683). In
brief, the parameters of a model photon spectrum are varied to minimize the
difference---quantified by $\chi^2$---between the calculated count spectrum
(the model photon spectrum folded through the detector response) and the
observed spectrum. We define $\chi^2$ with model variances, requiring
additional terms in the gradients used by the Marquardt-Levenberg algorithm
(see the appendix of Ford et al. 1995).
The continuum is usually fit with the four parameter ``GRB'' (for lack of a
better name) model (Band et al. 1993),
\begin{eqnarray}
N_C(E) =& A (E/100)^\alpha
\exp\left[-\left({E\over{E_0}}\right)\right]\quad , \quad
&E\le (\alpha-\beta)E_0 \\
=& A^\prime (E/100)^\beta \qquad , \quad &E> (\alpha-\beta)E_0 \quad
,\nonumber
\end{eqnarray}
where $A^\prime$ is chosen so that $N_C(E)$ is continuous at
$E=(\alpha-\beta)E_0$. In a few cases where we find $E_0$ to be very large or
$\beta$ is nearly equal to $\alpha$, we simplify the continuum model, either by
using a simple power law or by dropping the high energy power law in eq.~2 (the
COMP model); the decision to simplify the continuum can be justified in terms
of the effect on the reduced $\chi^2$. An apparent absorption line is usually
modeled as a multiplicative line,
\begin{equation}
F(E)=N_C(E)\, \exp\left(-\phi\left(E\right)\right) \quad ,
\end{equation}
and an emission line as an additive line,
\begin{eqnarray}
F(E)=N_C(E) + \phi(E) \quad ,
\end{eqnarray}
where $\phi(E)$ is the Gaussian line profile (with three parameters: amplitude,
line centroid and line width). For absorption lines whose intrinsic width
appears to be much narrower than the instrumental resolution we also use a
simple two parameter ``black'' line model with no flux over an energy range
centered on the line centroid; for very narrow lines the instrument is unable
to resolve the line profile.
We always fit the spectra over the greatest energy range possible. The low
energy cutoff is chosen to avoid the SLED while the highest available energy
channel is included. In Paper~III we found that extending the fitted energy
range to energies much greater than the features of interest increased the
significance of the features, in part because the continuum is better
determined. On the other hand, we were concerned that the choice of the low
energy cutoff $E_1$ could make a large difference in the calculated
significances. The exact value of $E_1$ which should be used is not automatic
because the upper end of the SLED artifact is not clear-cut, and analysts may
choose different values (in particular there is a tendency to choose round
numbers, e.g., 20~keV when the SLED ends at 17~keV). To test the effect of
varying $E_1$, we calculated the significance of the 54~keV candidate
absorption line in GB~920307 using different values of $E_1$, as is shown by
Figure~1. As can be seen, there is not a meaningful change in significance for
$E_1$ between 15 and 30~keV, and the significance decreases ($P(\ge F)$
increases) only as $E_1$ approaches the candidate line energy. Thus small
changes in $E_1$ should not make a large difference. Of course, the
significance will be sensitive to the choice of $E_1$ if the line energy is
near the low energy cutoff.
\section{RESULTS}
Table~1 describes the candidates and the fits to them, while Figure~2 presents
the distribution of their significances. Few candidates are more significant
than $P(\ge F)\le 0.01$, and none exceeds our detection threshold of $P(\ge F)
= 10^{-4}$, although two nearly satisfy the significance criterion; these two
are discussed in greater detail below.
The significance distribution of the line candidates in Figure~2 is shaped by a
number of components. First are the inevitable line-like statistical
fluctuations, with many low significance fluctuations and few at the high end
(i.e., small values of $P(\ge F)$). The normalization of this distribution is
proportional to the number of trials in the burst dataset, a poorly determined
quantity. Second is the distribution produced by real spectral lines (if lines
exist). These lines are characterized by a (currently unknown) distribution of
parameters. Paper~III showed that as a result of the fluctuations in any given
observation, the same line may be observed by identical detectors with very
different significances. Thus the observed significance distribution is the
convolution of the line parameter distribution and the distribution of
significances for a given set of line parameters. Note that it is entirely
possible that many of our low significance (large $P(\ge F)$ value) candidates
which do not satisfy our detection criteria may be real lines. Finally,
incompleteness truncates the low significance end of the observed significance
distribution; most candidates with $P(\ge F)\sim 0.1$ will not be identified by
the search.
We suspect that most or all of our line candidates are fluctuations. Our
intention was to identify those features which appeared to have $P(\ge F)\le
0.01$; the large number of candidates with $P(\ge F)\ge 0.1$ leads us to
suspect the accuracy of the qualitative (i.e., ``eyeball'') assessment of the
significance of a line. We would expect equal numbers of emission line-like
and absorption line-like fluctuations. However, absorption features were more
readily recorded because of the previous reports of absorption lines in the
15-100~keV range (as discussed in Paper~II), although emission features up to
511~keV were also flagged (emission lines have also been reported by previous
missions). In addition, absorption lines may be more apparent to the eye than
emission lines of comparable strength because of the curvature of the burst
count spectra. Thus we find that the number of absorption line candidates (67)
exceeds that of emission line candidates (13).
The visual search does not provide the statistics of the lines which could have
been detected in the searched spectra. These statistics are necessary to
compare different missions (e.g., BATSE with \hbox{\it Ginga}) and to place constraints
on the frequency with which different types of lines occur (Paper~II). In
addition, the search does not produce measures of its completeness (e.g., the
fraction of features with $P(\ge F)\sim 10^{-2}$ which were identified by the
search). Since the visual search relies on human judgement which can be
affected by fatigue and distractions, there is always the fear that an
interesting candidate may have been missed. While we are confident that we
would have detected features with $P(\ge F)\le 10^{-4}$ in the inspected
spectra, we are relying on identifying less significant features in order to
detect lines which persist over many of the searched spectra. Thus we cannot
conclude definitively that detectable lines do not exist in the BATSE spectra.
For these reasons we have begun a computerized search (Schaefer et al. 1994)
which is unaffected by human subjectivity and which will provide relevant
statistics. In addition, this computerized search will consider almost all
possible spectra which can be formed from consecutive SHERB spectra.
\subsection{The Feature in GB~920315}
The most significant candidate absorption line is found in GB~920315 with
$P(\ge F) = 1.6 \times 10^{-4}$, just above our detection threshold. The
candidate is a feature at 80~keV in a spectrum accumulated between 1.152 and
1.728~s from SD1 (i.e., the SD in module 1), which is the 2nd rank detector
with a burst angle of 58.3$^\circ$. The 1st rank detector, SD3, was in a low
gain setting (and therefore did not cover 80~keV), and the 3rd and 4th rank
detectors, SD6 and SD7, had burst angles of 121.7$^\circ$ and 105.4$^\circ$,
respectively. Therefore, SD1 was the only SD capable of seeing a feature below
$\sim 300$~keV, making consistency between detectors automatic; had the feature
in SD1 met our significance criterion, it would have technically been a
detection. Features are not evident at 80~keV in LAD spectra (with poorer
spectral resolution but better statistics than SD spectra), although the
sensitivity of the LADs to lines of this type has not yet been determined. As
can be seen from the count spectrum in Figure~3, the candidate at 80~keV draws
much of its significance from one particularly low channel; however, this low
channel is surrounded by a number of channels which dip below the general
trend.
As we argued above, there is a probability of order $\sim 30$\% of a
fluctuation with a significance of $\sim 10^{-4}$ somewhere within our data
set, and therefore this feature can be attributed to a statistical fluctuation.
Indeed, we would not have been comfortable declaring this candidate a detection
had its significance been a factor of 1.6 smaller because of the absence of
confirmation by other detectors and the importance of one low point. Our first
definitive detection must be more convincing.
\subsection{The Feature in GB~930506}
The candidate in GB~930506 has attracted a great deal of attention (Ford et al.
1994; Preece et al. 1994; Freeman \& Lamb 1994; Paper~I). An absorption
feature was found at 55 keV in SD2; Figure~4 presents the count spectrum. The
feature has been evaluated with fits by various continuum and line models over
different energy ranges; in many cases the resulting $P(\ge F)$ satisfied our
detection criterion.
However, the feature in SD2 is absent in the spectra from other BATSE
detectors. Two other SDs and two LADs, all with smaller burst angles than SD2,
also viewed the burst (see Table~2). Unfortunately, the SD with the smallest
burst angle (SD3) was at a low gain setting, and therefore did not cover the
energy of the line candidate. In SD7 the upper edge of the SLED fell at the
energy of the line in SD2. Using the methodology of Band et al. (1992), Ford
et al. (1994) developed an {\it ad hoc} correction for the SLED in SD7. With
this correction, no line is evident at 55~keV in SD7; fitting the SD7 spectrum
with a continuum+line model found a weak line at 55.9~keV with $P(\ge F)=0.73$.
The probability that the line observed in SD2 (i.e, with the parameters fit to
the SD2 spectrum) would be as nearly unobservable as the line fit to SD7 is
less than one percent; specifically, none of 200 simulated SD7 spectra using
the parameters of the SD2 fit resulted in $P(\ge F)$ as large as 0.73.
Similarly, a joint fit to SD2 and SD7 does not find a significant line feature
(Briggs 1995). No lines are evident in the two LADs with small burst angles
(Preece et al. 1994).
We reanalyzed the feature in SD2 with the procedures used to evaluate the other
line candidates; the results are presented by Table~3. We tried three
different continuum models---a power law, COMP and GRB---and two different
line
models---multiplicative and ``black.'' The background varied significantly
over the $\sim 2000$~s before and after the burst, and there were a
number of reasonable choices of background spectra from which we could
interpolate the background at the time of the candidate. Consequently we
calculated the line significances for the six continuum+line models (three
continuum and two line models) for two different interpolated backgrounds; as
can be seen from Table~3, the resulting line significances are qualitatively
similar using the two background spectra. From the reduced $\chi^2_\nu$ for
the different fits, also presented in Table~3, it is clear that the power law
fits are poor, while the COMP and GRB fits are acceptable ($\chi^2_\nu\sim 1$).
While the power law fits are highly significant, $P(\ge F)\sim 10^{-7}$, the
COMP and GRB fits do not meet our significance criterion for a detection. It
is our experience that when a continuum model does not fit a spectrum well, the
line model assists in fitting the observed continuum; hence a poor continuum
model results in a significant line fit. Thus we conclude that the high
significance of the power law fits is not an accurate representation of the
line significance. The fitted line is narrower than the
instrumental resolution (a resolution of $\sim 11$~keV as opposed to a line
width of $\sim3.9$~keV), and therefore using the ``black'' line model, with one
fewer parameter than the multiplicative line model, is more significant by a
factor of $\sim 4$. Nonetheless, the line does not satisfy our significance
criterion for any of the acceptable line fits.
In their extensive analysis of this candidate, Freeman \& Lamb (1994) performed
fits to SD2 and joint fits to SD2+SD7, SD2+SD3, and SD2+SD3+SD7 using both the
COMP and GRB continua. The significances of the joint fits to SD2+SD3 are
$1.6\times 10^{-5}$ and $7.4\times 10^{-5}$ for the COMP and GRB continua,
respectively. We do not believe this affects our conclusion that the feature
is not significant. First, the relevant significance is the largest $F$-test
probability that results from a reasonable continuum model; to show the feature
is significant we must prove that it cannot be produced by a reasonable
continuum shape. Thus we should use the significance from the joint fit with
the GRB continuum which just barely satisfies our first criterion. Second,
adding a second detector doubles the number of degrees-of-freedom. As
discussed in Paper~III, extending the energy range of a fit can improve the
$F$-test probability both by better determining the continuum, thereby
increasing $\Delta \chi^2$ between the continuum and continuum+line fits, and
by increasing the degrees-of-freedom. The first effect is a true reflection of
the line significance, while the second effect is an artifact of the $F$-test.
We found in Paper~III that a factor of two increase in the number of
degrees-of-freedom decreased the $F$-test probability by an order of magnitude;
therefore we conclude that most of the improvement of the significance of the
SD2+SD3 joint fit over the SD2 fit results from this degrees-of-freedom effect.
We therefore conclude that the line candidate in GB~930506 meets neither of our
detection criteria. The $F$-test probability for the line feature in SD2 is
small but not so small as to lead us to reject the possibility that the feature
is a rare statistical fluctuation.
\section{SUMMARY}
We have evaluated the line candidates identified by the still-ongoing visual
search of burst spectra observed by the BATSE SDs. We find that none of the
candidates satisfy our detection criteria; indeed most of the candidates are
not very significant.
Taken at face value, this evaluation of line candidates from the visual search
leaves us with no BATSE line detections, while KONUS, {\it HEAO~1} and \hbox{\it Ginga}\
all reported detections. However, our approximate assessment of the
consistency between BATSE and \hbox{\it Ginga}\ (Paper~II), accurate to a factor of
$\sim$2, indicates that the apparent discrepancy between these two missions is
not yet compelling; a more definitive calculation is in progress. Thus based
on the BATSE observations there is no reason yet to question the detections
reported by previous missions. Clearly, we will be forced to conclude that
BATSE is discrepant with previous missions if no lines are detected after many
additional intense BATSE bursts are observed.
The visual search has not provided us with measures of its sensitivity; we do
not know what part of the space of line parameters has been probed. In
addition, the visual search is affected by the subjectivity of the human eye,
and there is always the fear that the search may have missed a significant
line. Consequently we have begun a new computerized search which should
overcome these limitations. Also, this search will consider progressively
longer accumulations, thereby increasing the range of lines which can be found.
Both the ongoing visual search and this computerized search will undoubtedly
generate new candidates, and perhaps ultimately a few detections.
\acknowledgments
We thank P.~Freeman and D.~Lamb for insightful discussions regarding GB~930506.
The work of the UCSD group is supported by NASA contract NAS8-36081.
\clearpage
\addtocounter{page}{2}
\setcounter{table}{1}
\begin{table*}
\begin{center}
\smallskip
\begin{tabular}{l c c}
\tableline
\tableline
Detector & Burst Angle\tablenotemark{a} &
Energy Range (keV)\tablenotemark{b} \\
\tableline
SD3 & 43.7$^\circ$ & 250-25000 \\
SD7 & 56.2$^\circ$ & 55-4800 \\
SD2 & 73.7$^\circ$ & 15-1350 \\
LAD3 & 28.8$^\circ$ & 35-1800 \\
LAD7 & 42.8$^\circ$ & 35-1800 \\
\end{tabular}
\end{center}
\caption{Detectors viewing the GB~930506 candidate}
\tablenotetext{a}{The angle between the detector normal and the burst.}
\tablenotetext{b}{The energy range covered by the detector.}
\end{table*}
\clearpage
\begin{table*}
\begin{center}
\begin{tabular}{l c c c c c}
\tableline
\tableline
\multicolumn{2}{c}{Continuum Model} &
\multicolumn{2}{c}{Black Line\tablenotemark{a}} &
\multicolumn{2}{c}{Multiplicative Line\tablenotemark{b}} \\
Type\tablenotemark{c} & $\chi^2_\nu$ & $P(\ge F)$ & $\chi^2_\nu$ &
$P(\ge F)$ & $\chi^2_\nu$ \\
\tableline
PL & 1.96688 & $1.28\times10^{-7}$ & 1.72630 & $6.62\times10^{-7}$ &
1.73506 \\
& 1.97315 & $1.04\times10^{-7}$ & 1.72861 & $2.85\times10^{-7}$ &
1.72738 \\
\tableline
COMP & 1.00083 & $2.46\times10^{-4}$ & 0.93837 & $9.30\times10^{-4}$ &
0.94320
\\
& 0.99953 & $1.87\times10^{-4}$ & 0.93489 & $7.98\times10^{-4}$ &
0.94063 \\
\tableline
GRB & 0.99894 & $3.20\times10^{-4}$ & 0.93850 & $1.18\times10^{-3}$ & 0.94324
\\
& 0.99772 & $2.41\times10^{-4}$ & 0.93502 & $9.74\times10^{-4}$ &
0.94043 \\
\end{tabular}
\end{center}
\caption{Significance of GB~930506 candidate in SD2}
\tablenotetext{a}{A line model with a rectangular profile
centered on the line centroid (2 parameters).}
\tablenotetext{b}{The line model is an exponential of a Gaussian line profile
(eqn.~3---3 line parameters).}
\tablenotetext{c}{The type of continuum model: PL---power law (2 parameters);
COMP---power law with exponential break (3 parameters); GRB---power law with
exponential break flattening to high energy power law (eqn.~2---4 parameters).
Note that PL is a subset of COMP, which in turn is a subset of GRB.}
\tablecomments{Fits to the line candidate at 55~keV in the SD2 spectrum
accumulated over 6.080-9.920~s. There are two entries for each case
corresponding to two different background spectra. The spectrum was fit over
232 channels covering 15.3--1351~keV.}
\end{table*}
\clearpage
|
\section{Introduction}
\ Instanton-induced processes in the standard electroweak theory are known
to lead to baryon and lepton number violation. Although 't Hooft \cite{h1}
showed several years ago that such phenomena are utterly suppressed by the
factor $e^{-8\pi ^2/g^2}$ ($g$ is the $SU(2)_L$ gauge coupling constant),
several authors \cite{c2} explored the possibility that this exponential
suppression factor can be overcome at high energies by the phase space which
corresponds to multiparticle production. The key observation is that the $%
SU(2)_L$- instanton induces to leading semiclassical approximation effective
point-like interactions which involve all the fermionic left-handed $SU(2)_L$%
-doublets of the theory (four per generation) and any number of Higgs and
gauge bosons.
The inclusive cross section of the baryon and lepton number violating two
fermion scattering can then be calculated as the imaginary part of the
forward $2\rightarrow 2$ scattering amplitude depicted in fig.1 \cite{p3}.
As a result, this inclusive cross section appears to grow exponentially with
energy and can conceivably become unsuppressed at energies of the order of
the sphaleron \cite{m4} mass. This leading order behavior may, however, be
drastically altered by higher order corrections well before the energy
reaches the sphaleron mass and consequently, these phenomena may remain
unobservable at all energies. Several authors \cite{a5} actually, have
suggested that this could be the case if multiinstanton corrections were to
be taken into account. Corrections to the $2\rightarrow 2$ scattering
amplitude consisting of linear instanton-antiinstanton chains in alternating
order were considered, with particle exchange allowed only between
successive instantons and antiinstantons. Dorey and Mattis \cite{d6},
however, using the valley method \cite{k7}, pointed out that inclusion of
non-nearest neighbor instanton-antiinstanton as well as instanton-instanton
and antiinstanton-antiinstanton interactions could render these chain graphs
unimportant at the relevant energies. This, however, may not happen if
Dorey's result on the non linear $O(3)$ $\sigma -$model applies in the
realistic case too. The imaginary part of the chain graphs being
ultraviolately divergent,
requires the introduction of an appropriate cut-off. Finally, these graphs
do not include initial state corrections and thus, are not expected to alter
the high energy behavior of the leading semiclassical approximation \cite{k7}%
,\cite{m8}.
In this work we choose to deal with a class of ladder graphs shown in fig.2.
The imaginary part of these graphs turns out to be finite and can, in
principle, be unambiguously calculated. They can, in some sense, be thought
as including initial state corrections too since the incoming particles
enter in different instanton vertices. In addition, such ladder graphs are
known to dominate the high-energy behavior in ordinary field theories.We
should emphasize, however, that including the ladder graphs of fig.2 does
not solve the problem of initial state corrections. Indeed, the entire
picture of separated instanton-antiinstanton chains is an uncontrolled
approximation at energies where such chains actually become important.
\section{The leading semiclassical approximation.}
Consider the inclusive cross section, $\sigma _{inc}$, of the B and L
violating reaction
\begin{equation}
\label{eq1}q+q\rightarrow (3n_g-2)\overline{q}+n_g\overline{l}%
+\,any\,\#\,of\,Higgses,
\end{equation}
where $n_g$ is the number of fermion generations ($n_g\geq 1$), $q$ and $l$
represent quarks and leptons respectively and we have ignored for simplicity
the production of gauge particles. In the leading instanton approximation, $%
\sigma _{inc}$ can be determined by first calculating the forward $%
2\rightarrow 2$ scattering amplitude in Euclidean space as shown in fig.1.
Then, $\sigma _{inc}$ is given by the imaginary part acquired by this
amplitude after rotating the total incoming Euclidean particle momentum $%
p=p_1+p_2$ to Minkowski space $\ (p^2\rightarrow e^{-\imath \pi }\,p^2)$.
The expression corresponding to the Euclidean space forward scattering
amplitude is \cite{c2},\cite{p3}
\begin{equation}
\label{eq2}C\int_0^\infty d\rho ^2\,\rho ^{2\alpha }\,e^{-\pi ^2v^2\rho
^2}\,\int_0^\infty d\tilde \rho ^2\,\tilde \rho ^{2\alpha }\,e^{-\pi ^2v^2%
\tilde \rho ^2}\,\int d^4x\,e^{-\imath px}\,F(x^2,\rho ^2,\tilde \rho ^2).
\end{equation}
$C$ is a constant given by
\begin{equation}
\label{eq3}C\propto \frac 1{\pi ^2}\,(32\pi ^2)^{4n_g-1}\,\left( \frac{8\pi
^2}{g^2}\right) ^8\,\mu ^{\frac{43-8n_g}3}\,e^{-\frac{16\pi ^2}{g^2(\mu )}%
},\ \alpha =(28n_g+7)/12,
\end{equation}
$\rho $ and $\tilde \rho $ are the scale sizes of the instanton and the
antiinstanton, \thinspace $x_\mu $ is their Euclidean separation, $v=246GeV$
the electroweak breaking scale, g the gauge coupling and $\mu $ is the
renormalization point. The function $F(x^2,\rho ^2,\tilde \rho ^2)$, to
leading semiclassical order (or for $x^2\rightarrow \infty $), can be
written as
\begin{equation}
\label{eq4}F(x^2,\rho ^2,\tilde \rho ^2)\equiv F(x^2)=e^{\kappa /x^2}\,\frac
1{(x^2)^n},
\end{equation}
where $\kappa =\pi ^2\rho ^2\tilde \rho ^2v^2$ and $n=3(2n_g-1)\geq 3$. The
exponential factor in the formula above, corresponds to the Higgses in the
final state of reaction (1), whereas the second factor corresponds to the $%
4n_g-2$ fermions which are also being produced. Since we are interested in
the high energy behavior of $\sigma _{inc}$, we can assume throughout this
work that all fermions and Higgs bosons are effectively massless.
We will first consider the integral over the instanton-antiinstanton
separation
\begin{equation}
\label{eq5}I^0(p^2)\equiv \int d^4x\,e^{-\imath px}\,F(x^2)\quad .
\end{equation}
This integral converges at infinity for $n\geq 3$ , but diverges badly at $%
x=0$. This virulent ultraviolet divergence in Euclidean space is due to the
attractive nature of the Coulomb potential $V(x^2)=-\kappa /x^2$ resulting
from Higgs particle exchange between the instanton and the antiinstanton and
is an artifact of the leading semiclassical approximation. We, thus, define
a regularized integral
\begin{equation}
\label{eq6}I_\delta ^0(p^2)=\int_{x^2\geq \delta ^2}\,d^4x\,e^{-\imath
px}\,F(x^2)
\end{equation}
by removing from the range of integration a four-dimensional disc of finite
radius $\delta >0$ centered at the origin. Performing the angular
integrations, we obtain
\begin{equation}
\label{eq7}I_\delta ^0(p^2)=2\pi ^2\int_\delta ^\infty dr\,e^{\kappa
/r^2}\,r^{3-2n}\,G_{0\,2}^{1\,0}(\frac{p^2r^2}4|0,-1),
\end{equation}
\noindent where
\begin{equation}
\label{eq8}G_{0\,2}^{1\,0}(\frac{p^2r^2}4|0,-1)=\frac 1{2\pi \imath }%
\,\int_Ldz\left( \frac{p^2r^2}4\right) ^z\frac{\Gamma (-z)}{\Gamma (2+z)}
\end{equation}
is the well-known Meijer function. The contour $L$ is a loop starting and
ending at $+\infty $ and encircling the poles of $\Gamma (-z)$ once in the
negative direction. Since $p^2r^2/4$ is positive, one can show that the
contour $L$ can be distorted to become parallel to the imaginary axis and
lying in the strip $-1/2<Re(z)<0$. Then substituting eq.(8) in eq.(7) and
interchanging the order of integrations, we get
\begin{equation}
\label{eq9}I_\delta ^0(p^2)=\frac 1{2\pi \imath }\,\int_{c-\imath \infty
}^{c+\imath \infty }\,dz\,\left( \frac{p^2}4\right) ^z\,\frac{\Gamma (-z)}{%
\Gamma (2+z)}\,\pi ^2\,(-\kappa )^{z-n+2}\,\gamma (n-z-2,-\kappa /\delta
^2),
\end{equation}
$$
-1<c<0.
$$
Here, $\gamma (\alpha ,x)$ is the incomplete gamma function which can be
expressed as
\begin{equation}
\label{eq10}\gamma (\alpha ,x)=x^\alpha \,\Gamma (\alpha )\,\gamma
^{*}(\alpha ,x)\ ,
\end{equation}
with $\gamma ^{*}(\alpha ,x)$ being an analytic function of $\alpha $ and $x$%
.Eq.(9) then becomes
\begin{equation}
\label{eq11}I_\delta ^0(p^2)=\frac{\pi ^2\delta ^{4-2n}}{2\pi \imath }%
\,\int_{c-\imath \infty }^{c+\imath \infty }\,dz\left( \frac{p^2\delta ^2}4%
\right) ^z\,\frac{\Gamma (-z)\,\Gamma (n-2-z)}{\Gamma (2+z)}\,\gamma
^{*}(n-z-2,-\kappa /{\delta ^2}),
\end{equation}
$$
-1<c<0.
$$
Notice, that after interchanging the order of integrations, the range of $c$
can be extended.The imaginary part acquired by $I_\delta ^0(p^2)$ after
rotating $p^2$ to Minkowski space $(p^2\rightarrow e^{-\imath \pi }p^2)$,
comes from the $\ln p^2$ terms in the series expansion of the right-hand
side of eq.(11). These terms are produced by the double poles of the
integrand in eq.(11) at $z=m$,$~m=n-2,n-1,n,\ldots $ \nolinebreak. The
result is
$$
ImI_\delta ^0(e^{-\imath \pi }p^2)=\pi ^3\delta ^{4-2n}\sum_{m=n-2}^\infty
\left( \frac{p^2\delta ^2}4\right) ^m\,\frac{(-1)^{m-n+2}}{%
m!\,(m+1)!\,(m-n+2)!}\,\times
$$
\begin{equation}
\label{eq12}\gamma ^{*}(n-m-2,-\kappa /{\delta }^2).
\end{equation}
Using eq.(10), one can show that
\begin{equation}
\label{eq13}\gamma ^{*}(-\alpha ,x)=x^\alpha ,\quad \alpha =0,1,2,\ldots ,
\end{equation}
which implies
\begin{equation}
\label{eq14}ImI_\delta ^0(e^{-\imath \pi }p^2)=\pi ^3{\kappa }%
^{2-n}\,\sum_{m=n-2}^\infty \left( \frac{p^2\kappa }4\right) ^m\,\frac 1{%
m!\,(m+1)!\,(m-n+2)!}\,\cdot
\end{equation}
We now perform the integrals over the sizes of the instanton and the
antiinstanton to get the well-known semiclassical result.
\begin{equation}
\label{eq15}\sigma _{inc}^0=\frac{\pi ^3C}{(\pi ^2v^2)^{4+2\alpha -n}}%
\,\sum_{m=n-2}^\infty \left( \frac{p^2}{4\pi ^2v^2}\right) ^m\,\frac{[\Gamma
(\alpha +m-n+3)]^2}{m!\,(m+1)!\,(m-n+2)!}\,\ \cdot
\end{equation}
It is important to note that the $\delta $-dependence of the imaginary part
of $I_\delta ^0(e^{-\imath \pi }p^2)$ has completely disappeared in eq.(15).
Consequently, $\sigma _{inc}^0$ is $\delta $-independent for any $\delta >0$
and its exponential growth with energy results only from the boundary at
infinity of the Euclidean $x$-space in eq.(6).The virulent ultraviolate
divergence, as well as the contribution of any ''finite'' part of the
Euclidean $x$-space in eq.(6), do not seem to play any essential role.
\section{The ladder graphs.}
We will now turn to the calculation of the Euclidean space ladder graphs
shown in fig.2. The graphs, after rotating $p^2\rightarrow e^{-\imath \pi
}\,p^2$ and taking the imaginary part, constitute an important class of
multiinstanton corrections to the leading semiclassical approximation of $%
\sigma _{inc}$. The forward scattering amplitude which corresponds to the
ladder graph with $k+1$ rungs $(k=0,2,4,\ldots )$ is given by
\begin{equation}
\label{eq16}C^{k+1}\,\prod_{\imath =0}^k\,\prod_{\jmath =1}^k\,\int_0^\infty
\,d\rho _\imath ^2\,\rho _\imath ^{2\alpha }\,e^{-\pi ^2v^2\rho _\imath
^2}\,\int_0^\infty \,d\widetilde{\rho _\imath }^2\,\widetilde{\rho _\imath }%
^{2\alpha }\,e^{-\pi ^2v^2\widetilde{\rho }_\imath ^2}\times
\end{equation}
$$
\int \,d^4q_\jmath \,\frac 1{q_\jmath ^2}\,\int \,d^4x_\imath \,e^{-\imath
(q_{\imath }+q_{\imath +1})x_\imath }\,F(x_\imath ^2,\rho _\imath ^2,
\widetilde{\rho }_\imath ^2)\quad ,
$$
where $q_0=p_1$, $q_{k+1}=p_2$ and the four momenta $q_\jmath \,(\jmath
=1,2,\ldots ,k)$ are indicated in fig.2, $\rho _\imath \,,\widetilde{\rho
_\imath }\,\,\,(\imath =0,1,\ldots ,k)$ are the scale sizes of the $i$-th
instanton and the $i$-th antiinstanton respectively and $x_\imath $ is their
Euclidean separation.
We will first consider the integrals over the instanton-antiinstanton
separations
\begin{equation}
\label{eq17}I_\delta ^k(p^2)=\prod_{\jmath =1}^k\,\int \,d^4q_\jmath \,\frac
1{q_\jmath ^2}\,\,\prod_{\imath =0}^k\,\int\limits_{x_\imath ^2\geq \delta
^2}\,d^4x_\imath \,e^{-\imath (q_\imath +q_{\imath +1})}\,F(x_\imath ^2,\rho
_\imath ^2,\widetilde{\rho }_\imath ^2)\quad ,
\end{equation}
where the Euclidean space ultraviolet divergences are again regularized by
restricting the $x_i$-integrations to $x_\imath ^2\geq \delta ^2$. Repeating
the analysis of the previous section, we obtain\
$$
I_\delta ^k(p^2)=\frac{\pi ^{2(k+1)}\,\delta ^{(4-2n)(k+1)}}{(2\pi \,i)^{k+1}%
}\,\prod_{\imath =0}^k\,\int_{c_\imath -\imath \infty }^{c_\imath +\imath
\infty }\,dz_\imath \,\left( \frac{\delta ^2}4\right) ^{z_\imath }\,\frac{%
\Gamma (-z_\imath )\,\Gamma (n-2-z_\imath )}{\Gamma (2+z_\imath )}\times
$$
\begin{equation}
\label{eq18}\gamma ^{*}(n-z_\imath -2,-\frac{\kappa _\imath }{\delta ^2}%
)\,\,A_k(z_0,z_1,\ldots ,z_k)
\end{equation}
where $\kappa _\imath =\pi ^2\rho _\imath ^2\,\widetilde{\rho }_\imath
^2\,v^2$ and (see fig.2)
\begin{equation}
\label{eq19}A_k(z_0,z_1,\ldots ,z_k)=\prod_{\jmath =1}^k\,\int \,d^4q_\jmath
\,\frac 1{q_\jmath ^2}\,\prod_{\imath =0}^k\,(q_\imath +q_{\imath
+1})^{2z_\imath }\quad .
\end{equation}
The constants $c_\imath \ (\imath =0,1,\ldots ,k)$ which satisfy the
inequalities $-1<c_\imath <\nolinebreak0\,$(see eq.(11)) may and in fact
will have to be further restricted in eq.(18) so that $A_k(z_0,z_1,\ldots
,z_k)$ and the $z_\imath $ -integrals exist. For the moment we just assume
that there is some region of $z_\imath $ 's in which $A_k(z_0,z_1,\ldots
,z_k)$ exist and we restrict ourselves in this region. This assumption will
be proved to be correct a posteriori (see below). The $s=(p_1+p_2)^2$
dependence of $A_k$ at $p_1^2=p_2^2=0$ can be easily found from dimensional
arguments (there are no infrared divergences)\footnote{%
This semieuristic argument can be further corroborated by an explicit
although tedious calculation.}. We get
\begin{equation}
\label{eq20}A_k(z_0,z_1,\ldots ,z_k)=F_k(z_i)\,s^{k+\sum_i\,z_i}.
\end{equation}
It is easily seen that, for $p_1^2=p_2^2=0,$we also have
\begin{equation}
\label{eq21}p_2\cdot \frac \partial {\partial p_2}A_k=s\,\frac \partial {%
\partial s}A_k,
\end{equation}
and, thus,
\begin{equation}
\label{eq22}p_2\cdot \frac \partial {\partial p_2}A_k=(k+\sum_i\,z_i)A_k=z_k%
\,A_k-z_k\,A_k^{\,q_k}(z_0,\,z_1,\ldots ,z_k-1).
\end{equation}
\noindent $A_k^{\,q_k}$ denotes the expression $A_k$ with the $q_k$-
propagator ommited. Eq.(22 ) then gives
\begin{equation}
\label{eq23}A_k=-\frac{z_k}{k+\sum_{i\neq k}\,z_i}\,A_k^{\,q_k}(z_0,\,z_1,%
\ldots ,z_k-1)\ ,\ k+\sum_{i\neq k}\,z_i\neq 0\ \cdot
\end{equation}
The $q_k$-integration in $A_k^{\,q_k}$ can now be performed :
$$
\int d^4q_k\,[(q_k+p_2)^2]^{z_k-1}\,[(q_{k-1}+q_k)^2]^{z_{k-1}}=
$$
\begin{equation}
\label{eq24}\pi ^2\frac{\Gamma (-1-z_k-z_{k-1})}{\Gamma (1-z_k)\,\Gamma
(-z_{k-1})}\,{\rm B}(1+z_k,2+z_{k-1})\,[(q_{k-1}-p_2)^2]^{1+z_k+z_{k-1}}{\rm %
,}
\end{equation}
\noindent for $-1<Re(z_k)<1\,,\ -2<Re(z_{k-1})\,,\ Re(z_k)+Re(z_{k-1})<-1.$%
We then obtain the reccurence formula
$$
A_k(z_0,z_1,\ldots ,z_k)=\frac{\pi ^2}{\,[k+\sum_{j\neq k}z_j]}\,\frac{%
\,\Gamma (-1-z_k-z_{k-1})}{\,\Gamma (-z_k)\,\Gamma (-z_{k-1})}\times
$$
\begin{equation}
\label{eq25}\,B(1+z_k,\,2+z_{k-1})\,A_{k-1}(z_0,z_1,\ldots
,z_{k-2},\,1+z_k+z_{k-1})\ ,
\end{equation}
where $Re(z_k)<0$ and $A_k\quad (k=1,3,\ldots )$ is also defined by eq.(19)
but with $q_{k+1}=-p_2$ . Note that eq.(25) obviously holds for $%
k=1,3,5,\ldots $ too. Introducing the function
\begin{equation}
\label{eq26}D_k(z_0,z_1,\ldots ,z_k)=\prod_{\imath =0}^k\,\frac{\Gamma
(-z_i) }{\Gamma (2+z_\imath )}\,A_k(z_0,z_1,\ldots ,z_k)\quad k=0,1,2,\ldots
,
\end{equation}
the reccurence formula in eq.(25 ) takes the simple form
\begin{equation}
\label{eq27}D_k(z_0,z_1,\ldots ,z_k)=\frac{\pi ^2}{(1+z_k)(k+\sum_{\jmath
\neq k}z_\jmath )}\,\,D_{k-1}(z_0,z_1,\ldots ,z_{k-2},1+z_{k-1}+z_k)
\end{equation}
and can be easily solved to give
$$
D_k(z_0,z_1,\ldots ,z_k)=\pi ^{2k}\,\prod_{m=1}^k\,\frac 1{(m+z_0+\cdots
+z_{m-1})(k+1-m+z_m+\cdots +z_k)}
$$
\begin{equation}
\label{eq28}\times \frac{\Gamma (-k-\sum_{\imath =0}^k\,z_\imath )}{\Gamma
(2+k+\sum_{\imath =0}^k\,z_\imath )}\,s^{k+\sum_iz_\imath }.
\end{equation}
This formula holds provided that
$$
-2<Re(z_m)\quad m=0,1,\ldots ,k-1\ ;
$$
$$
-1<k-m+Re(z_m+\cdots +z_k)<0\quad m=1,2,\ldots ,k)\ ;
$$
$$
k+Re(z_0+\cdots +z_k)<0\ ;
$$
\begin{equation}
\label{eq29}m+z_0+\cdots +z_{m-1}\neq 0\quad m=1,2,\ldots ,k
\end{equation}
as can be easily deduced from the restrictions which follow eqs.(19), (24)
and (25). Substituting eq(28) in eq.(18) we obtain
$$
I_\delta ^k(s)=\frac{4^k\pi ^{2(2k+1)}}{(2\pi \imath )^{k+1}}\,\prod_{\imath
=0}^k\,\int_{c_\imath -\imath \infty }^{c_\imath +\imath \infty }\,dz_\imath
\,(\delta ^2)^{2-n+z_\imath }\,\Gamma (n-2-z_i)\,\gamma ^{*}(n-2-z_i,-\frac{%
\kappa _\imath }{\delta ^2})\times
$$
\begin{equation}
\label{eq30}\prod_{m=1}^k\frac 1{(m+z_0+\cdots +z_{m-1})(k+1-m+z_m+\cdots
+z_k)}\times
\end{equation}
$$
\frac{\Gamma (-k-\sum_{\imath =0}^kz_\imath )}{\Gamma (2+k+\sum_{\imath
=0}^kz_\imath )}\,\left( \frac s4\right) ^{k+\sum_\imath z_\imath }
$$
with $-1<c_\imath <0\quad (\imath =0,1,\ldots ,k)$ and $k+c_0+\cdots +c_k<0$.
The $z_\imath $- integrals can be evaluated by collapsing their contours to
the right and using residue calculus. Since we are only interested in the
imaginary part acquired by the amplitude when $s\rightarrow e^{-\imath \pi
}s $ , we only keep contributions proportional to $\ln s$. The relevant
contributions to the first $k\,$ $z_\imath $-integrals $(\imath =0,1,\ldots
,k-1)$ then come from the simple poles of the functions $\Gamma
(n-2-z_i)\,(\imath =0,1,\ldots ,k-1)$ whereas the $z_k$-integral gets
contributions from the double poles of the product $\,\Gamma (n-2-z_k)\Gamma
(-k-\sum_{\imath =0}^kz_\imath )$ where the first $k\,\ z_\imath $'s $%
(\imath =0,1,\ldots ,k-1)$ have already been substituted by integers. The
final result is%
$$
Im\,I_\delta ^k(se^{-\imath \pi })=4^k\pi ^{4k+3}\sum_{n-1\leq
l_0,l_1,\ldots ,l_k}\,\prod_{\imath =0}^k\,\frac{(\kappa _\imath )^{l_\imath
+1-n}}{(l_\imath +1-n)\,!}\times
$$
\begin{equation}
\label{eq31}\prod_{m=1}^k\,\frac 1{[l_0+\cdots +l_{m-1}][l_m+\cdots +l_k]}%
\times \frac{\left( s/4\right) ^{\sum_{\imath =0}^k\,l_\imath -1}}{%
[\sum_{\imath =0}^k\,l_\imath -1]\,!\,[\sum_{\imath =0}^k\,l_\imath ]\,!}
\end{equation}
and turns out to be again $\delta $-independent. Performing the $\rho
_\imath ^2\,,\widetilde{\rho }_\imath ^2$ -integrals in eq.(16) we finally
obtain the contribution of the ladder graph with $k+1$ rungs $%
(k=0,2,4,\ldots )$ to $\sigma _{inc}$ :%
$$
\sigma _{inc}^k=\frac 1{\pi \,s}\,\left[ \,4\pi ^4\,C\,(\pi
^2v^2)^{n-3-2\alpha }\right] ^{k+1}\,\,\sum_{n-1\leq l_0,l_1,\ldots
,l_k}\,\prod_{\imath =0}^k\frac{\left[ \Gamma (l_\imath +\alpha +2-n)\right]
^2}{\Gamma (l_\imath +2-n)}\,\times
$$
\begin{equation}
\label{eq32}\prod_{m=1}^k\,\frac 1{[l_0+\cdots +l_{m-1}][l_m+\cdots +l_k]}%
\times \frac{\left( s/4\pi ^2v^2\right) ^{\sum_{\imath =0}^k\,l_\imath }}{%
[\sum_{\imath =0}^k\,l_\imath -1]\,!\,[\sum_{\imath =0}^k\,l_\imath ]\,!}
\end{equation}
The multiple Series found for $\sigma _{inc}^k$, as it stands, looks very
complicated to be handled. We shall attempt to get an estimate by finding
suitable upper and lower bounds. In order to achieve this, we shall make
extensive use of the inequalities%
$$
l_0\,l_1\cdots l_k\leq \frac 1{(k+1)!}\,(l_0+l_1+\cdots +l_k)^{k+1}\quad ,
$$
\begin{equation}
\label{eq33}\frac 1{l_0\,l_1\cdots l_k}>\frac{(k+1)!}{(l_0+l_1+\cdots
+l_k)^{k+1}}\,\cdot
\end{equation}
Taking into account that $\alpha >2$ and defining $a=[\alpha ]+1$ we find
that
\begin{equation}
\label{eq34}\sigma _l^k<\sigma _{inc}^k<\sigma _u^k\,,
\end{equation}
where
$$
\sigma _l^k=\frac 1{\pi \,s}\,D^{k+1}\,2^k\,\,\,\sum_{n-1\leq l_0,l_1,\ldots
,l_k}\,\frac 1{\left[ \sum_{\imath =0}^kl_\imath \right] ^{n(k+1)-1}}\,\frac{%
\prod_{\imath =0}^k\Gamma (l_\imath )}{\left[ \Gamma (\sum_{\imath
=0}^kl_\imath )\right] ^2}\left[ \frac s{4\pi ^2v^2}\right] ^{\sum l_\imath
}\,,
$$
$$
\sigma _u^k=\frac 1{\pi \,s}\,D^{k+1}\frac 1{(n-1)^{2k}}\frac 1{(k+1)^{2a}}%
\frac 1{(k!)^{2(a+1)}}\left( \frac{a^a}{a!}\right) ^{2(k+1)}\times
$$
\begin{equation}
\label{eq35}\sum_{n-1\leq l_0,l_1,\ldots ,l_k}\,\left[ \sum_{\imath
=0}^kl_\imath \right] ^{2a(k+1)-1}\,\frac{\prod_{\imath =0}^k\Gamma
(l_\imath )}{\left[ \Gamma (\sum_{\imath =0}^kl_\imath )\right] ^2}\left[
\frac s{4\pi ^2v^2}\right] ^{\sum l_\imath },
\end{equation}
with $D=4\pi ^4\,C\,(\pi ^2v^2)^{n-3-2\alpha }\,.$
It is now clear that the next step must be the study of the multiple Series
\begin{equation}
\label{eq36}\Sigma _k(x)=\frac 1x\sum_{l_\imath =1}^\infty \frac{\Gamma
(l_0)\cdots \Gamma (l_k)}{\left[ \Gamma (l_0+l_1+\cdots +l_k)\right] ^2}%
\,x^{l_0+l_1+\cdots +l_k}\,,\quad x=\frac s{4\pi ^2v^2}\,\,\cdot
\end{equation}
Then, $\sigma _l^k$ and $\sigma _u^k$ could be recovered by integrating or
differentiating $\Sigma _k(x)$ with respect to $\ln (x)$ a suitable number
of times. The fact that the $l_\imath $-summations in the definition of $%
\sigma _l^k$ and $\sigma _u^k$ start at $n-1$ cannot change the results in
any fundamental way.
The Laplace transform of $\Sigma _k(x)$ is%
$$
S_k(t)=\int_0^\infty dx\,e^{-xt}\Sigma _k(x)=
$$
\begin{equation}
\label{eq37}\sum_{l_\imath =1}^\infty \frac{\Gamma (l_0)\,\Gamma (l_1)\cdots
\Gamma (l_k)}{\Gamma (l_0+l_1+\cdots +l_k)}\,t^{-(l_0+l_1+\cdots +l_k)}
\end{equation}
We can now use the integral representation for the generalized beta function
\begin{equation}
\label{eq38}\frac{\Gamma (l_0)\,\Gamma (l_1)\cdots \Gamma (l_k)}{\Gamma
(l_0+l_1+\cdots +l_k)}=\int_0^1\,\prod_{\imath =0}^k\,d\alpha _{\imath
\;}\prod_{\jmath =0}^k\,\alpha _\jmath ^{l_\jmath -1}\,\delta (1-
{\textstyle \sum }\alpha_\imath )
\end{equation}
to get
\begin{equation}
\label{eq39}S_k(t)=\int_0^1\,\frac{\prod d\alpha _\imath }{\prod \alpha
_\imath }\delta (1-{\textstyle \sum } \alpha _\imath
)\,\sum_{l_\imath =1}^\infty \left[ \frac{\alpha _0}t\right] ^{l_0}\cdots
\left[ \frac{\alpha _k}t\right] ^{l_k}.
\end{equation}
The summations are now decoupled and can be readily performed provided that $%
\mid \alpha _\imath /t\mid <1.\quad $ We obtain
\begin{equation}
\label{eq40} S_k(t)=\int_0^1 \frac
{ \prod d\alpha _i\,\delta (1-{\textstyle \sum } \alpha _\imath ) }
{(t-\alpha _0)\,(t-\alpha _1)\cdots (t-\alpha _k) }
\quad \cdot
\end{equation}
The inverse Laplace transform of the expression above can be found and the
answer is
\begin{equation}
\label{eq41} S_k(x)=\int_0^1 \prod d\alpha _i\,\delta (1-{\textstyle \sum }
\alpha _\imath) \sum_{m=0}^k\,\frac {e^{\alpha _mx}}{P_m(\alpha _m)}
\end{equation}
where $P_m(y)$ is a polynomial given by
\begin{equation}
\label{eq42}P_m(y)=\,\frac{\prod_{\imath =0}^k(y-\alpha _\imath )}{y-\alpha
_m}\quad \cdot
\end{equation}
Now the $\alpha _\imath -$ integrations can be performed and we end up with
the following recursive formula
\begin{equation}
\label{eq43}\Sigma _k(x)=e^{\frac xk}\int_0^xdz\,e^{-\frac zk}\int_0^1d\rho
\,\Sigma _{k-1}(\rho \,z)\,\quad \cdot
\end{equation}
This integral equation can be transformed into an integrodifferential
equation
\begin{equation}
\label{eq44}\frac d{dx}\Sigma _k(x)=\frac 1k\,\Sigma _k(x)+\frac 1x%
\,\int_0^x\,\Sigma _{k-1}(z)\,dz\,\,
\end{equation}
or a differential equation
\begin{equation}
\label{eq45}x\,\frac{d^2}{dx^2}\,f_k(x)\,+\,\left[ \left( 2-\frac 1k\right)
x+1\right] \,\frac d{dx}f_k(x)\,+\,\left( 1-\frac 1k\right) \left(
x+1\right) \,f_k(x)=\,f_{k-1}(x)\,,
\end{equation}
where
\begin{equation}
\label{eq46}f_k(x)=e^{-x}\,\Sigma _k(x)\,\cdot
\end{equation}
When $x$ is large, the differential equation reduces to
\begin{equation}
\label{eq47}\,\frac{d^2}{dx^2}\,f_k(x)\,+\,\left[ 2-\frac 1k\right] \,\frac d%
{dx}f_k(x)\,+\,\left[ 1-\frac 1k\right] \,f_k(x)=\frac{\,f_{k-1}(x)}x\quad
\cdot
\end{equation}
One particular integral of this equation is
\begin{equation}
\label{eq48}f_k(x)=\frac k{k-1}\,\frac 1{1+\frac d{dx}}\,\frac 1{1+\frac k{%
k-1}\frac d{dx}}\,\frac{\,f_{k-1}(x)}x
\end{equation}
with $\,f_1(x)=1\,.$
\noindent It is now obvious that the leading order solution of eq.(45 ) can
be written as
\begin{equation}
\label{eq49}f_k(x)=\frac k{x^{k-1}}
\end{equation}
leading to
\begin{equation}
\label{eq50}\Sigma _k(x)=k\ \frac{e^x}{x^{k-1}}\,\cdot
\end{equation}
We have already pointed out that the upper and lower bounds for $\sigma
_{inc}^k$ which were defined in eq.( 35) can be related to derivatives or
integrals of $\Sigma _k(x)$ with respect to$\ \ln x$. Such operations,
however, cannot modify the form of $\Sigma _k(x)$ in an essential way since
the exponential growth cannot be affected . We do expect a change in the
leading power of $x$ and of course the constant coefficient will be
different. We conclude that $\sigma _l^k\sim c_l^k\,e^x\,x^{-m_1},\,\sigma
_u^k\sim c_u^k\,e^x\,x^{-m_2}$ where the constants $c_l^k$, $c_u^k$ , $m_1$
, $m_2$ can in principle be calculated. The cross section $\sigma _{inc}^k$
being bound between two exponentials, can only behave exponentially, possibly
modified by an asymptotic Series of inverse powers of $s$ . Taking into
account that $\sigma _{inc}^k$ for $k>1$ is highly supressed by the small $%
D^{k+1}$ factor containing the instanton 't Hooft factor, we deduce that the
contribution of all ladder graphs for $k>1$ is negligible and cannot alter
the high energy behavior of the leading order result. In particular they do
not affect the possible validity of the ZMS picture, that is based on the
instanton-antiinstanton chain graphs only.
This result is not totally unexpected. It is known that to each shaded blob
of fig.2, which represents exchange of any number of bosons and $4n_g-2$
fermions between an instanton and an intiinstanton, corresponds an
exponentially growing factor, while, to each instanton or antiinstanton, an
exponentially small 't Hooft factor. Since the number of instanton or
antiinstanton vertices outnumbers the number of the multiparticle-exchange
blobs by a factor of two, we can expect that the contribution of such
ladders is supressed compared to the leading semiclassical result. Moreover,
the fact that not all the momentum flows through each rung makes at the end
the ladder graphs grow only exponentially with $s$. This is in contrast to
the case of a linear chain where all momentum flows between the instanton
and the antiinstanton, creating thus an exponential growth that can
counterbalance the suppression effect of the 't Hooft factors.
We thank Q. Shafi and C. Bachas for collaborating in early stages of this
work.\
|
\section{Introduction}
The Canada-France Redshift Survey (CFRS) is
aimed at the detection of faint galaxies among objects
with 17.5$\le I_{AB} \le$ 22.5, with no
discrimination against any other properties.
The final catalog comprises 943 objects, of which 736 are
extragalactic.
The most common methods used to
produce samples of quasars or active galactic nuclei (AGN)
exploit their unusual colours, strong emission lines, variability, radio
emission, X-ray emission, etc., and therefore are always
biassed in some sense. Hewett and Foltz (1994) stress
that if a completely unbiassed sample is not possible, an accurate
assessment of the probability of detection as a function of absolute
magnitude, redshift and spectral energy distribution must be
made, and this is frequently lacking.
Even though very effective and efficient survey techniques have been
developed, there is always a danger that there
may exist a subset of the population
that might have been overlooked. Indeed, Webster(1995)
has recently suggested that many more red quasars exist
than previously supposed. Although the results of
surveys based on techniques other than the popular
UVX method can be examined to see if any
such red population exists, surveys like CFRS
can be considered
as the ultimate check since spectroscopic
observations are made for every object regardless of colour, morphology
or spectroscopic properties. For example, the CFRS survey
could detect red, radio-quiet, X-ray quiet
quasars should they exist. On the other
hand, since quasars constitute only $\sim$ 1\% of the surface
density of objects on the sky at these magnitudes, CFRS
cannot be expected to be an efficient survey for quasars.
The fundamental limit to CFRS as a quasar survey is the signal-to-noise
at which a broad emission line object could be identified.
In this paper, the properties of the six objects which would
commonly be classified as quasars in major surveys are compared with
those from other samples. Although distinction is often made
between quasars and other AGNs, particularly Seyfert galaxies, such
a distinction is not usually made in the basic surveys: any point-like
object with strong, broad emission lines is usually classified as a quasar.
Subsequently, further subdivisions are made based on luminosities and
whether or not the object is extended under typical
ground-based seeing conditions. In this paper, we adopt the
traditional survey definition, and simply refer to
all six strong emission line objects as quasars, partly to avoid confusion
with other galaxies in our survey which exhibit ``AGN activity'' in
their nuclei (e.g., Tresse et al. 1994).
Our CFRS quasar sample, although very small, is relatively unique and hence
it is of interest to compare
it with samples derived by other methods. Colless et al. (1991) reported
a similar analysis based on detection of two quasars in a
faint blue-selected sample, $21 < b_{J} < 23.5$.
Throughout this paper
it is assumed that
$H_\circ=50$ km sec$^{-1}$ Mpc$^{-1}$ and $q_{\circ}=0.5$.
\section{Observations}
As part of the Canada-France Redshift Survey, spectra were
obtained of over 1000 objects with 17.5$\le I_{AB} \le$ 22.5 in five
different high galactic latitude fields selected to produce a fair sample
of extragalactic objects (Lilly et al 1995a).
The CFHT MOS spectrograph was used to cover
the spectral region 4200 -- 8600\AA\ with a resolution
of 7\AA\ per CCD pixel, and slits were used which yielded a spectral resolution of
FWHM $\sim$35\AA. $BVIK$ photometry was also obtained for most objects
in the spectroscopic sample (Lilly et al. 1995a).
Details of all the observations,
reduction techniques and sensitivity are presented by Le F\`evre et al. (1995a), Hammer et al. (1995),
Lilly et al. (1995b) and Crampton et al. (1995).
Eight one hour exposures were usually obtained of all targets so that
spurious features such as cosmic rays were easily rejected,
and features such as strong emission lines were readily apparent in all the individual spectra.
The final CFRS catalog includes 591 galaxies with z $<$ 1.3, 200 stars, 146
unidentified objects and 6 quasars. Colour and morphological information
indicates that most of the unidentified objects are likely to be
galaxies (Crampton et al. 1995) but 7 objects are
quite compact and have colours not very different from the quasars.
The minimum equivalent width of emission lines that
would have been detected in the spectra of these seven objects is $\sim$30\AA.
Emission lines in the observed frame of at most a few percent of quasars
at z $>$ 2 would have been missed, if the equivalent width distributions
of Francis et al. (1992) and Hartwick and Schade (1990) are assumed.
The Mg II equivalent width distribution indicates that
only $\sim$10\% of low redshift (z $<$ 1) quasars
might have been missed, so it is unlikely that any quasars
reside among the unidentified objects.
Thus, only 6 quasars were detected among
a total of 736 extragalactic objects.
The total
effective area surveyed was 112 arcmin$^2$.
Table 1 summarizes the properties of the quasars. The first column gives the CFRS number (the first two digits represent the
right ascension of the field). The second and third columns give 2000
coordinates, followed by the isophotal
I$_{AB}$ magnitude ($I_{AB} = I$ + 0.48), redshift, absolute
magnitude, $(V-I)_{AB}$ colour, spectral index and rest-frame
equivalent widths of strong emission lines. The errors of the
equivalent widths are estimated to be $\pm10$\AA\ for most of the lines,
and $\pm5\AA$ for the lines of CFRS14.0198.
Absolute magnitudes in the rest-frame B band
were computed assuming a power-law continuum
slope, $f_\nu=\nu^\alpha$, with $\alpha$ determined for each object
from its $(V-I)_{AB}$ colour corrected for emission line contamination.
As indicated above, not all of these ``quasars'' lie above
the canonical division
between quasars and Seyfert galaxies at $M_B = -23$ for ($H_\circ=50$
km sec$^{-1}$ Mpc$^{-1}$). It is interesting to note that
no very high redshift quasars were discovered. Since our targets were
selected by their $I$ magnitudes, there is certainly no selection bias
against high redshift quasars, and since Ly$\alpha$ would not disappear
from our spectroscopic bandpass until z $>$ 6, identification of high redshift
quasars would have been straightforward had they been present.
Spectra
of the six quasars are presented in Figure 1 with the strongest emission lines
marked.
Finding charts from our deep $I$ band images are given in Figure 2. At least
three of the quasars may be in groups or large structures of galaxies,
as noted below.
\subsection{Notes on individual quasars}
\noindent
{\bf CFRS00.0207 z=1.352} is surrounded by many faint galaxies of similar magnitude; the five CFRS galaxies indicated in Figure 2 have isophotal
$I_{AB} \sim23.4\pm0.3$mag. It is conceivable that this quasar is embedded in a
cluster similar to those recently proposed by Hutchings et al. (1994)
and Matthews et al. (1994) around other z $>$ 1 quasars.
\noindent
{\bf CFRS03.0106 z=2.070} is 2\farcs3 from a $\sim$1.3 mag fainter galaxy.
\noindent
{\bf CFRS03.0603 z=1.048} has a companion (615) only 16$^{\prime\prime}$ away which has
identical redshift (1.048). In the larger CFRS field in this direction,
there are three more galaxies with redshifts within $\Delta$z = 0.01
bringing the total to
five. We have spectra of only one other object shown in the
field; CFRS03.0602 is a star. The image of CFRS03.0603 appears to be slightly
extended compared to the nearby stars, and the
measured ``compactness parameter'', Q , (Le Fevre, et al. 1986)
also indicates that it is resolved. Typically, CFRS point source objects
have Q $\le$1.3 (Crampton et al. 1995), while Q = 1.55 for this quasar.
Since the absolute magnitude
of this quasar is M$= -23.0$, it is at the border of the canonical Seyfert
-- quasar classification. Nevertheless, since z $\sim$1, it must reside
in a bright host galaxy.
\noindent
{\bf CFRS14.0198 z=1.6034:} The bright object (163) is an M star.
\noindent
{\bf CFRS14.1303 z=0.9859} appears to be part of a very large structure
(spread over our entire field corresponding to 6.5 $h_{50}^{-1}$ Mpc
projected dimension for $q_\circ=0$,
and greater than 900 km/s in redshift space)
containing an estimated 30 bright galaxies at
z = 0.985 (Le F\`evre et al. 1994). The nearest of these is CFRS14.1262, only
17$^{\prime\prime}$ away. CFRS14.1275 is at z = 0.763 and
CFRS 14.1327 is at z = 0.932. Strong lines of Ne III and Ne V are
present in the spectrum of CFRS14.1303 (Fig. 1).
\noindent
{\bf CFRS14.1567 z=0.4787} is slightly extended, indicating that the host
galaxy is visible.
Since M$= -22.0$, this quasar is more properly classified as a Seyfert.
Only two of the nearby galaxies have redshifts measured; both
1525 and 1541 have z $\sim$0.74.
Even though there are indications that at least three of the quasars
may be located in groups of galaxies, a more rigorous analysis
(Le F\`evre et al 1995b) shows that only the group around 14.1303 is
statistically significant.
\section{Discussion}
\subsection{Surface density}
The effective area of the CFRS survey is 0.031 deg$^{2}$ so that
the observed 6 quasars translate to a surface density of
200$^{+120}_{-80}$
deg$^{-2}$;
the highest surface density observed to date.
The errors enclose a 1$\sigma$ confidence interval (Gehrels 1986).
The survey magnitude
limit of $I_{AB}\le$22.5 corresponds to $B=23$ for quasars with a power-law
spectral energy distribution ($f_\nu\propto \nu^{\alpha}$) with $\alpha=-0.5$.
For comparison, the previously-faintest quasar surveys
reached a limiting magnitude of $B\sim$22. The Koo and
Kron (1988) survey has a limiting magnitude of m$_{J}$=22.5 ( $B \sim$m$_{J}+0.1$), but spectroscopic followup has not yet
been completed (Majewski, private communication).
Zitelli et al. (1992, hereafter Z92)
have surveyed 0.35 deg$^{2}$ to m$_{J}=22$ (and a smaller area to m$_{J}$=20.85).
Their survey yields a quasar surface density at m$_{J}$=22 of
115$\pm$17 deg$^{-2}$ and, for z$<$ 2.2, a surface density
of 86$\pm17$ deg$^{-2}$. Boyle, Jones, and Shanks (1991) surveyed
an area of 0.85 deg$^2$ to a similar limiting magnitude. Although
their survey suffers from incompleteness for z $>$ 2.2 due
to the UVX selection procedure, the
surface density for quasars with z$<$2.2, $68\pm9$ deg$^{-2}$,
is in reasonable agreement with the Z92 result.
In total, 118 new quasars were discovered in these two surveys. The Z92
result is also consistent with the surface density estimated
in a review of all major surveys by Hartwick and Schade (1990).
They derive a surface density of 160 quasars deg$^{-2}$ at $B$ = 22.5 and
z $<$ 3.3
(or 129 at z $<$ 2.2) by applying a completeness
correction to the work of Koo and Kron (1988).
Figure 3 shows a comparison of all of these results for z $<$ 3.3.
The surface density of quasars, log N $<$ B deg$^{-2}$,
from Hartwick and Schade (1990) (circles) is combined with the
newer results from Boyle et al. (1991) (open triangle) and Z92
(solid triangle). The faintest point (solid square)
from the present survey lies on an extrapolation of the data from
previous surveys, although the error
bar is large.
An extrapolation of the Z92 quasar surface density using
their slope of 0.40 for the log $N$ versus $M$ relation gives
an expected number of 300 deg$^{-2}$
at $B = 23$ for all redshifts, or 200 deg$^{-2}$ at z $<2.2$.
As noted above, although all of our quasars are at z $<2.2$, we clearly have no bias against
objects at much larger redshifts. Our observed surface density of 200$^{+120}_{-80}$
deg$^{-2}$ is at most
a mild contradiction to the extrapolation of the Z92
surface density in the sense that the expected number of low-redshift
(z $<2.2$) objects is observed, but there is some deficit
of high-redshift objects.
The faintest X-ray counts where optical identifications
have been made are dominated by active galactic nuclei (Shanks et al 1991).
Analysis of the deepest ROSAT counts yield surface density estimates of
413 X-ray sources deg$^{-2}$ to a limiting flux
of $2.5 \times 10^{-15}$ erg cm$^{-2}$ s$^{-1}$
(Hasinger et al. 1993) and fluctuation analysis (Barcons et al. 1994)
yields an estimate of 900 to 1800 discrete sources deg$^{-2}$
brighter than $7 \times 10^{-16}$ erg cm$^{-2}$ s$^{-1}$. These
counts (in the 0.5-2 KeV) band which might be taken as
estimates of the quasar surface density are well above our
own estimate.
\subsubsection{Comparison with Boyle (1991) model}
The expected number of quasars in various redshift bins (e.g, $0 <$ z $<0.5$, $0.5<$ z $<1.0$, ...) can be computed from the model of Boyle et al. (1991).
We find good agreement between the expectations derived
from the evolving model and our observations. Over the range $0 < $z$ < 3$ ,
4.98 quasars are predicted by the model compared to 6 observed.
The redshift distribution of the observed quasars also agrees
well with the model prediction, as shown in Figure 4.
Agreement in luminosity is also good: 2.4 objects or $\sim$50\% should be
AGN with $M_B>-23$, compared to the real sample which has 3, also 50\%.
In summary, the redshift and luminosity distributions of the quasars
in our sample agree very well with the predictions of the model of the
evolving luminosity function of Boyle et al. (1991), demonstrating consistency
of the model with observations in an $M-$z regime which represents a
significant
extrapolation beyond the regime where it was derived.
\subsubsection{Relative numbers of quasars and Seyferts}
In the preceding, we have made no distinction between luminous
quasars ($M_B<-23$, $H_\circ=50$, $q_\circ=0.5$) and lower luminosity
AGNs or Seyferts. An interesting point made by Z92 is that
the counts of the Seyfert galaxies are rising more steeply (with a slope
of $\sim0.7$ in the log $N$ vs $m$ relation) than the more
luminous objects (whose slope is $\sim 0.4$). They note that the
ratio of Seyfert to quasar counts increases from 0.1 at m$_{J}<20$ to
$\sim 0.3$ at m$_{J}<21$ and is $\sim 0.5$ at m$_{J}<22$. In the faintest
half-magnitude bin the ratio reaches unity.
In our very faint sample, the ratio of Seyfert to quasar
counts also equals 1, confirming the trend noted by Z92.
\subsection{Continuum properties}
The continua of active galactic nuclei (AGN) are frequently characterized
as power-law $f_\nu \propto \nu^\alpha$ spectra, although this
is a simplification when the continuum is considered over a large range in wavelength (e.g., Neugebauer et al. 1979). For example, the continuum
shape in the optical-ultraviolet region is dominated by the
``blue-bump'' feature which may be the signature of
a hot accretion disk (Elvis et al. 1994). Since many of the
quasars previously studied have been selected because they have blue continua
(e.g., in ultraviolet-excess surveys), there may be
biases in the statistics related to continuum shape.
Hence, it is of interest to
examine the continuum slopes in our sample where no such bias
can exist.
The optical-UV region of quasar spectra are crowded with emission
lines, Balmer continuum emission, and a number of FeII emssion features
(e.g., Francis et al. 1991, hereafter F91). This makes the definition of the
continuum difficult, particularly when only a restricted wavelength range
(and one which varies with redshift) is available.
In general, our spectra were well-fitted by power laws, but estimates of $\alpha$
made in this way are subject to considerable error, due to errors in the
centering of any individual object among the $\sim$80 objects
in our multi-slit masks.
Consequently, the continuum spectral index was estimated from our $V-I$
colours (corrected for the contribution of emission lines), assuming a power law.
The slopes, listed in Table 1, range from $\alpha=0.2$ to
$\alpha=-2.0$ with a mean of $-1.0$.
The error is estimated to be $\sim \pm 0.2$ in $\alpha$
One of the largest and most complete surveys
with which to compare our results is the Large Bright Quasar Survey (LBQS).
F91 give the distribution
of power-law slopes measured in the wavelength range
1450 -- 5050\AA\ for 688 LBQS (m $\le18.85$) quasars. They find a mean
$\alpha = - 0.32$. Francis et al. (1992) found a higher value using
a principal component analysis method
on a smaller subsample of the LBQS, but since the rest-frame continuum window used was very blue, the earlier analysis is better matched to our data.
Interestingly, the results from our small sample span nearly the entire
range in continuum slope evident in the much larger sample
of F91. Half of our spectral indices are redder than
$-1$ whereas at most 10\% of those in F91
are as red as this. Furthermore, 5/6 of our
spectral indices are redder than the median of the
sample of F91. However, the 95\% confidence interval (Gehrels 1986)
for the ratio $R$ of red-to-total objects derived from
our observations is $0.36 < R < 0.996$,
so that this result is not highly significant.
There is no reason to believe that the survey
technique employed in the LBQS
would have missed the reddest (or any) quasars seen here. On the contrary,
Francis et al. (1992) claim that any quasar with a continuum
bluer than $\alpha\sim -2$ and/or emission lines would have
been detected. Thus, although there is some indication that our
sample is rather red, it does not constitute a population
that would have escaped detection in the LBQS or other surveys.
It should be remembered, however,
that there are substantial
differences in the luminosity and redshift distributions between our
sample and the LBQS.
\subsection{Emission-line strengths}
The rest-frame
equivalent widths of the three principal emission lines common
to most of our spectra are listed Table 1. The equivalent widths
are consistent with the means given
by F91 and those compiled by Hartwick
and Schade (1990), indicating once again that the quasars in our
sample are not unusual. Since there has always been a lingering concern that
the average emission line strengths of quasars have been biased by their
detection methods, this is reassuring.
\section{Summary}
The main goal of this paper was to compare the properties of
this very faint (albeit small) sample of quasars
with the properties of previously known samples. This
sample was selected without any of the biases inherent in
the most common survey techniques and therefore could
potentially yield new insights into the population.
The surface density of AGN at $B=23$ is $200^{+120}_{-80}$
deg$^{-2}$, the highest observed to date.
This is consistent with an extrapolation of
the faint counts due to Z92 and
Boyle et al. (1991), and also with the lower limit found
by Colless et al. (1991) based on a similar method to ours.
We find that the space density of these quasars
as a function of redshift and luminosity agrees
very well with the predictions of the
successful evolutionary model of
Boyle (1991), even though our sample is deeper than any of the
samples from which the model is derived.
The colours of these faint AGN are within the range defined
by previously known samples and would not have escaped
detection in major surveys. We find an apparent surplus of red
objects compared with other studies, but the numbers are small
and the significance is low. The strengths
of the emission lines are entirely consistent with those
found in other surveys.
In summary, we find no remarkable differences between any of the
properties of these faint objects and the distributions
of the same properties derived from previous
surveys. All our quasars would have been found by previous
search techniques.
One of the quasars is in a structure consisting of at
least 30 galaxies at z $\sim0.985$, and there is evidence
that two others may be associated with groups
of galaxies at a similar redshift.
|
\section{Introduction}
\label{intro}
Let $L$ be a perfect field of characteristic $p$.
Recall that a polynomial $f(x) \in L[x]$ is called {\em additive} if $f(x+y)=f(x)+f(y)$ identically.
It is easy to see that a polynomial is additive if and only if it is of the form
$$f(x)=a_0 x + a_1 x^p + a_2 x^{p^2} + \cdots + a_n x^{p^n}.$$
The set of additive polynomials forms a noncommutative ring in which $(f \circ g)(x)=f(g(x))$.
This ring is generated by the scalar multiplications $x \mapsto ax$ for $a \in L$, and by $\tau(x)=x^p$.
In this ring we can write the $f$ above as $a_0 + a_1 \tau + a_2 \tau^2 + \cdots + a_n \tau^n$, and we will denote the ring by $L\{\tau\}$, using braces instead of brackets to remind ourselves that it is a {\em twisted} polynomial ring in $\tau$.
The indeterminate $\tau$ does not commute with elements of $L$ (which are acting as scalar multiplications); instead if $a \in L$ then $\tau a = a^p \tau$ as additive polynomials.
Following Ore~\cite{Or1}, we define the {\em adjoint} of $f$ to be the expression
$$f^\ast(x) = a_0 x + a_1^{1/p} x^{1/p} + a_2^{1/p^2} x^{1/p^2} + \cdots + a_n^{1/p^n} x^{1/p^n}.$$
This defines a function on $L$ (and its algebraic closure) since $L$ is perfect.
Goss observed that the kernel of $f$ (the set of zeros of $f$ in a fixed algebraic closure of $L$) generates the same field extension as the kernel of $f^\ast$, by using a formula of Ore~\cite{Or1} which expresses the zeros of $f^\ast$ explicitly in terms of determinants involving the zeros of $f$.
One might ask whether there is a natural Galois-equivariant isomorphism between $\ker f$ and $\ker f^\ast$.
In general, the answer is no.
But, as was discovered independently by the author and N. Elkies~\cite{elkies}, there is something which is just as good:
\begin{unnumbered}
There exists a natural Galois-equivariant pairing
$$\ker f \times \ker f^\ast \rightarrow {{\Bbb F}_p}.$$
\end{unnumbered}
One of the main results of this paper is the generalization of this theorem to the situation where $f$ is an additive {\em power series}.
Let $C$ be an algebraically closed field of characteristic $p$ which is complete with respect to a (non-archimedean) absolute value $|\;|$.
If $f \in C[[z]]$ is an additive power series, its adjoint is a ``linear fractional power series'' of the form
$$g(z) = b_0 z + b_1 z^{1/p} + b_2 z^{1/p^2} + \cdots$$
It is easy to see that such a series converges for all $z$ when the $b_i$ tend to zero.
In this case, we will show that $g$ defines a continuous group homomorphism $C \rightarrow C$, which is surjective and open if nonzero.
One can see immediately that if $b_0 \not = 0$ then $g(z) = b_0 z + \text{(smaller terms)} \not= 0$ for large $z$, so $\ker g$ is a closed and bounded subgroup of $C$.
In fact, we prove that $\ker g$ is compact, and if infinitely many $b_i$ are nonzero, it is isomorphic to ${{\Bbb F}^\omega_p}$ as a topological group.
Conversely, if $G$ is any compact subgroup of $C$, then $G$ is the kernel of such a $g$ if and only if the integral $\int_G v$ of the real-valued function $v(z)=-\log|z|$ with respect to Haar measure on $G$ equals $+\infty$.
Also, $G$ essentially determines $g$.
We next develop the theory of Newton polygons for such series.
Finally we prove the analogue of our pairing theorem above, namely that for any everywhere convergent additive power series $f \in C[[z]]$, there is a natural pairing
$$\ker f \times \ker f^\ast \rightarrow {{\Bbb F}_p}$$
which exhibits the compact group $\ker f^\ast$ as the Pontryagin dual of the discrete group $\ker f$.
Actually we will generalize further by considering ${{\Bbb F}_q}$-linear power series for any power $q$ of $p$, and more substantially by considering {\em bi-infinite series}
$$f(z) = \cdots + a_{-2} z^{1/q^2} + a_{-1} z^{1/q} + a_0 z + a_1 z^q + a_2 z^{q^2} + \cdots.$$
The kernels of these series are locally compact.
In fact, many of our results extend to non-linear series for which the set of allowable exponents is
$$E=\{mp^n : m,n \in {\Bbb Z}, m \ge 0 \}.$$
We will develop results in this general context, specializing when appropriate.
Finally we give some applications of our results to the theory of Drinfeld modules.
We construct a pairing for Drinfeld modules which behaves in many ways like the Weil pairing on abelian varieties.
In addition, our results let us describe the topological module structure of the kernel of the adjoint exponential function of a Drinfeld module.
We conclude with a few unanswered questions.
\section{The rings ${\cal P}$ and ${\cal F}$ of fractional power series}
\label{tworings}
Let $C$ be an algebraically closed field containing ${{\Bbb F}_q}$ which is complete with respect to a (non-archimedean) absolute value $|\;|$.
Let $V = \{ x \in C : |x| \le 1 \}$ be the valuation ring of $C$, let ${\frak m} = \{ x \in C : |x|<1 \}$ be its maximal ideal, and let $k=V/{\frak m}$ be the residue field.
We will consider ``fractional power series'' in which the exponents belong to
$$E=\{mp^n : m,n \in {\Bbb Z}, m \ge 0 \}.$$
Since we have unique $p$-th roots in $C$, for any $z \in C$ and $e = mp^n \in E$ we can interpret $z^e$ as $(z^{1/p^n})^m$.
Define ${\cal P}$ to be the set of series $\sum_{e \in E} a_e z^e$ with coefficients $a_e$ in $C$, which converge at all $z \in C$.
Since $E$ is countable, it is clear what convergence means: for each $z \in C$ and $r>0$, there must be only finitely many terms in the series of absolute value greater than $r$.
And since we are in the non-archimedean situation, convergence at some $z$ of absolute value $r$ implies uniform convergence on all of $\Delta_r$.
Furthermore, we define the set ${\cal F}$ of ``${{\Bbb F}_q}$-linear fractional power series'' as the set of bi-infinite series
$$f(z) = \cdots + a_{-2} z^{1/q^2} + a_{-1} z^{1/q} + a_0 z + a_1 z^q + a_2 z^{q^2} + \cdots$$
which converge for all $z \in C$.
Hence ${\cal F}$ is a subset of ${\cal P}$.
For $f \in {\cal F}$ we will also write
$$f = \cdots + a_{-2} \tau^{-2} + a_{-1} \tau^{-1} + a_0 + a_1 \tau + a_2 \tau^2 + \cdots,$$
thinking of $\tau$ as the operator $\tau(z)=z^q$ on $C$, and thinking of $a \in C$ acting as the scalar multiplication by $a$.
Note that for $a \in C$, $\tau a = a^q \tau$ and $\tau^{-1} a = a^{1/q} \tau^{-1}$, as maps on $C$.
We also define ${\cal F}^+$ as the set of $f \in {\cal F}$ such that $a_n=0$ for $n<0$ (these are just the convergent ${{\Bbb F}_q}$-linear power series), and ${\cal F}^-$ as the set of $f \in {\cal F}$ such that $a_n=0$ for $n>0$.
There is a simple criterion for the convergence of linear series:
\begin{prop}
\label{convergence}
Suppose $a_i \in C$ for $i \in {\Bbb Z}$.
The bi-infinite series
$$f(z) = \cdots + a_{-2} z^{1/q^2} + a_{-1} z^{1/q} + a_0 z + a_1 z^q + a_2 z^{q^2} + \cdots$$
converges for all $z \in C$ if and only if the following two conditions hold:
\begin{enumerate}
\item $\lim_{n \rightarrow -\infty} |a_n| = 0$.
\item $\lim_{n \rightarrow +\infty} |a_n|^{1/q^n} = 0$.
\end{enumerate}
\end{prop}
\begin{pf}
By the well-known formula for the radius of convergence of a power series, the right end of the series converges everywhere when $\lim_{n \rightarrow +\infty} |a_n|^{1/q^n} = 0$.
The left end converges at $z$ when $\lim_{n \rightarrow -\infty} |a_n z^{q^n}| = 0$.
But for $z$ nonzero, this is equivalent to $\lim_{n \rightarrow -\infty} |a_n| = 0$, since $\lim_{n \rightarrow -\infty} |z^{q^n}| = 1$.
\end{pf}
For $f,g \in {\cal P}$, we define $f+g$ and $f \cdot g$ in the obvious way, by expanding and grouping terms with the same exponent.
\begin{prop}
$({\cal P},+,\cdot)$ is a commutative ring.
\end{prop}
\begin{pf}
The only nontrivial part is to check that the sum and product are actually everywhere convergent.
In fact, this is easy as well, since the sum or product of two convergent series is a doubly infinite convergent series, even before grouping terms.
\end{pf}
If $f,g \in {\cal F}$, we can still use the definitions of addition and multiplication above, but $f \cdot g$ need not be an element of ${\cal F}$, so ${\cal F}$ is not a subring of ${\cal P}$.
Nevertheless, we will make ${\cal F}$ into a ring by using composition in place of multiplication.
In fact, we will define the composition $f \circ g$ of two general elements $f(z)=\sum_{d \in E} a_d z^d$ and $g(z)=\sum_{e \in E} b_e z^e$ of ${\cal P}$.
First, for integral $m \ge 0$, let $g^m$ denote $g$ multiplied by itself $m$ times.
For $n \in {\Bbb Z}$, define $g^{p^n} = \sum_{e \in E} b_e^{p^n} z^{ep^n}$, which is consistent with the previous sentence.
For $d=mp^n \in E$, define $g^d = \left( g^m \right)^{p^n}.$
Finally define
$$f \circ g = \sum_{d \in E} a_d g^d,$$
in which we expand and group terms of the same exponent.
If $f=\sum_{i \in {\Bbb Z}} a_i \tau^i$ and $g=\sum_{j \in {\Bbb Z}} b_j \tau^j$ are elements of ${\cal F}$, then the above definition simplifies to $f \circ g =\sum_{n \in {\Bbb Z}} c_n \tau^n$ where
$$c_n = \sum_{i+j=n} a_i b_j^{q^i}.$$
\begin{prop}
\label{composition}
For $f,g \in {\cal P}$, $f \circ g$ converges to an element of ${\cal P}$, and represents the composition of the maps $f$ and $g$.
\end{prop}
\begin{pf}
Fix $r>0$.
Since the series $g$ converges uniformly on $\Delta_r$, we may let $R = \sup |g(\Delta_r)|$.
Fix $z$ of absolute value $r$.
Then from the definition of $g^d$ it is clear that all terms in the expansion of $g^d$ (for fixed $d \in E$) are bounded in absolute value by $R^d$, with only finitely terms of absolute value greater than $s$, for each $s>0$.
Fix $t>0$.
There are only finitely many $d$ such that $|a_d|R^d > t$ (since $f$ is convergent), so there are only finitely many $d$ such that $a_d g^d$ has terms of absolute value greater than $t$, and the previous sentence implies that even for such $d$, there are at most finitely many such terms.
Hence even before grouping terms, the series for $f \circ g$ converges at $z$, and this implies that the grouping gives a convergent series for each coefficient of $f \circ g$ as well.
\end{pf}
\begin{prop}
\label{ring}
$({\cal F},+,\circ)$ is a noncommutative ring containing $C$.
The center of ${\cal F}$ is ${{\Bbb F}_q}$.
\end{prop}
\begin{pf}
By the previous proposition and the explicit form of the definition for linear series, if $f,g \in {\cal F}$, then $f \circ g \in {\cal F}$.
Hence the first part is clear.
The center of ${\cal F}$ clearly contains ${{\Bbb F}_q}$.
On the other hand, if $c \in C$ is transcendental, direct computation shows that its centralizer in ${\cal F}$ is $C$, and the only elements of $C$ which commute with $\tau$ are those in ${{\Bbb F}_q}$.
\end{pf}
\begin{prop}
\label{uniformcontinuity}
Any $f \in {\cal P}$ defines a continuous function $f: C \rightarrow C$.
The continuity is uniform on each bounded subset of $C$.
\end{prop}
\begin{pf}
It suffices to show $f$ is uniformly continuous on $\Delta_R$ for each $R>0$.
Given $\epsilon >0$, there are only finitely many terms $a_e z^e$ in $f$ for which $|a_e| R^e >\epsilon$, since $f$ is everywhere convergent.
The sum of these is clearly uniformly convergent on $\Delta_R$, and the sum of the rest of the terms is bounded in absolute value by $\epsilon$ for all $z \in \Delta_R$, so $f$ is uniformly continuous on $\Delta_R$ as well.
\end{pf}
\begin{prop}
\label{nonvanishing}
If $f \in {\cal P}$ is nonzero as a series, then the function it defines on $C$ is not identically zero.
\end{prop}
\begin{pf}
Since $f(1)$ converges, there are only finitely many coefficients of $f$ larger than a given size, and by scaling by an element of $C$, we may assume the largest coefficients have absolute value 1.
Then the reduction of $f$ modulo ${\frak m}$ is a fractional {\em polynomial} over the residue field $k$, i.e., the $p^n$-th root of a polynomial.
Since $k$ is infinite, this reduction is nonzero at some $\bar{\alpha} \in k$.
If $\alpha \in C$ is any lift of $\bar{\alpha}$, then $|f(\alpha)|=1$, so in particular $f(\alpha) \not= 0$.
\end{pf}
\begin{prop}
\label{additive}
If $f \in {\cal P}$ considered as a map on $C$ is ${{\Bbb F}_q}$-linear, then $f \in {\cal F}$.
\end{prop}
\begin{pf}
Write $f= \sum_{e \in E} a_e z^e$.
For each $c \in C$, the function $f(cz+z)-f(cz)-f(z)$ in ${\cal P}$ is identically zero, so by Proposition~\ref{nonvanishing} it is zero as a series.
Hence for each $e=mp^n$, the coefficient of $z^e$ in it is zero.
This coefficient is $a_e \left[ (c+1)^e - c^e - 1 \right]$.
If $e$ is not a power of $p$ and $a_e \not=0$, then the $p^n$-th power of this coefficient will be a nonzero polynomial in $c$, and hence will not vanish for some value of $c$.
This contradiction shows that $a_e=0$ for $e$ not a power of $p$.
A similar argument with $f(cz)-cf(z)$ for $c \in {{\Bbb F}_q}$ shows that $a_e=0$ for $e$ not a power of $q$; i.e., $f \in {\cal F}$.
\end{pf}
\section{The topology and norm on ${\cal P}$ and ${\cal F}$}
\label{topologynorm}
We give ${\cal P}$ the ``bounded-open topology.''
For real $r>0$, define
\begin{eqnarray*}
\Delta_r & = & \{\, x \in C : |x| \le r \,\}, \\
\Delta_r' & = & \{\, x \in C : |x| < r \,\}.
\end{eqnarray*}
If $r,s>0$ let
$${\cal P}(r,s)=\{\,f \in {\cal P} | f(\Delta_r) \subset \Delta_s \,\}.$$
Take these subsets of ${\cal P}$ as a subbasis for the neighborhoods of zero.
For each $\alpha,\beta \in C$ and $c \in {{\Bbb F}_q}$, the set of $f \in {\cal P}$ such that $f(\alpha+\beta)=f(\alpha)+f(\beta)$ and the set of $f \in {\cal P}$ such that $f(c \alpha) = c f(\alpha)$ are closed subsets of ${\cal P}$.
Hence by Proposition~\ref{additive}, ${\cal F}$ is a closed subset of ${\cal P}$.
We give ${\cal F}$ the subspace topology, which is also the bounded-open topology on ${\cal F}$.
For $f = \sum_{e \in E} a_e z^e \in {\cal P}$, define
$$\|f\| = \sup\; \{\, |a_e|:e \le 1 \,\} \cup \{\, |a_e|^{1/e}:e \ge 1 \,\}.$$
The convergence of $f$ implies that this is a finite real number, and that the supremum is attained.
If $f = \sum_{n \in {\Bbb Z}} a_n \tau^n \in {\cal F}$ this becomes
$$\|f\| = \sup\; \{\, |a_n|:n \le 0 \,\} \cup \{\, |a_n|^{q^{-n}}:n \ge 0 \,\}.$$
This norm satisfies an ultrametric triangle inequality: $\|f+g\| \le \max\{\|f\|,\|g\|\}$ for any $f,g \in {\cal P}$.
In showing that the norm is consistent with the bounded-open topology on ${\cal P}$ we will use the following.
\begin{lemma}
\label{termwise}
Let $f = \sum_{e \in E} a_e z^e \in {\cal P}$, and let $r,s>0$.
Then the following are equivalent:
\begin{enumerate}
\item $f \in {\cal P}(r,s)$.
\item $a_e z^e \in {\cal P}(r,s)$ for all $e \in E$.
\item $|a_e| r^e \le s$ for all $e \in E$.
\end{enumerate}
\end{lemma}
\begin{pf}
It is clear that~(2) and~(3) are equivalent, and that they imply~(1).
In showing~(1) implies~(3), we can assume $r$ is the absolute value of some element of $C$, because if not, it is an increasing limit of such $r$, and the implication follows if it is known for those $r$.
Then we can reduce to the case $r=1$ by composing $f$ with a scalar multiplication on the right.
Now what we must show is that if $f$ is a nonzero element of ${\cal P}(1,s)$, a largest coefficient $b$ of $f$ is at most $s$ in absolute value.
The mod ${\frak m}$ reduction of $b^{-1} f$ is the $p^n$-th root of a nonzero {\em polynomial}, so we can pick $x \in V$ which does not reduce to one of its roots, and find that $|b^{-1} f(x)|=1$.
Hence $|b| = |f(x)| \le s$, since $f \in {\cal P}(1,s)$.
\end{pf}
\begin{prop}
\label{agree}
The bounded-open topology on ${\cal P}$ is the same as that induced by the norm $\|\;\|$.
\end{prop}
\begin{pf}
It is clear from the definitions that if $0<s<1<r$,
$$\|f\| \le s/r \implies f \in {\cal P}(r,s).$$
On the other hand, by Lemma~\ref{termwise}, if $0<s<1$,
$$f \in {\cal P}(1,s) \cap {\cal P}(1/s,1) \implies \|f\| \le s.$$
\end{pf}
\begin{theorem}
\label{topologicalring}
$({\cal P},+,\cdot)$ is a complete topological ring, and the composition map ${\cal P} \times {\cal P} \rightarrow {\cal P}$ is continuous.
Also, $({\cal F},+,\circ)$ is a complete topological ring.
\end{theorem}
\begin{pf}
The triangle inequality for $\|\;\|$ implies that addition is continuous.
To check that multiplcation is continuous, we must show that if $f,g \in {\cal P}$, and $r,s>0$, then for $f_0,g_0$ sufficiently small,
$$(f+f_0)\cdot(g+g_0)-f\cdot g \in {\cal P}(r,s).$$
Since $f$ and $g$ are bounded on $\Delta_R$, this follows for $f_0,g_0 \in {\cal P}(r,\epsilon)$ for sufficiently small $\epsilon>0$.
Now let us show that ${\cal P}$ is complete.
It suffices to show that if $f_1,f_2,\ldots$ is a sequence in ${\cal P}$ tending to zero, then $\sum_{i=1}^\infty f_i$ converges to another element of ${\cal P}$.
For each $e \in E$, the coefficients of $z^e$ in $f_i$ tend to zero by Lemma~\ref{termwise}, so they sum to some $a_e \in C$.
Interchanging the order of summation shows that $\sum_{i=1}^\infty f_i$ converges to $f \stackrel{\text{def}}{=} \sum_{e \in E} a_e z^e$ uniformly on bounded subsets, and that $f \in {\cal P}$.
For the continuity of composition, we must show that for $f_0,g_0$ sufficiently small,
\begin{equation}
\label{compcont}
(f+f_0) \circ (g+g_0) - f \circ g \in {\cal P}(r,s).
\end{equation}
Let $R=\sup |g(\Delta_r)|$.
If $g_0 \in {\cal P}(r,R)$ and $f_0 \in {\cal P}(R,s)$, then
\begin{equation}
\label{firsthalf}
f_0 \circ (g+g_0) \in {\cal P}(r,s).
\end{equation}
By Proposition~\ref{uniformcontinuity}, $f$ is uniformly continuous on $\Delta_R$, so if $g_0 \in {\cal P}(r,\epsilon)$ for sufficiently small $\epsilon$,
\begin{equation}
\label{secondhalf}
f \circ (g+g_0) - f \circ g \in {\cal P}(r,s).
\end{equation}
Adding~(\ref{firsthalf}) and~(\ref{secondhalf}) yields~(\ref{compcont}).
By restriction from ${\cal P}$ to ${\cal F}$, the ring operations on ${\cal F}$ are continuous.
As remarked earlier, ${\cal F}$ is a closed subset of ${\cal P}$, so the completeness of ${\cal F}$ follows from that of ${\cal P}$.
\end{pf}
Although the ring operations for both ${\cal P}$ and ${\cal F}$ are continuous with respect the topology induced by $\|\;\|$, it is not true that $\|f \cdot g\| \le \|f\| \cdot \|g\|$ for all $f,g \in {\cal P}$; nor is it true that $\|f \circ g\| \le \|f\| \cdot \|g\|$ for all $f,g \in {\cal F}$.
In fact, these can fail even if $f \in C$.
The following proposition shows that this defect of $\|\;\|$ is unavoidable.
(This is different from the theory of Banach algebras, in which if one has a norm for which multiplication is continuous in each variable, one can define a new norm with the sub-multiplicative property. See Section~3.1 of~\cite{kadison}.)
\begin{prop}
There is no norm $\|\;\|'$ on ${\cal P}$ inducing the same topology as above and satisfying
\begin{enumerate}
\item $\|f\|' \ge 0$ with equality if and only if $f=0$.
\item $\|f \pm g\|' \le \|f\|' + \|g\|'$
\item $\|1\|' = 1$.
\item $\|f \cdot g\|' \le \|f\|' \cdot \|g\|'$.
\end{enumerate}
Similarly, there is no norm $\|\;\|'$ on ${\cal F}$ inducing the same topology and satisfying (1), (2), (3), and (4) with $\cdot$ replaced by $\circ$.
\end{prop}
\begin{pf}
Suppose there were such a norm on ${\cal P}$.
As $\epsilon \rightarrow 0$ in $C$, $\epsilon z \rightarrow 0$ in ${\cal P}$, so $\|\epsilon z\|'<1$ for some nonzero $\epsilon \in C$.
Then the series $1 + \epsilon z + (\epsilon z)^2 + \cdots$ should converge in ${\cal P}$ to an inverse of $1-\epsilon z$.
This is impossible since $1-\epsilon z$ is zero at $z=1/\epsilon$.
Similarly, since $1-\epsilon \tau$ has no inverse (with respect to composition), there cannot be a sub-multiplicative norm on ${\cal F}$.
\end{pf}
\section{A Class of Locally Compact Subspaces of $C$}
\label{subspaces}
For $z \in C^\ast$, let
$$v(z)=-\log |z| \in {\Bbb R}$$
be the (additive) valuation associated with $|\;|$.
By convention, $v(0)=+\infty$.
If $f \in {\cal F}^+$, then as is well-known, $\ker f$ is a discrete sub-${{\Bbb F}_q}$-vector space of $C$, which is ``concentrated near infinity'' in the sense that if it is infinite, it is countable and its elements tend to infinity.
If $f \in {\cal F}^-$, it will turn out that $G \stackrel{\text{def}}{=} \ker f$ is a {\em compact} sub-${{\Bbb F}_q}$-vector space of $C$, this time ``concentrated near zero'' in the sense that the integral $\int_G v$ of the valuation with respect to Haar measure on $G$ is $+\infty$.
The kernel of a general (i.e., possibly infinite in both directions) element of ${\cal F}$ will belong to a certain hybrid of these two classes of vector spaces, which we now define.
\begin{defn}
Fix a real number $r>0$.
Let ${\cal G}$ be the collection of sub-${{\Bbb F}_q}$-vector spaces $G$ of $C$ satisfying the following conditions:
\begin{enumerate}
\item $G \cap \Delta_r$ is compact.
\item $\int_{G \cap \Delta_r} v = +\infty$. The integral is with respect to (any) Haar measure on $G \cap \Delta_r$.
\item $G \cap \Delta_r$ has finite or countable index in $G$, and in the latter case, if $g_1,g_2,\ldots$ is a system of coset representatives, then $\lim_{n \rightarrow \infty} |g_n| = +\infty$.
\end{enumerate}
\end{defn}
\begin{defn}
Let $G$ be a sub-${{\Bbb F}_q}$-vector space of $C$.
Let $\{\lambda_n\}_{n \in {\Bbb Z}}$ be a bi-infinite collection of elements of $C$, possibly terminating on either side.
We say $\{\lambda_n\}_{n \in {\Bbb Z}}$ is a {\em descending basis} for $G$ if
\begin{enumerate}
\item The $\lambda_n$ are independent over ${{\Bbb F}_q}$ and the closure of their span is $G$.
\item For each $m \in {\Bbb Z}$ (for which $\lambda_m$ exists), $\lambda_m$ is a smallest nonzero element of the ${{\Bbb F}_q}$-vector space spanned by $\{\, \lambda_n \mid n \le m \,\}$.
\item The sequence terminates on the left, or $\lim_{n \rightarrow -\infty} |\lambda_n| = \infty$.
\item The sequence terminates on the right, or $\sum_{n=1}^\infty v(\lambda_n) q^{-n} = +\infty$.
\end{enumerate}
\end{defn}
\begin{prop}
\label{class}
Let $G$ be a sub-${{\Bbb F}_q}$-vector space of $C$.
Then
\begin{enumerate}
\item The definition of ${\cal G}$ does not depend on the choice of $r>0$.
\item $G$ has a descending basis if and only if $G \in {\cal G}$.
\item If $G \in {\cal G}$, then $G$ is locally compact.
\item For $G$ discrete, $G \in {\cal G}$ if and only if it is finite or countable, with its elements tending to infinity in the latter case.
\item For $G$ compact, $G \in {\cal G}$ if and only if $\int_G v = +\infty$.
\item If $G \in {\cal G}$, then any closed sub-${{\Bbb F}_q}$-vector space $H$ of $G$ also belongs to ${\cal G}$.
\end{enumerate}
\end{prop}
\begin{pf} $\left.\right.$
\noindent {\em Proof of (2): }
Suppose $G$ has a descending basis $\{\lambda_n\}$.
Then the $\lambda_n$ are decreasing in size by condition~(2) in the definition, and $G \cap \Delta_r$ is the closure of the span of the $\{\lambda_n : n \ge n_0\}$ where $\lambda_{n_0}$ is the first $\lambda$ of absolute value at most $r$.
If the sequence of $\lambda_n$ after $\lambda_{n_0}$ terminates, then $G \cap \Delta_r$ is finite, hence compact.
Otherwise, these $\lambda_n$ form a sequence tending to zero by condition~(4), and we get a topological isomorphism ${{\Bbb F}^\omega_q} \rightarrow G \cap \Delta_r$ which sends $(a_0,a_1,a_2,\ldots)$ to $\sum_{i=0}^\infty a_i \lambda_{n_0+i}$.
(This, together with Theorem~\ref{locallycompact}, proves the remark in the introduction regarding the topological group structure of the kernel of an element of ${\cal F}^-$.)
Hence $G \cap \Delta_r$ is compact in any case.
Thus condition~(1) of the definition of ${\cal G}$ is verified for $G$.
Assume that the Haar measure on $G$ is normalized so that the open subspace $G \cap \Delta_r$ has measure 1.
If $G \cap \Delta_r$ is finite, then $\int_{G \cap \Delta_r} v = +\infty$ since $0$ has positive measure.
Otherwise, under the isomorphism above, the set of $(a_0,a_1,a_2,\ldots) \in {{\Bbb F}^\omega_q}$ such that the first nonzero $a$ is $a_i$ corresponds to a subset of $G \cap \Delta_r$ of measure $q^{-i}-q^{-i-1}$, and the function $v$ takes the value $v(\lambda_{n_0+i})$ everywhere on this subset, since otherwise some nontrivial combination of the $\lambda$'s of the same size as $\lambda_{n_0+i}$ would be closer to zero than $\lambda_{n_0+i}$, violating condition~(2) in the definition of decreasing basis.
These sets cover ${{\Bbb F}^\omega_q}$ (except for the point 0), so
$$\int_{G \cap \Delta_r} v = \sum_{i=0}^\infty v(\lambda_{n_0+i}) (q^{-i}-q^{-i-1}).$$
This is $+\infty$, since it agrees up to a finite number of terms with
$$(1-q^{-1})q^{n_0} \sum_{n=1}^\infty v(\lambda_n) q^{-n}.$$
(Substitute $n=n_0+i$.)
Hence condition~(2) of the definition of ${\cal G}$ holds.
Also, the finite linear combinations of the $\lambda_n$ for $n<n_0$ are coset representatives for $G \cap \Delta_r$ in $G$, so condition~(3) of the definition of ${\cal G}$ follows from properties (2) and~(3) of a decreasing basis.
Hence $G \in {\cal G}$.
\medskip
Conversely, suppose $G \in {\cal G}$.
Since $G \cap \Delta_r$ is compact, only zero can be a limit point of $\{|x|:x \in G \cap \Delta_r\}$.
Combined with condition~(3) of the definition of ${\cal G}$, this implies that the nonzero absolute values of elements of $G$ consist of a decreasing sequence $\{r_n\}_{n \in {\Bbb Z}}$, possibly terminating in either direction.
For each $r_n$, choose an ${{\Bbb F}_q}$-basis for $(G \cap \Delta_{r_n})/(G \cap \Delta_{r_n}')$, which is finite, by condition~(1) in the definition of ${\cal G}$ when $r_n \le r$, and by condition~(3) when $r_n>r$.
We claim that the concatenation of these bases (lifted to $C$) is a descending basis for $G$.
Properties (1) and~(2) of a descending basis are clearly satisfied.
Property~(3) follows from condition~(3) in the definition of ${\cal G}$, and property~(4) is equivalent to condition~(2) in the definition of ${\cal G}$, as in the proof of the converse above.
\medskip
\noindent {\em Proof of (1), (3), (4), and (5): }
Part~(1) follows trivially from~(2).
Part~(3) follows since if $G \in {\cal G}$, the open subgroup $G \cap \Delta_r$ is compact.
For part~(4), we may assume by~(1) that $r$ is smaller than the absolute value of the smallest nonzero element of $G$, and see what the conditions in the definition of ${\cal G}$ say in this case.
Similarly, part~(5) follows by assuming $r$ is greater than the absolute value of all elements of $G$.
\medskip
\noindent {\em Proof of (6): }
In constructing a descending basis for $G$ as in the proof of~(2), we may assume that the chosen basis for $(G \cap \Delta_{r_n})/(G \cap \Delta_{r_n}')$ contains a basis for its subspace $(H \cap \Delta_{r_n})/(H \cap \Delta_{r_n}')$.
It is then easy to check that the concatenation of the latter bases (lifted to $C$) is a descending basis for $H$, so that $H \in {\cal G}$ by~(2).
\end{pf}
\section{Maps and Fibers}
\label{kernels}
Here we state basic results on the maps on $C$ defined by elements of ${\cal P}$ or ${\cal F}$, and in particular study their fibers.
The proofs will be postponed until the next section.
First we have a fractional version of ``Picard's Theorem,'' which in the non-archimedean situation asserts that non-constant everywhere convergent power series are surjective.
The openness of the map is a kind of local surjectivity.
\begin{theorem}
\label{nonlinearsurjective}
If $f \in {\cal P}$ is not a constant series, then $f:C \rightarrow C$ is a surjective open map.
\end{theorem}
This theorem has two easy consequences.
\begin{cor}
\label{ffdomain}
If $f,g \in {\cal P}$ are not constant, then $f \circ g$ is not constant.
In particular, the noncommutative ring ${\cal F}$ has no zero divisors.
\end{cor}
\begin{cor}
\label{distinct}
Distinct series of ${\cal P}$ yield distinct maps on $C$.
\end{cor}
\begin{theorem}
\label{locallycompact}
If $f = \sum_{e \in E} a_e z^e \in {\cal P}$ is not constant, then for each $c \in C$, $f^{-1}(c)$ is a locally compact subset of $C$, and is compact if and only if there is a largest $e$ for which $a_e \not=0$.
If $f = \sum_{n \in {\Bbb Z}} a_n \tau^n \in {\cal F}$ is nonzero, then $G \stackrel{\text{def}}{=} \ker f$ is in ${\cal G}$.
Moreover,
\begin{center}
$G$ is discrete $\iff$ $a_n=0$ for $n \ll 0$. \\
$G$ is compact $\iff$ $a_n=0$ for $n \gg 0$.
\end{center}
\end{theorem}
\noindent (The last two statements are dual in the sense of Section~\ref{adjointpd}.)
We have a converse for $f \in {\cal F}$:
\begin{theorem}
\label{existence}
Each $G \in {\cal G}$ is the kernel of some $f \in {\cal F}$.
The $f$ is essentially unique: any $g \in {\cal F}$ with the same kernel is of the form $\epsilon \tau^n \circ f$ for some $\epsilon \in C^\ast$ and $n \in {\Bbb Z}$.
(In other words, by Corollary~\ref{units} below, $g$ differs from $f$ only by a unit of ${\cal F}$.)
\end{theorem}
Because $C$ is not locally compact, Theorem~\ref{locallycompact} implies it is not the union of $f^{-1}(0)$ and $g^{-1}(0)$ for nonzero $f,g \in {\cal P}$.
This proves the following.
\begin{cor}
\label{ppdomain}
The commutative ring ${\cal P}$ has no zero divisors.
\end{cor}
Recall that for a power series $f(z) = \sum_{i=0}^\infty a_i z^i \in C[[z]]$, the Newton polygon is the lower convex hull of the set of points $(i,v(a_i))$ in the plane, and that the Newton polygon gives information on the valuations of the zeros of $f$.
(See~\cite{amice},~\cite{artin},~\cite{kneser},~\cite{koblitz}, or~\cite{neukirch}.)
Similarly, we define the Newton polygon of $f=\sum_{e \in E} a_e z^e \in {\cal P}$ to be the lower convex hull of the set of points $(e,v(a_e))$ in the plane.
Let $v_p$ denote the $p$-adic valuation on $\Bbb Q$.
For $f = \sum_{e \in E} a_e z^e \in {\cal P}$, and $n \ge 0$, define
\begin{equation}
\label{fndefinition}
f_n = \sum_{v_p(e) \ge -n} a_e z^e.
\end{equation}
Therefore $f_n$ is the $p^n$-th root of a power series.
Since $f$ is convergent everywhere, we see that the $f_n$ converge to $f$ uniformly on bounded subsets (hence also in ${\cal P}$) as $n$ tends to infinity.
\begin{theorem}
\label{nonlinearnewton}
Suppose $f \in {\cal P}$ is not constant, and $c \in C$.
Then there is a canonical measure $\mu$ on $f^{-1}(c)$, characterized by
$$\mu \left( f^{-1}(c) \cap (\alpha+\Delta_r) \right) = \lim_{n \rightarrow \infty} \frac{\text{\# zeros of $(f_n-c)^{p^n}$ in $(\alpha+\Delta_r)$}}{p^n},$$
for all $\alpha \in C$, $r>0$.
(Here zeros are to be counted with multiplicity.)
The horizontal length of the segment of the Newton polygon of $f-c$ having slope $s$ (if any) equals
$$\mu(\{\,z \in f^{-1}(c) \mid v(z)=-s \,\}).$$
Also,
$$\mu(f^{-1}(c)) = \sup \{\,e \mid a_e \not=0 \,\}.$$
(Both might be $+\infty$.)
If $f \in C[[z]]$, then $\mu$ is the ``counting measure''; i.e., $\mu(\{x\})$ is the multiplicity of $x$ as a zero of $f-c$.
Finally, if $f \in {\cal F}$ is nonzero and $c=0$, then $\mu$ is a Haar measure on the locally compact group $\ker f$.
\end{theorem}
\begin{cor}
\label{trivialkernel}
If $f \in {\cal P}$ vanishes only at 0, then $f(z)=a z^e$ where $a \in C^\ast$ and $e \in E$.
\end{cor}
\begin{pf}
By Theorem~\ref{nonlinearnewton}, any series with at least two monomials has a ero of nonzero valuation.
\end{pf}
\begin{cor}
\label{units}
The unit group of ${\cal F}$ consists of the elements of the form $a \tau^n$ where $a \in C^\ast$ and $n \in {\Bbb Z}$.
\end{cor}
\begin{pf}
Any unit of ${\cal F}$ must have trivial kernel, and so must be of the form in Corollary~\ref{trivialkernel}, with the exponent a power of $q$.
On the other hand, it is easy to write down inverses of elements of this form.
\end{pf}
\begin{theorem}
\label{factors}
Let $f,g$ be nonzero elements of ${\cal F}$.
Then $\ker g \subseteq \ker f$ if and only if there exists $h \in {\cal F}$ such that $f=h \circ g$.
In this case, the $h$ is unique, and $\ker h = g(\ker f)$.
\end{theorem}
\section{Proofs}
\label{proofs}
Here we provide proofs of the theorems stated in the previous section.
\begin{lemma}
\label{mean}
Suppose $f(z)=\sum_{v_p(e) \ge -n} a_e z^e \in {\cal P}$, and $j>0$ is such that $a_j \not= 0$.
Then there exists $x \in C$ such that $f(x)=0$ and $|x| \le |f(0)/a_j|^{1/j}$.
\end{lemma}
\begin{pf}
This follows from looking at the Newton polygon for the (ordinary) power series $f(z)^{p^n}.$
\end{pf}
Let $f_n$ be as in~(\ref{fndefinition}).
\begin{lemma}
\label{cosets}
Suppose $f \in {\cal P}$ is not a constant series.
Fix $r,R \in {\Bbb R}$, with $0<r<R$, and fix $c \in C$.
Then the images of $f^{-1}(c) \cap \Delta_R$ and $f_n^{-1}(c) \cap \Delta_R$ in $\Delta_R/\Delta_r$ coincide for $n \gg 0$.
\end{lemma}
\begin{pf}
By considering $f-c$, we may assume $c=0$.
By composing with scalar multiplcations on both sides, we may assume $R=1$, and that the largest coefficient of $f$ other than the constant coefficient is of absolute value 1.
Let $a_e z^e$ be the term of $f$ with largest $e$ for which $|a_e|=1$.
Let $\epsilon_n=\sup|(f_{n+1}-f_n)(\Delta_R)|$.
Since $f_n \rightarrow f$, $\epsilon_n \rightarrow 0$.
Suppose $\lambda_n \in f_n^{-1}(0) \cap \Delta_R$.
Then
$$|f_{n+1}(\lambda_n)| = |(f_{n+1}-f_n)(\lambda_n)| \le \epsilon_n.$$
Provided that $n+1 \ge -v_p(e)$, the coefficient of $z^e$ in $f_{n+1}(z+\lambda_n)$ has absolute value 1, by choice of $e$ and since $|\lambda_n| \le 1$.
Hence by Lemma~\ref{mean}, there exists $z_n$ such that $|z_n| \le \epsilon_n^{1/e}$ and
$$f_{n+1}(z_n+\lambda_n)=0.$$
Let $\lambda_{n+1}=z_n+\lambda_n$.
Provided $n$ was chosen large enough, we can repeat the argument to construct a sequence $\lambda_n,\lambda_{n+1},\lambda_{n+2},\ldots$ with $\lambda_i \in f_i^{-1}(0)$ and
$$|\lambda_{i+1}-\lambda_i| \le \epsilon_i^{1/e} \le r.$$
Since $\epsilon_i \rightarrow 0$, the $\lambda_i$ converge to a limit $\lambda_\infty$ with $|\lambda_\infty-\lambda_n| \le r$.
Also since $f_n \rightarrow f$ uniformly on $\Delta_R$,
$$f(\lambda_\infty)=\lim_{i \rightarrow \infty} f_i(\lambda_i)=0.$$
For the other direction, suppose $\lambda \in f^{-1}(0) \cap \Delta_R$.
By Lemma~\ref{mean}, for $n \gg 0$, there exists $z_n$ such that $f_n(z_n+\lambda)=0$ and
$$|z_n| \le |f_n(\lambda)|^{1/e} \rightarrow 0$$
as $n \rightarrow \infty$.
In particular, for sufficiently large $n$ (and how large does not depend on $\lambda$), $|z_n| \le r$, so $z_n+\lambda \in f_n^{-1}(0)$ is in the same coset of $\Delta_r$ as $\lambda$.
\end{pf}
From the preceding we can deduce two useful corollaries.
First of all, we see that Lemma~\ref{mean} extends to arbitrary non-constant $f \in {\cal P}$.
\begin{cor}
\label{supermean}
Suppose $f(z)=\sum_{e \in E} a_e z^e \in {\cal P}$, and $j>0$ is such that $a_j \not= 0$.
Then there exists $x \in C$ such that $f(x)=0$ and $|x| \le |f(0)/a_j|^{1/j}$.
\end{cor}
Next we have a finiteness result, which will be used to prove that fibers are locally compact.
\begin{cor}
\label{preimage}
If $f \in {\cal P}$ is non-constant, $0<r<R$, and $c \in C$, then the image of $f^{-1}(c) \cap \Delta_R$ in $\Delta_R/\Delta_r$ is finite.
\end{cor}
\begin{pf}
The result with $f$ replaced by $f_n$ is proved by applying the theory of Newton polygons to the power series $f_n^{p^n}-c^{p^n}$.
Now apply Lemma~\ref{cosets}.
\end{pf}
\begin{pf*}{Proof of Theorem~\ref{nonlinearsurjective}}
Applying Corollary~\ref{supermean} to $f-c$ where $f \in {\cal P}$ is not a constant series and $c \in C$ shows that there is a solution to $f(z)-c=0$, so $f$ is surjective.
In checking that $f$ defines an open map, we may reduce to proving $f$ is open at 0, by composing $f$ with a translation.
Moreover we may assume $f(0)=0$.
If $c \in C$ is small, Corollary~\ref{supermean} applied to $f-c$ shows that $f(z)-c=0$ has a solution near 0.
This is exactly what is needed to prove that $f$ is open at 0.
\end{pf*}
\begin{pf*}{Proof of Theorems \ref{locallycompact} and~\ref{nonlinearnewton}}
First let us show that for any $r>0$, $f^{-1}(c) \cap \Delta_r$ is compact.
Suppose $\alpha_1,\alpha_2,\ldots$ is a sequence in $f^{-1}(c) \cap \Delta_r$.
By Corollary~\ref{preimage}, these elements lie in finitely many cosets of $\Delta_{r/2}$, so we can find an infinite subsequence within one coset.
Next we can find a subsequence of this subsequence lying within a single coset of $\Delta_{r/3}$, and so on, with the $n$-th subsequence lying within a coset of $\Delta_{r/n}$.
By diagonalization we obtain a convergent subsequence of the original sequence.
Since $f$ is continuous, $f^{-1}(c)$ is closed (and so is $\Delta_r$), so the limit lies in $f^{-1}(c) \cap \Delta_r$.
Thus $f^{-1}(c) \cap \Delta_r$ is compact.
Taking larger and larger $r$ shows that $f^{-1}(c)$ is locally compact.
\medskip
Before completing the proof of Theorem~\ref{locallycompact}, let us turn to Theorem~\ref{nonlinearnewton}.
First we check that the limit in the definition of $\mu$ exists.
Without loss of generality we may assume $c=0$.
Also we may assume $\alpha=0$, by considering $f(z+\alpha)$.
Let $\ell_r(f)$ denote the total horizontal length of the segments of the Newton polygon of $f$ whose slope is less than $\log r$, which is the largest $x$-coordinate of a point of contact of the Newton polygon with a supporting line of slope $\log r$.
The fact that $f$ is everywhere convergent implies that below any line in ${\Bbb R}^2$ there are at most finitely vertices, so for each $r$, $\ell_r(f_n)=\ell_r(f)$ for $n \gg 0$.
The Newton polygon for $f_n^{p^n}$ is the dilation of that of $f_n$ by the factor $p^n$, so $\ell_r(f_n^{p^n})=p^n \ell_r(f_n)$.
The theory of Newton polygons for ordinary power series implies that $\ell_r(f_n^{p^n})$ counts the number of zeros of $f_n^{p^n}$ (with multiplicity) in $\Delta_r$, so the limit in the definition of $\mu$ (for $c=\alpha=0$) converges to $\ell_r(f)$.
To check that $\mu$ is truly a measure, it will suffice to check that the definition is consistent in the sense that when $r<R$, its value on $f^{-1}(c) \cap (\beta+\Delta_R)$ equals the sum of its values on $f^{-1}(c) \cap (\alpha+\Delta_r)$ with $\alpha+\Delta_r$ ranging over the cosets of $\Delta_r$ contained in $\beta+\Delta_R$.
By Lemma~\ref{cosets} and Corollary~\ref{preimage}, we need concern ourselves with only finitely many of these cosets.
Now the result is clear from the finite additivity of the right hand side of the definition of $\mu$.
By what we have shown so far,
$$\mu(\{\, z \in f^{-1}(0) : |z| \le r \,\}) = \ell_r(f).$$
Taking this and subtracting the same with $r$ replaced by $r-\epsilon$ (with $s=\log r$, and $\epsilon$ sufficiently small) shows that
$$\mu(\{\,z \in f^{-1}(c) \mid v(z)=-s \,\})$$
equals the horizontal length of the Newton polygon segment of slope $s$.
We now check the final statements of Theorem~\ref{nonlinearnewton}.
By what we have shown so far, $\mu(f^{-1}(c))$ is the sum of the horizontal lengths of all the segments, and this is clearly $\sup \{\,e \mid a_e \not=0 \,\}$.
If $f \in C[[z]]$, the theory of Newton polygons for power series tells us that the lengths of the horizontal segments are counting zeros, so $\mu$ is the counting measure.
Finally, if $f \in {\cal F}$ is nonzero and $c=0$, it is clear from the definition that $\mu$ is translation-invariant on the locally compact group $\ker f$.
Moreover, $\mu(\ker f)>0$ by the formula we just derived for the measure of the whole space.
Thus $\mu$ is a Haar measure on $\ker f$.
This completes the proof of Theorem~\ref{nonlinearnewton}.
\medskip
We now resume the proof of Theorem~\ref{locallycompact}.
First, if there is a largest $e$ for which $a_e \not=0$, then the slopes of the Newton polygon of $f-c$ are bounded above for each $c \in C$, and hence by Theorem~\ref{nonlinearnewton}, $f^{-1}(c) \subset \Delta_r$ for some $r>0$.
Thus $f^{-1}(c)=f^{-1}(c) \cap \Delta_r$, which we showed already was compact.
On the other hand, if there is no largest $e$ for which $a_e \not=0$, then the Newton polygon of $f-c$ has infinitely many segments of increasing slope, so there is a sequence of elements in $f^{-1}(c)$ of decreasing valuation.
Such a sequence cannot have a convergent subsequence, so $f^{-1}(c)$ is not compact.
From now on, we assume $f=\sum_{n \in {\Bbb Z}} a_n \tau^n \in {\cal F}$, $c=0$, and $G=\ker f$.
By the first paragraph of this proof, $G \cap \Delta_r$ is compact for any $r>0$, proving that $G$ satisfies condition~(1) in the definition of ${\cal G}$.
Also by Corollary~\ref{preimage}, the image of $\ker f \cap \Delta_{nr}$ in $\Delta_{nr}/\Delta_r$ is finite for all $n$, so condition (3) in the definition of ${\cal G}$ follows.
Now we check condition~(2); i.e., that if $r>0$, then $\int_{G \cap \Delta_r} v =+\infty$.
If we integrate only over the subset
$$\{\, z \in G \mid v(z)=-s \,\}$$
we get $-s$ times the Haar measure of this set, which, by what we just proved, is the same as $-s$ times the horizontal length of the segment of the Newton polygon of slope $s$, which is the {\em vertical} displacement as one moves along the segment from right to left.
If we sum over all the segments of slopes less than $\log r$, we deduce that $\int_{G \cap \Delta_r} v$ equals the vertical displacement as one moves along the section of the Newton polygon to the left, starting at some point depending on $r$.
This vertical displacement is $+\infty$, since $v(a_n) \rightarrow +\infty$ as $n \rightarrow -\infty$.
Thus $G \in {\cal G}$.
Finally, we check the last equivalences in Theorem~\ref{locallycompact}.
First, $G$ is discrete if and only if there are no zeros of large finite positive valuation.
By Theorem~\ref{nonlinearnewton}, this happens if and only if $a_n=0$ for $n \ll 0$.
Lastly, the criterion for $G$ to be compact follows from our earlier criterion for fibers of general elements of ${\cal P}$.
\end{pf*}
For future reference, we record the following well-known result.
\begin{lemma}
\label{known}
If the zeros of a separable polynomial $f(z) \in C[z]$ form a sub-${{\Bbb F}_q}$-vector space of $C$, then $f \in C\{\tau\}$.
\end{lemma}
\begin{pf}
Proposition~1.3 in~\cite{DH1} proves this for $q=p$, so $f$ is an additive polynomial.
Comparing zeros and linear coefficients shows that $f(cz)=cf(z)$ for all $z \in {{\Bbb F}_q}$, and this forces $f \in C\{\tau\}$.
\end{pf}
\begin{lemma}
\label{image}
If $G \in {\cal G}$ and $f \in {\cal F}$ has kernel $G \cap \Delta_r$ for some $r>0$,
then $f(G)$ is in ${\cal G}$ and is discrete.
\end{lemma}
\begin{pf}
Since $f$ has compact kernel, its coefficient of $\tau^n$ is zero for $n \gg 0$ by Theorem~\ref{locallycompact}.
Hence in the series for $f(z)$, the last term dominates for large $z$.
Thus $|f(z)| \rightarrow \infty$ as $|z| \rightarrow \infty$.
Let $g_1,g_2,\ldots$ be representatives for the cosets of $G \cap \Delta_r$.
These form a sequence tending to infinity, by condition (3) in the definition of ${\cal G}$.
Then
$$f(G)=\{f(g_1),f(g_2),\ldots\}$$
also consists of a sequence tending to infinity, so it is in ${\cal G}$ and is discrete.
\end{pf}
\begin{pf*}{Proof of Theorem~\ref{existence}, existence} $\left.\right.$
\noindent{\em Case 1:} $G$ discrete.
\noindent Then by Lemma~\ref{known}, we may take
$$f(z) = z \prod_{g \in G, g \not=0} (1-z/g).$$
\noindent{\em Case 2:} $G$ compact.
\noindent We may assume $G$ is infinite, since otherwise we are in Case 1.
Without loss of generality suppose $G \subset \Delta_1$.
(If for some small $c \in C$ we can get $cG$ as a kernel of $f$, then we can get $G$ as the kernel of $f(cz)$.)
By Proposition~\ref{class}, there exists a descending basis $\lambda_1,\lambda_2,\ldots$ for $G$.
(It terminates on the left by condition~(3) of the definition of a descending basis.)
Let $V_n$ be the ${{\Bbb F}_q}$-vector space generated by $\lambda_1,\ldots,\lambda_n$.
Define
$$g_n(z) = \prod_{\lambda \in V_n} (z-\lambda),
\qquad \qquad h_n(z) = g_n(z)^{1/q^n}.$$
By Lemma~\ref{known}, $g_n \in C\{\tau\}$, so $h_n \in {\cal F}^-$, and both have coefficients bounded in absolute value by 1, since $G \subset \Delta_1$.
We will eventually show that the $h_n$ converge to the desired $f \in {\cal F}^-$.
We have
$$g_{n+1}(z) = g_n(z)^q - g_n(\lambda_{n+1})^{q-1} g_n(z)$$
since both sides are monic ${{\Bbb F}_q}$-linear polynomials of degree $q^{n+1}$ which vanish on $V_{n+1}$.
Raising to the $1/q^{n+1}$ power yields
\begin{equation}
\label{recurrence}
h_{n+1}(z) = h_n(z) - c_n h_n(z)^{1/q}
\end{equation}
where $c_n=h_n(\lambda_{n+1})^{1-1/q}$.
We claim that $|c_n| \rightarrow 0$ as $n \rightarrow \infty$.
We have
\begin{eqnarray*}
v(h_n(\lambda_{n+1})) & = & 1/q^n \sum_{\lambda \in V_n} v(\lambda_{n+1}-\lambda) \\
& \ge & 1/q^n \sum_{\lambda \in V_n} \min\{v(\lambda_{n+1}),v(\lambda)\} \\
& \ge & 1/q^n \sum_{\lambda \in V_n, \lambda \not= 0} v(\lambda),
\end{eqnarray*}
since $v(\lambda_{n+1}) \ge v(\lambda)$ for all nonzero $\lambda \in V_n$, by the definition of descending basis.
As $n \rightarrow \infty$, this tends to $\int_G v = +\infty$.
Hence
$$|c_n| = \exp(-(1-1/q) v(h_n(\lambda_{n+1}))) \rightarrow 0.$$
By~(\ref{recurrence}), we see that
$$\|h_{n+1}(z) - h_n(z)\| = \|c_n h_n(z)^{1/q}\| \le |c_n| \rightarrow 0$$
as $n \rightarrow \infty$.
Hence the $h_n$ converge to some $f \in {\cal F}$.
If $z \in V_n$, then $h_i(z)=0$ for $i \ge n$, so $f(z)=0$.
Thus $\ker f$ contains $\bigcup_{n=1}^\infty V_n$ and its closure, which is $G$.
On the other hand, if $z \not\in G$, and we set
$$\delta = \inf_{\lambda \in G} |z-\lambda|,$$
then directly from the definitions of $g_n(z)$ and $h_n(z)$ we get
$$|g_n(z)| \ge \delta^{q^n}, \qquad \qquad |h_n(z)| \ge \delta,$$
for all $n$, so in particular, $f(z) = \lim_{n \rightarrow \infty} h_n(z) \not= 0$.
Hence $\ker f = G$, as desired.
\medskip
\noindent{\em Case 3:} $G \in {\cal G}$ arbitrary.
\noindent By Case 2, we can find $g \in {\cal F}$ with kernel $G \cap \Delta_1$.
By Lemma~\ref{image}, $g(G)$ is in ${\cal G}$ and is discrete, so by Case 1, there exists $h \in {\cal F}$ with kernel $g(G)$.
Let $f=h \circ g$.
Then
$$f(x)=0 \iff h(g(x))=0 \iff g(x) \in g(G) \iff x \in G,$$
since $\ker g \subseteq G$.
\end{pf*}
Before proving Theorem~\ref{factors} and the uniqueness part of Theorem~\ref{existence}, let us prove the following ``remainder'' lemma.
\begin{lemma}
\label{remainder}
If $f \in {\cal F}$, then there exists $q \in {\cal F}$ such that
$$f = q \circ (1 - \tau) + f(1).$$
\end{lemma}
\begin{pf}
Suppose $f = \sum_{n \in {\Bbb Z}} a_n \tau^n \in {\cal F}$.
Without loss of generality, assume $f(1)=\sum_{n \in {\Bbb Z}} a_n =0$.
Then the rate of growth of the $a_n$ imposed by Proposition~\ref{convergence} implies that if
$$b_n \stackrel{\text{def}}{=} \sum_{m=-\infty}^n a_m = -\sum_{m=n+1}^\infty a_m,$$
then $q \stackrel{\text{def}}{=} \sum_{n \in {\Bbb Z}} b_n \tau^n$ belongs to ${\cal F}$.
(Use the first definition of $b_n$ to get convergence on the left, and the second for convergence on the right.)
Now
$$q \circ (1-\tau) = \sum_{n \in {\Bbb Z}} (b_n-b_{n-1}) \tau^n = \sum_{n \in {\Bbb Z}} a_n \tau^n = f.$$
\end{pf}
\begin{pf*}{Proofs of Theorem~\ref{factors} and the uniqueness in Theorem~\ref{existence}}
In Theorem~\ref{factors}, it is clear that if $f=h \circ g$, then $\ker g \subseteq \ker f$, so we will concern ourselves with converse, that if $\ker g \subseteq \ker f$, then there exists $h \in {\cal F}$ such that $f = h \circ g$.
Then the uniqueness of $h$ is clear from Corollary~\ref{ffdomain}, and $\ker h$ must be $g(\ker f)$ by the surjectivity of $g$ from Theorem~\ref{nonlinearsurjective}.
\noindent{\em Step 1: } Prove Theorem~\ref{factors} for $g= \alpha_0 \circ (1-\tau) \circ \alpha_1 \circ (1-\tau) \circ \alpha_2 \cdots (1-\tau) \circ \alpha_d$, where $\alpha_0,\alpha_1,\ldots,\alpha_d \in C^\ast$.
\medskip
\noindent We use induction on $d$.
The base case $d=0$ is trivial.
Suppose $d \ge 1$.
Let $\tilde{g} = \alpha_1 \circ (1-\tau) \circ \alpha_2 \cdots (1-\tau) \circ \alpha_d$.
Then the inductive hypothesis implies that $f = j \circ \tilde{g}$ for some $j \in {\cal F}$.
Since $f$ kills $\ker g$, $j$ kills $\tilde{g}(\ker g) = \ker( \alpha_1 \circ (1-\tau)) = {{\Bbb F}_q}$.
In particular $j$ kills 1, so by Lemma~\ref{remainder}, $j=q\circ (1-\tau)$ for some $q$.
If we now let $h=q\circ \alpha_0^{-1}$, then $f=h \circ g$.
\medskip
\noindent{\em Step 2: } Every separable $g \in C\{\tau\}$ of $\tau$-degree $d$ is of the form in Step 1.
\medskip
\noindent Again we use induction on $d$.
The base case $d=0$ is trivial.
If $d \ge 1$, there exists $c \in \ker g$, $c \not=0$.
Then $g$ kills $\ker ((1-\tau) \circ c^{-1})$, so $g = h \circ (1-\tau) \circ c^{-1}$ for some $h \in {\cal F}$ by Step 1, and the proof of the existence of $h$ in fact shows that $h$ must be a separable element of $C\{\tau\}$.
Applying the inductive hypothesis to $h$ produces the desired factorization.
\medskip
\noindent{\em Step 3: } Theorem~\ref{factors} holds when $g \in C\{\tau,\tau^{-1}\}$.
\medskip
\noindent Let $\tau^n$ be the lowest (most negative) power occuring in $g$.
Then by Steps 1 and 2, we can write $f=h' \circ (\tau^{-n} \circ g)$, so we can take $h=h' \circ \tau^{-n}$.
\medskip
\noindent{\em Step 4: } If $g_n,h_n \in {\cal F}$ are nonzero, $f = h_n \circ g_n$ for each $n$, and the $g_n$ converge to some nonzero $g \in {\cal F}$, then the $h_n$ converge to some $h \in {\cal F}$ and $f = h \circ g$.
\medskip
\noindent Fix $x \in C$.
We have $h_n(x)=f(g_n^{-1}(x))$ where $g_n^{-1}(x)$ denotes any $y_n \in C$ such that $g_n(y_n)=x$.
(By Theorem~\ref{nonlinearsurjective}, such $y_n$ exist.)
We will construct a sequence of such $y_n$ which {\em converges}.
Fix a nonzero term $a_j \tau^j$ in $g$.
Choose some large $R>0$, then choose some large $n_0>0$.
(We'll specify how large as we go along; how large we need $n_0$ to be depends on how large $R$ was taken to be.)
If $n \gg n_0$, the coefficient of $\tau^j$ in $g_n$ has the same absolute value as $a_j$, provided $n_0$ was chosen large enough.
Then by Corollary~\ref{supermean}, there exists a solution $y_{n_0} \in \Delta_R$ to $g_{n_0}(y_{n_0})=x$, if $R$ was chosen large enough.
Now inductively define $\epsilon_n$ and $y_{n+1} \in \Delta_R$ for $n \ge n_0$ as follows.
Pick $\epsilon_n$ such that $g_{n+1}(\epsilon_n)=(g_n-g_{n+1})(y_n)$, and set $y_{n+1}=y_n+\epsilon_n$.
Since the $g_n$ converge uniformly on $\Delta_R$, and since $y_n \in \Delta_R$ by the inductive hypothesis, $(g_n-g_{n+1})(y_n)$ can be assumed to be arbitrary small, if $n_0$ was chosen large enough.
Then by Corollary~\ref{supermean}, $\epsilon_n \in \Delta_R$, $y_{n+1} \in \Delta_R$, and
\begin{eqnarray*}
g_{n+1}(y_{n+1}) & = & g_{n+1}(y_n) + g_{n+1}(\epsilon_n) \\
& = & g_{n+1}(y_n) + g_n(y_n) - g_{n+1}(y_n) \\
& = & x.
\end{eqnarray*}
Moreover Corollary~\ref{supermean} guarantees that the $\epsilon_n$ can be chosen going to zero, so the $y_n$ converge as desired.
Since $f$ is continuous, we see that the sequence $h_n(x)$ converges.
In fact, as is clear from the construction of the $y_n$, the rate of convergence depends not on $x$, but only on an upper bound for $|x|$, so the $h_n$ converge uniformly on every bounded subset of $C$.
By Theorem~\ref{topologicalring}, the $h_n$ converge to some $h \in {\cal F}$.
Taking the limit of $f=h_n \circ g_n$ as $n \rightarrow \infty$ yields $f=h \circ g$.
\medskip
\noindent{\em Step 5: } Uniqueness holds in Theorem~\ref{existence} when $G \in {\cal G}$ is compact or discrete.
Suppose $G \in {\cal G}$ is compact.
By the proof of existence part of Theorem~\ref{existence}, there is a sequence of elements $j_n \in C\{\tau,\tau^{-1}\}$ with $\ker j_n \subseteq G$, which converges to an element $j \in {\cal F}$ with kernel $G$.
Let $f$ be any other element of ${\cal F}$ with kernel $G$.
By Step 3, we may write $f=h_n \circ j_n$, and then by Step 4, we may write $f=h \circ j$.
But $j$ is surjective by Theorem~\ref{nonlinearsurjective}, so $\ker h=j(\ker f)=0$ and hence $h$ is a unit by Corollaries \ref{trivialkernel} and~\ref{units}.
If $G \in {\cal G}$ is discrete, we can find a sequence $j_n \in C\{\tau\}$ with $\ker j_n \subseteq G$, which converges to an element $j \in {\cal F}$ with kernel $G$, so the same proof works.
\medskip
\noindent{\em Step 6: } Theorem~\ref{factors} holds when $\ker g$ is compact or discrete.
Indeed, by Step 5, $g$ must be a limit of elements $g_n \stackrel{\text{def}}{=} h \circ j_n \in C\{\tau,\tau^{-1}\}$ with $\ker g_n = \ker j_n \subseteq \ker g$, and we can then apply Step 4.
\medskip
\noindent{\em Step 7: } Theorem~\ref{factors} holds.
By the existence part of Theorem~\ref{existence}, we can find $g_1 \in {\cal F}$ with compact kernel $G \cap \Delta_1$.
By Step 6, we can write $g=j \circ g_1$, and $\ker j = g_1(\ker g)$, which is discrete by Lemma~\ref{image}.
By Step 6 again, we can write $f = h_1 \circ g_1$, and
$$\ker j = g_1(\ker g) \subseteq g_1(\ker f) = \ker h_1.$$
By Step 6 yet a third time, we can write $h_1 = h \circ j$.
Then
$$f = h_1 \circ g_1 = h \circ j \circ g_1 = h \circ g.$$
\medskip
\noindent{\em Step 8: } Uniqueness holds in Theorem~\ref{existence}.
Indeed, suppose $f$ and $g$ are nonzero elements of ${\cal F}$ with the same kernel.
By Step 7, each of $f$ and $g$ is a left multiple of the other, so they differ by a unit.
\end{pf*}
\section{Pontryagin duality for $A$-modules}
\label{pda}
Here we recall and develop some results on locally compact topological modules to be used later in the paper.
Throughout, topological groups, rings, and modules are assumed Hausdorff, and each $\operatorname{Hom}$ consists of continuous homomorphisms and is given the compact-open topology.
If $G$ is a locally compact abelian topological group (LCA group), its {\em Pontryagin dual} is $\hat{G} \stackrel{\text{def}}{=} \operatorname{Hom}_{\Bbb Z}(M,{\Bbb R}/{\Bbb Z})$, which is again an LCA group.
The elements of $\hat{G}$ are called {\em characters} of $G$.
The main theorem of Pontryagin duality is that the natural map $G \rightarrow \Hat{\Hat{G}}$ is an isomorphism.
For an account of Pontryagin duality for groups, see~\cite{armacost}.
The following technical result on LCA groups is due to Kaplansky (unpublished) and first appears in a paper of Glicksberg~\cite{glicksberg}.
See Chapter~10 of~\cite{armacost} for a proof and discussion.
\begin{prop}
\label{topology}
Let $G$ be a group which becomes a LCA group under a topology ${\frak t}$.
Let ${\frak t}'$ be a strictly stronger locally compact group topology on $G$.
Then there is a ${\frak t}'$-continuous character $\gamma$ of $G$ which is not ${\frak t}$-continuous.
\end{prop}
\begin{cor}
If $f:G \rightarrow H$ is a surjective continuous homomorphism of LCA groups whose dual $\hat{f}:\hat{H} \rightarrow \hat{G}$ is also surjective, then $f$ is a topological isomorphism.
\end{cor}
\begin{pf}
Since $\hat{f}$ is surjective, $f$ is injective by~P.23(b) in~\cite{armacost}.
Thus $f$ is an isomorphism, except that $G$ may have a stronger topology than $H$.
If it were strictly stronger, then by the proposition above, there would be a character of $G$ not coming from a character of $H$, contradicting the surjectivity of $\hat{f}$.
Therefore $f$ is a topological isomorphism.
\end{pf}
\begin{cor}
\label{showdual}
Suppose $G$ and $H$ are LCA groups, and there is a continuous bilinear pairing
$$G \times H \rightarrow {\Bbb R}/{\Bbb Z}$$
such that the induced maps
$$f_1:G \rightarrow \hat{H} \;\;\;\;,\;\;\;\; f_2:H \rightarrow \hat{G}$$
are surjective.
Then $f_1$ and $f_2$ are topological isomorphisms.
\end{cor}
\begin{pf}
By a well-known property of the compact-open topology, the induced maps are continuous.
By Pontryagin duality (and chasing definitions), $f_2$ is the dual of $f_1$, so $f_1$ is a topological isomorphism by the previous corollary.
The symmetric argument works for $f_2$.
\end{pf}
We will make use of the theory of Pontryagin duality for topological modules developed by Flood~\cite{flood}.
Let $A$ be a locally compact commutative topological ring.
Consider the class $\cal A$ of locally compact topological $A$-modules.
If $M \in \cal A$, then the Pontryagin dual $\hat{M} \stackrel{\text{def}}{=} \operatorname{Hom}_{\Bbb Z}(M,{\Bbb R}/{\Bbb Z})$ of $M$ as a topological group has a natural $A$-module structure, so $\hat{M} \in \cal A$.
Let us restate the main theorem of~\cite{flood} for this situation:
\begin{theorem}[Pontraygin duality for $A$-modules]
\label{pdfora}
For each $M \in \cal A$, the topological $A$-modules $\hat{M}$ and $\operatorname{Hom}_A(M,\hat{A})$ are canonically isomorphic, and the canonical map
\begin{eqnarray*}
M & \rightarrow & \operatorname{Hom}_A(\operatorname{Hom}_A(M,\hat{A}),\hat{A}) \\
m & \mapsto & (f \mapsto f(m))
\end{eqnarray*}
is an isomorphism of topological $A$-modules.
\end{theorem}
If $M,N \in \cal A$, we say $N$ is the Pontryagin dual of $M$ as an $A$-module, if it is isomorphic to $\operatorname{Hom}_A(M,\hat{A})$ as a topological $A$-module.
The following corollary is an $A$-module version of Corollary~\ref{showdual}.
The next corollary is useful in verifying that two elements of $\cal A$ are Pontryagin duals (as $A$-modules).
\begin{cor}
\label{showdual2}
Suppose $M,N \in \cal A$ and there is a continuous $A$-module pairing
$M \times N \rightarrow \hat{A}$
such that the induced maps
$$f_1:M \rightarrow \operatorname{Hom}_A(N,\hat{A}), \;\;\;\;\;\;\;\; f_2:N \rightarrow \operatorname{Hom}_A(M,\hat{A})$$
are surjective.
Then $f_1$ and $f_2$ are topological $A$-module isomorphisms.
\end{cor}
\begin{pf}
If we compose the pairing with the evaluation-at-1 map $\hat{A} \rightarrow {\Bbb R}/{\Bbb Z}$, we get a pairing
$$M \times N \rightarrow {\Bbb R}/{\Bbb Z}$$
and the induced map $M \rightarrow \hat{N}$ is the same as $f_1$ once we identify $\hat{N}$ with $\operatorname{Hom}_A(N,\hat{A})$ as in Theorem~\ref{pdfora}.
Similarly $f_2$ is the same as the induced map $N \rightarrow \hat{M}$.
In particular, these induced maps are surjective, so by Corollary~\ref{showdual}, they are topological isomorphisms. But we know that $f_1$ and $f_2$ are $A$-module homomorphism as well, so the result follows.
\end{pf}
\begin{lemma}
\label{showdual3}
Suppose $M,N$ are locally compact topological ${{\Bbb F}_q}$-vector spaces.
If there is a continuous ${{\Bbb F}_q}$-linear pairing
$$\langle \;\; , \;\; \rangle : M \times N \rightarrow {{\Bbb F}_q}$$
such that the induced maps
$$M \rightarrow \operatorname{Hom}_{{\Bbb F}_q}(N,{{\Bbb F}_q}) \;\;\;\;,\;\;\;\; N \rightarrow \operatorname{Hom}_{{\Bbb F}_q}(M,{{\Bbb F}_q})$$
are surjective, then $N$ is the Pontryagin dual of $M$ as an ${{\Bbb F}_q}$-vector space.
If furthermore $A$ is a locally compact ${{\Bbb F}_q}$-algebra and $M$ and $N$ are topological $A$-modules such that $\langle am,n \rangle = \langle m,an \rangle$ for all $a \in A$, $m \in M$ and $n \in N$, then $\operatorname{Hom}_{{\Bbb F}_q}(A,{{\Bbb F}_q}) \cong \hat{A}$ as topological $A$-modules and the map
\begin{eqnarray*}
[ \;\; , \;\; ] : M \times N & \rightarrow & \operatorname{Hom}_{{\Bbb F}_q}(A,{{\Bbb F}_q}) \cong \hat{A} \\
m \; , \; n \;\; & \mapsto & (a \mapsto \langle am,n \rangle)
\end{eqnarray*}
is a continuous $A$-module pairing which exhibits $N$ as the Pontryagin dual of $M$ as an $A$-module.
\end{lemma}
\begin{pf}
By counting elements, ${\hat{\Bbb F}_q}$ is a one-dimensional ${{\Bbb F}_q}$-vector space, and we can explicitly identify ${{\Bbb F}_q}$ with ${\hat{\Bbb F}_q}$ by mapping $1$ to the character $\chi=(1/p) \cdot \operatorname{Tr}_{{{\Bbb F}_q}/{{\Bbb F}_p}}$ of ${{\Bbb F}_q}$.
Thus the first half of the proposition is just the special case of Corollary~\ref{showdual2} for the discrete ring ${{\Bbb F}_q}$.
There is an isomorphism $\operatorname{Hom}_{{\Bbb F}_q}(A,{{\Bbb F}_q}) \cong \hat{A}$ of topological ${{\Bbb F}_q}$-vector spaces by Theorem~\ref{pdfora} (with $M=A$, $A={{\Bbb F}_q}$), and it is easily checked that it preserves the $A$-module structure.
As in Corollary~\ref{showdual2}, the pairing $\langle \;\;,\;\; \rangle$ induces a pairing $M \times N \rightarrow {\Bbb R}/{\Bbb Z}$ which identifies $N$ with $\hat{M}$.
The condition $\langle am,n \rangle = \langle m,an \rangle$ ensures that this isomorphism is an isomorphism of $A$-modules.
Finally, if we identify $N \cong \hat{M}$ with $\operatorname{Hom}_A(M,\hat{A})$ using Theorem~\ref{pdfora}, we get a pairing
$$M \times N \rightarrow \hat{A}$$
exhibiting $N$ as the Pontryagin dual of $M$ as an $A$-module, and definition chasing shows that this pairing is $[\;\;,\;\;]$.
\end{pf}
Now let $X$ be a nonsingular projective curve over ${{\Bbb F}_q}$, let $\infty$ be a closed point, and let $Y=X \setminus \infty$.
Let $A$ be the Dedekind ring of regular functions on the affine curve $Y$, and give $A$ the discrete topology.
Let $K$ be its fraction field, and let $K_\infty$ be the completion of $K$ at $\infty$.
The K\"{a}hler differentials of $A$ over ${{\Bbb F}_q}$ (see~\cite{matsumura}), i.e., the differentials on $X$ which are regular away from $\infty$, form a rank one projective $A$-module $\Omega_A$ inside the one-dimensional $K$-vector space $\Omega_K$ of K\"{a}hler differentials of $K$ over ${{\Bbb F}_q}$.
Therefore there is an ideal $J$ of $A$ isomorphic to $\Omega$ as an $A$-module.
Let $\Omega_\infty$ be the completion of $\Omega_K$ at $\infty$.
This is a one-dimensional vector space over $K_\infty$.
By Theorem~3 of Chapter~II in~\cite{weilnt}, any nontrivial character of the locally compact field $K_\infty$ can be used to identify $K_\infty$ with its dual, and it follows that the map
\begin{eqnarray}
\label{residuepairing}
K_\infty \times \Omega_\infty & \rightarrow & {{\Bbb F}_q} \\
a \;\; , \;\; \omega \;\; & \rightarrow & \operatorname{Res}_\infty(a \omega) \nonumber
\end{eqnarray}
exhibits $K_\infty$ and $\Omega_\infty$ as Pontraygin duals of each other, as topological ${{\Bbb F}_q}$-vector spaces.
\begin{theorem}
\label{fractionaldual}
Let $I$ be a fractional ideal of $A$ (with the discrete topology).
Then the pairing~(\ref{residuepairing}) puts $I$ and $\Omega_\infty/(I^{-1} \Omega_A)$ in Pontryagin duality as $A$-modules, so the Pontryagin dual of $I$ as an $A$-module is $K_\infty/(I^{-1}J)$.
\end{theorem}
\begin{pf}
For $\omega \in \Omega_\infty$, we claim that $\omega \in \Omega_A$ if and only if $\operatorname{Res}_\infty(a \omega)=0$ for all $a \in A$.
For $n \ge 0$, let
\begin{eqnarray*}
L(n \infty) & = & \{\, a \in K \mid \operatorname{div}(a) \ge -n \infty \,\} \\
\Omega_{\ge n \infty} & = & \{\, \eta \in \Omega_\infty \mid \operatorname{div}(\eta) \ge n \infty \,\}.
\end{eqnarray*}
(By the divisor of an element of $\Omega_\infty$ we mean only the part at $\infty$, which is the only part that makes sense.)
By the Theorem at the bottom of page~160 in~\cite{eichler}, with $h=0$, ${\frak a}=n \infty$, and $dw$ the principal part system which equals an element of $\Omega_K$ differing from $\omega$ by an element of $\Omega_{\ge n \infty}$ at $\infty$, and equalling zero at other places, we have that $\operatorname{Res}_\infty(a \omega)=0$ for all $a \in L(n \infty)$ if and only if $\omega \in \Omega_A + \Omega_{\ge n \infty}$.
But $A = \bigcup_{n \ge 0} L(n \infty)$, so $\operatorname{Res}_\infty(a \omega)=0$ for all $a \in A$ if and only if
$$\omega \in \bigcap_{n \ge 0} (\Omega_A + \Omega_{\ge n \infty}) = \Omega_A,$$
since $\Omega_A$ is a discrete subgroup of $\Omega_\infty$, and the $\Omega_{\ge n \infty}$ form a decreasing neighborhood base of 0.
Hence $\operatorname{Res}_\infty(a \omega)=0$ for all $a \in I$ if and only if $\omega \in I^{-1} \Omega_A$.
By P.22(d) in~\cite{armacost}, the fact that the pairing~(\ref{residuepairing}) puts $K_\infty$ and $\Omega_\infty$ in Pontryagin duality implies that
$$\hat{I} \cong \operatorname{Hom}_{{\Bbb F}_q}(I,{{\Bbb F}_q}) \cong \Omega_\infty/(I^{-1} \Omega_A) \cong K_\infty/(I^{-1}J),$$
and it is trivial to check that all our isomorphisms respect the $A$-module structures.
\end{pf}
Let $A_{\frak q}$, $K_{\frak q}$ denote the completions of $A$, $K$, respectively, at a nonzero prime ${\frak q}$ of~$A$.
\begin{theorem}
\label{aqdual}
The $A_{\frak q}$-module $\operatorname{Hom}_{{\Bbb F}_q}(K_{\frak q}/A_{\frak q},{{\Bbb F}_q})$ is free of rank one.
\end{theorem}
\begin{pf}
Let $t \in K$ be a uniformizing parameter at ${\frak q}$.
Then $A_{\frak q} \cong {{\Bbb F}_\qq}[[t]]$, where ${{\Bbb F}_\qq}$ is the residue field at ${\frak q}$.
Now for the residue pairing
\begin{eqnarray*}
{{\Bbb F}_\qq}((t)) \times {{\Bbb F}_\qq}((t))dt & \rightarrow & {{\Bbb F}_q} \\
a \;\;\; , \;\;\; \omega \;\;\;\;\;\;\;\;\;\; & \mapsto & \operatorname{Res}_{\frak q}(a \omega)
\end{eqnarray*}
which exhibits the locally compact field $K_{\frak q}$ as its own Pontryagin dual, the set of $\omega$ which give zero when paired with any $a \in {{\Bbb F}_\qq}[[t]]$ is exactly ${{\Bbb F}_\qq}[[t]]dt$, by definition of the residue, so by~P.22(d) in~\cite{armacost}, the Pontryagin dual of $A_{\frak q}={{\Bbb F}_\qq}[[t]]$ is $K_{\frak q}/A_{\frak q}={{\Bbb F}_\qq}((t))/{{\Bbb F}_\qq}[[t]]$.
\end{pf}
\section{Adjoints and Pontryagin duality}
\label{adjointpd}
The {\em adjoint map} on ${\cal F}$ is the map $f \mapsto f^\ast$ defined by the following proposition.
\begin{prop}
There is a norm-preserving multiplication-reversing involution
\begin{eqnarray*}
{\cal F} & \rightarrow & {\cal F} \\
f=\sum_{n \in {\Bbb Z}} a_n \tau^n & \mapsto & f^\ast=\sum_{n \in {\Bbb Z}} {a_n}^{1/q^n} \tau^{-n}.
\end{eqnarray*}
\end{prop}
\begin{pf}
The map is well-defined by Proposition~\ref{convergence}, and it is immediate from the definition that $\|f^\ast\|=\|f\|$.
One can check $f^{\ast \ast}=f$ and $(f \circ g)^\ast = g^\ast \circ f^\ast$ directly by comparing coefficients, or perhaps more enlighteningly by noting that
$$f^\ast = \sum_{n \in {\Bbb Z}} \tau^{-n} a_n$$
is obtained from $f$ by interchanging $\tau$ and $\tau^{-1}$ and reversing the order of multiplication everywhere, and that this preserves the pair of commutation relations $\tau \circ a = a^q \circ \tau$ and $a \circ \tau^{-1} = \tau^{-1} \circ a^q$.
\end{pf}
We now construct an ${{\Bbb F}_q}$-linear pairing between $\ker f$ and $\ker f^\ast$, for nonzero $f \in {\cal F}$.
Suppose $\alpha \in \ker f$ and $\beta \in \ker f^\ast$.
Then $f \circ \alpha$ vanishes on 1 (recall that $f \circ \alpha$ denotes the composition of $f$ with the map $x \mapsto \alpha x$), so by Lemma~\ref{remainder},
\begin{equation}
\label{def}
f \circ \alpha = g_\alpha \circ (1 - \tau)
\end{equation}
for some $g_\alpha \in {\cal F}$, which is uniquely defined, by Corollary~\ref{ffdomain}.
Define
$$\langle \alpha,\beta \rangle_f = g_\alpha^\ast(\beta).$$
\begin{prop}
If $f \in {\cal F}$ is nonzero, then
$$\langle \;\;,\;\; \rangle_f : \ker f \times \ker f^\ast \rightarrow {{\Bbb F}_q}$$
is an ${{\Bbb F}_q}$-vector space pairing.
\end{prop}
\begin{pf}
First let us check that the image lands in ${{\Bbb F}_q}$.
Take adjoints of (\ref{def}):
\begin{equation}
\label{adjdef}
\alpha \circ f^\ast = (1 - \tau^{-1}) \circ g_\alpha^\ast.
\end{equation}
Apply both sides to $\beta$:
$$0 = (1 - \tau^{-1}) g_\alpha^\ast(\beta).$$
So
$$g_\alpha^\ast(\beta) \in \ker(1 - \tau^{-1}) = {{\Bbb F}_q}.$$
From the definition~(\ref{def}), $g_{\alpha+\beta}=g_\alpha+g_\beta$, so the pairing is linear on the left.
Linearity on the right is obvious.
\end{pf}
We now prove a series of results leading up to Theorem~\ref{duality} below.
\begin{lemma}
\label{remainder2}
If $f \in {\cal F}$, then there exists $q \in {\cal F}$ such that
$$f = (1 - \tau) \circ q + f^\ast(1).$$
\end{lemma}
\begin{pf}
Apply Lemma~\ref{remainder} to $f^\ast$ to get $h$ such that
$$f^\ast = h \circ (1-\tau) + f^\ast(1).$$
Then taking adjoints yields
\begin{eqnarray*}
f & = & (1-\tau^{-1}) \circ h^\ast + f^\ast(1) \\
& = & (1-\tau) \circ (-\tau^{-1}) \circ h^\ast + f^\ast(1),
\end{eqnarray*}
so we may take $q=-\tau^{-1} \circ h^\ast$.
\end{pf}
\begin{lemma}
\label{oneone}
If $f,g \in {\cal F}$ and $f \circ (1-\tau) = (1-\tau) \circ g$, then $f^\ast(1)=g(1)$.
\end{lemma}
\begin{pf}
Use Lemmas \ref{remainder} and~\ref{remainder2} to write
\begin{eqnarray*}
f & = & (1-\tau) \circ q + f^\ast(1) \\
g & = & h \circ (1-\tau) + g(1).
\end{eqnarray*}
Substitute these into the given relation:
\begin{equation}
\label{subst}
(1-\tau) \circ q \circ (1-\tau) + f^\ast(1) \circ (1-\tau) = (1-\tau) \circ h \circ (1-\tau) + (1-\tau) \circ g(1).
\end{equation}
Evaluating the given relation at 1 shows $g(1) \in \ker(1-\tau) = {{\Bbb F}_q}$, so $g(1)$ commutes with $1-\tau$ and we can cancel $1-\tau$ on the right in (\ref{subst}) to obtain
$$(1-\tau) \circ q + f^\ast(1) = (1-\tau) \circ h + g(1).$$
Taking adjoints and evaluating at 1 gives the desired result.
\end{pf}
\begin{prop}
\label{backwards}
If $f \in {\cal F}$ is nonzero, $\alpha \in \ker f$, and $\beta \in \ker f^\ast$, then $\langle \alpha,\beta \rangle_f = -\langle \beta,\alpha \rangle_{f^\ast}$.
\end{prop}
\begin{pf}
Write
\begin{eqnarray*}
f \circ \alpha = g_\alpha \circ (1 - \tau) \\
f^\ast \circ \beta = h_\beta \circ (1 - \tau).
\end{eqnarray*}
Multiply the first by $\beta$ on the left, and take the adjoint of the second and multiply by $\alpha$ on the right:
\begin{eqnarray*}
\beta \circ f \circ \alpha & = & \beta \circ g_\alpha \circ (1 - \tau) \\
\beta \circ f \circ \alpha & = & (1 - \tau^{-1}) \circ h_\beta^\ast \circ \alpha \\
& = & (1-\tau) \circ (-\tau^{-1}) \circ h_\beta^\ast \circ \alpha.
\end{eqnarray*}
Equate and apply Lemma~\ref{oneone} to get
\begin{eqnarray*}
(\beta \circ g_\alpha)^\ast(1) & = & ((-\tau^{-1}) \circ h_\beta^\ast \circ \alpha)(1) \\
g_\alpha^\ast(\beta) & = & - \tau^{-1} (h_\beta^\ast(\alpha)) \\
\langle \alpha, \beta \rangle_f & = & - \tau^{-1}(\langle \beta,\alpha \rangle_{f^\ast}) \\
& = & - \langle \beta,\alpha \rangle_{f^\ast},
\end{eqnarray*}
since $\langle \beta,\alpha \rangle_{f^\ast} \in {{\Bbb F}_q}$.
\end{pf}
\begin{theorem}
\label{duality}
If $f \in {\cal F}$ is nonzero, then the pairing
$$\langle \;\;,\;\; \rangle_f : \ker f \times \ker f^\ast \rightarrow {{\Bbb F}_q}$$
exhibits $\ker f$ as the Pontryagin dual of $\ker f^\ast$ as a topological ${{\Bbb F}_q}$-vector space.
\end{theorem}
\begin{pf}
First we check that the pairing is continuous.
Because of Proposition~\ref{backwards}, it will suffice to show that given any bounded subset $B$ of $\ker f^\ast$, there exists a neighborhood $U$ of 0 in $\ker f$ such that $\langle \alpha,\beta \rangle_f=0$ for $\alpha \in U$, $\beta \in B$.
Given $\alpha \in \ker f$, write
$$f \circ \alpha = g_\alpha \circ (1 - \tau)$$
as in the definition of the pairing.
As $\alpha$ tends to zero, each coefficient of $f \circ \alpha$ tends to zero, and so does each coefficient of $g_\alpha$, by the construction in the proof of Lemma~\ref{remainder}.
Then the same is true for the coefficients of $g_\alpha^\ast$.
Combined with the knowledge that $g_\alpha^\ast$ is everywhere convergent for any $\alpha$, this implies that $g_\alpha^\ast(\beta)$ tends to zero uniformly for $\beta \in B$ as $\alpha$ tends to zero, as desired.
Thus the pairing is continuous.
The map induced by the pairing is
\begin{eqnarray*}
\Psi: \ker f & \rightarrow & \operatorname{Hom}_{{\Bbb F}_q}(\ker f^\ast,{{\Bbb F}_q}) \\
\alpha & \mapsto & g_\alpha^\ast|_{\ker f^\ast}
\end{eqnarray*}
This is well-defined and continuous, since the pairing is.
We now show $\Psi$ is surjective.
If $\phi \in \operatorname{Hom}_{{\Bbb F}_q}(\ker f^\ast,{{\Bbb F}_q})$ is nonzero, then $H=\ker \phi$ is some open and closed subspace in $G=\ker f^\ast$, so $H \in {\cal G}$ by~(6) in Proposition~\ref{class}.
Thus by Theorem~\ref{existence}, there exists $h \in {\cal F}$ such that $\ker h = H$.
Since $H$ has codimension 1 in $G$ (as an ${{\Bbb F}_q}$-vector space), $h$ must map $G$ to a one-dimensional ${{\Bbb F}_q}$-vector space inside $C$.
By scaling $h$, we may assume $h(G)={{\Bbb F}_q}$.
Now $(1 - \tau) \circ h$ has kernel $G$, so by the uniqueness in Theorem~\ref{existence},
$$\tau^n \alpha \circ f^\ast = (1 - \tau) \circ h$$
for some $\alpha \in C^\ast$ and $n \in {\Bbb Z}$.
Take adjoints and multiply on the right by $\tau^n$:
\begin{eqnarray*}
f \circ \alpha & = & h^\ast \circ (1-\tau^{-1}) \circ \tau^n \\
& = & h^\ast \circ (-\tau^{n-1}) \circ (1-\tau).
\end{eqnarray*}
Evaluating at 1 shows that $\alpha \in \ker f$.
Comparing with~(\ref{def}) shows
\begin{eqnarray*}
g_\alpha & = & h^\ast \circ (-\tau^{n-1}) \\
g_\alpha^\ast & = & - \tau^{1-n} \circ h \\
g_\alpha^\ast(G) & = & - \tau^{1-n}(h(G)) \\
& = & \tau^{1-n}({{\Bbb F}_q}) \\
& = & {{\Bbb F}_q}.
\end{eqnarray*}
Thus $\phi$ and $\Psi(\alpha)=g_\alpha^\ast|_{\ker f^\ast}$ are nonzero ${{\Bbb F}_q}$-homomorphisms from $G$ to ${{\Bbb F}_q}$ with the same kernel $H$, so $\phi = c \Psi(\alpha) = \Psi(c \alpha)$ for some $c \in {{\Bbb F}_q}^\ast$.
Therefore $\Psi$ is surjective.
Applying the above argument to $f^\ast$ and invoking Proposition~\ref{backwards} shows that the other induced map
$$\ker f^\ast \rightarrow \operatorname{Hom}_{{\Bbb F}_q}(\ker f,{{\Bbb F}_q})$$
is surjective as well.
Now apply Lemma~\ref{showdual3}.
\end{pf}
\begin{rems}
If $f = \sum_{i=m}^n a_i \tau^i$ is a finite series with $a_i$ in {\em any} field $L$ of characteristic $p$, then the construction yields a perfect pairing between the kernels of $f$ and $f^\ast$ acting on $\overline{L}$.
Here, in addition, the pairing will be $\operatorname{Gal}(L^{\text{sep}}/L)$-equivariant, since its construction is canonical.
The same will holds for infinite series, when $L$ has a non-archimedean valuation and $L^{\text{sep}}$ is replaced by topological Galois closure.
\end{rems}
\medskip
Next we prove a compatibility result.
\begin{prop}
\label{compatible}
Let $f$ and $h$ be nonzero elements of ${\cal F}$.
Then for all $\alpha \in \ker(f\circ h)$ and $\beta \in \ker f^\ast \subseteq \ker (f \circ h)^\ast$,
$$\langle \alpha,\beta \rangle_{f \circ h} = \langle h(\alpha),\beta \rangle_f.$$
Similarly, for all $\alpha \in \ker f \subseteq \ker(h \circ f), \beta \in \ker(h \circ f)^\ast$,
$$\langle \alpha,\beta \rangle_{h \circ f} = \langle \alpha,h^\ast(\beta) \rangle_f.$$
\end{prop}
\begin{pf}
For the first part, write
\begin{eqnarray*}
(f \circ h) \circ \alpha & = & G_\alpha \circ (1 - \tau) \\
f \circ h(\alpha) & = & g_{h(\alpha)} \circ (1 - \tau)
\end{eqnarray*}
Subtract to get
$$f \circ (h \circ \alpha - h(\alpha)) = (G_\alpha-g_{h(\alpha)}) \circ (1 - \tau).$$
Since $h \circ \alpha - h(\alpha)$ kills $1$,
$$h \circ \alpha - h(\alpha) = q \circ (1 - \tau)$$
for some $q \in {\cal F}$, by Lemma~\ref{remainder}.
Substitute and cancel $1-\tau$ on the right:
$$f \circ q = G_\alpha - g_{h(\alpha)}.$$
Take adjoints and apply both sides to $\beta \in \ker f^\ast$ to get
\begin{eqnarray*}
0 & = & G_\alpha^\ast(\beta) - g_{h(\alpha)}^\ast(\beta) \\
& = & \langle \alpha, \beta \rangle_{f \circ h} - \langle h(\alpha),\beta \rangle_f
\end{eqnarray*}
as desired.
For the second part, we have
\begin{eqnarray*}
\langle \alpha,\beta \rangle_{h \circ f}
& = & - \langle \beta, \alpha \rangle_{f^\ast \circ h^\ast} \;\;\; \text{(by Proposition~\ref{backwards})} \\
& = & - \langle h^\ast(\beta), \alpha \rangle_{f^\ast} \;\;\; \text{(by what we just showed)} \\
& = & \langle \alpha,h^\ast(\beta) \rangle_f,
\end{eqnarray*}
by Proposition~\ref{backwards} again.
\end{pf}
\begin{prop}
\label{annihilator}
Suppose $h,g \in {\cal F}$ and $f=h \circ g$.
Then
$$\{\, \beta \in \ker f^\ast : \langle \alpha,\beta \rangle_f = 0 \;\; \forall \alpha \in \ker g \,\} = \ker h^\ast.$$
\end{prop}
\begin{pf}
Suppose $\beta \in \ker f^\ast$.
Then, by Proposition~\ref{compatible},
\begin{eqnarray*}
\langle \alpha,\beta \rangle_f = 0 \;\; \forall \alpha \in \ker g
& \iff & \langle \alpha,h^\ast(\beta) \rangle_g = 0 \;\; \forall \alpha \in \ker g \\
& \iff & h^\ast(\beta)=0 \;\;\; \text{ (by Theorem~\ref{duality})}.
\end{eqnarray*}
\end{pf}
Below are three more properties of the pairing.
These were discovered by Elkies~\cite{elkies} (for the case of additive polynomials), although some of our proofs are new.
We will not need these in the sequel, but they are of interest in their own right.
First we explain how the pairing changes if we change the underlying finite field.
\begin{prop}
Let $q'=q^m$, for some $m \ge 1$.
If $f \in {\cal F}$ is actually ${{\Bbb F}_{q'}}$-linear, then we can define another pairing $\langle \;\;,\;\; \rangle'_f$ between $\ker f$ and $\ker f^\ast$, this time taking values in ${{\Bbb F}_{q'}}$.
(Note that $f^\ast$ is the same whether $f$ is considered as ${{\Bbb F}_q}$-linear or ${{\Bbb F}_{q'}}$-linear.)
Then $\langle \alpha,\beta \rangle_f = \operatorname{Tr}_{{{\Bbb F}_{q'}}/{{\Bbb F}_q}} \langle \alpha,\beta \rangle'_f$.
\end{prop}
\begin{pf}
Write
\begin{eqnarray*}
f \circ \alpha & = & G_\alpha \circ (1 - \tau^m) \\
& = & G_\alpha \circ (1+ \tau + \cdots + \tau^{m-1}) \circ (1-\tau).
\end{eqnarray*}
By definition of the pairing,
\begin{eqnarray*}
\langle \alpha,\beta \rangle_f & = & (G_\alpha \circ (1+ \tau + \cdots + \tau^{m-1}))^\ast(\beta) \\
& = & (1+\tau + \cdots + \tau^{m-1})^\ast(G_\alpha^\ast(\beta)) \\
& = & (1+\tau^{-1} + \cdots + \tau^{-(m-1)})(\langle \alpha,\beta \rangle'_f) \\
& = & \operatorname{Tr}_{{{\Bbb F}_{q'}}/{{\Bbb F}_q}} \langle \alpha,\beta \rangle'_f, \\
\end{eqnarray*}
since $\operatorname{Gal}({{\Bbb F}_{q'}}/{{\Bbb F}_q}) = \{ 1, \tau^{-1},\ldots,\tau^{-(m-1)} \}$.
\end{pf}
\begin{theorem}
\label{skewsymmetric}
Suppose $f \in {\cal F}$ is nonzero and $f=f^\ast$.
Then
\begin{enumerate}
\item $\langle \alpha,\alpha \rangle_f = 0$ for all $\alpha \in \ker f$.
\item If $h \in {\cal F}$ and $f=h^\ast \circ h$, then $\ker h$ is a maximal isotropic closed sub-${{\Bbb F}_q}$-vector space of $\ker f$ under $\langle \;,\; \rangle_f$.
\item Conversely, if $H$ is a maximal isotropic closed sub-${{\Bbb F}_q}$-vector space of $\ker f$ under $\langle \;,\; \rangle_f$, then there exists $h \in {\cal F}$ such that $f=h^\ast \circ h$ and $\ker h = H$.
\end{enumerate}
\end{theorem}
\begin{pf}
If $p \not=2$, then part~(1) follows already from Proposition~\ref{backwards}.
Hence assume $p=2$.
Write $f \circ \alpha = g_\alpha \circ (1-\tau)$ as usual.
Multiply by $\alpha$ on the left, and apply Lemma~\ref{remainder2} to write
\begin{eqnarray*}
\alpha \circ f \circ \alpha & = & [(1-\tau) \circ q + (\alpha \circ g_\alpha)^\ast(1)] \circ (1-\tau) \\
& = & (1-\tau^{-1}) \circ j \circ (1-\tau) + g_\alpha^\ast(\alpha) \circ (1-\tau) \;\;\; \text{(where $j=\tau \circ q$).}
\end{eqnarray*}
Subtract this from its adjoint and use $f=f^\ast$ and $g_\alpha^\ast(\alpha) \in {{\Bbb F}_q}$ to get
$$0 = (1-\tau^{-1}) \circ (j^\ast-j) \circ (1-\tau) + g_\alpha^\ast(\alpha) \circ (\tau-\tau^{-1}).$$
Since $\tau-\tau^{-1}=(1-\tau^{-1}) \circ (1-\tau)$ in characteristic 2, we can cancel $1-\tau^{-1}$ on the left and $1-\tau$ on the right to obtain
$$0 = (j^\ast-j) + g_\alpha^\ast(\alpha).$$
Take coefficients of $\tau^0$ to deduce
$$\langle \alpha,\alpha \rangle_f = g_\alpha^\ast(\alpha) = 0.$$
This proves part~(1).
Now, for part~(2), $h$ is continuous, so $\ker h$ is closed.
Proposition~\ref{annihilator} says that
$$\{\, \beta \in \ker f^\ast : \langle \alpha,\beta \rangle_f = 0 \;\; \forall \alpha \in \ker h \,\} = \ker h,$$
which proves part~(2).
Finally let us prove~(3).
Assume $H$ is a maximal isotropic closed sub-${{\Bbb F}_q}$-vector space of $\ker f$ under $\langle \;,\; \rangle$.
Then $H \in {\cal G}$ by~(6) in Proposition~\ref{class}.
By Theorem~\ref{existence} we can find $g \in {\cal F}$ with kernel $H$, and by Theorem~\ref{factors}, we have $f = j \circ g$ for some $j \in {\cal F}$.
By Proposition~\ref{annihilator},
$$\ker j^\ast = \{\, \beta \in \ker f^\ast : \langle \alpha,\beta \rangle_f = 0 \;\; \forall \alpha \in H \,\} =H,$$
since $H$ is maximal isotropic.
By the uniqueness in Theorem~\ref{existence}, we have $j^\ast = \epsilon \circ \tau^n \circ g$ where $\epsilon \in C^\ast$ and $n \in {\Bbb Z}$.
Then
$$f = j \circ g = g^\ast \circ \tau^{-n} \circ \epsilon \circ g.$$
Taking adjoints and using $f=f^\ast$ shows that
$$f = g^\ast \circ \epsilon \circ \tau^n \circ g.$$
If we equate and cancel $g^\ast$ on the left and $g$ on the right, we obtain
$$\tau^{-n} \circ \epsilon = \epsilon \circ \tau^n,$$
which forces $n=0$. Now simply take $h = \sqrt{\epsilon} \circ g$ where $\sqrt{\epsilon}$ is any square root of $\epsilon$ in $C$.
\end{pf}
\begin{theorem}
\label{symmetric}
If $f \in {\cal F}$ and $f^\ast=-f$, then $\langle \alpha,\beta \rangle_f = \langle \beta,\alpha \rangle_f$ for all $\alpha,\beta \in \ker f$.
\end{theorem}
\begin{pf}
By definition of the pairing, $\langle \;,\; \rangle_{-f} = -\langle \;,\; \rangle_f$, so
\begin{eqnarray*}
\langle \alpha,\beta \rangle_f & = & -\langle \beta,\alpha \rangle_{f^\ast} \quad \text{(by Proposition~\ref{backwards})} \\
& = & -\langle \beta,\alpha \rangle_{-f} \\
& = & \langle \beta,\alpha \rangle_f.
\end{eqnarray*}
\end{pf}
\begin{rems}
(I thank Noam Elkies for these.)
The theory of additive polynomials shares much with the theory of differential operators, as pointed out by Ore~\cite{Or1}.
In fact the pairing $\langle \;,\; \rangle_f$ can be obtained an analogue of a known pairing on kernels of differential operators.
If $f$ is a differential operator and $f^\ast$ denotes its formal adjoint, then we have an identity $uf^\ast(v)-vf(u)=B'$ where $B$ is a bilinear form in $\{u,u',u'',\ldots,u^{(n-1)}\}$ and $\{v,v',v'',\ldots,v^{(n-1)}\}$.
(See p.\ 124 in~\cite{ince}, where $B$ is called a ``bilinear concomitant.'')
If $u \in \ker f$ and $v \in \ker f^\ast$, then $B(u,v)$ is a constant depending bilinearly on $u$ and $v$.
Similarly, if $f \in {\cal F}$, then one can write $uf^\ast(v)-vf(u)=B^q-B$ where $B$ is a (potentially infinite) linear combination of terms of the form $u^{q^m} v^{q^n}$ ($m,n \in {\Bbb Z}$).
One can check that $B$ converges to a continuous bilinear function on $C \times C$ by writing $B$ explicitly in terms of the coefficients of $f$, and using Proposition~\ref{convergence}.
Clearly $B(\alpha,\beta) \in {{\Bbb F}_q}$ if $\alpha \in \ker f$ and $\beta \in \ker f^\ast$.
\end{rems}
\begin{prop}
\label{ince}
If $\alpha \in \ker f$ and $\beta \in \ker f^\ast$, then $B(\alpha,\beta)=\langle \alpha,\beta \rangle_f$.
\end{prop}
\begin{pf}
Let $b_\alpha(v)=B(\alpha,v)$.
If we set $u=\alpha$ in the defining equation for $B$, we have as functions of $v$,
\begin{eqnarray*}
\alpha \circ f^\ast & = & (\tau-1) \circ b_\alpha, \\
f \circ \alpha & = & b_\alpha^\ast \circ (\tau^{-1}-1) \\
& = & b_\alpha^\ast \circ \tau^{-1} \circ (1 -\tau).
\end{eqnarray*}
Comparing with equation~(\ref{def}) shows that $g_\alpha = b_\alpha^\ast \circ \tau^{-1}$, so
\begin{eqnarray*}
\langle \alpha,\beta \rangle_f & = & g_\alpha^\ast(\beta) \\
& = & (\tau \circ b_\alpha)(\beta) \\
& = & \tau(B(\alpha,\beta)) \\
& = & B(\alpha,\beta),
\end{eqnarray*}
since $B(\alpha,\beta) \in {{\Bbb F}_q}$.
\end{pf}
This alternative definition of the pairing could clearly be used to give new proofs of the properties of the pairing.
For instance, Proposition~\ref{backwards} would be immediate, as would the first part of Theorem~\ref{skewsymmetric}.
\section{$A$-module pairings for Drinfeld modules}
\label{amodule}
Let us retain the assumption that $A$ is the affine ring of a nonsingular projective curve over ${{\Bbb F}_q}$ minus a closed point $\infty$, and suppose we have an ${{\Bbb F}_q}$-algebra homomorphism $\iota: A \rightarrow C$.
The kernel ${\frak p}$ of $\iota$ will be called the characteristic.
Let $a \mapsto \phi_a$ and $a \mapsto \psi_a$ be ring homomorphisms from $A$ to ${\cal F}$.
For example, $\phi$ and $\psi$ might be Drinfeld $A$-modules over $C$.
We say that a nonzero $f \in {\cal F}$ is an {\em isogeny} from $\phi$ to $\psi$ if $\psi_a \circ f = f \circ \phi_a$ for all $a \in A$.
If $f: \phi \rightarrow \psi$ is an isogeny, then $\ker f$ is an $A$-module via $\phi$ and $\ker f^\ast$ is an $A$-module via $\psi^\ast$.
\begin{prop}
\label{apairing}
Let $a \mapsto \phi_a$ and $a \mapsto \psi_a$ be ring homomorphisms from $A$ to ${\cal F}$.
Suppose $f \in {\cal F}$ is an isogeny from $\phi$ to $\psi$.
Then for all $a \in A$, $\alpha \in \ker f$, $\beta \in \ker f^\ast$, we have
$$\langle \phi_a(\alpha),\beta \rangle_f = \langle \alpha,\psi_a^\ast(\beta) \rangle_f,$$
and
\begin{eqnarray*}
[ \;\; , \;\; ]_f : \ker f \times \ker f^\ast & \rightarrow & \operatorname{Hom}_{{\Bbb F}_q}(A,{{\Bbb F}_q}) \\
\alpha \; , \; \beta \;\;\;\;\;\;\; & \mapsto & (a \mapsto \langle \phi_a(\alpha),\beta \rangle)
\end{eqnarray*}
is a pairing of $A$-modules which exhibits $\ker f^\ast$ as the Pontryagin dual of $\ker f$ as an $A$-module.
Furthermore, if $\Psi:A \rightarrow {\cal F}$ is a third ring homomorphism, and $h:\Psi \rightarrow \phi$ is another isogeny, then we have the following two compatibility relations:
if $\alpha \in \ker(f\circ h),\beta \in \ker f^\ast \subseteq \ker (f \circ h)^\ast$, then
$$[ \alpha,\beta ]_{f \circ h} = [ h(\alpha),\beta ]_f.$$
Similarly, for all $\alpha \in \ker h \subseteq \ker(f \circ h), \beta \in \ker(f \circ h)^\ast$,
$$[ \alpha,\beta ]_{f \circ h} = [ \alpha,f^\ast(\beta) ]_h.$$
\end{prop}
\begin{pf}
Given $a \in A$, $\alpha \in \ker f$, $\beta \in \ker f^\ast$, we have
\begin{eqnarray*}
\langle \phi_a(\alpha),\beta \rangle_f & = & \langle \alpha,\beta \rangle_{f \circ \phi_a} \;\;\; \text{(by Proposition~\ref{compatible})} \\
& = & \langle \alpha,\beta \rangle_{\psi_a \circ f} \\
& = & \langle \alpha,\psi_a^\ast(\beta) \rangle_f \;\;\; \text{(by Proposition~\ref{compatible} again}).
\end{eqnarray*}
Apply Lemma~\ref{showdual3} to get the duality property of $[\;\;,\;\;]_f$.
The compatibility relations follow from those for $\langle \;\;,\;\; \rangle$ proved in Proposition~\ref{compatible}.
\end{pf}
\begin{cor}
\label{weil}
Let $\phi$ be a Drinfeld $A$-module over a field $L$.
Fix a nonzero $a \in A$.
Let $\phi[a]$ be the kernel of $\phi_a$ on $\overline{L}$, and similarly define $\phi^\ast[a]=\ker \phi_a^\ast$.
There is a Galois-equivariant perfect pairing of finite $A$-modules
$$[ \;\;,\;\; ]_a: \phi[a] \times \phi^\ast[a] \rightarrow \widehat{(A/a)} \subset \hat{A}.$$
(Here $\operatorname{Gal}(L^{\text{sep}}/L)$ acts trivially on $\hat{A}$ and $\widehat{(A/a)}$.)
If we also have a nonzero $b \in A$, then for any $\alpha \in \phi[ab]$, $\beta \in \psi^\ast[a]$
$$[ \alpha,\beta ]_{ab} = [\phi_b(\alpha),\beta]_a.$$
Similarly, for $\alpha \in \phi[a]$, $\beta \in \psi^\ast[ab]$,
$$[ \alpha,\beta ]_{ab} = [\alpha,\psi^\ast_b(\beta)]_a.$$
\end{cor}
The properties above of $[\;\;,\;\;]_a$ should remind one of the Weil pairing for abelian varieties.
Another proof of the Galois-equivariant duality between $\phi[a]$ and $\phi^\ast[a]$ was discovered independently by Taguchi.
(See the appendix of~\cite{gossadjoint}.)
As another application of our pairings, we can describe the kernel of the adjoint of the exponential function associated with a Drinfeld module.
\begin{cor}
\label{estar}
Let $C$ be the completion of the algebraic closure of $K_\infty$ with $|\;|_\infty$, and let $\phi$ be a Drinfeld $A$-module over $C$.
Let $e(z) \in C[[z]]$ be the associated exponential function.
Then there is a natural pairing
$$\ker e \times \ker e^\ast \rightarrow \hat{A}$$
which exhibits $\ker e^\ast$ as the Pontryagin dual of $\ker e$ as an $A$-module.
If the lattice $\ker e$ is isomorphic to the direct sum of fractional ideals $I_1 \oplus \cdots \oplus I_r$, then $\ker e^\ast$ is isomorphic to $K_\infty/(I_1^{-1}J) \oplus \cdots \oplus K_\infty/(I_r^{-1}J)$ as a topological $A$-module.
(Here $J$ is as in Theorem~\ref{fractionaldual}.)
\end{cor}
\begin{pf}
Simply note that $e$ is an isogeny from $C$ with the standard $A$-module structure to $\phi$, and apply Proposition~\ref{apairing}.
The last assertion follows from Theorem~\ref{fractionaldual}.
\end{pf}
\begin{prop}
Let $\phi$ and $e$ be as in the previous corollary.
Then $\ker e^\ast$ is the closure of $\bigcup_{a \in A} \phi^\ast[a]$.
\end{prop}
\begin{pf}
Taking adjoints of
$$e \circ a = \phi_a \circ e$$
yields
$$a \circ e^\ast = e^\ast \circ \phi^\ast_a,$$
from which it is clear that $\phi^\ast[a] \subset \ker e^\ast$.
Corollary~\ref{estar} says that $\ker e^\ast$ as a topological $A$-module via $\phi^\ast$ is isomorphic to $K_\infty/C_1 \oplus \cdots \oplus K_\infty/C_r$, for some fractional $A$-ideals $C_1,\ldots,C_r$.
The torsion submodule in this module is $K/C_1 \oplus \cdots \oplus K/C_r$, which is dense, so the result follows.
\end{pf}
\section{Tate module pairings}
Because of the compatibility relations, we can use our pairings to construct a pairing between Tate modules, just as the Weil pairing gives rise to a pairing between Tate modules.
Let ${\frak q}$ be a nonzero prime of $A$ different from the characteristic ${\frak p}$ of the Drinfeld module.
Let $K$ be the fraction field of $A$, and let $A_{\frak q}$, $K_{\frak q}$ be the completions at the prime ${\frak q}$.
The Tate module of a Drinfeld $A$-module over $L$ is
$$T_{\frak q}(\phi) = \operatorname{Hom}_A(K_{\frak q}/A_{\frak q},\phi(\overline{L}))$$
where $\phi(\overline{L})$ denotes the additive group of $\overline{L}$ with the $A$-module structure given by $\phi$.
Similarly we define the Tate module of $\phi^\ast$ as
$$T_{\frak q}(\phi^\ast) = \operatorname{Hom}_A(K_{\frak q}/A_{\frak q},\phi^\ast(\overline{L})).$$
These are both free $A_{\frak q}$-modules of rank equal to the rank of the Drinfeld module.
\begin{prop}
\label{tate}
There is a Galois-equivariant continuous perfect pairing of $A_{\frak q}$-modules
$$[\;\;,\;\;]_{\frak q} : T_{\frak q}(\phi) \times T_{\frak q}(\phi^\ast) \rightarrow \operatorname{Hom}_{{\Bbb F}_q}(K_{\frak q}/A_{\frak q},{{\Bbb F}_q}) \cong A_{\frak q}$$.
\end{prop}
\begin{pf}
Given $\alpha \in T_{\frak q}(\phi)$, $\beta \in T_{\frak q}(\phi^\ast)$, define $[\alpha,\beta]_{\frak q} \in \operatorname{Hom}_{{\Bbb F}_q}(K_{\frak q}/A_{\frak q},{{\Bbb F}_q})$ by
$$[\alpha,\beta]_{\frak q}(b)=[\alpha(b),\beta(a^{-1})]_a$$
where $a \in A$ kills $b$ and generates an ideal power of ${\frak q}$.
As is well known, $A$ is a Dedekind domain with finite class number, so some power ${\frak q}^h$ of ${\frak q}$ is principal, so it is always possible to find such an $a$, given $b$.
The definition is independent of the choice of $a$, by the last compatibility relation in Corollary~\ref{weil}.
If we fix $a \in A$ generating a power of ${\frak q}$, we can also obtain the pairing as the inverse limit of the pairings
$$[\;\;,\;\;]_{a^n} : \phi[a^n] \times \phi^\ast[a^n] \longrightarrow \operatorname{Hom}_{{\Bbb F}_q}(A/a^n,{{\Bbb F}_q}) = \operatorname{Hom}_{{\Bbb F}_q}(a^{-n}A/A,{{\Bbb F}_q}),$$
with respect to the maps
\begin{eqnarray*}
\phi[a^{n+1}] & \stackrel{a}{\longrightarrow} & \phi[a^n] \\
\phi^\ast[a^{n+1}] & \stackrel{a}{\longrightarrow} & \phi^\ast[a^n] \\
\operatorname{Hom}_{{\Bbb F}_q}(a^{-(n+1)}A/A,{{\Bbb F}_q}) & \stackrel{\text{res}}{\longrightarrow} & \operatorname{Hom}_{{\Bbb F}_q}(a^{-n}A/A,{{\Bbb F}_q}).
\end{eqnarray*}
The Galois-equivariance, continuity, and perfectness of the pairing then follow from the properties of the pairings $[\;\;,\;\;]_{a^n}$.
Finally, $\operatorname{Hom}_{{\Bbb F}_q}(K_{\frak q}/A_{\frak q},{{\Bbb F}_q})$ is (non-canonically) isomorphic to $A_{\frak q}$ as an $A_{\frak q}$-module, by Theorem~\ref{aqdual}.
\end{pf}
As pointed out by Goss, this duality of Tate modules has the following corollary.
\begin{cor}
\label{semisimple}
Suppose the Drinfeld module $\phi$ is defined over a finite extension $L$ of $K$.
Then $$T_{\frak q}(\phi^\ast) \otimes_{A_{\frak q}} K_{\frak q}$$ is a semi-simple $K_{\frak q}[\operatorname{Gal}(L^{\text{sep}}/L)]$-module.
\end{cor}
\begin{pf}
Combine the previous theorem with the main theorem in~\cite{taguchi}.
\end{pf}
\section{Questions}
\label{questions}
Our results show that in two cohomological realizations (the period lattice and Tate module), the cohomology of the adjoint of a Drinfeld module is dual to the cohomology of the original Drinfeld module.
There is a third cohomological realization of a Drinfeld module, namely the de Rham cohomology~\cite{gekeler} developed by Anderson, Deligne, Gekeler, and Yu.
\begin{question}
Is it possible to give a reasonable definition for the de Rham cohomology of the adjoint of a Drinfeld module and prove that it is dual in some sense to the de Rham cohomology of the original Drinfeld module?
\end{question}
It would also be nice to generalize the applications of this paper to include $t$-modules, the higher dimensional analogues of Drinfeld modules.
\begin{question}
Can one prove results similar to those in this paper for fractional power series in more than one variable?
\end{question}
A reasonable approach is to work with the ring ${{\Bbb M}_d}({\cal F})$ of $d$-by-$d$ matrices with coefficients in ${\cal F}$.
Elements of this ring act on $C^d$ in an obvious way.
Moreover, there is a multiplication-reversing adjoint map ${{\Bbb M}_d}({\cal F}) \rightarrow {{\Bbb M}_d}({\cal F})$, which maps a matrix $A=(a_{ij})$ to $A^\ast=(a_{ji}^\ast)$, i.e., the transpose of the matrix obtained by taking the adjoint of each entry.
Hence, for example, it is possible to define the adjoint of a $t$-module
\begin{eqnarray*}
\phi: {{\Bbb F}_q}[t] & \rightarrow & {{\Bbb M}_d}(C\{\tau\}) \subset {{\Bbb M}_d}({\cal F}) \\
a & \mapsto & \phi_a
\end{eqnarray*}
as
\begin{eqnarray*}
\phi^\ast: {{\Bbb F}_q}[t] & \rightarrow & {{\Bbb M}_d}({\cal F}) \\
a & \mapsto & \phi_a^\ast.
\end{eqnarray*}
For each $A = (a_{ij}) \in {{\Bbb M}_d}({\cal F})$ it is also possible to define a natural bilinear pairing
$$\langle \;,\; \rangle_A : \ker A \times \ker A^\ast \rightarrow {{\Bbb F}_q}$$
by following the differential operator analogy.
Let $\cdot$ denote the standard inner product on $C^d$.
As in the remarks at the end of Section~\ref{adjointpd}, for each $f \in {\cal F}$, let $B_f(u,v)$ be the bilinear function of $u,v \in C$ such that
$$u f^\ast(v) - v f(u) = B_f^q - B_f.$$
Then, for $u,v \in C^d$, the bilinear function
$$B_A(u,v) = \sum_{i,j} B_{a_{ij}}(u_i,v_j)$$
satisfies
$$u \cdot A^\ast(v) - v \cdot A(u) = B_A^q-B_A,$$
and we may define
$$\langle u,v \rangle_A = B_A(u,v).$$
Thus for instance, one has a pairing between the $a$-torsion of a $t$-module and the $a$-torsion of its adjoint.
\section*{Acknowledgements}
Many thanks go to David Goss, who introduced me to adjoints of polynomials and Drinfeld modules, and asked some of the questions answered by this paper.
I thank T.\ Y.\ Lam, Hendrik Lenstra, and Ken Ribet for helping me find references for Newton polygons.
Thanks also to Noam Elkies for sharing his preprint with me, and to Michael Rosen for suggesting references of use in proving Theorem~\ref{fractionaldual}.
Finally, I thank the referee for many intelligent and helpful suggestions.
|
\setcounter{equation}{0}\section@ori{\setcounter{equation}{0}\setcounter{equation}{0}\section@ori@ori}
}
\makeatother
\makeatletter
\newlength{\lr@listTW}
\newlength{\lr@listLW}
\def\lr@list#1{\def\lr@listDL{#1}%
\settowidth{\lr@listLW}{\lr@listDL}%
\lr@listTW=\linewidth%
\addtolength{\lr@listTW}{-\lr@listLW}%
\hbox to\lr@listLW{}\begin{minipage}[t]{\lr@listTW
}
\def\lrlist{%
\def\lr@listPAR{\noindent}%
\def\lr@listPL[##1] {%
\lr@listPAR\raisebox{0pt}[0pt][0pt]{\llap{%
\parbox[t]{\lr@listLW}{##1}}}%
\def\lr@listPAR{\par\noindent}}%
\def\item{\@ifnextchar [{\lr@listPL}{\lr@listPDL}}%
\def\lr@listPDL{%
\lr@listPAR\raisebox{0pt}[0pt][0pt]{\llap{%
\parbox[t]{\lr@listLW}{\lr@listDL}}}%
\def\lr@listPAR{\par\noindent}}%
\lr@list}%
\def\end{minipage}{\end{minipage}}
\makeatother
\def\figurename{Fig.}
\setlength{\textfloatsep}{0.4\baselineskip plus 2pt minus 4pt}
\ifx\LRcaptionwidth\UNDEFINE
\newlength{\LRcaptionwidth}\else\fi
\LRcaptionwidth=\textwidth
\addtolength{\LRcaptionwidth}{-2\parindent}
\begin{document}
\hbadness=10000
\ifx\hyperref\undefined\else\errmessage{HYPER DISABLED}\fi
\setcounter{page}{1}
\thispagestyle{empty}
\begin{samepage}
\title{%
{\normalsize\sf%
{\sl August 1995\/}\hfill
\begin{tabular}[t]{l}
GSI-Preprint-95-46 \\
hep-ph/9508318
\end{tabular}}\\*[5ex]
{\bf Photon Intensity Interferometry\\ for Expanding Sources}
}
\author{Leonid V. Razumov%
\thanks{E.Mail: \sf\,\, razumov@clri6a.gsi.de\ \
(URL$\Rightarrow$ http://www.gsi.de/groups/the/razumov.html)}\,\,,
\quad Hans Feldmeier\setcounter{footnote}{3}%
\thanks{E.Mail: \sf\,\, <EMAIL>}
}
\date{Gesellschaft f\"ur Schwerionenforschung, Postfach 110552,\\
D-64220 Darmstadt, Germany}
\maketitle
\setcounter{footnote}{0}
\thispagestyle{empty}
\begin{center}
{\large\bf Abstract}\\[2ex]
\end{center}
{\narrower\narrower\par\noindent \small
Using Quantum Field Theory we derive a general formula for the
double inclusive spectra of photons radiated by a system in local
equilibrium. The derived expression differs significantly from the one
mostly used up to now in photon intensity interferometry of heavy--ion
collisions. We present a covariant expression for double inclusive spectra
adapted for usage in numerical simulations. Application to a schematic
model with a Bj\o rken type expansion gives strong evidence for the need
of reinvestigating photon--photon correlations for expanding sources.
\par}
\vskip 15mm
\noindent
\centerline{\large\sl Submitted to Physics Letters B\/.}
\end{samepage}
\newpage
\noindent
Double inclusive spectra of hard photons radiated at the early stage of
heavy--ion collisions have become an important subject of theoretical
\cite{Neuh86,SriKap93a,TimPlumRazWein94} and experimental
\cite{GANIL94} investigations.
Since these photons originate from very energetic collisions and interact
very little with the matter they are traversing, they may carry
signatures of the hot and dense phase of the matter, in particular of
the quark--gluon plasma.
Except for the work of D.~Neuhauser \cite{Neuh86} in which he treated
a static source, usually one is analyzing the correlations in two--photon
coincidence measurements with expressions based on the corresponding
ones for scalar bosons \cite{Pratt84} augmented with a degeneracy
factor which is supposed to take care of the spin structure of the
photons.
The aim of this work is to derive in a quantum--field treatment
covariant expressions for the two--photon density and herewith the
corresponding correlation function. In particular we pay attention to
charge conservation and to the fact that massless photons have only two
helicity eigenstates. Due to these properties averaging over all spin
directions results in a reduction factor of one--half compared to the
scalar case only if the momenta of the two photons are parallel. For a
finite angle, however, the interference is reduced, so that the deduced
size of the source is smaller than the one obtained from the
correlation function for scalar bosons.
For Bj\o{}rken type hydrodynamics
we estimate a difference in the apparent source size of about 30\% in
rapidity. Therefore, we conclude that the ``pion--inspired'' formulas for
the Bose--Einstein correlation function is not applicable for
photon intensity interferometry studies, especially if one deals with an
expanding source.
The finally measured number of photons per invariant momentum bin with
polarization $\,\lambda\,$ is given by the density matrix of the final
state $\rho_f$ which can be expressed in terms of the initial density
matrix $\rho_i$ and the $S$--matrix using the relation
\mbox{$\rho_f = S \rho_i S^{\dag}$}.
\begin{equation} \label{P1:def}
P_1^{\lambda}(\mbold{k})\ \equiv\ k^0 \frac{d N^{\lambda}}{d^3\mbold{k}}
= Tr\{ \rho_f c_{\lambda}^{\dag}(\mbold{k})
c_{\lambda}(\mbold{k})\}
= Tr\{ \rho_i S^{\dag} c_{\lambda}^{\dag}(\mbold{k})
c_{\lambda}(\mbold{k}) S \}
\end{equation}
The same consideration as above leads to the expression for the number
of photon pairs
\ba \label{P2:def}
P_2^{\lambda_1\;\lambda_2}(\mbold{k}_1,\mbold{k}_2)\
&\equiv&\ k^0_1k^0_2 \frac{d^2 N^{\lambda_1 \lambda_2}}%
{d^3 \mbold{k}_1 d^3 \mbold{k}_2}
= Tr\{\rho_f c_{\lambda_2}^{\dag}(\mbold{k}_2)
c_{\lambda_1}^{\dag}(\mbold{k}_1)
c_{\lambda_1}(\mbold{k}_1)
c_{\lambda_2}(\mbold{k}_2)\}\ = \nonumber\\
&=&\ Tr\{\rho_iS^{\dag} c_{\lambda_2}^{\dag}(\mbold{k}_2)
c_{\lambda_1}^{\dag}(\mbold{k}_1)
c_{\lambda_1}(\mbold{k}_1)
c_{\lambda_2}(\mbold{k}_2) S\}\;\ . \quad
\end{eqnarray}
If we separate the interaction part of the Lagrangian into
$L^{int}_s(x)$ for strong interactions and
$J^{\mu}(x)A_{\mu}(x)$ for the coupling to the electromagnetic field the
\mbox{$S$--matix} can be written in the interaction picture by the
time--ordered exponential \cite{BjorkenDrell65}
\begin{equation} \label {S:def}
S\ =\ {\cal T}\exp\left\{i\int d^4x
\Big( L^{int}_s(x) + J^{\mu}(x)A_{\mu}(x)\Big) \right\}\;\ .\quad
\end{equation}
To proceed further we assume \cite{Machl89} that the initial state does
not contain hard photons, which means
\begin{equation} \label{rhoi:def}
c_{\lambda}(\mbold{k})\rho_i\ =\
\rho_i c^{\dag}_{\lambda}(\mbold{k}) \ =\ 0 \;\ .\quad
\end{equation}
The condition (\ref{rhoi:def}) together with the identities
\ba
\Big[ c_{\lambda}(\mbold{k})\,,\, S \Big] \ &=&\
i {\cal T} \Big(J_{\lambda}(k) S \Big) \quad,
\label{commut:cS} \\\relax
\Big[ S^{\dag}\,,\, c^{\dag}_{\lambda}(\mbold{k}) \Big] \ &=&\
-i \tilde{\cal T} \Big(S^{\dag} J^{\,\,\dag}_{\lambda}(k) \Big)
\quad, \nonumber
\end{eqnarray}
where $k^{\mu}=(|\mbold{k}|,\mbold{k})$ and
$J_{\lambda}(k)$ is the Fourier transform of the transverse current
operator given by
\ba
J_{\lambda}(k)\ &\equiv&\
\epsilon^{\!\!^{\scriptstyle *}\lambda}_{\mu}(k)
\!\int\! d^4\!x\,{\rm e}^{ixk} J^{\mu}(x)
\label{J_lambda:def} \\
J^{\,\,\dag}_{\lambda}(k)\ &\equiv&\
\epsilon^{\lambda}_{\mu}(k)
\!\int\! d^4\!x\,{\rm e}^{-ixk} J^{\mu}(x) \quad, \nonumber
\end{eqnarray}
allow us to write one-- and two--body densities via the chronological
(${\cal T}$) and antichronological ($\tilde{\cal T}$) products of the
current operators:
\ba
P_1^{\lambda}(\mbold{k})\ &=&\ Tr \bigg\{ \rho_i %
\tilde{\cal T} \Big( S^{\dag} J^{\,\,\dag}_{\lambda}(k) \Big)
{\cal T}\Big( J_{\lambda}(k) S \Big) \bigg\}
\label{P1:S} \\
P_2^{\lambda_1 \lambda_2}(\mbold{k}_1, \mbold{k}_2)\ &=&\
Tr \bigg\{ \rho_i \tilde{\cal T} \Big( S^{\dag}%
J^{\,\,\dag}_{\lambda_2}(k_2) J^{\,\,\dag}_{\lambda_1}(k_1) \Big)
{\cal T} \Big(
J_{\lambda_1}(k_1) J_{\lambda_2}(k_2) S \Big)\bigg\} \quad .
\label{P2:S}
\end{eqnarray}
All currents in the expressions written above are operators in the {\sl
interaction} representation. Utilizing properties of time
ordering and of the $S$--matrix we can write the one-- and two--photon
spectra in terms of the current operators in the {\sl
Heisenberg} representation ($\hat{J})$ \cite{RazWein95} as
\ba
P_1^{\lambda}(\mbold{k})\ &=&\ Tr \bigg\{ \rho_i %
\hat{J}^{\,\,\dag}_{\lambda}(k) \hat{J}_{\lambda}(k) \bigg\}
\label{P1:H} \\
P_2^{\lambda_1 \lambda_2}(\mbold{k}_1, \mbold{k}_2)\ &=&\
Tr \bigg\{ \rho_i \tilde{\cal T} \Big( %
\hat{J}^{\,\,\dag}_{\lambda_2}(k_2)
\hat{J}^{\,\,\dag}_{\lambda_1}(k_1) \Big)
{\cal T} \Big(
\hat{J}_{\lambda_1}(k_1) \hat{J}_{\lambda_2}(k_2) \Big)\bigg\}
\quad. \label{P2:H}
\end{eqnarray}
Up to now we have made no approximation to tackle the unsolvable
many--body problem. Expressions (\ref{P1:H}) and (\ref{P2:H}) are still
exact but they are recast into a form which is better suited for
approximations than the original definitions (\ref{P1:def}) and
(\ref{P2:def}). Formally they contain only operators of the strongly
interacting system.
In the following we apply eqs. (\ref{P1:H}) and (\ref{P2:H}) to a highly
excited system which due to short--ranged hard collisions of the charged
constituents is radiating energetic photons. The collisions cause
locally rapid changes in the electric charge current density
$J^{\mu}(x)\,$, which result in non--zero Fourier components in
$J^{\mu}(k)\,$ for large $|\mbold{k}|\,$. The energetic photons are
weakly interacting so that, once they are created, they are assumed to
leave the hadron system without further interactions. Therefore, for
calculating the final photon distributions one needs a local production
rate which is then integrated over space and time. Eq. (\ref{P1:H}) is
of this type. In coordinate space it reads
\begin{equation} \label{P1:xdx}
P^{\lambda}_1(\mbold{k})=
\epsilon^{\!\!^{\scriptstyle *}\lambda}_{\mu_1}(k)
\epsilon^{\lambda}_{\mu_2}(k)\!\!
\int\!\!d^4\!\bar{x}\!\!\int\!\!d^4\!\Delta{}x\, {\rm e}^{-ik\Delta x}
\,Tr\Big\{\rho_i \hat{J}^{\mu_1}(\bar{x}+\frac{\Delta x}{2})
\hat{J}^{\mu_2}(\bar{x}-\frac{\Delta x}{2}) \Big\}
\end{equation}
where the local rate is given by the integral over $\Delta x\,$ of the
current--current correlator.
The following approximations, which will lead to an expression for the
correlator adapted for usage in numerical simulations for heavy--ion
collisions, are based upon three physical assumptions:
\begin{LRdescription}{$(iii)$\hbox{\,\ }}
\item[\hfill$(i)$\hbox{\,\ }] The hadronic correlations in the
system are of short range in space--time. The typical correlation length
is about the mean free path $\lambda_s$ of the strongly interacting
particles. Therefore,
\benn <\!A(x) B(y)\!> \approx <\!A(x)\!> <\!B(y)\!> \qquad
\mbox{\rm if}\;\ |\mbold{x} - \mbold{y}| \;\gap\; \lambda_s \;\
\mbox{\rm or}\;\ |x^0 - y^0| \;\gap\; \lambda_s\ .
\eenn
\item[\hfill$(ii)$\hbox{\,\ }] The hadronic mean free path
$\lambda_s$ is much less than the characteristic size of variations of
the macroscopic variables in space and time, denoted by $L\,$.
\item[\hfill$(iii)$\hbox{\,\ }] Only hard photons with high momenta
$|\mbold{k}|$ are considered
$\,(\,|\mbold{k}| \lambda_s \;\gap\; 1\,)\,$, for which
the slowly varying collective current
$<\!\!\hat{J}^{\mu}(x)\!\!>$ (proportional to the collective 4--velocity
field $u^{\mu}(x)\,$) does not contribute, i.e.
$<\!\!\hat{J}^{\mu}(k)\!\!> \approx 0 $ for high enough $|\mbold{k}|\,$.
\end{LRdescription}
For instance, applying these assumptions to (\ref{P1:xdx}) means that
the integral in $\Delta x$ effectively spreads over the 4--volume of the
order $\lambda_s^4$ and the single inclusive spectrum $P_1(\mbold{k})\,$
is therefore proportional to $L^4 \lambda^4_s\;$.
Inside the trace on the right--hand side of (\ref{P1:xdx}) the mean time
$\bar{x}^0\,$ can be moved over to the statistical operator such that
\begin{equation} \label{JJ:trick}
Tr\Big\{\rho_i \hat{J}^{\mu_1}(\bar{x}+\frac{\Delta x}{2})
\hat{J}^{\mu_2}(\bar{x}-\frac{\Delta x}{2}) \Big\} =
Tr\Big\{\rho(\bar{x}^0) \hat{J}^{\mu_1}(\tilde{x}+\frac{\Delta x}{2})
\hat{J}^{\mu_2}(\tilde{x}-\frac{\Delta x}{2}) \Big\}\ ,\qquad
\end{equation}
where $\tilde{x}=(0,\mbold{x})$ and
$\rho(\bar{x}^0)={\rm e}^{-iH\bar{x}^0}
\rho_i {\rm e}^{iH\bar{x}^0}$
is the solution of the Liouville equation for the density matrix. For
the case of local equilibrium, where the time--dependence enters the
density matrix $\,\rho(\bar{x}^0)\,$
only through the thermodynamic quantities, like for
example temperature $T(\bar{x}^0,\mbold{x})\,$, 4--velocity
$\,u^{\mu}(\bar{x}^0,\mbold{x})\,$ etc., the identity (\ref{JJ:trick})
is a starting point to express the current--current correlator in terms
of thermodynamic variables.
The double inclusive spectrum is given by the four--point correlator
(\ref{P2:H}) and like any other Green function can be written in terms
of connected parts as
\begin{samepage}
\ba
P_2^{\lambda_1 \lambda_2}(\mbold{k}_1, \mbold{k}_2) =&&
P_1^{\lambda_1}(\mbold{k}_1) P_1^{\lambda_2}(\mbold{k}_2)
+ |\!\!<\!\!\hat{J}^{\dag}_{\lambda_1}(k_1)
\hat{J}_{\lambda_2}(k_2)\!\!>\!\!|^2
+ |\!\!<\!\!{\cal T} \Big( \hat{J}_{\lambda_1}(k_1)
\hat{J}_{\lambda_2}(k_2) \Big)\!\!>\!\!|^2 \qquad
\nonumber \\
&& \;+\; \ll\!\! \tilde{\cal T} \Big(
\hat{J}^{\,\,\dag}_{\lambda_2}(k_2)
\hat{J}^{\,\,\dag}_{\lambda_1}(k_1) \Big)
{\cal T} \Big( \hat{J}_{\lambda_1}(k_1)
\hat{J}_{\lambda_2}(k_2) \Big) \!\!\gg\ .\qquad
\label{P2:JJ}
\end{eqnarray}
\end{samepage}
All terms in (\ref{P2:JJ}) involve four space--time integrations. In
coordinate space a connected correlator disappears if the distance
between any pair of its space--time points exceeds the correlation
length. The connected four--point correlator $\,\ll\!\!\cdots\!\!\gg\,$
(last term on the right--hand side of (\ref{P2:JJ})\,) is only non--zero
if the relative distance of all its space--time points are within
$\,\lambda_s\,$. Therefore, this term is proportional to $\,L^4
(\lambda^4_s)^3\,$ and hence, due to condition $\,(ii)\,$, is smaller by
factor of $\,(\lambda_s/L)^4\,$ when compared to all other terms which are
proportional to $\,(L^4)^2(\lambda^4_s)^2\,$.
The third term on the r.h.s of (\ref{P2:JJ}),
$\,|\!\!<\!\!{\cal T}
\Big( \hat{J}_{\lambda_1}(k_1)
\hat{J}_{\lambda_2}(k_2) \Big)\!\!>\!\!|^2\,$,
contains an integral $\,\int\!d^4\!\bar{x}\,
\exp(i(|\mbold{k}_1|+|\mbold{k}_2|)\bar{x}^0)\,\cdots \;$
in which the energies of the two photons add up in the phase
factor. This implies that only momenta which fulfill the condition
$\,(|\mbold{k}_1|+|\mbold{k}_2|)\cdot L \;\lap\; 1 \,$ contribute
appreciably to this term.
These momenta are much smaller than required by
conditions $\,(ii)\,$ and $\,(iii)\,$ and hence we shall drop this term.
After these approximations based on the statistical properties of the
emitting source the one-- and two--photon spectra (\ref{P1:H}) and
(\ref{P2:H}) are given by the current--current correlator
$\,<\!\!\hat{J}^{\mu_1 \dag}(k_1)\hat{J}^{\mu_2}(k_2)\!\!>\,$ only.
Because the electrical current is conserved
$(\partial_{\mu}\hat{J}^\mu(x)=0)$ the correlator
must be transverse ``from both sides'':
\begin{equation} \label{kJJ=0}
k_{1 \mu_1}<\!\!\hat{J}^{\mu_1 \dag}(k_1)\hat{J}^{\mu_2}(k_2)\!\!> =
<\!\!\hat{J}^{\mu_1 \dag}(k_1)\hat{J}^{\mu_2}(k_2)\!\!>k_{2 \mu_2} = 0
\;\ .\qquad\end{equation}
In analogy to (\ref{P1:xdx}) the correlator can be written as
\ba
<\!\!\hat{J}^{\mu_1 \dag}(k_1)\hat{J}^{\mu_2}(k_2)\!\!> =
\sum\limits_{n}&&{\rm e}^{-i(k_1-k_2)\bar{x}_n}\!\!
\int\limits_{\Omega_n}\!\!d^4\!\bar{x}
{\rm e}^{-i(k_1-k_2)(\bar{x}-\bar{x}_n)}\!\!\int\!\!d^4\!\Delta x
{\rm e}^{-i\Delta x (k_1+k_2)/2}\,
\hbox{\qquad}\nonumber\\
\times&&Tr\Big\{\rho_i \hat{J}^{\mu_1}(\bar{x}+\frac{\Delta x}{2})
\hat{J}^{\mu_2}(\bar{x}-\frac{\Delta x}{2})\Big\}\;\ ,
\qquad \label{JJ:xdx}
\end{eqnarray}
where the whole space--time region occupied by the source is divided
into cells with 4--volume $\,\Omega_n\,$ located at mean--coordinates
$\,\bar{x}_n\,$.
Under the assumptions $\,(i)-(iii)\,$ the size of the space-time cells
should be chosen to be about $\lambda_s$ so that the radiation of hard
photons from different cells can be considered as entirely
independent processes. This implies that the currents are conserved for
each cell separately and that the contribution of the cells to the total
current correlator is additive. Therefore, we parametrize (\ref{JJ:xdx})
as
\begin{equation} \label{JJ:W}
<\!\!\hat{J}^{\mu_1 \dag}(k_1)\hat{J}^{\mu_2}(k_2)\!\!>\
\stackrel{ansatz}{=\!=\!=} \
\sum\limits_{n} \Omega_n{\rm e}^{-i\bar{x}_n(k_1 - k_2)}
Q^{\mu_1\mu_2}(k_1,k_2|\bar{x}_n) w(k_1,k_2|\bar{x}_n)\;\ ,\quad
\end{equation}
where the tensor $Q^{\mu_1 \mu_2}$ carries the tensorial structure
of the current--current correlator and the function $w$ describes the
strength.
The $Q$--tensor is explicitly transverse
\begin{equation} \label{kQ=0}
\,k_{1\mu_1}Q^{\mu_1\mu_2}(k_1,k_2|x) =
Q^{\mu_1\mu_2}(k_1,k_2|x)k_{2\mu_2}\nolinebreak=\nolinebreak{}0\,
\end{equation}
and, therefore, ensures current conservations for every cell.
The tensor $Q^{\mu_1\mu_2}$ and the function $w$
are hermitian in the sense that
$Q^{\mu_1\mu_2}(k_1,k_2|x)^{\mbox{\boldmath$*$}}
=Q^{\mu_2\mu_1}(k_2,k_1|x)$
and $w(k_1,k_2|x)^{\mbox{\boldmath$*$}}
=w(k_2,k_1|x)$. It is convenient to normalize the $Q$--tensor by
$\,Q^{\mu}_{\mu}(k,k|x) \equiv -2\,$.
From now on we will write an integral instead of the sum in
(\ref{JJ:W}).
Altogether single and double inclusive photon spectra (no polarization
measured) can be expressed with the help of (\ref{JJ:W}) as follows%
\footnote{We take the sum over photon polarisations by means of
\benn
\!\sum\limits^2_{(\lambda=1)}\!%
\epsilon^{\!\!^{\scriptstyle *}\lambda}_{\mu}(k)%
\epsilon^{\lambda}_{\nu}(k)
= - g_{\mu\nu} - k_{\mu}k_{\nu}/(sk)^2
+ (k_{\mu}s_{\nu} + s_{\mu}k_{\nu})/(sk)
\eenn
($\,s^{\mu}\,$ is some reference 4-vector $\,s^2=1\,$),
where due to the transversality condition (\ref{kQ=0}) only the first term
on the right--hand side $\,(-g_{\mu\nu})\,$ really contributes to
(\ref{P1:W}), (\ref{P2:W}).
}:
\ba
P_1(\mbold{k})\ &\equiv\ & \!\sum^2_{(\lambda=1)} P_1^{\lambda}(\mbold{k})
= 2 \!\int\! d^4\!x\, w(k,k|x)
\label{P1:W} \\
P_2(\mbold{k}_1,\mbold{k}_2) &\equiv&
\!\sum^2_{(\lambda_1,\lambda_2=1)}\!\!
P_2^{\lambda_1 \lambda_2}(\mbold{k}_1,\mbold{k}_2) =
P_1(\mbold{k}_1) P_1(\mbold{k}_2)
\label{P2:W} \\
&+& \!\int\!d^4\!x d^4\!y \,
{\rm e}^{-i(x-y)(k_1-k_2)}
w(k_1,k_2|x) w(k_2,k_1|y)
Q^{\mu_1\mu_2}(k_1,k_2|x)
Q_{\mu_2\mu_1}(k_2,k_1|y)
\quad \nonumber
\end{eqnarray}
To find an explicit form of the $Q$--tensor we assume that the radiating
system is in local equilibrium, that is, the radiation from an elementary
cell is isotropic in its local rest frame and therefore
$Q^{i_1i_2}=\delta^{i_1i_2}$ (where $\,i_1,i_2=1,2,3$).
The other components of the $Q$--tensor are then uniquely determined by the
transversality condition (\ref{kQ=0}) as
$\,Q^{00}=\mbold{k}_1\mbold{k}_2/k^0_1k^0_2\,$,
$\,Q^{i_10}=k^{i_1}_2/k^0_2\,$, $\,Q^{0i_2}=k^{i_2}_1/k^0_1\,$.
In local thermal equilibrium the only other 4--vector entering is the
4--velocity $u^{\mu}(x)$ of the emitting cell which is located at point
$x\,$. Therefore, the only hermitian and transversal $Q$--tensor which
can be constructed from $k^{\mu}_1\,,\; k_2^{\mu}\,\;$ and $u^{\mu}(x)\,$,
and which in the rest frame of a cell (\,$u^{\mu}=(1,\mbox{\bf 0})$\,)
reproduces the isotropic form given above, reads as follows:
\begin{equation} \label{Q:leq}
Q^{\mu_1\mu_2}(k_1,k_2|x) = -g^{\mu_1\mu_2}
- u^{\mu_1}(x)u^{\mu_2}(x)\frac{(k_1k_2)}{(k_1u(x))(k_2u(x))}
+ \frac{u^{\mu_1}(x)k_1^{\mu_2}}{(u(x)k_1)}
+ \frac{k_2^{\mu_1}u^{\mu_2}(x)}{(u(x)k_2)}\
\end{equation}
In eq.~(\ref{Q:leq}) we do not include the term proportional to
$\,k_1^{\mu_1}k_2^{\mu_2}\,$ because it does not contribute to
observables and can be removed by an appropriate gauge transformation.
After defining the $Q$--tensor all dynamic information on the photon
production is contained in the function $w(k_1,k_2|x)$. Inserting
(\ref{Q:leq}) into (\ref{P2:W}) one gets
\begin{equation} \label{P2:leq}
\!\!\!%
P_2(\mbold{k}_1,\mbold{k}_2) \!=\! P_1(\mbold{k}_1) P_1(\mbold{k}_2)
\!+\!\!\int\!d^4\!x d^4\!y \cos(\Delta{}x\Delta{}k)
R(k_1,k_2|x,y) w(k_1,k_2|x) w(k_2,k_1|y)\
\end{equation}
where we introduce the abbreviation
\ba
&&R(k_1,k_2|x,y) \equiv
Q^{\mu_1 \mu_2}(k_1,k_2|x)Q^{\mu_2 \mu_1}(k_2,k_1|y) =
\frac{\big(u(x)k_1\big)\big(u(y)k_2\big) -
2\big(u(x)u(y)\big)\big(k_1k_2\big)}{%
\big(u(x)k_2\big)\big(u(y)k_1\big)}
\qquad \nonumber \\
&&+\ \frac{\big(k_1k_2\big)}{\big(k_1u(x)\big)\big(u(x)k_2\big)} \,+\,
\frac{1}{2}\frac{\big(u(x)u(y)\big)^2\big(k_1k_2\big)^2}{%
\big(k_1u(x)\big)\big(u(x)k_2\big)\big(k_1u(y)\big)\big(u(y)k_2\big)}
\ +\ \Big(x \Leftrightarrow y \Big)\;\ .
\quad \label{R:leq}
\end{eqnarray}
The $R$--function reflects the fact that photons are massless spin--1
particles and in this point expression (\ref{P2:leq}) differs
significantly from the one obtained in \cite{Pratt84} for pions. The
latter has been used for photons of one single polarization in
\cite{SriKap93a,TimPlumRazWein94,SriKap93,SriKap94}. Since the
experimental measurements will involve an averaging over polarizations,
a comparison with data requires the new result (\ref{P2:leq}).
Being very sensitive to the 4--velocity field $\,u^{\mu}(x)\,$ the
$R$--function can change drastically the interference of two photons for
a relativistically expanding source. But even for a static source the
polarization properties of photons are very important.
To illustrate this let us consider two photons with momentum
$\,\mbold{k}_1\,$ and $\,\mbold{k}_2\,$ which are emitted from two cells
having the same 4--velocity $\,u(x)=u(y)=(1,\mbold{0})\,$. In that
case \cite{Neuh86} the $R$--function reduces to
\begin{equation}
R(k_1,k_2|x,y) = 1 + (\cos\theta)^2 =
\sum\limits^2_{(\lambda_1,\lambda_2 = 1)}\Big(
\mbold{\epsilon}^{\lambda_1}(k_1)
\mbold{\epsilon}^{\lambda_2}(k_2)
\Big)^2\;\ ,
\qquad \label{R:ux=uy}
\end{equation}
where
$\cos\theta=\mbold{k}_1\mbold{k}_2/(|\mbold{k}_1||\mbold{k}_2|)\,$.
The physical reason is easily understood for this special case when one
considers the last part of equation (\ref{R:ux=uy}). The summation over
polarizations (here we choose the Coulomb gauge:
$\,\epsilon^{\lambda}(k)=(0,\mbold{\epsilon}^{\lambda}(k))\,$ and
$\,\mbold{k}\mbold{\epsilon}^{\lambda}(k)=0\,$) does not just lead to a
factor of two as implicitly assumed in
\cite{SriKap93a,TimPlumRazWein94,SriKap93,SriKap94}.
As illustrated by \figurename~\ref{fig:pol} only one direction of the
linear polarization can be chosen equal for both photons, whereas the
other polarization directions differ by the angle $\,\theta\,$ between
the two momenta. Therefore, the polarization overlap involves
$\,(\cos\theta)^2\,$ and is less than 2 for $\,\theta \neq 0\,$ (see
(\ref{R:ux=uy})). In a realistic situation the radiation comes of
course from many cells with different four--velocities and the full
structure of the $Q$--tensor has to be employed. Anyhow, for a
non--vanishing angle between the observed photon momenta the correlation
is reduced compared to the ``pion inspired'' recipe
(\ref{P2:wrong}).
\begin{figure}
\begin{center}
\centerline{\epsfig{file=f_pol.ps,%
bbllx=70bp,bblly=580bp,bburx=465bp,bbury=766bp,clip=}}
\vskip 3mm
\begin{minipage}{\LRcaptionwidth}
{\refstepcounter{figure}\label{fig:pol}}
\sbox{\tmpbox}{\bf\figurename~\thefigure:\hbox{\ }}
\begin{LRdescription}{\usebox{\tmpbox}}
\item \footnotesize
The linear polarization vectors $\mbold{\epsilon}^{\lambda}(k)$
for two photons with momenta
$\mbold{k}_1\,$ and $\mbold{k}_2\,$.
\end{LRdescription}
\end{minipage}
\end{center}
\end{figure}
For application in dynamical models we recommend to use the following
formula:
\begin{equation} \label{P2:JmuJnu}
P_2(\mbold{k}_1, \mbold{k}_2) =
P_1(\mbold{k}_1) P_1(\mbold{k}_2)
+ \!<\!\!\hat{J}^{\mu_1 \dag}(k_1) \hat{J}^{\mu_2}(k_2)\!\!>
<\!\!\hat{J}^{\dag}_{\mu_2}(k_2)
\hat{J}_{\mu_1}(k_1)\!\!>\!
\end{equation}
and calculate the current--current correlator with (\ref{JJ:W}) keeping
the explicit form of the $Q$--tensor as given in (\ref{Q:leq}). This is
equivalent to (\ref{P2:leq}) but involves only one $\,d^4\!x\,$
integration instead of $\,d^4\!x\,d^4\!y\,$ in (\ref{P2:leq}).
The dependence of $\,w(k_1,k_2|x)\,$ on $\,\Delta k = k_1 -k_2\,$ and
$\,\bar{k}=(k_1 + k_2)/2\,$ has to be derived from a microscopic model
for the photon production. The only general statement which can be made
at this point is that due to condition $\,(i)\,$ the current--current
correlator decays in the local restframe as function of $\,\Delta k\,$
on the long scale $\,\lambda_s^{-1}\,$. In this paper we do not
consider a specific microscopic model for $\,w(k_1,k_2|x)\,$ but rather
investigate the role of the tensorial structure of the correlator on
photon--photon interferometry. Therefore we write
$\,w(k_1,k_2|x) = w(\bar{k},\bar{k}|x) +
{\cal O}(\Delta k^{\mu}\Delta k^{\nu})\;$. For simplicity we drop terms
of second order in $\,\Delta k\,$.
The strength of photon production $\,w(\bar{k},\bar{k}|x)\,$ can to some
extend be estimated from the single inclusive cross section
\cite{KapLicSeib92}.
Based on the
discussion above we propose to reexamine the photon intensity
interferometry in the spirit of
\cite{SriKap93a,TimPlumRazWein94,SriKap93,SriKap94} but using
(\ref{P2:JmuJnu}) for the double inclusive cross section.
Unfortunately until now the Bose--Einstein correlations of photons have
been studied using the formula for double inclusive spectra where the
photons are assumed to be scalar massless particles, but normalizing the
correlation function to $3/2$ instead of $2$
\cite{SriKap93,SriKap93a,SriKap94}.
That formula is just expression
(\ref{P2:leq}) with $R(k_1,k_2|x,y) = 2\;$ and reads
\begin{equation} \label{P2:wrong}
P_2^{wrong}(\mbold{k}_1,\mbold{k}_2) = P_1(\mbold{k}_1) P_1(\mbold{k}_2)
+ 2 \!\int\!d^4\!x d^4\!y \, \cos(\Delta{}x\Delta{}k)
w(\bar{k},\bar{k}|x)w(\bar{k},\bar{k}|y)\;\ . \qquad
\end{equation}
In order to demonstrate the size of the effect which arises from the
vector nature of the photon and the conservation of electric charge we
have performed numerical studies of photon intensity interferometry in
Bj\o rken hydrodynamics (equation of state $p = \epsilon/3$) using both
our formula (\ref{P2:JmuJnu}) (or equivalently (\ref{P2:leq})) and the
wrong one (\ref{P2:wrong}). We assume that the photons are produced
from the expanding source in local equilibrium and parameterize the
photon production rate as
$\,w(k_1,k_2|x) \cong w(\bar{k},\bar{k}|x) = N\cdot (T(x))^2
\exp(-(\bar{k}u(x))/T(x))\,$ (for a more precise expression
cf. \cite{KapLicSeib92}). The plot of the correlation function defined
as $C_2(k_1,k_2) \equiv P_2(k_1,k_2)/P_1(k_1)P_1(k_2) $ shows a
significant difference in the results based on the different formulas
(see \figurename~\ref{fig:hyd}). The correct expression gives less
correlations for non--zero relative rapidities which means that the
source size deduced from experimental data is overestimated by the wrong
formula (\ref{P2:wrong}).
\begin{figure
\begin{center}
\centerline{\epsfig{file=f_hyd.ps,%
bbllx=140bp,bblly=390bp,bburx=500bp,bbury=780bp,clip=}}
\vskip 3mm
\begin{minipage}{\LRcaptionwidth}
{\refstepcounter{figure}\label{fig:hyd}}
\sbox{\tmpbox}{\bf\figurename~\thefigure:\hbox{\ }}
\begin{LRdescription}{\usebox{\tmpbox}}
\item \footnotesize
Bose--Einstein correlation plot as a function of rapidity
difference $\Delta y$ at fixed transverse momenta
$\mbox{\bf k}_1^{\perp}=\mbox{\bf k}_2^{\perp}=100\; \mbox{\rm MeV} $
for Bj\o rken hydrodynamics with
$T_i=200\; \mbox{\rm MeV}$, $T_f=140\; \mbox{\rm MeV}$ and initial
proper--time $\tau_i=0.3\; fm/c\;$.\\
Solid line corresponds to the photon--photon correlations
(\ref{P2:JmuJnu}) (or equivalently (\ref{P2:leq})). \\
Dashed line represents the wrong ``pion--inspired''
expression (\ref{P2:wrong}).
\end{LRdescription}
\end{minipage}
\end{center}
\end{figure}
In summary we should like to stress that the derivation of a basic
equation for the double inclusive spectrum is significantly modified by the
fact that photons are massless particles with spin 1 and that they are
produced by a conserved electric current. Under the conditions
$\,(i)-(iii)\,$ which are fulfilled for high--energy photons in
relativistic heavy--ion collisions, the expression (\ref{P2:JmuJnu}) is
a suitable starting point for photon intensity interferometry studies.
We would like to acknowledge the fruitful discussions with G.~Bertsch
and J.~Knoll as well as the comments of D.~Seibert. Especially we are
grateful to R.~Weiner and M.~Pl\"umer for critical remarks.
\newpage
\def\noopsort#1{} \def\SL#1{#1}
|
\section{Introduction}
The decays of $B$ mesons to two charmless hadrons can be described by
a $b \rightarrow u$ tree-level spectator diagram
(Figure~\ref{fig:feynman}a),
or a $b \rightarrow sg$ one-loop ``penguin-diagram''
(Figure~\ref{fig:feynman}b) and to a lesser extent, by the
color-suppressed tree (Figure~\ref{fig:feynman}c)
or CKM-suppressed $b\rightarrow dg$ penguin diagrams.
Although such decays can also include contributions from
$b \rightarrow u$
$W$-exchange (Figure~\ref{fig:feynman}d), annihilation
(Figure~\ref{fig:feynman}e), or vertical $W$~loop
(Figure~\ref{fig:feynman}f) processes,
these contributions are expected to be
negligible in most cases.
Decays such as $B^0 \rightarrow \pi^+\pi^-$
and $B^0 \rightarrow \pi^\pm\rho^\mp$ are expected to be dominated by
the $b \rightarrow u$ spectator transition, and measurements of their
branching fractions could be used to extract a value for $|V_{ub}|$.
The decay mode $B^0 \rightarrow \pi^+\pi^-$ can be used to measure $CP$
violation in the $B$~sector at both asymmetric
$B$~factories~\cite{bfact} and hadron colliders\cite{hadcol}.
Since the $\pi^+\pi^-$ final state is a $CP$ eigenstate, $CP$
violation can arise from interference between the amplitude for
direct decay and the amplitude for the process in which the $B^0$
first mixes into a $\bar{B}^0$ and then decays. Measurement
of the time evolution of the rate asymmetry leads to a measurement of
$\sin 2\alpha$, where $\alpha$ is one of the angles in the unitarity
triangle\cite{unitarity_triangle}. If the
$B^0 \rightarrow \pi^+\pi^-$ decay has a non-negligible contribution
from the $b \rightarrow dg$ penguin,
interference between the spectator and penguin contributions
will contaminate the measurement of $CP$ violation via
mixing~\cite{pollution}, an effect
known as ``penguin pollution.'' If this is the case, the penguin and
spectator effects can be disentangled by also measuring the
isospin-related decays
$B^0\rightarrow \pi^0\pi^0$ and $B^\pm\rightarrow \pi^\pm\pi^0$
\cite{pipicp}.
Alternatively,
SU(3) symmetry can be used to relate $B^0\rightarrow \pi^+\pi^-$ and
$B^0\rightarrow K^+\pi^-$ \cite{silva,oits566}.
Penguin and spectator effects may then
be disentangled \cite{silva}
once the ratio of the two branching fractions
and $\sin 2\beta$\cite{unitarity_triangle} are
measured.
Decays such as $B^0 \rightarrow K^+\pi^-$ and $B^0 \rightarrow K^{*+}\pi^-$
are expected to be dominated by the $b \rightarrow sg$ penguin process,
with a small contribution from a Cabibbo-suppressed $b \rightarrow u$ spectator
process. Interference between the penguin and spectator amplitudes can
give rise to direct $CP$ violation, which will manifest itself
as a rate asymmetry for
decays of $B^0$ and $\bar{B}^0$~mesons, but the presence of
hadronic phases
complicates the extraction of the $CP$ violation parameters.
There has been discussion in recent literature about extracting the
unitarity angles using precise time-integrated measurements
of $B$ decay rates. Gronau,
Rosner, and London have proposed~\cite{kpicp} using isospin relations
and flavor SU(3) symmetry to extract, for example, the unitarity angle
$\gamma$ by measuring the rates of $B^+$ decays to
$K^0\pi^+$, $K^+\pi^0$, and $\pi^+\pi^0$ and their
charge conjugates. More recent publications
\cite{deshpande,GRLSU3,EWPrebuttal,EWPmore} have questioned whether
electroweak penguin contributions
($b \rightarrow s\gamma$, $b\rightarrow sZ$)
are large enough to invalidate isospin relationships
and whether SU(3) symmetry-breaking effects can be
taken into account.
If it is possible to extract unitarity angles from rate measurements alone,
the measurements could be made at either symmetric or
asymmetric $B$~factories (CESR, KEK, SLAC),
but will
require excellent particle identification to distinguish between the
$K\pi$ and $\pi\pi$ modes.
Decays such as $B \rightarrow K\phi$ and $B^+ \rightarrow K^0\pi^+$
cannot occur via a spectator process and are expected to be
dominated by the penguin process. Measurement of these decays will
give direct information on the strength of the penguin amplitude.
Various extensions or alternatives to the Standard Model have been suggested.
Such models characteristically involve hypothetical high mass particles,
such as fourth generation quarks, leptoquarks, squarks,
gluinos, charged Higgs, charginos, right-handed $W$'s, and so on.
They have negligible effect on tree diagram dominated $B$ decays, such as
those involving $b\rightarrow cW^-$ and $b\rightarrow uW^-$, but can
contribute significantly to loop processes like $b\rightarrow sg$ and
$b\rightarrow dg$.
Since non-standard models can have
enhanced $CP$ violating effects relative to predictions based on
the standard Kobayashi-Maskawa mechanism \cite{RKM,RNirQuinn},
such effects
might turn out to
be the key to the solution of the baryogenesis problem, that is,
the obvious asymmetry in the abundance of baryons over antibaryons
in the universe.
Many theorists believe
that the KM mechanism
for $CP$ violation is not sufficient to generate the observed
asymmetry or even to maintain an initial asymmetry through
cool-down \cite{RcosmicCP}.
Loop processes
in $B$ decay may be our most sensitive probe of physics
beyond the Standard Model.
This paper reports results on the decays $B \rightarrow \pi\pi$,
$B \rightarrow K \pi$, $B\rightarrow KK$, $B\rightarrow
\pi\rho$, $B\rightarrow K \rho$,
$B\rightarrow K^*\pi$, $B \rightarrow K\phi$, $B \rightarrow K^*\phi$,
and $B \rightarrow \phi \phi$ \cite{chargeconjugation}.
Recent observations of the sum of the two-body charmless hadronic
decays $B^0\rightarrow \pi^+\pi^- {\rm and}\ K^+\pi^-$~\cite{kpi} and
of the electromagnetic penguin decay $B \rightarrow K^*
\gamma$~\cite{kstargamma}, indicate that we have reached the
sensitivity required to observe such decays.
The size of the data
set and efficiency of the
CLEO detector allow us to
place upper limits on the branching fractions in
the range $10^{-4}$ to $10^{-6}$.
\section{Data sample and event selection}
The data set used in this analysis was collected with the CLEO-II
detector~\cite{detector} at the Cornell Electron Storage Ring (CESR).
It consists of $2.42~{\rm fb}^{-1}$ taken at the $\Upsilon$(4S)
(on-resonance) and $1.17~{\rm fb}^{-1}$ taken at a center of mass energy
about 35~MeV below $B\bar{B}$ threshold. The
on-resonance sample contains 2.6~million $B\bar{B}$ pairs.
The below-threshold sample
is used for continuum background estimates.
The momenta of charged particles are measured in a
tracking system consisting of a 6-layer straw
tube chamber, a 10-layer precision
drift chamber, and a 51-layer main drift chamber, all operating
inside a 1.5 T superconducting solenoid. The main drift chamber
also provides a measurement of the specific ionization loss, $dE/dx$,
used for particle identification.
Photons are detected
using 7800 CsI crystals, which are also inside the magnet. Muons are
identified using proportional counters placed at various depths in the
steel return yoke of the magnet.
The excellent efficiency and resolution of the CLEO-II detector for
both charged particles and photons are crucial in
extracting signals and suppressing both continuum
and combinatoric backgrounds.
Charged tracks are
required to pass track quality cuts based on the average hit residual
and the impact parameters in both the $r-\phi$ and $r-z$ planes.
We require that charged track momenta be greater than 175~MeV/$c$ to
reduce low momentum combinatoric background.
Pairs of tracks with vertices displaced from the primary interaction
point are taken as $K_S^0$ candidates. The secondary vertex is
required to be displaced from the primary interaction point by at
least 1~mm for candidates with momenta less than 1~GeV/$c$ and
at least 3~mm for candidates with momenta greater than
1~GeV/$c$. We make a momentum-dependent cut on
the $\pi^+\pi^-$ invariant mass.
Isolated showers with energies greater than
$30$~MeV in the central region of the CsI detector, $|\cos\theta| < 0.71$,
where $\theta$ is the angle with respect to the beam axis,
and greater than $50$~MeV elsewhere, are defined to be photons.
Pairs of photons with an invariant mass within two standard
deviations of the nominal $\pi^0$ mass \cite{pdb94}
are kinematically fitted with the mass constrained to the
$\pi^0$ mass. To reduce combinatoric backgrounds we
require that the $\pi^0$ momentum be greater than $175$~MeV/$c$, that the
lateral shapes of the showers be consistent with those from photons, and that
$|\cos\theta^*|<0.97$, where $\theta^*$ is the angle between the
direction of flight of the $\pi^0$ and the photons in the $\pi^0$ rest
frame.
We form $\rho$ candidates from
$\pi^+\pi^-$ or $\pi^+\pi^0$ pairs with an invariant mass within $150$~MeV
of the nominal $\rho$ masses.
$K^*$ candidates are selected from $K^+ \pi^-$, $K^+\pi^0$,
$K_S^0\pi^+$ or $K_S^0\pi^0$ pairs \cite{kstars} with
an invariant mass within $75$~MeV of the nominal
$K^*$ masses. We form $\phi$ candidates from $K^+K^-$ pairs with
invariant mass within $6.5$~MeV of the nominal $\phi$ mass.
Charged particles are identified
as kaons or pions according to $dE/dx$.
We first reject electrons
based on $dE/dx$ and the ratio of the track momentum to the associated shower
energy in the CsI calorimeter.
We reject muons
by requiring that the tracks not penetrate the steel absorber to a
depth of five nuclear interaction lengths.
We define $S$ for a particular hadron hypothesis as
\begin{equation}
S_{\rm hypothesis} =
{{dE \over dx}|_{\rm measured} - {dE \over dx}|_{\rm hypothesis} \over \sigma}
\end{equation}
where $\sigma$ is the expected resolution, which depends
primarily on the number of
hits used in the $dE/dx$ measurement.
We measure the $S$ distribution in data for kaons and pions using
$D^0\rightarrow K^-\pi^+$ decays where the $D^0$ flavor is tagged
using $D^{*+}\rightarrow D^0\pi^+$ decays. In particular, we are
interested in separating pions and kaons with momenta near 2.6 GeV/$c$.
The $S_\pi$ distribution for
the pion hypothesis is shown in Figure~\ref{fig:dedx} for pions and
kaons with momenta
between 2.3 and 3.0~GeV/$c$. At these momenta, pions and kaons are
separated by $1.8\pm0.1$ in $S_\pi$.
\section{Candidate selection}
\subsection{Energy Constraint}
Since the $B$'s are
produced via $e^+e^-\rightarrow
\Upsilon (4S)\rightarrow B\bar B$, where the $\Upsilon (4S)$ is at rest in
the lab frame, the energy of either of the two $B$'s is given by the
beam energy, $E_{\rm b}$. We define
$\Delta E = E_1 + E_2 - E_{\rm b}$ where $E_1$ and $E_2$
are the energies of the daughters of the $B$ meson candidate.
The $\Delta E$ distribution for signal peaks at $\Delta E =0$,
while the background distribution falls linearly
in $\Delta E$ over the region of interest.
The resolution of $\Delta E$ is mode dependent
and in some cases helicity angle dependent (see section III.C)
because of the difference in energy resolution between
neutral and charged pions.
For modes including high momentum
neutral pions in the final state, the $\Delta E$
resolution tends to be asymmetric because of energy loss out of the
back of the CsI crystals.
The $\Delta E$ resolutions for the modes in this paper,
obtained from Monte Carlo simulation, are listed in Tables~\ref{tab:deltae}
and \ref{tab:mlfits}.
We check that the Monte Carlo accurately reproduces the data in
two ways. First,
the r.m.s.\ $\Delta E$ resolution for $B^0\rightarrow h^+h^-$
(where $h^\pm$ indicates a $\pi^\pm$ or $K^\pm$) is given by
$\sigma_{{\Delta E}_{h^+h^-}} = \sqrt{2}\sigma_p$ where $\sigma_p$ is
the r.m.s.\ momentum resolution at $p=2.6$\ GeV$/c$.
We measure the momentum resolution at $p=5.3$ GeV$/c$ using muon
pairs and in the range $p=1.5$--2.5 GeV$/c$ using the modes
$B \rightarrow \psi K$,
$B\rightarrow D\pi$, and $B \rightarrow D^*\pi$.
We find
$\sigma_{\Delta E_{h^+h^-}} = 24.7\pm 2.3 ^{+1.4}_{-0.7}$ MeV,
where the first error is statistical and the second is
systematic. This result is in good agreement with the Monte Carlo prediction.
We also test our Monte Carlo simulation
in the modes $B^+ \rightarrow \bar D^0\pi^+$ and $B^0\rightarrow D^-\pi^+$
(where
$\bar D^0 \rightarrow K^+\pi^-$, $\bar D^0 \rightarrow K_S^0\pi^0$, and
$D^-\rightarrow K_S^0\pi^-$)
using an analysis similar to our $B \rightarrow K^*\pi$
analysis. Again, $\Delta E$ resolutions for data and Monte Carlo are
in good agreement.
The energy constraint also
helps to distinguish between modes of
the same topology. When a real $K$ is reconstructed as a
$\pi$, $\Delta E$ will peak below zero by an amount dependent on
the particle's momentum. For example, $\Delta E$
for $B\rightarrow K^+\pi^-$, calculated assuming $B\rightarrow\pi^+\pi^-$,
has a distribution which is centered at $-42$~MeV, giving
a separation
of $1.7\sigma$ between $B \rightarrow K^+\pi^-$ and
$B \rightarrow \pi^+\pi^-$.
\subsection{Beam-Constrained Mass}
Since the energy of a $B$ meson is
equal to the beam energy, we use
$E_{\rm b}$\ instead of the reconstructed energy of
the $B$ candidate to calculate the beam-constrained $B$ mass:
$M_B = \sqrt{E_{\rm b}^2 - {\bf p}_B^2}$.
The beam constraint improves the mass resolution by
about an order of magnitude, since $|{\bf p}_B|$ is only $0.3~$GeV/$c$
and the beam energy is known to
much higher precision than the measured energy of the $B$ decay products.
Mass resolutions range from 2.5 to 3.0 MeV, where the larger
resolution corresponds to decay modes with high momentum $\pi^0$'s.
Again, we verify the accuracy of our Monte Carlo by studying
fully reconstructed $B$ decays.
The $M_B$ distribution for continuum background is described by the
empirical shape
\begin{equation}
f(M_B) \propto M_B\sqrt{1-x^2}\exp\left(-\xi(1-x^2)\right)
\end{equation}
where $x$ is defined as $M_B/E_{\rm b}$ and $\xi$ is a parameter to be fit.
As an example,
Figure~\ref{fig:argusfcn} shows the fit for $B \rightarrow h^+\pi^0$
background from
data taken below $B\bar B$ threshold.
\subsection{Helicity Angle}
The decays $B \rightarrow \pi\rho$,
$B \rightarrow K\rho$, $B \rightarrow K^*\pi$,
and $B \rightarrow K\phi$ are of the form
pseudoscalar $\rightarrow$ vector + pseudoscalar. Therefore
we expect the helicity angle, $\theta_H$, between a resonance
daughter direction and the $B$ direction in the resonance rest
frame to have a $\cos^2\theta_H$ distribution. For these decays
we require $|\cos \theta_H| > 0.5$.
\subsection{$D$ Veto}
We suppress
events from the decay $B^+\rightarrow \bar{D}^0 \pi^+$
(where $\bar{D}^0 \rightarrow K^+\pi^-$ or $\bar{D}^0 \rightarrow
K_S^0\pi^0$) or $B^0 \rightarrow D^-\pi^+$ (where $D^-\rightarrow
K_S^0\pi^-$) by rejecting any candidate that can be
interpreted as $B\rightarrow \bar{D} \pi$, with a $K\pi$
invariant mass within $2\sigma$ of the nominal $D$
mass. We expect less than half an event background per mode from
$B\rightarrow \bar D \pi$ events after this veto.
The vetoed $D\pi$ signal is used
as a cross-check of signal distributions and efficiencies.
\section{Background Suppression using Event Shape}
The dominant background in all modes is from continuum production, $e^+e^-
\rightarrow q\bar{q}\ (q=u,d,s,c)$. After the $D$ veto,
background from $b \rightarrow c$
decays is negligible
in all modes because final state particles in such
decays have maximum momenta lower than what is required for the decays of
interest here. We have also studied backgrounds from the
rare processes
$b \rightarrow s \gamma$ and $b \rightarrow u\ell\nu$ and find
these to be negligible as well.
Since the $B$ mesons are approximately at rest in the lab,
the angles of the decay products of the two $B$ decays
are uncorrelated
and the event looks spherical. On the other hand,
hadrons from continuum $q\bar{q}$ production
tend to display a two-jet structure.
This event shape distinction is exploited in two ways.
First, we calculate the angle, $\theta_T$,
between the thrust axis of the $B$ candidate and the thrust
axis of all the remaining charged and neutral particles in the event.
The distribution of $\cos \theta_T$ is strongly peaked
near $\pm1$ for $q\bar q$ events and is nearly flat for $B\bar B$ events.
Figure~\ref{fig:costhrcomp} compares the $\cos \theta_T$ distributions
for Monte Carlo signal events and background data.
We require $|\cos\theta_T| < 0.7$ which
removes more than $90\%$ of the continuum background with
approximately $65\%$ efficiency for signal events \cite{thrustnotflat}.
Second, we characterize the event shape by
dividing the space around the candidate thrust axis into nine
polar angle intervals
of $10^\circ$ each, illustrated in Figure~\ref{fig:vcal};
the $i^{th}$ interval covers angles
with respect to the candidate thrust axis
from $(i-1)\times 10^{\circ}$\ to $i\times 10^{\circ}$. We fold
the event such that the forward and backward intervals are combined.
We then define the momentum flow, $x_i$ ($i=1,9$), into the
$i^{th}$ interval as the scalar sum of the momenta of all charged tracks and
neutral showers pointing in that interval.
The $10^\circ$ binning was chosen to enhance the
distinction between $B \bar B$
and continuum background events.
Angular momentum conservation considerations provide additional distinction
between $B\bar B$ and continuum $q\bar q$ events.
In $q\bar q$ events, the direction of the candidate thrust axis,
$\theta_{q\bar q}$, with respect to the beam axis in the lab frame
tends to maintain the
$1+\cos^2 \theta_{q\bar q}$ distribution of the primary
quarks. The direction of the candidate thrust
axis for $B \bar B$ events is random.
The candidate $B$ direction, $\theta_B$, with respect to
the beam axis exhibits a $\sin^2 \theta_B$ distribution for
$B \bar B$ events and is random for $q \bar q$ events.
A Fisher discriminant\cite{kendall}
is formed from these eleven
variables: the nine
momentum flow variables, $|\cos \theta_{q\bar q}|$,
and $|\cos \theta_B|$.
The discriminant, $\cal F$, is the linear combination
\begin{equation}
{\cal F} = \sum^{11}_{i=1} \alpha_i\ x_i
\label{eqn:xfdef}
\end{equation}
of the input variables, $x_i$, that maximizes the separation between
signal and background.
The Fisher discriminant parameters, $\alpha_i$, are given by
\begin{equation}
\alpha_i = \sum_{j=1}^{11} (U_{ij}^b + U_{ij}^{s})^{-1} \times
(\mu_j^{b} - \mu_j^{s}).
\label{eqn:alphadef}
\end{equation}
where $U_{ij}^s$\ and $U_{ij}^{b}$\ are the covariance matrices of the
input variables for
signal and background events, and $\mu_j^s,\ \mu_j^{b}$\
are the mean values of the input variables.
We calculate $\alpha_i$ using Monte Carlo samples of
signal and background events
in the mode
$B \rightarrow \pi^+\pi^-$.
Figure~\ref{fig:fdstan} shows the $\cal F$ distributions for Monte Carlo
signal in the mode $B^0\rightarrow \pi^+\pi^-$, and data signal in
the modes $B \rightarrow \bar D\pi$. Figure~\ref{fig:fdstan} also shows
the $\cal F$ distributions for Monte Carlo background in the mode
$B\rightarrow h^+\pi^-$ and
below-threshold background
data for modes comprising three charged tracks
or two charged tracks and a $\pi^0$.
The $\cal F$
distribution for signal is fit by a Gaussian distribution,
while the $\cal F$ distribution for background data is best fit
by the sum of two Gaussians with the same mean but different variances and
normalizations.
The separation between signal and background means is approximately 1.3 times
the signal width.
We find that the Fisher coefficients calculated for $B^0 \rightarrow
\pi^+\pi^-$ work equally well for all other decay modes presented in this
paper.
Figure~\ref{fig:fdmiracle} shows the remarkable consistency of the
means and widths of the $\cal F$
distributions for signal and background Monte Carlo for the modes in
this study.
\section{Analysis}
For the decay modes $B \rightarrow \pi\pi$, $B \rightarrow K\pi$,
and $B \rightarrow KK$,
we extract the signal yield using a maximum likelihood fit. For the other
decay modes, we use a simple counting analysis. Both techniques
are described below.
\subsection {Maximum Likelihood Fit}
We perform unbinned
maximum likelihood fits using $\Delta E$, $M_B$, $\cal F$,
and $dE/dx$ (where appropriate) as input information for each candidate
event to determine the signal yields for
$B^0\rightarrow \pi^+\pi^-,\ K^+\pi^-,\ K^+K^-,\
\pi^0\pi^0,\ K^0\pi^0$, and
$B^+\rightarrow \pi^+\pi^0,\ K^+\pi^0,\
K^0\pi^+$.
Five different fits are performed as listed in
Table~\ref{tab:mlfits}.
For each fit a likelihood function $\cal L$ is defined as:
\begin{equation}
{\cal L} =
\prod_{i=1}^{N} P\left( f_1, ..., f_m;
\left(\Delta E, M_B, {\cal F}, dE/dx\right)_i\right)
\label{eqn:likefun}
\end{equation}
where $P\left( f_1, ..., f_m;
\left(\Delta E, M_B, {\cal F}, dE/dx\right)_i\right)$
is the probability density function
evaluated at the measured point
$(\Delta E,\ M_B,\ {\cal F},\ dE/dx)_i$ for a
single candidate event, $i$,
for some assumption of the values
of the yield fractions, $f_j$, that are determined
by the fit. $N$
is the total number of events that are fit.
The fit includes all the candidate events
that pass the selection criteria discussed above as well as
$|\cos\theta_T|<0.7$, and $0<{\cal F}<1$.
The $\Delta E$ and $M_B$ fit ranges are given in Table~\ref{tab:mlfits}.
For the case of $B\rightarrow h^+h^-$, the
probability $P_i = P\left( f_1, ..., f_m;
\left(\Delta E, M_B, {\cal F}, dE/dx\right)_i\right)$
is then defined by:
\begin{eqnarray}
P_i &= & f^S_{\pi\pi} P^S_{\pi\pi} + f^S_{K\pi} P^S_{K\pi} +
f^S_{KK} P^S_{KK}
+ (1- f^S_{\pi\pi} - f^S_{K\pi} - f^S_{KK}) P^C \\
P^C &= & f^C_{\pi\pi} P^C_{\pi\pi} + f^C_{K\pi} P^C_{K\pi}
+ (1- f^C_{\pi\pi} - f^C_{K\pi}) P^C_{KK} \nonumber
\end{eqnarray}
where, for example, $P^S_{\pi\pi}$ ($P^C_{\pi\pi}$)
is the product of the individual
probability density functions for $\Delta E$, $M_B$, $\cal F$, and
$dE/dx$ for $\pi^+\pi^-$
signal (continuum background). The signal yield in
$B^0 \rightarrow \pi^+\pi^-$, for example, is then given by
$N_{\pi\pi} \equiv f_{\pi\pi}^S\times N$.
The central values of the individual signal yields
from the fits are given in
Table~\ref{tab:mlresults}. None of the individual modes shows a
statistically compelling signal. To illustrate the fits,
Figure~\ref{fig:mass_2body} shows $M_B$ projections for events
in a signal region defined by
$|\Delta E| < 2\sigma_{\Delta E}$ and ${\cal F} < 0.5$
and Figure~\ref{fig:de_2body} shows the $\Delta E$ projections
for events within a 2$\sigma$ $M_B$ cut and
${\cal F} < 0.5$.
The modes are sorted by $dE/dx$ according to the most likely
hypothesis and are shown in the plots with different shadings.
Overlaid on these plots are the projections of the
fit function integrated over the remaining variables
within these cuts. (Note that these curves are not fits to these
particular histograms.)
Our previous publication \cite{kpi} reported a significant signal in the
sum of $B^0 \rightarrow \pi^+\pi^-$ and $B^0 \rightarrow K^+\pi^-$.
While our current analysis confirms this result, we now focus on
separating the two modes. We separate
the systematic errors
that affect the total yield from those that affect the
separation of the two modes.
We do this by repeating the likelihood fit using
$N_{\rm sum}\equiv N_{\pi\pi} + N_{K\pi}$,
$R \equiv N_{\pi\pi}/N_{\rm sum}$, and fixing $N_{KK} = 0$, its most
likely value.
We find:
\begin{eqnarray*}
N_{\rm sum}
= && 17.2^{+5.6\ +2.2}_{-4.9\ -2.5} \nonumber \\
R = &&0.54^{+0.19}_{-0.20}\pm 0.05
\end{eqnarray*}
where the first error is statistical and the second is
systematic (described below).
The result of this fit is shown in Figure~\ref{fig:contour}.
This figure shows a contour plot (statistical errors only) of
$N_{\rm sum}$ {\it vs.}\ $R$ in which the solid curves represent the
$n\sigma$ contours ($n=$1--4) corresponding to decreases in the
log likelihood by $0.5n^2$. The dashed curve represents the
$1.28\sigma$ contour, from which estimates of the 90\%
confidence level limits
can be obtained.
The central value of $N_{\rm sum}$ has a statistical significance of
$5.2\sigma$. The significance is reduced to
$4.2\sigma$\ if all parameters defining $\cal L$\
are varied coherently so as to minimize $N_{\rm sum}$.
Further support for the statistical significance of
the result is obtained by using Monte Carlo to draw 10000
sample experiments, each
with the same number of events as in the
data fit region but no signal events.
We then fit each of these sample experiments to determine
$N_{\rm sum}$\ in the same way as done for data.
We find that none of the 10000 sample experiments leads to
$N_{\rm sum}>10$.
None of the physical range of $R$\ can be excluded
at the $3\sigma$\ level. However the systematic error
of $R$ is only 10\% (see below and Table~\ref{tab:mlsyst}).
We therefore conclude that our analysis technique has sufficient
power to distinguish the $\pi^+\pi^-$
mode from $K^+\pi^-$, but at this time
we do not have the statistics to do so.
Since none of our fits has a statistically significant signal, we
calculate the $90\%$ confidence level upper limit yield
from the fit, $N^{90}$, given by
\begin{equation}
{\int_0^{N^{90}} {\cal L}_{\rm max} (N) dN
\over
\int_0^{\infty} {\cal L}_{\rm max} (N) dN}
= 0.90
\label{eqn:upplim}
\end{equation}
where ${\cal L}_{\rm max}(N)$ is
the maximal $\cal L$\ at fixed $N$\ to
conservatively account for possible correlations among
the free parameters in the fit.
The upper limit yield is then increased by the systematic error
determined by varying the parameters defining
$\cal L$\ within their systematic uncertainty as discussed below.
Table~\ref{tab:mlresults} summarizes upper limits on the yields for
the various decay modes.
To determine the systematic effects on the yield due
to uncertainty of the shapes used in the likelihood fits,
we vary the parameters
that define the likelihood functions.
The variations of the yields are given in Table~\ref{tab:mlsyst}.
The largest contribution
to the systematic error arises from the variation of the $M_B$
background shape.
For this
shape, $f(M_B) \propto M_B\sqrt{1-x^2}\exp(-\xi(1-x^2))$
($x \equiv M_B/E_{\rm b}$), we vary $E_{\rm b}$ by $\pm1$ MeV,
consistent with
observed variation; we vary $\xi$ by the amount allowed by a fit
to background data (below-threshold and on-resonance
$\Delta E$ sideband) which pass all other selection criteria. To be
conservative, we allow for correlated variations of $E_{\rm b}$ and
$\xi$.
\subsection {Event-Counting Analyses}
In the event-counting analyses we make cuts on
$\Delta E$, $M_B$, $\cal F$, and $dE/dx$.
The cuts for $\Delta E$ and $M_B$
are mode dependent and are listed in Table~\ref{tab:deltae}.
We require ${\cal F} < 0.5$. Tracks are
identified as kaons and/or pions if their specific ionization loss,
$dE/dx$, is within three standard deviations of the expected value.
For certain topologies, candidates can have multiple
interpretations under different particle hypotheses. In these cases
we use a strict identification scheme where a track is positively
identified as a kaon or a pion depending on which $dE/dx$ hypothesis
is more likely:
we sort the modes with two charged tracks plus a $\pi^0$
($\pi^+\rho^-$, $\pi^0\rho^0$, $K^+\rho^-$,
$K^{*+}\pi^-$, and $K^{*0}\pi^0$)
by requiring strict identification for both charged tracks.
For modes with three charged tracks
($\pi^+\rho^0$, $K^+\rho^0$, and $K^{*0}\pi^+$)
we require strict identification of the two like-sign tracks,
while the unlike-sign track \cite{chargestrangeness} is
required to be consistent with the pion hypothesis within two
standard deviations.
We separate modes with one charged track plus two $\pi^0$'s
($\rho^+\pi^0$ and $K^{*+}\pi^0$)
by requiring strict identification of the charged track.
Figures~\ref{fig:mass_pirho}--\ref{fig:mass_kphi}
show $M_B$ distributions for $B \rightarrow \pi\rho$,
$B \rightarrow K\rho$, $B\rightarrow K^*\pi$, $B\rightarrow K\phi $,
$B\rightarrow K^*\phi$ and $B\rightarrow \phi\phi$
candidates (after making the cuts on $\Delta E$, $\cal F$, and particle
identification
described above.) The numbers of events in the signal regions are
listed in Table~\ref{tab:evcount_results}.
In order to estimate the background in our
signal box, we look in a large
sideband region in the $\Delta E~vs.~M_B$ plane:
$5.20 < M_B < 5.27$ GeV and $| \Delta E | < 200$ MeV.
The expected background in the signal region is obtained by
scaling the number of events seen in the
on-resonance and below-threshold
sideband regions (weighted appropriately for luminosity).
Scale factors are found using a continuum Monte Carlo
sample which is about five times the continuum data on-resonance.
In many modes, the
backgrounds are so low that there are insufficient statistics in the
Monte Carlo to adequately determine a scale factor. For these modes,
we calculate upper limits assuming all observed events are signal
candidates. The estimated background for each mode
is also listed in Table~\ref{tab:evcount_results}.
Although we find that there are slight excesses
above expected background
in some modes, no excess is statistically compelling.
We therefore calculate upper limits on the numbers of
signal events using the procedure outlined in the
Review of Particle Properties~\cite{pdb94} for evaluation of upper
limits in the presence of background.
To account for the uncertainties in the estimated continuum
background we reduce the background estimate by its uncertainty
prior to calculating the upper limit on the signal yield.
\section{Efficiencies}
The reconstruction efficiencies were determined using events generated
with a GEANT-based Monte Carlo simulation program~\cite{geant}.
Systematic uncertainties were determined using data wherever possible.
Some of the largest systematic errors
come from uncertainties in the efficiency of the $|\cos\theta_T| < 0.7$
cut (6\%), the uncertainty in
the $\pi^0$ efficiency (7\% per $\pi^0$), and
the uncertainty in the $K_S^0$ efficiency (8\% per $K_S^0$).
In higher
multiplicity modes, substantial contributions come from the
uncertainty in the tracking efficiency (2\% per track). In the
$B \rightarrow \pi\rho,\ K\rho,\ K^*\pi$ analyses,
the simulation of the efficiency
for the particle identification method has a systematic error of
15\%. For the event-counting analyses, the uncertainty in the
${\cal F} < 0.5$ cut is 5\%.
The total detection efficiency, ${\cal E}$,
is given by ${\cal E} \equiv {\cal E}_r \times {\cal E}_d$,
where ${\cal E}_r$ is the reconstruction efficiency and
${\cal E}_d$ is the product of the appropriate daughter branching
fractions. The efficiencies, with systematic errors, are listed in
Tables~\ref{tab:mlresults} and \ref{tab:eff}.
\section{Upper Limit Branching Fractions}
Upper limits on the branching fractions are given by
$N_{\rm UL}/({\cal E} N_B)$
where $N_{\rm UL}$ is the upper limit on the signal yield,
${\cal E}$ is the total detection efficiency, and $N_B$ is the
number of
$B^0$'s or $B^+$'s produced, 2.6 million, assuming
equal production of charged and
neutral $B$ mesons. To conservatively account for the
systematic uncertainty in our efficiency, we reduce the efficiency by
one standard deviation. The upper limits on the branching fractions
appear in Tables \ref{tab:mlresults} and \ref{tab:evcount_results}.
\section{Summary and Conclusions}
We have searched for rare hadronic $B$ decays in many modes and
find a signal only in the sum of $\pi^+\pi^-$ and $K^+\pi^-$.
The combined branching fraction, ${\cal B}(\pi^+\pi^- + K^+\pi^-)
= (1.8^{+0.6+0.2}_{-0.5-0.3}\pm0.2) \times 10^{-5}$,
is consistent with our previously published result.
We have presented new upper limits on the branching
fractions for a variety of charmless hadronic decays of $B$~mesons
in the range $10^{-4}$ to $10^{-6}$.
These results are significant improvements over those
previously published. Our sensitivity is at the level of Standard
Model
predictions for the modes $\pi^+\pi^-,\ K^+\pi^-,\ \pi^+\pi^0,\
K^+\pi^0,\ \pi^{\pm}\rho^{\mp},\ K^+\phi$,\ and $K^{*0}\phi$.
\acknowledgments
We gratefully acknowledge the effort of the
CESR staff in providing us with
excellent luminosity and running conditions.
J.P.A., J.R.P., and I.P.J.S. thank
the NYI program of the NSF,
G.E. thanks the Heisenberg Foundation,
K.K.G., M.S., H.N.N., T.S., and H.Y. thank the
OJI program of DOE,
J.R.P thanks the A.P. Sloan Foundation,
and A.W. thanks the
Alexander von Humboldt Stiftung
for support.
This work was supported by the National Science Foundation, the
U.S. Department of Energy, and the Natural Sciences and
Engineering Research Council of Canada.
|
\section{INTRODUCTION}
In recent years the Compton Gamma Ray Observatory (CGRO) and other
instruments have provided major new discoveries and detailed observations
of isolated $\gamma$-ray pulsars, including the Crab (Nolan et al. 1993),
Vela (Kanbach et al. 1994), Geminga (Halpern \& Holt 1992, Bertsch et
al. 1992, Mayer-Hasselwander et al. 1994), PSR B1509-58 (Wilson et al. 1992),
PSR B1706-44 (Thompson et al. 1992), PSR B1055-52 (Fierro et al. 1993),
and most recently PSR B1951+32 (Ramanamurthy et al. 1995).
Models of these objects must now account for a variety of detailed features
in the emission, especially from the most intense sources (Crab, Vela,
Geminga). Current models have in fact already encountered problems
in explaining how these sources can show both remarkable similarities
and puzzling variations in their light curves and phase-resolved energy
spectra. These difficulties are even more severe if models of the
$\gamma$- ray emission must also be consistent with the radiation observed
at radio, optical, and X-ray wavelengths. As the observational statistics
for the weaker sources improve, these theoretical challenges may become
even more formidable.
At present two general types of $\gamma$- ray pulsar models are popular
in the literature. The Polar Cap (PC) model, first proposed by Sturrock (1971)
and later investigated by numerous authors (see for example Ruderman
and Sutherland 1975, Harding 1981, Daugherty and Harding 1982, Arons 1983)
assumes that the emission is produced by electrons accelerated to high
energies just above the surface of a magnetized rotating neutron star (NS),
in the vicinity of the magnetic poles. In contrast, the Outer Gap model
(Cheng, Ho, and Ruderman 1986a, 1986b) places the acceleration regions
much higher in the NS magnetosphere, in vacuum gaps formed within
a charge-separated plasma.
In a previous paper (Daugherty and Harding 1994, hereafter DH94)
we proposed a version of the PC model based on the following
principal assumptions:
(a) The gamma emission is initiated by the acceleration of electrons
from the NS surface, just above the magnetic PC regions
which enclose the {\it open\/} magnetic field lines extending to the
velocity-of-light cylinder (LC).
(b) The emission originates as curvature radiation (CR) produced by
the electrons as they follow the curvature of the open magnetic
field lines.
(c) The processes of direct $1-\gamma$ pair conversion (see for example
Erber 1966) by the NS magnetic field and synchrotron radiation (SR) by
the emitted pairs produce photon-pair cascades, from which the observed
$\gamma$ radiation emerges.
(d) The rotational and magnetic axes of the radiating NS are nearly
aligned, so that the inclination $\alpha$ is small enough to be comparable
with the PC half-angle $\theta_{pc}$. More precisely, the model requires
that $\alpha \sim \theta_b$ where $\theta_b$ is the half-angle of the
$\gamma$-beam emerging from the PC.
Assumptions (a)-(c) comprise essentially the original postulates of the
PC model (Sturrock 1971). They describe the overall physics of
the cascade process and in combination they determine the form of the
production spectra for the gamma rays and the pairs. The final assumption
(d) primarily affects the viewing geometry. It implies that randomly
oriented observers should see emission from at most one PC. However,
since CR-induced cascades are intrinsically hollow-cone sources which
produce their most intense
emission near the PC rim, observers viewing a single PC may detect light
curves with either single or double peaks (DH94).
Sterner and Dermer (1994) independently noted a similar effect in a
model of PC cascades initiated by Comptonization rather than CR.
The assumption that $\alpha \sim \theta_b$ allows the phase separation
between double peaks to become large enough to match the observed values
($\sim 0.4$ for the Crab and Vela, $\sim 0.5$ for Geminga).
In the present work we refine assumption (d) by requiring {\it only\/} that
$\alpha \sim \theta_b$, not that $\alpha$ itself be necessarily small.
Hence in place of the Nearly Aligned Rotator (NAR) model described in
DH94, we consider here a more general Single Polar Cap (SPC) model.
In addition, we introduce a further assumption which allows
$\theta_b$ (and $\alpha$) to have significantly larger values
than $\theta_{pc}$ itself:
(e) the acceleration of the electrons occurs over an extended distance
above the PC surface, so that they reach their peak energies at heights
of a few NS radii. Above these heights, the acceleration is
cut off by an overlying force-free plasma.
In DH94 we neglected the height of the acceleration
region and simply supplied the electrons with an injection energy at the NS
surface, then traced their CR energy losses as they escaped outward along
field lines for which ${\bf E \cdot B} \sim 0$. We have since noted that
the assumption
of an extended acceleration region provides a solution to a serious
difficulty with our previous model, namely the ``observability'' problem.
This refers to the fact that if conventional estimates of PC dimensions
are accurate, $\gamma$-beams emitted by energetic electrons just above the NS
surface would be so small that there would be a low probability
($\lesssim 10^{-2}$)
that they could be detected by randomly oriented observers.
In our previous work we noted that the usual estimate for the PC radius
$R_{pc}$ may in fact be too small, although moderate increases in
$R_{pc}$ cannot by themselves resolve the observability problem.
However, the outward flaring of the magnetic field lines implies that
the half-angle $\theta_{b}$ of the (hollow) cascade $\gamma$-beams
increases rapidly with height above the NS surface. Thus the effect
of extending the acceleration zone up to heights of a few NS radii,
especially if combined with moderately increased ($\lesssim 2$)
PC dimensions, can produce rotating beams whose edges sweep
over a much larger solid angle.
In DH94 we also noted that in order to produce double
peaks as narrow as those observed from the Crab, Vela, and Geminga,
we had to assume that the surface density of the electrons drawn from
the NS surface is concentrated near the PC rim. In the present work
we suggest a physical basis for this empirical observation, namely
the acceleration of secondary cascade electrons created near the rim.
More precisely, the excess rim density may be supplied by a multistep
process initiated by the reversed acceleration of secondary positrons
created just below the acceleration cutoff height. These particles can
produce downward-oriented cascades, creating new pairs near the NS surface.
A fraction of the electrons from these pairs may then be accelerated
upward along with the true primary electrons, adding to the net outward
flow. We argue that this sort of cascade feedback process should occur
preferentially near the PC rim, where the open magnetic field lines have
their maximum curvature.
\newpage
\section{EXTENDED PAIR CASCADES}
In our treatment of the NAR Model in DH94, we assumed that
the acceleration of PC electrons starts at the NS surface and is cut off
sharply at a height $h \ll R_{pc}$ by an overlying force-free pair
plasma. This assumption was made partly for simplicity, and also
because there is still no firmly established, self-consistent
electrodynamical model for magnetospheric acceleration, either near the
NS surface or elsewhere. However, we note that a significant problem with
these models may be resolved if the accelerating potential $\Phi(h)$
extends upward to heights $h \sim 2 \-- 3$ NS radii or higher.
We will first discuss our motivation for exploring extended
acceleration regions, then describe our model results based on specific
empirical choices for $\Phi(h)$.
For simplicity we retain our NAR model assumption that each PC is almost
circular with radius $R_{pc} = R_{ns} \theta_{pc}$, where $R_{ns}$ is the
NS radius. While the more general SPC model allows larger values of
$\alpha$ and hence noncircular PCs, this approximation should be still
adequate for our present treatment. For purely dipolar fields,
the conventional estimate for the half-angle $\theta_{pc}$ is just
\begin{equation} \label{tpc}
\sin\theta_{pc} = \left({R_{ns} \over R_{lc}}\right)^{1/2}
= \left({R_{ns} \Omega\over c}\right)^{1/2}
\end{equation}
where $R_{lc} = c/\Omega$ denotes the distance to the velocity-of-light
cylinder and $\Omega$ is the NS angular rotation frequency. Eq. (\ref{tpc})
assumes that a dipole field line, emerging from a point near the PC rim,
should close just inside the light cylinder. However, as we noted in DH94
this estimate ignores all plasma effects and thus should be regarded only
as a lower limit on $\theta_{pc}$. For example, Michel (1982, 1991) has
found that the presence of a force-free, rigidly corotating plasma
(even without inertial effects or outward current flow) causes a
distortion of the field lines which increases the PC radius by a
factor $\sim 1.3$. Hence we argue that a more realistic model could be
expected to increase $\theta_{pc}$ by a factor $\sim 2$ over Eq. (\ref {tpc}).
If we make the usual assumption that the magnetic field is purely dipolar,
the equation describing a given field line emerging from the PC is just
\begin{equation} \label{dpl}
r = k \sin^{2}\theta
\end{equation}
where k is constant. At a given point on the field line, the angle $\psi$
of the local tangent (measured from the magnetic axis) is given by
\begin{equation} \label{tnl}
\tan\psi = {{3 \sin\theta \cos\theta} \over {3 {\cos^{2}\theta- 1}}}
\end{equation}
If the gamma beam size is approximately determined by the locus of tangents
to the outermost open field lines, for $\theta_{pc} \ll 1$ a cascade gamma
beam originating from the NS surface would have a half angle
$\theta_{b} \sim \tan\psi \sim {3 \over 2} \theta_{pc}$.
In general we can use Eqs. (\ref{dpl}) and (\ref{tnl}) to estimate
the increase in beam width $\theta_{b} \sim \psi$ with height, for a
given PC radius.
Figure 1 illustrates this height dependence by plotting the tangent angle
$\psi$ vs. radial distance along the field lines, for the case of the
Vela pulsar ($P = 0.89$ ms). The curves labeled 1, 2,... denote field
lines emerging from the NS surface at the corresponding multiples of
$\theta_{pc}$ as given by Eq. (\ref{tpc}). It is evident
that if the cascade gamma emission extends upward to heights exceeding
$\sim 3$ NS radii, $\theta_{b} \sim \psi$ can become significantly
larger than ${3 \over 2} \theta_{pc}$. This effect is even more
pronounced if $\theta_{pc}$ is taken to be $\sim 2$ or more times
the standard estimate (\ref{tpc}).
\section{ACCELERATION AND ENERGY LOSSES ABOVE POLAR CAPS}
We have shown that from the standpoint of viewing geometry, extended PC
cascades may provide a viable solution to the observability problem.
The obvious next step is to examine the possiblity that the acceleration
of electrons from the PC surface might be sustained up to heights of
several NS radii. This question also requires us to consider in detail
the energy loss mechanisms which may affect the net acceleration.
Due to the intense ($\sim 10^{12} G$) NS magnetic fields, electrons
accelerated from the PC surface are constrained by rapid SR losses
to follow the field lines. Hence they obey a one-dimensional equation
of motion, which may be expressed as an energy-balance equation:
\begin{equation} \label{erg}
{d\gamma \over ds} = (\beta c)^{-1}
\left[
\left({d\gamma \over dt}\right)_{acc}
- \left({d\gamma \over dt}\right)_{cr}
- \left({d\gamma \over dt}\right)_{cs}
- \left({d\gamma \over dt}\right)_{other}
\right]
\end{equation}
Here $\gamma$ denotes the electron Lorentz factor, $\beta = v/c$, and
$s$ is the distance traversed along the field line. The subscripts
labeling the component energy gain and loss rates are defined as follows.
The subscript {\it acc\/} denotes the energy gain due to electrostatic
acceleration in regions where ${\bf E \cdot B}$ is nonzero.
We assume this term is proportional to ${\bf E_{\parallel}}$, the component
of ${\bf E}$ parallel to ${\bf B}$, at each point along the particle
trajectory (magnetic field line). Unfortunately, current models of
pulsar magnetospheres do not agree on the behavior of ${\bf E_{\parallel}(r)}$
near the PC surface. Hence the energy-gain term in Eq. (\ref{erg}) must
be regarded as unknown. However, we can at least assume various simple
models for the accelerating potential (e.g. Ruderman and Sutherland 1975,
Arons 1983) in our simulations and compare the results for each model
with observations. In Sections 6 and 7 we show that we have
been able to find self-consistent models of extended cascades which
yield light curves and spectra similar to the observed values. We have
also identified significant constraints on the accelerating field which
are critical to the viability of these models.
In contrast to the gain-rate term, the principal loss-rate terms in
Eq. (\ref{erg})are reasonably well
understood. The subscripts {\it cr\/}, {\it cs\/}, and {\it other\/}
denote energy losses due to CR, Compton upscattering (Dermer 1990,
Chang 1995), and other scattering processes respectively. One example
of the latter is triplet pair production (Mastichiadis et al. 1986,
Mastichiadis 1991, Dermer and Schlickeiser 1991). Sturner (1995) has
recently provided a systematic treatment of PC electron acceleration
which considers these energy-loss processes in detail. We have used
his expressions for the CS loss terms in our simulations, although his
treatment involves a number of simplifying approximations.
The CR loss rate has the simple form (see for example Jackson 1975)
\begin{equation} \label{crl}
\left({d\gamma \over dt}\right)_{cr} =
{2 \over 3} {e^2 \over {m c}}
{\gamma^4 \over {\rho_{c}}^2}
\end{equation}
where $\rho_{c}$ is the local radius of curvature of the magnetic field
line. For a purely dipolar field, the exact
expression for $\rho_{c}$ is just
\begin{equation} \label{crd}
\rho_{c} = {{k (\sin^4 \theta + \sin^2 2 \theta)}
\over
{\sin^4 \theta + 2 \sin^2 2 \theta - 2 \sin^2 \theta \cos 2 \theta}}
\end{equation}
Since Eq. (\ref{crd}) yields values $~\sim 10^7$ cm for standard
PC model parameters, the CR loss rate only becomes significant for
$\gamma \gtrsim 10^6$. At higher energies it is by far the dominant loss
mechanism.
The CS loss rate results from upscattering of ambient photons by the
accelerated electron beam. In our model the photon background consists of
thermal emission from the NS surface, and hence the CS loss rate should
only be significant only at heights $h \lesssim R_{ns}$ above the surface.
Pulsed X-ray observations of Geminga (Halpern and Ruderman 1993)
and PSR B1055-52 (\"Ogelman and Finley 1993)
suggest that for at least some sources the thermal
background may include multiple components at distinct temperatures
(e.g., emission from both the cooling NS surface and hotter
regions in the vicinity of the PCs).
The CS loss rate is found from the general expression (Dermer 1990,
Sturner 1995)
\begin{equation} \label {cs1}
\left({d\gamma \over dt}\right)_{cs} =
c \int d\epsilon \int d\Omega n_{ph}(\epsilon, \Omega)
(1 - \beta \cos \Psi)
\int d\epsilon'_s
\int d\Omega'_s
{d\sigma' \over {d\epsilon'_s d\Omega'_s}}
(\epsilon_s - \epsilon)
\end{equation}
where $\epsilon = \hbar \omega / {m_e c^2}$ is the incident photon energy
in units of the electron rest energy, $n_{ph}(\epsilon, \Omega)$ is the
number density of incident background photons within energy and solid-angle
increments $d\epsilon$ and $d\Omega$, and $\Psi$ denotes
the angle between these photons and the local electron beam direction.
The quantity $d\sigma' / d\epsilon'_s d\Omega'_s$ is the magnetic Compton
scattering cross section in the local electron rest frame (ERF), where
the primes denote quantities evaluated in the ERF and the subscript
$s$ labels scattered photon quantities.
In the strong magnetic field the CS cross section includes both nonresonant
and resonant components (Herold 1979, Daugherty and Harding 1986,
Bussard et al. 1986). Dermer (1990) has derived a nonrelativistic
approximation for the loss rate based on the magnetic Thomson cross section
in the ERF (Herold 1979), resolving the total loss rate into component
terms which he labels `angular`, `nonresonant`, and `resonant`.
Sturner (1995) has applied further simplifying assumptions to these terms
in order to derive convenient expressions for the CS loss rate.
His results are summarized in his equations (4)-(9), which we have
incorporated into our acceleration tracing algorithm.
Sturner (1995) notes that for $\gamma \gtrsim 10^3$, the incident thermal
photon energies above the cyclotron resonance may become relativistic
$(\epsilon' \gtrsim 1)$. In this case he replaces the nonresonant component
of the cross section by a relativistic (but nonmagnetic) Klein-Nishina
expression given by his equations (10)-(14). In this work we have
included these expressions, although we note that a more accurate treatment
will require the use of the magnetized (resonant) Klein-Nishina cross section
(Daugherty and Harding 1986, Bussard et al. 1986).
The only loss term which Sturner (1995) includes under the `other` label
in Eq. (\ref {erg}) arises from electron-photon scattering events in which
the scattered photon is replaced by an emergent $e^+/e^-$ pair. Using
cross sections found by Mastichiadis et al. (1986) and Mastichiadis (1991)
for the nonmagnetic form of this process, Sturner (1995) applies a
monoenergetic photon approximation to derive a loss rate given by
his equation (16). For our model parameters this term is never dominant,
but for generality we have also included it in our simulation. As in the case
of his Klein-Nishina CS loss rate, however, we note that in future work the
magnetic form of this process should be investigated since it may also
exhibit resonant behavior which may increase its signifance.
\newpage
\section{SIMULATIONS OF EXTENDED PC CASCADES}
The basic features of our cascade simulation code are described in DH94.
The version used in this work includes several major improvements.
These include revised adaptive algorithms for tracing photon propagation,
which allow more accurate localization of near-threshold pair conversion
events. We have also improved the tracing of synchrotron/cyclotron
emission, which now more accurately simulates both recoil and angular
distribution effects in the cyclotron regime ($\gamma \gtrsim 1$).
However, the most significant improvement for this work is the algorithm
for tracing electron acceleration through extended regions above the PC.
In our current version, each primary electron emerges from the surface
with an initial Lorentz factor $\gamma_0 \gtrsim 1$. Assuming
specific parameters for both the energy gain and loss mechanisms
as described in Section 3 above, the calculation then traces the net
acceleration of the electron
as it escapes outward along the local magnetic field line.
For this purpose we have developed an adaptive numerical technique to
integrate Eq. (\ref {erg}) which accomodates a wide range of
energies and distance scales. To estimate the significance of Compton
losses due to thermal photons from the NS surface we have used a model
similar to that employed by Sturner (1995),
in which the PC has a uniform surface temperature $T_6$ (in units of $10^6 K$)
within a circle of radius $R_{tpc}$ centered on each magnetic pole.
This region is defined as the {\it thermal\/} PC. Note that $R_{tpc}$ may
differ from the PC radius as defined by the locus of the outermost open
field lines. In fact we treat both $T_6$ and $R_{tpc}$ as parameters
in the model. At present we ignore any softer emission which may be
emitted from the overall surface.
Figure 2 shows sample acceleration profiles $\gamma(h)$, where $h$
is the height above the NS surface in stellar radius units.
Curve (a) shows a case in which the accelerating
field is assumed to be constant, namely
$(d\gamma / ds)_{acc} = 5$ ${\rm cm}^{-1}$,
from the surface up to a sharp cutoff at height $h_c = 3$.
Curves (b) and (c) both assume that the gain rate is a linearly
increasing function $(d\gamma / ds)_{acc} = 5h$, over this same region.
They differ only in the assumed values for the Comptonization
parameters, namely the thermal PC temperature $T_6$
and radius $R_{tpc}$ (measured in NS radius units).
Curve (b) assumes $T_6 = 1$ and $R_{tpc} = 0.1$,
corresponding to a cool, small thermal PC. The opposite case of a
hot, large PC, is shown by curve (c) which assumes $T_6 = 2$ and
$R_{tpc} = 0.5$. We note that the constant-acceleration curve (a)
is not affected by these variations of the Comptonization parameters,
since in this case
the gain rate greatly exceeds the loss terms in Eq. (\ref{erg}).
We also observe that even the linear-acceleration profiles are sensitive
to the Comptonization parameters only for heights $h \ll 1$,
and they have little effect on the peak energies reached at the
cutoff height $h_c$.
As $\gamma$ exceeds values $\gtrsim 10^6$ the primary CR emission
reaches gamma ray energies, resulting in photon-pair cascades.
The calculation, as described in DH94, recursively traces the full cascade
development and accumulates 3D tables of emergent $\gamma$-ray counts
vs. energy and solid angle, from which we derive spectra and light curves
of the emission as seen from arbitrary viewing directions.
In this work we have accumulated photon counts from ensembles
of primary electrons distributed in concentric rings over the PC surface.
We have assumed that the primary beam current is
axisymmetric with respect to the magnetic axis, hence the electrons in
each ring are spaced uniformly in azimuth. However, our analysis facility
allows us to assign arbitrary weights to the $\gamma$ counts from each ring.
This technique allows us to vary the assumed radial dependence of the
primary electron current density without requiring new runs of the
simulation.
Finally we should point out that our current simulation is based strictly
on a CR-initiated cascade, i.e. it considers Comptonization as
an energy loss mechanism acting on the primary electrons but it does
not yet include the upscattered photons as a source
of high-energy $\gamma$-rays which may themselves initiate cascades.
This is in obvious contrast to the cascade model proposed by
Sturner and Dermer (1994), in which Comptonization
provides {\it all \/} the high-energy input photons.
Under our model assumptions the primaries reach much higher peak energies
($\gamma \gtrsim 10^6$) than the values they assume ($\gamma \sim 10^5$),
so that in our case CR should initiate the bulk of the cascade emission.
However, we recognize that Comptonization may add a measurable contribution
to the emergent $\gamma$-emission and in a separate work we will extend
the cascade simulation to trace the CS upscattered photons as well.
At the same time, we note that the CS contribution may be expected
to produce a narrower
$\gamma$-beam than the extended CR component we consider here, since it
should originate closer to the PC surface. Thus it is possible that
PC cascades initiated by CR and CS photons may be distinguishable both
spatially and energetically.
\newpage
\section{ELECTRON CURRENTS NEAR THE PC RIM}
In DH94 we showed that single magnetic poles
can exhibit doubly peaked light curves with phase separations
$\delta\phi \lesssim 0.5$ if $\alpha \sim \theta_b$ and the observer
angle $\zeta \sim \alpha$. However, in order
to reproduce the small duty cycles of the double peaks seen in the Crab,
Vela, and Geminga, we had to impose an additional {\it ad hoc\/}
assumption that the primary electron density is strongly concentrated
near the PC rim. We also noted that there are two possibilities for
obtaining doubly peaked profiles with $\delta\phi < 0.5$, in which
the designations of {\it leading\/} and {\it trailing\/} peaks are
reversed. In DH94 we considered in detail the case in which the first
peak corresponds to the phase at which the observer
viewpoint emerges from the interior of the (hollow) $\gamma$-beam, while the
second peak marks the point of reentry. This case, which we denote
as the Exterior Scenario (ES), can produce $\delta\phi < 0.5$ if the
rotational axis is contained within the $\gamma$-beam ($\alpha < \theta_{b}$).
By combining the ES with the assumption that the primary current is
concentrated near the PC rim, we could account for both the short
duty cycles and the lack of emission outside the peaks (since this would
be the phase interval during which the observer viewpoint penetrates
the interior of the hollow beam).
In this scenario we associated the finite emission observed
between the peaks with residual, higher-altitude cascades, which would
produce emission with larger beam widths.
In work following DH94 we have compared our model predictions in detail
with CGRO observations of phase-resolved spectra for the Vela pulsar
(Kanbach et al. 1994). We have concluded that the ES does
not provide uniformly consistent fits to the spectra, especially for
the phase intervals between the main peaks.
In the ES model the high-altitude cascades which produce
the interpeak emission do tend to produce harder spectra below their
characteristic high-energy turnovers, since a smaller fraction of the
hard CR emission is converted to softer cascade photons. By itself
this trend is at least qualitatively consistent with the observations.
However, the peak CR energy ($\propto \gamma^3$) also decreases rapidly
as the primaries lose energy above the acceleration zone, with the result
that the turnovers in the interpeak cascade spectra drop to lower
energies compared to the peak spectra. In this respect the model
prediction is opposite to the observed trend.
This problem with the ES has led us to reexamine the alternative
labeling of the leading and trailing peaks, in which the PC interior
is identified as the source of the interpeak emission. We refer to
this case as the
Interior Scenario (IS). In order to produce finite interpeak emission
in this case, we must abandon the phenomenological DH94 model of a pure
rim distribution for the primary electrons. However, if we replace the
pure rim model with a two-component model which includes a uniform
interior current, it turns out that the IS allows a more consistent
overall agreement with the observations than the ES. Moreover, in this
scenario we can suggest a tentative physical interpretation for a
two-component primary current. In particular, the uniform component
is a simple
approximation of a Goldreich-Julian (GJ) current
$I_{GJ} = \pi {R_{pc}}^2 c \rho_{0}$ (Goldreich and Julian 1969), where
\begin{equation} \label{gjc}
\rho_{0} \sim {{-\bf{\Omega \cdot B}} \over {2 \pi c}}
\end{equation}
which
should be valid if $\theta_{pc} \ll 1$. We propose that this component
includes all the true {\it primary\/} electrons drawn from the
NS surface. In this view the extra rim component consists of secondary
electrons from pairs preferentially created near the PC rim, where the
increasing field-line curvature produces more rapid $\gamma$-pair
conversions.
If any secondary pairs contribute to the PC current of high-energy
particles which initiates cascades, the pairs must themselves
be accelerated to energies comparable with the peak primary energies.
This in turn would require at least some pairs to be created well below
the acceleration cutoff height. If (as we assume here) the primaries
are negative electrons ($e^{-}$), each $e^{-}$ secondary would then move
{\it upward\/} and thus add to the GJ primary current, while the $e^{+}$
would be accelerated {\it downward\/} along the local field line toward the
surface. In fact the model $\gamma$-ray light curves we present in Section 6
show that if just a small fraction ($\sim 10^{-2}$ or less) of the
cascade pairs created near the rim can be boosted to $\gamma \gtrsim 10^6$,
a two-component current model shows good agreement with observations.
In spite of these results, we must first consider a fundamental
theoretical objection to the acceleration of secondary pairs.
The problem is that the onset of cascade pair production
is expected to produce a sharp cutoff in the acceleration of the
primaries at a height $h_c$, which marks the boundary of the
overlying pair plasma (e.g. Ruderman and Sutherland 1975, Arons 1983).
Our own simulation results confirm that the quenching of
$E_{\parallel}$ above $h_c$ should be an abrupt process, since
the density of created pairs is found to rise sharply with height.
This is demonstrated in Figure 3, which plots typical growth curves of
the multiplicity $M = (N_s^+ + N_s^-)/N_p$ where $N_p$ and $N_s$
denote the numbers of primaries and secondaries
respectively. Thus even if pairs created at the lowest heights can
be accelerated by a decreasing $E_{\parallel}$ within a finite
transition zone, the growth curves indicate that this zone
is too short for any $e^{-}$ secondaries to reach energies
comparable to those of the primaries. This appears to eliminate the
most obvious model for enhancing the PC current near the rim, in which
the negative pair members are accelerated outward with the primaries.
However, the positron ($e^{+}$) component in such a transition zone
must also be subject to acceleration. The key point here
is that these particles may be drawn {\it downward\/} from the
transition zone back into the acceleration zone, following the local
field lines back toward the NS surface. In fact they should traverse
a distance comparable to the full extent of the acceleration zone,
and thus reach energies sufficient to create (tertiary) pairs by a
variety of possible mechanisms (e.g. $\gamma-B$ pair production,
triplet pair production). The result would be the creation
of pairs deep within the acceleration zone, whose $e^{-}$
members could be accelerated outward with the primaries to reach
similar peak energies.
This sort of cascade feedback process should be most likely to occur
above those regions of the PC where the original upward-directed cascades
initiated by the primaries commence at the lowest heights. Unless the
electrostatic acceleration varies greatly over the PC interior regions,
the increasing curvature of the field lines from the pole to the rim
implies that the primary cascades develop first near the rim (cf. Figure 3).
Hence we argue that reverse $e^+$-acceleration and downward-oriented
cascades occur preferentially around the rim.
As a first test of this hypothesis we have generalized our acceleration
tracing algorithm to follow secondary positrons downward from creation points
just below the cutoff height, back toward the NS surface.
The results confirm that these particles can be boosted to
$\gamma \lesssim 10^7$ at heights $h \gtrsim R_{ns}$ above the surface,
allowing their CR spectra to reach pair-conversion energies and
initiate downward-oriented cascades. In a separate work we will refine
our complete simulation code to investigate the development of these
cascades in detail. We anticipate that their presence may impact our model
in several respects, since in addition to providing a new source of
electrons these cascades can influence the behavior of the acceleration
process just above the surface. In particular, if the cascades
create a sufficiently dense layer of pair plasma overlying the surface
they can retard acceleration below the effective height of this layer.
In addition, it is possible that energetic downstreaming cascade photons
can impose severe Comptonization losses on upward-directed electrons.
As described in the following sections, in this work we will allow for
these possibilities by considering simple models in which the acceleration
may effectively commence at finite heights above the NS surface.
\section{GAMMA-RAY LIGHT CURVES}
The 3D photon count tables accumulated by the simulation may be summed over
energy bins to produce 2D sky maps of the $\gamma$-emission between
arbitrary energy
limits. An example is shown in Figure 4, which plots a grayscale contour map
of emission over 100 MeV. Any horizontal line drawn across this plot
corresponds to a specific value of the polar angle $\zeta$ for a given
viewing direction, and the counts distributed along this line define
the $\gamma$-ray light curve as seen from this viewpoint.
Following the arguments in Section 5 we present sample results for
the Vela pulsar using a simple two-component primary current model,
which we obtain by superimposing simulation datasets for concentric
rings of primaries as discussed in Section 4. In each case we have
included a total of 10 rings spaced at equal radial increments
to cover the PC interior. Since each ring contains 180 particles
with a uniform 2-degree azimuthal spacing, the inner rings are
weighted $\propto r^{-1}$ to approximate a uniform interior density.
To simulate test cases with a moderate rim component, we have weighted
the outermost ring by arbitrary factors in the range 3 to 5. Physically
this corresponds to the acceleration of a few secondary electrons for
each primary electron on this ring, which is a small fraction of the
$10^{2}-10^{3}$ cascade pairs created per primary near the rim.
All the datasets we have accumulated to date assume the following
general form for the accelerating field, namely
\begin{equation} \label{acc1}
E_{\parallel}(h) = {{m c^2} \over e} \left({d\gamma \over ds}\right)_{acc}
= {{m c^2} \over e} [a_0 + a_1 (h - h_0)] \Theta(h - h_0) \Theta(h_c - h)
\end{equation}
We choose units for Eq. (\ref{acc1}) such that the path length $s$ is
measured in cm, while the height $h = (R - R_{ns})/R_{ns}$ is in NS
radius units from the PC surface, $\Theta (x)$
is the unit step function ($0$ for $x < 0$, $1$ for $x > 0$),
and the constants $a_0$, $a_1$, and $h_0$ are taken as free parameters
in our model. Their values effectively determine the height at which
cascades commence above the PC rim, which Arons (1983) denotes as the
``pair formation front''. In the following we take the height at
which the cascade multiplicity exceeds unity (cf. Figure 3)
as a reasonable measure of the acceleration cutoff height $h_c$.
Thus $h_c$ is a function of $(a_0, a_1, h_0)$ but is {\it not\/}
itself a free parameter. In practice we determine $h_c$ from trial
simulations before generating complete datasets.
The quantity $h_0 \ge 0$ in Eq. (\ref{acc1}) denotes the height at which
acceleration commences. We introduce $h_0$ to allow for the possibility
that downward-oriented cascades may prevent or impede acceleration just
above the NS surface. As noted in Section 5, this can occur
either if the cascades create a sufficiently dense layer of pair plasma
overlying the surface, or if downstreaming cascade photons impose strong
Comptonization losses on upward-moving electrons. In a separate study
of downward-oriented cascades we will investigate both of these effects
in order to put physical constraints on the choice of $h_0$, but here
we simply explore the effects of varying $h_0$ in sample models.
If we let $a_1 = 0$ in Eq. (\ref{acc1}) we obtain a constant-field
approximation, resembling vacuum gap acceleration models of the type
proposed by Ruderman and Sutherland (1975). If instead we set
$a_0 = 0$, we have a crude approximation for the potential suggested by
Arons (1983) in his slot-gap model. We have generated datasets for the
Vela pulsar using both of
these limiting forms. In each case we have empirically chosen combinations
of the parameters $(a_0,a_1,h_0)$ such that the primary electrons
reach their peak energies ($\gamma \gtrsim 10^6$) rapidly enough to
initiate cascades. We note that in this work we have assumed no dependence
of either $a_0$, $a_1$, or $h_0$ on the magnetic polar angle $\theta$.
We have used a further simplifying assumption here, namely that
the cutoff height $h_c$ has the same value over the PC interior
as determined by the onset of cascades near the rim. While this assumption
must be questioned in a more refined treatment, we show below that it
leads to encouraging agreement with observations.
Figure 5(a,b,c,d) shows model light curves obtained under these assumptions
for the acceleration function Eq. (\ref{acc1}), using sample parameters
$(a_0,a_1,h_0) = (5,0,0)$, $(0,5,0)$, $(0,20,1)$, and $(50,0,2)$,
which we denote as models A, B, C, and D respectively.
Table 1 lists additional simulation parameters which are common to all
these models. In models A-C a common weight factor of 5
was assigned to the outermost ring of primary electrons to represent the
excess rim current, while a factor 3 was used for model D. (The simulation
would assign a weight factor of 1 to this ring for a uniform distribution.)
In each case the rim weights were chosen to obtain reasonable fits to the
observed Vela light curve. Since each simulation includes a total of 10
concentric primary current rings covering the PC interior, these rim
weight factors increase the total PC currents above their uniform component
values by factors of roughly 1.7 for models A-C and 1.4 for model D.
For comparison, in each of these plots the light curves which would be
produced by the uniform current alone (without the excess rim component)
are shown in gray.
If we compare these model results with the observed Vela
light curve (Kanbach et. al 1994) shown in Figure 9, we see that
the acceleration parameters which best match the observations are those
for which the acceleration near the surface is low. In fact, satisfactory
fits are obtained only if the
primaries do not reach $\gamma \gtrsim 10^6$ until after they have
attained heights $h \gtrsim 1$. If they exceed these energies at altitudes
too far below the cutoff height $h_c$, the total cascade emission which
they produce over the full acceleration region and beyond is spread
over large solid angles, yielding broad pulse peaks.
In particular, this tendency rules out constant-acceleration models
($a_1 = 0$) such as that shown in Figure 5(a), except in cases where
$h_0 \gtrsim 2$ as in Figure 5(d). A comparison of Figures 5(b), 5(c)
shows that even for linear acceleration ($a_0 = 0$), the fits are
significantly improved by introducing nonzero values of $h_0$.
Among the sample runs shown in Figure 5, models C and D show peak duty
cycles which are in the best agreement with the observed values. Moreover,
in each of these cases the first half of the interpeak emission
resembles both the magnitude and slope seen in the data. This example
shows that the two-component model for the primary current can produce
consistent agreement with a significant portion of the total light curve.
Unfortunately the agreement breaks down for the
trailing interpeak component, but since our model assumes axisymmetric
current rings it cannot account for any strong asymmetry in the light curve.
Finally we note that all these models predict a low but finite level of
emission
throughout the phase interval between Peak 2 and Peak 1 (i.e., over regions
outside the PC rim). This emission
is due to the residual, high-altitude cascades which we suggested in DH94
might be the source of the interpeak emission. Kanbach et al.
(1994) find no detectable emission in this phase interval for Vela,
and no evidence for unpulsed emission. Given their stated estimates
for the EGRET detector sensitivity, however, their findings are not in
conflict with our model results for the sample datasets C and D described
above. However, the observations do impose an additional constraint on the
relative weight factors for the two-component PC current distribution.
For example a uniform PC current, without any rim current enhancement,
would produce significantly more emission outside the peaks than the
observations allow.
\section{PHASE-RESOLVED ENERGY SPECTRA}
The same choices of parameters (model C and D) which best match the
observed light curves in Section 6 also produce the best fits for the
energy spectra. In spite of the similar appearance of their light curves,
however, model D produces better spectral fits than model C. In fact,
as shown in Figure 6 model D provides the closest match to the observed
total (phase-averaged) spectrum across five decades in energy,
The spectral differences among these models
are principally due to their varying extent of cascade development.
In models A and B the primary electrons reach maximum energies of
$7.5 \times 10^6$ and $1.2 \times 10^7$ respectively, compared to
$1.7 \times 10^7$ for Model C and $2.0 \times 10^7$ for Model D.
The values for models A and B especially are too low to supply
either the photons up to 3 GeV or the level of emission observed below
100 MeV. Clearly, the observed Vela total emission is not the result
of curvature radiation alone.
Figure 7 shows that this agreement for model D applies not only to the
total spectra, but also to the phase-resolved spectra observed by EGRET
(Kanbach et. al. 1994). These plots show fits for various phase intervals
defined by these authors in their power law-fits to the Vela
phase-resolved spectra for energies between 70 and 4000 MeV.
The normalization factors were determined separately at each phase interval
to match the data and differ by less than a factor of 2.
We note that the model
reproduces the tendency for the (quasi) power-law spectra at the phase
intervals of the two peaks to become significantly softer than the spectra for
the interpeak subintervals. In the Interior Scenario (Section 5)
this trend is expected since the interpeak emission is due to the interior
primary electron current, whose hard CR emission is less efficiently
converted to softer cascade photons (cf. Figure 3). The IS model also
reproduces the
observational feature that the high-energy turnovers in the Vela spectra
occur at lower energies for the peaks vs. the interpeaks. The sharpness of
the high-energy turnovers in the P1 and P2 spectral intervals, due to
magnetic one-photon pair production attenuation, are also reproduced,
especially in P1.
The model D spectra in the phase intervals LW1 and TW2, the emission just
outside the peaks, turnover more gradually and at energies below 500 MeV.
This emission is primarily curvature radiation at high altitudes from primary
electrons that have lost a significant amount of their maximum energy.
These phase intervals are thus predicted to have the softest spectra,
consistent with both the data and the high indices of the power law
fits of Kanbach et al. (1994). In model D, the hard spectra in
intervals I1 and I2 extend to energies below 10 MeV, predicting that the
interpeak emission should decrease relative to that of the peak emission
at lower energies. This appears to be verified by the 0.07 - 0.6 MeV
light curves measured by OSSE (Strickman et al. 1995), where no interpeak
emission was detected.
One quantitative measure of the spectral evolution during each pulse
is the hardness ratio $H$, defined here as the ratio of the flux over
300 MeV to the flux between 100 and 300 MeV. Figure 8 shows the model
D hardness ratio vs. pulse phase for $\zeta = 16^{\deg}$, corresponding
to the phase-resolved spectra in Figure 7. The trend toward harder spectra
during the interpeak phase interval is clear and appears to be consistent
with EGRET Vela observations (Fierro et al. 1995).
\newpage
\section{TOTAL GAMMA FLUX ESTIMATES}
If we identify the uniform component of our model PC current with the
GJ current predicted by Eq. (\ref {gjc}), we can estimate an upper limit on
the absolute $\gamma$-ray flux levels expected from our model sources
within any specified energy range $\Delta E_\gamma$. The required inputs
are the dataset sky map counts, the pulse period $P$, and the estimated
distance $D$ (which we take to be 500 pc for Vela). We outline the procedure
briefly as follows.
First we derive the effective number of primaries traced in the simulation,
taking into account the weight factors assigned to each concentric ring
of electrons. Following the arguments in Section 5, we resolve this total
number of primaries into two components representing uniform and rim
distributions respectively. As noted above, our total flux estimate
(uniform plus rim components) assumes that the uniform component is
a GJ current. For the flux estimate, the quantity of interest is the
number of GJ primaries in the simulation.
After summing the full 3D photon arrays over the energy range
$\Delta E_{\gamma}$ to produce the appropriate 2D sky maps, we find
the number $\Delta N_\gamma$ of photons accumulated along a 1-bin strip
of constant $\zeta$ and angular width $d \zeta$ during one full pulse
($\Delta \phi = 2 \pi$). The phase-averaged $\gamma$-ray flux $F_\gamma$
{\it per primary electron\/} at the distance D is then given by
\begin{equation} \label{flx}
F_\gamma = \Delta N_\gamma / 2 \pi \sin \zeta d \zeta P D^2 N_{GJ}
\end{equation}
where $N_{GJ}$ denotes the effective number of GJ primaries in the
dataset (excluding the excess rim component). Finally we obtain an absolute
total flux estimate by multiplying Eq. (\ref{flx}) by the (maximal) current
of GJ primaries from the PC surface as given by Eq. (\ref{gjc}).
The predicted fluxes for our Vela models A,B,C,D at energies $ > 100$ MeV
as found from this procedure are
$7.3 \times 10^{-5}$,
$1.6 \times 10^{-4}$,
$2.8 \times 10^{-4}$,
$2.8 \times 10^{-4}$
photons ${\rm cm}^{-2} {\rm s}^{-1}$ respectively.
It turns out that these values are all an order of magnitude
higher than the average flux observed by EGRET (Kanbach et al. 1994),
namely $(7.8\pm 1.0) \times 10^{-6}$ photons ${\rm cm}^{-2} {\rm s}^{-1}$
for $E_{\gamma} > 100$ MeV. Our high model flux levels, which obviously
are due to strong beaming factors of the hollow-cone emission,
are not by themselves a problem for our model since the GJ estimate
should properly be regarded only as an upper limit on the PC current.
We note, however, that the model flux estimate does fall closer to the
GJ limit as the $\gamma$-beam half-angle $\theta_b$ is increased.
In this respect
the excess predicted flux shows that even larger PC dimensions and/or
acceleration cutoff heights can be allowed within the framework of
the model.
\section{COMPARISON WITH OBSERVATIONS AT OTHER WAVELENGTHS}
In the preceding sections we have applied the SPC model specifically to
the Vela pulsar, in part because both the $\gamma$-ray light curves and
phase-resolved spectra for this object have been observed in
considerable detail. However, our model results for Vela can also
account in general terms for the $\gamma$-ray emission from other
pulsars with doubly-peaked profiles such as the Crab, Geminga, and
PSR B1951+32 (Ramanamurthy et al. 1995). The second general class
of light curves predicted by the SPC model, namely those with only
a single broad peak, may describe PSR B1055-52 (but see below).
At present the only source whose $\gamma$-ray light curve may be
difficult to accomodate is PSR B1706-44 (Thompson et al. 1992), since
recent EGRET observations (Thompson et al. 1995) suggest that this object
may have a triply-peaked pulse.
However, we must consider whether the SPC $\gamma$-ray model is
also compatible with observations of pulsed emission at other wavelengths
from Vela and the other known $\gamma$-ray pulsars. Our primary
concern here involves the possible implications of these observations
regarding the viewing geometry for each source. In this context we focus
especially on three $\gamma$-ray pulsars for which we also have strong
evidence of thermal X-ray emission from the NS surface, namely
Vela itself (\"Ogelman et al. 1993), Geminga (Halpern and Holt 1993),
and PSR B1055-52 (\"Ogelman and Finley 1993).
These objects are of particular interest since the modulation and phase
behavior of the X-ray emission should be directly related to the magnetic
field geometry at the NS surface.
To facilitate the discussion of these sources, in Figure 9 we have assembled
their light curves at various wavelengths using a common phase origin for
each source. It turns out that each object presents a distinct set of
challenges for our model, which we analyze separately below.
Although it shows no evidence of surface thermal X-ray emission we must
also consider observations at other wavelengths from the Crab pulsar.
The Crab has the distinction of having doubly peaked light curves in
phase at all observed wavelengths. However, its optical emission exhibits
polarization swings which cause special problems for the SPC model.
An additional challenge is presented by recent HST and ROSAT imaging
of the inner Crab nebula, which strongly suggest an observer
angle $\zeta \lesssim 60^{\deg}$ (Hester et al. 1995).
(a) Vela (PSR B0833-45)
As shown in Figure 9(a), the pulsed radio emission from Vela
(see for example Manchester and Taylor 1977) exhibits a single narrow peak
which leads the first $\gamma$-ray peak by
$\sim 0.12$ in phase (Kanbach et al. 1994). The radio pulse shows a high
degree of linear polarization with an unusually wide
swing ($\gtrsim 90^{\deg}$) in the polarization angle $\psi$ across the
pulse. This behavior has been interpreted (Radhakrishnan and Cooke 1969,
see also Michel 1991) in terms of the rotating projection of a dipolar
magnetic field in the plane orthogonal to the viewing
direction. In this model $\psi$ is given as a function of $\alpha$,
$\zeta$, and the pulse phase angle $\phi$ by
\begin{equation} \label{pol}
\tan \psi = \sin \alpha \sin \phi
/ (\sin \zeta \cos \alpha - \cos \zeta \sin \alpha \cos \phi)
\end{equation}
Several authors (e.g. Lyne and Manchester 1988, Rankin 1990) have attempted
to invert this relation to determine the values of $\alpha$ and $\zeta$ for
various pulsars, although the results to date are subject to controversy
(Michel 1991, Miller and Hamilton 1993). However, Eq. (\ref{pol}) does
imply that the maximum rate of the polarization
swing $R \equiv |d(\tan \psi) / d(\sin \phi)|$ occurs at the phase
corresponding to the closest approach of the magnetic axis to the observer
direction, which we denote by $\phi_{M}$.
If this model is correct, the rapid, extended swing for Vela
($R \sim 5.9$) indicates that the observer viewpoint approaches a
magnetic pole to within a few degrees.
As may be seen from Figure 10, the values of $\alpha$ and $\zeta$ used
in the Vela model datasets discussed in Sections 6 and 7
do not produce polarization swings which are either as rapid or extended
as the observed values. However, the real challenge in accounting
for the radio pulse in our model is not simply to find better combinations
of these parameters. The key point is that if the radio pulse does indeed
mark the phase of closest approach to either of the magnetic poles,
in the case of Vela its location relative to the $\gamma$-ray peaks is
inconsistent with the SPC model. In particular, the Interior Scenario
requires $\phi_{M}$ to lie midway between the two $\gamma$ peaks, whereas
in the Exterior Scenario it is displaced from the midpoint by $0.5$
in phase. In contrast, Kanbach et al. (1994) find the phases of the
$\gamma$ peaks (relative to the phase of the radio peak, $\phi_0 = 0$)
to be
$\phi_{p1} = 0.12$ and $\phi_{p2} = 0.54$ respectively.
Hence the standard PC model of the radio pulse asserts that
$\phi_{M} = \phi_0 = 0$, while the IS predicts
$\phi_{M} = (\phi_{p1} + \phi_{p2})/2 = 0.33$ and the ES has
$\phi_{M} = 0.83$. Thus the standard model of the Vela radio pulse
is inconsistent with the SPC $\gamma$-ray model.
On the other hand, it turns out that both the optical and X-ray light curves
for Vela fit much more naturally within the geometry of the IS.
As shown in Figure 9(a), the optical emission (Wallace et al. 1977)
has a doubly peaked light curve with a smaller peak-to-peak phase separation
($\sim 0.2$) than that seen in the $\gamma$-ray regime. Moreover, the
$\gamma$-ray peaks enclose the optical
peaks in the sense that the leading optical peak follows the leading $\gamma$
peak, while the opposite occurs for the trailing peaks (see for example
Manchester and Taylor 1977). In the IS, this sort of optical/$\gamma$ phase
relationship would hold if the optical and $\gamma$ emission were beamed in
coaxial hollow cones from the PC, with beam angles
$\theta^{opt}_{b} < \theta^{\gamma}_{b}$. This in turn suggests
that the optical emission might either be associated with interior
PC currents, or that it might be produced by the rim current at lower
heights than the $\gamma$-emission.
Figure 9(a) also shows the pulsed X-ray emission from Vela detected
by the ROSAT satellite (\"Ogelman et al. 1993), which consists of a
broad pulse trailing
the radio peak, with the bulk of the emission
occurring between the two $\gamma$-ray peaks. The harmonic content of the
pulse suggests a complex nonsinusoidal structure, although the available
X-ray data do not show firm correlations with the optical or $\gamma$
peaks (or clear evidence of more than one peak). The statistics
are unfortunately limited by the fact that the emission contains
contributions from the compact nebula as well as the pulsar,
and the pulsed fraction of the latter is only about 11\%.
\"Ogelman et al. (1993) obtain their best fit to the pulsed component
with a soft blackbody spectrum ($T_6 \sim 1.5-1.6$). They also note that
the total point source (pulsed plus unpulsed) can either be fit with a
blackbody spectrum at a similar temperature or with a steep power law
($\Gamma \sim -3.3$), compared to a harder power law ($\Gamma \sim 2.0$)
which fits the surrounding compact nebula. \"Ogelman et al. (1993)
suggest that if the pulsed component is actually thermal emission,
the modulation may be due either to a nonuniform surface
temperature distribution or to anisotropic radiation transfer
effects in the magnetosphere. In either case the key point
for our model is that the pulsed X-ray emission should then be concentrated
near the phase $\phi_{M}$ of closest approach of the observer direction to
a magnetic pole (Page 1995). To the extent that the bulk of the emission
does occur between the $\gamma$ peaks, the Vela X-ray light curve appears
compatible with the IS $\gamma$-ray model.
In summary it appears that the observed optical, X-ray, and $\gamma$-ray
light curves for Vela all seem mutually consistent with the IS,
whereas the radio polarization swing cannot have the usual interpretation
based on Eq. (\ref{pol}) in either the IS or the ES.
At present we have no satisfactory way to account for the phase of the
Vela radio pulse within
the general framework of any SPC model, unless we invoke the
possibility of nondipolar magnetic fields near the NS surface.
However, we should point out that this incompatiblity is not simply a
problem for our $\gamma$-ray model. The same conflict already exists
between the standard radio model and the entire class of thermal X-ray
models (e.g. Page 1995) in which the peak(s) in the pulsed emission
coincide with the closest approach of the magnetic pole(s) to the
observer viewpoint.
(b) Geminga (PSR B0630+178)
Although Geminga has long been known to be a strong $\gamma$-ray source
(Kniffen et al. 1975), it was first discovered to be a pulsar from
X-ray observations (Halpern and Holt 1992). Shortly thereafter
$\gamma$-ray pulses were detected at the X-ray period (Bertsch et al. 1992).
To date no pulsed emission has been found at either radio or optical
wavelengths, although an optical counterpart has been identified
(Bignami et al. 1993).
While the lack of optical and radio light curves prevent the sort of
phase comparisons we can make for other sources, both the X-ray and
$\gamma$-ray data are relatively rich in detail. Figure 9(b) shows
the light curves for Geminga at both hard and soft X-ray energies
from ROSAT observations (Halpern and Ruderman 1993) as well as in
the EGRET $\gamma$-ray regime (Mayer-Hasselwander et al. 1994,
Ramanamurthy 1995).
As in the case of Vela, the $\gamma$-ray light curve above 100 MeV
exhibits a two-peak
structure with significant interpeak (bridge) emission. The peaks have
duty cycles only moderately larger than in Vela, with a phase separation
of $0.5$. In contrast, Halpern and Ruderman (1993) find
that the X-ray light curves at both soft (0.07-0.53 keV) and hard
(0.53-1.50 keV) energies consist of broad single pulses. The hard
component is somewhat narrower, but perhaps most remarkably the soft
and hard components are $\sim 105^{\deg}$ out of phase.
Halpern and Ruderman (1993) have fit the hard and soft components of the
pulsed X-ray spectrum to two blackbody sources at temperatures
$T_{6} \sim 0.5$ and $\sim 3$ respectively. These authors suggest
that the soft emission is from the overall NS surface, while the hard
component arises from hotter regions around a PC. However, they also
note that within the available statistics a power-law fit for the harder
component is nearly as good as the blackbody fit, which leaves open
the possibility of magnetospheric emission mechanisms. In any event
the hot PC model of the hard X-ray emission appears to be consistent
with the SPC $\gamma$-ray model, as in the case of Vela, since as seen
in Figure 9(b) the bulk of the hard X-ray pulse from Geminga also lies
between the double $\gamma$-ray peaks (Halpern and Ruderman 1993).
Unfortunately the modulation of the soft X-ray component and its phase
shift relative to the hard component complicate this model.
In fact the hard and soft components may not be consistently
explained within the framework of {\it any\/}
NS heating/cooling models which assume dipolar magnetic-field symmetry.
This point has led Halpern and Ruderman (1993) to suggest
an off-axis dipole model in the case of Geminga.
(c) PSR B1055-52
This source has been detected by EGRET at energies above 300 MeV
(Fierro et al. 1993). Figure 9(c) shows that in contrast to the doubly
peaked radio pulse, the $\gamma$-ray light curve appears to exhibit
a single broad peak. However, the available statistics are insufficient
to rule out a multipeaked substructure. The limited data makes it difficult
to analyze the phase relationship between the radio and $\gamma$ pulses,
although it may be significant that the precursor of the main radio pulse
appears just at the trailing end of the $\gamma$ peak. It is noteworthy
that the radio profile has some similarity to that of the Crab,
including a peak-to-peak phase separation $\gtrsim 0.4$ which would
require an off-axis dipole in an orthogonal rotator model.
PSR 1055-52 has the distinction of exhibiting the hardest phase-averaged
$\gamma$-spectrum of all the $\gamma$-ray pulsars known to date,
with a photon spectrum index of $\sim 1.2$. It is worth noting here that
PC cascades can definitely exhibit such hard spectra, although they tend
to do so only when both the electron CR losses and pair-conversion rates are
comparatively low. These conditions are most likely to apply in specific
regions of the magnetosphere, especially close to the magnetic axes and/or
at heights of several NS radii above the surface. However, both more
detailed $\gamma$-ray observations and further modeling of this source
will be required to determine how the hardness of the spectrum may
constrain the SPC model.
Pulsed X-rays have also been detected from PSR B1055-52
by ROSAT (\"Ogelman and Finley 1993). As in the case of Geminga, the emission
exhibits distinct hard and soft components above and below $\sim 0.5$ keV,
both of which exhibit broad single pulses. Figure 9(c) shows the phase
relationships between the X-ray light curves and the pulses at
radio and $\gamma$-ray energies. As in the $\gamma$-ray regime, evidence
for substructure in either X-ray component is limited by the available
statistics. Another striking similarity with Geminga is the large
relative phase shift between the hard and soft X-ray peaks, with the hard
component in this case leading
by $\sim 120^{\deg}$. \"Ogelman and Finley (1993) obtain satisfactory
spectral fits using two-component blackbody models,
although they find that the hard component may also be fit by a power law
which extrapolates up to flux levels in the $\gamma$-ray regime
comparable with the EGRET observations.
If PSR B1055-52 does in fact have only one $\gamma$-ray peak, then its
relationship to the X-ray emission may be difficult to explain within
the SPC model. The key problem is that the model identifies the phase
of a single $\gamma$ peak with the phase $\phi_M$ of closest approach
of the PC.
However, if the hard X-ray component is due to PC heating as proposed
for Geminga (Halpern and Ruderman 1993), the X-ray peak indicates a value
for $\phi_M$ in apparent conflict with the $\gamma$-ray location.
While this difficulty does not arise if the hard X-rays have a
magnetospheric origin as \"Ogelman and Finley (1993) suggest, their phase
shift relative to the $\gamma$-ray pulse is still problematical.
As in the case of Geminga, however, the modulation of the soft X-ray
component and its phase shift relative to the hard component
complicate the picture. The fact that the radio pulse for PSR B1055-52
has two peaks, with noteworthy similarities to the Crab radio profile,
is also puzzling. However, the principal question regarding
the viability of the SPC model for this source is whether the
$\gamma$-ray light curve is singly peaked. Hopefully further analysis
of EGRET data will be able to resolve this question.
(d) The Crab Pulsar (PSR B0531+21)
In constrast to all other $\gamma$-ray pulsars, the light curve of the
Crab exhibits a doubly peaked structure at at all wavelengths observed
to date, with the peaks appearing at essentially the same phase positions
throughout the entire spectrum. In purely geometric terms this phase
synchronization seems to suggest that a variety of emission processes,
which may occur in distinct magnetospheric regions of other pulsars,
are spatially coincident in the Crab. In the context of SPC $\gamma$-ray
models it appears to motivate a search for radio, optical, and
X-ray emission mechanisms involving the cascade pairs.
Unfortunately, this approach leads to at least one serious difficulty
for the SPC model, namely the optical polarization swings found to occur
across each peak (Smith et al. 1988). If both the optical and $\gamma$
peaks do originate from the same PC rim regions, then the optical
swings cannot be due to the sort of rotational projection effect
described by Eq. (\ref{pol}) since the extent of the swing
through the phase intervals containing each $\gamma$-peak cannot exceed
a few degrees (cf. Figure 10). However, SPC models for the Crab appear
to be compatible in this respect with the radio pulses, which do not
exhibit significant polarization swings.
In addition to this problem, a significant constraint on SPC models of
the Crab pulsar is posed by recent HST and ROSAT observations of the inner
nebula (Hester et al. 1995). These observations appear to confirm
numerous earlier
suggestions that the observer angle $\zeta$ for the Crab is considerably
larger ($\lesssim 60\deg$) than the values ($\sim 15\deg$) used in our
sample Vela datasets. However, this finding does not by itself rule out
the SPC model for the Crab, since it turns out that such large values of
$\zeta$ can be accomodated if we allow the PC dimensions to be
$\sim 4 \-- 5$ times larger than the standard estimate Eq. (\ref{tpc}),
as opposed to the factor $2$ used in our model datasets for Vela.
Somewhat smaller values are also adequate if the cascades are assumed
to extend up to heights $\gtrsim 3$ NS radii. Thus in the case of the
Crab especially, the dimensions of the PC are critical to our model.
\newpage
\section{DISCUSSION}
The model we have presented here has at least one significant advantage
over an alternative SPC model (Sturner and Dermer 1994, Sturner et al. 1995)
in which PC cascades are initiated by Comptonizaton of primaries by soft
photons from the NS surface rather than CR emission. As we have shown,
extended primary acceleration can easily generate CR-induced cascades
at heights reaching up to several NS radii. In contrast, cascades due
to Comptonization should be confined to significantly lower regions
unless some mechanism for strong beaming of the soft photons is invoked.
Assuming that similar PC dimensions are used in both models, the
Comptonization model has a more limited ability to overcome the
observability problem.
The best results we have obtained to date from the extended cascade SPC model
are for those cases in which the net electron acceleration becomes significant
only at heights $h \gtrsim R_{ns}$ above the NS surface.
However, we have shown in Section 3 that neither resonant Compton scattering
of thermal photons from the NS surface nor other known energy loss processes
considered in previous PC models can effectively counteract
accelerating potentials of the types we have considered over distances
of this order. This applies in particular to resonant Compton scattering,
even if we assume the highest plausible values for both the surface
temperatures and thermal PC radii. Thus it is obviously important to
investigate the possibilility noted in Section 5, namely that
downward-oriented cascades initiated by reversed secondary
acceleration can prevent or impede acceleration just above the surface.
An obvious next step in the exploration of the SPC
model is to trace the development of downward-oriented cascades in detail,
and if possible to estimate both their significance as a source of
energetic Comptonizing photons and the depth of the surface plasma layer
which they may create.
The discussion in Section 9 shows that the phase relationships between
light curves at different wavelengths are in fact quite complex.
The problem of accounting for all these observations in a self-consistent
manner may eventually force us to consider models with asymmetric magnetic
field geometries. One initial step in this direction would be to consider
off-axis dipolar models of the type suggested by Halpern and Ruderman (1993)
in more detail. In such models we anticipate that the modulation of
thermal X-ray emission from, say, the PC surface may be significantly
out of phase with magnetospheric emission produced above the surface
and directed along the open field lines.
\section{ACKNOWLEDGEMENTS}
We are indebted to Joe Fierro, Gottfried Kanbach, Peter Michelson,
P.V. Ramanamurthy, and David Thompson for valuable discussions regarding
EGRET observations, and to Mark Strickman for information regarding OSSE
and COMPTEL results. We also thank Hakki \"Ogelman and John Finley for
discussions on the pulsed X-ray emission from Vela. Michal Marko provided
valuable assistance in the development of our visualization and analysis
software. We gratefully acknowledge support for this work from NASA CGRO
Guest Investigator Grants for Phases 3 and 4 (JKD, AKH), and from the
NASA Astrophysics Theory Program (AKH).
\newpage
|
\section{Introduction}
Observations of anisotropy in the Cosmic Microwave Background
Radiation (CMBR) yield valuable clues about the formation of
large-scale structure in the early universe.
A particularly interesting angular scale for
observing CMBR anisotropy is near 0\fdg5, where the first ``Doppler peak''
(or adiabatic peak)
enhancement of the fluctuation power spectrum
is expected to be observable (\cite{white94}).
The Medium Scale Anisotropy
Measurement (MSAM) is an experiment designed to measure CMBR anisotropy at
this angular scale. This paper reports the initial results from the
second flight of this experiment.
A number of detections of anisotropy at angular scales near 0\fdg5
have been reported recently. Observations by ARGO
(\cite{debernardis94}), the Python experiment (\cite{dragovan94}), the
fourth flight of the MAX experiment (\cite{devlin94,clapp94}), SK94
(\cite{netterfield94}), and SP94 (\cite{gundersen94}) all report
detections of
anisotropy near this angular scale.
Quantifying CMBR anisotropy at the level of these detections
is an extremely challenging observational task (\cite{wilkinson95}).
Many potential systematic errors cannot be
unequivocally ruled out at the necessary levels, with the result
that any single observation cannot prudently be accepted without
an independent confirmation.
The results in this paper are our attempt to confirm the results
of our previous work.
By observing the same region of the sky with a second balloon flight,
we demonstrate the repeatability of our measurements in the
presence of potential atmospheric noise and contamination from Earthshine.
We have reported earlier (\cite{cheng94}, hereafter Paper~I) our
observations of anisotropy of the CMBR from the first flight of MSAM
in 1992.
Our results from those observations were 1) a
positive detection of anisotropy, with the caveat that we could not
rule out foreground contamination by bremsstrahlung;
2) the identification of two particular
bright spots that were consistent with being unresolved sources.
This paper reports our first results from the 1994 flight of MSAM, which
observed an overlapping field.
\section{Instrument Description}
This instrument has been briefly described in Paper~I;
we give only an overview here.
It has four spectral bands at 5.6, 9.0, 16.5, and
22.5~${\rm cm}^{-1}$, giving sensitivity to CMBR and Galactic dust. The
off-axis Cassegrain
telescope forms a 30\arcmin\ beam on the sky. The chopping secondary
mirror moves this beam in a step motion 40\arcmin\ left and right of center.
The beam moves center, left, center, right with a
period of 0.5~s. The detectors are sampled at 32~Hz, synchronously with
the chop.
The telescope is mounted on a stabilized balloon-borne platform.
The absolute pointing reference is provided by a star camera; positions between
camera fixes are interpolated using a gyroscope.
The telescope is shielded with aluminized panels so that the dewar feed horn,
the secondary and most of the primary have
no direct view of the Earth.
The gondola superstructure was changed between the 1992 and 1994
flights. The previous superstructure as viewed from the telescope
had a substantial cross-section
of reflective material; in spite of
our efforts to shield it we were concerned about the telescope
being illuminated by reflected Earthshine. The new
design is a cable suspension with considerably lower cross section
above the telescope.
Ground measurements indicate that rejection of signals from sources
near the horizon is better than 75~dB in our longest wavelength
channel.
\section{Observations}
The package was launched from Palestine, Texas at 00:59~UT 2~June~1994,
and reached its float altitude of 39.5~km at about 03:25~UT.
Science observations
ended with sunrise on the package at 12:04~UT. During the flight
we observed Jupiter to calibrate the instrument and map the
telescope beam, scanned M31 (which will be reported in a future {\sl
Letter\/}), and integrated on the same CMBR field observed during the
1992 flight for 3.5 hours.
The CMBR observations were made as described in Paper~I.
The telescope observes near the meridian 8\arcdeg\ above the north
celestial
pole, and scans in azimuth $\pm 45\arcmin$ with a period of 1
minute. The
scan is initially centered on a point 21\arcmin\ to the east of
meridian. We track to keep this point centered in our scan until
it is 21\arcmin\ to the west of meridian, then jog 42\arcmin\ to the
east. Each scan takes about 20 minutes, and half of each scan
overlaps the preceding scan.
We completed
4.5 such scans from 05:12 to 06:38~UT (we call
this section 1 of the data), and
completed an additional 7 scans from 07:22 to 09:43~UT (section 2). The
observed field is two strips at declination $81\fdg8 \pm 0\fdg1$,
from right ascension 15\fh27 to 16\fh84,
and from 17\fh57 to 19\fh71 (all coordinates are J1994.5).
Fig.~\ref{f_fields} shows the
fields observed in the 1992 and 1994 flights.
The overlap between the fields is better than half a beamwidth
throughout the flight.
Our ability to observe exactly the same position on the sky is
currently limited by the error in determining the position of the IR
beam center during the initial in-flight calibration, i.e., our
real-time determination of pointing is not as accurate as our post-flight
determination.
\section{Data Analysis}
The signal from the detectors is contaminated by spikes induced by cosmic rays
striking the detectors; we remove these spikes.
The data are calibrated by our observation of Jupiter.
The absolute pointing is
determined from star camera images. The detector data
are analyzed to provide measurements of brightness in our four spectral
channels as a function of sky position. These are then fit to a
spectral model to produce measurements of CMBR anisotropy and dust
optical depth. These analyses and their results are described
in the following sections.
\subsection{Pointing}
We determined the pointing by matching star camera images
against a star catalog. This fixes the position of the camera frame
at the time the exposure was taken. Between exposures, position is
interpolated with the gyroscope outputs plus a small linear correction to
make the gyroscope readings consistent with the camera fixes. This
correction is typically 2\arcmin\ in 20~minutes. The relative
orientation of the camera frame and the IR telescope beam is fixed by
a simultaneous observation of Jupiter with the camera and the IR
telescope. The resulting absolute pointing is accurate to 2\farcm5,
limited by the gyroscope drift correction. The pointing analysis
was done in an
identical way for the 1992 flight, and has similar accuracy.
\subsection{Detector Data Reduction}
The instrument is calibrated by in-flight observations of Jupiter. The
brightness temperatures of Jupiter for our four spectral channels are
172, 170, 148, and 148~K, derived from the spectrum of Jupiter
observed by \cite{griffin86}. The apparent diameter of Jupiter during the
1994 flight is 42\arcsec. The uncertainty in the absolute
calibration is 10\%,
dominated by uncertainty in the antenna temperature of Jupiter. The relative
calibration uncertainty between the 1992 and 1994 flights is 5\%,
due to noise in the observations of Jupiter.
The detector signal contains spikes, at a rate of 0.25--0.5~${\rm s}^{-1}$,
consistent with the hypothesis that they are due to cosmic
rays striking the detectors (\cite{charakhchyan78}), and with the rate
reported in Paper~I. Cosmic rays deliver an
unresolved energy impulse to the detector; we remove them by fitting
the data to the impulse response function of the
detector/amplifier/filter chain.
We give here our results for the 5.6~${\rm cm}^{-1}$\ channel; the numbers for the
other channels are similar. Candidate spike locations are identified
using a $1.5\,\sigma$ threshold. The data within 1~s (5 detector time
constants) are fit to a model of the response function. About 2\% of
the spikes require a second spike 2--10 samples separated from the
first to be added to the fit. If the resulting spike amplitude has
less than $3\,\sigma$ significance, the data is left as-is. If the
fit is good, and the spike amplitude has more than $3\,\sigma$
significance, the spike template is subtracted. 5065 spikes are
subtracted out of 504,000 time samples. (We allow either positive or
negative amplitudes; 90\% of the spikes have positive amplitude.) If
the fit is poor, and the spike amplitude is significant, full data
records (64 samples, or 2 sec) before and after the spike are deleted.
317 spikes were eliminated this way, removing a total of about 6\% of
the data.
We estimate the instrument noise by measuring the variance in the
demodulated, deglitched data after removing a slow drift in time and
the mean in each sky bin. This estimate is made for each 20~minute
segment of data, and is then propagated throught the remaining
processing. All $\chi^2$ reported below are with respect to this
error estimate.
We divide the sky into bins that are small compared to the beamsize.
The bins are 0\fh057 in right ascension and 0\fdg12 in declination.
Due to sky rotation, the data
also need to be divided by angular orientation of the beam throw on
the sky; the bin size for this coordinate is 10\arcdeg. The data are then
fit to a signal in each sky bin plus a model of long-term drift formed
from a cubic spline with knots every 12 minutes (2.5 minutes for the
16.5~${\rm cm}^{-1}$\ channel), plus terms for
gondola inclination, roll, and air pressure. The simultaneous fit of long-term
drift and sky signal ensures that this fit does not bias our
observations of the sky. This fit is done separately on each channel
and section of the flight. The
resulting sky signals have bin-to-bin correlation, and we propagate a full
covariance matrix through the remainder of the analysis.
Sky bins containing less than 4~s of integration are deleted.
So that our error estimate, described in the preceding
paragraph, is unbiased by sky signal, we form the estimate from the
residuals of this fit, and iterate to obtain a consistent solution.
The data are demodulated in two different ways.
The double difference demodulation corresponds to
summing the periods when the secondary is in the central position, and
subtracting the periods when it is to either side.
This demodulation is least sensitive to
atmospheric gradients and gondola swinging.
The single difference demodulation is formed by
differencing the period when the secondary is to the right from that when
it is to the left, and ignoring the periods when the secondary is in the
center. We use the scan
over Jupiter to deduce optimal demodulations of the infrared signal.
The
binned dataset contains 90\% of all the data originally taken, with an
achieved sensitivity in each of the four channels of 240, 150, 80,
and 230~\hbox{$\mu$K}~$\sqrt{\rm s}$
Rayleigh-Jeans. For channels 1 and 2 this is 490 and 850~\hbox{$\mu$K}~$\sqrt{\rm
s}$ CMBR. The offsets in the demodulated data for the different
channels
and demodulations range from 1 to 6~mK~RJ, smaller than those
reported in Paper~I.
\subsection{Spectral Decomposition}
At each sky bin, we fit the four spectral channels to a
model consisting of a CMBR anisotropy plus emission from warm Galactic dust.
The results are not very sensitive to the parameters of the dust model;
we use a dust temperature of
20~K and an emissivity index of 1.5 (consistent with \cite{wright91}).
The fit is
done separately for the single and double difference demodulations.
The $\chi^2/$DOF for the fit is 408/430 (double difference) and
448/430 (single difference).
Fig.~\ref{f_dust} shows the resulting fitted dust optical depth at
22.5~${\rm cm}^{-1}$. For clarity this figure has been binned more coarsely and
does not distinguish between points at slightly different declination
or chop orientation; our analyses, however,
do not ignore these details. We
have fit our observations to the {\sl IRAS}\ Sky Survey Atlas at 100~\micron\
(\cite{wheelock93}) convolved with our beam patterns, with
amplitude and offset as free parameters. The resulting fit is
superimposed on Fig.~\ref{f_dust}. The $\chi^2/$DOF of this fit is
262/210 for the double difference demodulation and 310/210 for the
single difference. The ratio of optical depths between IRAS and our
data is consistent with an average dust emissivity spectral index
between our bands and 100~\micron\ of $\alpha = 1.40 \pm 0.16$
(still assuming a dust temperature of 20~K).
Our measurements of CMBR anisotropy are plotted in Fig.~\ref{f_cmbr}.
Superimposed are the measurements from 1992. As noted earlier, there
is non-negligible correlation between the error bars on different sky
bins. In making Fig.~\ref{f_cmbr} we have fit out the two largest
eigenmodes of the covariance matrix, and used error bars formed from
the diagonal of the covariance matrix after removing the two largest
eigenmodes; the result is that the error bars shown in the figure can
be approximately treated as uncorrelated.
(This procedure is similar to that used in \cite{fixsen94b} for the
{\sl COBE\/}/FIRAS calibration.)
The data have also been binned more
coarsely, as in Fig.~\ref{f_dust}. We stress that these
steps are taken only for producing representative figures;
in all quantitative analyses we use
the full dataset and the full covariance matrix.
We are in the process of calculating the correlation for the MSAM1-92 data;
the 1992 data plotted here are identical to those in Paper~I.
\subsection{CMBR Anisotropy}
To set limits on anisotropy in the CMBR, we assume Gaussian
fluctuations with a Gaussian-shaped correlation function. We set 95\%
confidence level upper and lower bounds on the total rms fluctuation
over the sky $(\sqrt{C_0})$, assuming this correlation function with a
given correlation angle $\theta_c$. The method used is described in
Paper~I, though we now use a full covariance matrix for the
instrument noise on the observations.
The upper and lower bounds from these observations for the
single and double difference demodulations are shown in
Fig.~\ref{f_deltat}. The bounds for the correlation angles at which
the two demodulations are most sensitive are summarized in
Table~\ref{t_deltat}, which also shows results for the two sections of
the flight separately.
The confidence intervals for both demodulations are consistent with
those in Paper~I.
\section{Conclusions}
We observed the same field in our 1992 and 1994 flights in order to
determine if the detected signal was due to sidelobe pickup,
atmospheric noise, or
other systematic effects, or was in fact present in the sky. While we
are still in the process of completing
a detailed quantitative comparison of the two
datasets, it is apparent that the double difference
CMBR anisotropy features reproduce quite well.
This encourages us to believe
that the signal we see
in the double difference is present on the sky, and that
contamination from atmosphere or sidelobes is small compared to the
sky signal.
The single difference CMBR signal does not appear to reproduce as well.
Pending
the completion of the more thorough comparison,
we cannot rule out contamination in the single difference channel.
In Paper~I we pointed out that the anisotropy we
observe could be due to diffuse Galactic bremsstrahlung. This
possibility remains, and will be addressed by our MSAM2 experiment,
which will observe the same fields in five bands over 65--170~GHz.
In Paper~I we raised the possibility that the ``sources'' at R.A. 19~h and
15~h were either foreground sources of a previously unknown population,
or non-Gaussian CMBR fluctuations. This speculation was prompted by
our belief that such features were inconsistent with Gaussian
statistics. More careful analysis by us
and independently by \cite{kogut94} has indicated that
features like these are in fact consistent with a variety of plausible
correlation functions.
Observations by \cite{church95} at 4.7~${\rm cm}^{-1}$\ rule out the source MSAM15$+$82
being more compact than 2\arcmin.
Therefore removal of these regions in studies of CMBR
anisotropy, as we recommended in Paper~I, are a biased edit
of the data, and we no longer recommend it.
Our current conclusion is that the double difference, whole flight numbers
in Table~\ref{t_deltat} are a reliable estimate of CMBR anisotropy in the
observed regions.
When we include the 10\% uncertainty in the calibration, the resulting
limits are
$\Delta T/T = 1.9^{+1.3}_{-0.7}\times 10^{-5}$
(90\% confidence interval)
for total rms fluctuations. In the band power
estimation of (\cite{bond95}), this is $\langle {\cal C}_l \rangle_B
= 2.1^{+1.5}_{-0.9}\times 10^{-10}$ ($1\,\sigma$ limits), with
$\langle l \rangle = 263$.
The CMBR anisotropy channel, Galactic dust channel, pointing,
covariance matrices, and beammaps are publicly available. For more
information, read {\tt
ftp://cobi.gsfc.nasa.gov/pub/data/msam-jun94/README.tex}.
\acknowledgments
We would like to thank the staff of the National Scientific Balloon
Facility, who
remain our willing partners in taking the calculated risks that result in
extremely successful flights.
W.~Folz and J.~Jewell traveled with us to the NSBF
to help with flight preparations.
T.~Chen assisted in building and testing our new star camera system.
We are grateful to M.~Devlin and S.~Tanaka for providing cappuccino at
the crucial moment in Palestine.
The Free Software Foundation provided the
cross-development system for one of the flight computers.
This research was supported by the NASA Office of Space Science,
Astrophysics Division.
\clearpage
\begin{deluxetable}{rccrrcrr}
\tablecolumns{8}
\tablecaption{Upper and lower bounds on total rms CMBR anisotropy ($\protect\sqrt{C_0}$)
\label{t_deltat} }
\tablehead{
\colhead{} & \colhead{} & \multicolumn{3}{c}{MSAM1-94} &
\multicolumn{3}{c}{MSAM1-92} \\
\cline{3-5}\cline{6-8}
\colhead{} & \colhead{} &
\colhead{} & \colhead{Upper} & \colhead{Lower} &
\colhead{} & \colhead{Upper} & \colhead{Lower} \\
\colhead{$\theta_c$} & \colhead{Section} &
\colhead{R.A.} & \colhead{Bound} & \colhead{Bound} &
\colhead{R.A.} & \colhead{Bound} & \colhead{Bound} \\
\colhead{} & \colhead{} &
\colhead{(h)} & \colhead{(\hbox{$\mu$K})} & \colhead{(\hbox{$\mu$K})} &
\colhead{(h)} & \colhead{(\hbox{$\mu$K})} & \colhead{(\hbox{$\mu$K})} }
\startdata
\cutinhead{Single Difference}
0\fdg5 & 1 & 15.27--16.84 & 163 & 40 \nl
& 2 & 17.57--19.71 & 75 & 17 \nl
& All & 15.27--19.71 & 79 & 30 & 14.44--20.33 & 116 & 53 \nl
\cutinhead{Double Difference}
0\fdg3 & 1 & 15.27--16.84 & 132 & 44 \nl
& 2 & 17.57--19.71 & 74 & 24 \nl
& All & 15.27--19.71 & 78 & 34 & 14.44--20.33 & 97 & 50\nl
\enddata
\tablecomments{The limits in this table do not include the calibration
uncertainty.}
\end{deluxetable}
\clearpage
\bibliographystyle{aas}
|
\section{Introduction} %
The four-dimensional self-dual Einstein equation (SdE) has been given
attention for a long time both in physics and mathematics, as well as
the self-dual Yang-Mills equation. Among a number of works associated
with the SdE \cite{EGH}, an interesting and important subject is to
connect it to other (possibly simple) field equations. Well-known
examples of it are Plebanski's heavenly forms \cite{plebanski}, there
the SdE is given in terms of one function of space-time coordinates.
Q-Han Park \cite{park} and Ward \cite{ward} have shown that the SdE is
derived from several two-dimensional sigma models with the gauge group
of area preserving diffeomorphisms, SDiff(${\cal N}_2)$. Park also has
clarified the correspondence between the sigma models and first and
second heavenly forms. On the other hand, by Ashtekar's canonical
formulation for general relativity \cite{ashtekar1}, the SdE has been
reformulated as the Nahm equation \cite{ashtekar2}, and its covariant
version is given in Ref.\,\cite{mason}.
Through this formulation, Husain has arrived at one
of sigma models, that is, the principal chiral model \cite{husain}.
Also by several reduction methods, other interesting models, e.g. the
SL$(\infty)$ (affine) Toda equation \cite{boyer}\cite{park}, the KP
equation \cite{castro} etc., are obtained.
\par
Although we have various examples connected to the SdE, their relation
is rather unclear since their derivations from the SdE are more or less
complicated and separated. Such a link of the models, however, should be
investigated in order to understand the SdE further and in particular to
develop the quantization of self-dual gravity.
\par
In this paper, we describe the self-dual Einstein space by a trio of
differential form equations for simple two-forms and derive several
integrable theories quickly. This formulation elucidates their relation
and may indicate the possibility to find further a large class of models
connected to the SdE.
\vskip 0.4cm
\section{Self-dual Einstein equation}%
We start from the observation that the SdE is expressed as closed-ness
conditions of basis of the space of anti-self-dual two-forms,
\begin{equation}
d ( e^0 \wedge e^i - {1 \over 2} \epsilon_{ijk} e^j \wedge e^k ) = 0 \ ,
\qquad i = 1,2,3, \label{eq: cls}
\end{equation}
where $e^{0,i} = e^{0,i}_{\mu} dx^{\mu}$ are tetrad one-forms on
four-manifold. This formulation was employed by Plebanski to reduce
the SdE to the heavenly forms \cite{plebanski}, there the indices
$i,j,k$ are replaced with spinor ones $A,B$ by the Pauli matrices
${(\sigma^i)_A}^B$.
The equations in (\ref{eq: cls}) appear also in
Ref.\,\cite{capovilla}\cite{abe}.
We rewrite (\ref{eq: cls}) by defining a null basis,
\begin{eqnarray}
&& {\cal Z} = e^0 + i e^1, \qquad \ {\bar {\cal Z}} = e^0 - i e^1,
\nonumber \\
&& {\cal \chi} = e^2 - i e^3, \qquad \ {\bar {\cal \chi}}
= e^2 + i e^3 \ .
\end{eqnarray}
Then (\ref{eq: cls}) becomes
\begin{equation}
d ( {\cal Z} \wedge {\cal \chi}) = 0 \ ,
\qquad
d ( {\bar {\cal Z}} \wedge {\bar {\cal \chi}}) = 0 \ ,
\qquad
d ( {\cal Z} \wedge {\bar {\cal Z}} + {\cal \chi}
\wedge {\bar {\cal \chi}}) = 0 \ . \label{eq: nc}
\end{equation}
\par
As for symmetry, adding to the diffeomorphism invariance, (\ref{eq: nc})
is invariant under the local SL$(2,C)$ (self-dual) transformation,
in matrix form,
\begin{equation}
\left[ \begin{array}{l} {\cal Z}^{'} \\
{\cal \chi}^{'} \end{array} \right] =
\left[ \begin{array}{cc} a \ & b \ \\ c \ & d \ \end{array} \right]
\left[ \begin{array}{l} {\cal Z} \\ {\cal \chi} \end{array} \right] \ ,
\qquad
\left[ \begin{array}{l} {\bar {\cal Z}}^{'} \\ {\bar {\cal \chi}}^{'}
\end{array} \right]
=
\left[ \begin{array}{cc} d \ & -c \ \\ -b \ & a \ \end{array} \right]
\left[ \begin{array}{l} {\bar {\cal Z}}
\\ {\bar {\cal \chi}} \end{array}
\right] \ ,
\quad ad-bc=1 \ , \label{eq: sl}
\end{equation}
and also invariant under the global SL$(2,C)$ (anti-self-dual) transformation
for the pairs
$({\bar {\cal Z}}, {\cal \chi})$, $({\cal Z}, {\bar {\cal \chi}})$
with the same form as (\ref{eq: sl}).
\par
The two-form in the last equation in (\ref{eq: nc}) is of rank-four, while
others are of rank-two, but we can re-express the last equation
by defining two one-forms ${\cal P} = {\cal Z} - {\bar {\cal \chi}}$,
${\cal Q} = {\cal \chi} + {\bar {\cal Z}}$.
Using them, we obtain the following equations equivalent
to (\ref{eq: nc}),
\begin{equation}
d ( {\cal Z} \wedge {\cal \chi}) = 0 \ ,
\qquad
d ( {\bar {\cal Z}} \wedge {\bar {\cal \chi}}) = 0 \ ,
\qquad
d ( {\cal P} \wedge {\cal Q}) = 0 \ . \label{eq: smcl}
\end{equation}
Since all two-forms in (\ref{eq: smcl}) are simple and closed,
they can be written as, on a local coordinate system,
\begin{equation}
{\cal Z} \wedge {\cal \chi} = dz \wedge dx \ ,
\qquad
{\bar {\cal Z}} \wedge {\bar {\cal \chi}} = d\bar z \wedge d\bar x \ ,
\qquad
{\cal P} \wedge {\cal Q} = dp \wedge dq \ , \label{eq: sm}
\end{equation}
where $(z,x,\bar z,\bar x,p,q)$ are functions. Although we use the notation
suitable for the real, Euclidean case, we generally consider the complex
SdE, so the bars in (\ref{eq: sm}) do not mean the complex conjugation in
usual. {}From the definition of ${\cal P}$ and ${\cal Q}$, two identities
are obtained,
\begin{eqnarray}
{\cal Z} \wedge {\cal \chi} \wedge {\cal P} \wedge {\cal Q} &&
= dz \wedge dx \wedge dp \wedge dq = dz \wedge dx \wedge d\bar z \wedge d\bar x
= {\cal Z} \wedge {\cal \chi} \wedge {\bar {\cal Z}} \wedge {\bar {\cal \chi}} \ ,
\nonumber \\
{\bar {\cal Z}} \wedge {\bar {\cal \chi}} \wedge {\cal P} \wedge {\cal Q} &&
= d\bar z \wedge d\bar x \wedge dp \wedge dq = dz \wedge dx \wedge d\bar z \wedge d\bar x
= {\cal Z} \wedge {\cal \chi} \wedge {\bar {\cal Z}} \wedge {\bar {\cal \chi}} \ .
\label{eq: id}
\end{eqnarray}
(\ref{eq: smcl}), (\ref{eq: sm}) and (\ref{eq: id}) are key equations
in our formulation. For later use, we define the notation
$x^a = (z, \bar z)$ and $x^k = (p,q)$.
\vskip 0.4cm
\section{ Integrable theories derived from the SdE} %
(a)${\it \, The \ principal \ chiral \ model}$ \\
At first, let us choose $(z,\bar z,p,q)$ as four coordinate variables and
$(x,\bar x) = (A_z, A_{\bar z})$ as functions of them.
Then (\ref{eq: id}) reads
\begin{equation}
dz \wedge dA_z \wedge dp \wedge dq = dz \wedge dA_z \wedge d\bar z \wedge dA_{\bar z} \ ,
\quad
d\bar z \wedge dA_{\bar z} \wedge dp \wedge dq = dz \wedge dA_z \wedge d\bar z \wedge dA_{\bar z} \ ,
\label{eq: pcm0}
\end{equation}
from which we can quickly derive the following equations after expanding
$dA_{z(\bar z)} = \partial_a A_{z(\bar z)} dx^a + \partial_k A_{z(\bar z)}dx^k$
and rescaling $A_{z(\bar z)}$ by $-2$,
\begin{equation}
\partial_z A_{\bar z} + \partial_{\bar z} A_z = 0 \ ,
\qquad
F_{z\bar z}= \partial_z A_{\bar z} -\partial_{\bar z} A_z
+ \{A_z, A_{\bar z} \} = 0 \ , \label{eq: pcm}
\end{equation}
where $\{A_z, A_{\bar z} \}$ is the Poisson bracket with respect to $(p,q)$.
Using the potentials $A_{z(\bar z)}$, we define generators
${\cal A}_{z(\bar z)} = \{\ \,, A_{z(\bar z)}\}
= \partial_q A_{z(\bar z)} \partial_p - \partial_p A_{z(\bar z)} \partial_q$ of the
algebra of area preserving diffeomorphisms sdiff(${\cal N}_2$), which, at each
space-time point $(z, \bar z)$, act on function on the internal surface ${\cal N}_2$
parametrized by the coordinates $(p,q)$.
The second equation in (\ref{eq: pcm}) means the generators ${\cal A}_{z(\bar z)}$
are pure-gauge, ${\cal A}_{z(\bar z)} = g^{-1} \partial_{z(\bar z)} g$, where $g$
is an element of the group SDiff(${\cal N}_2)$.
Substituting them into the first equation, we obtain
\begin{equation}
\partial_{\bar z}(g^{-1}\partial_z g)
+ \partial_z (g^{-1}\partial_{\bar z} g) = 0 \ .
\end{equation}
This is precisely the chiral model equation on the $(z, \bar z)$ space-time
with $(p,q)$ treated as coordinates on ${\cal N}_2$.
Let us solve $A_{z(\bar z)}$ for a single scalar function $\Theta$ by
the first equation in (\ref{eq: pcm}), that is,
$A_z = 2 \partial_z \Theta$ and $A_{\bar z} = - 2 \partial_{\bar z} \Theta$.
Then the second equation becomes
\begin{equation}
\Theta_{z,\bar z} + \Theta_{z,p} \Theta_{\bar z, q}
- \Theta_{z,q} \Theta_{\bar z, p} = 0 \ ,
\end{equation}
where $\Theta_{z, p} = \partial_z \partial_p \Theta$.
By an adequate gauge condition for the local SL$(2,C)$ symmetry,
tetrads can take the form,
\begin{eqnarray}
{\cal Z} = h^{1 \over 2} dz \ , \qquad &&
{\cal \chi} = - h^{1 \over 2} d\bar z
- h^{-{1 \over 2}} \Theta_{z,k} dx^k \ ,
\nonumber \\
{\bar {\cal Z}} = h^{1 \over 2} d\bar z \ , \qquad &&
{\bar {\cal \chi}} = \ \, h^{1 \over 2} dz +
h^{-{1 \over 2}} \Theta_{\bar z, k}dx^k \ ,
\qquad h = \{\Theta_{\bar z}, \Theta_z \} \ ,
\end{eqnarray}
and the line element is
\begin{equation}
ds^2 = {\cal Z} \otimes {\bar {\cal Z}}
+ {\cal \chi} \otimes {\bar {\cal \chi}}
= - \Theta_{a,k} \, dx^a \otimes dx^k
+ {1 \over \{\Theta_z, \Theta_{\bar z} \}}
\Theta_{z,k} \Theta_{\bar z,l} \, dx^k \otimes dx^l \ .
\end{equation}
It is obvious that all self-dual metrices are obtained from this sigma
model. In the case of $\{A_z,A_{\bar z}\}=0$, the volume form
${1 \over 4}({\cal Z} \wedge {\bar {\cal Z}}
\wedge {\cal \chi} \wedge {\bar {\cal \chi}})$
vanishes, which corresponds to a degenerate space-time.
\vskip 0.2cm
(b)${\it \, The \ topological \ model \ with \ the \ WZ \ term \ only}$
\\
Next we take $(z,x,\bar z,\bar x)$ as our coordinates and $(p,q) = (B_0, B_1)$
as functions of them. After changing the notation
$(z,x,\bar z,\bar x)=(z,\bar z,p,q)$, (\ref{eq: id}) gives
\begin{equation}
\{B_0, B_1 \}_{(z,\bar z)} = 1 \ , \qquad \{B_0, B_1 \} = 1 \ .
\end{equation}
The bracket in the first equation is defined with respect to $(z,\bar z)$.
These equations are rather unfamiliar, but if we define
$\partial_k A_{z(\bar z)} = \{B_0, B_1 \}_{(z(\bar z), x^k)}$, we can
easily check the integrability $\partial_{[k} \partial_{l]} A_{z(\bar z)}
= 0$ and
\begin{equation}
\partial_z A_{\bar z} - \partial_{\bar z} A_z = 0 \ ,
\qquad
\{A_z, A_{\bar z} \} = 1 \ . \label{eq: tm2}
\end{equation}
Also in this case, generators ${\cal A}_{z(\bar z)} = \{\ \,, A_{z(\bar z)}\}$
of sdiff(${\cal N}_2$) are pure-gauge,
${\cal A}_{z(\bar z)} = g^{-1} \partial_{z(\bar z)} g$,
and from the first equation, we obtain
\begin{equation}
\partial_{\bar z}(g^{-1}\partial_z g)
- \partial_z (g^{-1}\partial_{\bar z} g) = 0 \ .
\end{equation}
This is the topological model derived from
the lagrangian of the Wess-Zumino term only \cite{park}.
The potentials $A_{z(\bar z)}$ are given in terms of one
function $\Omega$ by the first equation in (\ref{eq: tm2}).
Then $A_{z(\bar z)}= \partial_{z(\bar z)} \Omega$
and the second equation becomes
\begin{equation}
\Omega_{z, p} \Omega_{\bar z, q} - \Omega_{z, q} \Omega_{\bar z, p} = 1 \ ,
\label{eq: pl}
\end{equation}
which is Plebanski's first heavenly form \cite{plebanski}.
With a gauge condition for the local SL$(2,C)$ symmetry,
tetrads are given by
\begin{equation}
{\cal Z} = dz \ , \quad {\cal \chi} = d\bar z \ , \quad
{\bar {\cal Z}} = \Omega_{z,k} dx^k \ ,
\quad {\bar {\cal \chi}} = \Omega_{\bar z,k} dx^k \ ,
\end{equation}
and the line element is $ds^2 = \Omega_{a,k} dx^a \otimes dx^k$.
\vskip 0.2cm
(c) ${\it \, The \ WZW \ model}$ \\
The equation of the Wess-Zumino-Witten model is obtained by dropping the
term in the right-hand-side of the first equation in (\ref{eq: pcm0}),
\begin{equation}
dz \wedge dA_z \wedge dp \wedge dq = 0 \ , \qquad
d\bar z \wedge dA_{\bar z} \wedge dp \wedge dq = dz \wedge dA_z \wedge d\bar z \wedge dA_{\bar z} \ .
\label{eq: wzw}
\end{equation}
After relabeling $(z,A_z,\bar z,A_{\bar z},p,q)$ as $(A_z,z,p,q,\bar z,A_{\bar z})$,
(\ref{eq: wzw}) becomes
\begin{equation}
\partial_z A_{\bar z} - \partial_{\bar z} A_z= 0 \ , \qquad
\{A_z, A_{\bar z} \} = 0 \ .
\end{equation}
Solving the potentials as in the case (b), we have
\begin{equation}
\Omega_{z, p} \Omega_{\bar z, q} - \Omega_{z, q} \Omega_{\bar z, p} = 0 \ .
\end{equation}
Comparing it with the first heavenly form (\ref{eq: pl}), we see that the
WZW model describes a fully degenerate space-time \cite{park}.
\vskip 0.2cm
(d) ${\it The \ Higgs \ Bundle \ equation}$ \\
Here, note that it is not necessary to choose all four coordinate
variables from the set $(z,x,\bar z,\bar x,p,q)$. Instead, we can regard
more than two variables in it as functions. This observation enables
us to obtain a further large class of models connected to the SdE.
\par
Let us examine the observation by taking $(z,\bar z)$ as two coordinate
variables and $(x,\bar x,p,q) = (\phi,{\bar \phi},B_0,B_1)$ as functions.
Then (\ref{eq: id}) reads
\begin{equation}
dz \wedge d\phi \wedge dB_0 \wedge dB_1 = dz \wedge d\phi \wedge d\bar z \wedge d{\bar \phi}
\ , \quad
d\bar z \wedge d{\bar \phi} \wedge dB_0 \wedge dB_1
= dz \wedge d\phi \wedge d\bar z \wedge d{\bar \phi} \ . \label{eq: hb}
\end{equation}
As like the case (b), we introduce
$\partial_k A_{z(\bar z)} = \{B_0, B_1 \}_{(z(\bar z), x^k)}$, in which
the coordinates $(p,q)$ are defined as the condition $\{B_0, B_1\}=1$
is satisfied. This condition ensures the integrability
$\partial_{[k} \partial_{l]} A_{z(\bar z)} = 0$.
Using $\phi, {\bar \phi}$ and $A_{z(\bar z)}$, (\ref{eq: hb}) and another
identity ${\cal P} \wedge {\cal Q} \wedge {\cal P} \wedge {\cal Q} = 0$ become
\begin{equation}
\partial_{\bar z} \phi + \{A_{\bar z}, \phi \} = - \{\phi, \bar \phi \} \ ,
\quad
\partial_{z} {\bar \phi} + \{A_z, {\bar \phi} \} = \{\phi, \bar \phi \} \ ,
\quad
F_{z \bar z}(A) = 0 \ ,
\end{equation}
or, changing $A_z \rightarrow A_z + \phi$ and
$A_{\bar z} \rightarrow A_{\bar z} + {\bar \phi}$,
\begin{equation}
\partial_{\bar z} \phi + \{A_{\bar z}, \phi \} = 0 \ ,
\qquad
\partial_{z} {\bar \phi} + \{A_z, {\bar \phi} \} = 0 \ ,
\qquad
F_{z \bar z} = - \{\phi, {\bar \phi}\} \ .
\end{equation}
This is the two-dimensional Higgs bundle equation with the group
SDiff(${\cal N}_2$), which is mentioned in Ref.\,\cite{ward}.
A self-dual Einstein metric is given by a solution of the model
through the tetrads,
\begin{eqnarray}
{\cal Z} = (\partial_p \phi + {g \over h} \partial_p \bar \phi) dz
+ {1 \over h} \partial_p \bar \phi \partial_{I} \phi \, dx^I \, ,
&&
{\cal \chi} = (\partial_q \phi + {g \over h} \partial_q \bar \phi) dz
+ {1 \over h} \partial_q \bar \phi \partial_{I} \phi \, dx^I \, ,
\\
{\bar {\cal Z}} = (\partial_q \bar \phi
- {{\tilde g} \over h} \partial_q \phi) d\bar z
- {1 \over h} \partial_q \phi \partial_{J} \bar \phi \, dx^J \, ,
&&
{\bar {\cal \chi}} = - (\partial_p \bar \phi
- {{\tilde g} \over h} \partial_p \phi)d\bar z
+ {1 \over h} \partial_p \phi \partial_{J} \bar \phi \, dx^J \, ,
\nonumber
\end{eqnarray}
where $x^I$ and $x^J$ mean the sets of variables, $x^I=(\bar z,p,q)$,
$x^J=(z,p,q)$, and $h= \{\phi,\bar \phi \}$, $g=\{\phi,A_z \}$ and
${\tilde g} = \{\bar \phi, A_{\bar z}\}$.
The Higgs bundle equation was originally derived from a dimensional
reduction of the four-dimensional self-dual Yang-Mills
theory \cite{hitchin}.
\vskip 0.2cm
(e) Let us consider another example
by choosing $(p,q)$ as two coordinates
and $(z,x,\bar z,\bar x)= (C_0,C_1,D_0,D_1)$
as functions of $(p,q)$ and other
$(z, \bar z)$. Also in this case we define
$\partial_k A_{z(\bar z)} = \{C_0, C_1 \}_{(z(\bar z), x^k)}$,
$\partial_k B_{z(\bar z)} = \{D_0, D_1 \}_{(z(\bar z), x^k)}$ and impose the
conditions $\{C_0, C_1\}=1$, $\{D_0, D_1\}=1$ to ensure the
integrability of $\partial_k A_{z(\bar z)}$ and $\partial_k B_{z(\bar z)}$.
Then we obtain
\begin{equation}
F_{z \bar z}(A)= F_{z \bar z}(B)= 0 \ , \
\epsilon^{ab} (\partial_a B_b + \{A_a, B_b \} ) = 0 \ , \
\epsilon^{ab} (\partial_a A_b + \{B_a, A_b \} ) = 0 \ . \label{eq: new}
\end{equation}
The first two equations result from
${\cal Z} \wedge {\cal \chi} \wedge {\cal Z} \wedge {\cal \chi} =
{\bar {\cal Z}} \wedge {\bar {\cal \chi}}
\wedge {\bar {\cal Z}} \wedge {\bar {\cal \chi}}=0$.
Tetrads in this case are, for example,
\begin{eqnarray}
&&{\cal Z} = h^{1 \over 2} dz
- h^{- {1 \over 2}} \partial_k A_{\bar z} dx^k \ , \ \
\ {\cal \chi} = h^{1 \over 2}d\bar z
+ h^{- {1 \over 2}} \partial_k A_z dx^k \ ,
\ h=\{A_z,A_{\bar z}\} \ ,
\nonumber \\
&&{\bar {\cal Z}} = - g^{1 \over 2}d\bar z
- g^{- {1 \over 2}} \partial_k B_z dx^k \ ,
\ {\bar {\cal \chi}} = g^{1 \over 2}dz
- g^{- {1 \over 2}} \partial_k B_{\bar z} dx^k \ ,
\ \, g=\{B_z,B_{\bar z}\} \ .
\end{eqnarray}
In this model, the flat potentials $A_{z(\bar z)}$ and $B_{z(\bar z)}$ interact
with each other through the third and fourth equations in (\ref{eq: new}).
\vskip 0.4cm
\section{Conclusion} %
In this paper, we have shown a formulation
of the self-dual Einstein space
which leads to low-dimensional field theories quickly and clearly.
Now the relation among those theories is rather clear.
For example, if we obtain a solution of (a) the principal chiral model,
then it is straightforward, at least formally, to derive
the corresponding solution of (b) the topological model or the first
heavenly form (\ref{eq: pl}); solving the potentials $A_{z(\bar z)}$
in (\ref{eq: pcm}) for $(p,q)$, relabeling $(z,A_z,\bar z,A_{\bar z},p,q)$ as
$(z,\bar z,p,q,B_0,B_1)$ and following the step in the case (b).
Here, for the opposite direction from (b) to (a),
we give a simple example.
A solution $\Omega$ of (\ref{eq: pl}) corresponding to the $k=0$ (flat)
Gibbons-Hawking metric \cite{gibbons}\cite{EGH} is given by
\begin{equation}
\Omega = 2 \epsilon^{-1} \sqrt{\bar z q} \sinh{z} \sinh{p}
+ 2 \epsilon \sqrt{\bar z q} \cosh{z} \cosh{p} \ , \label{eq: ex}
\end{equation}
where $\epsilon$ is an arbitrary constant.
We can make $\Omega$ real by setting the complex conjugate condition
$z^{\ast}=p$, ${\bar z}^{\ast}=q$.
Through the relation $\partial_k A_{z(\bar z)}
= \{B_0, B_1 \}_{(z(\bar z), x^k)}$,
the pair $(B_0, B_1)$ can take the form,
\begin{equation}
B_0 = \epsilon^{-1/2} \sqrt{2\bar z} \sinh{z}
- \epsilon^{1/2} \sqrt{2q} \cosh{p} \ , \
B_1 = \epsilon^{1/2} \sqrt{2\bar z} \cosh{z}
+ \epsilon^{-1/2} \sqrt{2q} \sinh{p} \ . \label{eq: bb}
\end{equation}
By changing the notation as noted above and solving for $(A_z,A_{\bar z})$,
we obtain a solution of the principal chiral model (\ref{eq: pcm}),
\begin{eqnarray}
A_z && = - [{{\epsilon^{-1/2} p \sinh{\bar z}
+ \epsilon^{1/2} q \cosh{\bar z} }
\over {\epsilon^{-1} \sinh{z} \sinh{\bar z}
+ \epsilon \cosh{z} \cosh{\bar z}}}]^2 \ ,
\nonumber \\
A_{\bar z} && = - [{{\epsilon^{1/2} p \cosh{z}
- \epsilon^{-1/2} q \sinh{z} }
\over {\epsilon^{-1} \sinh{z} \sinh{\bar z}
+ \epsilon \cosh{z} \cosh{\bar z}}}]^2 \ .
\end{eqnarray}
Also it may be interesting to discuss
various reduction procedures through
this formulation. For example, in the case (b) with the above complex
conjugate condition, suppose that $(A_z, A_{\bar z})$ are functions of
$\bar z, q$ and the imaginary part of $(z,p)$ only.
Then changing the imaginary part and $i A_z$, we obtain the
three-dimensional Laplace equation, which is just the `translational'
Killing vector reduction by Boyer and Finley \cite{boyer}.
Also for the `rotational' case, the SL$(\infty)$
Toda equation is derived
from (\ref{eq: id}) quickly.
It is intriguing to pursue a further large class
of models connected to the
SdE by arranging the functions $(z,\bar z,x,\bar x,p,q)$ suitably
and also to investigate the relation among the models.
\par
To check the coordinate variables which permit real, Euclidean metrices,
let us write down the form of ${\cal P} \wedge {\cal Q}$ explicitly,
\begin{equation}
{\cal P} \wedge {\cal Q} =
({\cal Z} \wedge {\cal \chi} + {\bar {\cal Z}} \wedge {\bar {\cal \chi}})
+ ({\cal Z} \wedge {\bar {\cal Z}} + {\cal \chi} \wedge {\bar {\cal \chi}}) \ .
\label{eq: rc}
\end{equation}
If all tetrads are real, the real, Euclidean case, the first term in
(\ref{eq: rc}) is real, while the second term pure-imaginary.
In the case of (a) the principal chiral model,
suppose $(p,q)$ are real or
complex conjugate to each other.
According to it, ${\cal P} \wedge {\cal Q} = dp \wedge dq $ becomes real or
pure-imaginary.
But then either the first or the second term in (\ref{eq: rc}) vanishes,
which corresponds to a degenerate space-time, not an interesting case.
Therefore, with such $(p,q)$, the chiral model
permits only complex metric,
or signature $(+,+,-,-)$ real metric in which
case
two of four tetrad one-forms
may be taken as pure-imaginary.
In fact $(B_0, B_1)$ in (\ref{eq: bb}),
which are $(p,q)$ in the case (a),
are neither real nor complex conjugate to each other.
\par
The infinite-dimensional group SDiff(${\cal N}_2$) is known to be
realized as a large N limit of the SU(N) group when the surface
${\cal N}_2$ is the sphere or the torus \cite{sun}.
Hence an approach to the SdE is to start from the SU(N) principal
chiral model, there its explicit
classical
solutions can be determined by the
Uhlenbeck's uniton construction \cite{uhlenbeck}.
Adding to the SDiff(${\cal N}_2$), several (hidden) symmetrical
structures in the SdE have been studied by the sigma model approach
\cite{park}\cite{husain}\cite{morales}.
Our formulation may be useful to investigate the structure since,
from our key equations (\ref{eq: sm}),(\ref{eq: id}), we can easily see
fundamental symmetries of our models, e.g. the conformal invariance on
the $(z,\bar z)$ space and
SDiff(${\cal N}_2$), before we derive their field equations explicitly.
\newpage
I am grateful to Q-Han Park, S. Nam, Ryu Sasaki
and S. Odake for discussions.
This work is supported by the department of research in Kyunghee
University and the Japan Society for the Promotion of Science.
\vskip 1.0cm
|
\section{Introduction}
The recent discovery of anisotropy in the cosmic microwave background (CMB)
lends support to the hypothesis that structure in the universe
formed via gravitational instability
from small density perturbations. Furthermore, the observed anisotropy
allows us to characterize the initial perturbations statistically
with enough precision to draw important cosmological conclusions.
In particular, the COBE detection (Smoot et al.~1992, Bennett et al.~1994)
shows that
density perturbations existed on scales
larger than the horizon at the epoch of recombination.
Furthermore, the data are consistent
with the hypothesis that the perturbations are Gaussian distributed and
can be used to place constraints on the power spectrum of the initial
density perturbations.
CMB anisotropy is already one of the most important pieces of cosmological
data, and in coming years its importance will only increase.
In recent years, many cosmologists have favored theories in which
the density parameter $\Omega_0$ is equal to one. In particular,
this value for $\Omega_0$ is predicted by
the simplest versions of the popular inflationary scenario.
However, many dynamical measurements indicate
a low-density universe (Peebles 1993, Ratra {$\&$ } Peebles 1994a).
It is well known that the standard COBE-normalized Cold Dark Matter (CDM)
scenario, in which universe is flat and contains only baryons and
CDM particles, and in which the initial density perturbations
are of the Harrison-Zel'dovich type, predicts fluctuations that are
too large in amplitude on scales of galaxy clusters and below,
although a number of slight variants on this model can
be devised that fit the data better ({\it e.g.}, Bunn, Scott, \& White
1995; White et al.~1995).
There are two classes of low-density cosmological model
that can be motivated by the inflationary universe scenario.
One is the $\Lambda$-model. In this model
the matter density is small, so $\Omega_0<1$; however,
the cosmological constant contributes to the total mean
density required to make the universe flat: $\Omega_0+\Omega_\Lambda=1$.
The usual inflationary
scenarios predict that the universe is flat, so
that $\Omega_0+\Omega_\Lambda$ must be
extremely close to unity (Kashlinsky, Tkachev, {$\&$ } Frieman 1994).
The cosmological constant is equal to the vacuum energy
density of the Universe. Although we have no reason to be certain
that
it must be zero, the values that are of interest to cosmologists
({\it i.e.}, the values that make $\Omega_\Lambda$ of order unity)
are unnaturally small from the point of view of particle physics
(Weinberg 1989), and so these $\Lambda$-models are often regarded
as unappealing.
The other class of low-density inflationary scenario
is the open model, in which the universe has
negative spatial curvature. Recently, the possibility
of realizing an open universe has been discussed
in the context of inflation theory (Bucher, Goldhaber,
{$\&$ } Turok 1994; Yamamoto, Sasaki, {$\&$ } Tanaka 1995;
Linde 1995; Linde {$\&$ } Mezhlumian 1995).
The essential idea is based on the semiclassical picture
of a bubble nucleation, which is described by a bounce
solution (Coleman {$\&$ } De~Luccia 1980).
One bubble nucleation process can be regarded as the creation of
a homogeneous and isotropic spacetime with negative spatial
curvature inside the bubble due to the $O(4)$-symmetry of the
bounce solution.
It is of great interest to ask whether an open universe created
in this one-bubble inflationary scenario is observationally
acceptable or not. Open CDM models have been investigated
by many authors (Lyth {$\&$ } Stewart 1990; Ratra {$\&$ } Peebles
1994ab; Sugiyama {$\&$ } Silk 1994; Kamionkowski et al. 1994;
G\'orski et al. 1995; Liddle et al. 1995).
Their investigations are based on the simple assumption
that the quantum state of a scalar field is in the conformal
vacuum state at the inflationary stage; however,
this is unlikely
to be the prediction of the one-bubble
inflationary scenario. It has been pointed out that
if we take the Bunch-Davies vacuum state as the state
of scalar field, the CMB
anisotropy in a low $\Omega_0$ universe appears quite different
due to the
super-curvature mode (Yamamoto, Sasaki, {$\&$ } Tanaka 1995).
Following the usual inflationary picture in which
the quantum fluctuation of a scalar field generates
the density perturbation, we must investigate the
quantum state of fields inside the bubble.
Attempts have been made to study this problem
by developing a field-theoretical formalism based on
a multidimensional tunneling wave function
(Tanaka, Sasaki, {$\&$ } Yamamoto 1994; Sasaki et al. 1994; Tanaka {$\&$ }
Sasaki 1994; Yamamoto, Tanaka, {$\&$ } Sasaki 1995; Hamazaki et al. 1995).
Bucher et al. have also considered this problem
(Bucher, Goldhaber {$\&$ } Turok 1994; Bucher {$\&$ } Turok 1995).
However, this problem requires further investigation.
In this paper, we consider the simple case in which the quantum
state of a scalar field is in the Bunch-Davies vacuum state,
and compare the predictions of this model with several cosmological
observations.
This situation is physically definite and clear, and as long
as
the bubble nucleation occurs in the de Sitter inflationary
background (where the initial inflationary period
has lasted sufficiently long), the field can be
approximated by the Bunch-Davies vacuum state, provided that
the effect of the bubble nucleation process is negligible.
In section 2, we first consider the initial spectrum by
solving for the evolution of cosmological perturbations in the
open inflationary stage.
In section 3 we use this initial power spectrum to
calculate various cosmological quantities and
compare these predictions
with observations.
We also compare these results with those of a previous
analysis of the open inflationary model based on the conformal vacuum state,
and a $\Lambda$-model with Harrison-Zel'dovich spectrum.
Section 4 is devoted to a discussion of our results.
We will work in units where $c=1$ and $\hbar=1$.
\section{Initial Conditions}
In this section, we consider the evolution of cosmological
perturbations in an open inflationary stage and derive the
initial spectrum of perturbations.
Ratra {$\&$ } Peebles (1994b) have investigated the evolution of
cosmological perturbations in an open inflationary universe
with gauge fixed. Bucher, Goldhaber {$\&$ } Turok (1994)
have also investigated cosmological perturbations in an
inflationary stage in the gauge-invariant formalism.
In the first half of this section, we follow the work by
Bucher, Goldhaber {$\&$ } Turok (1994).
In an open universe the line element can be written as
\begin{equation}
ds^2=a^2(\eta)\biggl[
-d\eta^2+d\chi^2+\sinh^2\chi d\Omega^2_{(2)}\biggr],
\label{metric}
\end{equation}
where $\eta$ is conformal time. We consider a scalar
field $\phi$ with potential $V(\phi)$, which drives
inflation. Writing the scalar field as a homogeneous part and
a small inhomogeneous part, {i.e.}, $\phi=\phi_0+\delta\phi$,
the equations for the homogeneous part are
\begin{equation}
\phi_0''+2{\cal H}\phi_0'+a^2 {\partial V\over\partial\phi}=0,
\end{equation}
and
\begin{equation}
{\cal H}^2 -1={8\pi G\over 3}
\biggl[{\phi_0'^2\over 2}+a^2V\biggr],
\end{equation}
where ${\cal H}:=a'/a$ and the prime
denotes differentiation with respect to $\eta$.
The small fluctuation $\delta\phi$ gives rise to a
metric perturbation. As we are interested in a
scalar perturbation in a scalar field dominated universe,
the metric perturbation can be written,
\begin{equation}
ds^2=a^2(\eta)\biggl[
-\bigl(1+2\Phi\bigr)d\eta^2+\bigl(1-2\Phi\bigr)
\biggl(d\chi^2+\sinh^2\chi d\Omega^2_{(2)}\biggr)
\biggr].
\end{equation}
$\Phi$ corresponds to the curvature perturbation or
gravitational potential. Then we get the evolution
equations for the perturbation
(Mukhanov, Feldman, {$\&$ } Brandenberger 1992),
\begin{eqnarray}
&&\Phi''+2\Bigl({\cal H}-{\phi_0''\over\phi_0'}\Bigr)\Phi'
+\Bigl(-{\bf L}^2+4
+2{\cal H}'-2{\cal H}{\phi_0''\over\phi_0'}\Bigr)\Phi=0,
\label{eqPhi}
\\
&&\delta\phi={1\over4\pi G\phi_0'}
\bigl(\Phi'+{\cal H}\Phi\bigr),
\label{connect}
\end{eqnarray}
where
\begin{equation}
{\bf L}^2:={1\over\sinh^2\chi}{\partial\over\partial\chi}
\biggl(\sinh^2\chi{\partial\over\partial\chi}\biggr)+
{1\over\sinh^2\chi}{\bf L}^2_\Omega,
\end{equation}
and ${\bf L}^2_\Omega$ is the Laplacian on the unit sphere.
Now let us consider an inflationary stage of the universe
inside a bubble. To solve the above equations analytically,
we assume that the potential is nearly flat, and
use the approximation, $V\simeq V_0+V'\phi$,
where $V'={\rm const}$.
We also assume that the background spacetime is approximated
by de Sitter spacetime, that is, $a(t)=-1/H\sinh\eta$.
Then the field equation for the homogeneous part is
\begin{equation}
\phi_0''-2{\rm coth}\,\eta\phi_0' =
{-V' \over H^2\sinh^2\eta},
\label{phi0eq}
\end{equation}
with $H^2={8\pi G V_0/3}$.
Eq. (\ref{phi0eq}) can be integrated, giving
\begin{equation}
\phi_0'={-V'\over H^2}
{-\cosh^3\eta+3\cosh\eta\sinh^2\eta+2\sinh^3\eta\over 3\sinh\eta},
\end{equation}
and the perturbation equation (\ref{eqPhi}) reduces to,
\begin{equation}
\Phi''-{6(1-e^{2\eta})\over 3-e^{2\eta}}\Phi'+
\biggl(-{\bf L}^2 +4-{4(3+e^{2\eta})\over3-e^{2\eta}}\biggr)
\Phi=0.
\label{eqPhib}
\end{equation}
To solve these equations, we need the initial values.
To determine the initial values, we must investigate
the problem of what the quantum state is inside the bubble.
As mentioned before, this is a very important problem,
which demands further investigation.
We here consider the case in which the scalar field
is in the Bunch-Davies vacuum state.
This is the case provided that the effect of bubble
nucleation process is small and negligible. But we should
keep in mind that this point needs examination in the various
models of one-bubble inflation scenario, taking into consideration
the effect of
bubble nucleation.
Recently, quantum field theory in de Sitter space-time
associated with the open chart has been investigated
(Sasaki, Tanaka, {$\&$ } Yamamoto 1995). According to this analysis,
a quantized scalar field with mass $m^2\ll H^2$ in the
Bunch-Davies Vacuum state is described in the second
quantized manner as,
\begin{equation}
\delta\phi=\int_0^\infty dp\sum_{\sigma,l,m}
\chi_{p,\sigma}(\eta)Y_{plm}({\chi},\Omega_{(2)})\hat a_{p\sigma lm} +
\sum_{lm}v_{(*)lm}(t,{\chi},\Omega_{(2)}) \hat a_{(*)lm} +
{\rm h.c.},
\end{equation}
where
\begin{eqnarray}
\chi_{\sigma,p}(\eta)&=&{-1\over\sqrt{8p(p^2+1)\sinh\pi p}}
\nonumber
\\
&&\biggl[e^{\pi p/2}(ip+\coth\eta)e^{-ip\eta}
+\sigma e^{-\pi p/2}(ip-\coth\eta)e^{ip\eta}
\biggr]{1\over a(\eta)},
\label{chi}
\end{eqnarray}
\begin{eqnarray}
v_{(*)lm}(t,\chi,\Omega_{(2)})
&=&{H\over2}\sqrt{\Gamma(l+2)\Gamma(l)}
{P^{-l-1/2}_{1/2}(\cosh\chi)\over \sqrt{\sinh\chi}}
Y_{lm}(\Omega_{(2)})
\nonumber
\\
&=:&{H\over2}W_{(*)l}(\chi)Y_{lm}(\Omega_{(2)}),
{\hspace{14mm}} (l>0),
\label{vstar}
\end{eqnarray}
$\sigma$ takes on the values $\pm1$, $\hat a$ is the annihilation
operator, and $\Gamma(z)$ is the gamma function.
The orthonormal harmonics on a three-dimensional unit
hyperboloid $Y_{plm}(\chi,\Omega)$ are
\begin{equation}
Y_{plm}(\chi,\Omega_{(2)})
=\Bigg\vert{p\Gamma(ip+l+1)\over\Gamma(ip+1)}\Bigg\vert
{P^{-l-1/2}_{ip-1/2}(\cosh\chi)\over \sqrt{\sinh\chi}}
Y_{lm}(\Omega_{(2)}),
\end{equation}
with normalization
\begin{equation}
\int_0^\infty d\chi\int d\Omega_{(2)}
\sinh^2\chi Y_{p_{1}l_{1}m_{1}}(\chi,\Omega_{(2)})
\overline{Y_{p_{2}l_{2}m_{2}}(\chi,\Omega_{(2)})}
=\delta(p_{1}-p_{2}) \delta_{l_{1}l_{2}} \delta_{m_{1}m_{2}}.
\end{equation}
In the above expression, the usual harmonics behave as
$Y_{plm}\propto e^{-\chi}$ at scales larger than the
curvature scale, $\chi\gg1$, although $v_{(*)lm}$ is constant for
$\chi\gg1$. Thus $v_{(*)lm}$ represents a fluctuation
larger than the curvature scale, so we call this mode a
super-curvature mode. The necessity of super-curvature modes
for a complete description of a random field in an open
universe has also been discussed by Lyth {$\&$ } Woszczyna
(1995).
Next, let us consider the curvature perturbation,
which can be written in the mode-expanded form
\begin{equation}
\Phi=\int_0^\infty dp \sum_{\sigma,l,m}
\Phi_{p,\sigma}(\eta) Y_{plm}(\chi,\Omega_{(2)})+
\sum_{l,m}\Phi_{(*)}(\eta)W_{(*)l}(\chi)Y_{lm}(\Omega_{(2)}).
\end{equation}
For the continuous mode $(p,l,m)$, Eq.(\ref{eqPhi})
reduces to
\begin{equation}
\Phi_p''-{6(1-e^{2\eta})\over 3-e^{2\eta}}\Phi_p'+
\biggl(p^2 +5-{4(3+e^{2\eta})\over3-e^{2\eta}}\biggr)
{\Phi_p}=0.
\end{equation}
The solution that behaves like
$\tilde\Phi_p\rightarrow e^{(1-ip)\eta}$ as
$\eta\rightarrow -\infty$ is (Bucher {$\&$ } Turok 1995)
\begin{equation}
\tilde\Phi_p=e^{(1-ip)\eta}\biggl(1+{1+ip\over 1-ip}
{e^{2\eta}\over3}\biggr).
\end{equation}
The equation for the super-curvature mode is
\begin{equation}
\Phi_{(*)}''-{6(1-e^{2\eta})\over 3-e^{2\eta}}\Phi_{(*)}'+
\biggl(4-{4(3+e^{2\eta})\over3-e^{2\eta}}\biggr)
{\Phi_{(*)}}=0,
\end{equation}
and we find a solution
\begin{equation}
\tilde\Phi_{(*)}=e^{2\eta}.
\end{equation}
To determine the amplitude of $\Phi$, we use
Eq.(\ref{connect}). From the behavior at
$\eta\rightarrow-\infty$, we find for the continuous mode
\begin{equation}
\Phi_{p,\sigma}(\eta)={2\pi G V'\over H}
{1\over\sqrt{8p(p^2+1)\sinh\pi p}}
\biggl\{
{e^{\pi p/2}}{1-ip\over2-ip}
\tilde\Phi_{p}
+\sigma
{e^{-\pi p/2}}{1+ip\over2+ip}
\tilde\Phi_{-p}\biggr\},
\end{equation}
and for the super-curvature mode
\begin{equation}
\Phi_{(*)}(\eta)={2\pi G V'\over H}
{1\over3}\tilde\Phi_{(*)}(\eta).
\end{equation}
We therefore have the following spectrum at the
end of inflation, by taking the limit $\eta\rightarrow0$,
\begin{eqnarray}
\Phi_p(0)^2
&:=&\lim_{\eta\rightarrow0}
\sum_{\sigma=\pm1} \Phi_{p,\sigma}(\eta)^2
\nonumber
\\
&=&\biggl({2\pi G V'\over H}\biggr)^2{\coth\pi p\over 2p(p^2+1)}
{p^2+1\over p^2+4}
\lim_{\eta\rightarrow0}
\Bigl\vert \tilde\Phi_p(\eta)\Bigr\vert{}^2
\nonumber
\\
&=&\biggl({4\pi G V'\over 3H}\biggr)^2
{\coth\pi p\over 2p(p^2+1)},
\label{specnm}
\end{eqnarray}
and
\begin{eqnarray}
\Phi_{(*)}(0)^2
&=&\lim_{\eta\rightarrow0}\Phi_{(*)}(\eta)^2
\nonumber
\\
&=&\biggl({4\pi G V'\over 3H}\biggr)^2{1\over4}.
\label{specsc}
\end{eqnarray}
For comparison, we also investigate the case when
the scalar field is assumed to be in the conformal
vacuum state (Lyth {$\&$ } Stewart 1990, Ratra {$\&$ } Peebles
1994b). In this case, the scalar field is written
\begin{equation}
\delta\phi=\int_0^\infty dp\sum_{l,m}
\chi_{p}(\eta)Y_{plm}({\chi},\Omega_{(2)})\hat b_{plm}
+{\rm h.c.} ~,
\end{equation}
where
\begin{equation}
\chi_{p}(\eta)={(ip+\coth\eta)e^{-ip\eta}\over\sqrt{2p(p^2+1)}}
{1\over a(\eta)}.
\end{equation}
After a similar analysis, we get the
spectrum of curvature perturbations at the end of inflation,
\begin{equation}
\lim_{\eta\rightarrow0}\Phi_p(\eta)^2
=\biggl({4\pi G V'\over 3H}\biggr)^2 {1 \over 2p(p^2+1)}.
\label{spev}
\end{equation}
The conformal vacuum case differs from the
Bunch-Davies vacuum case in two ways, the factor $\coth\pi p$
and the super-curvature mode.
Lyth {$\&$ } Stewart (1990) have investigated perturbations
in an open inflationary universe, and have given a relation
to relate the curvature perturbation and the scalar field
perturbation ${\cal R}\simeq -(H/\dot\phi)\delta\phi$,
though a paper justifying this relation has never been published.
But the above investigation shows the correctness of their result
on all scales except for the small difference of the
former coefficient.
\section{Observational Confrontations}
Now we start testing the predictions of the open universe
in the context of CDM cosmology with the initial conditions
obtained above. The matter-dominated open universe
has the line element (\ref{metric}) with $a(\eta)=\cosh\eta-1$.
In this section we use $\eta(>0)$ as the conformal time in the
matter-dominated universe.
\begin{center}
\underline{(1) CMB Anisotropies}
\end{center}
Let us first consider the CMB temperature fluctuation.
Having obtained an initial perturbation spectrum, we can
compute the temperature fluctuations in the gauge-invariant
formalism (Sugiyama {$\&$ } Gouda 1992).
As usual, we write the temperature
autocorrelation in the form,
\begin{equation}
C(\alpha)={1\over4\pi}\sum_{l}(2l+1)C_l{\rm P}_l(\cos\alpha).
\end{equation}
Figure 1(a) shows the power spectrum of temperature fluctuations,
$l(l+1)C_l\times10^{10}/2\pi$, for various values of $\Omega_0$
with initial conditions associated with the Bunch-Davies
vacuum state. We have taken $\Omega_B h^2=0.0125$ and Hubble
parameter $h=0.75$, $0.70$, $0.65$, $0.65$, $0.60$, for
$\Omega_0=0.1$, $0.2$, $0.3$, $0.4$, $0.5$, respectively,
to take the age problem into consideration.
Figure 1(b) shows the same quantities with the
conformal vacuum state (Kamionkowski et al. 1994; G\'orski et al. 1995).
For reference, we show the corresponding results for a $\Lambda$-model with
a Harrison Zel'dovich spectrum in Figure 1(c) (Sugiyama 1995).
Here the parameter $\Omega_B h^2=0.0128$ and $h=0.8$.
We also show the results of several CMB experiments, taken
from the paper by Scott, Silk {$\&$ } White (1995).
Open models may have trouble fitting the data near the ``Doppler
peak'' on degree scales, although assessing the significance of this
problem will require very careful investigation
(Ratra et al. 1995).
The differences between Fig.1(a) and Fig.1(b) at low multipoles
come almost entirely from the contribution of the super-curvature mode
(Yamamoto, Sasaki, {$\&$ } Tanaka 1995),
\begin{equation}
C_{(*)l}=\biggl({2\pi G V'\over 3H}\biggr)^2
\biggl\{{1\over3}f(\eta_{LS})W_{(*)l}(\eta_0-\eta_{LS})
+2\int_{\eta_{LS}}^{\eta_0}d\eta' {df(\eta')\over d\eta' }
W_{(*)l}(\eta_0-\eta') \biggr\}^2,
\end{equation}
where $\eta_{LS}$ and $\eta_0$ are the recombination time and the present
time, respectively, $W_{(*)l}$ is defined in Eq.(\ref{vstar}), and
$f(\eta)$ is the decay factor of the curvature perturbation,
\begin{equation}
f(\eta)=5~{\sinh^2\eta-3\eta\sinh\eta+4\cosh\eta-4\over(\cosh\eta-1)^3}.
\label{feta}
\end{equation}
The most accurate and reliable CMB anisotropy data at the present
time come from the COBE DMR experiment. In addition to providing us
with accurate estimates of the fluctuation amplitude, the data from
this experiment can be used to constrain the shape of the power spectrum.
We have used the two-year COBE data (Bennett et al.~1994)
to place constraints on open inflationary
models, following a procedure based on the Karhunen-Lo\`eve transform
(Bunn, Scott, \& White 1994; Bunn \& Sugiyama 1995; White \& Bunn 1995;
Bunn 1995ab). This procedure gives results that are generally
consistent with the spherical-harmonic technique devised by G\'orski
(1994).
We will now describe this procedure.
In inflationary cosmological models, the CMB anisotropy is a realization
of a Gaussian random field. If we expand the anisotropy in spherical
harmonics,
\begin{equation}
\Delta T(\hat{\bf r}) = \sum_{l=2}^\infty\sum_{m=-l}^l a_{lm}Y_{lm}
(\hat{\bf r}),
\end{equation}
then each coefficient $a_{lm}$ is an independent Gaussian random variable
of zero mean. Furthermore, the variance of $a_{lm}$ is simply $C_l$.
With this information, we can in principle compute the probability
density $p(\vec d \,|\, C_l)$ of getting the actual COBE data $\vec d$ given
a power spectrum $C_l$: since each data point is a linear combination
of Gaussian random variables, the probability distribution $\vec d$
is simply a multivariate Gaussian,
\begin{equation}
p(\vec d \,|\, C_l)\propto\exp\left(-{1\over 2}\vec d^TM^{-1}\vec d\right),
\label{GaussProb}
\end{equation}
where the covariance matrix $M$ can be written in terms of the
power spectrum and the noise covariance matrix.
Let us restrict our attention to a few-parameter family of possible
power spectra. We will denote the parameters generically by $\vec q$.
In this paper, for example, we will consider
power spectra that are parameterized by two parameters, the
density parameter $\Omega_0$ and the power spectrum normalization
$Q\equiv\sqrt{5C_2/4\pi}$, and so $\vec q$ will be a two-dimensional
vector. The probability density in
equation (\ref{GaussProb}) is then simply the probability density $p(\vec d\,|
\,\vec q)$
of the data $\vec d$ given the parameters $\vec q$.
If we adopt a Bayesian view of statistics,
we can convert this into a probability density for the parameters
given the data:
\begin{equation}
p(\vec q \,|\, \vec d)\propto p(\vec d \,|\, \vec q) p(\vec q),
\end{equation}
where $p(\vec q)$ is the prior probability density we choose to
adopt. $p(\vec q \,|\,\vec d)$ is generally denoted $L(\vec q)$
and called the likelihood.
The choice of prior distribution is a notoriously troublesome issue.
In practice, one generally chooses a prior
that is a smooth, slowly-varying function of the
parameters. In this paper, we will adopt a prior that is uniform
in $\Omega_0$ and one that is uniform in $\ln Q$. (This prior
is approximately equivalent to one that is uniform in the power spectrum
normalization $C_{10}$ near the ``pivot point.'' It differs slightly
from one that is uniform in $Q$, although not enough to affect our
results significantly.)
Unfortunately, in order to compute the probability density $p(\vec
d\,|\,\vec q)$, and hence the likelihood $L$, one must invert a matrix of
dimension equal to the number of data points. For the COBE DMR data,
this number is of order 4000. Such exact likelihoods have been
computed for a small class of models (Tegmark \& Bunn 1995); however,
this is quite a time-consuming procedure. The Karhunen-Lo\`eve
transform allows us to ``compress'' the data from 4000 numbers to only
400 in a way that throws away very little of the actual cosmological
signal. The likelihoods estimated from the transformed data
approximate the true likelihoods well, and are much more efficient to
compute. For details on how the Karhunen-Lo\`eve transform is
performed, see White \& Bunn (1995) and Bunn (1995ab).
Once we know $L$, it is quite easy to place constraints on the parameters
$\vec q$. Since $L$ is a probability distribution for $\vec q$, it
should be normalized so that
\begin{equation}
\int L(\vec q)\,d\vec q = 1.
\end{equation}
Now suppose that we choose some subset $R$ of possible parameter values.
Then if
\begin{equation}
\int_R L(\vec q)\,d\vec q=c
\end{equation}
then we can say that $\vec q$ lies in the region $R$ with probability
$c$. If we want to find a 95\% confidence interval, we simply
find a region $R$ such that $c=0.95$. One frequently chooses $R$
to be the region enclosed by a contour of constant likelihood.
If one of the parameters is deemed to be uninteresting, the standard
practice is to ``marginalize'' over it. For example, if we are interested
in constraining $\Omega_0$ but not $Q$, then we replace $L(\Omega_0,Q)$
by
\begin{equation}
L_{\rm marg}(\Omega_0)=\int L(\Omega_0,Q)\,dQ.
\end{equation}
This is a natural thing to do: if $L$ is the joint probability density
for $\Omega_0$ and $Q$, then $L_{\rm marg}$ is the probability density
for $\Omega_0$ alone.
Figure 2 shows the contours of the likelihood $L(\Omega_0,Q)$ for open models
associated with the Bunch-Davies vacuum state. In computing these
likelihoods, we use a linear combination of the 53 and 90 GHz
two-year COBE maps, with weights chosen to minimize the noise. We
use the ecliptic-projected maps; maps that were made in Galactic
coordinates, and therefore have different pixelization, give normalizations
that are generally lower by a few percent (Stompor, G\'orski \& Banday
1995; Bunn 1995a).
The choice of pixelization
appears to affect primarily the overall normalization of models;
likelihood ratios of models with power spectra of different shapes
are less affected (Bunn 1995ab).
Figure 3 shows the marginal likelihoods for $\Omega_0$ for both
the Bunch-Davies and the conformal vacuum open models.
In the Bunch-Davies case, we find that $\Omega_0>0.34$ at 95\%
confidence and $\Omega_0>0.15$ at 99\% confidence.
We also show the confidence levels for various $\Omega_0$
in Table 1.
For the conformal vacuum models, the likelihood is bimodal,
and so the allowed regions are not connected.
If we take a cut at small $\Omega_0$ and only consider the region
$\Omega_0\geq0.03$, we can state that at 95\% confidence
either $\Omega_0<0.085$ or $\Omega_0>0.36$, and 99\% confidence,
either $\Omega_0<0.14$ or $\Omega_0>0.23$.
There are difficulties associated with the interpretation of the
likelihoods in this case
(G\'orski et al. 1995).
Of course, the likelihood $L(\Omega_0,Q)$ provides us with accurate
normalizations in addition to shape constraints. For any particular
value of $\Omega_0$, we find the value of $Q$ that maximizes the
likelihood and use this value as the power spectrum normalization.
The normalizations determined in this way have typical one-sigma
fractional uncertainties of approximately $7.5\%$.
The maximum-likelihood normalizations computed in this way
are listed in the second column of Table 2(a) for the
Bunch-Davies vacuum case. For comparison, we have also
computed the conformal vacuum case; these normalizations are given in
the
second column of Table 2(b).
\vspace{3mm}
\begin{center}
\underline{(2) Linear density power spectrum}
\end{center}
We next consider the matter inhomogeneities
using the COBE normalization as described above. As the density
perturbation $\Delta$ is related to the curvature perturbation
$\Phi$ by the gravitational Poisson equation (Kodama {$\&$ } Sasaki 1984),
\begin{equation}
(p^2+4)\Phi_p(\eta)=4\pi G\rho(\eta) a^2(\eta)\Delta_p(\eta),
\end{equation}
in linear perturbation theory, we can write
the power spectrum of the matter perturbation in an open
universe from Eq.(\ref{specnm}),
\begin{eqnarray}
a_0^3P(k)~\Bigl(:=a_0^3\Delta_p^2\Bigr)
&=&\biggl({2(1-\Omega_0)\over 3\Omega_0}\biggr)^2
(p^2+4)^2a_0^3\Phi_p^2(0) f^2(\eta_0) T(k)^2
\nonumber
\\
&=:&{\cal A}(p^2+4)^2{\coth \pi p \over p(p^2+1)}T(k)^2,
\label{Pk}
\end{eqnarray}
where $p=a_0 k$, $a_0=1/H_0\sqrt{1-\Omega_0}$,
and $H_0=100h{\rm km/s/Mpc}$.
The super-curvature mode does not contribute
on small scales. The model based on the
conformal vacuum state leads to the same form
but without the factor $\coth\pi p$.
In the CDM cosmology, the following transfer function is useful
(Bardeen et al. 1986, Sugiyama 1995),
\begin{equation}
T(k)={\log(1+2.34q)\over 2.34q}
\Bigl[1+3.89q+(16.1q)^2+(5.46q)^3+(6.71q)^4\Bigr]^{-1/4},
\label{transfersu}
\end{equation}
with
\begin{equation}
q=\Big({2.726\over2.7}\Bigr)^2
{k\over\Omega_0h\exp(-\Omega_B-\sqrt{2h}\Omega_B/\Omega_0)}
~h{\rm Mpc}^{-1} .
\end{equation}
This transfer function is for a flat universe; however, since we are
interested in small-scale perturbations, the
curvature of the universe can be neglected and the
transfer function above is acceptable.
The COBE DMR normalization determines the amplitude of the
fluctuation. We give the numerical value of ${\cal A}$
determined from the likelihood normalization
in the second column in Table 2(a).
The second column in Table 2(b) gives the value for
the conformal vacuum case, in which the power spectrum
is obtained by Eq.(\ref{Pk}) without the factor $\coth\pi p$.
We show in Figure 4(a) the density power spectrum $a_0^3P(k)$,
for various $\Omega_0$, with initial conditions
based on the Bunch-Davies vacuum state. The Hubble parameter
is the same as that in Fig.1(a).
The points are from Peacock {$\&$ } Dodds (1994).
Figure 4(b) shows the density power spectra for $\Lambda$-models with
Harrison-Zel'dovich power spectrum, in which the
Hubble parameter is same as that in Fig.1(c).
Given the density perturbation spectrum $P(k)$, we are able to
calculate $\sigma_8^2=$$(\delta M$$/M)^2_{8 h^{-1}{\rm Mpc}}$,
the variance of the mass fluctuation in a
sphere of a radius $R=8h^{-1}{\rm Mpc}$,
\begin{equation}
\sigma^2(R)={1\over2\pi^2}\int k^2dk P(k) W^2(kR),
\end{equation}
where the top-hat window function $W$ is defined by
$W(x)=3(\sin x -x\cos x)/x^3$.
In Tables 2(a) and 2(b), we give
$\sigma_8$ for various $\Omega_0$ and $ h$ in the open model
associated with Bunch-Davies vacuum and the conformal vacuum respectively.
We also show values for the $\Lambda$ model
in Table 2(c) (Sugiyama 1995; Stompor et al. 1995).
The difference between the Bunch-Davies case
and the conformal vacuum case is very small. The
difference is $10$ percent at $\Omega_0=0.05$, but
is a few percent even at $\Omega_0=0.1$. This is because
the COBE normalization based on a likelihood analysis gives
more weight to the behavior of the power spectrum at $l\simeq10$,
and less weight at lower multipoles (White {$\&$ } Bunn 1995).
When we take a $\sigma(10^\circ)$ normalization,
there is a $10$ percent difference between the two cases
at $\Omega_0=0.1$.
The values obtained are consistent with those in G\'orski et al.
(1995).
A precise comparison between predictions and observations
of the matter power spectrum is difficult. One of the primary
problems is that we do not know whether the galaxy distribution is
an unbiased tracer of the mass distribution. However, if we
make the reasonable assumptions that the galaxies are not anti-biased
and are not extremely strongly biased (say $b\equiv\sigma_8^{-1}\mathrel{\mathpalette\oversim<}
2.5$), then
these calculations suggest that $0.3\mathrel{\mathpalette\oversim<}\Omega_0\lsim0.5$. Note
that the values of $\Omega_0$ preferred by the COBE likelihood
analysis tend to be higher than this range; however, one might
be inclined to argue that a model with $\Omega_0\simeq 0.4-0.5$
passes both tests.
\vspace{3mm}
\begin{center}
\underline{(3)Large-scale bulk velocity}
\end{center}
Next, we consider large-scale bulk velocities,
which are given by the following expression,
\begin{equation}
v_R^2= {H_0^2a_0^2\over 2\pi^2} \Omega_0^{1.2} \int dk P(k) W(kR)^2
\exp(-k^2R_s^2),
\end{equation}
where $W(kR)$ is the window function, and $R_s=12h^{-1}{\rm Mpc}$
is the Gaussian smoothing length for comparison with the observational
data.
In Table 3, we have summarized the computation of $v_R^2$
with $R=40h^{-1}{\rm Mpc}$ for various $\Omega_0$ for open
models with the initial conditions associated with the Bunch-Davies vacuum
state. These results are consistent with those of G\'orski et al. (1995).
We can compare this results with the recent data from the POTENT
analysis (Dekel~1994; Liddle et al.~1995):
$v_{R=40h^{-1}{\rm Mpc}}=373\pm50{\rm km/s}$.
Large values of $\Omega_0$ clearly provide a better fit to the velocities.
It appears difficult
to reconcile models with $\Omega_0\mathrel{\mathpalette\oversim<} 0.3$ with these data; however,
it is difficult to make precise statistical statements based on these
observations.
As is discussed by Liddle et al. (1995),
this measurement of the bulk velocity contains additional uncertainty due to
cosmic variance. In addition, it is quite difficult to assess the
uncertainties and potential biases in the POTENT analysis, and
one should therefore be reluctant to draw firm conclusions
on the basis of such a comparison.
\vspace{3mm}
\begin{center}
\underline{(4) epoch of galaxy formation}
\end{center}
Liddle et al. (1995) have performed a detailed
investigation of abundances of galaxy clusters
and damped Lyman-alpha systems in open CDM models using
Press-Schechter theory.
In this paper we will rely on
a simple and rough estimate of the epoch of galaxy formation,
following the work of Gottl\"ober, M\"uchet, $\&$ Starobinsky (1994)
and Peter, Polarski, $\&$ Starobinsky (1994).
According to Press-Schechter theory, the fraction of the matter in
the universe which is in gravitationally bound objects above
a given mass $M_R$ at a redshift $z$ has the form
\begin{equation}
F(>M_R)={\rm erfc}\Bigl({\delta_c\over\sqrt{2}\sigma(M_R,z)}\Bigr),
\end{equation}
where
\begin{equation}
\sigma(M_R,z)=\sigma(R){1\over 1+z}
{f(\eta_0)\over f(\eta(z))},
\end{equation}
where $M_R=(4/3)\pi R^3\rho$, and $f(\eta)$ is the decay
factor of the curvature perturbation.
The choice of $\delta_c$ depends on the collapse model.
The spherical collapse of a top-hat perturbation
gives $\delta_c=1.69$, although non-spherical collapse models
suggest other values. Here let us consider the range
$(1.33<\delta_c<2)$ (Gottl\"ober et al. 1994).
Observations suggest that many galaxies seems to have formed at
$z=1$, then, assuming $F(>10^{12}{\rm M}_\odot) \mathrel{\mathpalette\oversim>} 0.1$ at $z=1$, we have
$\sigma(M_R=10^{12}{\rm M}_\odot,z=1)\mathrel{\mathpalette\oversim>} 2\pm0.4$.
Figure 5(a) shows a contour plot of $\sigma(M_R=10^{12}{\rm M}_\odot,z=1)$
in the $\Omega_0-h$ plane for the open model. Figure 5(b) is same but
for the $\Lambda$-model with a Harrison Zel'dovich spectrum.
If we take the age problem into consideration, it indicates
a lower bound $\Omega_0\mathrel{\mathpalette\oversim>} 0.4$ for the open model.
Note that the above estimate is very rough,
although a similar constraint has been obtained from
the exact estimation of cluster abundances (Liddle et al. 1995).
The bound is weaker in the $\Lambda$-model than in the open model.
\section{Discussion}
A low-density universe is well motivated from several dynamical
observations of galaxies and clusters. The simplest such
low-density models are those in which the universe is open.
In the context of inflation theory, however, we need a special idea
such as the
one-bubble inflationary scenario in order to produce an open universe.
In this paper, motivated by the one-bubble inflationary universe
scenario, we have examined the cosmological predictions based on the
assumption that the scalar field is initially in
the Bunch-Davies vacuum state.
The initial perturbation spectrum has been derived by
considering the evolution of perturbations in an open inflationary
stage. Then the CMB anisotropies and the matter inhomogeneities
have been examined.
As the first test, we have performed a likelihood analysis
for the CMB anisotropies by using the COBE DMR data.
Interestingly, the COBE likelihood analysis
gives severe constraints on the model. Models with
$\Omega_0\leq0.4$, $\Omega_0\leq0.5$ are excluded at confidence
levels of
$92\%$, $83\%$, respectively.
In a previous analysis associated with the conformal
vacuum state (G\'orski et al. 1995), the likelihood function
has another steep peak below $\Omega_0\lsim0.15$.
This complicates the statistical interpretation of the results
(G\'orski et al. 1995).
In the case of the Bunch-Davies vacuum state, no such peak
appears in the range of $\Omega_0$ we are interested in, and so the
likelihood analysis gives clear results.
The COBE likelihood analysis is therefore a powerful probe of
these open models.
We have used the
the COBE DMR maximum-likelihood normalization to predict the
amplitude of matter fluctuations. According to this
normalization method, there is little difference between the predictions
of the Bunch-Davies vacuum and conformal vacuum cases. Even for
the case $\Omega_0=0.1$, the discrepancy of $\sigma_8$
is a few percent. We obtain results that are similar to previous
analyses:
the power spectrum of the mass fluctuation fits the observations of
galaxies and clusters for
$0.3\mathrel{\mathpalette\oversim<}\Omega_0\lsim0.5$.
The required bias is unacceptably high
for $\Omega_0\lsim0.1$, while high values of $\Omega_0$
demand anti-biasing. For example, $\Omega_0\gsim0.6$
needs anti-biasing when $h=0.65$.
On the other hand, the $\Lambda$-models with Harrison Zel'dovich
spectrum have higher amplitude compared with open
models. The $\Lambda$-model needs anti-biasing for $\Omega_0\gsim0.4$
even when $h=0.65$.
The $\Lambda$-model therefore needs low $h$ or a tilted spectrum
(Ostriker {$\&$ } Steinhardt 1995).
For the range $0.3\mathrel{\mathpalette\oversim<}\Omega_0\lsim0.5$, open models
give unacceptably small bulk velocities compared with the
POTENT analysis. However, given the present quality of the velocity
data and
the problem of cosmic variance, one might be reluctant to draw
strong conclusions from this fact.
The rough estimation associated with the
galaxy formation gives lower bound of $\Omega_0$
consistent with the value discussed above.
It is very interesting that the COBE likelihood analysis
has given the most severe constraint on this open model.
The COBE likelihood analysis strongly prefers a high value of $\Omega_0$.
The peak value is around $0.7\mathrel{\mathpalette\oversim<}\Omega_0\lsim0.8$, and we can
state that
$\Omega_0\geq0.5$ with $83\%$ confidence in this model.
Considering both the COBE analysis and the matter inhomogeneity,
we are led to prefer a value of $\Omega_0\simeq 0.5$ if the
one-bubble inflationary scenario is correct. Such a model is
consistent with the Press-Schechter analysis of the epoch of galaxy
formation and is marginally consistent with the bulk velocity data.
As the CMB data continue to improve, particularly on degree scales,
we should be able to test this model.
It is premature to rule out low $\Omega_0$ inflationary models on the
basis of this
investigation at present, because we do not include
the effect of bubble nucleation in the calculation of
initial density power spectrum.
Previous analysis indicates that the bubble nucleation effect
in general excites fluctuations, and amplifies the perturbations on
scales larger than curvature scale (Yamamoto, Tanaka, {$\&$ } Sasaki 1995;
Hamazaki et al. 1995). One might therefore expect low-density models
to fit the COBE data even more poorly once this effect is taken
into account; however, since the calculation has not been done, we
cannot be certain. In particular, the status of the super-curvature
mode is still quite uncertain.
Various modifications of the open model may also be viable.
We must investigate the effect of gravity waves
in an open inflationary universe. One might also consider the effect of
tilting the primordial power spectrum; however, in order
to improve the fit to the data one would probably
need to tilt the power spectrum to have increased power on small scales,
and such ``blue'' power spectra are not naturally produced by
inflation. Such a model is probably too contrived to be plausible.
\vspace{3mm}
\begin{center}
\bf Acknowledgments
\end{center}
We would like thank N. Sugiyama for providing us with the
CMB anisotropy power spectra and for helpful discussions
and comments.
We are grateful to M. Sasaki and T. Tanaka for discussions and comments.
One of us (K.Y.) would like to thank Professor J. Silk
and the people at the Center for Particle Astrophysics,
University of California, Berkeley, where many parts of this
work were done for their hospitality.
He would like to thank Professor H. Sato for continuous
encouragement.
This work was supported in part by Ministry of Education
Grant-in-Aid for Scientific Research No. 2841.
\vspace{10mm}
\begin{center}
\bf References
\end{center}
\noindent
Bardeen, J. M., Bond, J. R., Kaiser, N., {$\&$ } Szalay, A. S.
1986, ApJ, {\bf 304}, 15.
\\
Bennett, C.L, et al. 1994, ApJ, {\bf 436}, 423.
\\
Bucher, M., Goldhaber, A. S., {$\&$ } Turok, N. 1994, PUTP-1507,
hep-ph/9411206.
\\
Bucher, M., {$\&$ } Turok, N. 1995, PUTP-1518, hep-ph/9503393.
\\
Bunn, E.F. 1995a, Ph.D. dissertation, U.C. Berkeley Physics Department.
\\
Bunn, E.F. 1995b, in preparation.
\\
Bunn, E.F., Scott, D., \& White, M. 1995, ApJ, {\bf 441}, L9.
\\
Bunn, E.F., \& Sugiyama, N. 1995, ApJ, in press.
\\
Coleman, S., {$\&$ } De Luccia, F. 1980, Phys. Rev. {\bf D21}, 3305.
\\
Dekel, A. 1994 ARA $\&$ A, 32, 371.
\\
G\'orski, K. M. 1994, ApJ, {\bf 430}, L85.
\\
G\'orski, K. M., Ratra, B., Sugiyama, N., {$\&$ } Banday, A. J.
1995, astro-ph/9502034, ApJ, in press.
\\
Gottl\"ober, S., M\"uchet, J. P., {$\&$ } Starobinsky, A. A. 1994,
ApJ, {\bf 434}, 417.
\\
Hamazaki, T., Sasaki, M., Tanaka, T., {$\&$ } Yamamoto, K. 1995,
KUNS1340.
\\
Kamionkowski, M., Ratra, B., Spergel, D. N., {$\&$ } Sugiyama, N.
1994, ApJ, {\bf 434}, L1.
\\
Kashlinsky, A., Tkachev, A. A., {$\&$ } Frieman, J. 1994,
Phys. Rev. Lett., {\bf 73}, 1582.
\\
Kodama, H., {$\&$ } Sasaki, M. 1984, Prog. Theor. Phys. Suppl. {\bf 78}, 1.
\\
Liddle, A. R., Lyth, D. H., Roberts, D., {$\&$ } Voana, P. T. 1995,
SUSSEX-AST 95/6-2, astro-ph/9506091, MNRAS, in press.
\\
Linde, A. 1995, preprint SU-ITP-95-5, hep-th/9503097,
Phys. Lett. {\bf B}, in press.
\\
Linde, A., {$\&$ } Mezhlumian, A. 1995, preprint SU-ITP-95-11,
astro-ph/9506017.
\\
Lyth, D. H., {$\&$ } Stewart, E. D. 1990, Phys. Lett. {\bf B252}, 336.
\\
Lyth, D. H., {$\&$ } Woszczyna, A. 1995, Lancaster preprint, astro-ph/9408069.
\\
Mukhanov, V. F., Feldman, H. A., {$\&$ } Brandenberger, R. H. 1992,
Phys. Rep. {\bf 215}, 203.
\\
Ostriker, J. P., {$\&$ } Steinhardt, P. J. 1995, astro-ph/9505066.
\\
Peacock, J. A., {$\&$ } Dodds, S. J. 1994, MNRAS, 267, 1020.
\\
Peebles, P. J. E. 1993, {\it Principle of Physical Cosmology},
Princeton Univ. Press.
\\
Peter, P., Polarski, D., {$\&$ } Starobinsky, A. A. 1994, Phys. Rev.
{\bf D50}, 4827.
\\
Ratra, B., et al. 1995, in preparation.
\\
Ratra, B., {$\&$ } Peebles, P. J. E. 1994a, ApJ, {\bf 432}, L5.
\\
Ratra, B., {$\&$ } Peebles, P. J. E. 1994b, preprint PUPT-1444.
\\
Sasaki, M., Tanaka, T., Yamamoto, K., {$\&$ } Yokoyama, J. 1994,
Prog. Theor. Phys. 90, 1019.
\\
Sasaki, M., Tanaka, T., {$\&$ } Yamamoto, K. 1995, Phys. Rev.
{\bf D51}, 2979.
\\
Scott, D., Silk, J., {$\&$ } White, M. 1995, preprint.
\\
Smoot, G. F. et al. 1995, ApJ. {\bf 396}, L1.
\\
Stompor, R., G\'orski, K. M., {$\&$ } Banday, A. J. 1995, preprint.
\\
Sugiyama, N. 1995, preprint CfPA-TH-94-62.
\\
Sugiyama, N., {$\&$ } Gouda, N. 1992, Prog. Theor. Phys., 88, 803.
\\
Sugiyama, N., {$\&$ } Silk, J. 1994, Phys. Rev. Lett., {\bf 73}, 509.
\\
Tanaka, T., Sasaki, M., {$\&$ } Yamamoto, K. 1994,
Phys. Rev. {\bf D49}, 1039.
\\
Tanaka, T., {$\&$ } Sasaki, M. 1994, Phys. Rev. {\bf D50}, 6444.
\\
Tegmark, M., \& Bunn, E. F. 1995, Berkeley preprint.
\\
Weinberg, S. 1989, Rev. Mod. Phys., 61, 1.
\\
White, M., {$\&$ } Bunn, E. F. 1995, CfPA-95-TH-02, ApJ, in press.
\\
White, M., Scott, D., Silk, J., \& Davis, M. 1995, preprint.
\\
Yamamoto, K., Sasaki, M., {$\&$ } Tanaka, T. 1995, KUNS-1309,
ApJ, in press.
\\
Yamamoto, K., Tanaka, T., {$\&$ } Sasaki, M. 1995, Phys. Rev.
{\bf D51}, 2968.
\def{Q}{{Q}}
\def{\mu {\rm K}}{{\mu {\rm K}}}
\vspace{12mm}
\centerline{Table 1}
\begin{centerline}
{Confidence levels for Bunch-Davies open model }
\end{centerline}
\begin{center}
\begin{tabular}{c@{\hspace{1pc}}c}
\hline\hline
{\ \ $\Omega_0(>)$ \ \ }&{confidence level $(\%)$}
\\
\hline
$ 0.1$ & $99.4$ \\
$ 0.2$ & $98.4$ \\
$ 0.3$ & $96.2$ \\
$ 0.4$ & $91.8$ \\
$ 0.5$ & $83.2$ \\
\hline
\end{tabular}
\end{center}
\vspace{2mm}
\centerline{Table 2(a)}
\begin{centerline}
{Amplitude of density perturbation for Bunch-Davies open model }
\end{centerline}
\begin{center}
\begin{tabular}{c@{\hspace{1pc}}c@{\hspace{1pc}}c@{\hspace{1pc}}c@{\hspace{1pc}}c}
\hline\hline
{\ \ $\Omega_0$ \ \ }&{$\cal A$} & {\ \ } & {$\sigma_8$} & {\ \ }
\\
{\ \ } & {$(h^{-1}{\rm Mpc})^3$} & {h=0.5} & {h=0.65} & { h=0.8}
\\ \hline
$ 0.05$ & $1.96({Q}/27.9{\mu {\rm K}})^2~\times10^{2}$ & $0.0041$ & $0.011$& $0.021$ \\
$ 0.1 $ & $2.42({Q}/28.8{\mu {\rm K}})^2~\times10^{2}$ & $0.032$ & $0.063$ & $0.099$ \\
$ 0.2 $ & $2.83({Q}/27.8{\mu {\rm K}})^2~\times10^{2}$ & $0.15$ & $0.25$ & $0.35$ \\
$ 0.3 $ & $2.94({Q}/25.8{\mu {\rm K}})^2~\times10^{2}$ & $0.31$ & $0.48$ & $0.66$ \\
$ 0.4 $ & $2.90({Q}/23.4{\mu {\rm K}})^2~\times10^{2}$ & $0.49$ & $0.74$ & $1.00$ \\
$ 0.5 $ & $2.72({Q}/21.1{\mu {\rm K}})^2~\times10^{2}$ & $0.69$ & $1.01 $ & $1.33$ \\
$ 0.6 $ & $2.43({Q}/19.3{\mu {\rm K}})^2~\times10^{2}$ & $0.89$ & $1.26 $ & $1.63$ \\
$ 0.7 $ & $2.03({Q}/18.3{\mu {\rm K}})^2~\times10^{2}$ & $1.05$ & $1.45 $ & $1.90$ \\
$ 0.8 $ & $1.54({Q}/18.3{\mu {\rm K}})^2~\times10^{2}$ & $1.20$ & $1.66 $ & $2.10$ \\
$ 0.9 $ & $1.00({Q}/18.9{\mu {\rm K}})^2~\times10^{2}$ & $1.31$ & $1.80 $ & $2.25$ \\
\hline
\end{tabular}
\end{center}
\newpage
\centerline{Table 2(b)}
\begin{centerline}
{Amplitude of density perturbation for conformal vacuum open model }
\end{centerline}
\begin{center}
\begin{tabular}{c@{\hspace{1pc}}c@{\hspace{1pc}}c@{\hspace{1pc}}c@{\hspace{1pc}}c}
\hline\hline
{\ \ $\Omega_0$ \ \ } & {$\cal A$} & {\ \ } & {$\sigma_8$} & {\ \ }
\\
{\ \ } & {$(h^{-1}{\rm Mpc})^3$} & {h=0.5} & {h=0.65} & { h=0.8}
\\ \hline
$ 0.05$ & $2.45({Q}/18.7{\mu {\rm K}})^2~\times10^{2}$ & $0.0045$ &$0.012$ & $0.023$ \\
$ 0.1 $ & $2.54({Q}/23.1{\mu {\rm K}})^2~\times10^{2}$ & $0.032$ & $0.064$ & $0.10$ \\
$ 0.2 $ & $2.89({Q}/26.5{\mu {\rm K}})^2~\times10^{2}$ & $0.15$ & $0.25$ & $0.36$ \\
$ 0.3 $ & $3.01({Q}/25.9{\mu {\rm K}})^2~\times10^{2}$ & $0.31$ & $0.49$ & $0.67$ \\
$ 0.4 $ & $2.96({Q}/23.5{\mu {\rm K}})^2~\times10^{2}$ & $0.50$ & $0.75$ & $1.01$ \\
$ 0.5 $ & $2.77({Q}/20.6{\mu {\rm K}})^2~\times10^{2}$ & $0.69$ & $1.02$ & $1.34$ \\
$ 0.6 $ & $2.46({Q}/18.3{\mu {\rm K}})^2~\times10^{2}$ & $0.88$ & $1.27$ & $1.64$ \\
$ 0.7 $ & $2.04({Q}/17.2{\mu {\rm K}})^2~\times10^{2}$ & $1.06$ & $1.50$ & $1.90$ \\
$ 0.8 $ & $1.55({Q}/17.5{\mu {\rm K}})^2~\times10^{2}$ & $1.20$ & $1.67$ & $2.10$ \\
$ 0.9 $ & $1.00({Q}/18.7{\mu {\rm K}})^2~\times10^{2}$ & $1.31$ & $1.80$ & $2.25$ \\
\hline
\end{tabular}
\end{center}
\vspace{5mm}
\centerline{Table 2(c)}
\begin{centerline}
{Amplitude of density perturbation for Harrison-Zel'dovich $\Lambda$-model }
\end{centerline}
\begin{center}
\begin{tabular}{c@{\hspace{1pc}}c@{\hspace{1pc}}c@{\hspace{1pc}}c}
\hline\hline
{\ \ $\Omega_0$ \ \ } & {\ \ } & {$\sigma_8$} & {\ \ }
\\
{\ \ } & {h=0.5} & {h=0.65} & { h=0.8}
\\ \hline
$ 0.1 $ & $0.15$ & $0.29$ & $0.46$ \\
$ 0.2 $ & $0.41$ & $0.68$ & $0.98$ \\
$ 0.3 $ & $0.65$ & $1.0$ & $1.4$ \\
$ 0.4 $ & $0.85$ & $1.3$ & $1.7$ \\
$ 0.6 $ & $1.2 $ & $1.7$ & $2.2$ \\
\hline
\end{tabular}
\end{center}
\vspace{5mm}
\centerline{Table 3}
\begin{centerline}
{Large scale bulk velocity for Bunch-Davies open model }
\end{centerline}
\begin{center}
\begin{tabular}{c@{\hspace{1pc}}c@{\hspace{1pc}}c@{\hspace{1pc}}c}
\hline\hline
{\ \ $\Omega_0$ \ \ } & {\ \ } &
{$v_{R=40h^{-1}{\rm Mpc}} ({\rm km/s})$ } & {\ \ }
\\
{\ \ \ \ } & {h=0.5} & {h=0.65} & { h=0.8}
\\ \hline
$ 0.05$ & $6.3$ & $8.6$ & $ 11$ \\
$ 0.1 $ & $ 21$ & $ 29$ & $ 37$ \\
$ 0.2 $ & $ 71$ & $ 90$ & $106$ \\
$ 0.3 $ & $130$ & $160$ & $180$ \\
$ 0.4 $ & $200$ & $230$ & $260$ \\
$ 0.5 $ & $260$ & $300$ & $330$ \\
$ 0.6 $ & $320$ & $360$ & $390$ \\
$ 0.7 $ & $370$ & $410$ & $450$ \\
$ 0.8 $ & $410$ & $450$ & $490$ \\
$ 0.9 $ & $440$ & $480$ & $510$ \\
\hline
\end{tabular}
\end{center}
\newpage
\begin{center}
\underline{Figure Captions}
\end{center}
\noindent
{Figure 1.}
Power spectra of the CMB temperature anisotropy
$l(l+1)C_l\times10^{10}/2\pi$ for (a) the Bunch-Davies vacuum
open model with $\Omega_0=0.1,0.2,0.3,0.4,0.5$.
The data were provided by N. Sugiyama.
These theoretical curves are normalized by the COBE likelihood.
The curves are in descending order of $\Omega_0$ as
one moves down at $l=50$.
The results of several CMB experiments are also shown,
taken from the paper by Scott, Silk, {$\&$ } White (1995).
To compare with degree-scale observations, careful
investigations are required (Ratra et al. 1995).
\vspace{3mm}
\noindent
{Figure 1(b).} CMB power spectra for the conformal vacuum
open model (G\'orski et al. 1995). The curves and points are as in
Figure 1(a).
\vspace{3mm}
\noindent
{Figure 1(c).} Power spectrum of the CMB temperature anisotropy
$l(l+1)C_l\times10^{10}/2\pi$ for the Harrison-Zel'dovich
$\Lambda$-model with $\Omega_0=0.1,0.2,0.3,0.4,1.0$.
(Sugiyama 1995)
\vspace{3mm}
\noindent
{Figure 2.} Contour plot of the likelihood function $L(\Omega_0,Q)$
for the Bunch-Davies open model. The contour range is from
$L=0.25$ to $L=1.5$, where the likelihoods are scaled
so that $L=1$ corresponds to a flat Harrison-Zel'dovich
spectrum with maximum-likelihood normalization.
\vspace{3mm}
\noindent
{Figure 3.} The marginal likelihood $L_{\rm marg}(\Omega_0)$
as a function of $\Omega_0$ for both
the Bunch-Davies open model (solid line)
and the conformal vacuum open model (dashed line).
\vspace{3mm}
\noindent
{Figure 4(a).} Power spectrum of density perturbation $a_0^3 P(k)$
for the open model with $\Omega_0=0.1$, $0.2$, $0.3$, $0.4$, $0.5$, for
$h=0.75, 0.70, 0.65, 0.65, 0.60$, respectively.
We have taken $\Omega_Bh^2=0.0125$.
The points are from Peacock {$\&$ } Dodds (1994).
\vspace{3mm}
\noindent
{Figure 4(b).} Same figure as Fig.4(a) but for the Harrison-Zel'dovich
$\Lambda$-model with $\Omega_0=$ $0.1$, $0.2$, $0.3$, $0.4$, $1.0$.
We have taken $\Omega_Bh^2=$$0.0128$ and $h=0.8$.
\vspace{3mm}
\noindent
{Figure 5(a).} Contours of $\sigma(M_R=10^{12} {\rm M}_\odot,z=1)$,
in the $(\Omega_0-h)$ plane for the Bunch-Davies open models.
The contour range is from
$\sigma=1.0$ to $\sigma=6.0$. The dashed lines are $\sigma=1.6$ and
$2.4$.
\vspace{3mm}
\noindent
{Figure 5(b).} Same figure as Fig.5(a) but for
Harrison-Zel'dovich $\Lambda$-models.
\end{document}
|
\section*{\large\bf 1. Introduction}
In molecular dynamics simulations of
heavy-ion collisions at intermediate energies (300 MeV/u -- 2 GeV/u)
it is still an open question
question, which kind of forces should be used for the long range part of the
nucleon-nucleon interaction which builds up the mean field. The Skyrme forces,
which work very well for lower energies are not relativistically invariant.
There are attempts to extend non-relativistic two-body potentials
to higher velocities by requiring approximate Lorentz
invariance~\cite{Fel1,Schm89,Stac76,Bodm80}.
Our aim is to derive from a
relativistic lagrangian of the Walecka type a covariant Lagrange function
for point-like particles
which is suited for molecular dynamics and still has the
saturation property of the original lagrangian. It should then be
applicable in transport theories as for example
QMD or BUU and should give reasonable results even at lower energies (50 --
200 MeV/u).
In Section~2 the relativistic Hamilton function is derived in the small
acceleration approximation \cite{Weber} from a many-body Lagrange function
and it is expressed in terms of positions
and canonical momenta of the particles.
In Section~3 the QMD procedure is described, while in
Section~4 we present our QMD calculations for fragment distributions in
O + Br collisions at different energies and compare the calculated results with
the experimental data. A few years ago the analogue calculation was
done~\cite{kn:judqmd} with
the Skyrme interaction, which is explicitly density dependent. It is
interesting to compare
those results with the new forces presented in this paper in order to see how
far they can describe saturation and low energy phenomena.
The intention is, however, to go to intermediate energies, where QMD is more
appropriate because quantum aspects like the Pauli principle or the uncertainty
relation are expected to be less important.
\section*{\large\bf 2. Relativistic equations of motion for nucleons}
In this section we are deriving the relativistic equations of motion
for the mean 4-positions and the mean 4-momenta of the nucleons.
We start from a field theoretical lagrangian of the following type
\begin{eqnarray}\label{Lagr1}
{\cal L}(x) & = & \bar{\psi}(x) \Big( \gamma^\alpha i\partial_\alpha
- m^*\left( \phi(x)\right) \Big)
\psi(x) \nonumber \\
& - & g_{\rm v} \bar{\psi}(x) \gamma^\alpha \psi (x) A_\alpha (x) \nonumber \\
& - & \frac{1}{2} \phi(x) (\partial_\alpha \partial^\alpha + \mu^2_{\rm s})
\phi (x) \nonumber \\
& + & \frac{1}{2} A^\alpha(x) (\partial_\beta \partial^\beta + \mu^2_{\rm v})
A_\alpha (x) \ ,
\end{eqnarray}
where we allow for two different ways to couple the scalar
field~\cite{Lindner}. If
$m^*(\phi) = m - g_{\rm s} \phi$ we obtain the Walecka lagrangian, whereas for
$m^*(\phi) = \left( m + g_{\rm s} \phi \right)^{-1}$ we deal with the
lagrangian proposed by Zim\'anyi and Moszkowski~\cite{kn:zimmos}. (In the
following, if not expressed explicitly otherwise, we use units such that
$\ \hbar = c = 1$.)
The field equations are
\begin{equation}\label{Dirac}
\Big\{(i \partial_\alpha - g_{\rm v} A_\alpha(x)) -
m^*\left( \phi(x)\right)\Big\}\ \psi(x) = 0
\end{equation}
for the nucleons,
\begin{equation}\label{Wavephi}
(\partial^\beta \partial_\beta + \mu^2_{\rm s})\ \phi(x) =
-\ {d \over{ d \phi }} m^* \left( \phi(x) \right) \
\bar{\psi}(x)\psi(x)
\end{equation}
for the scalar field and
\begin{equation}\label{WaveA}
(\partial^\beta \partial_\beta + \mu_{\rm v}^2)\ A^\alpha(x) =
g_{\rm v} \, \bar{\psi}(x) \gamma^\alpha \psi(x)
\end{equation}
for the vector field.
In the mean-field approximation the source terms are replaced
by their expectation values
$\bar{\psi}(x)\psi(x) \Rightarrow \rho_{\rm s}(x)$
and
$\bar{\psi}(x) \gamma^\alpha \psi(x) \Rightarrow j^\alpha(x)$
so that the scalar and vector fields
become classical fields. For classical fields, however,
an arbitrarily weak time dependence in the scalar density
$\rho_{\rm s}(x)$ or current density $j^\alpha(x)$ results in a radiation
field which travels away from the nucleons. The total field energy
in the radiation may even be less than the masses of the field quanta.
Since the fields are only effective fields, and for the scalar field there
is no corresponding elementary particle, the physical meaning of this
radiation is obscure.
As we are using the relativistic scalar and vector fields only
for the mean-field part of the nuclear interaction we follow the
suggestion
\cite{Fel1,Fel2} to exclude radiation of fictitious mesons
from the very beginning by
using the action-at-a-distance formulation with the
symmetric Green's function of Wheeler and Feynman \cite{Wheeler}.
\begin{equation}
G_{\rm s,v}(x-y) = \frac{1}{2}
\Big( G^{advanced}_{\rm s,v} (x-y)+ G^{retarded}_{\rm s,v} (x-y)\Big) \
\end{equation}
The formal solution of the wave Eq.~(\ref{Wavephi})
(for $m^* = m - g_{\rm s} \phi$)
\begin{equation}\label{Phigreens}
\phi(x) = g_{\rm s} \int\! d^4 y \
G_{\rm s} (x-y) \bar{\psi}(y)\psi(y)
\end{equation}
and of Eq. (\ref{WaveA})
\begin{equation}\label{Agreens}
A^\alpha(x) = g_{\rm v} \int\! d^4 y \
G_{\rm v} (x-y) \bar{\psi}(y)\gamma^\alpha\psi(y)
\end{equation}
fulfills the desired boundary condition of vanishing incoming and outgoing
free fields.
Eliminating the fields from the lagrangian (\ref{Lagr1}) leads to the
non-local action which contains only nucleon variables:
\begin{eqnarray}\label{Action}
\int d^4x\ {\cal L}(x) & = &
\int d^4x\ \bar{\psi}(x) (\gamma^\alpha i\partial_\alpha - M)\psi(x)
\nonumber \\
&+& \frac{1}{2}g_{\rm s}^2 \int d^4x\ d^4y\ \bar{\psi}(x)\psi(x)\,
G_{\rm s}(x-y)\,
\bar{\psi}(y)\psi(y) \nonumber \\
&-& \frac{1}{2}g_{\rm v}^2 \int d^4x\ d^4y\ \bar{\psi}(x)\gamma^\alpha\psi(x)
\,G_{\rm v}(x-y)\, \bar{\psi}(y)\gamma_\alpha\psi(y)
\end{eqnarray}
If quantum effects are negligible the scalar density and vector current density
can be represented in terms of the world lines of the nucleons as
\begin{eqnarray}\label{rhos}
\rho_{\rm s}(x):= \bar{\psi}(x)\psi(x) &\rightarrow&
\sum_{\rm j=1}^{\rm A}
\int\! d \tau^{\ }_{\rm j} \ \delta^4 (x-r_{\rm j}(\tau^{\ }_{\rm j})) \\
j^\mu(x):= \bar{\psi}(x)\gamma^\mu\psi(x) &\rightarrow&
\sum_{\rm j=1}^{\rm A}
\int\! d \tau^{\ }_{\rm j} \ \delta^4 (x-r_{\rm j}(\tau^{\ }_{\rm j}))\,
u^\mu_{\rm j}(\tau^{\ }_{\rm j})\ ,
\label{rhos2}
\end{eqnarray}
where $r_{\rm j}(\tau^{\ }_{\rm j})$ denotes the world line of the nucleon
$j$ at its proper time $\tau^{\ }_{\rm j}$ and $u_{\rm j}(\tau^{\ }_{\rm j})$ its
4-velocity. Nucleons in nuclei can however not be localized well enough for
treating them as classical particles on world lines. The Fermi momentum sets a
limit to the radius of the nucleon wave packet in coordinate space which is of
the order of 2 fm. These effects, the Pauli principle and other quantum
effects, are properly taken care of in Antisymmetrized Molecular Dynamics
(AMD)~\cite{kn:amd} and in Fermionic Molecular Dynamics (FMD)~\cite{kn:fmd}.
However, presently neither the AMD nor the FMD equations of motion can be
solved numerically for large systems like gold on gold. Therefore, we mimic the
finite size effect of the wave packets by folding the finally obtained forces
with gaussian density distributions for the nucleons. This "smearing" provides
also the prescription how to eliminate strong forces at short distances which
are taken care of by the random forces of the collision term.
Therefore, the forces to be derived in the following are only meant to
represent the long range part of the nuclear interactions which constitute a
mean field.
In this sense $r_{\rm j}(\tau^{\ }_{\rm j})$ is the mean world line of the nucleon and
$u_{\rm j}(\tau^{\ }_{\rm j})$ its mean 4-velocity.
Using the representation (\ref{rhos})
the scalar field at a space-time point
$x$, for example is given by integrals over past and future proper
times $\tau^{\ }_{\rm j}$ as
\begin{equation}\label{Phiworld}
\phi(x) = g_{\rm s} \sum_{\rm j=1}^{\rm A}
\int\! d\tau^{\ }_{\rm j} \ G_{\rm s}(x-r_{\rm j}(\tau^{\ }_{\rm j})) \ .
\end{equation}
By identifying the following expressions with their
representation in terms of mean positions and mean momenta of the
nucleons
\begin{eqnarray}
\int d^4x\ d^4y\ \bar{\psi}(x)\psi(x)\,G_{\rm s}(x-y)\,
\bar{\psi}(y)\psi(y) &\rightarrow& \nonumber \\
&& \mbox{\hspace*{-6cm}}
\sum^{\rm A }_{\stackrel{{\scriptstyle \rm i,j=1}}{\rm i\neq j}}
\int\! d\tau_i d\tau^{\ }_{\rm j}\ G_{\rm s}(r_{\rm i}(\tau^{\ }_{\rm i})-
r_{\rm j}(\tau^{\ }_{\rm j})) \quad ,\\
\int d^4x\ d^4y\ \bar{\psi}(x)\gamma^\alpha\psi(x)
\,G_{\rm v}(x-y)\, \bar{\psi}(y)\gamma_\alpha\psi(y)
&\rightarrow& \nonumber \\
&&\mbox{\hspace*{-6cm}}
\sum^{\rm A}_{\stackrel{{\scriptstyle \rm i,j=1}}{\rm i\neq j}}
\int\! d\tau^{\ }_{\rm i} d\tau^{\ }_{\rm j}\ G_{\rm v}(r_{\rm i}(\tau^{\ }_{\rm i})-
r_{\rm j}(\tau^{\ }_{\rm j}))\,
u_{{\rm i}\alpha}(\tau^{\ }_{\rm i}) u_{\rm j}^\alpha(\tau^{\ }_{\rm j})
\end{eqnarray}
in the action
(\ref{Action}) one obtains the non-instantaneous
action-at-a-distance \cite{Wheeler}
\begin{eqnarray}\label{Action1}
\int\! d^4x \ {\cal L}(x) &\rightarrow& \nonumber \\
{\cal A} &=& - \frac{1}{2} \sum^{\rm A}_{\rm i=1} \int\! d\tau^{\ }_{\rm i}
\left(M-\lambda_{\rm i}(\tau^{\ }_{\rm i})\right) u_{\rm i}(\tau^{\ }_{\rm i})^2
\nonumber \\
&+&\frac{1}{2} g^2_{\rm s}
\sum^{\rm A}_{\stackrel{{\scriptstyle \rm i,j=1}}{\rm i\neq j}}
\int\! d\tau^{\ }_{\rm i} d\tau^{\ }_{\rm j} \ G_{\rm s}(r_{\rm i}(\tau^{\ }_{\rm i}) -
r_{\rm j}(\tau^{\ }_{\rm j})) \\
&-&\frac{1}{2} g^{2}_{\rm v}
\sum^{\rm A}_{\stackrel{{\scriptstyle \rm i,j=1}}{\rm i\neq j}}
\int\! d\tau^{\ }_{\rm i} d\tau^{\ }_{\rm j} \ G_{\rm v}(r_{\rm i}(\tau^{\ }_{\rm i}) -
r_{\rm j}(\tau^{\ }_{\rm j}))\
u_{{\rm i}\alpha}(\tau^{\ }_{\rm i}) u_{\rm j}^\alpha(\tau^{\ }_{\rm j}) \ .\nonumber
\end{eqnarray}
Lagrange multipliers $\lambda_{\rm i}(\tau^{\ }_{\rm i})$ are introduced
to ensure $u_{\rm i}(\tau^{\ }_{\rm i})^2=1$.
The equations of motion which result from this Wheeler-Feynman action
are equivalent to the classical field equations, provided
the system does not radiate \cite{Wheeler}.
However, the Wheeler Feynman equations of motion cannot be
solved in general
as one needs to know the world lines for all the
past and future times in order to calculate the fields which enter
the equations for the world lines~\cite{BelMartin}. Furthermore, the
no-interaction theorem states that, except in the case of
non-interacting particles, there exist no covariant
equations of motion for world lines in which only the 4-positions and
the 4-velocities at a given time enter, as is the case in
non-relativistic mechanics.
In the following this no-interaction theorem is circumvented by
introducing an approximative solution to the non-local Wheeler Feynman
equations of motion. This is achieved by the so called
small acceleration approximation which does not assume small
velocities.
\subsection*{\large\sl 2.1. Action-at-a-distance in the small acceleration
approximation}
In order to conserve manifest covariance of the equations of motion
we introduce first the concept of a scalar time. For that the
whole Minkowski space is chronologically ordered by a set of
space like surfaces $S(x)$ which attribute to each 4-position
$x$ a scalar time $t$ by
\begin{equation}\label{Isochrone}
S(x)=t \ \ , \ \ \ \mbox{with} \ \ \ \partial^0 S(x) > 0 \ ,\
\partial^0 \partial^0 S(x) \ge 0 \ .
\end{equation}
The simplest choice for these isochrones, which we shall take
in the following, is
$S(x)=\eta_\alpha x^a$, where $\eta_\alpha$ is a position
independent time like 4-vector.
Unlike in a collision the nucleons are not strongly accelerated by the
action of the mean-field. Therefore, in order to describe the motion
in the mean-field one may expand each world line
around the proper time $\tau^s_{\rm j}$ at which the particle is at
a given isochrone, i.e. $t=S(r_{\rm j}(\tau^s_{\rm j}))$ for each $j$.
\begin{equation}\label{worldline}
r_{\rm j}^\alpha(\tau^{\ }_{\rm j})=r_{\rm j}^\alpha(\tau^s_{\rm j})
+(\tau^{\ }_{\rm j}-\tau^s_{\rm j})
u_{\rm j}^\alpha(\tau^s_{\rm j})+
\frac{1}{2}(\tau^{\ }_{\rm j}-\tau^s_{\rm j})^2 \,
a_{\rm j}^\alpha(\tau^s_{\rm j})+\cdots \ .
\end{equation}
This way we are defining a synchronization prescription for all particles,
which does not
depend on the frame. For calculating the fields we neglect of the
quadratic term with the acceleration
$a_{\rm j}^\alpha(\tau^s_{\rm j})$ and all higher powers
which leaves us with a straight world line in the vicinity of the
synchronizing time $t=S(r_{\rm j}(\tau^s_{\rm j}))$ (c.f. Fig.~1).
This is called ''small acceleration approximation"
\begin{figure}[t]
\vspace*{-10mm}
\insertplot{ff-wline.ps}
\vspace*{-10mm}
\caption{World lines and synchronizing hypersurface $S(x)$}
\end{figure}
The small acceleration approximation is of course best in the vicinity
of $r_{\rm j}^\alpha(\tau^s_{\rm j})$ and becomes worse further away.
As sketched in Fig. 1 only those parts of the world lines (thick lines),
which are outside the light cone (centered at $x$),
contribute to the field strength at point $x$.
A world line which hits the light cone far away from $x$ may be badly
approximated by Eq. (\ref{worldline}), but for short range interactions
a distant particle does not contribute anymore to the field at $x$.
Thus, the first condition for the validity of the approximation is
that the range $\mu^{-1}$ is small compared to the curvature of the
world lines, i.e. the inverse of the acceleration.
The second condition is weak radiation, which is fulfilled when the
acceleration is small compared to the meson mass $\mu$. Both
conditions are actually the same, namely
\begin{equation}\label{smallacc}
\mid a_\mu a^\mu \mid \ \ll \ \mu^2 .
\end{equation}
The assumption of small accelerations is justified if the $\phi$
and $A^\alpha$ fields are only meant to be the mean-field part of the
nucleon-nucleon interaction in a hadronic surrounding. The hard collisions
between individual nucleons which are due to the repulsive core will
cause large accelerations and also create new particles.
These hard collisions cannot be described within Walecka type
mean-field models.
Therefore, it is consistent to regard $\phi (x)$ and $A^\alpha(x)$ as Hartree
mean-fields which bring about only small accelerations and which
are not radiated away from their sources.
Inserting the straight world line into Eq.
(\ref{Phiworld})
results in the easily understood situation that the field at a point $x$ is
just the sum of Lorentz-boosted Yukawa potentials which are traveling
along with the charges:
\begin{equation}\label{Phiyuk}
\phi(x) = \frac{g_{\rm s}}{4\pi} \sum_{\rm j}
\frac{\exp\left\{-\mu_{\rm s} \sqrt{R_{\rm j}(x)^2}\right\} }
{\sqrt{R_{\rm j}(x)^2}}\ ,
\end{equation}
where $R_{\rm j}(x)^2$ is given by
\begin{equation}
R_{\rm j}(x)^2 = - (x-r_{\rm j}(\tau^s_{\rm j}))^2 +
\left[(x^\alpha-r_{\rm j}^\alpha(\tau^s_{\rm j}))\,
u_{{\rm j}\alpha}(\tau^s_{\rm j})\right]^2
\end{equation}
The vector field is derived in an analogue fashion as
\begin{equation}\label{Ayuk}
A^\alpha(x) = \frac{g_{\rm v}}{4\pi} \sum_{\rm j}
\frac{\exp\left\{-\mu_{\rm v}\sqrt{R_{\rm j}(x)^2}\right\} }
{\sqrt{R_{\rm j}(x)^2}}\, u_{\rm j}^\alpha(\tau^s_{\rm j}) \ .
\end{equation}
At this level of the approximation the causality problem with the advanced
part of the Green function is not present because the retarded and the
advanced fields are identically the same when they are
created by charges moving on straight world lines.
Therefore one may regard the fields as retarded only.
In addition the unsolved problem of radiation reaction, where
the radiation acts back on the world lines \cite{Jackson},
does not occur because there are no radiation fields any more.
\subsection*{\large\sl 2.2. Instantaneous action-at-a-distance}
In the spirit of the small acceleration assumption discussed in the previous
section one can use the straight line expansion in the action
(\ref{Action1}) and perform the integration over
$\tau^{\ }_{\rm j}$. This results in an instantaneous
action-at-a-distance where the Lorentz-boosted Yukawa fields appear again
and there is only one time, the scalar synchronizing
time $t$.
\begin{eqnarray}\label{Action2}
{\cal A} = \int \! dt \sum_{\rm i} \frac{1}{\partial_\alpha
S(r_{\rm i}(t))u_{\rm i}^\alpha(t)}
\Bigg[
&-&\frac{1}{2} \left(M-\lambda_{\rm i}(t)\right)u_{\rm i}(t)^2 \nonumber\\
&+& \frac{1}{2} \sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}
\frac{g^2_{\rm s}}{4\pi}
\frac{\exp\{-\mu_{\rm s} \sqrt{R_{\rm ij}(t)^2}\}}{\sqrt{R_{\rm ij}(t)^2}} \\
&& \mbox{\hspace*{-2cm}}-
\frac{1}{2} \sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}
\frac{g^2_{\rm v}}{4\pi}
\frac{\exp\{-\mu_{\rm v} \sqrt{R_{\rm ij}(t)^2}\}}{\sqrt{R_{\rm ij}(t)^2}}\,
u_{{\rm i}\alpha}(t) u_{\rm j}^\alpha(t)\,\Bigg]
\nonumber
\end{eqnarray}
The four-positions $r_{\rm i}^\alpha(t)\equiv
r_{\rm i}^\alpha(\tau^{\ }_{\rm i}(t))$ and
4-velocities $u_{\rm i}^\alpha(t) \equiv u_{\rm i}^\alpha(\tau^{\ }_{\rm i}(t))$
are to be taken at the same scalar synchronizing time $t$ and
\begin{equation}
R_{\rm ij}^2(t) := - (r_{\rm i}(t) - r_{\rm j}(t))^2 +
\Big[(r_{{\rm i}\beta}(t) -
r_{{\rm j}\beta}(t)) u_{\rm j}^\beta (t)\Big]^2
\end{equation}
The small acceleration approximation together with the introduction of a
synchronizing hypersurface $S(x)$ leads to an equal time lagrangian which is
Lorentz-scalar and written in a manifestly covariant way.
Giving up explicit covariance and choosing $\eta=(1,0,0,0)$ in a special
coordinate frame, the positions and velocities take the form
\begin{equation}\label{eq:rcm}
r_{\rm i}(\tau^{\ }_{\rm i}(t)) = \Big(t\, , \,\vec{r}_{\rm i}(t)\Big)
\ \ \ \mbox{and} \ \ \ u_{\rm i}(\tau^{\ }_{\rm i}(t))
= \frac{1}{\sqrt{1-\vec{v}_{\rm i}^{\,2}(t)}} \Big( 1\, ,\,
\vec{v}_{\rm i}(t) \Big) \ .
\end{equation}
With that a Lagrange function ${\cal L} (\vec{r}_{\rm i}(t),
\vec{v}_{\rm i}(t))$
can be defined which depends only on the independent variables and one time.
Even the Lagrange multipliers $\lambda_{\rm i}(t)$ are not needed anymore if
the
variation is with respect to $\vec{v}_{\rm i}(t)$ instead of all four
$u_{\rm i}^\alpha(t)$.
The advantage of the instantaneous Lagrangian is that one can
define easily the hamiltonian and the total momentum, which are then strictly
conserved by the equations of motion.
\subsection*{\large\sl 2.3. Hamilton equations of motion}
In the following we want to express the total hamiltonian as a function of the
positions $\vec{r}_{\rm i}$ and the canonical momenta $\vec{p}_{\rm i}$.
Using relations
(\ref{Action2})-(\ref{eq:rcm}) the Lagrange function is
\begin{eqnarray}
\label{eq:lang}
{\cal L} = &-&\sum_{\rm i} \ \left[ {m \over{ u_{\rm i}^0 }} + {1 \over 2}
\sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}\
{g^2_{\rm s} \over{ 4\pi u_{\rm i}^0 }}
{{\rm exp}\left\{ -\mu_{\rm s} \sqrt{R_{\rm ij}^2(t)} \right\} \over{
\sqrt{R_{\rm ij}^2(t)} }} \right. \\
&-& \left. {1 \over 2}
\sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}\
{g^2_{\rm v} \over{ 4\pi
u_{\rm i}^0 }}
{{\rm exp}\left\{ -\mu_{\rm v} \sqrt{R_{\rm ij}^2(t)} \right\} \over{
\sqrt{R_{\rm ij}^2(t)} }} u_{\rm i\alpha}(t) u_{\rm j}^{\alpha}(t)
\right] \, . \nonumber
\end{eqnarray}
The canonical momenta and the energy are determined from the Lagrange function
above as
\begin{eqnarray}
\vec{p}_{\rm i} &=& {\partial {\cal L} \over{ \partial \vec{v}_{\rm i} }}
\quad , \\
E &=& \sum_{\rm i} {\partial {\cal L} \over{ \partial \vec{v}_{\rm i} }}
\vec{v}_{\rm i} - \cal{L} \quad .\nonumber
\end{eqnarray}
Performing these derivatives, the energy is given by
\begin{eqnarray}
\label{eq:enu}
E &=& \sum _{\rm i} \ m\ u_{\rm i}^0 - {1\over 2}
\sum_{\stackrel{\scriptstyle\rm i,j}{\rm j \neq i}}\
f_{\rm ij} u_{\rm i}^0 - {1\over 2}
\sum_{\stackrel{\scriptstyle\rm i,j}{\rm j \neq i}}\
\widehat{f}_{\rm ji}\ (\vec{r}_{\rm ij} \vec{u}_{\rm i})^2
{\left( u_{\rm i}^0 \right)^2 \over{ u_{\rm j}^0 }} \nonumber \\
&+& {1\over 2}
\sum_{\stackrel{\scriptstyle\rm i,j}{\rm j \neq i}}\
g_{\rm ji}\ u_{\rm i}^0 - {1\over 2}
\sum_{\stackrel{\scriptstyle\rm i,j}{\rm j \neq i}}\
g_{\rm ji} \left[
u_{\rm i}^0 \vec{u}_{\rm i}^2 - {\left( u_{\rm i}^0 \right)^2
\over{ u_{\rm j}^0 }} \vec{u}_{\rm i} \vec{u}_{\rm j} \right] \\
&+& {1\over 2}
\sum_{\stackrel{\scriptstyle\rm i,j}{\rm j \neq i}}\
\widehat{g}_{\rm ji}\ (\vec{r}_{\rm ij} \vec{u}_{\rm i})^2 \left[
\left( u_{\rm i}^0 \right)^3 - {\left( u_{\rm i}^0 \right)^2
\over{ u_{\rm j}^0 }}
\vec{u}_{\rm i} \vec{u}_{\rm j} \right] , \nonumber
\end{eqnarray}
\noindent where we use the abbreviations
\begin{eqnarray}
\label{eq:fdef1}
f_{\rm ij} \equiv f(R_{\rm ij}) = {g_{\rm s}^2\over{4 \pi}}\ {{\rm exp}\left\{
-\mu_{\rm s} \rm R_{ij}\right\} \over{R_{\rm ij}}} \quad &,& \quad
g_{\rm ij} = {g_{\rm v}^2\over{4 \pi}}\ {{\rm exp}\left\{ -\mu_{\rm v}
\rm R_{ij}\right\} \over{R_{\rm ij}}} \, , \nonumber \\
\widehat{f}_{\rm ji} \equiv -{1 \over{ R_{\rm ji} }} {d \over{ d R_{\rm ji} }}
f(R_{\rm ji}) = {g_{\rm s}^2\over{4 \pi}}\ (1 &\hspace*{-5mm}+& \hspace*{-5mm}
\mu_{\rm s} R_{\rm ji})
{{\rm exp}\left\{ -\mu_{\rm s} \rm R_{ji}\right\} \over{ R_{\rm ji}^3 }}\, , \\
R_{\rm ij}^2 = \vec{r}_{\rm ij}^2 + (\vec{r}_{\rm ij}
\vec{u}_{\rm j})^2 \quad &,& \quad
\vec{r}_{\rm ij} = \vec{r}_{\rm i} - \vec{r}_{\rm j} \, ,
\label{eq:fdef2}
\end{eqnarray}
and $\widehat{g}_{\rm ji}$ is defined in the same way as $\widehat{f}_{\rm
ji}$.
It is worthwhile to note that $f_{\rm ij} \neq f_{\rm ji}$.
The momentum of a particle looks as
\begin{eqnarray}
\label{eq:mom}
\vec{p}_{\rm i} &=& m\vec{u}_{\rm i} - {1\over 2} \left(
\sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}\
f_{\rm ij} \right) \vec{u}_{\rm i} - {1\over 2}
\sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}\
\widehat{f}_{\rm ji} {u_{\rm i}^0 \over{ u_{\rm j}^0 }} \left[
\vec{r}_{\rm ij} (\vec{r}_{\rm ij} \vec{u}_{\rm i})
+ \vec{u}_{\rm i} (\vec{r}_{\rm ij}
\vec{u}_{\rm i})^2 \right] \nonumber \\
&+& {1\over 2}
\sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}\
\left( g_{\rm ij} + {u_{\rm i}^0
\over{ u_{\rm j}^0 }} g_{\rm ji} \right) \vec{u}_{\rm j}
- {1\over 2}
\sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}\
g_{\rm ji} \left[
\left( u_{\rm i}^0 \right)^2 - {u_{\rm i}^0 \over{ u_{\rm j}^0 }}
\vec{u}_{\rm i} \vec{u}_{\rm j} \right] \vec{u}_{\rm i} \\
&+& {1\over 2}
\sum_{\stackrel{\scriptstyle\rm j}{\rm j \neq i}}\
\widehat{g}_{\rm ji} \left[
\left( u_{\rm i}^0 \right)^2 - {u_{\rm i}^0 \over{ u_{\rm j}^0 }}
\vec{u}_{\rm i} \vec{u}_{\rm j} \right]
\left[
\vec{r}_{\rm ij} (\vec{r}_{\rm ij} \vec{u}_{\rm i}) +
\vec{u}_{\rm i} (\vec{r}_{\rm ij} \vec{u}_{\rm i})^2
\right] \, . \nonumber
\end{eqnarray}
Let us first prove that for isotropic nuclear matter, like in the original
mean-field picture~\cite{kn:serwal}
the vector potential will not contribute to the canonical momentum.
In that case the summation can be replaced by
an integration over space and momentum, folded with the phase space
distribution
function. Since for isotropic nuclear matter the distribution function is
independent of the position, the space integral can be carried out
immediately and the remaining part is written as a summation over momenta.
This way we get for the Yukawa terms in nuclear matter
\begin{eqnarray}
\sum_{\rm j}\ f_{\rm ij} &\rightarrow&
\sum_{\rm j}\
\int d^3r_{\rm j} \ f(R_{\rm ij}) = \left( {g_{\rm s} \over{
\mu_{\rm s}}} \right)^2 \sum_{\rm j}\
{1\over{u_{\rm j}^0}} \nonumber \, . \\
\sum_{\rm j}\ \widehat{f}_{\rm ji}\ (\vec{r}_{\rm ij} \vec{u}_{\rm i})^2
&\rightarrow&
\sum_{\rm j}\
\int d^3r_{\rm j}\ \widehat{f}(R_{\rm ji}) (\vec{r}_{\rm ij}
\vec{u}_{\rm i})^2 =
\left( {g_{\rm s} \over{ \mu_{\rm s}}} \right)^2
{\vec{u}_{\rm i}^2\over{\left( u_{\rm i}^0 \right)^3 }} \
\sum_{\rm j}\ 1 \, .
\label{eq:nms}
\end{eqnarray}
Substituting the above expressions into the momentum~(\ref{eq:mom}) the
$\widehat{f}_{\rm ij}$ term gives the same contribution as the
$f_{\rm ij}$, while
the $g_{\rm ij}$ terms cancel each other and the momentum can be written as
\begin{equation}
\vec{p}_{\rm i} = m \vec{u}_{\rm i} - \left( {g_{\rm s}\over{
\mu_{\rm s} }} \right)^2 \vec{u}_{\rm i}
\sum_{\stackrel{{\scriptstyle\rm j}}{\rm j\neq i}} \
{1\over{u_{\rm j}^0}} + \left( {g_{\rm v}\over{\mu_{\rm v} }}
\right)^2
\sum_{\stackrel{{\scriptstyle\rm j}}{\rm j\neq i}} \
{\vec{u}_{\rm j}\over{ u_{\rm j}^0}} \, .
\label{eq:pnm0}
\end{equation}
The last term gives zero, since the average of $\vec{u}_{\rm j}$ is zero in
nuclear matter and Eq.~(\ref{eq:pnm0}) can be written as
\begin{equation}
\vec{p}_{\rm i} = m^* \vec{u}_{\rm i} \quad ,
\label{eq:pnm1}
\end{equation}
\noindent where
\begin{eqnarray}
m^* &=& m - \left( {g_{\rm s}\over{\mu_{\rm s} }} \right)^2
\sum_{\stackrel{{\scriptstyle\rm j}}{\rm j\neq i}} \
{1\over{u_{\rm j}^0}} \nonumber \\
&=& m - \left( {g_{\rm s}\over{\mu_{\rm s} }} \right)^2
\sum_{\rm j} \ {m^* \over{\sqrt{{m^*}^2 + \vec{p}_{\rm j}^2}}} \, .
\label{eq:msnm}
\end{eqnarray}
Similarly, the energy density (\ref{eq:enu}) has the form
\begin{equation}
\epsilon = \sum_{\rm i}\ \sqrt{{m^*}^2 + \vec{p}_{\rm i}^2}
+ {1\over 2} \left( {g_{\rm v}\over{\mu_{\rm v} }} \right)^2
\left( j^0 \right)^2 + {1\over 2} \left( {\mu_{\rm s}
\over{ g_{\rm s} }} \right)^2
(m - m^*)^2 \quad .
\label{eq:ednm}
\end{equation}
\noindent Eqs.~(\ref{eq:msnm}-\ref{eq:ednm}) are just
the mean-field results of the Walecka model~\cite{kn:serwal}.
It is worthwhile to mention that the small acceleration approximation which
introduces $R_{\rm ij}$ instead of $\mid \vec{r}_{\rm ij}\mid$ in
Eqs.~(\ref{eq:fdef1}-\ref{eq:fdef2}), is
definitely needed to get back the relativistic mean-field result for nuclear
matter.
Therefore, for the coupling strengths $g_{\rm s}$ and $g_{\rm v}$ we can simply
take the values obtained from the saturation properties of the nuclear
matter~\cite{Lindner}.
For the Hamilton equations we need the energy as the function of the
canonical momenta instead of the 4-velocities $u_{\rm i}$. For that we have
to invert
the $p(u)$ equation (\ref{eq:mom}) into a $u(p)$ one. One can do this
approximatively by first expanding the 4-velocities in the $p(u)$ relation
up to the third order in the quantities $(\vec{p}/m)^2$, $f_{\rm ij}$ and
$g_{\rm ij}$, and then substituting that expression for $u(p)$ into the
energy Eq.~(\ref{eq:enu}). At the end we will
check the validity of this approximation. After tedious calculations
the energy can be expressed as follows (in the following we include explicitly
$c$ and $\hbar$)
\begin{eqnarray}
\label{eq:en}
{E\over{mc^2}} &=& \sum_{\rm i}\ \sqrt{1+\vec{\tilde{p}}_{\rm i}^2}
- {1\over 2}
\sum_{\stackrel{{\scriptstyle \rm i,j}}{\rm j \neq i}}\
{\tilde{f}_{\rm ij} \over{
\sqrt{1+\vec{\tilde{p}}_{\rm i}^2}}} + {1\over 2}
\sum_{\stackrel{{\scriptstyle \rm i,j}}{\rm j \neq i}}\
\tilde{g}_{\rm ji} \sqrt{1+\vec{\tilde{p}}_{\rm i}^2} \nonumber \\
&-& {1\over 2}
\sum_{\stackrel{{\scriptstyle \rm i,j}}{\rm j \neq i}}\
\tilde{g}_{\rm ji}
{(\vec{\tilde{p}}_{\rm i} \vec{\tilde{p}}_{\rm j}) \over{
\sqrt{1+\vec{\tilde{p}}_{\rm i}^2}}} + {1\over 2}
\sum_{\rm i}\ \left(
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
\tilde{f}_{\rm ji} \vec{\tilde{p}}_{\rm i}
- \sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
\tilde{g}_{\rm ji} \vec{\tilde{p}}_{\rm j}
\right)^2 \quad ,
\end{eqnarray}
\noindent where
\begin{eqnarray}
\tilde{f}_{\rm ij} &=& {1\over{4\pi}} \left( {g_{\rm s} m \over{
\mu_{\rm s}}} \right)^2 \left( {\hbar \over{ mc }} \right)
{{\rm exp}\left\{ -{\hbar \over{ \mu_{\rm s} c }}
\tilde{R}_{\rm ij}\right\} \over{\tilde{R}_{\rm ij}}} \quad {\rm and}
\nonumber \\
\tilde{R}_{\rm ij}^2 &=& \vec{r}_{\rm ij}^2 + ( \vec{r}_{\rm ij}
\vec{\tilde{p}}_{\rm j} )^2 \quad ,
\end{eqnarray}
\noindent with $\vec{\tilde{p}}_{\rm i} = \vec{p}_{\rm i}/mc$.
$\tilde{g}_{\rm ij}$ is defined similarly to $\tilde{f}_{\rm ij}$. The last
term is already of third order and hence small.
{}From Eq.~(\ref{eq:en}) the Hamilton equations
\begin{equation}
{d \over{ dt }} \vec{r}_{\rm i} = -\
{\partial E\over{\partial \vec{p}_{\rm i}}} \qquad
{d \over{ dt }} \vec{p}_{\rm i} =
{\partial E\over{\partial \vec{r}_{\rm i}}}
\end{equation}
are calculated easily.
\section*{\large\bf 3. Details of the QMD calculations}
Molecular dynamics is a classical many-body theory in which some quantum
features
due to the fermionic nature of nucleons are simulated. For the details of
the theory we would like to refer the reader to the works of
Aichelin~\cite{kn:aich} and the
Frankfurt group~\cite{kn:frank}. In our calculation we introduced some
modifications which are described in~\cite{kn:judqmd,kn:paula}.
In the following this
model is used in a relativistic treatment with scalar and vector forces.
\subsection*{\large\sl 3.1. The mean-field forces}
The scalar and vector forces are the relativistic two-body forces derived in
the previous section which give saturation for
nuclear matter even without an explicit density dependent term. However,
Walecka's values for the coupling constants result in a too
high compressibility. For this reason it seems to be more adequate to use
instead the Zim\'anyi-Moszkowski (ZM) lagrangian~\cite{kn:zimmos} which
provides a more reasonable compressibility. The different coupling
of the mean-field ($m^* = ( m + g_{\rm s} \phi )^{-1}$)
modifies the scalar part of the total energy.
In the used approximation, the expression
\begin{equation}
-{1\over 2} \sum_{\stackrel{\scriptstyle \rm i,j}{
\scriptstyle \rm j \neq i}} \
{\tilde{f}_{\rm ij}
\over{
\sqrt{1 + \vec{\tilde{p}}_{\rm i}^2} }}
\left( {1 \over{
1 + 2 {\displaystyle \sum_{\stackrel{\scriptstyle \rm k}{\rm k \neq j}}\
\tilde{f}_{\rm jk} } }} \right) \quad ,
\end{equation}
replaces the second term on the r.h.s. of Eq.~(\ref{eq:en}). Here we see the
effect of the nonlinear coupling in the ZM lagrangian. The strength of the
scalar potential which is felt by particle j is weakened.
We compare the Walecka forces with the ZM ones and study
their effect on nuclear multifragmentation. The coupling strengths $g_{\rm s}$
and $g_{\rm v}$ are determined from the ground state properties of nuclear
matter~\cite{Lindner,kn:serwal}. In order to include the symmetry
energy we take also
the $\varrho$ meson into account. To
simplify the calculations the mass of the $\varrho$ meson is taken to be equal
to the one of the $\omega$ meson, and the coupling constants are fitted to get
the appropriate symmetry energy. In Table~1 we list all coupling
constants used.
\begin{table}[t]
\begin{center}
\begin{tabular}{||l|c|c|c|c|c||} \hline
& $g_{\rm s}$ & $g_{\rm v}$ & $g_{\varrho}$ & $\mu_{\rm s} c^2$ &
$\mu_{\rm v} c^2$ = $\mu_{\varrho} c^2$ \\ \hline \hline
W & 11.04 & 13.74 & 7.0 & 550 MeV & 783 MeV \\ \hline
ZM & 7.84 & 6.67 & 7.0 & 550 MeV & 783 MeV \\ \hline
\end{tabular}
\end{center}
\caption{Parameters of the Walecka and the Zim\'anyi-Moszkowski forces}
\end{table}
Since the relativistic two-body forces are meant to describe the long
range part of the interaction, the Yukawa functions are folded with a gaussian
density distributions for the nucleons.
For nuclear matter one gets a smooth density distribution if the
gaussians $e^{-\alpha^2 (\vec{r} - \vec{r}_1 )^2}$ have a width parameter of
$\alpha=0.5$ fm$^{-1}$. We use that value for
folding the Yukawa forces, analogue to our earlier QMD
calculations~\cite{kn:judqmd,kn:paula}, with $R^2_{\rm ij}$ replacing
$\vec{r}^{\ 2}_{\rm ij}$ everywhere.
\subsection*{\large\sl 3.2. Initial conditions}
For the initial coordinate and momentum distribution of the particles we
use the ones, which were prepared for the non-relativistic
calculations~\cite{kn:judqmd} .
The main aspect when generating initial positions and momenta is to get a
smooth
phase-space distribution. The ground-state energies calculated with
distributions which fit nuclear charge densities
are found to be good for both, the Walecka and the ZM forces.
\subsection*{\large\sl 3.3. Cross sections}
\begin{figure}[t]
\vspace*{-10mm}
\insertplot{ff-pdis.ps}
\vspace*{-10mm}
\caption{Distribution of nucleon-nucleon relative momenta up to time $t$
in Ca+Ca central collision}
\end{figure}
The multifragmentation depends on the cross sections used. According to Cugnon,
his parameterization~\cite{kn:cug} gives a good fit only for relative
momenta $\ge 0.8$ GeV/c in the two particle center of mass frame.
For Ca + Ca central
collisions at 200, 400, 600 and 800 MeV/u initial beam energies
Fig.~2 shows however, that the vast majority of the collisions occur at
relative momenta below $0.2$ GeV/c. To see the effect of the cross section
on multifragmentation
we fit the experimental free nucleon-nucleon
cross sections~\cite{kn:land-b} as a function of the relative momentum
with second order polynomials
and use that fit in our calculations. However, due to the screening
effect of the other nucleons we do not allow collisions at
distances larger than $r_{\rm coll}$. The screening length is somewhat
arbitrary, so we work with two parameterization, $r_{\rm coll}$ = 1.6 fm
and $r_{\rm coll}$ = 2.4 fm.
The low energy cross section, we use in
the calculation, is displayed in Fig.~3. The dependence of the
fragment distribution on the different cross sections is discussed
in the next section.
\begin{figure}
\vspace*{-10mm}
\insertplot{ff-cross.ps}
\vspace*{-10mm}
\caption{Low energy nucleon-nucleon cross section}
\end{figure}
\subsection*{\large\sl 3.4. Effective mass in two-body collisions}
In relativistic many-body calculations single-particle energies
$e_{\rm i}$ cannot be defined such that ($e_{\rm i},\vec{p}_{\rm i}$)
form a 4-vector.
This holds true only for the total momentum $\vec{P}$ and total energy $E$.
However, we still try to deduce a single particle energy, which we use for
prescribing the conservation of energy in each collision. If one considers
only
the scalar potential in the lowest order the solution is easy. The
equation (\ref{eq:en}) in the leading order can be written as
\begin{equation}
\label{eq:efm}
{E\over{mc^2}} = \sum_{\rm i}\ \sqrt{1+\vec{\tilde{p}}_{\rm i}^2} - {1\over 2}
\sum_{\stackrel{{\scriptstyle \rm i,j}}{\rm j \neq i}}\
{\tilde{f}_{\rm ij} \over{ 1 + 2 {\displaystyle
\sum_{\stackrel{{\scriptstyle \rm k}}{\rm k \neq j}}\
\tilde{f}_{\rm jk} } }}
{1 \over{ \sqrt{1+\vec{\tilde{p}}_{\rm i}^2}}} + {1\over 2}
\sum_{\stackrel{{\scriptstyle \rm ij}}{\rm j \neq i}}\
\tilde{g}_{\rm ji} \sqrt{1+\vec{\tilde{p}}_{\rm i}^2}
\end{equation}
There are two ways in our case to define the effective mass $m^*$ appearing in
the collisions:
Case A: If we neglect the momentum dependence in $\tilde{f}_{\rm ij}$ and
$\tilde{g}_{\rm ij}$, the energy of the i$^{\rm th}$ particle can be written
as
\begin{eqnarray}
\label{eq:sinen}
{E(A) - E_{\rm i}(A-1) \over{ mc^2 }} &=& \sqrt{ 1 + \vec{\tilde{p}}_{\rm i}^2
}
- {1 \over 2}
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
{\tilde{f}_{\rm ij} \over {
\sqrt{ 1 + \vec{\tilde{p}}_{\rm i}^2 } }}
{1 \over{ 1 + 2 {\displaystyle
\sum_{\stackrel{\scriptstyle \rm k}{\rm k \neq j}}\
\tilde{f}_{\rm jk} } }} \\
&+& {1 \over 2}
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
\tilde{g}_{\rm ji} \sqrt{ 1 + \vec{\tilde{p}}_{\rm i}^2 }
+ \mbox{terms independent of }\ \vec{\tilde{p}}_{\rm i} \ ,\nonumber
\end{eqnarray}
where $E_{\rm i}(A-1)$ means the energy of the system when particle i is taken
out.
We can expand Eq.~(\ref{eq:sinen}) in $\vec{\tilde{p}}_{\rm i}^2$ and get
$$
\epsilon_{\rm i} = {1 \over 2} \vec{\tilde{p}}_{\rm i}^2 \left[ 1 + {1 \over 2}
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
{\tilde{f}_{\rm ij} \over{ 1 + 2 {\displaystyle
\sum_{\stackrel{{\scriptstyle \rm k}}{\rm k \neq j}}\
\tilde{f}_{\rm jk} } }} + {1 \over 2}
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
\tilde{g}_{\rm ji} \right] + \epsilon_{\rm i0} \quad ,
$$
where $\epsilon_{\rm i0}$ is independent of $\vec{\tilde{p}}_{\rm i}$.
This can be written as
$$
\epsilon_{\rm i} = \sqrt{ \left( m^*_{\rm i} \right)^2 +
\vec{\tilde{p}}_{\rm i}^2 } + \epsilon^{\prime}_{\rm i0} \quad ,
$$
where the effective mass $m^*_{\rm i}$ turns out to be
\begin{equation}
\label{eq:msr}
m^*_{\rm i} = 1
- {1 \over 2}
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
{\tilde{f}_{\rm ij} \over{
1 + 2 {\displaystyle
\sum_{\stackrel{{\scriptstyle \rm k}}{\rm k \neq j}}\
\tilde{f}_{\rm jk} } }}
- {1 \over 2}
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
\tilde{g}_{\rm ji} \quad .
\end{equation}
\begin{figure}[t]
\vspace*{-10mm}
\insertplot{ff-fdti2.ps}
\vspace*{-10mm}
\caption{The evolution of the fragment distribution in time in 200 MeV/u
O + Br central collisions}
\end{figure}
Case B: If we assume, that the momentum dependence of
$\tilde{f}_{\rm ij}$ and $\tilde{g}_{\rm ji}$ is the same as in nuclear matter,
that is
\begin{equation}
E = \sum_{\rm i}\
\sqrt{ 1 + \vec{\tilde{p}}_{\rm i}^2 }
- {1\over 2} \sum_{\stackrel{{\scriptstyle \rm i,j}}{\rm j \neq i}}\
{f(r_{\rm ij}) \over{ \sqrt{ 1 + \vec{\tilde{p}}_{\rm i}^2 }
\sqrt{ 1 + \vec{\tilde{p}}_{\rm j}^2 } }}
{ 1 \over { 1 + 2 {\displaystyle
\sum_{\stackrel{{\scriptstyle \rm k}}{\rm k \neq j}}\
f(r_{\rm jk}) } }}
+ {1\over 2} \sum_{\stackrel{{\scriptstyle \rm i,j}}{\rm j \neq i}}\
g(r_{\rm ji}) \quad ,
\end{equation}
(see Eq.~(\ref{eq:nms})),then the effective mass for finite systems turns out
to be
\begin{equation}
\label{eq:msi}
m^*_{\rm i} = 1 -
\sum_{\stackrel{{\scriptstyle \rm j}}{\rm j \neq i}}\
{f(R_{\rm ij}) \over{
1 + 2 {\displaystyle
\sum_{\stackrel{{\scriptstyle \rm k}}{\rm k \neq j}}\
f(R_{\rm jk}) } }} \quad .
\end{equation}
We found in our calculations that the definition~(\ref{eq:msr}) gave better
conservation of the total energy.
\subsection*{\large\sl 3.5. The treatment of a relativistic collision}
\begin{figure}
\vspace*{-10mm}
\insertplot{ff-fdti5.ps}
\vspace*{-10mm}
\caption{The evolution of the fragment distribution in time in 50 MeV/u
O + Br central collision}
\end{figure}
The individual two-body collision is calculated in a relativistically covariant
way, as given in
\cite{kn:wolf}. The
particles are allowed to collide
if the impact parameter
\begin{eqnarray}
b &=& \sqrt{R_{12}^2 - h_{12}^2/v_{12}^2} \qquad \mbox{with} \\
R_{12}^2 &=& -(x_1 - x_2)^2 + \left( {p_1 (x_2 - x_1) \over{ m_1 c}} \right)^2
\quad , \nonumber \\
h_{12} &=& {p_1 (x_2 - x_1) \over{ m_1 c}} - {p_2 (x_2 - x_1) m_1 c \over{
p_1 p_2 }} \quad \mbox{and} \nonumber \\
v_{12}^2 &=& 1 - \left( {m_1 m_2 c^2 \over{ p_1 p_2 }} \right)^2 \nonumber
\end{eqnarray}
is smaller than the collision radius $\sqrt{\sigma/\pi}$. In
the model the collision takes place at equal time in the CM of the total
system, provided the closest approach of the world-lines occurs within the
following time step. After the collision we only allow final momenta which do
not let the particles approach each other further.
\subsection*{\large\sl 3.6. Pauli blocking}
Pauli blocking is treated as in Ref.~\cite{kn:judqmd}.
We calculate the averaged phase-space distribution function at the location
of each particle. The collision can occur if the new phase-space distribution
function is smaller than 1
for the colliding particles and for all other particles after the collision.
Since the distribution function at the phase space location of a particle
depends on the momenta of all the other particles, without this condition it
may occur that the phase
space density becomes greater than one even for particles which did not
take part in the collision.
\subsection*{\large\sl 3.7. Definition of fragments}
We consider the particles to belong to the same fragment if they stay together
in coordinate and momentum space and are bound. In our
calculation we follow the evolution of the system up to 360 fm/c for higher,
and up to 720 fm/c for lower energies. We see in the Figs.~4 and 5
that in the last 120 fm/c the so determined fragment distribution does
not change anymore.
In order to calculate the energy of a cluster we calculate the $f_{\rm ij}$
and $g_{\rm ij}$ in the total CM frame because they are Lorentz-scalars. Only
the momenta appearing in the momentum dependent terms have to be Lorentz
transformed to the center of momentum frame of the cluster.
The radii and the energies of the
fragments are very near to the experimental values, although some fragments can
shrink a little. However, we do not get energies lower than $-8.8$ MeV/u
for any case.
\begin{table}[t]
\begin{center}
\begin{tabular}{||c|c|c|c|c||} \hline
& Scalar & Vector & $\vec{p}_{\rm i} \vec{p}_{\rm j}$ term & 3. order term \\
\hline \hline
$\varrho_0$ & -153 & 122 & -0.01 & 2.2 \\
1 MeV/u & -56 & 28 & 0.005 & 0.5 \\ \hline
3 $\varrho_0$ & -647 & 529 & 0.3 & 38 \\
1 MeV/u & -90 & 59 & 0.04 & 2.19 \\ \hline
$\varrho_0$ & -110 & 175 & -85 & 17 \\
2 GeV/u & -40 & 40 & -19 & 14 \\ \hline
3 $\varrho_0$ & -235 & 381 & -169 & 80 \\
2 GeV/u & -66 & 89.7 & -39 & 51.8 \\ \hline
\end{tabular}
\end{center}
\caption{The contribution of the different energy terms of
\protect{Eq.~(\ref{eq:en})} in MeV per particle
for gold on gold systems for very extreme cases. The upper numbers refers to
the Walecka, the lower ones to the ZM forces. The calculations
are made for 1 MeV/u and 2 GeV/u energies at normal and 3 times
normal nuclear matter density.}
\end{table}
\section*{\large\bf 4. Multifragmentation in the O\ +\ Br collision}
Among the experiments carried out to study nuclear multifragmentation, emulsion
measurements were the first to give an almost exclusive
identification of the atomic numbers of the fragments emitted in heavy ion
collisions.
The investigation of the same system at several bombarding energies is
particularly useful for testing models. Such data are available for the O + Br
system at bombarding energies 50, 75, 100, 150 and 200 MeV/u \cite{kn:jakob}.
Nowadays modern experimental techniques allow mass production of such
measurements \cite{kn:FOPI}. Few years ago we
carried out calculations for the emulsion experiment with a
non-relativistic force to test the treatment of the Pauli
blocking~\cite{kn:judqmd}. Now we have made the analogue calculation with
relativistic forces at different energies to check their applicability at lower
energies. These forces are meant to be used at higher energies, but they should
also have good non-relativistic properties.
\begin{figure}
\vspace*{-10mm}
\insertplot{ff-fdis.ps}
\vspace*{-10mm}
\caption{Fragment distribution in O + Br central collisions at energies 50,
75, 100, 150 and 200 MeV/u}
\end{figure}
As a first step we checked the validity of the expansion in $f_{\rm ij}$ and
$g_{\rm ij}$. In Table~2 we give the values of different energy terms of
Eq.~(\ref{eq:en}) for a gold on gold system which was artificially compressed
to different densities and given different relative energies.
One can see, that for the ZM case the contribution of the
meson fields are much smaller than the nuclear rest mass $mc^2$. Even for the
Walecka case, except maybe for the high compression with $j^0=3 \varrho_0$, the
expansion is applicable. Thus
our approximation in the expansion (\ref{eq:mom}) of the momentum is justified.
In Fig.~6 the fragment distributions of the O + Br collisions are displayed for
different energies. The contribution from the
largest fragments are very small. Due to the computer time needed, the figures
present results averaged over only 500 events. Thus the statistical fluctuation
for events with small yields are still large.
In Fig.~7 we compare the results obtained
with the Walecka and ZM forces using the Cugnon parameterization \cite{kn:cug}
and our fitted cross sections for 50 MeV/u and 200 MeV/u bombarding energies.
The ZM force gives a somewhat better agreement with data than the forces
derived from the Walecka lagrangian. The effect is larger for higher energies.
However, using the Cugnon
cross section gives more or less the same results as using the one fitted
to the low energy free nucleon-nucleon scattering.
The fragment distributions shown in Fig.~6 do not differ significantly from
the earlier ones calculated with non-relativistic Skyrme forces, however
one should mention, that for lower energies the experimental cut does not
correspond to the central events. In Fig.~8 we display the distribution of
impact parameters for the 10\% highest multiplicities. For
the 200 MeV/u collision the multiplicity cut selects the central events (impact
parameter is less than 3 fm), but for 50 MeV it results in broad distribution.
The yields for 50 MeV/u and 200 MeV/u beam energies with the multiplicity cut
are given in Fig.~9.
\section*{\large\bf 5. Conclusion}
\begin{figure}
\vspace*{-10mm}
\insertplot{ff-fdforc.ps}
\vspace*{-10mm}
\caption{Fragment distribution in O + Br central collisions at 200 MeV/u for
different cross sections and forces}
\end{figure}
{}From an effective field theory lagrangian with scalar and vector mesons
we deduced in the small acceleration approximation
coordinate and momentum dependent relativistic forces,
which are suitable for transport
equation calculations. Even without an explicit
density dependence, saturation of nuclear matter is obtained at proper
densities and energies
due to relativistic effects. The Zim\'anyi-Moszkowski forces, where a
derivative coupling is used, can produce saturation even without
relativistic effects.
The proposed forces will be used in the intermediate energy domain
(500 MeV/u -- 2 GeV/u), where the mean-field plays still an important role and
relativistic forces are necessary. However, they should work
at lower energies too. As a first check we studied their low energy
behavior in a QMD calculation. We investigated both, the more static features
of these forces in the ground-states of nuclei and their dynamical
properties in the fragmentation process in heavy ion collisions.
The fragment distributions in the O + Br (50 -- 200 A MeV) collision were
calculated. The result we obtained
is of similar quality as the one with the non-relativistic
Skyrme forces~\cite{kn:judqmd,kn:paula}.
That encourages us to use these forces for beam energies in the domain of
500 MeV/u -- 2 GeV/u, where non-relativistic Skyrme type forces cannot be used
anymore.
A further test of the model should be for example the comparison of
the calculated flow (which is especially sensitive to the momentum
dependence) for different energies with experimental results of the
FOPI collaboration~\cite{kn:FOPI}.
Such calculations are under progress.
\begin{figure}
\vspace*{-10mm}
\insertplot{ff-bdis.ps}
\vspace*{-10mm}
\caption{The distribution of impact parameters to the
highest multiplicities. The solid line corresponds to 50 MeV/u collisions
and selects 11.3\% of the events, while the dashed line represents
9.5\% for 200 MeV/u}
\end{figure}
\begin{figure}
\vspace*{-10mm}
\insertplot{ff-50b.ps}
\vspace*{-10mm}
\insertplot{ff-200b.ps}
\caption{The yields for 50 MeV/u (upper) and 200 MeV/u (lower) beam energies
with trigger to the events with the 10\% highest multiplicities. The solid line
displays the yield obtained for a central collision (b=0.5 fm), the dashed line
displays the average over the highest multiplicity events.
The experimental data are shown by crosses}
\end{figure}
\section*{\large\bf Acknowledgments}
One of the authors (J.N.) should like to express her thanks to Prof.
N\"orenberg
for his kind hospitality at the GSI, where part of this work was done. G.P.
thanks Gy. Wolf for fruitful discussions.
This work was supported in part by the Hungarian Research Foundation OTKA.
\section*{\large\bf Notes}
E-mail: H. Feldmeier: <EMAIL> ; J. N\'e\-meth:
<EMAIL> ;\\
G. Papp: <EMAIL>; WWW:http://www.gsi.de/\verb+~+papp .\\
$^\dagger$: Permanent address
|
\section{Introduction}
The intention of this talk is to summarize the application of linearized
gravity, in the specific form of the theory of black hole perturbations,
to the study of the collision of black holes. Most of the
results are already present in the literature, and the rest of the work
is still in progress so I present here only a brief survey.
The motivation for studying black hole collisions is quite clear. In
the next few decades gravitational wave detectors will come online
that will require ``templates'' of possible waveforms from different
sources. The collision of black holes is one of the main candidates
for observable sources of gravitational radiation. Although the
initial and advanced LIGO detectors will not quite have the frequency
range to detect the waves produced in the final moments of the most
common collisions, it is expected that future detectors will, and
knowing the waveform for the final moments can also lead to insights
into the waveforms emitted earlier on.
The presence of this strong motivation from the experimental side has
led to the formation of an alliance of numerical relativity groups
(the ``binary black hole grand challenge collaboration'') with the
goal of numerically simulating the collision of two black holes using
supercomputers. The degree of difficulty of this project is reflected
in the fact that several established numerical relativity groups have
decided to team efforts in order to tackle it.
Here we will like to offer a much more modest approach, which is based
on a simple idea: when a collision of two black holes starts with the
holes so close to each other that they are surrounded by a common
horizon, the problem looks from the point of view of an external
observer as a single distorted black hole. It can therefore be treated
with perturbation theory. Although one expects this approach to only
yield results in a small range of initial separation, it provides ---at
least for that range--- a benchmark against which one can calibrate
numerical codes of the fully numerical approach. In reference
\cite{PrPu} an explicit calculation was carried out using this
idea. We took the initial data for the head-on collision of two black
holes given by the Misner \cite{Mi} solution and re-wrote it in such a
way that in the case that the two black holes are close to each other
it explicitly looks like ``Schwarzschild plus something small". We
took the ``something small" and evolved it using the equations of
linearized gravity (the Zerilli equation) and computed the radiated
energy. The results are shown in figure 1, where we plot the energy
radiated in the collision as a function of the initial separation and
compare with the results of the NCSA group \cite{NCSA} using a
numerical integration of the full Einstein equations. We see that the
close approximation works very well until the holes are no longer
surrounded by a common apparent horizon ($\mu_0= 1.3$) and works
within the correct order of magnitude up to when the holes are no
longer surrounded by an event horizon ($\mu_0=2.0$). Also shown is a
``far approximation'' based on a particle-membrane paradigm
\cite{collab}. Comparisons of waveforms have also been performed
\cite{collab} and they also show very good agreement between the
linearized theory and the full numerical simulations.
\begin{figure}
\vspace{-2.5cm}
\epsfxsize=200pt \epsfbox{fig1.ps}
\vspace{-0.5cm}
\caption{Comparison of results for the radiated
energy vs initial separation in a collision for the close
approximation, the fully numerical results and the far
approximation. Vertical scale is logarithmic. }
\end{figure}
All this shows that the use of linearized gravity in the close limit
can be a valuable aid to full numerical evolutions of the two black
hole problem. It is therefore quite tempting to apply the linearized
treatment to more interesting situations, specifically the
in-spiraling collision of two black holes with angular
momentum. There are two main obstacles to doing this computation and we
will detail them in the next two sections.
\section{Second order perturbations: giving the formalism
error bars}
Assuming initial data for a black hole collision is given, we can
rather easily evolve and compute energies in linearized theory. Why
therefore not do it for the in-spiraling collision? The main reason
is that for that case there are no numerical results with which to
compare and the linearized formalism does not have a measure of error
in it: it therefore has little predictive power. There is no
consistent way to say when the close approximation breaks down. In
fact, this example teaches us a valuable lesson about perturbation
theory: when is linearized perturbation valid? The obvious answer
``when perturbations are small'' is clearly naive. To begin with,
``small'' should be characterized in a coordinate invariant
way. Moreover, as this example shows, perturbations can be ``large''
and perturbation theory can still be valid: it just needs to happen
that the perturbations be large in regions of spacetime that do not
contribute in a significant way to the physics of interest. In the two
black hole example, such a region is the interior of the horizon and
regions close to it, in which perturbations mostly fall into the black
hole.
How is one to characterize when to trust the approximation? The answer is
simple: work out the second order perturbations, compute the physical
quantities of interest and use how much the first and second order
calculation differ as a measure of the accuracy of the first order
results. The advantage of this answer is that it is phrased in terms of
what one is exactly interested in: the physical quantities. In the case
of the collision of two black holes these are the radiated waveforms
and energies.
The formalism for second order perturbations of black holes has not been
worked out in the past. It can be studied in detail as we do in reference
\cite{zeri2}. Here I just sketch some of the outstanding points. It
turns out that all the information can be coded into a single variable,
exactly as in the first order perturbation case and that that variable
satisfies a ``Zerilli equation'',
\begin{equation}
-{\partial \psi^{(2)} \over \partial t} +
{\partial \psi^{(2)} \over \partial r_*} +V(r) \psi^{(2)} = S
\end{equation}
where $r_*=r+\log(r/2M-1)$ and the Zerilli function $\psi^{(2)}$ is a
coordinate invariant combination of the perturbed metric coefficients.
This equation is exactly the same as the one satisfied by the first
order perturbations (including the ``potential'' V(r), which can be seen
in reference \cite{PrPu}). However, there is an important difference:
the right-hand side is not zero but a ``source'' term S, which is listed
explicitly in reference \cite{zeri2} and which is a complicated function
quadratic in the first order perturbations and their derivatives. The
way in which we derived this equation is to compute a particular
combination of the Einstein equations, writing the perturbed metric in a
particular coordinate system, the so called ``Regge-Wheeler'' gauge.
This, in turn is a way of deriving the original Zerilli equation. The
expression we get for $\psi^{(2)}$ is therefore a representation in that
gauge of a gauge invariant quantity. The explicitly gauge invariant form
of $\psi^{(2)}$ can also be computed.
We therefore are in a position to evolve to second order the problem of
black hole collisions and therefore to endow the first order predictions
with ``error bars''. This will be crucial for the inspiralling case,
where numerical results are not expected for some time.
\section{Initial data in the close approximation}
In the head-on collision case we were lucky to have an exact solution to
the initial value problem that we could evolve. For the more realistic
cases there are no exact solutions available at present and it is
unlikely that they will be easily found in the future.
There is an immediate alternative at hand. There exist already well
tested numerical codes \cite{Cook} for solving the initial value
problem in general relativity in the context of black hole collisions.
One could simply take these initial data evaluated for the case in which
the black holes are close and ``read off'' from them the departures from
Schwarzschild to be evolved using the linearized theory. This is
certainly possible and has already been illustrated for
Brill-Lindquist-type initial data by Abrahams and Price \cite{AbPr}.
Apart from the possibility of using numerical initial data for
realistic collisions it is interesting to notice that one can, up to a
certain extent, solve the initial value problem analytically if one is
only interested in initial data for the close approximation. The idea
is simple: in the close approximation the initial data for a black
hole collision departs a small amount from the initial data for a
Schwarzschild spacetime for a single black hole with mass equal to the
sum of the masses of the colliding holes. Therefore one can develop an
approximation technique for the initial data starting from the initial
data of Schwarzschild and adding small corrections proportional to the
separation of the holes. We illustrate here only the zeroth order
results, details will be given in a forthcoming paper in collaboration
with John Baker.
The initial value problem of general relativity can be conveniently cast
in the conformal formalism \cite{Yo}. One is interested in solving the
momentum and Hamiltonian constraints
\begin{eqnarray}
\nabla^a (K_{ab} - g_{ab} K) &=& 0\\
{}^3R-K_{ab} K^{ab} + K^2 &=&0
\end{eqnarray}
where $g_{ab}$ is the spatial metric, $K_{ab}$ is the extrinsic
curvature and ${}^3R$ is the scalar curvature of the three metric. One
proposes a three metric that is conformally flat
$g_{ab} = \psi^4 \delta_{ab}$, with $\psi^4$ the conformal factor
and a decomposition of the extrinsic curvature
$\widehat{K}_{ab} = \psi^{-2} K_{ab}$.
The constraints become,
\begin{eqnarray}
\widehat{\nabla}^a \widehat{K}_{ab} &=& 0\\
\widehat{\nabla}^2 \psi &=& \psi^{-7} \widehat{K}_{ab} \widehat{K}^{ab}.
\end{eqnarray}
where $\widehat{\nabla}$ is a derivative with respect to the flat
spacetime. Since the momentum constraint is linear, one can propose as
a solution for it for the case of two black holes the sum of the
solutions for the case of individual holes\footnote{The particular
solution chosen depends on the boundary conditions imposed. This may
add other terms to the simple ones we list here for brevity, but they
all behave in a similar fashion with respect to the approximations we
will consider.} with momentum $P_a$,
\begin{equation}
\hat{K}_{ab} = {3 \over 2 r^2} \left[ P_{(a} n_{b)} -(\delta_{ab}
-n_a n_b)P^c n_c\right]
\end{equation}
where $n_{b}$ is a unit normal in the direction of $\vec{r}$ and all
vector fields are defined in the flat background spacetime.
One now can put this solution in the Hamiltonian constraint and one is
left with an elliptic, highly non-linear equation for $\psi$. This is
the equation that is usually solved numerically. There exist
situations, however, where one can make some progress
analytically. Consider the case in which the momenta of the holes is
small \cite{Yo}. In that case one can neglect the right-hand side of
the Hamiltonian constraint and one only needs to solve a vacuum
Laplace equation for $\psi$. The solution can therefore be very simply
found, the difficulty depending on the boundary conditions one chooses
for the problem (typically a ``symmetrized'' boundary condition is
imposed, which complicates calculation quite a bit in certain cases,
see \cite{Cook} for details).
Another situation in which one can obtain an approximate solution is in
the ``close approximation''. In that case one has two black holes of
momenta equal and opposite $P^{(1)}_a=-P^{(2)}_a$, and since the black
holes are close, the unit normals appearing in the form for the
extrinsic curvature for each hole are approximately equal. That implies
that the extrinsic curvature for the problem is approximately zero (as
it should, since in the close limit the problem looks like a
Schwarzschild black hole at rest.) Therefore one can again neglect the
right-hand side of the Hamiltonian constraint and one is again left with
a Laplace equation. Let us compare this approximation with the full
numerical results. In order to do this we will compare the ADM energy of
initial data for a collision of two holes of momentum $P$. The ADM
energy in the conformal formalism is given by
\begin{equation}
E=-{1\over 2 \pi}\oint_\infty \nabla_i\psi \, d^2 S^i
\end{equation}
and we notice that it does not depend explicitly on the extrinsic
curvature (it does implicitly via the constraints). Therefore at the
approximation we are working, in which the constraints do not couple the
conformal factor and the extrinsic curvature, the energy is independent
of the extrinsic curvature and therefore independent of the momenta of
the holes. We compare this prediction with the full numerical results of
Cook in figure \ref{cook}.
\begin{figure}
\vspace{-2.5cm}
\epsfxsize=200pt \epsfbox{fig2.ps}
\vspace{-0.5cm}
\caption{The ADM energy of
initial data for collisions of black holes of momenta
$P$. The dots are the full numerical results of Cook, for different
values of the initial separation $\beta$. We see that for small
separations, the energy is approximately independent of the holes
momenta, which coincides with the close approximation prediction,
depicted by the solid line.}
\label{cook}
\end{figure}
An interesting aspect is that one can advance this approximation one
step further. One can input the extrinsic curvature and the conformal
factor found as a fixed ``source'' in the equation determining the
conformal factor and one can obtain a correction through the integration
of a Poisson equation. Comparison of this approximation with the
numerical data is currently in progress. Details are complicated by the
particular boundary conditions that are usually chosen in the numerical
computations.
It is evident that the ``close approximation'' can work in many other
cases, apart from the head-on, equal momenta holes we considered here.
The only changes will be that the solution one obtains in the ``close
limit'' rather than being a slice of Schwarzschild will be a slice of
Kerr or boosted Schwarzschild if the net result of the collision has
angular momentum or linear momentum.
\section{Summary}
We have seen that the use of the ``close approximation'' can be a
valuable aid to full numerical computations of the collision of two
black holes. With the introduction of a second order scheme we are now
in a position of offering reliable estimates of energies and waveforms
that we expect people working on the full numerical simulations will
find of use to calibrate codes and design strategies for better
integrating the Einstein equations in this problem of great current
physical interest.
\section*{Acknowledgments}
The work described here is in collaboration with Richard Price, John
Baker, Reinaldo Gleiser and Oscar Nicasio. I acknowledge support of NSF
through grants PHY94-06269 PHY93-96246, funds of the Pennsylvania State
University, its Office for Minority Faculty Development, and the Eberly
Family research fund.
|
\section{\@startsection {section}{1}{\z@}{-2.5ex plus-1ex minus
\newcommand{V_{ub}}{V_{ub}}
\newcommand{V_{cb}}{V_{cb}}
\newcommand{b\to u\ell\nu}{b\to u\ell\nu}
\newcommand{b\to c\ell\nu}{b\to c\ell\nu}
\newcommand{\Upsilon(4S)}{\Upsilon(4S)}
\newcommand{GeV$/c$}{GeV$/c$}
\newcommand{MeV$/c$}{MeV$/c$}
\newcommand{\pi\ell\nu}{\pi\ell\nu}
\newcommand{\rho\ell\nu}{\rho\ell\nu}
\newcommand{\omega\ell\nu}{\omega\ell\nu}
\newcommand{\rho^-\ell^+\nu}{\rho^-\ell^+\nu}
\newcommand{\rho^0\ell^+\nu}{\rho^0\ell^+\nu}
\newcommand{\omega\ell^+\nu}{\omega\ell^+\nu}
\newcommand{\pi^-\ell^+\nu}{\pi^-\ell^+\nu}
\newcommand{\pi^0\ell^+\nu}{\pi^0\ell^+\nu}
\newcommand{\rho^\mp\ell^\pm\nu}{\rho^\mp\ell^\pm\nu}
\newcommand{\rho^0\ell^\pm\nu}{\rho^0\ell^\pm\nu}
\newcommand{\omega\ell^\pm\nu}{\omega\ell^\pm\nu}
\newcommand{\pi^0\ell^\pm\nu}{\pi^0\ell^\pm\nu}
\newcommand{\pi^\mp\ell^\pm\nu}{\pi^\mp\ell^\pm\nu}
\newcommand{{\cal B}(B^0\to\pimlv)}{{\cal B}(B^0\to\pi^-\ell^+\nu)}
\newcommand{{\cal B}(B^0\to\rhomlv)}{{\cal B}(B^0\to\rho^-\ell^+\nu)}
\newcommand{\Gamma(B^0\to\pimlv)}{\Gamma(B^0\to\pi^-\ell^+\nu)}
\newcommand{\Gamma(B^0\to\rhomlv)}{\Gamma(B^0\to\rho^-\ell^+\nu)}
\newcommand{ B\to\pilv}{ B\to\pi\ell\nu}
\newcommand{ D\to K\ell\nu}{ D\to K\ell\nu}
\newcommand{ D\to\pilv}{ D\to\pi\ell\nu}
\newcommand{ B\to\rholv}{ B\to\rho\ell\nu}
\newcommand{ D\to\rholv}{ D\to\rho\ell\nu}
\newcommand{ D\to K^{*}\ell\nu}{ D\to K^{*}\ell\nu}
\newcommand{q^2_{max}}{q^2_{max}}
\newcommand{E_{miss}}{E_{miss}}
\newcommand{\vec{p}_{miss}}{\vec{p}_{miss}}
\newcommand{\Delta E}{\Delta E}
\newcommand{m_B}{m_B}
\newcommand{\vec{p}_\nu}{\vec{p}_\nu}
\newcommand{\vec{p}_\pi}{\vec{p}_\pi}
\newcommand{\vec{p}_\ell}{\vec{p}_\ell}
\newcommand{M_{miss}^2}{M_{miss}^2}
\newcommand{N_{\pi^\pm\ell^\mp\nu}}{N_{\pi^\pm\ell^\mp\nu}}
\newcommand{N_{\rho^\pm\ell^\mp\nu}}{N_{\rho^\pm\ell^\mp\nu}}
\newcommand{\e}[1]{\times10^{#1}}
\newcommand{\theta^*_{\pi\ell}}{\theta^*_{\pi\ell}}
\newcommand{\cos\thepil}{\cos\theta^*_{\pi\ell}}
\newcommand{fb$^{-1}$}{fb$^{-1}$}
\newcommand{B\bar{B}}{B\bar{B}}
\newcommand{{\it et al.}}{{\it et al.}}
\newcommand{\plb}[1]{Phys. Lett. {\bf B#1}}
\newcommand{\prl}[1]{Phys. Rev. Lett. {\bf #1}}
\newcommand{\npb}[1]{Nucl. Phys. {\bf B#1}}
\newcommand{\prd}[1]{Phys. Rev. {\bf D#1}}
\newcommand{\zpc}[1]{Z. Phys. {\bf C#1}}
\newcommand{\nim}[1]{Nucl. Instrum. Methods Phys. Res. Sect. A {\bf #1}}
\newcommand{\begin{equation}}{\begin{equation}}
\newcommand{\end{equation}}{\end{equation}}
\newcommand{\hspace*{0.5em}}{\hspace*{0.5em}}
\newcommand{\hspace*{0.5em}}{\hspace*{0.5em}}
\renewcommand{\topfraction}{1.0}
\renewcommand{\bottomfraction}{1.0}
\renewcommand{\textfraction}{0.0}
\renewcommand{\floatpagefraction}{0.75}
\setcounter{totalnumber}{2}
\begin{document}
\begin{titlepage}
\begin{center}
\vspace*{1.0 truecm}
{\elevenrm \hfill UR-1427}\\
{\elevenrm \hfill June, 1995}\\
\vspace*{2.5 truecm}
{\twelvebf PRELIMINARY RESULTS FOR EXCLUSIVE \boldmath$b\to u\ell\nu$\\
DECAYS FROM CLEO}\\
\vspace{1 truecm}
{\elevenrm Lawrence Gibbons}\\
{\elevenit University of Rochester, Rochester, NY 14627}\\
\newlength{\abskip}
\setlength{\abskip}{4.0 truecm}
\addtolength{\abskip}{-6\baselineskip}
\vspace*{\abskip}
\begin{abstract}
{ \noindent
A preliminary analysis of exclusive $b\to u\ell\nu$ decays to the final states
$\pi^\mp\ell^\pm\nu$, $\pi^0\ell^\pm\nu$, $\rho^\mp\ell^\pm\nu$, $\rho^0\ell^\pm\nu$ and $\omega\ell^\pm\nu$ based on $2.2\e{6}$
$B\bar{B}$ decays collected at CLEO is presented. We have measured the first
exclusive $b\to u\ell\nu$ branching fraction ${\cal B}(B^0\to\pimlv)=[1.19\pm0.41\pm0.21\pm0.19]\e{-4}$
($[1.70\pm0.51\pm0.31\pm0.27]\e{-4}$), with the ISGW (WSB) model used for
efficiency determination. A 90\% C.L. upper limit on ${\cal B}(B^0\to\rhomlv)$ similar to the
previous CLEO limit is obtained. The ratio $\Gamma(B^0\to\rhomlv)/\Gamma(B^0\to\pimlv)<3.4$ at the 90\%
confidence level for both the ISGW and WSB models. This ratio provides some
discrimination between form factor models.
}
\end{abstract}
\vspace*{2.0 truecm}
{\elevenit To appear in the Proceedings of the XXXth Rencontres de Moriond\\
``Electroweak Interactions and Unified Theories'',\\
Les Arcs, France, March, 1995}
\end{center}
\end{titlepage}
\section{Introduction}
This talk will focus on a preliminary CLEO analysis of $b\to u\ell\nu$ decays to the
exclusive final states $\pi\ell\nu$, $\rho\ell\nu$ and $\omega\ell\nu$. The ultimate goal of
this analysis is to improve our knowledge of $|V_{ub}|$.
ARGUS \cite{bb:argus_inclusive} and CLEO \cite{bb:cleo_inclusive} have already
demonstrated that $|V_{ub}|>0$ by examining the inclusive lepton momentum
spectrum from $B$ decays at the $\Upsilon(4S)$. They observe events beyond 2.4 GeV$/c$,
which is kinematically forbidden for the copious $b\to c\ell\nu$ processes, but is
still accessible to $b\to u\ell\nu$ decays. While these analyses clearly establish an
excess in this endpoint region, and hence that $|V_{ub}|>0$, extracting a
reliable value of $|V_{ub}|$ is difficult because of the theoretical uncertainty
in extrapolating from the observed rate in the endpoint region to the total
$b\to u\ell\nu$ rate. Values of $|V_{ub}/V_{cb}|$ obtained from these analyses are now in
the 7\% to 11\% range, with the theoretical uncertainty dominating.
\section{Exclusive $b\to u\ell\nu$}
An alternate route to $|V_{ub}|$ is through the study of exclusive $b\to u\ell\nu$
channels. The best previous information concerning such channels is the
upper limit set by CLEO \cite{bb:UCSBlimit} in the combined modes $\rho^-\ell^+\nu$,
$\rho^0\ell^+\nu$ and $\omega\ell^+\nu$. The CLEO result
corresponds to an upper limit of ${\cal B}(B^0\to\rhomlv)<3.2\e{-4}$ at the 90\% confidence
level (ISGW model \cite{bb:isgw}).
The preliminary analysis presented here studies the two pseudoscalar modes
$\pi^-\ell^+\nu$ and $\pi^0\ell^+\nu$, the three vector modes $\rho^-\ell^+\nu$, $\rho^0\ell^+\nu$ and
$\omega\ell^+\nu$, and the charge conjugate modes. At a fixed $|V_{ub}|$, the existing
form factor models predict a wide range of partial widths for these modes, as
Table~\ref{tab:exclpred} shows. Unfortunately, measured branching
fractions depend on the form factor model used to evaluate the experimental
efficiencies, as does the extraction of $|V_{ub}|$. We therefore need to
discriminate between the different models.
\begin{table}[b]
\centering
\caption{Predictions for the exclusive partial widths $\Gamma(B^0\to\pimlv)$ and $\Gamma(B^0\to\rhomlv)$
and the ratio $\Gamma(B^0\to\rhomlv)/\Gamma(B^0\to\pimlv)$. The partial width units are
$10^{12}|V_{ub}|^2$ sec$^{-1}$.}
\label{tab:exclpred}
\begin{tabular}{lccc} \hline\hline
Model & $\Gamma(B^0\to\pimlv)$ & $\Gamma(B^0\to\rhomlv)$ & $\Gamma(B^0\to\rhomlv)/\Gamma(B^0\to\pimlv)$\\ \hline
WSB \cite{bb:WSB} & 6.3 -- 10.0 & 18.7 -- 42.5 & 3.0 -- 4.3 \\
KS \cite{bb:KS} & 7.25 & 33.0 & 4.6 \\
ISGW \cite{bb:isgw} & 2.1 & 8.3 & 4.0 \\
ISGW II \cite{bb:ISGWII} & 9.6 & 14.2 & 1.5 \\ \hline
\end{tabular}
\end{table}
The ratio $\Gamma(B^0\to\rhomlv)/\Gamma(B^0\to\pimlv)$ provides one means of discrimination. Because the
$\pi\ell\nu$ rate is helicity-suppressed when the daughter meson is at rest in the B
meson rest frame (at $q^2_{max}$), where the form factors for the decay are
largest, while the $\rho\ell\nu$ rate is not, we expect the ratio to be larger than
one. The exact value for the ratio will depend on the $q^2$-behavior of the
form factors. In Table~\ref{tab:exclpred}, we see that the predictions of the
ratio span a fairly broad range, so the ratio should prove useful.
\section{Neutrino ``Measurement'' and Exclusive $b\to u\ell\nu$}
Experimentally, semileptonic decays are troublesome because of the undetected
neutrino. This analysis takes advantage of the excellent hermeticity and
resolution of the CLEO II detector located at the Cornell Electron Storage Ring
(CESR) to obtain information about the neutrino in semileptonic $b\to u\ell\nu$
decays. Three concentric tracking devices provide a momentum resolution of
$\sigma_p/p=0.005\oplus 0.0015p$ ($p$ in GeV$/c$), while covering 95\% of the
$4\pi$ solid angle. The CsI calorimeter located inside of the CLEO solenoid
provides an energy resolution well approximated by
$\sigma_E/E=0.019+0.0035/E^{0.75}-0.001E$ ($E$ in GeV), while covering 98\% of
$4\pi$. The detector is described in detail elsewhere \cite{bb:nim}. This
analysis is based on a data sample with a luminosity of $2.09$ fb$^{-1}$\
(about $2.2\e{6}$ $B\bar{B}$ decays).
The underlying idea is very simple: the $B\bar{B}$ system is at rest at CLEO
and the beam energy is known very precisely, so we can ``measure'' the neutrino
four momentum by measuring the missing energy and momentum of an event. We
define
\begin{eqnarray}
E_{miss} & \equiv & 2E_{beam} - \sum_i E_i \\
\vec{p}_{miss} & \equiv & -\sum_i \vec{p}_i,
\label{eq:missdef}
\end{eqnarray}
where the index $i$ runs over all charged tracks and all showers in the
calorimeter that pass cuts designed to reject false tracks and spurious
showers from hadronic interactions.
In events with no extra missing particles,
$\vec{p}_{miss}$ can be reliably associated with the momentum $\vec{p}_\nu$ of the signal
mode neutrino. The $b\to u\ell\nu$ decay can then be fully reconstructed: the energy
difference $\Delta E\equiv E_{beam} - (E_h+E_\ell+|\vec{p}_\nu|)$, where $h$ is the
candidate hadron, should be zero, and the beam-energy constrained mass
$m_B\equiv\sqrt{E_{beam}^2-|\vec{p}_h+\vec{p}_\ell+\vec{p}_\nu|^2}$ should reconstruct at the
$B$ mass. Signal events that are reconstructible show resolutions of
approximately 260 MeV on $E_{miss}$ and 110 MeV on $|\vec{p}_{miss}|$.
Signal events with particles missing in addition to the neutrino usually fail
the reconstruction criteria. On the other hand, those background events that
pass the criteria do so because they have extra particles missing.
Consequently, we reject events with multiple leptons or a non-zero total charge
because they indicate a second neutrino or a missed charged particle,
respectively. Most remaining events with extra missing particles are
eliminated by requiring that $M_{miss}^2\equivE_{miss}^2 - |\vec{p}_{miss}|^2$ be consistent
with zero. The criterion $M_{miss}^2/2E_{miss}<350$ MeV is used since the $M_{miss}^2$
resolution varies approximately as $2E_{miss}\sigma_{E_{miss}}$.
Continuum background is suppressed using standard event shape variables. The
$b\to u\ell\nu$ processes are enhanced over $b\to c$ by requiring the leptons to have
momenta larger than 1.5 GeV$/c$\ (2.0 GeV$/c$) in the $\pi\ell\nu$
(vector) modes. The lower cut is used in the $\pi$ modes because
these modes are expected to have a softer lepton momentum spectrum.
Both electrons and muons are used in this analysis. We combine information
from specific ionization, energy/momentum measurements from the calorimeter
and tracking systems, and position matching from these two systems to identify
electrons down to 600 MeV$/c$. Muon candidates must register hits in muon
counters at least 5 interaction lengths deep, limiting the muon momentum range
to approximately 1.4 GeV$/c$. The probability that a hadron is misidentified as
a lepton (a ``fake lepton'') is of the order 0.1\% (1\%) for electron (muon)
identification.
Candidate $2\pi$ ($3\pi$) combinations must have an invariant mass within 90
(30) MeV of the nominal $\rho$ ($\omega$) mass. A $\pi^0$ candidate must have
a 2-photon invariant mass within 2 standard deviations (about 12 MeV) of the
$\pi^0$ mass. Within any one of the five modes, we pick the meson candidate in
each event that yields the smallest value of $|\Delta E|$.
We require the lepton, neutrino and meson candidates to satisfy $-250 \mbox{
MeV}<\Delta E<150\mbox{ MeV}$. The cut is asymmetric because the $b\to c$
backgrounds increase rapidly as $\Delta E$ increases. The range $5.265 \mbox{
GeV}<m_B<5.2875 \mbox{ GeV}$ defines the signal region. The $m_B$ distribution
for data after all cuts, including the $\Delta E$ cut, is shown in
Figure~\ref{fig:masses} for the combination of the $\pi^\mp\ell^\pm\nu$ and $\pi^0\ell^\pm\nu$
modes, and for the combination of the three vector modes. There is a clear
excess above the background in the signal region for the $\pi\ell\nu$ modes. The
fit yielding the background levels shown is described in the next section.
The dominant background in both the $\pi$ and the vector modes comes from
$b\to c\ell\nu$ decays in events containing either an undetected $K_L$ or a second
neutrino. The small backgrounds in each mode from fake leptons and from
continuum processes are measured with the data. In the $\pi$ modes, Monte
Carlo studies indicate that feed-across from the $\rho\ell\nu$ modes should
contribute the next largest background. In the vector modes, $b\to u\ell\nu$ decays to
higher mass and non-resonant final states form the other major background
component. Our fits do not make any requirement on the distribution of events
inside our mass windows, so resonant and non-resonant final states are not
distinguished. Consequently only an upper limit on ${\cal B}(B^0\to\rhomlv)$ will be
obtained. We derive the limit conservatively by assuming zero background from
the non-resonant and higher mass decays.
\begin{figure}[t]
\centering
\leavevmode
\epsfbox{morart_mass.eps}
\caption{Beam constrained mass distributions for the combined $\pi^-\ell^+\nu$ and
$\pi^0\ell^+\nu$ modes (left) and the combined vector modes (right). The points are
continuum- and fake-subtracted data. The histograms show the contribution from
$b\to c$ (shaded), $u\ell\nu$ crossfeed (hatched) and signal (hollow).}
\label{fig:masses}
\end{figure}
\section{Extracting the Yields}
\begin{table}[t]
\centering
\caption{Backgrounds, efficiencies and fit results for the $\pi\ell\nu$ analysis.
The $\chi^2$ for the fits using the ISGW and WSB signal models were
respectively 10.8 and 10.3 for $20-7$ degrees of freedom. Note that the errors
on the signal yields and crossfeed backgrounds in the $\pi^-\ell^+\nu$ and $\pi^0\ell^+\nu$
modes are completely correlated because of the isospin constraints.}
\label{tab:pifit}
\begin{tabular}{lcccc} \hline\hline
& \multicolumn{2}{c}{$\pi^-\ell^+\nu$} &
\multicolumn{2}{c}{$\pi^0\ell^+\nu$} \\ \cline{2-5}
& ISGW & WSB &
ISGW & WSB \\ \hline
Raw Data & \multicolumn{2}{c}{30} &
\multicolumn{2}{c}{15} \\
Continuum Bkg. & \multicolumn{2}{c}{$2.3\pm0.8$} &
\multicolumn{2}{c}{$1.0\pm0.5$} \\
Fake Lepton Bkg. & \multicolumn{2}{c}{$1.2\pm0.3$} &
\multicolumn{2}{c}{$0.7\pm0.2$} \\
other $u\ell\nu$ Bkg. & \multicolumn{2}{c}{0.6} & \multicolumn{2}{c}{0.2} \\
Efficiency & 2.9\% & 2.1\% & 1.9\% & 1.4\% \\
Signal Yield & $15.6\pm5.3$ & $16.3\pm5.3$ & $5.0\pm1.7$ & $5.3\pm1.7$ \\
$b\to c$ Bkg. & $9.8\pm1.1$ & $9.8\pm1.1$ & $1.8\pm0.5$ & $1.7\pm0.5$ \\
$\rho/\omega$ Bkg.&$3.8\pm1.7$ & $3.4\pm1.4$ & $1.8\pm0.8$ &
$1.6\pm0.7$ \\ \hline
\end{tabular}
\end{table}
After subtracting the continuum and fake lepton backgrounds, we fit the
beam-constrained mass distributions in our five reconstructed $b\to u\ell\nu$ modes
simultaneously, which allows the data in the vector modes to constrain the
$\rho\ell\nu$ background in the $\pi\ell\nu$ modes. In addition to the signal shapes and
the feed-across shapes between the five modes, the fit includes $b\to c$ and
other $b\to u\ell\nu$ background components. The isospin relations
$\frac12\Gamma(B^0\to\pi^-\ell^+\nu) = \Gamma(B^+\to\pi^0\ell^+\nu)$ and
$\frac12\Gamma(B^0\to\rho^-\ell^+\nu) = \Gamma(B^+\to\rho^0\ell^+\nu) \approx
\Gamma(B^+\to\omega\ell^+\nu)$ constrain the neutral meson rates relative to the charged
meson rates. We therefore obtain two yields, $N_{\pi^\pm\ell^\mp\nu}$ and $N_{\rho^\pm\ell^\mp\nu}$, from the fit.
The $b\to c\ell\nu$ and feed-across background shapes in $m_B$ are obtained from Monte
Carlo simulation. The $b\to c\ell\nu$ background level floats independently in each of
the five modes, while the feed-across rates between the five modes are tied to
the signal yields $N_{\pi^\pm\ell^\mp\nu}$ and $N_{\rho^\pm\ell^\mp\nu}$.
Monte Carlo simulation also provides the $m_B$ distributions for the
non-resonant and higher mass $b\to u\ell\nu$ backgrounds. The inclusive lepton yield at
high momentum fixes this background level. We vary the physical model and the
rate by hand to estimate the systematic uncertainty in this procedure.
The results of the fit from which the $\pi\ell\nu$ yield (and the background levels
in Figure~\ref{fig:masses}) is obtained are summarized in
Table~\ref{tab:pifit}. The efficiencies and crossfeed probabilities have been
determined using the ISGW and WSB models. We obtain similar $\pi\ell\nu$ yields for
the two models, but obtain efficiencies that differ by approximately 30\%. The
$b\to c\ell\nu$ background levels in the five modes are all consistent with absolute
Monte Carlo predictions based on the luminosity. Correcting for acceptance and
averaging the electron and muon samples, we obtain the preliminary branching
fraction ${\cal B}(B^0\to\pimlv)=[1.19\pm0.41]\e{-4}$ ($[1.70\pm0.55]\e{-4}$) for the ISGW (WSB)
model, where the errors are statistical only. We obtain consistent results if
we fit using the $\Delta E$ distributions, having resolved multiple candidates
using $m_B$.
To obtain upper limits for the vector modes, we perform a similar fit assuming
no non-resonant or higher mass $b\to u\ell\nu$ backgrounds. This fit gives the
same $\pi\ell\nu$ yield. We obtain the efficiency-corrected numbers
of $834\pm337$ ($1248\pm484$) $\rho^\mp\ell^\pm\nu$ decays for the ISGW (WSB) model, and a
$\Gamma(B^0\to\rhomlv)/\Gamma(B^0\to\pimlv)$ ratio of $1.56^{+1.29}_{-0.76}$ ($1.63^{+1.21}_{-0.75}$).
\begin{figure}[b]
\centering
\leavevmode
\epsfbox{morart_lepthe.eps}
\caption{Charged lepton spectrum (left) and $\cos\thepil$ distribution (right) for
the combined $\pi^-\ell^+\nu$ and $\pi^0\ell^+\nu$ modes. The
points are continuum- and fake-subtracted data. The top histogram is the total
prediction using rates from the yield fit, with components $b\to c$ (shaded),
$u\ell\nu$ crossfeed (hatched) and signal (dashed).}
\label{fig:lepspec}
\end{figure}
Many distributions have been examined for consistency with the $\pi\ell\nu$
hypothesis. The charged lepton momentum spectrum for $\pi^\mp\ell^\pm\nu$ and $\pi^0\ell^\pm\nu$
candidates in the $m_B$ signal region is shown in Figure~\ref{fig:lepspec}.
The spectrum obtained from the data is quite stiff, with a sizeable fraction of
events beyond the $b\to c\ell\nu$ endpoint. The sum of the signal and background
distributions, scaled according to the fit results, shows good agreement with
the data. The $\pi$ and $\nu$ momentum spectra are also consistent with the
results of the fit.
For $B\to\pi\ell\nu$, the $V-A$ interaction predicts that the angle between the
$\pi$ and the lepton in the $W$ rest frame, $\theta^*_{\pi\ell}$, should have a
$\sin^2\theta^*_{\pi\ell}$ distribution. The observed $\cos\thepil$ distribution, also shown
in Figure~\ref{fig:lepspec}, is in good agreement with this expectation. We
estimate the probability, including systematic uncertainties, that the
background processes could fluctuate to give the observed $m_B$ and $\cos\thepil$
distributions in the combined $\pi\ell\nu$ modes, and obtain $6.4\e{-5}$. This
corresponds to a 3.8 standard deviation significance for a Gaussian
distribution.
\section{Systematics}
\begin{table}[t]
\centering
\caption{Summary of systematic uncertainties on the yields and efficiencies in
the $\pi\ell\nu$ and $\rho\ell\nu$ modes. The numbers in parentheses in the background
levels indicate the uncertainty in the background as a fraction of that
background.}
\label{tab:syst}
\begin{tabular}{lrr||lrrr} \hline\hline
On yields: & $\pi\ell\nu$ & $\rho\ell\nu$ &
On Efficiencies: & $\pi\ell\nu$ & $\rho\ell\nu$ & $\rho/\pi$ ratio
\\ \hline
$b\to c$ bkg. & (20\%) 13\% & (20\%) 23\% &
$\nu$-measurement & 15\% & 15\% & 15\% \\
$\rho/\omega\ell\nu$ bkg.& (36\%)\hspace*{0.5em}\ 8\%& (63\%)\hspace*{0.5em}\ 7\% &
$\pi/\rho/\omega$ finding& 3\% & 6\% & 7\% \\
other $u\ell\nu$ bkg. & 8\% & --- &
$\rho/\omega$ polarization & --- & 10\% & 10\% \\
cont.+fake bkg. & (20\%)\hspace*{0.5em}\ 6\%& (24\%)\hspace*{0.5em}\ 7\% &
lepton fake rates & 4\% & 4\% & 4\% \\
lepton finding & 2\% & 2\% &
lepton finding & 4\% & 4\% & 4\% \\
& & &
Luminosity & 2\% & 2\% & --- \\ \hline
{\bf Total} & {\bf 18\%} & {\bf 25\%} &
{\bf Total} & {\bf 16\%} & {\bf 20\%} & {\bf 20\%} \\ \hline
\end{tabular}
\end{table}
\begin{figure}[b]
\centering
\leavevmode
\epsfbox{morart_dstar.eps}
\caption{$|\vec{p}_{miss}|$ spectrum (left) $m_B$ distribution (center) and $\Delta E$
distribution (right) for $D^{*\pm}\ell^\mp\nu$ reconstruction. The points are
continuum and combinatoric background-subtracted data. The histograms are
signal Monte Carlo distributions normalized to equal area.}
\label{fig:dstar}
\end{figure}
The systematic uncertainties on the yields and efficiencies are summarized in
Table~\ref{tab:syst}. The dominant uncertainty in the yields comes from the
uncertainty in the shapes of the background $m_B$ distributions. The shapes
have been checked in a variety of ways: examining the shapes in $\Delta E$
sidebands and in signal-free modes (eg., $K_S\ell\nu$), and varying the
misreconstruction behavior of the Monte Carlo simulation.
The uncertainty in the efficiencies is dominated by the
neutrino-measurement simulation. One method of estimating this uncertainty is
to use this technique to measure the branching fraction for $B\to
D^{*\pm}\ell^\mp\nu$ via the modes $D^{*\pm}\to\pi^\pm D^0$, $D^0\to
K^\mp\pi^\pm$. We find that the simulation of $E_{miss}$, $\vec{p}_{miss}$, $\Delta E$ and
$m_B$ agrees well with the data (Figure~\ref{fig:dstar}), and that ${\cal
B}(B^0\to D^{*-}\ell^+\nu)=4.66\pm0.65\%$. This agrees with the published CLEO
result \cite{bb:dstar} of $4.49\pm0.32\pm0.32\%$, which used a
higher statistics technique. The 15\% statistical uncertainty is taken as the
systematic uncertainty; other studies indicate that this is a conservative
estimate. We expect this systematic to cancel in the $\rho/\pi$ ratio,
but retain a preliminary 15\% uncertainty.
\section{Conclusion}
Combining the $\pi\ell\nu$ yields and the systematic uncertainties, we obtain the
preliminary branching fraction ${\cal
B}(B^0\to\pi^-\ell^+\nu)=[1.19\pm0.41\pm0.21\pm0.19]\e{-4}$
($[1.70\pm0.51\pm0.31\pm0.27]\e{-4}$) using the ISGW (WSB) model to evaluate
efficiencies. The errors are statistical, systematic on the yield, and
systematic on the efficiency, respectively. This is the first measurement of
any exclusive $b\to u\ell\nu$ branching fraction. The probability
of a background fluctuation resulting in the observed signal is $6.4\e{-5}$.
Assuming no non-resonant or high mass $u\ell\nu$ background, we obtain a
conservative 90\% C.L. upper limit of ${\cal B}(B^0\to\rho^-\ell^+\nu) < 3.1\e{-4}$ for
the ISGW model and ${\cal B}(B^0\to\rho^-\ell^+\nu) < 4.6\e{-4}$ for the WSB model.
The statistical and systematic uncertainties have been combined in quadrature
in evaluating these limits. The results are comparable to the previous CLEO
upper limits for the vector modes.
Finally, we find $\Gamma(B^0\to\rhomlv)/\Gamma(B^0\to\pimlv)<3.4$ at the 90\% confidence level for both
the ISGW and WSB models. Again, statistical and systematic uncertainties have
been combined in quadrature. Comparing to the predictions in
Table~\ref{tab:exclpred}, the WSB model is compatible with this limit, but it
is difficult to reconcile the ISGW model with this limit.
These preliminary measurements herald a new era for the study of $V_{ub}$. CLEO
is still refining these measurements, with 50\% more data soon to be available
and work in progress on the separation of the vector modes from non-resonant
modes.
\section{References}
|
\section{Introduction}
\label{secintro}
The description of small-$x$\ partons within hadrons has attracted a great
deal of interest, especially after the measurements of the proton
structure function $F_2$ at $x$\ values down to $10^{-4}$ at HERA,
where a substantial increase was found \cite{H1struct93,ZEUSstruct93}.
Although the small-$x$\ rise was first predicted by the so-called BFKL
evolution equation \cite{BFKL77:1,BFKL77:2} it soon turned out
\cite{BallForte941,BallForte942} that it could also be explained in
terms of the conventional Altarelli--Parisi (DGLAP) evolution
equations \cite{DGLAP:1,DGLAP:2,DGLAP:3,DGLAP:4}. Instead, much of the
focus has been directed towards the study of hadronic final states in
deep inelastic lepton--hadron scattering (DIS) events at small $x$,
and it has been suggested that the large flow of transverse energy in
the proton direction found in such events is a signal of BFKL dynamics
\cite{MARTINetflow94}.
Much can also be learned from comparing data with different models
implemented in Monte Carlo event generators. So far it has been shown
that generators built around a conventional DGLAP-inspired
initial-state parton showers, such as \pythia\ \cite{JETPYT94,JETPYT93}
and \lepto\ \cite{LEPTO91}, with strong ordering in virtuality,
completely fail to describe things like the forward transverse energy
flow at small $x$, while a generator such as \ariadne\ \cite{ARIADNENOW}
-- although not implementing BFKL evolution, but sharing with it the
feature that emissions are unordered in transverse momenta --
describes such event features quite well \cite{H1flow94,H1flow95}.
Besides deep inelastic lepton--hadron scattering, Drell--Yan\ production in
hadron--hadron collisions is one of the cleanest probes of hadronic
structure. Recent results from the D0 collaboration \cite{GEOFFcdm94} at
the Tevatron shows a surprising feature of events with Drell--Yan-produced
$W$ bosons, namely the decorrelation in rapidity between the $W$ and
the associated jets. Although the typical $x$-values probed in $W$
events at the Tevatron is on the order of {\small $\sqrt{m_W^2/S}$}
$\approx 80/1800 \approx 0.04$, for large rapidities of the $W$, one
of the incoming partons has a much smaller momentum fraction of the
proton (e.g.\ $y_W \approx 2$ gives $x_1 \approx 6 \times 10^{-3}$ and
$x_2 \approx 0.3$). Therefore it could be worth while to take the
experience gained from studying small-$x$\ final states at HERA and try to
apply it to large rapidity $W$-production at the Tevatron.
In this paper the Dipole Cascade Model (DCM)
\cite{CDMinit86,CDMplain88}, on which the \ariadne\ program is built,
is extended to also model the jet structure of Drell--Yan\ production
events. The main feature of the DCM for DIS \cite{CDMdis89} is that
gluon emission is treated as final state radiation from the colour
dipole formed between the struck quark and the proton remnant as in
fig.\refig{figdisdip}a. In this way there is no explicit initial state
radiation, and the proton structure enters only in the way the dipole
radiation is suppressed due to the spatial extension of the proton
remnant. This approach has some problems when it comes to describing
features particular to the initial state, such as the initial splitting
of a gluon into a $\qq$-pair.
\begin{figure}
\setlength{\unitlength}{0.07mm}
\begin{picture}(2000,600)(-100,0)
\thicklines
\put(200,125){\line(1,0){600}}
\thinlines
\put(200,150){\line(1,0){200}}
\put(400,150){\vector(1,2){50}}
\put(400,150){\line(1,2){100}}
\put(500,350){\line(2,1){300}}
\put(500,350){\vector(2,1){150}}
\qbezier(500,350)(450,350)(450,390)
\qbezier(450,390)(450,430)(400,430)
\qbezier(400,430)(350,430)(350,470)
\qbezier(350,470)(350,510)(300,510)
\qbezier(300,510)(250,510)(250,550)
\qbezier(250,550)(250,590)(200,590)
\put(450,550){\makebox(0,0){$\gamma/Z^0$}}
\qbezier(850,450)(900,300)(850,150)
\put(850,450){\vector(-1,4){0}}
\put(850,150){\vector(-1,-4){0}}
\thicklines
\put(1200,125){\line(1,0){600}}
\put(1200,575){\line(1,0){600}}
\thinlines
\put(1200,150){\line(1,0){200}}
\put(1400,150){\vector(1,2){50}}
\put(1400,150){\line(1,2){100}}
\put(1500,350){\line(-1,2){100}}
\put(1500,350){\vector(-1,2){50}}
\put(1400,550){\line(-1,0){200}}
\multiput(1500,350)(50,0){5}{\line(1,0){25}}
\put(1650,300){\makebox(0,0){$W$}}
\qbezier(1825,150)(1875,350)(1825,550)
\put(1825,150){\vector(-1,-4){0}}
\put(1825,550){\vector(-1,4){0}}
\put(400,50){\makebox(0,0){(a)}}
\put(1400,50){\makebox(0,0){(b)}}
\end{picture}
\fcap{The colour dipoles that initiate the dipole cascade
in (a) DIS and (b) Drell--Yan\ production of $W$.}
\label{figdisdip}
\end{figure}
The simplest extension of the DCM to also treat Drell--Yan\ production would
be to treat gluon emission as final-state radiation from the colour
dipole formed between the two remnants, as in fig.\refig{figdisdip}b.
However, as is seen from fig.\refig{figordalp}, the leading order
$W$+jet diagrams all correspond to initial-state radiation (except for
the last one, which is the least important). And in particular it is
clear that if the gluon emission is treated as final-state radiation
between the two remnants, it would be difficult to explain the
contribution to the transverse momentum of the $W$ from the
diagrams in fig.\refig{figordalp}.
\begin{figure}
\setlength{\unitlength}{0.07mm}
\begin{picture}(2000,600)(-100,0)
\put(0,50){\line(1,1){150}}
\put(150,200){\line(0,1){200}}
\put(150,400){\line(-1,1){150}}
\multiput(150,400)(40,40){4}{\begin{picture}(70,70)
\qbezier(0,0)(30,30)(60,0)
\qbezier(60,0)(70,-10)(60,-20)
\qbezier(60,-20)(50,-30)(40,-20)
\qbezier(40,-20)(10,10)(40,40)\end{picture}}
\multiput(150,200)(40,-40){4}{\begin{picture}(20,20)
\qbezier(0,0)(10,-10)(20,-20)\end{picture}}
\put(0,550){\vector(1,-1){75}}
\put(150,400){\vector(0,-1){100}}
\put(150,200){\vector(-1,-1){75}}
\put(355,300){\makebox(0,0){{\large +}}}
\put(355,20){\makebox(0,0){(a)}}
\put(400,50){\line(1,1){150}}
\put(550,200){\line(0,1){200}}
\put(550,400){\line(-1,1){150}}
\multiput(550,200)(40,-40){4}{\begin{picture}(70,70)
\qbezier(0,0)(30,-30)(60,0)
\qbezier(60,0)(70,10)(60,20)
\qbezier(60,20)(50,30)(40,20)
\qbezier(40,20)(10,-10)(40,-40)\end{picture}}
\multiput(550,400)(40,40){4}{\begin{picture}(20,20)
\qbezier(0,0)(10,10)(20,20)\end{picture}}
\put(400,550){\vector(1,-1){75}}
\put(550,400){\vector(0,-1){100}}
\put(550,200){\vector(-1,-1){75}}
\multiput(1000,40)(40,40){4}{\begin{picture}(70,70)
\qbezier(0,0)(30,30)(60,0)
\qbezier(60,0)(70,-10)(60,-20)
\qbezier(60,-20)(50,-30)(40,-20)
\qbezier(40,-20)(10,10)(40,40)\end{picture}}
\put(1160,200){\line(0,1){200}}
\multiput(1160,400)(40,40){4}{\begin{picture}(20,20)
\qbezier(0,0)(10,10)(20,20)\end{picture}}
\put(1160,400){\line(-1,1){160}}
\put(1320,40){\line(-1,1){160}}
\put(1000,560){\vector(1,-1){80}}
\put(1160,400){\vector(0,-1){100}}
\put(1160,200){\vector(1,-1){80}}
\put(1370,300){\makebox(0,0){{\large +}}}
\put(1370,20){\makebox(0,0){(b)}}
\multiput(1420,140)(40,40){4}{\begin{picture}(70,70)
\qbezier(0,0)(30,30)(60,0)
\qbezier(60,0)(70,-10)(60,-20)
\qbezier(60,-20)(50,-30)(40,-20)
\qbezier(40,-20)(10,10)(40,40)\end{picture}}
\put(1580,300){\line(-1,1){160}}
\put(1780,300){\line(-1,0){200}}
\multiput(1780,310)(40,40){4}{\begin{picture}(20,20)
\qbezier(0,0)(10,10)(20,20)\end{picture}}
\put(1940,140){\line(-1,1){160}}
\put(1420,460){\vector(1,-1){80}}
\put(1580,300){\vector(1,0){100}}
\put(1780,300){\vector(1,-1){80}}
\end{picture}
\fcap{The leading-order Feynman diagrams contributing to $W$ + jet
production corresponding to the annihilation (a) and Compton (b)
diagrams.}
\label{figordalp}
\end{figure}
In previous work \cite{CDMdispom94} the DCM was extended to also
include the boson--gluon fusion diagram in DIS, which can be viewed as
a special case of initial-state gluon splitting into $\qq$. In this
paper, this approach is further developed into a general inclusion of
initial-state gluon splitting into $\qq$. Also, a way is presented of
taking into account the contribution to the transverse momentum of the
$W$ from the gluon emission, formulated it terms of radiation from the
colour dipole between the two proton remnants.
The layout of this paper is as follows. In sections \ref{secglu} and
\ref{seccmp} the treatment of gluon emission and initial state gluon
splitting is presented. In section \ref{secres}, results for the
$W$--jet rapidity correlation from the improved DCM is compared to a
leading-order calculation and with the conventional DGLAP-inspired
initial-state parton shower approach of \pythia. Also some predictions
are given for the transverse energy flow in high rapidity $W$ events
at Tevatron energies. Finally, in section \ref{secsum}, the
conclusions are presented.
\section{Gluon Emission}
\label{secglu}
The DCM for $\ee$\ annihilation and deep inelastic lepton--hadron
scattering is described in detail in refs.\
\cite{CDMinit86,CDMplain88,CDMdis89} and only a brief summary of the
features important for this paper will be given here.
The emission of a gluon $g_1$ from a $\qq$\ pair created in an $\ee$\
annihilation event can be described as radiation from the colour
dipole between the $q$ and $\bar{q}$. A subsequent emission of a softer
gluon $g_2$ can be described as radiation from two independent colour
dipoles, one between the $q$ and $g_1$ and one between $g_1$ and
$\bar{q}$. Further gluon emissions are given by three independent dipoles
etc.
In DIS, the gluon emission comes from the dipole stretched between the
quark, struck by the electro--weak probe, and the hadron remnant. The
situation is the same as in $\ee$\ above, except that, while $q$ and
$\bar{q}$\ are both point-like in the case of $\ee$, the hadron remnant in
DIS is an extended object. In an antenna of size $l$, radiation with
wavelengths $\lambda \ll l$ are strongly suppressed. In the DCM, this
is taken into account by only letting a fraction
\begin{equation}
a=\mu/k_\perp
\label{eqremfrac1}
\end{equation}
of the hadron remnant take part in the emission of a gluon with
transverse momentum $\kt$, where $\mu$ is a parameter corresponding to
the inverse (transverse) size of the hadron.
The phase space available in dipole emission is conveniently pictured
by the inside of a triangle in the $\kappa$--$y$ plane, where
$\kappa\equiv\ln{k_\perp^2}$ and $y$ is the rapidity of the emitted gluon
as in fig.\refig{figphase1}. In these variables the dipole emission
cross section also takes a particularly simple approximate form:
\begin{equation}
d\sigma\propto\alpha_Sd\kappa dy.
\label{eqMEdip1}
\end{equation}
In DIS, assuming that the hadron is coming in with momentum (using
light-cone coordinates) ($P_+,0,\vec{0})$, and is probed by a virtual
photon $(-Q_+,Q_-,\vec{0})$, the triangular area comes from the
trivial requirement
\begin{eqnarray}
k_{+g} & \equiv & k_\perp e^y < P_+\nonumber\\
k_{-g} & \equiv & k_\perp e^{-y} < Q_-.
\end{eqnarray}
The condition that only a fraction of the remnant participates in an
emission means that
\begin{equation}
k_{+g}<(\mu/k_\perp)P_+
\label{eqphasecut}
\end{equation}
and translates into an extra cutoff in the phase space corresponding
to the thick line in fig.\refig{figphase1}. This should be compared to
the initial-state parton shower scenario, where gluon emission is
given by
\begin{equation}
d\sigma_{q}=\frac{2\alpha_S}{3\pi}
\frac{1+z^2}{1-z}\frac{f_q(x/z)}{f_q(x)}\frac{dz}{z}\frac{dQ^2}{Q^2}.
\label{eqiniqsplit1}
\end{equation}
Identifying the ratio of structure functions in eq.\req{eqiniqsplit1}
(corresponding to the dotted line of equal suppression in
fig.\refig{figphase1}) with the extra cutoff (\ref{eqphasecut}) in the
DCM, the two models are equivalent in the low-$\kt$\ limit.
\begin{figure}
\setlength{\unitlength}{0.07mm}
\begin{picture}(2000,700)(-100,0)
\put(1050,650){\makebox(0,0){$\kappa$}}
\put(1650,100){\makebox(0,0){$y$}}
\put(600,50){\vector(-1,0){50}}
\put(625,50){\makebox(0,0)[l]{{\footnotesize direction of struck quark}}}
\put(400,100){\vector(1,0){1200}}
\put(1000,100){\vector(0,1){600}}
\put(500,100){\line(1,1){500}}
\put(1500,100){\line(-1,1){500}}
\thicklines
\put(1520,100){\line(-2,1){800}}
\qbezier[40](1300,100)(1100,300)(900,500)
\end{picture}
\fcap{The phase space available for gluon emission in DIS (thin
lines) and the extra restriction due to the extendedness of the
proton remnant (thick line). The dotted line corresponds to a line
of equal suppression due to the ratio of parton density functions
entering into a conventional parton shower scenario.}
\label{figphase1}
\end{figure}
As mentioned in the introduction, the simplest way of extending the
DCM to describe gluon emissions in Drell--Yan\ events is to describe it as
radiation from the colour dipole between the two hadron remnants. One
problem with this approach is what to do with the transverse recoil
from the gluon emission. In $\ee$, this recoil is shared by the $q$ and
$\bar{q}$. In DIS, since only a fraction of the remnant is taking part in
the emission, only that fraction is given a transverse recoil,
resulting in an extra, so-called recoil gluon \cite{CDMdis89}. The
corresponding procedure for Drell--Yan\ would be to introduce two recoil
gluons, one for each remnant. However in that way it is impossible to
reproduce the transverse momentum of the $W$ as given by the \tordas\
matrix element.
The \tordas\ matrix element for $q+\bar{q}\rightarrow W+g$ production
takes the form \cite{Halzen78}
\begin{equation}
M^{q\bar{q}\rightarrow Wg} \propto
\frac{\hat{t}^2+\hat{u}^2+2m_W^2\hat{s}}{\hat{t}\hat{u}},
\label{eqMEWg1}
\end{equation}
where $\hat{s}$, $\hat{t}$ and $\hat{u}$ are the ordinary Mandelstam
variables satisfying $\hat{s}+\hat{t}+\hat{u}=m_W^2$. In order to
reproduce this in a parton shower scenario, where the gluon is emitted
``after'' the $W$ is produced, this has to be convoluted with the
parton density functions, and the lowest-order $W$-production matrix
element, again convoluted with the relevant parton densities, has to
be factored out. This introduces some ambiguities, which are solved by
assuming that the rapidity of the $W$ is the same before and after the
gluon emission, resulting in the following cross section, expressed in
the transverse momentum $\k2t$\ and rapidity $y_g$ of the emitted gluon
\begin{eqnarray}
\frac{d\sigma_g}{dy_gdk_\perp^2} & = & \frac{2\alpha_S}{3\pi}
\frac{f_q(x_q')}{f_q(x_q)}
\frac{f_{\bar{q}}(x_{\bar{q}}')}{f_{\bar{q}}(x_{\bar{q}})} \times \nonumber
\\ & & \frac{(k_\perp^2+m_\perp^2+k_\perpm_\perp e^{\Delta y})^2+(k_\perp^2+m_\perp^2+k_\perpm_\perp
e^{-\Delta y})^2}{(k_\perp^2+k_\perpm_\perp e^{\Delta y})(k_\perp^2+k_\perpm_\perp
e^{-\Delta y})(k_\perp^2+m_\perp^2+k_\perpm_\perp(e^{\Delta y}+e^{-\Delta y}))},
\label{eqMEWg2}
\end{eqnarray}
where $\Delta y = y_g-y_W$, $m_\perp^2=k_\perp^2+m_W^2$, $y_W$ the rapidity of
the $W$ and $x_i$ and $x_i'$ the energy--momentum fractions carried by
the incoming partons before and after the gluon emission so that
\begin{eqnarray}
x_q = \frac{m_W e^{y_W}}{\sqrt{S}}, & & x_q' = \frac{m_\perp e^{y_W} +
k_\perp e^{y_g}}{\sqrt{S}}, \\ x_{\bar{q}} = \frac{m_W
e^{-y_W}}{\sqrt{S}}, & & x_{\bar{q}}' = \frac{m_\perp e^{-y_W} + k_\perp
e^{-y_g}}{\sqrt{S}},
\end{eqnarray}
assuming the $q$ coming in along the positive $z$-axis and a total
invariant mass of $\sqrt{S}$.
In the limit $k_\perp^2\ll m_W^2$, eq.\req{eqMEWg2} reduces to the simple
dipole emission cross section in eq.\req{eqMEdip1}, so it is clear
that the strategy outlined above is a good leading log approximation.
It also turns out that it is fairly simple to correct the first gluon
emission so that, disregarding the ratios of parton densities,
eq.\req{eqMEWg2} is reproduced.
\begin{figure}
\setlength{\unitlength}{0.07mm}
\begin{picture}(2000,700)(-100,0)
\put(1050,650){\makebox(0,0){$\kappa$}}
\put(1650,100){\makebox(0,0){$y$}}
\put(400,100){\vector(1,0){1200}}
\put(1000,100){\vector(0,1){600}}
\put(500,100){\line(1,1){500}}
\put(1500,100){\line(-1,1){500}}
\thicklines
\put(1520,100){\line(-2,1){800}}
\put(480,100){\line(2,1){800}}
\put(1100,50){\vector(0,1){50}}
\put(1100,25){\makebox(0,0){$y_W$}}
\put(1400,400){\vector(-1,0){50}}
\put(1420,400){\makebox(0,0)[l]{$\ln{m_W^2}$}}
\put(600,350){\vector(1,0){50}}
\put(575,350){\makebox(0,0)[r]{$\ln{k_{\perp\max}^2}$}}
\qbezier[31](800,100)(950,250)(1100,400)
\qbezier[30](820,100)(965,245)(1110,390)
\qbezier[29](840,100)(980,240)(1120,380)
\qbezier[28](860,100)(995,235)(1130,370)
\qbezier[27](880,100)(1010,230)(1140,360)
\qbezier[26](900,100)(1025,225)(1150,350)
\qbezier[25](920,100)(1040,220)(1160,340)
\qbezier[24](940,100)(1055,215)(1170,330)
\qbezier[23](960,100)(1070,210)(1180,320)
\qbezier[22](980,100)(1085,205)(1190,310)
\qbezier[21](1000,100)(1100,200)(1200,300)
\qbezier[20](1020,100)(1115,195)(1210,290)
\qbezier[19](1040,100)(1130,190)(1220,280)
\qbezier[18](1060,100)(1145,185)(1230,270)
\qbezier[17](1080,100)(1160,180)(1240,260)
\qbezier[16](1100,100)(1175,175)(1250,250)
\qbezier[15](1120,100)(1190,170)(1260,240)
\qbezier[14](1140,100)(1205,165)(1270,230)
\qbezier[13](1160,100)(1220,160)(1280,220)
\qbezier[12](1180,100)(1235,155)(1290,210)
\qbezier[11](1200,100)(1250,150)(1300,200)
\qbezier[10](1220,100)(1265,145)(1310,190)
\qbezier[9](1240,100)(1280,140)(1320,180)
\qbezier[8](1260,100)(1295,135)(1330,170)
\qbezier[7](1280,100)(1310,130)(1340,160)
\qbezier[6](1300,100)(1325,125)(1350,150)
\qbezier[5](1320,100)(1340,120)(1360,140)
\qbezier[4](1340,100)(1355,115)(1370,130)
\qbezier[3](1360,100)(1370,110)(1380,120)
\qbezier[2](1380,100)(1385,105)(1390,110)
\end{picture}
\fcap{The phase space available for gluon emission in $W$ production
(thin lines) and the extra restriction due to the extendedness of
the proton remnants (thick lines). The shaded triangle corresponds
to the phase space area covered by the $W$.}
\label{figphase2}
\end{figure}
The ratio of parton densities in eq.\req{eqMEWg2} is instead
approximated by the suppression of the phase space introduced for DIS
in eq.\req{eqphasecut}, which in this case corresponds to suppressions
on both sides of the triangle, as in fig.\refig{figphase2}. One problem
with this procedure is that the $\k2t$\ of the gluon and hence of the
$W$ is limited by this suppression to
\begin{equation}
k_\perp^2<\mu\sqrt{S/4},
\end{equation}
which, with $\mu\approx 1$ GeV, gives $k_\perp\,\lower3pt\hbox{$\buildrel < \over\sim$}\, 30$ GeV. To be able to
describe high-$\kt$\ $W$ production, it is clear that the sharp cutoff
in fig.\refig{figphase2}, which in any case is an oversimplification,
must be replaced by a smooth suppression. In \cite{CDMdis89} it was
shown that introducing a power suppression in the disallowed regions
in fig.\refig{figphase1} does not influence the general event shapes
in DIS; it is clear, however, that such a power suppression would
greatly influence the high-$\kt$\ spectrum of the $W$.
First, however, the way of obtaining a transverse momentum of the $W$
in the gluon emissions must be formalized. It is clear that, in the
first emission, the gluon corresponds unambiguously to initial state
radiation, and hence the $\kt$\ of the gluon must be balanced by the
$\kt$\ of the $W$. In further emissions this is not the case, as the
dipole radiation is a coherent sum of the emission from the incoming
partons and the outgoing, previously radiated, gluon. It is therefore
argued that only gluon radiation that takes place close to the $W$ in
phase space should be able to influence the $\kt$\ of the $W$; only
gluons emitted in the shaded region of fig.\refig{figphase2},
corresponding to $P_{g+}<P_{W+}$ and $P_{g-}<P_{W-}$, will have their
transverse recoil absorbed by the $W$. Outside this region the
transverse recoil will be treated as in the DIS case above.
\begin{figure}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2880,1728)(-500,-100)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
0 1426 V
LTb
600 251 M
63 0 V
2034 0 R
-63 0 V
600 394 M
31 0 V
2066 0 R
-31 0 V
600 478 M
31 0 V
2066 0 R
-31 0 V
600 537 M
31 0 V
2066 0 R
-31 0 V
600 583 M
31 0 V
2066 0 R
-31 0 V
600 621 M
31 0 V
2066 0 R
-31 0 V
600 653 M
31 0 V
2066 0 R
-31 0 V
600 680 M
31 0 V
2066 0 R
-31 0 V
600 705 M
31 0 V
2066 0 R
-31 0 V
600 726 M
63 0 V
2034 0 R
-63 0 V
600 869 M
31 0 V
2066 0 R
-31 0 V
600 953 M
31 0 V
2066 0 R
-31 0 V
600 1013 M
31 0 V
2066 0 R
-31 0 V
600 1059 M
31 0 V
2066 0 R
-31 0 V
600 1096 M
31 0 V
2066 0 R
-31 0 V
600 1128 M
31 0 V
2066 0 R
-31 0 V
600 1156 M
31 0 V
2066 0 R
-31 0 V
600 1180 M
31 0 V
2066 0 R
-31 0 V
600 1202 M
63 0 V
2034 0 R
-63 0 V
600 1345 M
31 0 V
2066 0 R
-31 0 V
600 1428 M
31 0 V
2066 0 R
-31 0 V
600 1488 M
31 0 V
2066 0 R
-31 0 V
600 1534 M
31 0 V
2066 0 R
-31 0 V
600 1572 M
31 0 V
2066 0 R
-31 0 V
600 1603 M
31 0 V
2066 0 R
-31 0 V
600 1631 M
31 0 V
2066 0 R
-31 0 V
600 1655 M
31 0 V
2066 0 R
-31 0 V
600 1677 M
63 0 V
2034 0 R
-63 0 V
600 251 M
0 63 V
0 1363 R
0 -63 V
862 251 M
0 63 V
0 1363 R
0 -63 V
1124 251 M
0 63 V
0 1363 R
0 -63 V
1386 251 M
0 63 V
0 1363 R
0 -63 V
1649 251 M
0 63 V
0 1363 R
0 -63 V
1911 251 M
0 63 V
0 1363 R
0 -63 V
2173 251 M
0 63 V
0 1363 R
0 -63 V
2435 251 M
0 63 V
0 1363 R
0 -63 V
2697 251 M
0 63 V
0 1363 R
0 -63 V
600 251 M
2097 0 V
0 1426 V
-2097 0 V
600 251 L
LT0
2394 1514 M
180 0 V
810 251 M
0 447 V
52 0 V
0 495 V
53 0 V
0 -52 V
52 0 V
0 -56 V
52 0 V
0 -49 V
53 0 V
0 -53 V
52 0 V
0 -45 V
53 0 V
0 -44 V
52 0 V
0 -51 V
53 0 V
0 -32 V
52 0 V
0 -42 V
52 0 V
0 -25 V
53 0 V
0 -49 V
52 0 V
0 -25 V
53 0 V
0 -33 V
52 0 V
0 -49 V
53 0 V
0 6 V
52 0 V
0 -39 V
52 0 V
0 -43 V
53 0 V
0 -54 V
52 0 V
0 -28 V
53 0 V
0 -9 V
52 0 V
0 -8 V
52 0 V
0 -47 V
53 0 V
0 -37 V
52 0 V
0 7 V
53 0 V
0 -50 V
52 0 V
0 33 V
53 0 V
0 -68 V
LT1
2394 1414 M
180 0 V
600 1551 M
52 0 V
0 80 V
53 0 V
0 -76 V
52 0 V
0 -104 V
53 0 V
0 -113 V
52 0 V
0 -111 V
53 0 V
0 -123 V
52 0 V
0 -122 V
52 0 V
0 -124 V
53 0 V
0 -130 V
52 0 V
0 -123 V
53 0 V
0 -143 V
52 0 V
0 -80 V
53 0 V
0 -131 V
LT3
2394 1314 M
180 0 V
600 1563 M
52 0 V
0 40 V
53 0 V
0 -91 V
52 0 V
0 -99 V
53 0 V
0 -94 V
52 0 V
0 -81 V
53 0 V
0 -74 V
52 0 V
0 -68 V
52 0 V
0 -62 V
53 0 V
0 -59 V
52 0 V
0 -54 V
53 0 V
0 -46 V
52 0 V
0 -42 V
53 0 V
0 -55 V
52 0 V
0 -33 V
52 0 V
0 -26 V
53 0 V
0 -59 V
52 0 V
0 -41 V
53 0 V
0 -36 V
52 0 V
0 -24 V
53 0 V
0 -33 V
52 0 V
0 -32 V
52 0 V
0 -29 V
53 0 V
0 -37 V
52 0 V
0 -16 V
53 0 V
0 -51 V
52 0 V
0 -16 V
52 0 V
0 -18 V
53 0 V
0 -7 V
52 0 V
0 -41 V
53 0 V
0 -4 V
52 0 V
0 -24 V
stroke
grestore
end
showpage
}
\put(2334,1314){\makebox(0,0)[r]{DCM gluons $\beta=2$}}
\put(2334,1414){\makebox(0,0)[r]{DCM gluons $\beta=\infty$}}
\put(2334,1514){\makebox(0,0)[r]{LO $W+g$}}
\put(1648,-49){\makebox(0,0){$k_{\perp W}$ (GeV)}}
\put(220,964){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$d\sigma/dk_{\perp W}$ (pb/GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(2697,151){\makebox(0,0){80}}
\put(2435,151){\makebox(0,0){70}}
\put(2173,151){\makebox(0,0){60}}
\put(1911,151){\makebox(0,0){50}}
\put(1649,151){\makebox(0,0){40}}
\put(1386,151){\makebox(0,0){30}}
\put(1124,151){\makebox(0,0){20}}
\put(862,151){\makebox(0,0){10}}
\put(600,151){\makebox(0,0){0}}
\put(540,1677){\makebox(0,0)[r]{1000}}
\put(540,1202){\makebox(0,0)[r]{100}}
\put(540,726){\makebox(0,0)[r]{10}}
\put(540,251){\makebox(0,0)[r]{1}}
\end{picture}
\fcap{The transverse momentum spectrum of the $W$ at the Tevatron.
The full line is the prediction of the \tordas\ $W+g$ matrix
element as implemented in \pythia\ using CTEQ2L parton density
functions. The matrix element calculation was cut off at $k_{\perp
W}=10$ GeV to avoid divergences. The dashed and dotted lines are
the predictions of the gluon emission in the DCM with
$\beta=\infty$ and $\beta=2$, respectively.}
\label{figptWg}
\end{figure}
Figure \ref{figptWg} shows the \tordas\ $W+g$ matrix element
prediction (as implemented in \pythia) of the $\kt$\ spectrum of the
$W$ at the Tevatron, compared to the modified DCM described above.
With a sharp cutoff in the phase space, it is clear that the DCM
cannot describe the high-$\kt$\ tail of the spectrum. Instead, a smooth
suppression can be introduced by changing eq.\req{eqremfrac1},
allowing a larger fraction $a'$ of the remnant to take part in the
emission with the probability
\begin{equation}
P(a') \propto
\frac{\beta(\frac{a'}{a})^\beta}{a'(1+(\frac{a'}{a})^\beta)^2},
\end{equation}
corresponding to a smoothening of the theta function suppression in
fig.\refig{figphase2}, giving a power-suppressed tail. As seen in
fig.\refig{figptWg}, using $\beta=2$ describes well the high-$\kt$\
spectrum obtained from the leading-order calculation using the
CTEQ2L\footnote{The CTEQ2L structure function parametrization is used
in all analyses in this paper where applicable. None of the
conclusions in this paper were found to be sensitive to this
choice.} \cite{CTEQ2} structure function parametrization. In the
following, this value of $\beta$ will be used, unless stated
otherwise.
This concludes the description of the $W$ + gluon jet in the DCM.
However, at small $x$, the gluon density in the proton becomes very
large, and the Compton diagrams in fig.\refig{figordalp}b are
dominating.
\section{The Compton Diagrams}
\label{seccmp}
The matrix element for the Compton diagrams looks like \cite{Halzen78}
\begin{equation}
M^{gq\rightarrow Wq} \propto
-\frac{\hat{s}^2+\hat{t}^2+2m_W^2\hat{u}}{\hat{s}\hat{t}}.
\label{eqMEWq1}
\end{equation}
The $s$-channel diagram is, of course heavily suppressed for small
$\kt$, and looking only at the $t$-channel diagram, convoluting with
parton densities and factoring out the zeroth order $W$ production
cross section, as in the gluon emission case above, the normal leading
log initial-state parton shower cross section for the splitting of an
incoming gluon into a $\qq$\ pair \cite{TSinips88} is obtained:
\begin{equation}
d\sigma_{q}=\frac{\alpha_S}{4\pi}
(z^2+(1-z)^2)\frac{f_g(x/z)}{f_q(x)}\frac{dz}{z}\frac{dQ^2}{Q^2},
\label{eqinigsplit1}
\end{equation}
where $Q^2=-\hat{t}$ and $z=m_W^2/\hat{s}$.
In the DCM, however, there is no initial state gluon splitting into
$\qq$, and, just as in the case of final-state $\g2qq$\ splitting
\cite{CDMsplit90}, this process has to be added by hand to the DCM.
The simplest way is to introduce the initial-state $\g2qq$\ splitting
in the same way as in \cite{CDMsplit90}, as a process competing with
the DCM gluon emission described above. The competition is as usual
governed by the Sudakov form factor using ordering in $\k2t$.
Rewriting eq.\req{eqinigsplit1} in terms of the transverse momentum
$\k2t$\ and rapidity $y_q$ of the outgoing quark, the probability of
the {\em first} emission to be an initial-state $\g2qq$\ splitting at a
certain $\k2t$\ and $y_q$ is given by
\begin{eqnarray}
\frac{dP_{q}(k_\perp^2,y_q)}{dk_\perp^2 dy_q} & = &
\frac{d\sigma_{q}(k_\perp^2,y_q)}{dk_\perp^2 dy_q} \times \nonumber \\ & &
\exp{-\int_{k_\perp^2}^{k_{\perp\max}^2}dk_\perp^{2\prime}\int dy_q'
\left(\frac{d\sigma_{q}(k_\perp^{2\prime},y_q')}{dk_\perp^{2\prime}
dy_q'}+\frac{d\sigma_{g}(k_\perp^{2\prime},y_q')}{dk_\perp^{2\prime}
dy_q'}\right)},
\end{eqnarray}
where the second factor is the Sudakov form factor, corresponding to
the probability {\em not} to have any emission of gluons {\em or} gluon
splittings above the scale $\k2t$.
Technically, the extra process is implemented as follows. If the quark
going into the hard interaction on one side is a sea-quark, the
remnant on that side is allowed to ``radiate'' the corresponding
antiquark according to eq.\req{eqinigsplit1}. After such an emission,
the remnant is split in two parts according to the prescription
described in ref.\ \cite{CDMdis89}, one of which forms a dipole with
the ``emitted'' antiquark while the other retains the dipole colour
connection of the original remnant.
If the first emission is a $\g2qq$\ splitting, the full \tordas\ matrix
element is used, and the rapidity of the $W$ is assumed to be the same
before and after the emission, as in the gluon emission case. A
splitting later on in the cascade, the kinematic is fixed by requiring
the non-radiating remnant to be unchanged. In all cases, the
transverse momentum of the struck system will of course balance the
$\kt$\ of the emitted antiquark. Note that only one initial state
$\g2qq$\ splitting is allowed per remnant. This is a good
approximation, since a second such splitting is heavily suppressed by
the parton density functions.
This procedure can be used not only in the case of $W$ production, but
for all processes with a hadron remnant present. In particular it can
be (and is\footnote{This is the default in \ariadne\ version 4.06 and
later.}) applied in the DIS case.
\begin{figure}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2880,1728)(-500,-400)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
0 1326 V
LTb
600 251 M
63 0 V
2034 0 R
-63 0 V
600 351 M
31 0 V
2066 0 R
-31 0 V
600 409 M
31 0 V
2066 0 R
-31 0 V
600 451 M
31 0 V
2066 0 R
-31 0 V
600 483 M
31 0 V
2066 0 R
-31 0 V
600 509 M
31 0 V
2066 0 R
-31 0 V
600 531 M
31 0 V
2066 0 R
-31 0 V
600 550 M
31 0 V
2066 0 R
-31 0 V
600 567 M
31 0 V
2066 0 R
-31 0 V
600 583 M
63 0 V
2034 0 R
-63 0 V
600 682 M
31 0 V
2066 0 R
-31 0 V
600 741 M
31 0 V
2066 0 R
-31 0 V
600 782 M
31 0 V
2066 0 R
-31 0 V
600 814 M
31 0 V
2066 0 R
-31 0 V
600 840 M
31 0 V
2066 0 R
-31 0 V
600 863 M
31 0 V
2066 0 R
-31 0 V
600 882 M
31 0 V
2066 0 R
-31 0 V
600 899 M
31 0 V
2066 0 R
-31 0 V
600 914 M
63 0 V
2034 0 R
-63 0 V
600 1014 M
31 0 V
2066 0 R
-31 0 V
600 1072 M
31 0 V
2066 0 R
-31 0 V
600 1114 M
31 0 V
2066 0 R
-31 0 V
600 1146 M
31 0 V
2066 0 R
-31 0 V
600 1172 M
31 0 V
2066 0 R
-31 0 V
600 1194 M
31 0 V
2066 0 R
-31 0 V
600 1213 M
31 0 V
2066 0 R
-31 0 V
600 1230 M
31 0 V
2066 0 R
-31 0 V
600 1246 M
63 0 V
2034 0 R
-63 0 V
600 1345 M
31 0 V
2066 0 R
-31 0 V
600 1404 M
31 0 V
2066 0 R
-31 0 V
600 1445 M
31 0 V
2066 0 R
-31 0 V
600 1477 M
31 0 V
2066 0 R
-31 0 V
600 1503 M
31 0 V
2066 0 R
-31 0 V
600 1526 M
31 0 V
2066 0 R
-31 0 V
600 1545 M
31 0 V
2066 0 R
-31 0 V
600 1562 M
31 0 V
2066 0 R
-31 0 V
600 1577 M
63 0 V
2034 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
1019 251 M
0 63 V
0 1263 R
0 -63 V
1439 251 M
0 63 V
0 1263 R
0 -63 V
1858 251 M
0 63 V
0 1263 R
0 -63 V
2278 251 M
0 63 V
0 1263 R
0 -63 V
2697 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
2097 0 V
0 1326 V
-2097 0 V
600 251 L
LT0
2394 1414 M
180 0 V
768 251 M
0 686 V
42 0 V
0 354 V
42 0 V
0 -29 V
42 0 V
0 -38 V
42 0 V
0 -30 V
41 0 V
0 -35 V
42 0 V
0 -23 V
42 0 V
0 -33 V
42 0 V
0 -25 V
42 0 V
0 -28 V
42 0 V
0 -16 V
42 0 V
0 -23 V
42 0 V
0 -21 V
42 0 V
0 -23 V
42 0 V
0 -29 V
42 0 V
0 -16 V
42 0 V
0 -15 V
42 0 V
0 -30 V
42 0 V
42 0 V
0 -24 V
42 0 V
0 -18 V
42 0 V
0 -22 V
41 0 V
0 -12 V
42 0 V
0 -3 V
42 0 V
0 -26 V
42 0 V
0 -29 V
42 0 V
0 -7 V
42 0 V
0 -12 V
42 0 V
0 -33 V
42 0 V
0 -8 V
42 0 V
0 -8 V
42 0 V
0 -16 V
42 0 V
0 -25 V
42 0 V
0 -1 V
42 0 V
0 -32 V
42 0 V
0 -18 V
42 0 V
0 -8 V
42 0 V
0 -24 V
41 0 V
0 7 V
42 0 V
0 5 V
42 0 V
0 -40 V
42 0 V
0 -19 V
42 0 V
0 -8 V
42 0 V
0 -25 V
42 0 V
42 0 V
0 10 V
42 0 V
0 -37 V
LT3
2394 1314 M
180 0 V
600 1437 M
42 0 V
0 50 V
42 0 V
0 -41 V
42 0 V
0 -46 V
42 0 V
0 -44 V
42 0 V
0 -46 V
42 0 V
0 -40 V
42 0 V
0 -35 V
42 0 V
0 -36 V
41 0 V
0 -28 V
42 0 V
0 -24 V
42 0 V
0 -32 V
42 0 V
0 -34 V
42 0 V
0 -14 V
42 0 V
0 -18 V
42 0 V
0 -31 V
42 0 V
0 -11 V
42 0 V
0 -36 V
42 0 V
0 -12 V
42 0 V
0 -27 V
42 0 V
0 -24 V
42 0 V
0 -27 V
42 0 V
0 2 V
42 0 V
0 -40 V
42 0 V
0 -2 V
42 0 V
0 -45 V
41 0 V
0 8 V
42 0 V
0 -15 V
42 0 V
0 -20 V
42 0 V
0 -12 V
42 0 V
0 -42 V
42 0 V
0 30 V
42 0 V
0 -42 V
42 0 V
0 -8 V
42 0 V
0 -7 V
42 0 V
0 -53 V
42 0 V
0 -12 V
42 0 V
0 36 V
42 0 V
0 -53 V
42 0 V
0 -3 V
42 0 V
0 7 V
42 0 V
0 -75 V
41 0 V
0 -12 V
42 0 V
0 69 V
42 0 V
42 0 V
0 -82 V
42 0 V
0 -93 V
42 0 V
0 63 V
42 0 V
0 -9 V
42 0 V
0 -54 V
42 0 V
0 100 V
stroke
grestore
end
showpage
}
\put(2334,1314){\makebox(0,0)[r]{Full DCM}}
\put(2334,1414){\makebox(0,0)[r]{Leading order}}
\put(1648,1677){\makebox(0,0){(a)}}
\put(1648,-49){\makebox(0,0){$k_{\perp W}$ (GeV)}}
\put(280,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$d\sigma/dk_{\perp W}$ (pb/GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(2697,151){\makebox(0,0){100}}
\put(2278,151){\makebox(0,0){80}}
\put(1858,151){\makebox(0,0){60}}
\put(1439,151){\makebox(0,0){40}}
\put(1019,151){\makebox(0,0){20}}
\put(600,151){\makebox(0,0){0}}
\put(540,1577){\makebox(0,0)[r]{1000}}
\put(540,1246){\makebox(0,0)[r]{100}}
\put(540,914){\makebox(0,0)[r]{10}}
\put(540,583){\makebox(0,0)[r]{1}}
\put(540,251){\makebox(0,0)[r]{0.1}}
\end{picture}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2880,1728)(-500,-100)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
0 1326 V
LTb
600 251 M
63 0 V
2034 0 R
-63 0 V
600 351 M
31 0 V
2066 0 R
-31 0 V
600 409 M
31 0 V
2066 0 R
-31 0 V
600 451 M
31 0 V
2066 0 R
-31 0 V
600 483 M
31 0 V
2066 0 R
-31 0 V
600 509 M
31 0 V
2066 0 R
-31 0 V
600 531 M
31 0 V
2066 0 R
-31 0 V
600 550 M
31 0 V
2066 0 R
-31 0 V
600 567 M
31 0 V
2066 0 R
-31 0 V
600 583 M
63 0 V
2034 0 R
-63 0 V
600 682 M
31 0 V
2066 0 R
-31 0 V
600 741 M
31 0 V
2066 0 R
-31 0 V
600 782 M
31 0 V
2066 0 R
-31 0 V
600 814 M
31 0 V
2066 0 R
-31 0 V
600 840 M
31 0 V
2066 0 R
-31 0 V
600 863 M
31 0 V
2066 0 R
-31 0 V
600 882 M
31 0 V
2066 0 R
-31 0 V
600 899 M
31 0 V
2066 0 R
-31 0 V
600 914 M
63 0 V
2034 0 R
-63 0 V
600 1014 M
31 0 V
2066 0 R
-31 0 V
600 1072 M
31 0 V
2066 0 R
-31 0 V
600 1114 M
31 0 V
2066 0 R
-31 0 V
600 1146 M
31 0 V
2066 0 R
-31 0 V
600 1172 M
31 0 V
2066 0 R
-31 0 V
600 1194 M
31 0 V
2066 0 R
-31 0 V
600 1213 M
31 0 V
2066 0 R
-31 0 V
600 1230 M
31 0 V
2066 0 R
-31 0 V
600 1246 M
63 0 V
2034 0 R
-63 0 V
600 1345 M
31 0 V
2066 0 R
-31 0 V
600 1404 M
31 0 V
2066 0 R
-31 0 V
600 1445 M
31 0 V
2066 0 R
-31 0 V
600 1477 M
31 0 V
2066 0 R
-31 0 V
600 1503 M
31 0 V
2066 0 R
-31 0 V
600 1526 M
31 0 V
2066 0 R
-31 0 V
600 1545 M
31 0 V
2066 0 R
-31 0 V
600 1562 M
31 0 V
2066 0 R
-31 0 V
600 1577 M
63 0 V
2034 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
1019 251 M
0 63 V
0 1263 R
0 -63 V
1439 251 M
0 63 V
0 1263 R
0 -63 V
1858 251 M
0 63 V
0 1263 R
0 -63 V
2278 251 M
0 63 V
0 1263 R
0 -63 V
2697 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
2097 0 V
0 1326 V
-2097 0 V
600 251 L
LT0
2394 1414 M
180 0 V
600 678 M
42 0 V
0 202 V
42 0 V
0 106 V
42 0 V
0 90 V
42 0 V
0 81 V
42 0 V
0 46 V
42 0 V
0 7 V
42 0 V
0 -16 V
42 0 V
0 -19 V
41 0 V
0 -28 V
42 0 V
0 -18 V
42 0 V
0 -24 V
42 0 V
0 -23 V
42 0 V
0 -19 V
42 0 V
0 -22 V
42 0 V
0 -24 V
42 0 V
0 -14 V
42 0 V
0 -22 V
42 0 V
0 -12 V
42 0 V
0 -28 V
42 0 V
0 -14 V
42 0 V
0 -35 V
42 0 V
0 7 V
42 0 V
0 -49 V
42 0 V
0 19 V
42 0 V
0 -18 V
41 0 V
0 -29 V
42 0 V
0 5 V
42 0 V
0 -8 V
42 0 V
0 -49 V
42 0 V
0 2 V
42 0 V
0 -23 V
42 0 V
0 -8 V
42 0 V
0 -17 V
42 0 V
0 -34 V
42 0 V
0 -7 V
42 0 V
0 -19 V
42 0 V
42 0 V
0 -71 V
42 0 V
0 53 V
42 0 V
0 -53 V
42 0 V
0 -20 V
41 0 V
0 33 V
42 0 V
0 -100 V
42 0 V
0 31 V
42 0 V
0 25 V
42 0 V
0 -106 V
42 0 V
0 58 V
42 0 V
0 -58 V
42 0 V
0 12 V
42 0 V
0 -38 V
LT1
2394 1314 M
180 0 V
600 1504 M
42 0 V
0 -19 V
42 0 V
0 -57 V
42 0 V
0 -53 V
42 0 V
0 -42 V
42 0 V
0 -35 V
42 0 V
0 -35 V
42 0 V
0 -37 V
42 0 V
0 -30 V
41 0 V
0 -27 V
42 0 V
0 -33 V
42 0 V
0 -18 V
42 0 V
0 -33 V
42 0 V
0 -29 V
42 0 V
0 -21 V
42 0 V
0 -25 V
42 0 V
0 -15 V
42 0 V
0 -33 V
42 0 V
0 -27 V
42 0 V
0 -21 V
42 0 V
0 -22 V
42 0 V
0 -21 V
42 0 V
0 -18 V
42 0 V
0 -38 V
42 0 V
0 -5 V
42 0 V
0 -43 V
41 0 V
0 -33 V
42 0 V
0 -43 V
42 0 V
0 -38 V
42 0 V
0 -8 V
42 0 V
42 0 V
0 -67 V
42 0 V
0 -13 V
42 0 V
0 -48 V
42 0 V
0 -43 V
42 0 V
0 -209 V
42 0 V
0 81 V
42 0 V
0 -22 V
42 0 V
0 -73 V
LT3
2394 1214 M
180 0 V
600 1437 M
42 0 V
0 50 V
42 0 V
0 -41 V
42 0 V
0 -46 V
42 0 V
0 -44 V
42 0 V
0 -46 V
42 0 V
0 -40 V
42 0 V
0 -35 V
42 0 V
0 -36 V
41 0 V
0 -28 V
42 0 V
0 -24 V
42 0 V
0 -32 V
42 0 V
0 -34 V
42 0 V
0 -14 V
42 0 V
0 -18 V
42 0 V
0 -31 V
42 0 V
0 -11 V
42 0 V
0 -36 V
42 0 V
0 -12 V
42 0 V
0 -27 V
42 0 V
0 -24 V
42 0 V
0 -27 V
42 0 V
0 2 V
42 0 V
0 -40 V
42 0 V
0 -2 V
42 0 V
0 -45 V
41 0 V
0 8 V
42 0 V
0 -15 V
42 0 V
0 -20 V
42 0 V
0 -12 V
42 0 V
0 -42 V
42 0 V
0 30 V
42 0 V
0 -42 V
42 0 V
0 -8 V
42 0 V
0 -7 V
42 0 V
0 -53 V
42 0 V
0 -12 V
42 0 V
0 36 V
42 0 V
0 -53 V
42 0 V
0 -3 V
42 0 V
0 7 V
42 0 V
0 -75 V
41 0 V
0 -12 V
42 0 V
0 69 V
42 0 V
42 0 V
0 -82 V
42 0 V
0 -93 V
42 0 V
0 63 V
42 0 V
0 -9 V
42 0 V
0 -54 V
42 0 V
0 100 V
stroke
grestore
end
showpage
}
\put(2334,1214){\makebox(0,0)[r]{Full DCM}}
\put(2334,1314){\makebox(0,0)[r]{Parton shower}}
\put(2334,1414){\makebox(0,0)[r]{LO+PS}}
\put(1648,1677){\makebox(0,0){(b)}}
\put(1648,-49){\makebox(0,0){$k_{\perp W}$ (GeV)}}
\put(280,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$d\sigma/dk_{\perp W}$ (pb/GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(2697,151){\makebox(0,0){100}}
\put(2278,151){\makebox(0,0){80}}
\put(1858,151){\makebox(0,0){60}}
\put(1439,151){\makebox(0,0){40}}
\put(1019,151){\makebox(0,0){20}}
\put(600,151){\makebox(0,0){0}}
\put(540,1577){\makebox(0,0)[r]{1000}}
\put(540,1246){\makebox(0,0)[r]{100}}
\put(540,914){\makebox(0,0)[r]{10}}
\put(540,583){\makebox(0,0)[r]{1}}
\put(540,251){\makebox(0,0)[r]{0.1}}
\end{picture}
\fcap{The transverse momentum spectrum of the $W$ at the Tevatron. In
(a) the full line is the prediction of the full \tordas\ $W+$jet
matrix element as implemented in \pythia. The dotted line is the
prediction of the full DCM with $\beta=2$ and initial-state $\g2qq$\
splitting. In (b) the full line is as in (a) but with the parton
shower of \pythia\ added after the first emission, the dashed line
is \pythia\ using only parton showers and the dotted line is the
same as in (a).}
\label{figptW1}
\end{figure}
In fig.\refig{figptW1}a, the $W$ $\kt$\ spectrum at the Tevatron is
shown, using the full \tordas\ matrix element (as implemented in
\pythia) and using the modified DCM with the initial state $\g2qq$\
splitting as implemented in \ariadne\footnote{All results labelled
\ariadne\ or DCM are actually generated using the zeroth-order $W$
production in \pythia, with the CDM added and, where indicated,
using the string fragmentation implemented in \jetset\
\cite{JETPYT94}.}. Clearly the DCM does a good job of reproducing
the high-$\kt$\ tail of the distribution. In fig.\refig{figptW1}b the
DCM is compared with the two parton shower approaches of \pythia, one
using only parton showers and one using first-order matrix elements
with parton shower added. Since the DCM is a leading-log cascade,
except that the first emission is uses the full matrix element, it
smoothly interpolates between the pure parton shower description,
which should be a good approximation for small $\kt$, and the matrix
element description, which is good for high $\kt$, but has to be cut
off at small $\kt$\ to avoid divergences.
\section{Results and Predictions}
\label{secres}
In ref.\ \cite{GEOFFcdm94}, it was found that, when looking at the
rapidity of the balancing jet in high-$\kt$\ $W$ events at the
Tevatron, no correlation with the $W$ rapidity was found, while a
leading order and a next to leading order calculation predicted a
strong correlation. It was also found that a preliminary
implementation of the DCM model described here reproduced data fairly
well and only gave a very weak correlation\footnote{The results
presented here differ from the ones in ref.\ \cite{GEOFFcdm94} due
to a bug introduced in the initial-state $\g2qq$\ splitting in the
preliminary version of \ariadne\ used in that paper.}.
\begin{figure}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2880,1728)(-500,0)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
2097 0 V
600 251 M
0 1426 V
LTb
600 251 M
63 0 V
2034 0 R
-63 0 V
600 536 M
63 0 V
2034 0 R
-63 0 V
600 821 M
63 0 V
2034 0 R
-63 0 V
600 1107 M
63 0 V
2034 0 R
-63 0 V
600 1392 M
63 0 V
2034 0 R
-63 0 V
600 1677 M
63 0 V
2034 0 R
-63 0 V
600 251 M
0 63 V
0 1363 R
0 -63 V
1019 251 M
0 63 V
0 1363 R
0 -63 V
1439 251 M
0 63 V
0 1363 R
0 -63 V
1858 251 M
0 63 V
0 1363 R
0 -63 V
2278 251 M
0 63 V
0 1363 R
0 -63 V
2697 251 M
0 63 V
0 1363 R
0 -63 V
600 251 M
2097 0 V
0 1426 V
-2097 0 V
600 251 L
LT0
1667 1534 M
180 0 V
810 415 M
419 355 V
420 315 V
419 216 V
419 97 V
LT1
1667 1434 M
180 0 V
810 363 M
419 298 V
420 201 V
419 131 V
419 188 V
LT3
1667 1334 M
180 0 V
810 369 M
419 152 V
420 144 V
419 129 V
419 -77 V
stroke
grestore
end
showpage
}
\put(1607,1334){\makebox(0,0)[r]{DCM}}
\put(1607,1434){\makebox(0,0)[r]{LO+PS}}
\put(1607,1534){\makebox(0,0)[r]{Leading order}}
\put(1648,51){\makebox(0,0){$y_W$}}
\put(280,964){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$\langle\eta_{\mbox{jet}}\rangle$}}%
\special{ps: currentpoint grestore moveto}%
}
\put(2697,151){\makebox(0,0){2.5}}
\put(2278,151){\makebox(0,0){2}}
\put(1858,151){\makebox(0,0){1.5}}
\put(1439,151){\makebox(0,0){1}}
\put(1019,151){\makebox(0,0){0.5}}
\put(600,151){\makebox(0,0){0}}
\put(540,1677){\makebox(0,0)[r]{1}}
\put(540,1392){\makebox(0,0)[r]{0.8}}
\put(540,1107){\makebox(0,0)[r]{0.6}}
\put(540,821){\makebox(0,0)[r]{0.4}}
\put(540,536){\makebox(0,0)[r]{0.2}}
\put(540,251){\makebox(0,0)[r]{0}}
\end{picture}
\fcap{The average jet pseudo-rapidity $\eta$ vs.\ the rapidity of the
$W$ $y_W$ at the Tevatron. The jets are reconstructed with a cone
algorithm using a radius of $0.7$, and in each event a jet with
$E_\perp>20$ GeV, and $|\eta|<3$ on the oposite side in azimuth
w.r.t.\ the $W$, is selected. In case of several such jets, the one
with $\et$\ closest to the $\kt$\ of the $W$ is chosen. The full line
is the leading order calculation as implemented in \pythia, the
dashed line is the same, but with parton showers and fragmentation
added, and the dotted line is the full DCM also with fragmentation
added. (The ``kinkiness'' of the lines are due to limited statistics
in the simulations.)}
\label{figWjet1}
\end{figure}
Figure \ref{figWjet1} is an attempt to reconstruct the measurement in
ref.\ \cite{GEOFFcdm94}\footnote{The details in the jet reconstruction
may differ from that of ref.\ \cite{GEOFFcdm94}. In addition, the
experimental ambiguity in the $y_W$ determination is not taken into
account here.} on the generator level. As expected from eqs.\
(\ref{eqMEWg1}) and (\ref{eqMEWq1}), which are both symmetric around
the $W$ rapidity, the leading-order calculation gives a more or less
linear correlation between the mean jet pseudo-rapidity and the
rapidity of the $W$, although $\langle\eta_{\mbox{\tiny
jet}}\rangle\neq y_q$ due to the smearing of the structure
function convolution and the limited kinematical acceptance for jets.
When parton showers and fragmentation are added to the leading-order
calculation, the smearing is increased. Also, since the phase space
available for emissions is larger on the side where the $x$ of the
incoming parton is smaller, the jets for large $y_W$ are ``dragged''
somewhat towards the centre, destroying the correlation. In the
DCM, this dragging is more pronounced due to the ordering in the
cascade as follows.
\begin{figure}
\setlength{\unitlength}{0.063mm}
\begin{picture}(2400,700)(0,0)
\put(50,100){\vector(1,0){1100}}
\put(600,100){\vector(0,1){550}}
\put(600,50){\makebox(0,0){(a)}}
\put(100,100){\line(1,1){500}}
\put(1100,100){\line(-1,1){500}}
\put(750,350){\makebox(0,0){{\footnotesize $\bullet$}}}
\put(950,450){\makebox(0,0){{\footnotesize $(y_W,\ln{m_W^2})$}}}
\put(750,350){\vector(-1,-2){30}}
\put(720,290){\vector(-1,-1){60}}
\put(660,230){\vector(-2,-1){100}}
\put(560,180){\vector(-4,-1){150}}
\put(735,275){\makebox(0,0){{\footnotesize 1}}}
\put(675,215){\makebox(0,0){{\footnotesize 2}}}
\put(575,160){\makebox(0,0){{\footnotesize 3}}}
\put(425,120){\makebox(0,0){{\footnotesize 4}}}
\put(650,675){\makebox(0,0){{\footnotesize $\kappa$}}}
\put(1175,70){\makebox(0,0){{\footnotesize $y$}}}
\put(1250,100){\vector(1,0){1100}}
\put(1800,100){\vector(0,1){550}}
\put(1800,50){\makebox(0,0){(b)}}
\put(1300,100){\line(1,1){500}}
\put(2300,100){\line(-1,1){500}}
\put(1300,100){\line(2,1){667}}
\put(2300,100){\line(-2,1){667}}
\put(1850,675){\makebox(0,0){{\footnotesize $\kappa$}}}
\put(2375,70){\makebox(0,0){{\footnotesize $y$}}}
\put(1950,350){\makebox(0,0){{\footnotesize $\bullet$}}}
\put(2150,450){\makebox(0,0){{\footnotesize $(y_W,\ln{m_W^2})$}}}
\put(1950,350){\vector(-1,-3){60}}
\put(1890,150){\makebox(0,0){{\footnotesize 3}}}
\put(1890,170){\vector(-1,1){100}}
\put(1780,290){\makebox(0,0){{\footnotesize 1}}}
\put(1790,270){\vector(-1,-1){120}}
\put(1680,130){\makebox(0,0){{\footnotesize 4}}}
\put(1670,150){\vector(-2,1){100}}
\put(1550,200){\makebox(0,0){{\footnotesize 2}}}
\end{picture}
\fcap{Example of paths, tracing emissions backwards from the hard
interaction $(y_W,\ln{m_W^2})$ on the small-$x$\ side of high-rapidity
$W$ events for (a) the initial-state parton shower in \pythia,
where the emissions are ordered both in $x$ and $\k2t$, and (b) the
DCM, where the emissions, although the cascade is ordered in
$\k2t$, when traced backwards from the hard interaction in this
way, are ordered in $x$ but not in $\k2t$.}
\label{figphase3}
\end{figure}
In the parton shower in \pythia, each step in the backward evolution
of the initial-state shower is ordered in both $x$ and virtuality
\tq2; thus even if the phase space is larger on the small-$x$\ side of the
$W$, the shower quickly runs out of phase space due to the ordering in
\tq2\ (resulting also in an ordering in $\k2t$) as in
fig.\refig{figphase3}a. The DCM, although ordered in $\k2t$, is not
ordered in $x$, or, if the final state partons are traced backwards in
colour from the hard interaction, ordered in $x$ but {\em not} in
$\k2t$ as in fig.\refig{figphase3}b. In this respect, the DCM is
similar to the BFKL evolution, and it gives a good description of the
large transverse energy flows in small-$x$\ events at HERA, which has been
suggested as a signal for the BFKL evolution \cite{MARTINetflow94}.
Because of this, the DCM can better use the increased phase space on
the small-$x$\ side of the $W$ and the ``dragging'' effect is larger than for
conventional parton showers in fig.\refig{figWjet1}, and the result
closer to, if not consistent with, the measurement in ref.\
\cite{GEOFFcdm94}.
The increase in transverse energy flow at small $x$ found at HERA
should also be visible at the Tevatron in high-rapidity $W$ events. In
fig.\refig{figEtflow1} the predictions for the $\et$\ flow from the
parton shower model of \pythia\footnote{Since no high-$\kt$\ jets are
required and the bulk of the events are at low $k_{\perp W}$, the
matrix element plus parton shower approach in fig.\refig{figWjet1}
is not adequate here.} and the DCM of \ariadne\ are shown for
inclusive $W$ events at the Tevatron for two $W$ rapidity intervals.
The two models are fairly similar at central $W$ rapidities, while for
high $y_W$ the CDM gives more transverse energy, despite the fact that
the hard interaction scale ($m_W^2\approx 6400$ $\mbox{GeV}^2$) is much
larger here than at HERA ($\langle Q^2 \rangle \,\lower3pt\hbox{$\buildrel < \over\sim$}\,
100$ $\mbox{GeV}^2$).
\begin{figure}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2160,1728)(100,0)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
1377 0 V
-688 0 R
0 1326 V
LTb
600 251 M
63 0 V
1314 0 R
-63 0 V
600 781 M
63 0 V
1314 0 R
-63 0 V
600 1312 M
63 0 V
1314 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
830 251 M
0 63 V
0 1263 R
0 -63 V
1059 251 M
0 63 V
0 1263 R
0 -63 V
1289 251 M
0 63 V
0 1263 R
0 -63 V
1518 251 M
0 63 V
0 1263 R
0 -63 V
1748 251 M
0 63 V
0 1263 R
0 -63 V
1977 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
1377 0 V
0 1326 V
-1377 0 V
600 251 L
LT0
1693 649 M
180 0 V
657 745 M
772 852 L
887 980 L
115 143 V
114 73 V
115 68 V
115 30 V
115 -45 V
114 -100 V
115 -119 V
1805 887 L
1920 778 L
LT3
1693 549 M
180 0 V
657 827 M
115 78 V
115 88 V
115 94 V
114 99 V
115 93 V
115 -32 V
115 -50 V
114 -77 V
1690 995 L
115 -91 V
115 -75 V
stroke
grestore
end
showpage
}
\put(1633,549){\makebox(0,0)[r]{DCM}}
\put(1633,649){\makebox(0,0)[r]{PS}}
\put(1288,1677){\makebox(0,0){(a) $0.0<y_W<0.5$}}
\put(1288,51){\makebox(0,0){$\eta$}}
\put(400,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$dE_\perp/d\eta$ (GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(1977,151){\makebox(0,0){3}}
\put(1748,151){\makebox(0,0){2}}
\put(1518,151){\makebox(0,0){1}}
\put(1289,151){\makebox(0,0){0}}
\put(1059,151){\makebox(0,0){-1}}
\put(830,151){\makebox(0,0){-2}}
\put(600,151){\makebox(0,0){-3}}
\put(540,1312){\makebox(0,0)[r]{4}}
\put(540,781){\makebox(0,0)[r]{2}}
\put(540,251){\makebox(0,0)[r]{0}}
\end{picture}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2160,1728)(100,0)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
1377 0 V
-688 0 R
0 1326 V
LTb
600 251 M
63 0 V
1314 0 R
-63 0 V
600 781 M
63 0 V
1314 0 R
-63 0 V
600 1312 M
63 0 V
1314 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
830 251 M
0 63 V
0 1263 R
0 -63 V
1059 251 M
0 63 V
0 1263 R
0 -63 V
1289 251 M
0 63 V
0 1263 R
0 -63 V
1518 251 M
0 63 V
0 1263 R
0 -63 V
1748 251 M
0 63 V
0 1263 R
0 -63 V
1977 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
1377 0 V
0 1326 V
-1377 0 V
600 251 L
LT0
1693 649 M
180 0 V
657 613 M
115 33 V
115 46 V
115 58 V
114 105 V
115 40 V
115 46 V
115 51 V
1575 890 L
115 -92 V
115 -58 V
115 -47 V
LT3
1693 549 M
180 0 V
657 845 M
115 50 V
115 61 V
115 60 V
114 67 V
115 48 V
115 22 V
115 19 V
114 -95 V
115 -60 V
1805 888 L
1920 784 L
stroke
grestore
end
showpage
}
\put(1633,549){\makebox(0,0)[r]{DCM}}
\put(1633,649){\makebox(0,0)[r]{PS}}
\put(1288,1677){\makebox(0,0){(b) $2.0<y_W<2.5$}}
\put(1288,51){\makebox(0,0){$\eta$}}
\put(400,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$dE_\perp/d\eta$ (GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(1977,151){\makebox(0,0){3}}
\put(1748,151){\makebox(0,0){2}}
\put(1518,151){\makebox(0,0){1}}
\put(1289,151){\makebox(0,0){0}}
\put(1059,151){\makebox(0,0){-1}}
\put(830,151){\makebox(0,0){-2}}
\put(600,151){\makebox(0,0){-3}}
\put(540,1312){\makebox(0,0)[r]{4}}
\put(540,781){\makebox(0,0)[r]{2}}
\put(540,251){\makebox(0,0)[r]{0}}
\end{picture}
\fcap{The transverse energy flow in inclusive $W$ events at the
Tevatron for (a) $0.0<y_W<0.5$ and (b) $2.0<y_W<2.5$. The full and
dotted lines are the predictions of the DCM in \ariadne\ and of the
parton shower in \pythia, respectively.}
\label{figEtflow1}
\end{figure}
In $pp$ collisions we also have to worry about underlying events. In
fig.\refig{figWjet1}, this does not give large effects since a
large-$\et$\ jet is required, but for fig.\refig{figEtflow1} the
underlying event would give an extra contribution to the $\et$\ flow.
This extra contribution should however be evenly spread out in $\eta$
and independent of $y_W$, and the differences between the parton
shower and DCM approaches should survive. To take this contribution
into account, the multiple interaction model implemented in \pythia\
\cite{TSmult87} has been used. Note, however, that in the case of the
DCM, only the qualitative features of the contribution are completely
relevant, as the parameters of the multiple interaction model probably
need to be retuned to fit the DCM.
\begin{figure}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2160,1728)(100,0)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
1377 0 V
600 251 M
0 1326 V
LTb
600 251 M
63 0 V
1314 0 R
-63 0 V
600 630 M
63 0 V
1314 0 R
-63 0 V
600 1009 M
63 0 V
1314 0 R
-63 0 V
600 1388 M
63 0 V
1314 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
875 251 M
0 63 V
0 1263 R
0 -63 V
1151 251 M
0 63 V
0 1263 R
0 -63 V
1426 251 M
0 63 V
0 1263 R
0 -63 V
1702 251 M
0 63 V
0 1263 R
0 -63 V
1977 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
1377 0 V
0 1326 V
-1377 0 V
600 251 L
LT0
1486 1198 M
180 0 V
738 680 M
275 46 V
276 27 V
275 -3 V
275 -39 V
LT3
1486 1098 M
180 0 V
738 718 M
275 65 V
276 44 V
275 33 V
275 20 V
stroke
grestore
end
showpage
}
\put(1426,1098){\makebox(0,0)[r]{DCM}}
\put(1426,1198){\makebox(0,0)[r]{PS}}
\put(1288,1677){\makebox(0,0){(a)}}
\put(1288,51){\makebox(0,0){$y_W$}}
\put(400,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$dE_\perp/d\eta$ (GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(1977,151){\makebox(0,0){2.5}}
\put(1702,151){\makebox(0,0){2}}
\put(1426,151){\makebox(0,0){1.5}}
\put(1151,151){\makebox(0,0){1}}
\put(875,151){\makebox(0,0){0.5}}
\put(600,151){\makebox(0,0){0}}
\put(540,1388){\makebox(0,0)[r]{6}}
\put(540,1009){\makebox(0,0)[r]{4}}
\put(540,630){\makebox(0,0)[r]{2}}
\put(540,251){\makebox(0,0)[r]{0}}
\end{picture}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2160,1728)(100,0)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
1377 0 V
600 251 M
0 1326 V
LTb
600 251 M
63 0 V
1314 0 R
-63 0 V
600 630 M
63 0 V
1314 0 R
-63 0 V
600 1009 M
63 0 V
1314 0 R
-63 0 V
600 1388 M
63 0 V
1314 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
875 251 M
0 63 V
0 1263 R
0 -63 V
1151 251 M
0 63 V
0 1263 R
0 -63 V
1426 251 M
0 63 V
0 1263 R
0 -63 V
1702 251 M
0 63 V
0 1263 R
0 -63 V
1977 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
1377 0 V
0 1326 V
-1377 0 V
600 251 L
LT0
1486 630 M
180 0 V
738 1094 M
275 50 V
276 18 V
275 -15 V
275 -81 V
LT3
1486 530 M
180 0 V
738 1213 M
275 73 V
276 46 V
275 66 V
275 -44 V
stroke
grestore
end
showpage
}
\put(1426,530){\makebox(0,0)[r]{DCM+MI}}
\put(1426,630){\makebox(0,0)[r]{PS+MI}}
\put(1288,1677){\makebox(0,0){(b)}}
\put(1288,51){\makebox(0,0){$y_W$}}
\put(400,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$dE_\perp/d\eta$ (GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(1977,151){\makebox(0,0){2.5}}
\put(1702,151){\makebox(0,0){2}}
\put(1426,151){\makebox(0,0){1.5}}
\put(1151,151){\makebox(0,0){1}}
\put(875,151){\makebox(0,0){0.5}}
\put(600,151){\makebox(0,0){0}}
\put(540,1388){\makebox(0,0)[r]{6}}
\put(540,1009){\makebox(0,0)[r]{4}}
\put(540,630){\makebox(0,0)[r]{2}}
\put(540,251){\makebox(0,0)[r]{0}}
\end{picture}
\fcap{The transverse energy flow two units of rapidity ``behind'' the
$W$ for inclusive $W$ events, at the Tevatron, i.e.\ for the
interval $1.0<y_W<1.5$ the $\et$\ flow in the pseudo-rapidity
interval $-1.5<\eta<-1.0$. The full line is the \pythia\ parton
shower and the dotted line is the DCM without (a) and with (b)
multiple interactions.}
\label{figEtflow2}
\end{figure}
The differences between the parton shower and the DCM approaches are
most significant when the $\et$\ flow is measured as a function of
$y_W$ as in fig.\refig{figEtflow2}, where the $\et$\ flow, two units of
rapidity away from the $W$, is shown for both models, with and without
the multiple interaction model for the underlying event implemented in
\pythia\ \cite{TSmult87}. It is clear that the underlying event
introduces an extra $\et$\ for both models, but that the dependence of
the $\et$\ flow on $y_W$ is still different for the two models. As
expected the $\et$\ flow is more or less constant for the parton shower
approach, but increases slightly for the DCM due to the increase in
phase space at large $y_W$.
\begin{figure}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2160,1728)(100,0)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
1377 0 V
600 251 M
0 1326 V
LTb
600 251 M
63 0 V
1314 0 R
-63 0 V
600 630 M
63 0 V
1314 0 R
-63 0 V
600 1009 M
63 0 V
1314 0 R
-63 0 V
600 1388 M
63 0 V
1314 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
875 251 M
0 63 V
0 1263 R
0 -63 V
1151 251 M
0 63 V
0 1263 R
0 -63 V
1426 251 M
0 63 V
0 1263 R
0 -63 V
1702 251 M
0 63 V
0 1263 R
0 -63 V
1977 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
1377 0 V
0 1326 V
-1377 0 V
600 251 L
LT0
1486 1198 M
180 0 V
738 680 M
275 -30 V
276 -39 V
275 -38 V
275 -40 V
LT3
1486 1098 M
180 0 V
738 647 M
275 -5 V
276 1 V
275 -3 V
275 -3 V
stroke
grestore
end
showpage
}
\put(1426,1098){\makebox(0,0)[r]{DCM}}
\put(1426,1198){\makebox(0,0)[r]{PS}}
\put(1288,1677){\makebox(0,0){(a)}}
\put(1288,51){\makebox(0,0){$y_W$}}
\put(400,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$dE_\perp/d\eta$ (GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(1977,151){\makebox(0,0){2.5}}
\put(1702,151){\makebox(0,0){2}}
\put(1426,151){\makebox(0,0){1.5}}
\put(1151,151){\makebox(0,0){1}}
\put(875,151){\makebox(0,0){0.5}}
\put(600,151){\makebox(0,0){0}}
\put(540,1388){\makebox(0,0)[r]{6}}
\put(540,1009){\makebox(0,0)[r]{4}}
\put(540,630){\makebox(0,0)[r]{2}}
\put(540,251){\makebox(0,0)[r]{0}}
\end{picture}
\setlength{\unitlength}{0.1bp}
\special{!
/gnudict 40 dict def
gnudict begin
/Color false def
/Solid false def
/gnulinewidth 5.000 def
/vshift -33 def
/dl {10 mul} def
/hpt 31.5 def
/vpt 31.5 def
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/vpt2 vpt 2 mul def
/hpt2 hpt 2 mul def
/Lshow { currentpoint stroke M
0 vshift R show } def
/Rshow { currentpoint stroke M
dup stringwidth pop neg vshift R show } def
/Cshow { currentpoint stroke M
dup stringwidth pop -2 div vshift R show } def
/DL { Color {setrgbcolor Solid {pop []} if 0 setdash }
{pop pop pop Solid {pop []} if 0 setdash} ifelse } def
/BL { stroke gnulinewidth 2 mul setlinewidth } def
/AL { stroke gnulinewidth 2 div setlinewidth } def
/PL { stroke gnulinewidth setlinewidth } def
/LTb { BL [] 0 0 0 DL } def
/LTa { AL [1 dl 2 dl] 0 setdash 0 0 0 setrgbcolor } def
/LT0 { PL [] 0 1 0 DL } def
/LT1 { PL [4 dl 2 dl] 0 0 1 DL } def
/LT2 { PL [2 dl 3 dl] 1 0 0 DL } def
/LT3 { PL [1 dl 1.5 dl] 1 0 1 DL } def
/LT4 { PL [5 dl 2 dl 1 dl 2 dl] 0 1 1 DL } def
/LT5 { PL [4 dl 3 dl 1 dl 3 dl] 1 1 0 DL } def
/LT6 { PL [2 dl 2 dl 2 dl 4 dl] 0 0 0 DL } def
/LT7 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 1 0.3 0 DL } def
/LT8 { PL [2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 2 dl 4 dl] 0.5 0.5 0.5 DL } def
/P { stroke [] 0 setdash
currentlinewidth 2 div sub M
0 currentlinewidth V stroke } def
/D { stroke [] 0 setdash 2 copy vpt add M
hpt neg vpt neg V hpt vpt neg V
hpt vpt V hpt neg vpt V closepath stroke
P } def
/A { stroke [] 0 setdash vpt sub M 0 vpt2 V
currentpoint stroke M
hpt neg vpt neg R hpt2 0 V stroke
} def
/B { stroke [] 0 setdash 2 copy exch hpt sub exch vpt add M
0 vpt2 neg V hpt2 0 V 0 vpt2 V
hpt2 neg 0 V closepath stroke
P } def
/C { stroke [] 0 setdash exch hpt sub exch vpt add M
hpt2 vpt2 neg V currentpoint stroke M
hpt2 neg 0 R hpt2 vpt2 V stroke } def
/T { stroke [] 0 setdash 2 copy vpt 1.12 mul add M
hpt neg vpt -1.62 mul V
hpt 2 mul 0 V
hpt neg vpt 1.62 mul V closepath stroke
P } def
/S { 2 copy A C} def
end
}
\begin{picture}(2160,1728)(100,0)
\special{"
gnudict begin
gsave
50 50 translate
0.100 0.100 scale
0 setgray
/Helvetica findfont 100 scalefont setfont
newpath
-500.000000 -500.000000 translate
LTa
600 251 M
1377 0 V
600 251 M
0 1326 V
LTb
600 251 M
63 0 V
1314 0 R
-63 0 V
600 630 M
63 0 V
1314 0 R
-63 0 V
600 1009 M
63 0 V
1314 0 R
-63 0 V
600 1388 M
63 0 V
1314 0 R
-63 0 V
600 251 M
0 63 V
0 1263 R
0 -63 V
875 251 M
0 63 V
0 1263 R
0 -63 V
1151 251 M
0 63 V
0 1263 R
0 -63 V
1426 251 M
0 63 V
0 1263 R
0 -63 V
1702 251 M
0 63 V
0 1263 R
0 -63 V
1977 251 M
0 63 V
0 1263 R
0 -63 V
600 251 M
1377 0 V
0 1326 V
-1377 0 V
600 251 L
LT0
1486 725 M
180 0 V
738 1094 M
275 -44 V
276 -40 V
275 -56 V
275 -79 V
LT3
1486 625 M
180 0 V
738 1213 M
275 -3 V
276 -12 V
275 -5 V
275 -52 V
LT1
1486 525 M
180 0 V
738 1391 M
275 -58 V
276 -87 V
275 -57 V
275 -45 V
LT4
1486 425 M
180 0 V
738 1138 M
275 5 V
276 5 V
275 -3 V
275 -34 V
stroke
grestore
end
showpage
}
\put(1426,425){\makebox(0,0)[r]{DCM+MI $\beta=0$}}
\put(1426,525){\makebox(0,0)[r]{ME+PS+MI}}
\put(1426,625){\makebox(0,0)[r]{DCM+MI}}
\put(1426,725){\makebox(0,0)[r]{PS+MI}}
\put(1288,1677){\makebox(0,0){(b)}}
\put(1288,51){\makebox(0,0){$y_W$}}
\put(400,914){%
\special{ps: gsave currentpoint currentpoint translate
270 rotate neg exch neg exch translate}%
\makebox(0,0)[b]{\shortstack{$dE_\perp/d\eta$ (GeV)}}%
\special{ps: currentpoint grestore moveto}%
}
\put(1977,151){\makebox(0,0){2.5}}
\put(1702,151){\makebox(0,0){2}}
\put(1426,151){\makebox(0,0){1.5}}
\put(1151,151){\makebox(0,0){1}}
\put(875,151){\makebox(0,0){0.5}}
\put(600,151){\makebox(0,0){0}}
\put(540,1388){\makebox(0,0)[r]{6}}
\put(540,1009){\makebox(0,0)[r]{4}}
\put(540,630){\makebox(0,0)[r]{2}}
\put(540,251){\makebox(0,0)[r]{0}}
\end{picture}
\fcap{The transverse energy flow in the interval $-2.5<\eta<-2.0$ for
$W$ events at the Tevatron as a function of $y_W$. The full line is
the \pythia\ parton shower and the dotted line is the DCM without
(a) and with (b) multiple interactions. In (b) the dashed line is
\pythia\ using \tordas\ matrix elements with addition of parton
showers and multiple interactions, and the dash-dotted line is the
DCM with $\beta=\infty$.}
\label{figEtflow3}
\end{figure}
Similarly, looking at the $\et$\ flow in a fixed rapidity interval for
varying $y_W$, the DCM predicts a fairly constant value, while in the
parton shower approach, the flow decreases with increasing $y_W$ as in
fig.\refig{figEtflow3}. To check that the differences are not due to
the fact that the DCM has the correct \tordas\ matrix element in the
first emission, fig.\refig{figEtflow3}b also contains a line with the
matrix element plus parton shower option in \pythia. It is clear that,
although the $\et$\ flow is higher since only events with $k_{\perp
W}>10$ GeV are included, it has the same $y_W$ dependence as the
plain parton shower approach.
Also in fig.\refig{figEtflow3}, the changing of the $\beta$ parameter
in the DCM is shown to have some effect on the $\et$\ flow; however,
the dependence on $y_W$ is still the same.
\section{Conclusions}
\label{secsum}
The model presented in this paper is not perfect. The treatment of
initial-state $\g2qq$\ splitting is a bit foreign to the original
Colour Dipole Model, and so is the transfer of transverse recoil to
the $W$ for gluon emissions. But as a leading log approximation it
should be just as good as the conventional parton shower approach, and
it has the advantage of correctly describing also high-$\kt$\ $W$
production. In addition it gives the opportunity of studying effects
of unordered parton evolution on the hadronic final state, and,
although there are some uncertainties in the overall $\et$\ level due
to multiple interactions and the $\beta$ parameter in the DCM, it
would be very interesting to compare the predictions for the $y_W$
dependence of the $\et$\ flow presented in this paper with data from
the Tevatron.
Although only $W$ production at the Tevatron has been discussed in
this paper, the model can of course be applied to any Drell--Yan-like
process in any hadron--hadron collision. And, as pointed out above, the
initial-state $\g2qq$\ splitting can be used also in deep inelastic
lepton--hadron scattering.
The model presented here is also the ``last piece'' to complete the
DCM description of QCD cascades for all standard processes in $\ee$,
$ep$ and $pp$ collisions. This is reflected in the fact that \ariadne\
now is fully interfaced to all hard sub-processes in \pythia. However,
some care must be taken when using the multiple interaction model of
\pythia\ together with the DCM, as mentioned above.
\section*{Acknowledgements}
I would like to thank Bo Andersson and G\"{o}sta Gustafson for
valuable discussions.
|
\section{Introduction}
\label{introduction}
The study of the phase diagram of fundamental--adjoint pure gauge systems,
\begin{equation}
S = \beta_f \sum_{P} [1 - \frac{1}{N} {\rm Re} {\rm Tr} U_{P} ]
+ \beta_a \sum_{P} [1 - \frac{1}{N^2} {\rm Tr} U^\dagger_{P} {\rm Tr}
U_{P}] ,
\label{eq:aaction}
\end{equation}
in the early 80's revealed a non-trivial phase structure with first order
(bulk) transitions in the region of small $\beta_f$
\cite{Greensite81,Bhanot82}. In particular a line of first order
transitions points towards the fundamental axis and, after terminating,
extends as a roughly straight line of bulk crossovers to the fundamental
axis and beyond. This non-trivial phase structure, and in particular the
critical endpoint, has been argued to be associated with the dip in the
discrete $\beta$-function of the theory with standard Wilson action, which
occurs in the region where the bulk crossover line crosses the fundamental
axis.
In a couple of recent papers Gavai, Grady and Mathur \cite{Gavai94} and
Mathur and Gavai \cite{Gavai94b} returned to investigating the behavior of
pure gauge SU(2) theory in the fundamental--adjoint plane at finite,
non-zero temperature. They raised doubts about the bulk nature of the phase
transition and claimed that their results were consistent with the
transitions, for lattices with temporal extent $N_t = 4$, 6 and 8, to be of
thermal, deconfining nature, displaced toward weak coupling with increasing
$N_t$. On the transition line for a fixed $N_t$ there is then a switch from
second order behavior near the fundamental axis to first order behavior at
larger adjoint coupling. In a Landau Ginzburg model of the effective action
in terms of Polyakov lines, Mathur \cite{Mathur95} reported that he could
reproduce the claimed behavior seen in the numerical simulations. These
results, should they be confirmed, are rather unsettling, since they
contradict the usual universality picture of lattice gauge theory with a
second order deconfinement transition for gauge group SU(2).
Puzzled by the finding of Ref.~\cite{Gavai94}, we studied the finite
temperature behavior of pure gauge SU(3) theory in the fundamental--adjoint
plane \cite{SU3_F_A,SU3_F_Ab}. We obtained results in agreement with the
usual universality picture: there is a first order bulk transition line
ending at
\begin{equation}
(\beta^*_f,\beta^*_a) = (4.00(7), 2.06(8)) .
\label{eq:endpoint}
\end{equation}
The thermal deconfinement transition lines for fixed $N_t$ (being of first
order for gauge group SU(3)) in the fundamental adjoint plane are
ordered such that the thermal transition for smaller $N_t$ occurs to the
left, at smaller $\beta_f$, than that for a larger $N_t$. In this order the
thermal transition lines join on to the bulk transition line. The thermal
transition line for $N_t = 4$ joins the bulk transition line very close to
the critical endpoint. This is shown in Figure 4 of Ref.~\cite{SU3_F_Ab},
reproduced here as Figure~\ref{fig:phas_diag_T}.
\begin{figure}
\begin{center}
\vskip 30mm
\leavevmode
\epsfysize=360pt
\epsfbox[90 40 580 490]{phas_diag_T.eps}
\end{center}
\caption{The phase diagram together with the thermal deconfinement
transition points for $N_t=2$, 4, 6 and 8 from
Ref.~\protect\cite{SU3_F_Ab}. The lower plot shows
an enlargement of the region around the end point of the bulk
transition.}
\label{fig:phas_diag_T}
\end{figure}
To solidify this picture, which is in agreement with the usual universality
scenario, we have continued the investigation studying zero-temperature
observables. We computed the string tension and the masses of some
glueballs, in particular the $0^{++}$ glueball, along the thermal transition
line for $N_t = 4$. For universal continuum behavior $\sqrt{\sigma}$ and
the glueball masses should be constant along the thermal transition line for
a fixed $N_t$, leading to constant ratios $T_c/\sqrt{\sigma}$ and
$m_g/\sqrt{\sigma}$.
Of course, at small $N_t$, corresponding to a large lattice spacing $a$,
we expect to see some deviations from this constant behavior. However, we
find large deviations for $m_{0^{++}}/\sqrt{\sigma}$ as we approach the
critical endpoint of the bulk transition along the $N_t=4$ thermal
transition line. The scalar glueball mass decreases dramatically, much more
than what could be expected from simple scaling violations at a large lattice
spacing. On the other hand, this is not really a surprise since at the
critical endpoint at least one mass in the $0^{++}$-channel has to vanish.
We elaborate on our findings in the next sections and then discuss the
implications for the scaling behavior along the fundamental (or Wilson) axis.
\section{Observables and Analysis}
\label{analysis}
We have made simulations of the model with action (\ref{eq:aaction}) along
the thermal transition line for $N_t=4$, and continued along the bulk
transition line, on a $12^4$ lattice. Experience
indicates this size to be sufficient to avoid significant finite size
effects. The simulations were carried out with a 10-hit Metropolis
algorithm tuned to an acceptance rate of about 50\%. Observables were
measured every 20 sweeps after thermalization. Close to the fundamental
axis this resulted in essentially statistically independent measurements,
while closer to the critical endpoint the autocorrelation time was
significantly increased. We computed finite $T$ approximants to the potential
from time-like Wilson loops constructed with `APE'-smeared spatial links
\cite{APE_smear} to increase the overlap to the ground state potential. On
and off axis spatial paths were considered with distances $R = n$,
$\sqrt{2} n$, $\sqrt{3} n$ and $\sqrt{5} n$, with $n = 1$, 2, \dots an
integer. The string tension was then extracted from the usual fit
\begin{equation}
V(\vec R) = V_0 - \frac{e}{R} + l \left( G_L(\vec R) - \frac{1}{R} \right)
+ \sigma R .
\label{eq:fit_form}
\end{equation}
Here $G_L(\vec R)$ is the lattice Coulomb potential, included in the fit
to take account of short distance lattice artefacts. Our fits are fully
correlated $\chi^2$-fits with the correlations estimated by bootstrap,
after binning to alleviate autocorrelation effects. The results
of the best fits are listed in Table~\ref{tab:fits}.
\begin{table}[ht]
\centerline{%
\begin{tabular}{|l|l|l|l|r|l|l|c|} \hline
$\beta_a$ & $\beta_{fc}(N_t=4)$ & $\beta_f$ & $L$ & $N_{meas}$ &
$\sqrt{\sigma}$ & $m_{0^{++}}$ & $m_{2^{++}}(t=1)$ \\ \hline
0.0 & 5.6925(2) & 5.7$^a$ & & & 0.4099( 12) & 0.97( 2) &
2.39(13) \\
0.5 & 5.25(5) & 5.25 & 12 & 500 & 0.4218( 28) & 0.93(11) &
2.29(13) \\
1.0 & 4.85(5) & 4.85 & 12 & 500 & 0.4024( 82) & 0.78(28) &
2.45(19) \\
1.5 & 4.45(5) & 4.45 & 12 & 500 & 0.3743( 51) & 0.56(17) &
2.13( 9) \\
2.0 & 4.035(5) & 4.03 & 12 & 1000 & 0.555 ( 11) & 0.34( 6) &
3.11(16) \\
2.0 & 4.035(5) & 4.035 & 12 & 2000 & 0.4725(128) & 0.20( 4) &
3.17(14) \\
2.0 & 4.035(5) & 4.04 & 12 & 1000 & 0.3750( 24) & 0.27( 8) &
2.37( 9) \\ \hline
2.25 & 3.8475(25) & 3.8475$^b$ & 12 & 500 & 0.619 ( 18) & 0.37(10) &
3.15(37) \\
2.25 & 3.8475(25) & 3.8475$^c$ & 12 & 500 & 0.3005( 22) & 0.61( 8) &
1.59( 7) \\
2.25 & 3.8475(25) & 3.8475$^c$ & 16 & 500 & 0.2965( 19) & 0.62( 5) &
1.72( 9) \\ \hline
\end{tabular}%
\caption{The results in the neighborhood of the $N_t=4$ thermal transition
line. Comments: (a) $\protect\sqrt{\sigma}$ at $\beta_a=0.0$
comes from \protect\cite{MTc_sig}, $m_{0^{++}}$ from
\protect\cite{GF11_gb}; we did not take their best value, but rather
the effective mass from the same distance as at $\beta_a=0.5$;
$m_{2^{++}}$ is from \protect\cite{dFSST86}; (b) in the disordered
phase and (c) in the ordered phase on the bulk transition.}
\label{tab:fits}
\medskip\noindent
\end{table}
We also computed glueball correlation functions in the $0^{++}$, $2^{++}$
and $1^{+-}$ channel that can be built from simple plaquette operators. In an
attempt to improve the signals we built these plaquettes not only from the
original, but also from the smeared links, already used for the computation
of the potential. Not surprisingly for computations done around the
critical coupling for the $N_t = 4$ thermal phase transition, we did not
obtain a significant signal in the $1^{+-}$ channel and only an effective
mass from time distances $t = 0/1$ in the $2^{++}$ channel -- we had 500
measurements everywhere, except for $\beta_a = 2.0$, near the critical
endpoint where the number was increased, as given in Table~\ref{tab:fits}.
In the $0^{++}$ channel we got a signal at distance $t = 1/2$ at small
$\beta_a$ and out to $t = 3/4$ at $\beta_a = 2.0$. Our best results are also
given in Table~\ref{tab:fits}.
The quantities $\sqrt{\sigma}$ and $m_{0^{++}}$ are shown in
Figure~\ref{fig:m_and_sig} plotted versus $\beta_a$. As can be seen,
$\sqrt{\sigma}$ remains approximately constant along the thermal transition
line for $N_t=4$ -- the errors shown in the figure are statistical only; no
estimate of the error from the uncertainty in the determination of
$\beta_{fc}$ has been attempted except for $\beta_a=2.0$. There, the
computation has been repeated for two nearby couplings, also listed in
Table~\ref{tab:fits}; the variation with $\beta_f$ becomes so rapid that
the error in the determination of $\beta_{fc}$ becomes the dominating factor.
While our estimate for $m_{2^{++}}$, albeit an unreliable estimate since we
had to use distances $t=0$ and 1 to obtain enough of a signal to extract an
effective mass, also remains approximately constant, $m_{0^{++}}$ shows a
remarkable decrease as the critical endpoint is approached. This observed
behavior suggests that the mass in the $0^{++}$ channel vanishes at the
critical endpoint, thereby giving strong additional evidence for the
existence of this critical endpoint, since at a critical point at least one
mass, in the $0^{++}$ channel to have a rotationally invariant continuum
limit, must vanish.
Since no other observable seems to be dramatically affected by the critical
endpoint, we conjecture that the continuum theory one would obtain there is
simply the (trivial) $\phi^4$ theory. To substantiate this claim somewhat,
we made a fit to the scalar mass of the form
\begin{equation}
m_{0^{++}} = A \left( \beta^*_a - \beta_a \right)^p
\label{eq:massfit}
\end{equation}
expected near a critical point. A 3-parameter fit gave $A=0.76(11)$,
$p=0.35(20)$ and $\beta^*_a=2.02(6)$ with a $\chi^2$ of $0.29$ for 2 dof.
Note that the estimate for $\beta^*_a$ is in agreement with the previous
estimate (\ref{eq:endpoint}) obtained in \cite{SU3_F_Ab} from fits to the
jump in the plaquette across the bulk transition line. Within its large
error, the exponent $p$ is compatible with the mean field value $0.5$ of
$\phi^4$ theory, up to logarithmic corrections. Since the errors of the
fit parameters are rather large we also made a fit with $\beta^*_a$ held
fixed at its value $2.06$ of (\ref{eq:endpoint}). This fit gave $A=0.71(3)$
and $p=0.44(5)$ with a $\chi^2$ of $0.47$ for 3 dof. Again, $p$ is
compatible with the mean field value. Finally, a fit with $\beta^*_a=2.06$
and $p=0.5$ both held fixed gave $A=0.68(1)$ with a $\chi^2$ of $1.50$ for
4 dof. This last, still very acceptable fit is included in
Figure~\ref{fig:m_and_sig}.
\begin{figure}
\begin{center}
\leavevmode
\epsfysize=360pt
\epsfbox[90 40 580 490]{m_and_sig.eps}
\end{center}
\caption{$\protect\sqrt{\sigma}$ (octagons) and $m_{0^{++}}$ (squares) as
a function of $\beta_a$ along the thermal transition line for
$N_t=4$. At $\beta_a=2.25$ we show the results from both phases
at the bulk transition. The dashed vertical line gives the
location of the critical endpoint, (\protect\ref{eq:endpoint}),
with the dotted lines indicating the error band. The curve is a
fit to $m_{0^{++}}=A(2.06-\beta_a)^{1/2}$.}
\label{fig:m_and_sig}
\end{figure}
\begin{figure}
\begin{center}
\leavevmode
\epsfysize=360pt
\epsfbox[90 40 580 490]{pot_b3p8475_ba2p25.eps}
\end{center}
\caption{The potential in both phases on the bulk transition line at
$(\beta_f,\beta_a)=(3.8475,2.25)$ on $12^4$ lattices ($\Box$ and
$\times$) and on a $16^4$ lattice ($\circ$).}
\label{fig:pot_ba2p25}
\end{figure}
While we believe to have assembled impressive further evidence for the bulk
nature of the phase transition line in the fundamental--adjoint plane and of
its critical endpoint, we decided to make one further test. We computed the
potential in both phases on the bulk transition line at $\beta_a = 2.25$.
This is about the location where the $N_t=6$ thermal transition line joins
onto the bulk transition line according to Ref.~\cite{SU3_F_Ab}. Hence, even
in the ordered phase a lattice of size $12^4$ should be large enough to
obtain a reliable potential. To make sure that this is indeed the case, we
repeated the computation in the ordered phase on a $16^4$ lattice. As can
be seen in Table~\ref{tab:fits} the finite size effect on the $12^4$
lattice does indeed appear negligible. On both sides of the bulk transition
-- at identical couplings: we did not observe even an attempt at tunneling
from one phase to the other -- we find a clearly confining potential.
Therefore the bulk transition does not, as expected, affect confinement,
provided the lattices are large enough.
Of course, the string tension (and glueball masses) jump from one side of
the bulk transition to the other. Indeed, as the bulk transition line is
approached from the fundamental axis the string tension varies more and
more rapidly -- and the thermal deconfinement transition lines for different
$N_t$ come closer together -- {\it i.e.,} the dip in the step $\beta$-function
becomes deeper until a jump is developed through the bulk transition line.
\section{Implications for Scaling}
\label{scaling}
In the previous section we have corroborated the existence of a first order
bulk phase transition line ending in a critical endpoint. We have provided
evidence that physical observables are little affected by this phase
transition line, except for a stronger dependence on $\beta_f$ at fixed
$\beta_a$ -- a deepening of the dip in the step $\beta$-function -- and
eventually a jump across the bulk transition line. The notable exception to
this is the glueball mass in the $0^{++}$ channel which decreases as the
critical endpoint is approached and vanishes there.
\begin{figure}
\begin{center}
\leavevmode
\epsfysize=360pt
\epsfbox[90 40 580 490]{m0_ov_sig.eps}
\end{center}
\caption{$10T_c/\protect\sqrt{\sigma}$ (octagons) and
$m_{0^{++}}/\protect\sqrt{\sigma}$ (squares) as a function of
$\beta$ for the fundamental (Wilson) action.}
\label{fig:m0_ov_sig}
\end{figure}
The influence of the critical endpoint on the $0^{++}$ glueball appears
still visible in the crossover region on the fundamental (Wilson) axis.
This has first been argued in Ref.~\cite{Rajan86}. The $0^{++}$ glueball is
lighter in the crossover region, leading to a different scaling behavior
than other observables. This can be seen in Figure~\ref{fig:m0_ov_sig}
where we show the latest data of $T_c/\sqrt{\sigma}$ from Ref.~\cite{Boyd95}
and $m_{0^{++}}/\sqrt{\sigma}$ with $\sqrt{\sigma}$ taken from
Refs.~\cite{MTc_sig,B-S_sig} and the glueball mass from
Refs.~\cite{GF11_gb,M-T_gb,Bali_gb}. While $T_c/\sqrt{\sigma}$ stays
approximately constant in the $\beta$ interval shown,
$m_{0^{++}}/\sqrt{\sigma}$ decreases visibly in the crossover region around
$\beta = 5.7$.
It has long been known that the scalar glueball mass scales differently in
the crossover region than $T_c$ and the string tension, whose scaling
behavior as a function of $\beta$ deviates from asymptotic scaling but
agrees with the step $\beta$-function found in MCRG computations. The
scalar glueball seemed more compatible with asymptotic scaling. However, in
view of our findings that the different behavior of the scalar glueball
mass comes from the influence of the nearby critical endpoint this
asymptotic scaling behavior appears to be accidental. It would be
interesting to determine the scaling behavior of other (pure gauge)
observables. We conjecture that they would behave more like $T_c$ or
$\sqrt{\sigma}$ than like the scalar glueball mass. Unfortunately no
reliable, large volume, glueball masses for $J^{PC}$ quantum numbers other
than $0^{++}$ for couplings in the crossover region exist in the literature
to check this conjecture.
\section*{Acknowledgements}
This work was partly supported by the DOE under grants
\#~DE-FG05-85ER250000 and \#~DE-FG05-92ER40742. The computations were
carried out on the cluster of IBM RS6000's at SCRI.
|
\section{The Gauged Linear Sigma Model}
We begin with a brief review of the gauged linear sigma model
(GLSM) construction of Witten\cite{phases}.
The model is formulated in (2,2)
superspace, and requires for its construction the specification
of a compact abelian group $G$, a faithful representation $\rho:G\toU(1)^n$,
and a $G$-invariant polynomial $W(x_1,\dots,x_n)$,
where $x_1, \dots, x_n$ are coordinates in a
complex vector space ${\bf} \def\bf{\bf C}^n$
on which $U(1)^n$ acts diagonally.
To construct the gauged linear sigma model, we begin with $n$ chiral
superfields $\Phi_i$ (satisfying $\overline{D}_+\Phi_i =
\overline{D}_-\Phi_i=0$) interacting via the holomorphic superpotential
$W(\Phi_1,\ldots,\Phi_n)$. The model is invariant under the action of
$G$ (via $\rho$) on $\vec\Phi$ and we gauge this action, preserving $N{=}2$
supersymmetry, by introducing the ${\bf g}$-valued vector multiplet
$V$ with invariant field strength
$\Sigma=\frac1{\sqrt{2}}\overline{D}_+D_-V$. This last field is
{\it twisted chiral}, which means that $\overline{D}_+\Sigma
=D_-\Sigma=0$. For each continuous $U(1)$ factor of $G$ we include
a Fayet-Illiopoulos $D$-term and a $\theta$-angle; these terms are
naturally written in terms of the complex combination $\tau=ir+\frac
1{2\pi}\theta$.
The resulting Lagrangian density is thus
\begin{eqnarray}
{\cal L}&=&\int d^4\theta\left(\|e^{R({V})}\vec{\Phi}\|^2
-\frac1{4e^2}\|{\Sigma}\|^2\right)\nonumber\\
&&+ \left(\int d\theta^+d\theta^- W(\Phi_1,\dots,\Phi_n) +
\mbox{ c.c. } \right)\label{GLSM}\\
&&+ \left( {i\over\sqrt 2}\,
\int d\theta^+d\bar\theta^- \langle\tau,\Sigma\rangle
+ \mbox{ c.c. } \right) ,\nonumber
\end{eqnarray}
where $R=-i\,d\rho:{\bf g}\to{\bf} \def\bf{\bf R}^n$ is the
derivative of the representation $\rho$ (with a factor of $-i$
to make it real-valued).
Concretely, the general compact abelian group takes the form $G =
U(1)^{n-d}\times\Gamma$ where $\Gamma$ is a product of finite cyclic groups.
Choosing a basis for the continuous part we have $n{-}d$ vector
multiplets $V_a$; the action on the fields is given by $R(V_a)\Phi_i =
Q_i^a\Phi_i$ for some integer charges $Q_i^a$.
Notice that the discrete group $\Gamma$ does not appear explicitly
in the Lagrangian, but it does affect the construction of the field
theory---the fields are sections of bundles
with structure group $G$.
In the case that $\Gamma$ is nontrivial,
we recover in this way an orbifold of
another theory (cf.\ Ref.~\citelow{phases}).
We will be interested in families of such models, parameterized by the
coefficients of $W$ and by the instanton factors $q_a := e^{2\pi
i\tau_a}$. (There will be a set of complex codimension one in the
parameter space along which the model is singular; we will study values
of the parameters away from this locus.)
A family is thus characterized by the group $G$ and the
collection of monomials appearing in $W$.
In order to specify these data, it is convenient
to introduce a $u\times n$ matrix $P$ of rank $d$ with nonnegative integer
entries, and a factorization $P=ST$ of $P$ as a product of integer matrices
$S$ and $T$, each of rank $d$. The rows $t_\alpha$ of $T$ can then be
used to construct a collection of Laurent monomials
$x^{t_\alpha}:=\prod x_i^{t_{\alpha i}}$, and the group $G$ is defined to
be the largest subgroup of $H=U(1)^n$ which leaves the monomials
$x^{t_\alpha}$
invariant. The monomials $x^{p_r}$ defined by the rows of $P$ are then
$G$-invariant by construction, thanks to the relation
$p_{ri}=\sum s_{r\alpha}t_{\alpha i}$.
Since $p_{ri}\ge0$ by assumption, we may use these monomials to specify
the family of interaction polynomials
\begin{equation}\label{W}
W(x_1,\dots,x_n):={{1\over 2\pi\sqrt 2}}\sum_{r=1}^u c_r x^{p_r}={{1\over 2\pi\sqrt 2}}\sum_{r=1}^u c_r
\prod_{i=1}^n x_i^{p_{ri}}\ .
\end{equation}
Alternatively, if we are given $G$ and a family of polynomials
$W$, it is not difficult to reconstruct
the matrices $P$, $S$, and $T$. (Actually, $S$ and $T$ are only well-defined
up to $(S,T)\mapsto (SL,L^{-1}T)$, with $L$ an invertible integer matrix.)
If we start with only $W$, as specified by the rows of the matrix
$P$, then the choice of a factorization $P=ST$ amounts to a choice of a
subgroup $G\subset U(1)^n$ (which can be less than maximal) under which $W$
is invariant.
On the other hand, if we start with only $G$ then the choice of
factorized matrix
$P=ST$ allows us to specify which of the $G$-invariant monomials
should be included in $W$. In particular, it is possible to omit
some monomials and in this way to study
subsets of the
maximal set of all $G$-invariant superpotentials. As we shall see,
this possibility is useful. (The assumption made above on the rank of
$S$ and $T$ restricts this choice---we must use ``enough'' monomials
to get the rank to be correct---but the restriction is not
essential and relaxing it would simply require a more cumbersome notation.)
Under these conditions, the
possible choices for factorizations (and hence for $G$) for a given $P$
are rather limited. In any factorization $P=ST$, the rows of $T$ will
generate an integral lattice of rank $d$, and the rows of $P$ will
generate a sublattice, also of rank $d$. This implies that the quotient
\[\mathop{\rm rowlattice}\nolimits(T)/\mathop{\rm rowlattice}\nolimits(P)\]
is a torsion subgroup of ${\bf} \def\bf{\bf Z}^n/\mathop{\rm rowlattice}\nolimits(P)$. There are only
finitely many of choices of such a torsion subgroup, once $P$
has been fixed.
\section{The $R$-Symmetry}
The gauged linear sigma model is not conformally invariant, and will
flow to strong-coupling in the infrared. Generically such a theory
would be expected to develop a mass gap, but
evidence has been found\cite{phases,SilWit} that the theory at
the IR fixed point will be nontrivial if we require that
the high-energy theory admit a non-anomalous chiral $R$-symmetry.
We consider a right-moving $R$-symmetry, acting in such a way that
$\theta^+$ has charge $1$ and $\theta^-$ is neutral.
Invariance of the kinetic terms requires that
the gauge fields ${V}$ be neutral, and their field strengths
${\Sigma}$ hence have charge $1$. Invariance of the superpotential
interaction requires that the superpotential have charge $1$. If we let
$\mu_i$ denote the $R$-charge of $\Phi_i$ (which may be a rational number)
this tells us that
\[\sum_{i=1}^n p_{ri} \mu_i=1 \quad \mbox{for all $r$}\ .\]
This chiral symmetry can be anomalous in the presence of the gauge
fields. A quick computation\cite{phases,summing} shows that
the anomaly is given by a function on the Lie
algebra proportional to $V\mapsto\mathop{\rm trace}\nolimits(R(V))$;
we require that this vanish identically, i.e., that the symmetry be
nonanomalous. Since the action of the continuous part
of $G$ on the monomial $x_1\cdots x_n$ is via
$\exp(\mathop{\rm trace}\nolimits(R(V)))$, this is the same as requiring
that $x_1\cdots x_n$
be invariant under the continuous part of $G$ (or in our explicit
coordinates that $\sum_i Q_i^a=0$ for all $a$). This in turn will hold
exactly when the vector of exponents $(1,\dots,1)$ is a linear combination
of the rows of the matrix $T$, with rational coefficients. (If we had wanted
$x_1\cdots x_n$ to be invariant under all of $G$, we would have insisted
upon integer coefficients.) That is, there is a rational
vector $\vec{\lambda}$ such that $\vec{\lambda}^{\rm T} T=(1,\dots,1)$.
Since we are assuming that the $d\times u$ matrix $S^{\rm T}$
has rank $d$, it
is possible to solve the equation $S^{\rm T}\vec{\nu}=\vec{\lambda}$ for
a rational vector $\vec{\nu}$. If we also set
$\vec{\mu}=(\mu_1,\dots,\mu_n)^{\rm T}$, then
we can write the conditions for an
unbroken $R$-symmetry as the existence of $\vec\mu,\,\vec\nu$ such
that
\begin{eqnarray}
P\vec{\mu}&=&(1,\dots,1)^{\rm T}\label{eq1}\\
\vec{\nu}^{\rm T} P&=&(1,\dots,1)\ .\label{eq2}
\end{eqnarray}
Finally, using the $R$-symmetry and
calculating as in Ref.~\citelow{SilWit}, one finds that
the central charge $c$ of the fixed-point CFT is determined by
\[d-(c/3)=2\sum_{i=1}^n\mu_i=2\,\vec{\nu}\,{}^{\rm T}
P\vec{\mu}=2\sum_{r=1}^u\nu_r\ .
\]
It may be useful to have some examples.
\vspace*{0.6cm}\noindent{\normalsize\it Example 1.}\par\vspace*{0.4cm}
Consider $1\times1$ matrices
$(p)=(s)(t)$ with $s$ and $t$ positive integers. Note that
\eqsref{eq1} and (\ref{eq2}) are trivially satisfied in this case. Since
$G$ is the group
which leaves $x^t$ invariant, we must have $G={\bf} \def\bf{\bf Z}/(t)$. The interaction
polynomial is $W(x)=x^p$, so our model is an orbifold of the
Landau-Ginsburg version of a minimal model.
\vspace*{0.6cm}\noindent{\normalsize\it Example 2.}\par\vspace*{0.4cm}
Consider a $u\times 6$ matrix $P$ of rank $5$ such
that $p_{0i}=p_{r0}=1$, (letting the $i$ and $r$ indices start at $0$
for this example) implementing \eqsref{eq1} and (\ref{eq2}) directly.
This leads to a polynomial of the form
\[W(x_0,\dots,x_5)=x_0\,W'(x_1,\dots,x_5)=
{{1\over 2\pi\sqrt 2}}\left(c_0\,x_0x_1x_2x_3x_4x_5+x_0\,f(x_1,\dots,x_5)\right),\]
where $f$ is a polynomial in $5$ variables containing precisely
$u{-}1$ monomials. The existence of a kernel for $p$, an integral
generator of which we denote by $(-k,\ell_1,\dots,\ell_5)^{\rm T}$,
means that $W$ is invariant under a $U(1)$-action on
$(x_0,\dots,x_5)$ with weights
$(-k,\ell_1,\dots,\ell_5)$---this implies that $f$ is
quasi-homogeneous of degree $k$ and that $k=\sum \ell_i$.
One possible factorization $P=ST$
would then be given by taking the rows of $T$ to be a basis for the
$U(1)$-invariant Laurent monomials, and expressing the monomials
appearing in $f$ in terms of those. By construction, this leads to
$G=U(1)$. In general (if $f$ is a generic quasi-homogeneous
polynomial) this is the only possible factorization. For special
subfamilies (determined by a subset of the rows of
the maximal $P$) the
polynomial is invariant under additional discrete symmetries (the
restriction on $\mathop{\rm rank}\nolimits(P)$ constrains the continuous part), and
other factorizations are possible, leading
to groups $G$ with a nontrivial discrete part.
As we will see below, the weights we have given specify a weighted
projective space ${\bf} \def\bf{\bf P}^{(\ell_1,\dots,\ell_5)}$, and
$\{f{=}0\}\subset {\bf} \def\bf{\bf P}^{(\ell_1,\dots,\ell_5)}$
defines a (singular) Calabi--Yau hypersurface in that space,
closely related to the GLSM built from the data $P=ST$.
A particularly interesting case is when $u=6$, and $P$ is $6\times6$.
The description of these models in terms of the matrix $P$ was
first given by Candelas, de la Ossa, and Katz\cite{cdk} in the
course of generalizing the Berglund--H\"ubsch\cite{BHub} construction.
\vspace*{0.6cm}\noindent{\normalsize\it Example 3.}\par\vspace*{0.4cm}
To make the previous example a bit more concrete, consider a Fermat-type
polynomial
\[f(x_1,\dots,x_5)=x_1^{a_1}+x_2^{a_2}+x_3^{a_3}+x_4^{a_4}
+x_5^{a_5}\]
with $\ell_j=k/a_j$ and suppose that $\ell_5=1$.
Choosing $G=U(1)$ as in example 2 corresponds to the factorization
\[
\left(\begin{array}{cccccc}
1&1&1&1&1&1\\ 1&a_1&0&0&0&0\\ 1&0&a_2&0&0&0\\ 1&0&0&a_3&0&0\\
1&0&0&0&a_4&0\\ 1&0&0&0&0&a_5
\end{array}\right)=
\left(\begin{array}{ccccc}
1&1&1&1&1\\ 1&a_1&0&0&0\\ 1&0&a_2&0&0\\ 1&0&0&a_3&0\\ 1&0&0&0&a_4\\
1&0&0&0&0
\end{array}\right)
\left(\begin{array}{cccccc}
1&0&0&0&0&k\\ 0&1&0&0&0&-\ell_1\\ 0&0&1&0&0&-\ell_2\\
0&0&0&1&0&-\ell_3\\ 0&0&0&0&1&-\ell_4
\end{array}\right).
\]
On the other hand, to maximize the group $G$ we should use the entire
row space of $P$ as the row space of $T$ which leads to the factorization
\[
\left(\begin{array}{cccccc}
1&1&1&1&1&1\\ 1&a_1&0&0&0&0\\ 1&0&a_2&0&0&0\\ 1&0&0&a_3&0&0\\
1&0&0&0&a_4&0\\ 1&0&0&0&0&a_5
\end{array}\right)=
\left(\begin{array}{ccccc}
1&0&0&0&0\\ 0&1&0&0&0\\ 0&0&1&0&0\\ 0&0&0&1&0\\ 0&0&0&0&1\\
k&-\ell_1&-\ell_2&-\ell_3&-\ell_4
\end{array}\right)
\left(\begin{array}{cccccc}
1&1&1&1&1&1\\ 1&a_1&0&0&0&0\\ 1&0&a_2&0&0&0\\ 1&0&0&a_3&0&0\\
1&0&0&0&a_4&0
\end{array}\right).
\]
This corresponds to an orbifold of the previous model; in fact, this
is the quotient found by Greene and Plesser\cite{gp} to correspond to
the mirror manifold. Thus we can reproduce the construction of
Ref.~\citelow{gp} in this context by appropriate choices of factorization.
\section{The Low-Energy Limit}
As explained in detail in Ref.~\citelow{phases}, the low-energy limit of this
theory can be described explicitly and---when the central charge is
an integer---often coincides with a sigma-model
on a Calabi--Yau space (for appropriate values of the parameters).
To study the low-energy limit one begins by mapping out the space of
classical vacua of the theory. To this end, we first solve the algebraic
equations of motion for the auxiliary fields $D_a$ (in the vector
multiplets) and $F_i$ (in the chiral multiplets)
\begin{eqnarray}
D_a &=& -e^2\left(\sum_{i=1}^n Q_i^a |\phi_i|^2 - r_a\right)\label{D}\\
F_i &=& -{\partial W\over\partial\phi_i}\ .\label{F}
\end{eqnarray}
The potential energy for the bosonic zero modes is then
\[U = {1\over 2e^2}\sum_{a=1}^{n-d}D_a^2 + \sum_{i=1}^n |F_i|^2 +
\sum_{a,b}\bar\sigma_a\sigma_b\sum_{i=1}^n Q_i^aQ_i^b|\phi_i|^2\ ,\]
where $\phi_i,\,\sigma_a$ are the lowest components of
$\Phi_i,\,\Sigma_a$ respectively. The space of classical vacua is the
quotient by $G$ of the set of zeros of $U$.
Neglecting for a moment the superpotential, the space of solutions
(setting $\sigma=0$) is $D^{-1}(0)/G$. For values of the instanton
factors $q_a$ in a suitable range this is a toric variety of dimension
$d$. In the ``geometric'' case in which $\vec{\mu}$ has $N$ $1$'s and $n{-}N$
$0$'s, this variety can be recognized as the total space of a sum of
$N$ line bundles over a {\it compact}\/ toric variety of dimension $d-N$.
The equations $F_i=0$ then determine (for generic values of the
parameters $c_r$) a complete intersection subvariety of codimension
$N$ in the base space, homologous to the intersection of the divisors
associated to the $N$ line bundles. The condition in \eqref{eq2}
implies that this subvariety is Calabi--Yau.
The fixed-point CFT is given by the nonlinear sigma model with this
target space. The moduli of this CFT are the marginal operators in
\eqref{GLSM}. Na\"{\i}vely, both the $q_a$ and the $c_r$ would appear to
be marginal
couplings. The latter, however, are not all independent. As is
well known, some of them can be absorbed in rescalings of the fields
$\Phi_i$. The true moduli are thus the $q_a$ and the scaling-invariant
combinations of the $c_r$. There will in general be other marginal
operators in the model which do not appear explicitly in the
Lagrangian; we restrict our attention to the subspace of those that do.
Of course, the linear (or more accurately, toric) structure with which we
have endowed our moduli space is an artifact of our description. In
particular, it is consistent with the natural complex structure that
this moduli space carries intrinsically, but bears no relation to the
K\"ahler structure determined by the Zamolodchikov metric. In other
words, the $q_a$ and $c_r$ are not the ``special'' coordinates on this
space.
For example, in example 2 above in which $G=U(1)$, the $D$-term equation is
\[-k\,|x_0|^2+\sum_{i>0}\ell_i|x_i|^2 - r=0,\]
so when $r>0$ we cannot have $x_i=0$ for all $i>0$. The space of
solutions $D^{-1}(0)/G$ can be recognized as the total space of the
canonical bundle over ${\bf} \def\bf{\bf P}^{(\ell_1,\dots,\ell_5)}$. If we impose
the $F$-term equation as well
\[\sum_{i=0}^5\left|\frac{\partial W}{\partial x_i}\right|^2
=|W'|^2+|x_0|^2\sum_{i=1}^5\left|\frac{\partial W'}{\partial x_i}\right|^2
=0,\]
then for a generic choice of the coefficients of $W$ (away from a
codimension-one subspace we avoid as promised above) the space of
solutions is given by $x_0=0$ and $W'(x)=0$, yielding a
Calabi--Yau hypersurface in ${\bf} \def\bf{\bf P}^{(\ell_1,\dots,\ell_5)}$. Further
study shows that the low-energy excitations are tangent to this, so
that the low-energy theory is a nonlinear sigma model with this Calabi--Yau
target space.
\section{Mirror Symmetry}
We are now in a position to state the mirror symmetry conjectures for
this class of models. Mirror symmetry relates two Calabi--Yau
manifolds which lead to isomorphic CFT's when used as target spaces
for supersymmetric nonlinear sigma models. The mirror isomorphism
reverses the sign of the right-moving
$R$-symmetry; many of the fascinating properties of mirror pairs can
be traced to this feature. Given a GLSM family determined by $P=ST$ we
will construct the mirror family in the same class of models.
We incorporate the sign change on $R$ by writing the
dual model using {\it twisted}\/ chiral matter fields coupled to twisted
gauge multiplets (with {\it chiral}\/ field strengths). We will call such
a model a {\it twisted}\/ gauged linear sigma model. As above, we
characterize the family of models by a factorized matrix of
nonnegative integers $\widehat P = \widehat S\widehat T$. We use this
data to construct a twisted superpotential $\widehat W$ and write a
Lagrangian density (compare \eqref{GLSM})
\begin{eqnarray}
{\cal L}&=&\int d^4\theta\left(-\|e^{\widehat R({\widehat V})}
\widehat {\vec{\Phi}}\|^2
+\frac1{4e^2}\|{\widehat \Sigma}\|^2\right)\nonumber\\
&&+ \left(\int d\theta^+d\bar\theta^- \widehat W({\widehat\Phi}) +
\mbox{ c.c. } \right)\label{TGLSM}\\
&&+ \left( {i\over\sqrt 2}\,
\int d\theta^+d\theta^- \langle\widehat \tau,\widehat \Sigma\rangle
+ \mbox{ c.c. } \right) ,\nonumber
\end{eqnarray}
where $\widehat\Sigma = -{1\over 2\sqrt 2}\bar D_+\bar D_-\widehat V$ is the
(chiral) field strength.
The {\it mirror conjecture}\/ states that if we set
\begin{eqnarray}
\widehat P &=& P^{\rm T}\nonumber\\
\widehat S &=& T^{\rm T}\label{mirp}\\
\widehat T &=& S^{\rm T}\nonumber
\end{eqnarray}
the two families of CFT's defined by the infrared limits of the two
linear models are isomorphic. This is essentially just a translation
of the conjecture of Batyrev and Borisov\cite{batbor}, versions of
which have appeared in Refs.~\citelow{ag:rigid,cdk}.
In fact, the statement given here generalizes the
conjecture somewhat since the factorization makes possible a map
between subfamilies.
One can make a more precise statement of the conjecture---specifying
explicitly the map between the two parameter spaces which relates
isomorphic low-energy limits
\begin{eqnarray}
q_a &=& \prod_{r=1}^u \widehat c_i^{Q_i^a}\nonumber\\
\widehat q_{\widehat a} &=& \prod_{i=1}^n c_r^{\widehat Q_r^{\widehat
a}}\ .\label{mondiv}
\end{eqnarray}
Note that as expected only the true moduli appear on the right-hand
side of \eqref{mondiv}.
The asymptotic limits of \eqref{mondiv} constitute
the {\it
monomial-divisor mirror map}, first proposed in Refs.\ \citelow{catp,mondiv}
(see also Ref.~\citelow{Bat:qu})
and extended in Refs.\ \citelow{ag:rigid,multdiv}.
The ability to
write a global version of this
map and not just an asymptotic form is related to the
coordinates we use on the moduli space (for a full discussion of this
including a proof of \eqref{mondiv} for a class of examples see
Ref.~\citelow{summing}).
We note that this statement of the conjecture is very reminiscent of
recent results on duality in four-dimensional supersymmetric gauge
theories\cite{SWI,SWII,S}. Two distinct gauge theories
with nontrivial dynamics lead to the same low-energy physics; further,
instanton effects in one model are reproduced by classical physics
in the other.
In example 3 above, it should be clear that this construction
interchanges the two choices of factorization, leaving the (symmetric)
matrix $P$ unchanged. As mentioned above, this reproduces the
orbifold construction of the mirror in Ref.~\citelow{gp}, clearly and
explicitly demonstrating how the general construction reduces to this
upon restricting to a subfamily of all possible $W$---precisely the
subfamily invariant under the maximal discrete group.
\section{Abelian Duality}
The formulation we have given for the mirror conjecture in the context of
the GLSM immediately suggests a relation to abelian duality. We recall
that given a theory with an abelian continuous global symmetry, one can
use duality to obtain an equivalent theory. The dual model is obtained
by gauging the global symmetry, introducing Lagrange multipliers which
constrain the connection to be pure gauge (ensuring that the theory is actually
unchanged). One must then eliminate the original degrees of freedom,
in the process generating a nontrivial effective
action for the Lagrange multipliers, which become the fundamental degrees
of freedom for the dual model. Classically this is accomplished by
gauge-fixing. In a theory with
$N{=}2$ supersymmetry and chiral charged fields, the Lagrange multipliers
mentioned above will be {\it twisted\/} chiral fields, and will appear
in the twisted superpotential as $\widehat\Lambda\Sigma$. Thus the dual
variables will naturally be twisted chiral\cite{GHR}, as expected for
the mirror model. This idea was pursued in Refs.\ \citelow{G,RV,GW,BH}, and in
special cases leads to an interpretation of mirror symmetry as abelian
duality.
In the general case, however, the approach runs into difficulties. The
most obvious one is that a generic Calabi--Yau manifold $M$ has no isometries,
hence the associated two dimensional CFT lacks suitable global symmetries.
If this approach is to lead to the equivalence of GLSM models which is
tantamount to the mirror conjecture, some modification will be required.
The two linear models are {\it not\/}
equivalent; the conjecture implies only that they become equivalent
in the extreme low-energy limit. This is consistent with the fact
that the model has nontrivial dynamics. As discussed above this is
similar to the recent discoveries in four-dimensional supersymmetric
gauge theories.
In the context of the GLSM, there is a natural symmetry
group one would attempt to use. This is the group $H = U(1)^n$ acting
by phases on the fields $\Phi_i$. The difficulty, of course, is that
this symmetry is explicitly broken by the superpotential \eqref{W} to the
subgroup $G$. Since this symmetry is gauged in \eqref{GLSM}, the
resulting model has no global symmetries at all (except the $R$-symmetry
which we cannot gauge).\footnote{This once led E. Witten to
describe the problem of understanding mirror symmetry in this model as
the question ``How to perform duality on a non-symmetry?''} We can
attempt to overcome this obstacle by recasting the symmetry-breaking
terms as anomalies. To implement this idea, introduce a set of
twisted chiral fields $\widehat\Psi_r$ into the model, coupled to
twisted gauge multiplets $\widehat V_r$ gauging the
group $\widehat H = U(1)^u$ which
acts by phases, i.e.\ consider the additional term in the
kinetic energy
\begin{equation}
{\cal L}_{\widehat k} = \int d^4\theta
\left( -\|\widehat{\vec\Psi} e^{\widehat{V}}\|^2 +{1\over 4e^2}\|{\widehat{\Sigma}}\|^2\right)\ .
\label{khat}
\end{equation}
We then replace the superpotential \eqref{W} by
\begin{equation}
W_s = {{1\over 2\pi\sqrt 2}} \sum_{r=1}^u{\widehat{\Sigma}}_r\left[\log\left(
c_r\prod_{i=1}^n\Phi_i^{p_{ri}}\right)+1\right]\ .
\label{Ws}
\end{equation}
In this form it is clear that $H$ is broken by $\widehat H$ anomalies, as
desired. Note that this interaction does not break the $R$-symmetry.
However, the couplings $c_r$ have nonzero beta functions,
which will cause them to grow large, so the $\widehat H$ gauge theory is
strongly coupled at low energies. In this limit, as discussed in
Refs.\ \citelow{phases,summing}, this sector of the model is in a confining
phase, in which the lowest component $\widehat\sigma$ of ${\widehat{\Sigma}}$ gets
a nonzero expectation value. The charged fields are then all massive,
with masses of order this expectation value, and the
light degrees of freedom are in the ${\widehat{\Sigma}}$ multiplet.
The effective action for these can be reliably computed at
one-loop order. Integrating out the $\widehat\Psi$'s leads to the effective
superpotential
\begin{equation}
W_{\rm eff} = {{1\over 2\pi\sqrt 2}}\sum_{r=1}^u {\widehat{\Sigma}}_r\left[\log\left(
c_r\prod_{i=1}^n\Phi_i^{p_{ir}}\right)+1 - \log{\widehat{\Sigma}}_r\right]\ .
\label{Wsef}
\end{equation}
In the effective theory we can treat ${\widehat{\Sigma}}$ as a chiral field; since
there are no charged fields we can forget its relation to a gauge
symmetry. At low energy, we can eliminate ${\widehat{\Sigma}}$ from \eqref{Wsef} to get
${\widehat{\Sigma}}_r = c_r\prod_{i=1}^n\Phi_i^{p_{ir}}$. When substituted in
\eqref{Wsef}, this leads to \eqref{W} as the effective superpotential for
$\Phi$. Thus at low energies this model is equivalent to the original
GLSM, while the symmetry-breaking terms are explicitly exhibited as
anomalies.
We now wish to perform a duality transformation with respect to the
anomalous symmetry. Gauging an anomalous symmetry appears inconsistent,
but in performing a duality transformation an anomalous abelian
symmetry can be restored
by assigning transformation properties (under the twisted symmetries) to
the Lagrange multipliers.\cite{GR} In the case at hand, gauging
the global symmetry means that the gauge group becomes all of $H$.
We should also introduce Lagrange multipliers, transforming under
$\widehat H$, to constrain the field strength to lie in $\bf
g\subset\bf h$. Classically, integrating these out will reproduce
the original model. This is not true quantum mechanically, however,
as is most easily seen by introducing a small kinetic term for the
Lagrange multipliers, and considering the limit as this term is
removed. It then becomes clear that the contribution of these to the
one-loop effective superpotential for ${\widehat{\Sigma}}$ is constant in the limit.
Introducing these extra fields would thus destroy the one-loop
calculation that led to \eqref{Wsef}.
This suggests that we
consider a related Lagrangian, in which the twisted chiral matter is
an accurate ``reflection'' of the chiral matter. This defines a model
that is manifestly mirror-symmetric. Our proposal for this new model
contains chiral
fields $\vec\Phi$ and twisted chiral fields $\widehat{\vec\Phi}$, and
(twisted) gauge multiplets gauging the entire group $H$ ($\widehat H$).
We begin with the Lagrangian density
\begin{eqnarray}
{\cal L} &=& \int d^4\theta
\left( \sum_i \left( J_i - {1\over 4e^2}|\Sigma_i|^2\right) -
\sum_r\left( \widehat J_r - {1\over 4e^2}|{\widehat{\Sigma}}_r|^2\right) +
{1\over 2\pi}\sum_{ri}p_{ri} \log J_i \log \widehat J_r \right)\nonumber\\
&& + \left({{i\over \sqrt 2}}\int d\theta^+d\theta^-\langle{\widehat{\Sigma}},\widehat\tau\rangle
+{\rm c.c}\right)
+ \left({{i\over \sqrt 2}}\int d\theta^+d{\bar\theta}^-\langle\Sigma,\tau\rangle
+{\rm c.c}\right) \ ,\label{our}
\end{eqnarray}
where $J_i = |\Phi_ie^{V_i}|^2$ and likewise for $\widehat J_r$.
This action is manifestly invariant under $H\times\widehat H$. It is
classically equivalent to the action we obtain by
integration by parts,
\begin{eqnarray}
{\cal L_{\rm eq}} &=& \int d^4\theta
\left( \sum_i \left( J_i - {1\over 4e^2}|\Sigma_i|^2\right) -
\sum_r\left( \widehat J_r - {1\over 4e^2}|{\widehat{\Sigma}}_r|^2\right) +
{2\over \pi}\sum_{ri}p_{ri} V_i\widehat V_r\right)\nonumber\\
&& + \left(\int d\theta^+d\theta^- W({\widehat{\Sigma}},\Phi)+{\rm c.c}\right)
+ \left(\int d\theta^+d{\bar\theta}^- \widetilde W(\Sigma,{\widehat\Phi})
+{\rm c.c}\right) \ ,\label{ourpi}
\end{eqnarray}
where the modified superpotentials are
\begin{eqnarray}
W({\widehat{\Sigma}},\Phi) &=& {{1\over 2\pi\sqrt 2}}\sum_{r=1}^u {\widehat{\Sigma}}_r\left[\log\left(
\widehat q_r\prod_{i=1}^n\Phi_i^{p_{ri}}\right)+1\right] \nonumber\\
\widetilde W(\Sigma,{\widehat\Phi}) &=& {{1\over 2\pi\sqrt 2}}\sum_{i=1}^n \Sigma_i\left[\log\left(
q_i\prod_{r=1}^u{\widehat\Phi}_r^{-p_{ri}}\right)+1\right]\ ,\label{Wt}
\end{eqnarray}
We see that the conditions for a
nonanomalous $R$-symmetry are precisely \eqsref{eq1} and (\ref{eq2}).
When these hold we expect to find at low energies an $N{=}2$
superconformal field theory where the $R$-symmetry defined above
becomes the chiral $U(1)$ contained in the superconformal algebra.
We will use the second formulation as our definition of the quantum
theory. (In the presence of instantons the integration by parts
requires the consideration of boundary terms, which are nontrivial.)
It is worthwhile noting the way in
which \eqref{ourpi} manages to be
gauge-invariant despite the manifestly non-invariant interactions.
Under a gauge transformation (in $H\times\widehat H$) the variation of
\eqref{Wt} is cancelled, up to a total derivative, by the variation of
the $V\widehat V$ term. Thus the full action is invariant under gauge
transformations approaching the identity at infinity, while
transformations with constant parameter are still anomalous. We note
that this cancellation holds precisely when the exponents in the two
superpotentials are related as in \eqref{Wt}. In a derivation of
mirror symmetry along these lines this would be the origin of the
first part of \eqref{mirp}.
At this point one still needs to show that the theory defined by
\eqref{ourpi} is equivalent to \eqref{GLSM}. This can be achieved
using the fact that the coefficients $q_i,\,\widehat q_r$ once more
contain redundant deformations which can be undone by field rescalings
(in fact this is precisely the holomorphic extension of the statement
above about anomalies). One can perform a redundant deformation of
the model to find a region in parameter space in which the ${\widehat{\Sigma}},{\widehat\Phi}$
system is in a confining phase and the one-loop approximation that led
to \eqref{Wsef} is valid. One can also show that in this region the
integration over ${\widehat\Phi}$ will lead to constraints on
$\Sigma$, restricting the gauge group to $G$. In fact, the last two
equations of \eqref{mirp} arise naturally in this context as well.
The manifest mirror
symmetry would then lead one to expect that \eqref{ourpi} will also
be equivalent
to \eqref{TGLSM} under the conditions \eqsref{mirp} and (\ref{mondiv}).
To see this one performs redundant deformations to a region in
parameter space in which the $\Sigma,\Phi$ system confines and follows
a ``mirror'' version of the argument sketched above. This naturally
incorporates all of the conditions in \eqsref{mirp} and (\ref{mondiv}).
We hasten to point out however that \eqref{ourpi} is not in fact
mirror symmetric. The reason for
this is the sign in the second of \eqref{Wt}.\footnote{This sign was
inadvertently omitted in earlier stages of this work, leading to
conclusions---reported in several talks by the
authors---differing from those presented here.}
Unexpectedly, this
sign, which is required for gauge invariance since it follows by
integration by parts from \eqref{our}, cannot
be removed (it is related to the relative sign of the kinetic terms
for the chiral and twisted chiral fields).
Similar signs appear often in discussions of duality, and
are usually harmless. In the case at hand, however, the sign
multiplies a logarithm, and its effect is to lead, at least
superficially, to a twisted superpotential polynomial in ${\widehat\Phi}^{-1}$
rather than ${\widehat\Phi}$ after performing the ``mirror'' argument.
This would seem to shatter the hopes for an understanding of mirror
symmetry for these models along the lines presented here. However,
the model we present does seem to incorporate naturally many of the
features required for a derivation of the conjectures in section four.
It is possible that some modification of this model will lead to the
desired result. Future study will tell if this is indeed the case.
\vspace*{0.6cm}
\noindent{\normalsize\bf Acknowledgements}
\vspace*{0.4cm}
We thank P. Candelas and X. de la Ossa for many discussions and for
collaboration in the early stages of this work, and
P. Argyres, P. Aspinwall, B. Greene, G. Moore, M. Ro\v cek,
and especially E. Witten and N. Seiberg for discussions.
The work of D.R.M.\ was supported in part by the United States Army
Research Office through the Mathematical Sciences Institute of
Cornell University, contract DAAL03-91-C-0027,
and in part by the National Science Foundation through
grants DMS-9401447 and PHY-9258582.
The work of M.R.P.\ was supported by National Science Foundation grant
PHY-9245317 and by the W.M. Keck Foundation.
\vspace*{0.6cm}
\noindent{\normalsize\bf References}
|
\section{INTRODUCTION}
Integrable models in two space-time dimensions provide a fascinating
arena for investigating non-perturbative phenomena in quantum field
theory \cite{Raj,Col,AAR}.
The integrable theories that we shall focus on here can be divided into
three broad classes, but they all have in common the properties of
asymptotic freedom and dynamical mass generation
which make
them akin to realistic models of particle interactions in four dimensions.
First, there are bosonic sigma models
based on a symmetric space $G/H$.
Any model of this sort is classically integrable, but unfortunately
anomalies can destroy integrability at the quantum level unless
$H$ is simple.
Second, there are fermionic
models of Gross-Neveu or Thirring type with four-fermion interactions.
Finally,
there are hybrids of the two previous classes in which fermions are
added to bosonic symmetric space models in various ways.
In some cases the addition of fermions can cancel the
anomalies encountered in the purely bosonic theory,
leading to interesting new examples of quantum integrable theories;
in particular this
is believed to occur for supersymmetric sigma models based on
a symmetric space $G/H$.
A review of these matters with detailed references can be found in
\cite{AAR}.
Exact S-matrices have been proposed for
many of the two-dimensional models which are thought to be integrable at the
quantum level \cite{AAR,ZZ,SMGN,SMCGN,SMPCM,SW,SMCPN}. These S-matrices
describe the scattering of some conjectured set of particles states,
and they are postulated on the basis of the specific symmetries
of the model in question together with the usual axioms of
S-matrix theory and the powerful constraint of factorization.
(There may also be additional information available about the theory, such as
the existence of bound-states.) However, such S-matrices
are always subject to CDD ambiguities \cite{CDD,ZZ} which cannot be further
constrained by these general considerations. What is needed is
some completely non-perturbative way of fixing the CDD ambiguities
by testing the proposed equivalence between an S-matrix
on the one hand and renormalized Lagrangian perturbation theory
on the other.
In this paper we shall explain such a programme
which can be applied to any of the types of theory mentioned above.
The technique, which was pioneered in \cite{HMN,HN,FNW} following earlier
work in \cite{TBA,PW,W,JNW,ZTBA}, can be applied to any integrable
model possessing
a group $G$ of global symmetries. The particle states in such a
theory fall into representations of $G$, and the symmetry is generated
by conserved charges in the Lie algebra.
The idea is to couple the theory to some particular conserved charge $Q$ by
modifying the Hamiltonian from $H$ to $H-hQ$, where $h$
is a coupling constant of mass dimension 1 (so that $Q$ is
dimensionless).
The corresponding change in the ground-state energy density
$\delta {\cal E}(h)={\cal E}(h)-{\cal E}(0)$ can be computed
straightforwardly in
perturbation theory. But in the special case of an integrable model,
it can also be computed by a very different
non-perturbative method starting from the exact S-matrix and using the
Thermodynamic Bethe Ansatz (TBA).
The comparison of these calculations provides a powerful check that
the S-matrix and the Lagrangian really do correspond to one another
and it also allows one to extract an {\it exact\/} expression for the mass
gap of the model. We now describe more quantitatively how this
comparison works.
The result of a perturbative calculation of the ground-state energy
density---which
is of course a renormalization group-invariant quantity---is an
expansion
\[{
\delta {\cal E}(h) =
h^2 \sum_{j=0}^\infty {\alpha}_{j} \, g(h/\Lambda)^{j-1}
}\]
in ascending powers of the running coupling $g(h / \Lambda)$
where the $\alpha_j$ are dimensionless numbers.
The running coupling can be found by integrating the usual beta-function
equation
\[
\mu \, {d g \over d \mu} = \beta(g)=
- \sum_{j=1}^\infty \beta_j \, g^{j+1}
\quad {\hbox{\rm to~obtain}} \quad
{1 \over g(\mu / \Lambda)}
= \beta_1 \ln {\mu \over \Lambda}
+ {\beta_2 \over \beta_1} \ln \ln {\mu \over \Lambda}
+ {\cal O} {\hbox{\Large (}} \ln \ln {\mu \over \Lambda} {\hbox{\Large /}}
\ln {\mu \over \Lambda} {\hbox{\Large )}} \, ,
\]
where the absence of a constant term in the solution defines the mass
scale $\Lambda$ for the particular renormalization
scheme used.
In the models we are considering, $\beta_1 > 0$ and we expect
perturbation theory to be valid in the asymptotic regime where
$\mu\gg\Lambda$.
Combining the expressions above we find that the
ground-state energy density is given by:
\begin{equation}
{\delta {\cal E}(h) \over h^2} = \alpha_{0}\beta_1\ln{h\over\Lambda}
+\alpha_{0}{\beta_2\over\beta_1}
\ln\ln{h\over\Lambda}+\alpha_1
+ {\cal O} {\hbox{\Large (}} \ln \ln {h \over \Lambda} {\hbox{\Large /}}
\ln {h \over \Lambda} {\hbox{\Large )}} \, .\label{eq;E1}
\end{equation}
Notice that if $\alpha_0 \neq 0$, there is a {\it classical\/} or
{\it tree-level\/} contribution to the ground-state energy density
and this quantity is consequently unbounded as $h$ becomes large.
Otherwise, $\alpha_0 = 0$ and the leading contribution
to the ground-state energy
is then the constant term $\alpha_1$.
We shall have more to say shortly about the important differences
between these situations.
Turning now to the TBA calculation: in order to extract the ground-state
energy from the S-matrix, one must confront a set of
coupled integral equations of Wiener-Hopf type.
The special circumstances of interest to us are those in which the
temperature is zero, with the coupling to the charge $Q$ acting like a
chemical potential.
In general the resulting equations cannot be solved exactly, but
it is possible, at least under certain simplifying assumptions, to
generate an expansion of the solution in $h/m$ valid when $h{\gg}m$,
where $m$ is some physical mass parameter occurring in the S-matrix.
The result is
\begin{equation}
{\delta {\cal E}(h) \over h^2}
= \kappa_0\ln{h\over m}+\kappa_1\ln\ln{h\over m}+\kappa_2
+ {\cal O} {\hbox{\Large (}} \ln \ln {h \over m} {\hbox{\Large /}}
\ln {h \over m} {\hbox{\Large )}} \, ,\label{eq;E2}
\end{equation}
where the $\kappa_j$ are dimensionless numbers.
The exact values of the parameters $\kappa_j$ depend sensitively on the
analytic structure of the S-matrix and, in particular, one finds that
the presence or absence of CDD factors alters dramatically the result for the
ground-state energy obtained in this way.
Equality of the expressions (1) and (2) gives a powerful
check that the Lagrangian and S-matrix descriptions are consistent.
The greater the accuracy to which we can calculate these
expressions, the more conclusive this check will be.
Even finding just the first few terms, however, can
provide strong evidence in favour of the proposed S-matrix used in the TBA
calculation, allowing us to argue
that any alteration by CDD factors would destroy the delicate
agreement with the perturbative result.
In addition, the consistency of (1) and (2)
clearly determines the mass gap $m / \Lambda$, at least if we have
calculated the expressions involved to sufficient accuracy.
At this stage it is useful to look more closely at the two possible
cases to which we drew attention earlier.
If $\alpha_0 \neq 0$, we see that the TBA calculation must
reproduce the first two coefficients of the beta-function through the
conditions:
\begin{equation}
\alpha_0 \beta_1 = \kappa_0 \ , \qquad
\alpha_0 \beta_2 / \beta_1 = \kappa_1 \ ,
\end{equation}
providing a highly non-trivial test of the S-matrix.
In these circumstances we can also read off the value of the mass gap in the
model:
\begin{equation}
\ln (m/\Lambda) = (\kappa_2 - \alpha_1)/\kappa_0 = (\kappa_2 -
\alpha_1) / \alpha_0 \beta_1 \, .
\end{equation}
We emphasize that these relations are deduced by comparing the terms
written {\it explicitly\/} in (1)
and (2) and they therefore rely on just a one-loop perturbative
calculation of the ground-state energy.
If, instead, $\alpha_0 = 0$, we must have
$\kappa_0 = \kappa_1 = 0$ and $\kappa_2 = \alpha_1$ for consistency
and we are then
unable to find the mass-gap by taking the expressions to the order
given explicitly in (1) and (2). To extract $m/\Lambda$ in this
situation one needs to extend both the TBA and perturbative
calculations to higher orders; in fact it is not hard to see
that this would involve at least a three-loop perturbative
calculation.
The strategy we have just outlined
has been applied to several series of integrable models, with results we shall
summarize in section 3.
The expressions for the exact mass-gaps derived in this way are
very useful, because they provide
bench-marks for the reliability of other non-perturbative approaches,
such as lattice simulations.
To minimize the amount of work involved, it is clearly desirable
to try to choose $Q$ so as to produce a classical term in the perturbative
expression for the ground-state energy (1) because, as we have
explained, the mass-gap can then be
found from the TBA analysis in conjunction with a one-loop
perturbative calculation.
It can be shown \cite{EH3} that such a choice of $Q$ is possible for any
bosonic or supersymmetric sigma model based on a symmetric space
$G/H$, a result which unifies the treatment of various examples
considered previously in \cite{HMN,HN,BNNW,THIII,EH1,EH2}.
For the purely fermionic Gross-Neveu models, however,
it seems that generically the classical contribution vanishes, and so
a three-loop calculation is necessary in order to find the mass
gap \cite{FNW,CGN}.
There is another crucial consideration to be borne in mind when
choosing $Q$.
In each of the cases considered so far,
an important simplifying assumption was made regarding the TBA
calculation, namely, that for very particular choices of the charge $Q$,
only a small number of particles---in fact those with the largest charge/mass
ratio---contribute to the new ground-state.
The original one-particle states can be chosen to be eigenvectors of
the new Hamiltonian $H-hQ$ with
eigenvalues $m_i\cosh\theta_i-hq_i$, where $q_i$ is the charge of the
particle labelled by $i^{\rm}$ and $\theta_i$ is its rapidity.
If $i=1$ labels the particle type with the largest charge/mass ratio,
then for sufficiently small values of $h$ such that
$hq_1<m_1$ it is not
energetically favourable to find any particles in the new ground-state, and
so $\delta {\cal E}(h)=0$. As
$h$ passes the threshold value $m_1/q_1$, it suddenly becomes favourable to
fill the original vacuum with particles. For large $h$, it would seem that any
arbitrary particle could appear in the new ground-state and hence the TBA
calculation would have to keep track of all particles. However, for
very particular choices of the charge $Q$ it seems that the particles with
largest charge/mass ratio actually repel other particles and so only
they appear in the ground state. This assumption
greatly simplifies the solution of the TBA
equations and in most of the existing papers it is taken as a working
hypothesis which is vindicated by the consistency of the final
results.
With more care, it can actually be proven from an analysis of the
full TBA equations \cite{EH3}.
Having introduced the general idea behind the technique, we shall
discuss in the next section how the TBA equations can be derived
under the simplifying assumption explained above.
We shall then summarize the known results, concluding with
a more detailed example which illustrates the most important points.
\section{THE TBA EQUATIONS}
In the cases where a single particle-type contributes to the ground state,
the TBA analysis is rather straightforward. The elastic scattering of two
particles of the same species is described
by an S-matrix element which is just a phase $S(\theta)$,
where $\theta$ is the rapidity difference of the
incoming particles. Since particle number and momenta are conserved on
interaction, it makes sense to consider single particle states. In the
dilute regime, where the particles are on average well separated, the
wavefunction of $N$ particles is built up from these single particle
states and has the form
\[
\Psi(x_1,\ldots,x_N)=\sum_{Q\in S_N}\Theta(x_Q)\zeta(Q)
\exp im (x_1\sinh\theta_1 + \ldots + x_N\sinh\theta_N )
\]
where the sum is taken over all permutations $Q=\{Q_1,\ldots,Q_N\}$ of
$\{1,\ldots,N\}$ with
\[
\Theta(x_Q)=\left\{\begin{array}{ll}1&{\rm if}\
x_{Q_1}<x_{Q_2}<\cdots<x_{Q_N}\\ 0&{\rm otherwise}\end{array}\right.
\]
and where $\zeta(Q)$ are numbers defined via the S-matrix so that
if $Q$ differs from $Q'$ only by an exchange of two specific
elements of the list
$Q_i$ and $Q_j$, say, then
\[
\zeta(Q')=S(\theta_i-\theta_j)\zeta(Q) \ .
\]
This means that the $\zeta(Q)$ are determined up to an overall factor.
We impose periodic boundary conditions on the wavefunctions:
\[
\Psi(x_1,\ldots,x_j,\ldots,x_N)=\Psi(x_1,\ldots,x_j{+}L,\ldots,x_N)\
\]
which imply that, for each fixed $j$,
\[
\exp({imL\sinh\theta_j})\prod_{k\neq j}S(\theta_j-\theta_k)=1\ .
\]
Taking the logarithm of these equations we have
\begin{equation}
mL\sinh\theta_j-i\sum_{k\neq j}\ln S(\theta_j-\theta_k)=2\pi n_j
\end{equation}
where the $\{n_j\}$ are a set of $N$ integers which we assume
determine uniquely the set of rapidities
$\{\theta_j\}$
(this is guaranteed in the dilute regime where
the $\sinh\theta_j$ dominates in (5)).
It is crucial that in all the integrable models we are considering here
the particles behave as fermions in rapidity space.
The condition (5) then proves to be the key ingredient when we come to
analyse the thermodynamics of such a system.
The thermodynamic limit can be taken in the
usual way by letting $N\rightarrow\infty$ and $L\rightarrow\infty$
with $N/L$ fixed. It now becomes meaningful to talk about
the density of one-particle states in rapidity space, $\varrho
(\theta)$, defined by $dn=L\varrho(\theta)d\theta$, and
also the density of {\it occupied\/}
states in rapidity space, $\sigma(\theta)$.
It is also convenient at this stage to define from the S-matrix the quantities
$K(\theta)$ and $R(\theta)$ given by
\begin{equation}
K(\theta) = \delta(\theta) - R(\theta)=
{1\over2\pi i}{d\over d\theta}\ln S(\theta) \, .
\end{equation}
By differentiating the key condition (5) we deduce
\begin{equation}
\varrho(\theta)={m\over2\pi}\cosh\theta+K*\sigma(\theta)
\end{equation}
(where $f*g(\theta)=\int_{-\infty}^\infty d \theta'
f(\theta') g(\theta-\theta')$ for any two functions $f$ and $g$).
When $K(\theta)=0$, (7) reduces to the usual expression for
the density-of-states in a free theory. But when there is an
interaction between the particles, (7) implies that
the total density and the density of occupied states are coupled
in a complicated manner.
Of course for overall consistency we must have
$\sigma(\theta)\leq\varrho(\theta)$, which is guaranteed in the dilute
regime since $\sigma(\theta)$ is then small compared with $\varrho(\theta)$.
For a macrostate specified by the density $\sigma(\theta)$, the
value of $H-hQ$ per unit length is
\[
{\cal E}(\sigma;h)=
\int_{-\infty}^\infty {d\theta\over2\pi}(m\cosh\theta-h)\sigma(\theta)
\, .
\]
(where we assume the charge is normalized to unity on our
preferred particle states).
To find the ground-state energy density we must minimize ${\cal E}(\sigma;h)$
with respect to $\sigma(\theta)$ subject to the constraint
(7). The solution to this
variational problem involves filling all the available states of
lowest rapidity first, so $\sigma(\theta)=\varrho(\theta)$ up to some
Fermi rapidity $\pm\theta_{\rm F}$.
It can be shown that this problem can be solved
in terms of an energy density
$\epsilon(\theta)$ which satisfies the TBA
equation at zero temperature:
\begin{equation}
\epsilon^+(\theta)+R*\epsilon^-(\theta)=m\cosh\theta-h
\end{equation}
with the notation
\[
f^\pm(\theta)=\left\{\begin{array}{ll}
f(\theta)&f(\theta){>\atop<}0\\ 0&{\rm otherwise}.\\\end{array}\right.
\]
The solution to (8) is concave and negative between the values
$\pm\theta_{\rm F}$, which are fixed by the condition
$\epsilon(\pm\theta_{\rm F})=0$.
The final result for the way in which the ground-state energy density
behaves as a function of
$h$ is then given simply by
\begin{equation}
\delta {\cal E}(h)={m\over2\pi}\int_{-\theta_{\rm F}}^{\theta_{\rm F}}d\theta
\, \cosh \theta \, \epsilon(\theta) \ .
\end{equation}
The problem of finding the ground-state energy density has thus been
reduced to solving the Wiener-Hopf integral equation (8).
It is not possible to find the
solution exactly for arbitrary $h$. But we require the solution only in the
asymptotic regime with $h\gg m$ and it is explained in \cite {FNW}
and \cite{BNNW} how to develop
such an asymptotic expansion for the solution based on the original
approach of \cite{JNW}.
The nature of the expansion depends
crucially on whether or not the Fourier transform $\hat R(\omega)$
of the kernel $R(\theta)$ defined via
\[
R(\theta)=\int_0^\infty{d\omega\over\pi}\cos(\omega \theta)\hat
R(\omega) \, ,
\]
vanishes at the origin. Suppose that $\hat R(0)=0$, and that
we can decompose
$1/ \hat R (\omega) = G_+(\omega) G_-(\omega)$ where $G_{\pm}(\omega)$
are analytic in the upper/lower half planes with $G_+(\omega) =
G_- (-\omega)$. It can be shown that if $G_+(i\xi)$ has an
expansion for small $\xi$ like
\begin{equation}
G_+(i\xi)={k\over\sqrt\xi}\exp(-a\xi\ln\xi)\left(1-b\xi+{\cal
O}(\xi^2)\right)
\end{equation}
then the ground-state energy density for $h\gg m$ takes the form
(2) with
\begin{equation}
\kappa_0 = -k^2/4 \ , \quad \kappa_1/\kappa_0 = a + 1/2 \ , \quad
{\kappa_2 / \kappa_0 = \ln (\sqrt{2\pi}/G_+(i)) - 1}
+\ln k +a(\gamma_{\rm E}-1+\ln8)-b \ .
\end{equation}
This matches the expansion in perturbation theory
{\it with\/} a classical term in (1). If $\hat R(0)\neq0$, on the
other hand,
then the asymptotic expansion proceeds slightly
differently and the result is an expression of the form (2) with
$\kappa_0 = \kappa_1 =0$.
This matches the perturbative result (1) in the case
when there is no tree-level contribution.
At this point we should remark that in the
regime $h\gg m$ considered above, the Fermi rapidity is large and so
the ground-state will contain a large number of particles. It
would seem therefore that the system is far from the dilute situation
used in our naive derivation of the TBA equations. It is one of the
miracles of the TBA that it seems, nevertheless, to be valid for many
systems in the deep ultra-violet regime, although the exact
reason for this is not understood.
The simple TBA analysis which we have presented also
assumed that there is only one particle type
that contributes to the new ground-state. In general many particles with
different quantum numbers and masses can contribute and the TBA
analysis then becomes much more complicated.
If the scattering of the
particles is purely elastic, the one-particle
analysis can be generalized in
an obvious way: there is a function $\epsilon_a(\theta)$ for each
particle type with mass $m_a$, and the TBA system (8) becomes a matrix
equation involving
\[
K_{ab}(\theta)=
{1\over2\pi i}{d\over d\theta}\ln S_{ab}(\theta),
\]
where $S_{ab}(\theta)$ is the S-matrix element between particles $a$
and $b$. If the scattering is not elastic, there are
still greater complications, with the net
result being that the TBA system involves
additional ``magnon'' degrees of freedom which
behave like particles with zero mass.
Nevertheless the analysis can sometimes be carried out successfully in
this situation too, as we shall see later for a particular
example.
\section{SUMMARY OF RESULTS}
We now summarize the results obtained for several families of
integrable models.
In each case we shall define the theory by a Lagrangian $\cal L$
displaying a global symmetry group $G$ or
$G{\times}G$, where $G$ = SU($n$), SO($n$) or Sp($n$). To express the
results compactly, we introduce the quantity
$1/\Delta$ = $n$, $n{-}2$ or $2n{+}2$ respectively
(this is the dual Coxeter number for SU($n$) and SO($n$) but
{\it twice\/} the dual Coxeter number for Sp($n$)).
The fields in the Lagrangian
transform in the defining representation of the symmetry group unless
we state otherwise. The mass-gap $m/\Lambda$ will be
given for particles which also belong to
the defining representation of the symmetry group. The renormalization
scheme is $\overline{\rm MS}$ except in one case.
{\bf (i)} O($n$) sigma model \cite{HN,HMN}:
\[
{\cal L}={1\over2g}\, \partial_\mu \phi_a\partial^\mu \phi_a
\]
where $\phi_a$ is an $n$-component real scalar field
obeying $\phi_a\phi_a=1$.
The TBA calculation confirms the S-matrix proposed in \cite{ZZ},
correctly predicting
$\beta_1 = 1/ 2\pi \Delta$ and $\beta_2 = 1/ 4\pi^2 \Delta$.
The mass gap
\[
m={(8 /e )^\Delta
\over\Gamma(1+\Delta)} \,
\Lambda_{\overline{\rm MS}}\
\]
is consistent with the $1/n$ expansion \cite{BCR}.
{\bf (ii)} $G{\times}G$ principal chiral model \cite{BNNW,THIII}:
\[
{\cal L}={1\over g} \, {\rm Tr}\left(\partial_\mu U
\partial^\mu U^{-1}\right)
\]
where $U$ is a $G$-valued field.
The TBA calculation with the S-matrices proposed in \cite{SMPCM}
correctly predicts $\beta_1 = 1/ 16 \pi x \Delta$, $\beta_2 = \beta_1^2 /2$
where $x$ is the Dynkin index of the defining representation of $G$
($x=1/2$ for SU($n$), Sp($n$); $x=1$ for SO($n$)).
The mass gap is
\[
m=
2^{(d \Delta + 1 /2)} \, {\sin(\pi\Delta)\over \sqrt{\pi e}\Delta} \,
\Lambda_{\overline{\rm MS}}
\]
where $d$ is the dimension of the defining representation
of $G$ ($d=n$ for SU($n$), SO($n$); $d=2n$ for Sp($n$)).
The mass gap has been measured on the lattice for $G$ = SU($3$) and the
agreement with the theoretical value is quite accurate \cite{HAS}.
{\bf (iii)} O($n$) Gross-Neveu model $(n > 4)$ \cite{FNW}:
\[
{\cal L}={i\over2}\bar\psi_a\gamma^\mu\partial_\mu\psi_a+{g\over8}
\left(\bar\psi_a\psi_a\right)^2
\]
where $\psi_a$ is an $n$-component Majorana spinor.
The S-matrices proposed in \cite{ZZ,SMGN} are found to be consistent with the
TBA/perturbation calculation which correctly predicts
$\beta_1 = 1/2\pi \Delta$ and $\beta_2 = -1/ 4\pi^2 \Delta$.
The mass gap is
\[
m= { \, \, (2e)^\Delta \over\Gamma(1- \Delta)} \,
\Lambda_{\overline{\rm MS}}
\]
which is consistent with the $1/n$ expansion \cite{FNWII}.
Note that our coupling constant $g$ and mass-scale
$\Lambda_{\overline{\rm MS}}$ differ
slightly from those used in \cite{FNW,FNWII}.
{\bf (iv)} SU($n$) chiral Gross-Neveu model \cite{CGN}:
\[
{\cal L}={i\over2}\bar\psi_a\gamma^\mu\partial_\mu\psi_a+{g\over8}
\left \{
\left(\bar\psi_a\psi_a\right)^2 -
\left(\bar\psi_a\gamma_5\psi_a\right)^2
\right \}
\]
where $\psi_a$ is an $n$-component complex spinor.
The S-matrix for the massive sector conjectured in \cite{SMCGN}
is consistent with the
TBA/perturbation calculation which correctly predicts
$\beta_1 = 1/ \pi \Delta$ and $\beta_2 = -1 /\pi^2 \Delta$.
The mass gap
\[
m= {\, \, (e/4)^{\Delta /2} \over \Gamma(1- \Delta)} \,
\Lambda
\]
agrees with the $1/n$ expansion \cite{CGN}.
See \cite{CGN} for full details of the renormalization scheme
and note also our slightly different definitions of
$g$ and $\Lambda$.
{\bf (v)} O($n$) supersymmetric sigma model $(n>4)$ \cite{EH1}:
\[
{\cal L}={1\over2g}\left\{(\partial_\mu \phi_a)^2
+i\bar\psi_a\gamma^\mu\partial_\mu\psi_a
+ {1\over4}\left(\bar\psi_a
\psi_a\right)^2\right \}
\]
where the fields $\phi_a$ and $\psi_a$ are an $n$-component real scalar
and spinor satisfying the constraints $\phi_a\phi_a=1$ and
$\phi_a\psi_a=0$. The model is invariant under $N=1$ supersymmetry
transformations mixing the bosons and fermions.
The S-matrix
conjectured in \cite{SW} is consistent with the TBA/perturbation
calculation which correctly predicts $\beta_1 = 1/2 \pi \Delta$ and
$\beta_2 = 0$. The mass gap
\[
m=2^{2 \Delta} \, {\sin ( \pi\Delta) \over \pi \Delta} \,
\Lambda_{\overline{\rm MS}}
\]
is consistent with the $1/n$ expansion \cite{JAG}.
{\bf (vi)} SU($n$) or ${\bf C}P^{n-1}$ supersymmetric sigma model \cite{EH2}:
\[
{\cal L}={1\over2g} \left \{ \, \vphantom{ {x \over x} }
\left | (\partial_\mu{-}A_\mu) z_a \right |^2
+ i \bar \psi_a \gamma^\mu (\partial_\mu{-}A_\mu) \psi_a
+ \, {1 \over 4} \left[ \left
(\bar \psi_a \psi_a\right)^2 - \left(\bar \psi_a \gamma_5 \psi_a\right)^2
- \left(\bar \psi_a \gamma_\mu \psi_a\right )^2\right]\right\}
\]
where $A_\mu = \frac{1}{2}( z^*_a \partial_\mu z_a - z_a \partial_\mu
z^*_a)$ and
the fields $z_a$ and $\psi_a$ are an $n$-component complex scalar and
Dirac spinor satisfying the constraints
$z_a^* z_a^{\phantom{*}} = 1$ and $z_a^* \psi_a^{\phantom{*}} = 0$.
The Lagrangian has a local U(1)
invariance under which all complex fields transform by a phase and
$A_\mu$ transforms as a gauge field. The model is also invariant under
$N=2$ supersymmetry.
The S-matrix proposed in \cite{SMCPN} correctly predicts
$\beta_1 = 1/ \pi \Delta$ and $\beta_2 = 0$. The formula for the mass-gap is
\[
m= {\sin ( \pi \Delta ) \over \pi \Delta} \,
\Lambda_{\overline{\rm MS}}
\]
which is consistent with the $1/n$ expansion \cite{MC} and with other
non-perturbative approaches specific to $N=2$ supersymmetric
theories \cite{CV}.
\section{DETAILED EXAMPLE: THE SUPERSYMMETRIC ${\bf C}P^{n-1}$ MODEL}
Most of the features of the general method we have outlined here
are nicely illustrated by the last example, the supersymmetric
${\bf C}P^{n-1}$ model \cite{INST,DDL2}, so we now discuss this case in more detail.
To introduce the S-matrix, we assume that in the quantum theory there
exist states $\vert a,i,\theta\rangle$ representing fundamental
particles and states $\vert\bar a,i,\theta\rangle$ representing
fundamental antiparticles. Here $\theta$ is rapidity;
$i=0,1$ distinguishs ``bosons'' and ``fermions'' (actually these particles
carry fractional statistics \cite{SMCPN,FI});
and indices $a$ and $\bar a$ label the
$n$ and $\bar n$ representations of $\SU$.
The fundamental anti-particles can be regarded as
bound-states of the fundamental particles, or vice-versa---an example
of ``nuclear democracy''.
There are additional bound states transforming in all the
antisymmetric representations of SU($n$).
The integrability of the model \cite{AAG} implies that
the S-matrix factorizes and that all S-matrix elements can be deduced from
the two-body ones; furthermore, the S-matrix for any
desired set of particles can be obtained from the S-matrix for the
fundamental particles.
Thus the entire S-matrix is specified by the amplitude
\begin{equation}
\langle c,k,\theta_2;d,l,\theta_1,{\rm out}\vert
a,i,\theta_1;b,j,\theta_2,{\rm in}\rangle
=S_{N=2}(\theta_1-\theta_2)_{ij}^{kl}
S_{\rm CGN}(\theta_1-\theta_2)_{ab}^{cd}
\end{equation}
which was first proposed by K\"oberle and Kurak
\cite{SMCPN}.
Following \cite{Schou,FI}, we have written this proposal in a factorized form
in which $S_{\rm CGN}$ is the S-matrix for the chiral Gross-Neveu model
\cite{SMCGN}, which specifies the scattering of the $\SU$ degrees of
freedom, and $S_{N=2}$ controls the $N=2$ supersymmetric degrees of
freedom, as described in \cite{FI} (although we have chosen to include the
physical strip pole in the CGN part of the
S-matrix). Explicit expressions for these factors are given in
\cite{EH2}.
The conjectured S-matrix is minimal in
the sense that it has
the minimum number of poles and zeros on the physical strip (the
region $0\leq{\rm Im}(\theta)\leq\pi$) consistent with the
requirements of symmetry, the existence of a bound-state
and the axioms of S-matrix theory. But this still leaves open
the possibility of adding CDD factors to the S-matrix \cite{CDD,ZZ};
these spoil none of the axioms, they introduce no new poles on the
physical strip and they passively respect the bootstrap equations.
For our model the CDD ambiguities correspond to multiplying the S-matrix of
the fundamental particles by factors of the form
\begin{equation}{
\sinh\left({\theta\over2}-{i\pi\over2 n}\alpha\right)
\sinh\left({\theta\over2}-{i\pi\over2 n}(2-\alpha)\right)
\over
\sinh\left({\theta\over2}+{i\pi\over2 n}\alpha\right)
\sinh\left({\theta\over2}+{i\pi\over2 n}(2-\alpha)\right)
} \ ,
\end{equation}
where $0<\alpha<2$.
One of the conclusions of our analysis will be that the minimal
form is the true S-matrix of the theory, so that all CDD
factors are ruled out.
We have emphasized that in following the general method for testing
the equivalence between an S-matrix and a Lagrangian, there
are at least two important points to be borne in mind when choosing
the coupling to a conserved charge $Q$.
First, in order to simplify the analysis of the TBA system we must
choose $Q$ so
that the new ground-state consists of a restricted number of particle
types. In the models considered in \cite{HMN,HN,FNW,CGN,BNNW,THIII} it was
possible to find a generator $Q$ such that the new ground-state
contained a single particle type. In the present theory, however,
we know that the lowest energy states must come in degenerate
supersymmetric multiplets, since supersymmetry commutes with the $\SU$
invariance.
We might be tempted
to make the same choice for $Q$ as in the $\SU$ principle
chiral model \cite{BNNW} and the chiral Gross-Neveu model
\cite{CGN}:
\begin{equation}
Q={\rm diag}\left(1,-{1\over n-1},\ldots,-{1\over n-1}\right),
\end{equation}
for which there will be a single fundamental doublet
$\vert1,j,\theta\rangle$ with the largest
charge/mass ratio.
But it turns out that this violates our second criterion regarding the
choice of $Q$, which is that there should be a classical or tree-level
term in the ground-state energy so that we can extract the mass-gap via a
one-loop, rather than a higher-loop, calculation.
We are therefore motivated to consider an
alternative choice:
\begin{equation}
Q={\rm diag}\left(1,-1,0,\ldots,0\right).
\end{equation}
This does indeed lead to the desired tree-level term, but the TBA
analysis is complicated by the fact that
there are now two fundamental doublets with the largest
charge/mass ratio,
namely
$\vert1,j,\theta\rangle$ and $\vert\bar2,j,\theta\rangle$,
Despite this complication, the calculation proves tractable with this
choice, as we shall describe below.
\subsection{Perturbation theory calculation}
The coupling of the theory to the charge (15) by a change in the Hamiltonian
$H \rightarrow H - h Q$ can be achieved by making a corresponding
replacement $\partial_0 \rightarrow
\partial_0 + i h Q$ in the Lagrangian.
Since we are interested in performing a one-loop calculation of the
change in the ground-state energy density as a function of $h$,
it is enough to expand the resulting Lagrangian to
quadratic order in an independent
set of fields, and we can drop all terms which are independent of $h$ to this
order.
By exploiting the local U(1) invariance of the
Lagrangian, we can take $z_1$ to be real and we can solve the bosonic
constraint $z_a z_a^* =1$ by writing
$z_1 = \{ (1 - \vert \pi \vert^2 )(\frac{1}{2} + \phi) \}^{1/2}$
and
$z_2 = e^{i \theta} \{ (1 - \vert \pi \vert^2)(\frac{1}{2} - \phi) \}^{1/2}$
where $\pi = (z_3 , \ldots , z_n)$ and $\theta$, $\phi$ are real.
The fermionic degrees of freedom and
the variable $\theta$ decouple to quadratic order and we are left with the
expression
\[
{\cal L}_{{\rm 1-loop}} =
{1 \over 2 g} \left \{ (\partial_\mu \phi)^2
+ \vert \partial_\mu \pi \vert^2 + h^2 - 4 h^2 \phi^2 - h^2 \vert \pi \vert^2
\right\}.
\]
We see that there is indeed a tree-level term, as desired.
Using standard dimensional regularization with the $\overline{\rm
MS}$-scheme gives
a one-loop expression for the ground-state energy
\[
\delta {\cal E}(h) = -{h^2 \over 2 g} - {h^2 \over \pi} \ln 2
+ {h^2 n \over 4 \pi} \left [ 1 - \ln {h^2 \over \mu^2} \right ].
\]
It is important that when we substitute for the running
coupling with the known values of the beta-function
coefficients
\begin{equation}
\beta_1 = n / \pi \ , \quad \beta_2 = 0 \ ,
\end{equation}
the $\mu$-dependence must cancel (to leading order) since the quantity we
are computing is renormalization group-invariant.
{}From above, the values of the other dimensionless numbers appearing in
(1) are, for this model,
\begin{equation}
\alpha_0 = -1/ 2 \ , \quad
\alpha_1 = (n/4 \pi) - (1/\pi) \ln2 \ .
\end{equation}
The explicit expression for the ground-state energy is thus,
to the required order,
\begin{equation}
\delta {\cal E}(h)=-{h^2n\over2\pi}\left[ \, \ln{h\over
\Lambda_{\overline{\rm MS}}}-{1\over2}+{2 \over n} \ln 2 \, \right].
\end{equation}
\subsection{TBA calculation}
We must now consider the TBA equations for the model and
solve them in the limit $h\gg m$.
Following the hypothesis introduced earlier, we assume
that only the multiplets
$\vert1,j,\theta\rangle$ and $\vert\bar2,j,\theta\rangle$ contribute to the
ground-state. Since the scattering of these multiplets is purely
elastic, it is not necessary to perform a diagonalization in the
space of $\SU$ quantum numbers (although this diagonalization can be
done \cite{THII}). The remaining difficulty is that the S-matrix
for these favoured states is still
non-diagonal in the supersymmetric subspace. Fortunately, this problem can
be solved---in fact it has been shown by
Fendley and Intriligator \cite{FI} that it is equivalent to
diagonalizing the transfer matrix of the six vertex model at the free fermion
point.
The TBA equations involve energy densities $\epsilon_1(\theta)$ and
$\epsilon_{\bar2}(\theta)$ for the two supermultiplets and two magnon
energy densities $\xi_0(\theta)$ and $\xi_{\bar 0}(\theta)$ which reflect
the non-elastic nature of the scattering amongst the supersymmetric degrees
of freedom (two magnons because we are dealing with $N=2$
supersymmetry).
The resulting TBA equations are, at $T=0$,
\begin{eqnarray}
\epsilon_a(\theta)-\phi_{ab}*\epsilon^-_b(\theta)-\phi_{al}*\xi^-_l(\theta)
\!\!\!&=&\!\!\!m\cosh\theta-h,\nonumber\\
\xi_l(\theta)-\phi_{al}*\epsilon^-_a(\theta)
\!\!\!&=&\!\!\!0,
\end{eqnarray}
where $a=1,\bar2$ and $l=0,\bar0$. In terms of these the ground-state
energy density is
\begin{equation}
\delta {\cal E}(h)={m\over2\pi}\int_{-\infty}^\infty d\theta
\left[\epsilon_1^-(\theta)+\epsilon_{\bar2}^-(\theta)\right]\cosh\theta.
\end{equation}
The kernels appearing in (19) are
\begin{eqnarray*}
\phi_{ab}(\theta)\!\!\!
&=&\!\!\!{1\over2\pi i}{d\over d\theta}
\ln S_{\rm CGN}(\theta)_{ab}^{ba},\\
\phi_{10}(\theta)\!\!\!
&=&\!\!\!\phi_{\bar2\bar0}(\theta)=
{1\over2\pi}{\sin(\pi/n)\over\cosh\theta-\cos(\pi/n)},\\
\phi_{\bar20}(\theta)\!\!\!
&=&\!\!\!\phi_{1\bar0}(\theta)=
{1\over2\pi}{\sin(\pi/n)\over\cosh\theta+\cos(\pi/n)},
\end{eqnarray*}
where $S_{\rm CGN}(\theta)_{ab}^{ba}$ are the appropriate S-matrix
elements of the chiral Gross-Neveu model.
To simplify (19) it is important to notice that $\phi_{al}(\theta)$ is a
positive kernel, which implies that the magnon variables are given by
$\xi^+_l(\theta)=0$ and $\xi^-_l(t)=\phi_{al}*\epsilon^-_a(\theta)$.
Furthermore, the solution does
not distinguish between the values of the favoured SU($n$) quantum numbers
and so we have $\epsilon_1(\theta)=\epsilon_{\bar2}(\theta)\equiv
\epsilon(\theta)$. The four equations in (19) then reduce
to a single equation for $\epsilon(\theta)$ of the form (8)
with kernel
\[
R(\theta)=\delta(\theta)-\phi_{11}(\theta)-\phi_{1\bar2}(\theta)
-\left[\phi_{10}+\phi_{1\bar0}
\right]*\left[\phi_{10}+\phi_{1\bar0}\right](\theta) \ ,
\]
and, from (20), each of the densities $\epsilon_1$ and $\epsilon_{\bar
2}$ contributes an amount to the ground-state energy given by (9).
We have therefore succeeded in reducing the
problem to the relatively simple case which we already know how to
handle.
Following the subsequent steps outlined in section 2, we find the
Fourier transform of the kernel:
\[
\hat R(\omega)
={\cosh(({1\over2}{-}{1\over n})\pi\omega)\sinh({1 \over n}\pi\omega)
\over\cosh^2(\frac{1}{2}\pi\omega)} \, \exp{\textstyle \frac{1}{2}\pi\omega} \ .
\]
This vanishes at the origin and we can decompose it as
$1/(G_+(\omega)G_-(\omega))$ where $G_\pm(\omega)$ are analytic
in the upper/lower half planes and $G_-(\omega)=G_+(-\omega)$.
The unique solution is
\begin{eqnarray*}
G_+(\omega)={\Gamma(\frac{1}{2}-(\frac{1}{2}{-}{1 \over n})i\omega)
\Gamma(1-{1 \over n}i\omega)\over
\Gamma^2(\frac{1}{2}-\frac{1}{2} i\omega)}
\exp \left \{ \, \textstyle \frac{1}{2} \ln n
-\frac{1}{2}(1{+}i\omega) \ln(-i\omega) \right. \\
\textstyle
\left. + \, i \omega \left [ \, \ln 2 + \frac{1}{2}
+ (\frac{1}{2}{-}{1\over n}) \ln (\frac{1}{2}{-}{1\over n})+{1\over n}\ln{1\over n}
\, \right ] \, \right \}
\end{eqnarray*}
and we can indeed find an expansion of
$G_+(i\xi)$ for small $\xi$ of the form (10)
with
\[
k=\sqrt{n/\pi}\ , \quad
a=-1/2\ , \quad
b= {\textstyle \frac{1}{2} (1{-}\gamma_{\rm E})- {1\over n} \ln 4n
+ ( \frac{1}{2}{-}\frac{1}{n} )
\ln ( {1\over 2}{-}{1\over n} ) } \ .
\]
Substituting in (11) and adding the contributions for the
densities $\epsilon_1$ and $\epsilon_{\bar 2}$ produces
\begin{equation}
\kappa_0 = -n/ 2 \pi \, , \quad \kappa_1 = 0 \, , \quad
\kappa_2 / \kappa_0 = \ln (n / \pi) + \ln \sin (\pi / n)
+ (2 \ln 2)/n - 1/2 \, ,
\end{equation}
or explicitly, to the required order,
\begin{equation}
\delta {\cal E}(h)=-{h^2n\over2\pi}\left[ \, \ln{h\over
m} - \frac{1}{2} + {2 \over n} \ln 2
+\ln\left({n\over \pi}\right) +\ln
\sin \left({\pi \over n}\right)
\right] \ .
\end{equation}
\subsection{Comparison of calculations}
By substituting the values (16), (17) and (21) into the
general relations (3) and (4)---or simply by comparing the explicit
final results (18) and (22)---we see that the TBA
calculation correctly reproduces the universal coefficients of the
beta-function and that it predicts the value of the mass-gap for
the supersymmetric ${\bf C}P^{n-1}$ model to be
$ m /\Lambda_{\overline{\rm MS}}=(n/\pi)\sin(\pi/n)$, as claimed.
It is instructive to consider what would have happened if we had
carried out our analysis using the other charge (14), rather than
(15). In that case only the multiplet
$\vert1,j,\theta\rangle$ appears in the ground-state and the
resulting TBA equations are simpler in as much as they
involve only this single doublet, rather than two doublets.
The system can be reduced to a single integral equation in a similar
way, but on doing so we find a different kernel:
\[
R(\theta)=\delta(\theta)-\phi_{11}(\theta)
-\phi_{10}*\phi_{10}(\theta)-
\phi_{1\bar0}*\phi_{1\bar0}(\theta) \ .
\]
The Fourier transform of this kernel does not vanish at the origin,
and so, as explained previously, we would need to go beyond one-loop
perturbation theory to carry out a non-trivial test of the S-matrix.
This matches precisely the
fact that with this different choice of charge the perturbative
expansion of the ground-state energy density is also markedly different with
no tree-level contribution.
It is also interesting to see for this particular example
how the calculation resolves
the problem of CDD ambiguities in the S-matrix.
An additional CDD factor of the form (13)
would alter the kernel appearing in the TBA equation from $\hat
R(\omega)$ to
\[
\hat R(\omega)-2{\cosh(({1\over2}{-}{1\over n})\pi\omega)
\cosh({1\over n}(\alpha-1)\pi\omega)\over
\cosh(\frac{1}{2}\pi\omega)} \ .
\]
But this expression fails to vanish at the origin and so the argreement with
the perturbative result is destroyed.
|
\section{Introduction}
The phenomenology of neutralino dark matter has been studied extensively
in the Minimal Supersymmetric extension of the Standard Model (MSSM)
\cite{Susy}.
This model incorporates the same gauge group as the Standard Model
and the supersymmetric extension of its particle content. The
Higgs sector is slightly modified as compared to that of the Standard
Model: the MSSM requires
two Higgs doublets $H_1$ and $H_2$ in order to give mass both to down-- and
up--type quarks and to cancel anomalies. After Electro--Weak Symmetry
Breaking (EWSB),
the physical Higgs fields consist of two
charged particles and three neutral ones: two scalar fields ($h$ and $H$) and
one pseudoscalar ($A$). The Higgs sector is specified at the tree level by
two independent parameters:
the mass of one of the physical Higgs fields and the ratio of the two vacuum
expectation values, usually defined as $\tan\beta=v_2/v_1 \equiv
<H_2> / <H_1>$.
The supersymmetric sector of the model introduces some other free parameters:
the mass parameters $M_1$, $M_2$ and $M_3$ for the supersymmetric partners
of gauge fields (gauginos), the Higgs--mixing parameter $\mu$ and, in general,
all the masses of the scalar partners of the fermions (sfermions).
In the MSSM it is generally assumed that
the gaugino masses are equal at the grand unification scale $M_{GUT}$:
$M_i(M_{GUT})\equiv m_{1/2}$ and hence are related at lower scales by
\begin{equation}
M_1 : M_2 : M_3 = \alpha_1 : \alpha_2 : \alpha_3
\label{eq:GUTgaugino}
\end{equation}
\noindent
where the $\alpha_i$ (i=1,2,3) are the coupling constants of the three
Standard Model gauge groups.
The neutralinos are mass--eigenstate linear superpositions of the
two neutral gauginos ($\tilde \gamma$ and $\tilde Z$) and the two
neutral higgsinos ($\tilde H_1$ and $\tilde H_2$)
\begin{equation}
\chi = a_1 \tilde \gamma + a_2 \tilde Z + a_3 \tilde H_1 + a_4 \tilde H_2\;.
\end{equation}
\noindent
The neutralino sector depends, at the tree--level, on the following
(low--energy)
parameters: $M_1= (5/3) \tan^2 \theta_W M_2$, $M_2 \simeq 0.8\, m_{1/2}$,
$\mu$ and
$\tan\beta$. Neutralino properties are naturally discussed in the
($m_{1/2}$, $\mu$) plane, for a fixed value of $\tan\beta$. As an
example, in Fig.1 the lines of constant mass for the lightest neutralino
($m_{\chi}$)
and constant gaugino fractional weight ($P \equiv a_1^2 + a_2^2$) are plotted
in the ($m_{1/2}$, $\mu$) plane for $\tan\beta=8$.
We observe that the mass of the lightest neutralino increases
from the bottom left to the top right, while the
neutralino composition changes from higgsino dominance in the top--left
region of the plane to gaugino dominance in the bottom--right.
The regions forbidden by accelerator data are also displayed in Fig.1.
The low--energy MSSM scheme is a purely phenomenological approach, whose
basic idea is
to impose as few model--dependent restrictions as possible.
In this approach the lightest neutralino is a favourite candidate for
cold dark matter. This scheme
has been employed extensively in the analysis of
the size and the relevance of
various possible signals of neutralino dark matter:
direct detection \cite{thdir,elflor,direct},
signals due to neutralino annihilation
in celestial bodies, namely the Earth and the Sun \cite{bodies,ap3},
and signals from neutralino
annihilation in the galactic halo \cite{halo}.
The MSSM provides a useful framework
in which neutralino phenomenology may be analysed without strong
theoretical prejudices which could, {\it a posteriori}, turn out to be
incorrect. This scheme is also frequently
employed in analyses of the discovery potential of future
accelerators \cite{Baer}.
At a more fundamental level, it is natural to implement this
phenomenological scheme within the supergravity framework
\cite{sugra,bbo,diehl}. One
attractive feature of the ensuing model is the connection between
soft supersymmetry breaking and EWSB, which would then be induced radiatively.
The essential elements of the model are described by
a Yang--Mills Lagrangian, the
superpotential, which contains all the Yukawa interactions
between the standard and supersymmetric fields, and by the soft--breaking
Lagrangian, which models the breaking of supersymmetry.
Here we only recall the soft supersymmetry breaking terms
\begin{eqnarray}
&-{\cal L}_{soft}& =
\displaystyle \sum_i m_i^2 |\phi_i|^2
\no \\
&+& \left\{\left[
A^{l}_{ab} h_{ab}^{l} \tilde L_a H_1 \tilde R_b +
A^{d}_{ab} h_{ab}^{d} \tilde Q_a H_1 \tilde D_b +
A^{u}_{ab} h_{ab}^{u} \tilde Q_a H_2 \tilde U_b +\mb{h.c.} \right] -
B \mu H_1 H_2 + \mb{h.c.} \right\}
\no \\
&+& \displaystyle \sum_i M_i
(\lambda_i \lambda_i + \bar\lambda_i \bar\lambda_i)
\label{eq:soft}
\end{eqnarray}
\noindent
where the $\phi_i$ are the scalar fields, the $\lambda_i$ are the
gaugino fields, $H_1$ and $H_2$ are the two Higgs fields,
$\tilde Q$ and $\tilde L$
are the doublet squark and slepton fields, respectively,
and $\tilde U$, $\tilde D$ and
$\tilde R$ denote the $SU(2)$--singlet fields for the up--squarks,
down--squarks and sleptons. In Eq.(\ref{eq:soft}), $m_i$ and $M_i$ are the mass
parameters of the scalar and gaugino fields, respectively, and $A$ and
$B$ denote trilinear and bilinear supersymmetry breaking parameters,
respectively. The
Yukawa interactions are described by the parameters $h$, which
are related to the masses of the standard fermions by the usual
expressions, {\em e.g.}, $m_t = h^t v_2$.
The supergravity framework is usually implemented with a
number of restrictive assumptions about unification at $M_{GUT}$:
i) Unification of the gaugino masses:
$M_i(M_{GUT}) \equiv m_{1/2}$,
ii) Universality of the scalar masses with a common mass denoted by
$m_0$: $m_i(M_{GUT})$ \hfill \break
\indent \phantom{ii)\ } $ \equiv m_0$,
iii) Universality of the trilinear scalar couplings:
$A^{l}(M_{GUT}) = A^{d}(M_{GUT}) = A^{u}(M_{GUT})$ \hfill \break
\indent \phantom{iii)\ } $\equiv A_0 m_0$. \hfill\break
\noindent
These conditions have strong consequences for low--energy supersymmetry
phenomenology, and in particular for the properties of the neutralino
as dark matter particle.
Typically,
the lightest neutralino is constrained to regions
of gaugino dominance, that entail a large relic abundance (in wide
regions of the parameter space $\Omega_{\chi} h^2$ exceeds the
cosmological upper bound) and a small direct detection rate for
neutralino dark matter.
Indirect signals from the neutralino, such as high--energy neutrinos
from the Earth and Sun, and the products of annihilation in the halo,
are practically undetectable \cite{diehl}.
The above assumptions, particularly ii) and iii), are not very solid,
since universality may occur at a scale higher
than $M_{GUT}$, {\em i.e.}, the Planck scale or string scale \cite{comm},
in which case renormalization above $M_{GUT}$ may weak universality in the
$m_i$, {\em e.g.}, between scalars in $\underline{\bar{5}}$ and
$\underline{10}$ representations of $SU(5)$ \cite{ehnt}. Moreover,
in many
string models the $m_i$'s are not universal even at the string scale.
In a number of recent works \cite{OlPok,others}, deviations
from some of the
unification conditions have been considered. In particular, in
Ref.\cite{OlPok} phenomenological consequences for neutralinos of a
relaxation
of assumption ii) have been analysed in the regime of large values of
tan $\beta$. It has been shown that deviations from condition ii) may
entail a changeover in neutralino composition from a gaugino--like state
to a higgsino--like state (or at least to a higgsino--gaugino mixed state),
with important consequences for neutralino phenomenology.
In this paper, we first explore,
over the full range of tan $\beta$,
the various scenarios which may occur when condition ii) is relaxed,
with an approach which is similar to the one adopted in the
large--$\tb$ analysis of Ref.\cite{OlPok}. We then
discuss in detail the ensuing consequences for neutralino dark
matter, with particular emphasis for its direct detection.
In the following, we first discuss
which constraints can be applied to the parameters
when specific physical requirements are imposed. In Sect.II, we
summarize the conditions implied by radiative EWSB and define the type
of departure from universality examined in this paper.
Then, in Sect.III we establish some upper bounds on the supergravity parameters
by requiring that radiative
EWSB does not occur with excessive fine tuning among different terms.
In Sect.IV we analyse in detail the
constraints due to the requirement that EWSB takes place radiatively.
Subsequently, in Sect.V
cosmological constraints, derived from the evaluation of the neutralino
relic abundance, are discussed. Other
constraints, from experimental data on $b \rightarrow s \gamma $ processes
and on the mass of the bottom quark $m_b$,
are applied in Sect.VI. In Sect.VII the
effects of these various constraints
are first displayed in the ($m_{1/2}$, $m_0$) plane for fixed $\tb$ and
$A_0$, and then shown in the ($m_{1/2}$, $\mu$) plane, which provides the
most useful representation for discussing neutralino
phenomenology. We recall some specific properties of the neutral Higgs
bosons in Sect.VIII. Finally, in Sect.IX event rates for direct detection
of neutralino dark matter are discussed. Conclusions are presented in
the last Section.
\section{Radiative EWSB}
We recall that the tree--level scalar potential for the neutral Higgs fields
may be written in the form
\begin{equation}
V_0 = (M_{H_1}^2+\mu^2) |H_1|^2 + (M_{H_2}^2+\mu^2) |H_2|^2 -B\mu (H_1 H_2 +
\mb{h.c.}) + \mb{quartic D terms.}
\label{eq:higgspot}
\end{equation}
\noindent
The parameters of this potential must obey the following physical conditions:
\begin{equation}
\sin 2 \beta = \frac {-2B\mu} {M_{H_1}^2+M_{H_2}^2+2 \mu^2}
\label{eq:s2beta}
\end{equation}
\begin{equation}
M_Z^2 = 2 \frac {M_{H_1}^2-M_{H_2}^2 \tan^2 \beta}
{\tan^2 \beta -1} - 2 \mu^2
\label{eq:mz}
\end{equation}
\begin{equation}
M_A^2 = M_{H_1}^2+M_{H_2}^2+2 \mu^2 > 0~.
\label{eq:pinco}
\end{equation}
\noindent Here $M_A$ is the mass of the CP--odd neutral Higgs boson (see
Sect.VIII below), and
eq.(\ref{eq:pinco}) must in fact be strengthened to
$M_A \geq (M_A)_{lb}$, where $(M_A)_{lb}$ is
the experimental lower bound \cite{LEP}.
For instance, for $\tb \gsim 3$, $(M_A)_{lb} \simeq 45$ GeV.
Notice that the sign of $\mu$ is defined according to the convention of
reference \cite{Susy}.
We remark that although
Eqs.(\ref{eq:higgspot}--\ref{eq:pinco}) are expressed
at the tree level,
in our actual calculations 1--loop corrections
to $V_0$ \cite{oneloop} have been included.
The $M_{H_i}$'s (as well as the sfermion and the gaugino masses and
the parameters {\it A, B} and $\mu$) evolve from
the $M_{GUT}$ scale down to the $M_Z$ scale according to the
Renormalization Group Equations (RGE's).
This is how Eq.(\ref{eq:mz}) may be satisfied, even
if $M_{H_1}$ and $M_{H_2}$ are equal at $M_{GUT}$.
In this work we
consider deviations from universality
in the scalar masses at $M_{GUT}$, which split $M_{H_1}$ from
$M_{H_2}$. This effect is parameterized as
\begin{equation}
M^2_{H_i}(M_{GUT}) = m_0^2(1+\delta_i)~.
\label{eq:nonuniv}
\end{equation}
\noindent
The parameters $\delta_i$ which quantify the departure from universality for
the $M^2_{H_i}$ will be varied in the range ($-1$,$+1$), but are taken to be
independent of the supersymmetry parameters. This is an
{\it Ansatz}, since, when evolving the scalar masses from the unification scale
(Planck scale or string scale) to the GUT scale $M_{GUT}$, the
deviation parameters are in general functions of all the supersymmetry
parameters \cite{mark}.
Following a common procedure, Eq.(\ref{eq:s2beta}) is used to replace the
parameter $B$ by $\tan \beta$. Thus the set of independent parameters
becomes
$m_{1/2}$, $m_0$, $A_0$, $\tan \beta$, and $\mu^2$ is given in terms of these
parameters by Eq.(\ref{eq:mz}), suitably corrected by 1--loop effects:
only the sign of $\mu$ remains
undetermined. Obviously, values of $\mu^2$ are accepted only if
they exceed the experimental
lower bound $\mu_{lb}^2$, which is derived from the lower limit
on the chargino mass \cite{LEP}: $|\mu_{lb}| \simeq 45$ GeV.
We have solved the RGE's using the 1--loop beta functions
including the whole supersymmetric particle spectrum
from the GUT scale down to $M_Z$,
neglecting the possible effects of intermediate thresholds.
Two--loop and threshold effects on the running of the gauge and Yukawa
couplings are known not to exceed 10\% of the final result \cite{lang}.
While this is
of crucial importance as far as gauge coupling unification is
concerned \cite{lang}, it is a second--order
effect on the evolution of the soft masses.
Since neutralino properties are studied
over a wide range of variation for the high scale parameters, such a degree of
refinement is not required here.
In order to specify the supersymmetry phenomenology,
boundary conditions for the gauge and Yukawa couplings have to be
specified.
Low--scale values for the gauge couplings and for the top--quark and the
tau--lepton Yukawa couplings are fixed using present experimental results.
In particular, we assign for the top mass the value $m_t=178$ GeV \cite{CDF}.
In addition, we require the unification of the bottom and tau Yukawa couplings
at the GUT scale, as would be suggested by a unifying group
that includes an $SU(5)$--like structure \cite{chano}.
The values of $M_{H_1}$ and $M_{H_2}$ at the $M_Z$ scale,
obtained from the RGE's, may be parameterized in the following way:
\begin{equation}
M^2_{H_i} = a_i m_{1/2}^2 +b_i m_0^2 + c_i A_0^2 m_0^2+
d_i A_0 m_0 m_{1/2}~.
\label{eq:polinom}
\end{equation}
\noindent
(Notice that, in our notation, all running quantities written without
any further specification are meant to denote their values at $M_Z$.)
The coefficients
in the expression (\ref{eq:polinom}) are functions of tan $\beta$
and of the $\delta_i$'s. They are displayed in Fig.2 (a,b) for the case
of universal scalar masses, ({\em i.e.}, $\delta_i = 0$).
The coefficients for $M_{H_2}^2$ turn out to be very stable as
functions
of $\tb$, except for small $\tb$. More precisely, $a_2 \sim -2.5$ for $\tb
\gsim
4$ with all the other coefficients much smaller (of order 0.1). As far as
$M_{H_1}^2$ is concerned, whereas $c_1$ and $d_1$ are again very
stable
(of order 0.1), $a_1$ and $b_1$ vary rapidly as functions of $\tb$.
This property of $a_1$ and $b_1$ is due to the very fast increase of
$h^b$ for increasing $\tb$.
When a departure from $m_0$ universality is introduced, the coefficients in
Eq.(\ref{eq:polinom}), except for $a_1$ and $a_2$,
become functions of the parameters $\delta_i$:
$b_1$, $c_1$ and $d_1$ depend on $\delta_1$ and $b_2$, $c_2$ and $d_2$ on
$\delta_2$. Whereas the $b_i$'s are rapidly--increasing functions of the
$\delta_i$'s, the $c_i$'s and the $d_i$'s are rather insensitive to these
parameters.
Stringent constraints on
the parameters $m_{1/2}, m_0, A_0$ and $\tan \beta$ follow from the
request that
the ${M^2_{H_i}}$'s, evaluated from Eq.(\ref{eq:polinom}),
satisfy Eqs.(\ref{eq:mz}--\ref{eq:pinco}). Explicitly, we require that
$\mu^2$ and $M_A^2$, given by the expressions
\begin{eqnarray}
\mu^2 &=& \frac {1} { \tan^2 \beta -1} \{(a_1 - a_2 \tan^2 \beta) m_{1/2}^2+
(b_1 - b_2 \tan^2 \beta) m_0^2 + \nonumber \\
& & (c_1 - c_2 \tan^2 \beta) A_0^2 m_0^2 + (d_1 -
d_2 \tan^2 \beta) A_0 m_0 m_{1/2}\}- \frac {M_Z^2} {2} \nonumber \\
&\equiv& J_1 m_{1/2}^2 + J_2 m_0^2 + J_3 A_0^2 m_0^2 + J_4 A_0 m_0
m_{1/2} - \frac {M_Z^2} {2}
\label{eq:mu}
\end{eqnarray}
\begin{eqnarray}
M_A^2 &=& (a_1 + a_2 +2 J_1) m_{1/2}^2+
(b_1 + b_2 +2 J_2 ) m_0^2 + \nonumber \\
& & (c_1 + c_2 +2 J_3) A_0^2 m_0^2 + (d_1 +
d_2 +2 J_4) A_0 m_0 m_{1/2} - M_Z^2 \nonumber \\
&\equiv& K_1 m_{1/2}^2 + K_2 m_0^2 + K_3 A_0^2 m_0^2 +
K_4 A_0 m_0 m_{1/2} - M_Z^2
\label{eq:M_A}
\end{eqnarray}
\noindent satisfy the conditions:
$\mu^2 \geq \mu_{lb}^2$, $M_A \geq (M_A)_{lb}$ mentioned earlier.
The coefficients $J_i$ and $K_i$ in
Eqs.(\ref{eq:mu},\ref{eq:M_A}) are plotted as functions of $\tb$
in Fig.2 (c,d) for the case of $m_0$ universality.
In Fig.2c we notice that all the $J_i$'s are positive, with $J_1$ dominating
over the others: for $\tb \gsim 4$, one has
$J_1 \simeq 2.4$.
As far as the coefficients $K_i$ are concerned, we see in Fig.2d that
only two of them are sizeable: $K_1$ and $K_2$. They are
both decreasing functions of $\tb$, with $K_1 > K_2$. At very large
$\tb$ these coefficients become
very small, and $K_2$ even becomes negative (but still small in
magnitude) at $\tb \gsim 50$.
In the case of non--universality, the coefficients $J_i$ and $K_i$, except
for $J_1$ and $K_1$, become functions of the parameters $\delta_i$.
We will see in Sect.IV that many important features of the supersymmetry
parameter space depend on the signs of the two coefficients $J_2$ and
$K_2$. We show in Figs.3 and 4 how their signs depend on the
values of the $\delta_i$'s. In Fig.3 the lines $J_2 = 0$
are plotted in the ($\delta_2$, $\delta_1$) plane for a few values of $\tb$:
for each value of $\tb$,
$J_2$ is negative in the region above the relevant $J_2 = 0$ line and
positive below.
Similarly, in Fig.4 the $K_2 = 0$ lines are displayed in the same
($\delta_2$, $\delta_1$) plane at fixed $\tb$: $K_2$ is negative above the
$K_2 = 0$ lines, and positive below.
We now make a few comments related to Eq.(\ref{eq:M_A}), since the
value of $M_A$ plays a very crucial role in a number of important
neutralino properties. This is due to the fact that many physical
processes involving neutralinos are mediated by the neutral
Higgs bosons. Thus the
value of $M_A$ determines the size of the relevant cross sections both
through $M_A^2$--dependence in propagators and, in an implicit way,
through the couplings of the $h$ and $H$ bosons to quarks and to
the lightest neutralino $\chi$ (see Sect.VIII).
As a consequence, a small value of $M_A$ has the effect of enhancing
the magnitude of the relevant cross sections.
What values of $M_A$ do we obtain from Eq.(\ref{eq:M_A})?
Because of the properties of the coefficients
$K_i$ previously analysed, $M_A$ turns out to
be a rapidly--decreasing function of $\tb$. In Fig.5, $M_A$ is
displayed at the representative point $m_0 = 50$ GeV, $m_{1/2} =
200$ GeV (1--loop corrections to $M_A$ have been included in the
calculation). One notices that $M_A$ is $O(M_Z)$ for
$\tb \gsim 45$. This feature provides one of the most appealing
scenarios for neutralino phenomenology.
\section{Constraints due to the absence of fine tuning}
Before we exploit fully the two constraints
$\mu^2 \geq \mu_{lb}^2$, $M_A \geq (M_A)_{lb}$
to restrict the parameter space, we apply the general criterion that
the expression (\ref{eq:mu}) is satisfied without excessive tuning among the
various terms \cite{nft,bbo}. In radiative EWSB the physical value of
$M_Z$, which sets the EW scale, may be written as
\begin{eqnarray}
M_Z^2 =
2( J_1 m_{1/2}^2 + J_2 m_0^2 + J_3 A_0^2 m_0^2 + J_4 A_0 m_0
m_{1/2} - \mu^2)~~.
\label{eq:mz2}
\end{eqnarray}
Accidental compensation (fine tuning)
among different terms in Eq.(\ref{eq:mz2}) may occur.
We explicitly require the absence of too--strong fine tuning, {\em i.e.},
cancellations
among exceedingly large values of the parameters $m_{1/2}$, $m_0$, $A_0$
and $\mu$. Denoting by $\eta_f$ a parameter which quantifies the
degree of fine tuning, we require \cite{nft} that
\begin{eqnarray}
\left |\frac {\Delta M^2_Z} {M^2_Z}\right| & < &
\eta_f \left |\frac {\Delta x^2_i}
{x^2_i}\right |
\label{eq:nft}
\end{eqnarray}
\noindent
where $x_i$ denotes any of the previous parameters. For instance, for
$A_0 = 0$, Eq.(\ref{eq:nft}) provides the following conditions
\begin{eqnarray}
m_{1/2}^2 < \frac {\eta_f} {2 |J_1|} M^2_Z, ~~
m^2_0 < \frac {\eta_f} {2 |J_2|} M^2_Z, ~~
\mu^2 < \frac {\eta_f} {2} M^2_Z \simeq (640\mb{ GeV})^2
\label{eq:pippo}
\end{eqnarray}
\noindent
where in the last approximate equality we have taken $\eta_f = 100$, which
means that we allow an accidental compensation at the 1\% level.
The upper bound on $m_0$ depends on $\tb$ and the $\delta_i$'s, whereas that
on $m_{1/2}$ varies only with $\tb$ (because of the nature of the {\it
Ansatz} (\ref{eq:nonuniv}): see the comment after Eq.(\ref{eq:nonuniv})).
For the sake of illustration, we give some numerical examples,
taking again $\eta_f = 100$.
For $\tb = 8$, we have, for $\delta_1 = \delta_2 = 0$,
$m_{1/2} \lsim 400$ GeV, $m_0 \lsim 1.5$ TeV. For two other pairs of
values of the $\delta_i$'s, which will be discussed later on, we
obtain $m_0 \lsim 2.4$ TeV for $\delta_1 = -0.2$, $\delta_2 = 0.4$ and
$m_0 \lsim 3.0$ TeV for $\delta_1 = -0.8$, $\delta_2 = 0.2$.
At $\tb = 53$ we have $m_{1/2} \lsim 415$ GeV and
$m_{0} \lsim (1.7-1.9)$ TeV, depending on the values for the $\delta_i$'s.
These inequalities imply for the neutralino
mass $m_\chi \lsim 170$ GeV.
In the following, when graphical representations for the parameter
space are shown, we display no--fine--tuning upper bounds
obtained from the general expression (\ref{eq:nft})
with $\eta_f = 100$. These upper bounds are denoted by dashed
lines in Figs.9--14.
\section{Constraints due to radiative EWSB}
The EWSB constraints are given by the set of
Eqs.(\ref{eq:s2beta}--\ref{eq:pinco}), or equivalently
by Eqs.(\ref{eq:mu}--\ref{eq:M_A}),
together with the conditions
$\mu^2 \geq \mu_{lb}^2$ and $M_A \geq (M_A)_{lb}$.
From these equations the values of
$m_0$ and $m_{1/2}$ (or $\mu$ and $m_{1/2}$) are constrained and thus some
domains in the ($m_{1/2}$, $m_0$) or ($m_{1/2}$, $\mu$) planes can be excluded.
Let us start this discussion by analyzing the condition
$M_A \geq (M_A)_{lb}$, with $M_A$ given by Eq.(\ref{eq:M_A}).
For the sake of simplicity, we put
$A_0 = 0$ for the moment. To discuss the role of $M_A \geq (M_A)_{lb}$ in
placing bounds on $m_{1/2}$ and $m_0$, we first rewrite it explicitly as
\begin{eqnarray}
K_1 m_{1/2}^2 + K_2 m_0^2 \geq M_Z^2 + (M_A)_{lb}^2~.
\label{eq:con1}
\end{eqnarray}
\noindent
The nature of this quadratic form in the ($m_{1/2}$, $m_0$) plane obviously
depends on the signs of the two coefficients $K_1$ and $K_2$.
As we have seen in Sect.II, it turns out that, whereas $K_1$ is always
positive, the sign of $K_2$ depends on the values of $\tb$ and of the
$\delta_i$'s.
Two different situations may occur, depending on the sign of $K_2$.
In the case $K_2 > 0$ the region allowed by (\ref{eq:con1}) is the
one above an elliptical
branch centered in the origin of the ($m_{1/2}$, $m_0$) plane. Therefore,
both parameters $m_{1/2}$ and $m_0$ are bounded from below.
When $K_2 < 0$, the region allowed by Eq.(\ref{eq:con1}) is the one
between the
$m_{1/2}$ axis and an upward--moving hyperbolic branch. Thus, whereas
$m_{1/2}$ is still bounded from below, $m_0$ is now constrained from
above. The upper bound on $m_0$ is particularly stringent when
$K_2$ is large and negative and $K_1$ is not large.
This occurs, for instance, at very large values of $\tb$ in the
case of $m_0$ universality.
This discussion may be extended straightforwardly to the case
$A_0 \neq 0$. In this case the constraint $M_A \geq (M_A)_{lb}$
may be written explicitly as
\begin{eqnarray}
K_1 m_{1/2}^2 + K_2 m_0^2 + K_3 A_0^2 m_0^2 +
K_4 A_0 m_0 m_{1/2} \geq M_Z^2 + (M_A)_{lb}^2~.
\label{eq:con2}
\end{eqnarray}
\noindent
The nature of this quadratic form depends on the sign of its
determinant. When this determinant is positive, an elliptical branch
in the ($m_{1/2}$, $m_0$) plane provides lower bounds on the two variables.
On the other hand,
a negative determinant entails an upward--moving hyperbolic branch
which places an upper bound on $m_0$. These branches are part of
conics whose axes are somewhat tilted with respect to the ($m_{1/2}$,
$m_0$) axes.
Similar implications follow from the constraint
$\mu^2 \geq \mu_{lb}^2$, which may be written explicitly as
(for $A_0 = 0$)
\begin{eqnarray}
J_1 m_{1/2}^2 + J_2 m_0^2 \geq \frac {M_Z^2} {2} + \mu_{lb}^2~.
\label{eq:con3}
\end{eqnarray}
\noindent
This quadratic form may be discussed in much the same way as the one
in Eq.(\ref{eq:con1}).
{}From the properties seen in Sect.II it turns out that the coefficient
$J_1$ is always positive,
whereas the coefficient $J_2$ is positive in the universal case,
but may be negative when deviations from $m_0$ universality are
introduced. Thus it follows that the condition $\mu^2 \geq \mu_{lb}^2$
puts lower bounds on $m_{1/2}$ and either lower or upper bounds on
$m_0$, depending on the sign of $J_2$ (due to analytic properties
identical to those discussed previously below Eq.(\ref{eq:con1})).
The condition $\mu^2 \geq \mu_{lb}^2$ sets a very stringent upper
bound on $m_0$, whenever $J_2$ is negative and large in
magnitude.
The extension to the case $A_0 \neq 0$ may be repeated here
in a way similar to the above discussion for Eq.(\ref{eq:con2}).
Thus we have seen that two important constraints,
$\mu^2 \geq \mu_{lb}^2$ and $M_A \geq (M_A)_{lb}$,
are at work in bounding
$m_{1/2}$ and $m_0$, when EWSB is required to occur radiatively.
When $J_2$ and $K_2$ are positive, the two conditions place lower
bounds on $m_{1/2}$ and $m_0$. Similar constraints are established
by the requirements that also the sfermion masses and $m_{\chi}$
satisfy the relevant experimental bounds. These last conditions are
not explicitly discussed here, but they are taken into account in our
evaluations.
It is worth emphasizing that the most dramatic impact of the
conditions
$\mu^2 \geq \mu_{lb}^2$ and $M_A \geq (M_A)_{lb}$
over the parameter space occurs when either $J_2$ or $K_2$ (or both of
them) are negative. Under these circumstances, as we have seen above,
$\mu^2 \geq \mu_{lb}^2$ and $M_A \geq (M_A)_{lb}$
may place stringent upper limits on $m_0$, bounding the neutralino
parameter space considerably. Which of the two conditions prevails
over the other depends on the specific regions of the full parameter
space and on the values of the $\delta_i$'s.
In Sect.VII we will illustrate the implications of these constraints
in a few specific examples.
\section{Cosmological constraint}
Let us turn now to the evaluation of the neutralino relic abundance
$\Omega_{\chi}h^2$ and to the requirement that the lightest neutralino is not
overproduced, {\em i.e.}, $\Omega_{\chi}h^2 \leq 1$.
The neutralino relic abundance $\Omega_{\chi} h^2$ is evaluated
following the standard procedure \cite{omega,omega_poles,omega1,ap1},
according to which $\Omega_{\chi} h^2$ is essentially given by
$\Omega_{\chi} h^2 \propto <\sigma_{\mbi{ann}} v>^{-1}_{\mbi{int}}$,
where $<\sigma_{\mbi{ann}} v>_{\mbi{int}}$ is the
thermally--averaged annihilation cross section, integrated from the
freeze--out temperature to the present temperature. The standard
expansion $<\sigma_{\mbi{ann}} v> = a + b x + ...$
may be employed, with $x=T/m_\chi$,
except at s--channel resonances ($Z,A,H,h$), where a more precise
treatment has to be used for the thermal average \cite{omega_poles}.
In the evaluation of
$<\sigma_{\mbi{ann}} v>$ the full set of annihilation final
states ($f \bar{f}$ pairs, gauge--boson pairs,
Higgs--boson pairs and Higgs--gauge boson pairs), as well as the
complete set of Born diagrams are taken into account
\cite{ap1}. We recall that one of the largest contributions to the annihilation
cross section is provided by diagrams with the exchange of the pseudoscalar
Higgs boson $A$. (More relevant properties of the Higgs bosons are discussed in
Section VIII.)
We note that the constraint $\Omega_{\chi} h^2 \leq 1$ is very effective
for small and intermediate values of $\tan \beta$, but is not restrictive for
large values of $\tan \beta$.
The strong restriction in the former case comes from the large value of $M_A$
implied by small and intermediate values of $\tan \beta$ (see
Fig.5) (also the couplings of $A$ to $\chi$ and fermions are small for these
values of $\tan \beta$).
We show in Figs.6--8 a few examples where $\Omega_{\chi} h^2$ is given
as a function of $m_{\chi}$ in the form of scatter plots. These
scatter plots have been
obtained by varying the parameters $m_0$ and $m_{1/2}$ on a
equally--spaced linear grid over the ranges $10~\mb{GeV} \leq m_0
\leq 2~\mb{TeV}$,
$45~\mb{GeV} \leq m_{1/2} \leq 500~\mb{GeV}$. Furthermore, we remark
that all evaluations presented in this paper are for positive values of
$\mu$, since negative values of $\mu$ are disfavoured by the
constraints due to $m_b$ and $b \rightarrow s \gamma$ processes
(see Sect.VI). The configurations shown in Figs.6--8 satisfy the
constraints due to radiative EWSB, discussed previously.
In Fig.6 is shown the case $\tb = 8$ and $\delta_i=0$.
Here, as expected
because of the intermediate value of $\tb$, many neutralino
configurations provide $\Omega_{\chi}h^2 > 1$,
whilst only a few give $\Omega_{\chi} h^2 \leq 1$.
(Also, $M_A$ is large here because of sizeable values of $K_2$
(see Fig.2d), which helps increase $\Omega_{\chi}h^2$.)
An exception occurs when
$m_\chi \simeq M_Z/2$, since in this case the annihilation cross section is
greatly enhanced due to the Z--pole contribution.
In Fig.7 we display $\Omega_{\chi} h^2$ in a case of
non--universality ($\delta_1 = -0.2, \delta_2 = 0.4$, for definiteness).
It is easier to find $\Omega_{\chi}h^2 \leq 1$
in this case, since here the departure from
$m_0$ universality implies a changeover of the neutralino composition
from the
gaugino dominance of the previous example to higgsino dominance
(this point will be elucidated in Sect.VII). This
implies a larger $\chi$--$\chi$ annihilation cross section and
consequently a smaller relic abundance. Thus only
a few neutralino configurations are excluded by the
$\Omega_{\chi} h^2 \leq 1$ condition.
An example for $\Omega_{\chi} h^2$ in the case of large $\tan \beta$
and $\delta_i = 0$
is shown in
Fig.8. We see that $\Omega_{\chi} h^2 \leq 1$ imposes no constraint
since, for this very large value of $\tb$, annihilation cross sections are
very large.
\section{Constraints from $\lowercase{b} \rightarrow \lowercase{s}
\gamma$ and $\lowercase{m_b}$}
In the evaluation of the $b \rightarrow s \gamma$ decay rate we
have included the supersymmetric
contributions arising from the charged Higgs loops and chargino loops given
in Ref.\cite{bsgamma1}. The Higgs term always adds to the
Standard Model value and usually entails too large a value for the rate. On the
other hand, the chargino contribution gives rise to a destructive
interference for $\mu > 0$ (in our convention for the sign of $\mu$).
At large $\tan\beta$ supersymmetric contributions may
be sizeable: unless the destructive interference protects the decay rate, it
can
very easily be driven out of the present experimental bounds. In the light
of this property, the positive $\mu$ scenario appears to be the favourite one
and, as already remarked, in this paper we only show results for this case.
In comparing our predictions with observations
we have taken into account that,
as discussed in Ref.\cite{bsgamma_uncert}, large theoretical uncertainties
are present, mainly due to QCD effects. In particular, predictions depend very
strongly on the choice of the renormalization scale, leading to an inaccuracy
of
order 25\%. To account for this effect we have relaxed the experimental bounds
of Ref.\cite{bsgamma_lim} by the same amount, keeping
the renormalization scale fixed
at the representative value of 5 GeV. Thus, our requirement is that the rate
of $b \rightarrow s \gamma$ decay falls into the range
$0.8 \times 10^{-4} \leq \mbox{BR}(b \rightarrow s \gamma) \leq 5.3 \times
10^{-4}~.$
The supersymmetric corrections to the bottom mass include
contributions from bottom--squark--gluino loops and from
top--squark--chargino loops \cite{carena}.
In the present analysis, the bottom mass is computed as a function of the other
parameters and required to be compatible with the present experimental bounds.
Theoretical uncertainties in the evaluations of $m_b$ arise both from the
running of the RGE's and from assumptions about Yukawa unification.
Since our choice
is to solve RGE's at the 1--loop level and without thresholds,
we estimate an uncertainty of the order of 10\% in our prediction for
$m_b$. In addition, a relatively small departure (see Ref.\cite{barger})
from bottom--$\tau$ Yukawa unification at the GUT scale may
significantly change the bottom mass result. To take into account such
uncertainties we have chosen to weaken the bounds on $m_b$ given in
\cite{Wright} by an amount of 10\%. Thus we require $m_b$
to fall into the range $2.7~ \mbox{GeV} \leq m_b(M_Z) \leq 3.4~ \mbox{GeV}$.
\section{Allowed regions in neutralino parameter space}
We discuss now in a few examples how the various constraints analysed
in the previous Sections complement each other in shaping the allowed
regions in the parameter space. We start with the ($m_{1/2}$, $m_0$)
representation, and later display our results in the
($m_{1/2}$, $\mu$) plane which provides the most useful representation for
neutralino phenomenology.
Let us first clarify a few graphical conventions adopted in our
($m_{1/2}$, $m_0$) and ($m_{1/2}$, $\mu$) plots.
Regions are left empty when at least one of the following constraints is
not satisfied: i) experimental bounds on Higgs, neutralino and sfermion masses
\cite{LEP,d0},
ii) the $\chi$ is the Lightest Supersymmetric Particle (LSP),
iii) radiative EWSB and $\mu^2 \geq \mu_{lb}^2$,
$M_A \geq (M_A)_{lb}$.
Regions forbidden by the cosmological constraint ($\Omega_{\chi} h^2 \leq
1$) are explicitly denoted by dots and those disallowed by the
$b \rightarrow s \gamma$, $m_b$ constraints
(but not by the previous ones) are denoted by crosses (crosses are displayed
only in the ($m_{1/2}$, $m_0$) plane, but not in the
($m_{1/2}$, $\mu$) plane, to simplify these plots).
The allowed domains are denoted by squares when they satisfy
$\Omega_{\chi} h^2 > 0.01$, or by diamonds otherwise in the ($m_{1/2}$, $m_0$)
plots. They are denoted by squares in the ($m_{1/2}$, $\mu$) plots,
independently of the $\Omega_\chi h^2$ value.
To simplify the discussion, we first take $A_0 = 0$.
We comment on the $A_0 \neq 0$ case at the end of this Section.
As a first example, let us consider the representative point $\tb =8$.
For this intermediate value of $\tb$,
the cosmological constraint is expected to be very effective in view of
the arguments discussed in Sect.V. This is actually the case for
universal $m_0$, when both $K_2$ and $J_2$ are positive
(see Fig.2), so that the conditions of radiative EWSB
do not set any upper limit on $m_0$ (Fig.9a).
The empty region in the lower part of these figures is forbidden by the
experimental bound on $m_{\chi}$.
As shown in this figure, in wide regions
(denoted by dots) $\Omega_{\chi} h^2 > 1$. Thus the cosmological constraint
places a very stringent upper bound on $m_0$ for
$m_{1/2} \gsim 150$ GeV. However, for smaller values of $m_{1/2}$, an
allowed horizontal region extends up to $m_0 \simeq 2$ TeV. In fact,
along this strip, $m_{\chi} \simeq M_Z/2$ and then
$\Omega_{\chi} h^2 \leq 1$
is satisfied (see the discussion in Sect.V).
Moving away from the universal point towards a region where $J_2$
is negative, we
expect $\mu^2 \geq \mu_{lb}^2$ to be effective in placing a
stringent upper bound on $m_0$. This is
actually the case in the example shown in Fig.10a, which refers to the
representative point
$\delta_1 = - 0.2$, $\delta_2 = 0.4$ ($J_2 = -0.07$). Here
it is the bound $\mu^2 \geq \mu_{lb}^2$
which provides the most stringent constraint in disallowing the
large (empty) domain on the
right side. Nevertheless,
$\Omega_{\chi} h^2 \leq 1$ is still effective in excluding an
internal region that would otherwise be allowed (see the discussion below).
Keeping $\tb = 8$, we complete our discussion by considering the representative
point
$\delta_1 = - 0.8,\delta_2 = 0.2$ shown in Fig.11a,
which gives an example where $J_2$ is very small.
The peculiarity of this example
will become clear when we discuss the relevant situation in the
($m_{1/2}$, $\mu$) plane, to which we now turn.
The shape and general properties of the physical region in the
($m_{1/2}$, $\mu$) plane are
dictated by the constraints previously derived, and they are determined
most notably by $J_2$.
It is convenient to distinguish the two cases i) $J_2 > 0$ and ii) $J_2 <0$.
For case (i)
at fixed $m_{1/2}$, $\mu$ increases for increasing $m_0$ with the
consequence that the allowed physical region extends to the right of the
$m_0=0$ line in the ($m_{1/2}$, $\mu$) plane, allowing for the neutralino
only a gaugino--dominated region.
In the case (ii) ($J_2 <0$), starting from the $m_0=0$ line and
increasing $m_0$ at fixed $m_{1/2}$,
one moves to the left and then one may reach regions of sizeable
higgsino--gaugino mixing or even of higgsino dominance.
Case i) applies in particular to the case of $m_0$ universality
($\delta_i = 0$) for any value of $\tb$. This is clear from Fig.3,
which shows that in the ($\delta_2$, $\delta_1$) plane the origin is below
any $J_2 = 0$ line. An example of this situation is displayed in
Fig.9b (for $\tb = 8$).
However, as we have seen in Sect.II, when the assumption of
$m_0$ universality is relaxed, then $J_2$, which in the universal
case is positive and small,
may very easily become negative and sizeable. In this case a changeover
in neutralino composition
from an originally gaugino--like state into a higgsino--like one occurs. This
remarkable property, discussed in Ref.\cite{OlPok} for
large $\tb$, is in fact valid over the whole range of $\tb$, if
the degree of non--universality is increased for decreasing $\tb$.
An example of case ii) ($J_2 < 0$) is shown in Fig.10b, where the
allowed region extends widely into the higgsino region.
It is instructive to compare Fig.9 with Fig.10. Looking at sections
a) of these figures, we notice that changing the values of the
$\delta_i$'s from the set $\delta_i = 0$ to the set
$\delta_1 = -0.2$, $\delta_2 = 0.4$ relaxes substantially
the cosmological constraint. Parts b) of these figures
provide the explanations for
this feature. In fact, whereas in the former case the neutralino is mainly a
gaugino,
in the latter case $\chi$ is higgsino--like or mixed. As we already
remarked, this
implies an increase of the $\chi$--$\chi$ annihilation cross section and a
reduction of the relic abundance.
The physical region also displays an extension to the right,
in the example of Fig.11b, but
here the effect is very tiny, due to a very small $J_2$ and to
the severe upper bound on $m_0$ for $m_{1/2} \gsim 180$ GeV.
This is the first case to show a very marked ($m_{1/2}$, $\mu$) correlation.
Now we turn to the case of
large $\tb$, where new features appear. First, the
$M_A \geq (M_A)_{lb}$ condition is no longer protected by large
values of $K_1$, and may become effective in restricting the parameter
space. Secondly,
the $m_b$ and $b \rightarrow s \gamma$ conditions are now rather
stringent over large domains and not only occasionally relevant as in the
smaller $\tb$ cases. Thirdly, the cosmological constraint is usually
overwhelmed by the other conditions. In Figs.12a, 13a, 14a we have,
for $\tb = 53$, the following sequence of examples.
i) $\delta_1 = 0, \delta_2 = 0$ (Fig.12a): here
$K_2 < 0$, $J_2 > 0$, and since $K_2$ is negative and sizeable in
magnitude, the constraint
$M_A \geq (M_A)_{lb}$ sets an extremely stringent upper bound on $m_0$
and thus forbids the wide (empty) region on the right.
ii) $\delta_1 = 0, \delta_2 = -0.2$ (Fig.13a): here one still has
$K_2 < 0$, $J_2 > 0$, but $|K_2|$ is smaller than in the previous case, so the
constraint $M_A \geq (M_A)_{lb}$ is still very effective but less compelling
than in the case (i). Also, the role of the $m_b$ and the
$b \rightarrow s \gamma$
conditions is more significant here.
iii) $\delta_1 = 0.7, \delta_2 = 0.4$ (Fig.14a): here
$K_2 > 0$, $J_2 < 0$,
$M_A \geq (M_A)_{lb}$ gives a lower
bound on $m_{1/2}$ and the $\mu^2 \geq \mu_{lb}^2$ condition
provides the frontier of the empty domain on the right.
The ($m_{1/2}$, $\mu$) representations for large $\tan \beta$ and
for the
representative $\delta_i$ points discussed
above are displayed in Figs.12b--14b.
We start from the universal case of Fig.12b. Here we expect gaugino--dominated
configurations. However, because the values of $m_0$ are strongly limited from
above (see Fig.12a), we have the extremely correlated states shown in Fig.12b.
In the
case of Fig.13b one has $J_2>0$, and gaugino--dominated states occur. No strong
($m_{1/2}$, $\mu$) correlation shows up in this case.
The opposite case, $J_2 < 0$, is shown in Fig.14b, where
higgsino--dominated configurations appear.
It is worth adding a few comments about the
examples of Figs.11 and 12, where
the physical regions in the ($m_{1/2}$, $\mu$) plane show a very
pronounced correlation in the two variables. This feature occurs whenever
$|J_2| m_0^2 \ll J_1 m_{1/2}^2$, {\em i.e.}, whenever $m_0$ is severely
bounded
from above and/or $|J_2|$ is very close to zero. As far as the values of
$|J_2|/J_1$ are concerned, we notice that in the universal case (see
Fig.2c), except for small values of $\tb$, $J_2/J_1 \simeq 0.04$
(in fact, for $\tb \gsim 4$,
$J_1 \simeq - a_2 \simeq 2.5$, $J_2 \simeq - b_2 \simeq 0.1$). Thus for
$\delta_i = 0$ a strong ($m_{1/2}$, $\mu$) correlation occurs whenever
$m_0 \lsim O(m_{1/2})$. This happens in the example of Fig.12,
where $m_0$ is severely bounded by the $M_A \geq (M_A)_{lb}$ condition,
and in the case of Fig.11, where the correlation is enforced by a very
small value of $J_2$: $J_2 = 0.06$. A ($m_{1/2}$, $\mu$) correlation is also
exhibited in Fig.9b for the range $m_{1/2} \gsim 150$ GeV, where $m_0$
is bounded by the cosmological constraint.
In general, we do not consider these physical regions with a strong
($m_{1/2}$, $\mu$) correlation as unnatural, since they are usually
realized without much tuning. We recall that the size of the
coefficients $J_1$ and $J_2$ is dictated by the RGE's
with their intrinsic cancellations, and that one naturally has
$J_1= O$(a few), $J_2= O(0.1-0.01)$. As we have seen, these properties,
combined with severe upper bounds on $m_0$, are sufficient to
generate the ($m_{1/2}$, $\mu$) correlation.
We turn now to the $A_0 \neq 0$ case. First we recall that $A_0$ is constrained
in the range $|A_0| \lsim 3$ from the absence of charge and color breaking
\cite{a_0}. Thus, allowing $A_0 \neq 0$ does not change essentially the general
picture previously discussed. The previous scenarios still occur, but at
different points in the parameter space. Two specific comments are in order
here: i) independently of its sign, $A_0$ disfavours the changeover from
gaugino dominance to higgsino dominance in the neutralino composition, ii) a
negative $A_0$ reduces the value of $M_A$ as compared to the $A_0=0$ case, and
so either provides a light $A$ boson (and hence interesting phenomenology) or
enforces a more stringent constraint on the parameter space.
\medskip
\section{Neutral Higgs Bosons}
Neutralino direct detection, to be discussed in the next Section,
is based on neutralino--nucleus scattering. In this process,
exchanges of neutral Higgs bosons play a dominant role, provided the
Higgs masses are not too heavy.
It is convenient to recall here some relevant properties
of the couplings of $\chi$ with matter via Higgs exchange. As was
already mentioned in the Introduction,
the two Higgs isodoublets $H_1$, $H_2$ yield 3 neutral Higgs
mass eigenstates: one $\mb{CP}$--odd ($A$)
state, whose mass $M_A$ is given by
expression (\ref{eq:M_A}) and two $\mb{CP}$--even
states (of masses $M_h$, $M_H$,
$M_h < M_H$), which are obtained from
$H_1^0$, $H_2^0$ by a rotation through an angle $\alpha$
\beqarr
H &=& \cos\alpha\,H_1^0 + \sin\alpha\,H_2^0 \nonumber \\
h &=& -\sin\alpha\,H_1^0 + \cos\alpha\,H_2^0 ~.
\eeqarr
\noindent
It is important to notice here that $\alpha$ depends very
sensitively on $M_A$, being very close to
zero for $\tb \gsim 4$ and rising very fast to $\pi/2$
for $M_A \lsim O(M_Z)$ (see Fig.15).
The angle $\alpha$ plays a crucial role in determining the size of the
neutral $h,H$--quark couplings. Here, as
we are interested in $\chi$--nucleus scattering,
we discuss explicitly only the couplings
involving the CP--even states, since $h,H$ are dominant compared to $A$.
The low--energy neutralino--quark
effective Lagrangian generated by Higgs exchange may be written
as follows \cite{[4]}
\beq
{\cal L}_{\mbi{eff}} = \sqrt2 G_F {m_Z \over m_{h,H}^2} F_{h,H}
\sum_q k_q m_q \bar{\psi}_{\chi} \psi_{\chi} \bar{q} q~.
\label{eq:el}
\eeq
\noindent
Here $F_{h,H}$ is the ratio of the Higgs--neutralino coupling to the $SU(2)$
gauge coupling, which depends on the composition of $\chi$
\beqarr
F_h &=& a_2 (a_3 \sin \alpha + a_4 \cos \alpha) \nonumber \\
F_H &=& a_2 (a_3 \cos \alpha - a_4 \sin \alpha)
\eeqarr
\noindent and the $k_q$ are given, for the up--type quarks and the
down--type quarks respectively, by
\begin{eqnarray}
\,\qquad &\qquad H \qquad & \qquad \qquad \quad h \qquad \nonumber \\
k_u \qquad & \qquad \sin\alpha / \sin\beta \qquad & \qquad
\phantom{-} \cos\alpha / \sin\beta \qquad \nonumber \\
k_d \qquad & \qquad \cos\alpha / \cos\beta \qquad & \qquad
- \sin\alpha / \cos\beta~.
\end{eqnarray}
\noindent
Note that, in general, since $\tb > 1$, the strength of the
coupling to the down--type quarks is bigger than the one to
the up--type quarks, and
${\cal L}_{\mbi{eff}}$ usually gets a sizeable contribution when
the $h$ boson is exchanged
($h$ is lighter than $H$ and is therefore favored
because of the propagator denominator in Eq.(\ref{eq:el})) and when
$\alpha \simeq \pi/2$, {\em i.e.}, when $M_A \lsim O(M_Z)$.
When this regime does not apply, the size of ${\cal L}_{\mbi{eff}}$ is
much suppressed.
The cross section for elastic neutralino--nucleus scattering
which follows
from the effective Lagrangian (\ref{eq:el}) will be given in Sect.IX.B.
\section{Direct detection}
Much experimental activity is under way in the direct search for
neutralino dark matter and the perspectives for
significant improvements in experimental sensitivities are
encouraging \cite{mosca}. In this class of experiments, a
relic neutralino would be detected by the amount of energy released by
its elastic scattering off nuclei in an appropriate apparatus. A
signature would be provided by a yearly modulation of the signal,
whose observations would require high statistics and
extremely good stability in the
detector response. Here we evaluate the event rates for
this process extending previous analyses to the non--universal $\delta_i \neq
0$ case.
Various materials are being used in the current experiments and others
are under investigation for future detectors. In this paper
we analyse two of
the most interesting materials: Ge (in its natural composition)
\cite{[15],[14],klapdor} and
$^{129}$Xe \cite{[17]}.
\subsection{Differential rates}
The nuclear recoil spectrum may be evaluated from the expression
\beq
\frac {dR}{dE_R}=
\sum_i \frac {R_{0,i}} {<E_R^{max}>} F_i^2(E_R)I(E_R)
\label{eq:dir1}
\eeq
\noindent
where
\beq
R_{0,i}=N_T \frac {\rho_{\chi}} {m_\chi} \sigma_i <v>~.
\label{eq:dir2}
\eeq
\noindent
In Eqs.(\ref{eq:dir1})--(\ref{eq:dir2}) we use the following
notations:
the subscript $i$ refers to the two cases of coherent and spin--dependent
effective interactions,
$N_T$ is the number of the target nuclei per unit of mass,
$\rho_\chi$ is the local neutralino matter density, and
$E_R$ is the nuclear recoil energy given by
$E_R={{m_{red}^2}}v^2(1-\cos \theta^*)/{m_N}$,
where $\theta^*$ is the scattering
angle in the neutralino--nucleus center--of--mass frame,
$m_N$ is the nuclear mass, $m_{\rm red}$ is the neutralino--nucleus
reduced mass and $v$ is the relative velocity. The maximum value of $E_R$ is
$E_R^{max}={{2m_{red}^2}}v^2/{m_N}$. Returning to
(\ref{eq:dir1}--\ref{eq:dir2}),
$F(E_R)$ denotes the nuclear form factor, and
$\sigma_i$ is the (coherent/spin--dependent) neutralino--nucleus cross section.
The factor $I(E_R)$ is given by
\beq
I(E_R)=\frac {<v^2>} {<v>}
\int_{v_{min}(E_R)}^{v_{max}} dv \frac {f(v)} {v}
\label{eq:dir3}
\eeq
\noindent
where $f(v)$ is the velocity distribution of neutralinos in the
Galaxy, as measured in the Earth's rest frame, and
$v_{\rm min}(E_R)$ is given by
$v_{\rm min}(E_R)=({{m_N E_R}/({2m_{\rm red}^2})})^{1/2}$.
The averages
appearing in Eqs.(\ref{eq:dir1})--(\ref{eq:dir3})
denote averages over the velocity distribution in
the Earth's rest frame. An explicit formula for $I(E_R)$ in the case of a
Maxwellian velocity distribution may be found in Ref.\cite{direct}.
The differential rates to be discussed below will be
expressed in terms of the electron--equivalent energy $E_{ee}$ rather than
in terms of $E_R$. These two variables are
proportional: $E_{ee}=Q E_R$ where $Q$ is called the quenching factor: typical
values of $Q$ will be discussed shortly.
\subsection{Neutralino--nucleus elastic cross sections}
The total cross sections for neutralino--nucleus elastic scattering
have been evaluated following standard procedures
\cite{elflor,direct,[4],[5],goodman}.
Here we only summarize some of the main properties. Neutralino--quark
scattering is described by amplitudes with
Higgs--boson exchanges and $Z$--boson exchange
in the t--channel,
and by amplitudes with
squark exchanges in the s-- and u--channels.
The neutral Higgs bosons considered here are the two
CP--even bosons: $h,H$ and the CP--odd
one: $A$, whose couplings were previously discussed in Sect.VIII.
The relevant properties for these amplitudes are: 1) Higgs--boson
exchanges contribute a coherent cross section which vanishes only
when there is no zino--higgsino mixture in the neutralino composition
\cite{[4]},
2) $Z$--boson exchange
provides a spin--dependent cross section which receives
contributions only from the higgsino components of $\chi$, 3) squark
exchanges contribute a coherent cross section (due to zino--higgsino
mixing) as well as a spin--dependent cross section (due mainly to the
gaugino components of $\chi$)\cite{[5]}.
As examples we recall here only the expressions for the coherent
cross section due to the exchange of a Higgs boson ($h$ or $H$) and the
spin--dependent one due to Z exchange.
The former cross section is easily evaluated from the effective
Lagrangian of Eq.(\ref{eq:el}) \cite{[4]}
\beq
\sigma_{\mbi{CH}} = \frac {8 G_F^2} {\pi} \frac {m_Z^2} {m_{h,H}^4}
\alpha^2_{h,H} m_{red}^2 A^2
\eeq
\noindent
where $A$ is the
nuclear mass number and $\alpha_{h,H}$ is given by
\beq
\alpha_{h,H} = F_{h,H} I~~~~,~~~~
I=\sum_q k_q m_q \langle N|\bar{q} q |N \rangle .
\eeq
\noindent
The quantity $I$ may be expressed conveniently in terms of the
$\pi N$ sigma--term $\sigma_{\pi N}$ and of a parameter
$a$ which is related to the strange--quark content of the nucleon
$y$ by
\beq
a = y (m_s/(m_u + m_d))~~~,~~~~~ y = 2 \frac {<N|\bar{s} s|N>}
{<N|\bar{u} u+\bar{d} d|N>}~.
\eeq
One has
\beq
I \simeq k_u g_u+k_d g_d
\eeq
\noindent
where
\beqarr
g_u &=& {4 \over {27}} \left( m_N + {{19}\over{8}} \sigma_{\pi N} -
a \sigma_{\pi N} \right)
\nonumber \\
g_d &=& {2 \over {27}} \left( m_N + {{23}\over{4}} \sigma_{\pi N} +
\frac {25} {2} a \sigma_{\pi N} \right) ~.
\eeqarr
\noindent
Unfortunately, the values of both the quantities $y$ and $\sigma_{\pi N}$
are somewhat uncertain. Here, for $y$ we use the central value of the
most recent evaluation: $y = 0.33 \pm 0.09$, obtained from a lattice
calculation \cite{liu}. For $\sigma_{\pi N}$, which is derived by
phase--shift analysis and dispersion relation techniques from
low--energy pion--nucleon scattering cross--sections \cite{chen,gass}, we
employ
the value of Ref.\cite{gass}: $\sigma_{\pi N} = 45$ MeV. We then find the
results: $g_u =123$ MeV, $g_d =288$ MeV (we use $2 (m_s/(m_u + m_d)) = 29$
\cite{bj}). We note that these values further
reinforce the role of the down--type quarks as compared to the up--type
ones.
We point out that the Higgs--nucleon couplings for nucleons
bound in a nucleus may be renormalized by the nuclear
medium. As a consequence, the strength of $I$ might in principle be
reduced to some extent \cite{brown}. However, this effect is neglected here.
Now let us turn to the spin--dependent cross section due to
Z exchange. This may be cast into the usual form
\cite{[5],goodman}
\beq
\sigma_{\mbi{SD}} = \frac {8 G_F^2} {\pi} (a_3^2-a_4^2)^2 m_{red}^2
(\sum_q T_{3L,q} \Delta q)^2 \lambda^2 J(J+1)~.
\label{eq:sd}
\eeq
\noindent
In this paper we use this formula for $^{73}$Ge (this isotope is present
at the level of 7.8 \% in the natural composition of Ge) and to $^{129}$Xe. For
these nuclei we employ the values of $\lambda^2$ obtained in the
odd--group model \cite{elflor}, where only the odd nuclear species in
odd--even nuclei are explicitly taken into account. The $\Delta q$'s in
Eq.(\ref{eq:sd}) denote the fractions of the nucleon spin carried
by the quarks $q$ in the
nucleon of the odd species, and the $T_{3L,q}$'s stand for the third
components of the quark weak isospin. The values for the $\Delta q$'s
are taken from Ref.\cite{ellkar}.
It is worth noticing that the
event rates for neutralino direct detection with the materials considered here
are largely dominated by coherent
effects in most regions of the parameter space. In the small domains
where spin--dependent effects dominate over the coherent ones the total
rates are usually too small to allow detection.
The experimental strategy of employing materials enriched in
heavy isotopes of high spin is interesting for a search for
hypothetical dark matter particles which interact with matter via
substantial spin--dependent interactions. However, this
approach does not appear to be very fruitful for neutralinos.
One more ingredient which enters the event rate in Eq.(\ref{eq:dir1})
is the nuclear form factor, which
depends sensitively on the nature of the effective
interaction involved in the neutralino--nucleus scattering.
For the coherent case, we simply employ the standard
parameterization \cite{[12]}
\beq
F(E_R)=3 \frac {j_1(qr_0)} {qr_0} e^{-\frac {1} {2}} s^2 q^2
\label{eq:ff}
\eeq
\noindent
where $q^2 \equiv\mid{\vec {q}}\mid^2=2m_NE_R$ is the squared
three--momentum transfer, $s \simeq 1~ \rm fm$ is the
thickness parameter for
the nucleus surface, $r_0 = (r^2-5s^2)^{1/2}$, $r=1.2~A^{1/3}$ fm and
$j_1(qr_0)$ is the spherical Bessel function of index 1.
The form factor in Eq.(\ref{eq:ff}) introduces a substantial suppression in the
recoil spectrum unless $q r_0 \ll 1$. A noticeable reduction in $dR/dE_R$ may
already occur at threshold $E_R=E_R^{\rm th}=E_{\mb{ee}}^{\rm th}/Q$ when
$r_0 \sqrt{2 m_N E_R^{\rm th}}$ is not small compared to unity. The actual
occurrence of this feature depends on parameters of the detector
material: nuclear radius, quenching factor, threshold energy
$E_{\mb{ee}}^{\rm th}$. The values
of these parameters for the nuclei considered in this
paper are reported in Table I \cite{mosca,[14],klapdor,[17]}, and
the values of $F^2(E_R^{\mb{th}})$ calculated from
Eq.(\ref{eq:ff}) are given in this same Table.
Since we consider in this paper mainly the value of the differential rate near
threshold, $F^2(E_R^{\mb{th}})$ is the most relevant quantity.
We see from the values in the Table that the reduction
introduced by the form factor is moderate
in Ge, but quite substantial in $^{129}$Xe.
In general, for the spin--dependent case there are no
analytic expressions for the form factors.
However, numerical analyses have been performed for a number of nuclei.
The general feature is that these form factors
have a much milder dependence on $E_R$ as compared to the coherent ones,
because only a few nucleons participate in the neutralino--nucleus
scattering in this case.
In our evaluations we use the results of Refs.\cite{[12],[19]} for
$^{131}Xe$ and $^{73}Ge$ respectively.
\bigskip
\begin{table}
\centering
\caption{
Characteristics of some current experiments.
In the second column is reported the quenching
factor $Q$, in the third column the electron--equivalent energy at
threshold, in the fourth the square of the form factor at threshold, and in
the last column the present experimental sensitivity.
}
\begin{tabular}{|r|r|r|r|r|} \hline
{\em \mb{Nucleus}} & $Q$ & $E_{ee}^{th}(\mb{KeV})$ & $F^{2}(E_{R}^{th})$ &
${\mb{evts/(Kg~d~ KeV)}}
$\\ \hline
Ge\cite{[14]} & 0.25 & 2 & 0.87 & 3.0 \\ \hline
Ge\cite{klapdor} & 0.25 & 12 & 0.41 & 0.2 \\ \hline
Xe\cite{[17]} & 0.80 & 40 & 0.07 & 0.8 \\ \hline
\end{tabular}
\end{table}
\subsection{Local Neutralino Density}
We denote the local halo density by $\rho_l$, for which we use the estimate
$\rho_l = 0.5$ GeV cm$^{-3}$ \cite{turner}.
For the value of the local neutralino density $\rho_\chi$ to be
used in the rate of Eq.(\ref{eq:dir2}), for each point of the model parameter
space we take into account the relevant value of the cosmological neutralino
relic density. When $\Omega_\chi h^2$ is larger than a minimal
$(\Omega h^2)_{\rm min}$ required by observational data and by large
scale structure calculations
we simply put $\rho_\chi=\rho_l$.
When $\Omega_\chi h^2$ turns out less than
$(\Omega h^2)_{\rm min}$, the neutralino may only provide a fractional
contribution
${\Omega_\chi h^2 / (\Omega h^2)_{\rm min}}\equiv \xi$
to $\Omega h^2$; in this case we take
$\rho_\chi = \rho_l \xi$.
The value to be assigned to $(\Omega h^2)_{\rm min}$ is
somewhat arbitrary. Here we set it equal to 0.1.
It is worth remarking here that, due to this scaling procedure, for the
direct detection rate one has: i) $R_{0,i} \propto \rho_l \sigma_i$
for ${\Omega_\chi h^2 \geq (\Omega h^2)_{\rm min}}$ and
ii) $R_{0,i} \propto \rho_l\xi\sigma_i \propto \rho_l\sigma_i/
<\sigma_{\mbi{ann}} v>_{\mbi{int}}$
for ${\Omega_\chi h^2 < (\Omega h^2)_{\rm min}}$.
Thus the rate $R_{0,i}$ is large in the regions of the parameter space
where $\sigma_i$ is large. This is trivial in case i), but it is also
true in case ii), since when $\sigma_i$ is large also $\sigma_{\mbi{ann}}$
increases but in such a way that usually the ratio
$\sigma_i/\sigma_{\mbi{ann}}$ increases too. Because of the relation
$\Omega_{\chi} h^2 \propto <\sigma_{\mbi{ann}} v>^{-1}_{\mbi{int}}$
it follows that $R_{0,i}$ is large for neutralino configurations with
modest values of the relic abundance, and {\em vice versa}.
\subsection{Results for detection rates}
The most significant quantity in comparing experimental data and
theoretical evaluations for direct detection is the differential rate
$dR/dE_{ee} = (dR/dE_R)/Q$ (with $dR/dE_R$ defined in Eq.(\ref{eq:dir1}))
rather than the total rates, obtained by integration over wide ranges
of $E_{ee}$. By using the differential rate
instead of the integrated ones, one obtains the best signal--to--background
ratio. Note that the experimental spectra, apart from
an energy interval around threshold, usually show a very flat
behaviour, whereas signals for light neutralinos are decreasing
functions of the nuclear recoil energy.
A complete procedure would then be to compare the
experimental and theoretical rates over the whole $E_{ee}$ range.
However, to simplify the presentation here,
we give our results in terms of the rate integrated over a narrow
range of 1 KeV at a specific value of $E_{ee}$, the one
which appears the most appropriate for each experiment:
typically it corresponds to a point close to the experimental
threshold. To be definite we consider the following cases:
i) Ge (natural composition). Among the various running experiments
\cite{mosca}, we select the
{two} which, at present, appear to provide the most stringent limits:
a) Caltech--PSI--Neuchatel \cite{[14]} with $E^{th}_{ee} = 2$ KeV,
differential rate $\simeq 3$ events/(Kg day KeV);
b) Heidelberg--Moscow \cite{klapdor} with $E^{th}_{ee} = 12$ KeV,
differential rate $ \simeq 0.2$ events/(Kg day KeV). Correspondingly,
for Ge we
have evaluated our rate by integrating $dR/dE_{ee}$ over the range
(2--3) KeV for experiment a) and over
(12--13) KeV for experiment b). It turns out that the case b) provides
the most stringent bound also for light neutralinos.
ii)$^{129}$Xe. In this case, taking into account the features of
the DAMA experiment \cite{[17]}, we have considered the rate R
integrated over the range (40--41) KeV.
Our results are shown in Figs.16--19. Figs.16--18 report the rate
for a Ge detector for the regions of the parameter space which are
depicted in Figs.12--14, respectively. In parts (a) and (b) of each
figure, R is displayed in the form of a scatter plot, in terms of
$m_{\chi}$ and of the relic abundance, respectively. The
horizontal line denotes the present level of sensitivity
in the Heidelberg--Moscow experiment. We notice that, in all
cases shown in these figures, the experimental sensitivity is
already, for some configurations, at the level of the predicted rate.
Some points of the supersymmetric parameter space, denoted by filled squares in
Figs.12--14, are even already excluded by present data.
The exploration potential
of this class of experiments as the sensitivity is improved is apparent from
these figures.
Fig.19 shows the rate R for $^{129}$Xe for the region of the parameter
space displayed in Fig.13: again the horizontal line gives the
present experimental sensitivity. A comparison of Fig.19 with
Fig.17 shows that the Ge experiments are currently more effective.
However, it has to be noticed that experiments with
liquid Xe may become extremely competitive in the future \cite{[17]}.
A few more remarks are in order here:
i) The cases displayed in Figs.16--18 present the common feature of
providing fair
chances for direct detection. This is not a surprise,
since these representative points all belong to the category of
configurations with small values of $M_A$. As was stressed before,
once we move away from these appealing physical regions of the
neutralino parameter space, the rates for direct detection may fall
far below (by many orders of magnitude) the detection sensitivities
(present or future). This unfortunate situation occurs, for instance,
typically as we move towards smaller values of $\tb$. However, one
should keep in mind that the regime of very large $\tb$, where signals may be
sizeable, represents a very interesting scenario,
deserving much attention and exploration. In fact this is one of the two
options, very small or very large $\tb$, which seem to fit
low--energy phenomenology at the best \cite{ch}.
ii) The scatter plots in parts b) of Figs.16--19 show
explicitly a property previously mentioned in Sect.IX.C,
namely that the scaling procedure adopted
to evaluate the neutralino local density implies a
R--$\Omega_{\chi} h^2$ correlation.
Configurations which provide a measurable R usually entail a low
$\Omega$ and viceversa. Only in a few cases the neutralino may be detectable
by direct detection and also provide a sizeable contribution to
$\Omega$.
\section{Conclusions}
In the present paper we have discussed some possible scenarios for
neutralino dark matter which originate from the relaxation of the
assumption of strict universality for soft scalar masses at $M_{GUT}$.
This approach derives from the general consideration that many crucial
theoretical points entering not only grand unified and supersymmetric
theories, not to mention the Standard Model, are far from being understood
and/or verified. For this reason, any new theoretical assumption
has to be fully
scrutinized. This is even more important because new assumptions in
supersymmetric models are often introduced not because of solid
arguments, but rather for the sake of simplicity and
for the need to reduce the large number of free parameters that
would otherwise prevent any firm prediction.
In our work we have discussed different scenarios, by considering
various physical constraints in a sort of
hierarchical order, giving top priority to the requirement of
radiative EWSB, implemented with a no--fine--tuning criterion, and to the
cosmological relic neutralino density constraint.
Some other assumptions, often introduced in the
literature, have been
relaxed in our work. This is in particular the case for universality
in the soft scalar masses. However, it has to be remarked that the type of
departure from universality that we have considered in our paper
is far from being
the most general one, as was noticed in Sect.II. In particular, it only
refers to the Higgs masses, and not to the sfermion masses.
The implications of the various scenarios on neutralino relic abundances
and rates for detection rates have been analysed, and the impact of a
non--universality in $m_0$ has been discussed for the whole range of
$\tb$. We have shown that the departure from $m_0$ universality is
particularly interesting in two respects:
i) Small values of
$M_A$ are allowed: this has in itself the dramatic consequence for
direct detection of generating a large value for the angle $\alpha$ and
large couplings to matter of the lightest neutralino $\chi$.
ii) Higgsino or mixed higgsino--gaugino configurations appear for all
$\tb$: this contrasts with the pure gaugino configurations
favoured by strict $m_0$ universality. \hfill \break
Consequences of such a departure from universality on the size of the
neutralino relic abundance have been
analysed for both large and small values of $\tan \beta$. It has been
shown that, because of the previous properties,
deviations from universality may reduce the value of
$\Omega_{\chi} h^2$.
The predicted rates for direct detection has been
analysed in detail and compared with current and foreseen experimental
sensitivities. The role of the previous properties in opening
interesting perspectives for this kind of search has been elucidated.
We find that presently--running experiments are already impacting interesting
regions of the neutralino parameters space in some of the non--universal
scenarios discussed here.
{\bf Acknowledgements.}
We wish to express our thanks to Uri Sarid for interesting discussions
and for contributions in the very early stages of our work. We also
gratefully acknowledge very useful conversations with Marek Olechowski
and informative discussions on experimental aspects of direct detection
with Pierluigi Belli, Rita Bernabei and Luigi Mosca. NF wishes to
express his gratitude to the Fondazione A. Della Riccia for a
fellowship. This work was supported in part by the Ministero della
Ricerca Scientifica e Tecnologica (Italy).
|
\section{Introduction}
\indent
Over the decades, as is well known, a wide variety of regularization schemes
have been developed in quantum field theory \cite{1}. However, as every schemes
heve their own distinct advantages and disadvantages, this topic is still one
of the important and fundamental issues under investigation. One of the most
challenging problem is perhaps how to preserve all properties of the original
action manifestly and consistently.
A few years ago, a new regularization method named intrinsic vertex
regularization was first proposed for the $\phi^4$ theory by Wang and Guo
\cite{2}. The key point of the method is, in fact,
based upon the following simple observation: For a given ultraviolet
divergent function at certain loop order in a renormalizable QFT, there always
exists a set of convergent functions at the same loop order such that their
Feynman graphs share the same loop skeleton and the main difference is that
the convergent ones have additional vertices of certain kind and the original
one is the case without these vertices. This is, in fact, a certain intrinsic
relation between the original ultraviolet divergent graph and the convergent
ones in the QFT. It is this relation that indicates it is possible to introduce
the regularized function for the divergent function with the help of those
convergent ones so that the potentially divergent integral of the graph can be
rendered finite while for the limiting case of the number of the additional
vertices $q\to 0$ the divergence again becomes manifest in pole(s) of $q$.
To be concrete, let us consider a 1PI graph with $I$ internal lines at one loop
order in the $\phi^4$ theory. Its superficial degree of divergences in the
momentum space is
$$
\delta=4-2I.
$$
When $I=1$ or $2$, the graph is divergent. Obviously, there exists such kind of
graphs that they have additional $q$ four-$\phi$-vertices in the internal
lines. Then the number of internal lines in these graphs is $I+q$ so that the
divergent degree of the new 1PI graphs become
$$
\delta'=4-2(I+q).
$$
If $q$ is large enough, the new ones are convergent and the original divergent
one is the case of $q=0$. Thus, a certain intrinsic relation has been reached
between the original divergent 1PI graph and the new convergent ones at the
same loop order.
However, application of this method to QED runs into a difficulty. The problem
is that, unlike the the $\phi^4$ theory, the electron-photon vertex in QED
carries a $\gamma$-matrix and is a Lorentz vector. As a result, simply
inserting the vertex would increase the rank of the function as Lorentz tensors
and would make the problem quite complicated. In
order to overcome this difficulty, in \cite{3,4} the authors introduced an
alternative method. We follow the example of the $\phi^4$ theory to demonstrate
the key point of this method: One shifts the mass term of
the $\phi$ field from $m^2$ to $m^2+\mu^2$ and regards $\mu^2\phi^2$ as a new
vertex in addition to the vertex $\lambda\phi^4$. By inserting the new vertex
into the internal $\phi$ lines in the graph of a given 1PI $n$-point divergent
function, a set of new convergent functions can be obtained provided the number
of inserted vertices, $q$, is large enough. Then one can introduce a new
convergent function, the regularized function, and the potential infinity in
the original 1PI $n$-point function may be recovered as the $q\to 0$ limiting
case of that function. Obviously, the mass shifting method can be easily
generalized to QED by simply shifting the electron mass from $m$ to $m+\mu$ and
regarding the term $-\mu\bar{\psi}{\psi}$ as a new vertex. In fact, as has been
shown in \cite{4}, it turns out to be successful to QED. Nevertheless, it is
not really
intrinsic since the Feynman rules of the theory have to be modified. As a
result, it may not completely work for non-Abelian gauge theories, {\it e.g.}, QCD,
because generally it is not clear whether the gauge symmetry can be preserved
manifestly for these theories, although such a proof for QED at one loop level
has been given \cite{6}.
Very recently, we presented an improved approach in \cite{7} to reexamine the
$\phi^4$ theory and QED, in which a new concept, {\it inserter}, was
introduced. An inserter is a vertex or a pair of vertices linked by an internal
line, in which the momenta of the external legs are all set to zero, and, if
there are any, all the Lorentz indices are contracted in pair by the spacetime
metric and all the internal gauge symmetry indices are contracted by the
Killing-Cartan metric in the corresponding representation, so that as a whole
an inserter always carries the vacuum quantum numbers, i.e. zero momentum,
scalar in the spacetime symmetry, and singlet in internal and gauge symmetries.
It is not hard to see that in any given QFT as long as a suitable kind of
inserters are constructed with the help of the Feynman rules of the theory,
some intrinsic relations between the divergent functions and convergent ones at
the same loop order will be found by simply regarding the convergent ones as
the ones given by suitably
inserting $q$-inserters in all internal lines in the given divergent ones.
The crucial point of this approach, therefore, is very simple but fundamental,
that is, the entire procedure is intrinsic in the QFT. There is nothing
changed, the action, the Feynman rules, the spacetime dimensions etc. are all
the same as that in the given QFT. This is a very important property which
should shed light on the challenging problem mentioned at the beginning of the
paper. Consequently, in applying to other cases all symmetries and topological
properties there should be preserved in principle.
In what follows, we concentrate on how to apply the inserter approach to QCD at
one loop order. We present the main steps and the results of the inserter
regularization procedure for it. We find that, as is expected, the gauge
invariance is preserved manifestly, and all results are the same as those
derived by means of other regularization methods.
\section{Intrinsic Regularization in QCD}
\indent
The QCD Lagrangian, including ghost fields and gauge fixing terms, can be
written as
\begin{eqnarray}
{\cal L}_{QCD} & = & -\frac{1}{4}
(\partial_{\mu} A^a_{\nu}-\partial_{\nu}A^a_{\mu}+g_cf^{abc}A^b_{\mu}A^c_{\nu})
(\partial^{\mu} A^{a\nu} -\partial^{\nu} A^{a\mu}+g_cf^{ade}A^{d\mu}A^{e\nu})
\nonumber \\
& & -\frac{1}{2\xi}(\partial_{\mu} A^{a\mu})^2 -\bar{\eta}^a
\partial_{\mu}(\partial^{\mu} \delta^{ac}-g_cf^{abc}A^{b\mu})\eta^c
+\bar{\psi}[i\gamma_{\mu}(\partial^{\mu}
- ig_cA^{a\mu} \frac{\lambda^a}{2})-m]\psi~,
\label{L} \end{eqnarray}
and the Feynman rules are well known.
The main steps of the inserter approach for QCD may be stated more concretely
as follows. First, we should construct the inserters in QCD. This work, with
regards to simplicity and consistency with other theories, {\it e.g.}, the
electroweak theory, may actually be done within a more general framework,
namely, within the framework of the standard model in which QCD is contained.
The explicit expressions of all inserters in the standard model have been
preestablished in \cite{7}. Here we merely list those relevant to QCD:
\begin{itemize}
\item
The gluon-inserter:
\begin{eqnarray}
I^{\{g\}ab}_{~~\mu\nu}(p)=-6ig_c^2C_2({\bf 8}) g_{\mu\nu} \delta^{ab}.
\label{gluon-ins} \end{eqnarray}
\item
The ghost-inserter:
\begin{eqnarray}
I^{\{gh\}}_{~~~a_1a_2}(p) = -ig_c^2 C_2({\bf 8}) \delta_{a_1a_2}.
\label{ghost-ins} \end{eqnarray}
\item
The quark-inserter:
\begin{eqnarray}
I^{\{q\}}(p)=-i\lambda_q.
\label{quark-ins} \end{eqnarray}
\end{itemize}
In eqs.(\ref{gluon-ins}) and (\ref{ghost-ins}), $C_2({\bf 8})$ is the second
Casimir operator valued in the adjoint representation of $SU_c(3)$ algebra.
In eq.(\ref{quark-ins}), $\lambda_q$ takes value $\frac{g}{2}\frac{m_q}{M_W}$
in the standard model, but here its value is irrelevant for our purpose. It
should be mentioned that here the quark inserters are constructed by borrowing
the fermion-Higgs-vertex of Yukawa type from the standard model, this is in
analogy with as occurs in QED. The issue has been discussed in detail in
\cite{7}.
For a given divergent 1PI amplitude
$\Gamma^{(n_f,n_g)}(p_1,\cdots, p_{n_f};k_1,\cdots, k_{n_g})$
at the one loop order with $n_f$ external fermion lines and $n_g$ external
photon lines, we consider a set of 1PI amplitudes
$\Gamma^{(n_f,n_g)} (p_1,\cdots, p_{n_f};k_1,\cdots, k_{n_g}; q)$
which correspond to the graphs with, if the loop contained in the graph purely
consists of fermion lines, all possible $2q$ insertions of the fermion inserter
in the internal fermion lines, or in other cases, all possible $q$ insertions
of
the corresponding inserter in the internal boson (ghost) lines in the original
graph. The divergent degree therefore becomes:
$$
\delta=4-I_f-2I_g-2q.
$$
If $q$ is large enough,
$\Gamma^{(n_f,n_g)}(p_1,\cdots,p_{n_f};k_1,\cdots, k_{n_g}; q)$
are convergent and the original divergent function is the case of $q=0$. Thus
we reach a relation between the given divergent 1PI function and a set of
convergent 1PI functions at the one loop order. In fact, the function of
inserting the inserter(s) into internal lines is simply to raise the power of
the propagator of the lines and to decrease the degree of divergence of given
graph.
In order to regularize the given divergent function with the help of this
relation, we need to deal with those convergent functions on an
equal footing and pay attention to their differences due to the insertions.
To this end, we introduce a new function:
\begin{eqnarray}
\begin{array}{ll}
\Gamma^{(n_f,n_g)}(p_1,\cdots, p_{n_f};k_1,\cdots, k_{n_g}; ~q;~\mu)
{}~~~~~~~\\[4mm]
{}~~~~~~~= (-i\mu)^{2q} (-i\lambda)^{-2q} \frac 1 {N_q} \sum
\Gamma^{(n_f,n_g)}(p_1,\cdots, p_{n_f};k_1,\cdots, k_{n_g}; ~q)
\end{array}
\end{eqnarray}
where $\mu$ is an arbitrary reference mass parameter,
the summation is taken over the entire set of such ${N_q}$ inserted functions,
and the factor $(-i\lambda)^{-2q}$ introduced here, in which $\lambda$ stands
for $\lambda_q$ for fermion loop and for $g_c$ for other cases, is to cancel
the ones coming from the inserters. It is clear that this function is the
arithmetical average of those convergent functions and has the same dimension
in mass, the same order in coupling constant with the original divergent 1PI
function. Then we evaluate it and analytically continue $q$ from the integer to
the complex number. Finally, the original 1PI function is recovered as its
$q\to 0$ limiting case:
\begin{eqnarray}
\Gamma^{(n_f,n_g)}(p_1,\cdots, p_{n_f};k_1,\cdots, k_{n_g})
=\lim_{q\to 0}
\Gamma^{(n_f,n_g)}(p_1,\cdots, p_{n_f};k_1,\cdots, k_{n_g};q;\mu),
\end{eqnarray}
and the original infinity appears as pole in $q$. Similarly, this procedure
should work for the cases at the higher loop orders in principle.
The divergent 1PI graphs at the one loop order in QCD are as follows: the gluon
self-energy $\Pi_{\mu \nu}^{ab}(k)$, the quark self-energy $\Sigma(p)$, the
ghost self-energy $\tilde{\Pi}^{ab}(p)$, renormalized by $Z_3$, $Z^F_3$ and
$\tilde{Z}_3$, the three-gluon vertex $\Gamma^{abc}_{\mu\nu\lambda}
(k_1,~k_2)$, the four-gluon vertex
$\Gamma^{abcd}_{\mu\nu\lambda\tau}(k_1,~k_2,~k_3)$, the quark-gluon vertex
$\Gamma^{a}_{\mu}(p^{\prime},p)$, and the ghost-gluon vertex
$\tilde{\Gamma}^{abc}_{\mu}(p^{\prime},p)$, with the renormalization constant
$Z_1$, $Z_4$, $Z^F_1$, $\tilde{Z}_1$. In addition, there is a mass shift for
the quark, which we shall ignore. All corresponding graphs are listed in
figures 1-7. As numerious diagrams are concerned, evaluating them one by one in
detail would be much lengthy and unnecessary. In the next section, we will
evaluate in detail the gluon self-energy $\Pi_{\mu \nu}^{ab}(k)$ as a typical
example to show the main step of the approach, paying special attentions to the
gauge inva
riance of the function. Then in the subsequent section, we will directly give
all the results corresponding to other involved diagrams to verify the
Slavnov-Taylor identities at one loop order.
\section{Regularization and Evaluation of the Gluon Self-Energy $\Pi_{\mu
\nu}^{ab}(k)$}
\indent
Before the detailed evaluations are presented, we should first refer to a
special problem which arises in any massless theories, {\it i.e.}, the genuine
infrared divergence in these theories. In the regularization schemes, this
problem usually appears as the lack of consistent definitions of the
regularized Feynman integrals for the ones which are both ultraviolet and
infrared divergent. For instance, in the dimensional regularization scheme,
let's consider the massive integral
\begin{eqnarray}
\int\frac{d^{2\omega}l}{(2\pi)^{2\omega}}\frac{1}{(l^2+m^2)^n}
=\frac{i\Gamma (n-\omega)}{(4\pi)^{\omega}\Gamma (n)(m^2)^{n-\omega}}
\equiv I(m,~\omega,~n)~, ~~~(~m^2\neq 0~)
\label{massive} \end{eqnarray}
which converges for $\omega$ complex; the parameter $n$ is arbitrary but fixed.
We note first of all that the limit
$\displaystyle\lim_{m^2\to0}I(m,~\omega,~n)$ may or may not exist, depending on
the relative magnitudes of $n$ and $\omega$. But even if it did exist, another
problem could arise as we approach four-space (provided the original integral
is infrared divergent to begin with), because in general
$$
\displaystyle\lim_{\omega\to2} [ \displaystyle\lim_{m^2\to0} I(m,~\omega,~n)]
\neq \displaystyle\lim_{m^2\to0} [\displaystyle\lim_{\omega\to2}
I(m,~\omega,~n)],
$$
so that the massless integral $\displaystyle\lim_{\omega \to
2}{\int\frac{d^{2\omega}l}{(2\pi)^{2\omega}}\frac{1}{(l^2)^n}}$
can not be derived unambiguously from the massive integral (\ref{massive}).
Furthermore, the trick of inserting a finite mass into the integral and then
allowing it to approach to zero at the end of the calculation is, in general,
not a satisfactory prescription yet, because it spoils the gauge symmetry in
the original theory, provided such a symmetry existed in the first place. To
avoid this difficulty, 't Hooft and Veltman naively comjectured that
\begin{eqnarray}
\displaystyle\lim_{\omega \to 2}{\int\frac{d^{2\omega}l}{(2\pi)^{2\omega}}
\frac{1}{(l^2)^n}}=0~,~~~for~ \omega,~n~complex.
\label{massless} \end{eqnarray}
It has been shown that no inconsistencies occur, {\it e.g.}, in the Slavnov-Taylor
identities \cite{10}, due to the acceptance of the above conjecture.
In our application of the present approach to QCD, as we will see, the same
problem occurs, {\it e.g.}, in calculating the tadpole diagram Fig.1d of the gluon
self-energy. To solve this problem, we employ a conjecture analogous to 't
Hooft and Veltman's:
\begin{eqnarray}
\displaystyle\lim_{q\to 0}{\int\frac{d^4l}{(2\pi)^4}
\frac{1}{[(k-l)^2]^{Aq}(l^2)^{Bq+n}}}=0,~~for~q,~n,~complex,~A\geq 0,~B\geq
0,~A+B=1~.
\label{conj} \end{eqnarray}
Likewise, we will see that no inconsistencies occur due to the acceptance of
the eq.(\ref{conj}).
Now we turn to evaluate in detail the gluon self-energy $\Pi_{\mu\nu}^{ab}(k)$
shown in Fig.1. The diagrams contributing to $\Pi_{\mu\nu}^{ab}(k)$ are four in
number, namely, the gluon loop diagram, the ghost loop diagram, the quark loop
diagrams, and the gluon tadpole. The integral expressions of the regularized
diagrams in the momentum space are given in the appendix ( For simplicity, we
take the Feynman gauge $\xi =1$. ).
First, we consider the gluon loop contribution. From (\ref{glse-a}), a little
bit of algebra yields
\begin{eqnarray}
\Pi_{(A)\mu\nu}^{~~~ab}(k;q;\mu)~=~-\frac{1}{2}g_c^{2}[-6C_2({\bf 8})\mu^{2}]^q
\delta^{ab}I_1 ~, \nonumber
\end{eqnarray}
with
\begin{eqnarray}
I_1~=~\frac{1}{N_q} \displaystyle{\sum_{i=0}^{q} \int\frac{d^4p}{(2\pi)^4}}
\frac{-2p^2g_{\mu\nu}-(5k^2-2p\cdot k)g_{\mu\nu}
-10p_{\mu}p_{\nu}+2k_{\mu}k_{\nu}+5p_{\mu}k_{\nu}+5k_{\mu}p_{\nu}}{
(p^2)^{i+1}[(p-k)^2]^{q-i+1}}~. \nonumber
\end{eqnarray}
In the present case, $N_q=q+1$.
Note that because of eq.(\ref{conj}), the contribution of the first term in the
numerator of the above equation actually vanishes, so it can be neglected.
Using the Feynman parameterization, we get
\begin{eqnarray}
I_1 & = & \frac{1}{q+1} \displaystyle{\sum_{i=0}^{q}} \frac{\Gamma (q+2)}{\Gamma
(i+1)\Gamma (q-i+1)}
\displaystyle{\int_0^1dx} x^{q-i}(1-x)^i \nonumber \\
& & \times
\displaystyle{\int\frac{d^4p}{(2\pi)^4}} \frac{(2x-5)k^2g_{\mu\nu}
-10p_{\mu}p_{\nu} -(10x^2-10x-2)k_{\mu}k_{\nu}}{[p^2+x(1-x)k^2]^{q+2}}~,
\label{I1}\end{eqnarray}
where we have made a momentun shift: $p \rightarrow p-kx$, and the linear terms
in $p$ in the numerator have been dropped since they do not contribute to the
integral. Now the integration over $p$ can be performed by using the following
formulas:
\renewcommand{\theequation}{\arabic{equation}\alph{abc}}
\setcounter{abc}{0}
\addtocounter{abc}{1}
\begin{eqnarray}
\int \frac{d^4 p}{(2\pi)^4} \frac{(p^2)^{\beta}}{(p^2+M^2)^A} & = &
\frac{i}{(4\pi)^2}
\frac{\Gamma(2+\beta)\Gamma(A-2-\beta)}{\Gamma(A)}(M^2)^{2+\beta-A},
\label{int-a} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
\int \frac{d^4 p}{(2\pi)^4} \frac{(p^2)^{\beta}p_{\mu}p_{\nu}}{(p^2+M^2)^A}
& = &
\frac{i}{(4\pi)^2}\frac{1}{4}g_{\mu\nu}
\frac{\Gamma(3+\beta)\Gamma(A-3-\beta)}{\Gamma(A)}(M^2)^{3+\beta-A},
\label{int-b} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
\int \frac{d^4 p}{(2\pi)^4} \frac{(p^2)^{\beta}p_{\mu}p_{\nu}p_{\rho}p_{\sigma}}
{(p^2+M^2)^A}
& = &
\frac{i}{(4\pi)^2}\frac{1}{24}(g_{\mu\nu}g_{\rho\sigma}+g_{\mu\rho}g_{\nu\sigma}
+g_{\mu\sigma}g_{\nu\rho}) \nonumber \\
& & \times
\frac{\Gamma(4+\beta)\Gamma(A-4-\beta)}{\Gamma(A)}(M^2)^{4+\beta-A}.
\label{int-c} \end{eqnarray}
\addtocounter{abc}{1}
\renewcommand{\theequation}{\arabic{equation}}
{}From (\ref{I1}), (\ref{int-a}), and (\ref{int-b}) we get:
$$
\begin{array}{ll}
I_1~ = & \frac{i}{(4\pi)^2} \displaystyle{\sum_{i=0}^{q}} \frac{\Gamma
(q-1)}{(q+1)\Gamma (i+1)\Gamma (q-i+1)} \displaystyle{\int_0^1dx}
x^{q-i}(1-x)^i \nonumber \\
& \times
\displaystyle{ \frac{(q-1)[(2x-5)k^2g_{\mu\nu}-(10x^2-10x-2)k_{\mu}k_{\nu}]
-5x(1-x)k^2g_{\mu\nu}}{[x(1-x)k^2]^{q}} } \nonumber \\
~~~~~ = &
\frac{i}{(4\pi)^2} \frac{\Gamma (q-1)}{\Gamma (q+2)} \displaystyle{\int_0^1dx}
\frac{[q(2x-5)+(5x^2-7x+5)]k^2g_{\mu\nu}-(q-1)(10x^2-10x-2)k_{\mu}k_{\nu}}
{[x(1-x)k^2]^{q}}~, \nonumber
\end{array}
$$
where in the last step the summation over $i$ has been performed with the help
of the binomial theorem. This expression also makes sense when we make an
analytical continuation of $q$ from integer to complex number. When $q\to 0$,
using the expansion
$$
\frac{\Gamma(q-1)}{\Gamma(q+2)}=-\frac{1}{q}-q+O(q^2),~~~
[x(1-x)k^2]^q=1+q\ln [x(1-x)k^2]+O(q^2)~,
$$
and making a rescaling of the parameter $\mu^2 \rightarrow -6C_2({\bf
8})\mu^2$, we get:
\begin{eqnarray}
\Pi_{(A)\mu\nu}^{~~~ab}(k;q;\mu) & = & \frac{ig_c^2C_2({\bf 8})}{(4\pi)^2}
\delta^{ab} [ (\frac{19}{12}k^2g_{\mu\nu}-\frac{11}{6}k_{\mu}k_{\nu})\frac{1}{q}
\nonumber \\
& &
-(\frac{19}{12}k^2g_{\mu\nu}-\frac{11}{6}k_{\mu}k_{\nu})\ln\Bigl(\frac{k^2}{\mu^2}\Bigr)+(\frac{47}{36}k^2g_{\mu\nu}-\frac{14}{9}k_{\mu}k_{\nu}) ]~.
\label{no13} \end{eqnarray}
Clearly $\Pi_{(A)\mu\nu}^{~~~ab}(k;q;\mu)$ does not conserve current, this is
due to our choice to use a covariant gauge (rather than, for example, an axial
gauge). For the sake of this choice, we have to introduce spurious gluon
polarization states. These spurious states must be removed by taking the ghost
loop contribution into account. We could have computed with an axial or ``ghost
free'' gauge, but it is usually much easier to use the simple Feynman gauge and
add in the ghost contribution.
To compute ghost loop contribution (\ref{glse-b}), we use the same Feynman
parameterization arriving at
\begin{eqnarray}
\Pi_{(B)\mu\nu}^{~~~ab}(k;q;\mu) ~=~ -g_c^2 C_2({\bf 8})[C_2({\bf 8})\mu^2]^q
\delta^{ab}I_2 ~, \nonumber
\end{eqnarray}
with
\begin{eqnarray}
I_2 ~=~ \displaystyle{{\int_0^1dx} \int\frac{d^4p}{(2\pi)^4}}
\frac{p_{\mu}p_{\nu}-x(1-x)k_{\mu}k_{\nu}}{[p^2+x(1-x)k^2]^{q+2}}~, \nonumber
\end{eqnarray}
where momentun shift: $p \rightarrow p-kx$ has been made, and the linear terms
in $p$ in the numerator have been dropped. After performing the integration
over $p$ and taking the limit $q\to 0$ subsequently, we obtain
\begin{eqnarray}
\Pi_{(B)\mu\nu}^{~~~ab}(k;q;\mu)& = & \frac{ig_c^2C_2({\bf 8})}{(4\pi)^2}
\delta^{ab} [ (\frac{1}{12}k^2g_{\mu\nu}+\frac{1}{6}k_{\mu}k_{\nu})\frac{1}{q}
\nonumber \\
& &
-(\frac{1}{12}k^2g_{\mu\nu}-\frac{1}{6}k_{\mu}k_{\nu})\ln\Bigl(\frac{k^2}{\mu^2}\Bigr)+\frac{5}{36}k^2g_{\mu\nu}+\frac{1}{9}k_{\mu}k_{\nu}) ]~.
\label{no14} \end{eqnarray}
Again, we find that $\Pi_{(B)\mu\nu}^{~~~ab}(k;q;\mu)$ also does not conserve
current. However, the non-conserving term in (\ref{no14}) exactly cancels that
in (\ref{no13}), and makes
$\Pi_{(A)\mu\nu}^{~~~ab}(k;q;\mu)+\Pi_{(B)\mu\nu}^{~~~ab}(k;q;\mu)$ gauge
invariant, namely,
$$
k^{\mu}[\Pi_{(A)\mu\nu}^{~~~ab}(k;q;\mu)+\Pi_{(B)\mu\nu}^{~~~ab}(k;q;\mu)]
=0~.
$$
Next, the quark loop contribution $\Pi_{(C)\mu\nu}^{~~~ab}(k;q;\mu)$ need not
be recalculated since it is just $\Pi_{\mu\nu}^{QED}(k;q;\mu)$, which can be
found in \cite{6,7}, multiplied by a color factor $N_fTr({\bf T}^a{\bf T}^b)$:
\begin{eqnarray}
\Pi_{(C)\mu\nu}^{~~~ab}(k;q;\mu) & = & N_fTr({\bf T}^a{\bf T}^b)
\Pi_{\mu\nu}^{QED}(k;q;\mu)=\frac{4ig_c^2N_fC_2({\bf 3})}{(4\pi)^2} \delta^{ab}
(k_{\mu}k_{\nu}-k^2g_{\mu\nu}) \nonumber \\
& & \times
[\frac{1}{3q}+\frac{2}{3}-2\int_{0}^{1} dx x(1-x)\ln\frac{k^2 x(1-x)-m^2}
{\mu^2}+\cdots]~.
\label{no15} \end{eqnarray}
Obviously, $\Pi_{(C)\mu\nu}^{~~~ab}(k;q;\mu)$ itself is gauge invariant.
Finally, as in dimensional regularization scheme, from eq.(\ref{conj}) we know
that the contribution of the gluon tadpole diagram vanishes:
\begin{eqnarray}
\Pi_{(D)\mu\nu}^{~~~ab}(k;q;\mu)=0~.
\label{no16} \end{eqnarray}
{}From (\ref{no13}), (\ref{no14}), (\ref{no15}), and (\ref{no16}) we obtain the
final expression of the regularized gluon self-energy
$\Pi_{\mu\nu}^{ab}(k;q;\mu)$:
\begin{eqnarray}
\Pi_{\mu\nu}^{ab}(k;q;\mu) & = & \Pi_{(A)\mu\nu}^{~~~ab}(k;q;\mu)
+\Pi_{(B)\mu\nu}^{~~~ab}(k;q;\mu)+\Pi_{(C)\mu\nu}^{~~~ab}(k;q;\mu) \nonumber \\
& = & \frac{ig_c^2}{(4\pi)^2}\delta^{ab}(k^2g_{\mu\nu}-k_{\mu}k_{\nu})
\{[\frac{5}{3}C_2({\bf 8})-\frac{4}{3}N_fC_2({\bf 3})]\frac{1}{q}+\cdots\}~,
\end{eqnarray}
which is the same as that derived by means of other regularization methods and
satisfies the gauge invariance condition $k^{\mu}\Pi_{\mu\nu}^{ab}(k;q;\mu)=0$.
\section{Renormalization Constants and Slavnov-Taylor Identities at
one Loop Order}
\indent
So far we have calculated the gluon self-energy $\Pi_{\mu\nu}^{ab}(k)$ in
detail by means of the intrinsic regularization method at one loop level. As we
have expected, the result turns out to be gauge invariant. However, this is not
enough for us to say that the method preserves the gauge invariance of QCD. As
a complement, we should further show that the Slavnov-Taylor identities between
the renormalization constants hold, namely,
\begin{eqnarray}
\frac{Z_1}{Z_3} ~=~ \frac{Z^F_1}{Z^F_3} ~=~ \frac{\tilde{Z}_1}{\tilde{Z}_3} ~=~
\frac{Z_4}{Z_1}~.
\label{stid} \end{eqnarray}
It is these identities that guarantee that the renormalized theory possess the
same gauge theory structure as the original one. Moreover, they are essential
for proving the renormalizability and unitarity of the theory.
To verify these identities, all divergent 1PI graphs, not only
$\Pi_{\mu\nu}^{ab}(k)$, but also others, must be taken into account. After
lengthy and tedious calculations, we get
\renewcommand{\theequation}{\arabic{equation}\alph{abc}}
\setcounter{abc}{0}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
Z_1=1+\frac{g_c^2}{(4\pi)^2}[\frac{2}{3}C_2({\bf 8})-\frac{4}{3}N_fC_2({\bf
3})]\frac{1}{q}
\label{const-a} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
Z_3=1+\frac{g_c^2}{(4\pi)^2}[\frac{5}{3}C_2({\bf 8})-\frac{4}{3}N_fC_2({\bf
3})]\frac{1}{q}
\label{const-b} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
Z_4=1+\frac{g_c^2}{(4\pi)^2}[-\frac{1}{3}C_2({\bf 8})-\frac{4}{3}N_fC_2({\bf
3})]\frac{1}{q}
\label{const-c} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
Z^F_1=1-\frac{g_c^2}{(4\pi)^2}[C_2({\bf 8})+\frac{8}{3}C_2({\bf 3})]\frac{1}{q}
\label{const-d} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
Z^F_3=1-\frac{g_c^2}{(4\pi)^2}\frac{8}{3}C_2({\bf 3})\frac{1}{q}
\label{const-e} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
\tilde{Z}_1=1-\frac{g_c^2}{(4\pi)^2}\frac{C_2({\bf 8})}{2}\frac{1}{q}
\label{const-f} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\vspace{-6mm}
\begin{eqnarray}
\tilde{Z}_3=1+\frac{g_c^2}{(4\pi)^2}\frac{C_2({\bf 8})}{2}\frac{1}{q}
\label{const-g} \end{eqnarray}
\addtocounter{abc}{1}
\renewcommand{\theequation}{\arabic{equation}}
It is easy to see that the Slavnov-Taylor identities (\ref{stid}) indeed holds.
For a close observation, we note that although these expressions are calculated
in a specific gauge therefore their gauge dependence are not explicit, they are
in fact all gauge dependent ( For instance, in the axial gauge, the effective
Lagrantian has no ghosts and has the same structure as in QED, resulting in the
identity $Z^F_1=Z^F_3$, which is patently untrue in the Feynman gauge.). We
also remark that the fermion contribution to the vector boson quartic
self-interaction must diverge if the Slavnov-Taylor identities (\ref{stid}) are
to hold because we explicitly see that the ratio $Z^F_1/Z^F_3Z_3^{\frac{1}{2}}$,
by which the bare coupling constant is related to the renormalized one,
contains a fermion contribution. On the contrary, the corresponding box diagram
of QED is finite. The reason is that in QED the box diagram's divergence
vanishes only upon symmetrization of the external photon lines denoted only by
their vector in
dices, while in QCD the symmetrization of the external lines can be performed
in two ways: by symmetrizing on both vector and group indices, which as in QED
gives no divergence, {\it or} by antisymmetrizing on both vector and group
indices. It is this new contribution which deverges. The same reason can be
applied to the fermion contribution to the triple gauge vertex.
\section{Concluding Remarks}
\indent
We have shown the main steps and results for the regularization of the
divergent 1PI functions at the one loop order in QCD by means of the inserter
proposal for the intrinsic regularization method. It turns out to be
satisfactory: The gauge invariance is preserved manifestly and the results are
the same as those derived by means of other regularization methods. It is
natural to expect that this proposal should be available to the cases at
higher loop orders in principle.
The renormalization of the QCD under consideration in this scheme should be the
same as in usual approaches. Namely, we may subtract the divergent part of the
$n$-point functions at each loop order by adding the relevant counterterms to
the action. The renormalized $n$-point functions are then evaluated from the
renormalized action. In the limiting case, we get the finite results for all
correlation functions.
It should be mentioned that the method presented here is somewhat analogous to
the analytic regularization method developed by Speer \cite{11,12}. However,
the two methods are in fact different: As is well known, the analytic
regularization method violates unitarity, this is due to the fact that it
actually continuing the power of propagators in an arbitrary way. While in our
method this not the case. Here we trie to find out a procedure of
regularization from some physical principles. That is, giving a ultraviolet
divergent process, one can always find a set of convergent function obtainable
from existing Feymann rules and for the limiting case it turns to be the
original ultra-divergent one. Or from another point of view, given a
ultra-divergent function, one can always extract a set of convergent functions
from a certain physical process, which is gauge invariant.
After taking the limitation of the convergent functions, the divergent
function is naturally regularized. There is nothing changed, the action, the
Feynman rules, the spacetime dimensions etc. are all the same as that in the
given QFT. From this viewpoint, one should have no doubt of gauge invariance
and the unitarity of the method since the sum of the convergent functions comes
from a certain physical process.
Application of our approach to gauge theories containing spontaneous symmetry
breaking such as the standard model should be straightforward. Also, it will be
much helpful to apply our approach to some other cases, such as anomalies, SUSY
theories {\it etc.}, since in these cases the symmetries and topological properties
are sensitive to the spacetime dimensions and the number of fermionic degrees
of freedom {\it etc.}, thus we are unable to tackle them consistently by means of the
hitherto well-known regularization methods such as dimensional regularization
method. It is reasonable to expect that the approach presented here should be
able to get rid of those problems. We will investigate these issues in detail
elsewhere.
\bigskip
\bigskip
{\raggedright{\Large\bf Appendix: Integral Expressions of the Divergent 1PI
Graphs at}}\\
{\raggedright{\Large\bf \* \* \* \* \* \* \* \* \* \* \* \* \* \* \* One Loop
Level in QCD}}\\
\indent
There are a number of divergent iPI graphs at one loop level in QCD, which
contribute to the gluon self-energy $\Pi_{\mu \nu}^{ab}(k)$, the quark
self-energy $\Sigma(p)$, the ghost self-energy $\tilde{\Pi}^{ab}(p)$, the
three-gluon vertex $\Gamma^{abc}_{\mu\nu\lambda} (k_1,~k_2)$, the four-gluon
vertex $\Gamma^{abcd}_{\mu\nu\lambda\tau}(k_1,~k_2,~k_3)$, the quark-gluon
vertex $\Gamma^{a}_{\mu}(p^{\prime},p)$, and the ghost-gluon vertex
$\tilde{\Gamma}^{abc}_{\mu}(p^{\prime},p)$ respectively. In what follows we
present the integral expressions of the regularized diagrams in the momentum
space (in Feynman gauge $\xi =1$).
\noindent
1. \hspace{0.3cm}The integral expressions of the regularized diagrams
contributing to $\Pi_{\mu \nu}^{ab}(k)$:
\renewcommand{\theequation}{\arabic{equation}\alph{abc}}
\setcounter{abc}{0}
\addtocounter{abc}{1}
\begin{eqnarray}
\Pi_{(A)\mu\nu}^{~~~ab}(k;q;\mu)& = & \frac{1}{2}g_c^{2-2q}
\mu^{2q}[-6ig_c^2C_2({\bf 8})]^q Tr({\bf F}^a{\bf F}^b)
\frac{1}{N_q}\sum_{i=0}^{q} \int\frac{d^4p}{(2\pi)^4} \nonumber \\
& & \times
G_{\mu\rho_1\sigma_1}(k,-p,p-k) G_{\nu\rho_2\sigma_2}(-k,p,k-p) \nonumber \\
& & \times
g^{\rho_1\rho_2} g^{\sigma_1\sigma_2} \Bigl(\frac{-i}{p^2}\Bigr)^{i+1}
\Bigl(\frac{-i}{(p-k)^2}\Bigr)^{q-i+1}~,
\label{glse-a} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{eqnarray}
\Pi_{(B)\mu\nu}^{~~~ab}(k;q;\mu)& = & g_c^{2-2q} \mu^{2q}[-ig_c^2C_2({\bf
8})]^q Tr({\bf F}^a{\bf F}^b) \frac{1}{N_q}\sum_{i=0}^{q} \int\frac{d^4p}{(2\pi)^4}
\nonumber \\
& & \times p_{\mu}(p-k)_{\nu}\Bigl(\frac{-i}{p^2}\Bigr)^{i+1}
\Bigl(\frac{-i}{(p-k)^2}\Bigr)^{q-i+1}~,
\label{glse-b} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{eqnarray}
\Pi_{(C)\mu\nu}^{~~~ab}(k;q;\mu)& = & g_c^{2}\mu^{2q} (-i\lambda_q)^{2q} N_f
Tr({\bf T}^a{\bf T}^b)\frac{1}{N_q}\sum_{i=0}^{2q} \int\frac{d^4p}{(2\pi)^4}
\nonumber \\
& & \times Tr[\gamma_{\mu}\Bigl(\frac{1}{\not{p}-\not{k}-m}\Bigr)^{i+1}
\gamma_{\nu}\Bigl(\frac{1}{ \not {p} -m}\Bigr)^{2q-i+1}]~,
\label{glse-c} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{eqnarray}
\Pi_{(D)\mu\nu}^{~~~ab}(k;q;\mu)& = &
-\frac{1}{2}ig_c^{2-2q}\mu^{2q}[-6ig_c^2C_2({\bf 8})]^q [f^{abe}f^{cde}
(g_{\mu\sigma}g_{\nu\rho}-g_{\mu\rho}g_{\nu\sigma}) \nonumber \\
& & +f^{ace}f^{dbe}(g_{\mu\rho}g_{\nu\sigma}-g_{\mu\nu}g_{\rho\sigma})+
f^{ade}f^{bce}(g_{\mu\nu}g_{\rho\sigma}-g_{\mu\sigma}g_{\nu\rho})] \nonumber \\
& & \times g^{\sigma\rho}\delta^{cd} \int\frac{d^4p}{(2\pi)^4}
\Bigl(\frac{-i}{p^2}\Bigr)^{q+1}~,
\label{glse-d} \end{eqnarray}
\addtocounter{abc}{1}
\renewcommand{\theequation}{\arabic{equation}}
where $N_f$ is the number of flavors, ${\bf T}^a$ are generators of $SU_c(3)$
in fundamental representation, $f^{abc}$ denote the structure constants of
$SU_c(3)$, $({\bf F}^a)^{bc}=-if^{abc}$ are generators of $SU_c(3)$ in adjoint
representation, and
$$
G_{\mu\nu\lambda}(p_1,p_2,p_3)~=~(p_1-p_2)_{\lambda}g_{\mu\nu}+(p_2-p_3)_{\mu}g_{\nu\lambda}+(p_3-p_1)_{\nu}g_{\lambda\mu}
$$
comes from the three-gluon vertex.
\bigskip
\noindent
2. \hspace{0.3cm}The integral expressions of the regularized diagrams
contributing to $\Gamma^{abc}_{\mu\nu\lambda} (k_1,~k_2)$:
\renewcommand{\theequation}{\arabic{equation}\alph{abc}}
\setcounter{abc}{0}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{\Gamma_{(A)\mu\nu\lambda}^{~~~abc}(k_1,~k_2;q;\mu) ~ = ~
ig_c^{3-2q} \mu^{2q}[-6ig_c^2C_2({\bf 8})]^q Tr({\bf F}^a{\bf F}^b{\bf F}^c)
\frac{1}{N_q}\sum_{i=0}^{q} \sum_{j=0}^{q-i} \int\frac{d^4p}{(2\pi)^4} } \nonumber
\\
{}~~~~~~~~~~~~~ \times
\displaystyle{g^{\rho_1\sigma_2} g^{\rho_2\sigma_3} g^{\rho_3\sigma_1}
G_{\nu\rho_2\sigma_2}(k_2,-k_2-p,p)
G_{\lambda\rho_3\sigma_3}(-k_1-k_2,k_1-p,k_2+p) } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{G_{\mu\rho_1\sigma_1}(k_1,-p,p-k_1)
\Bigl(\frac{-i}{p^2}\Bigr)^{i+1} \Bigl(\frac{-i}{(p+k_2)^2}\Bigr)^{j+1}
\Bigl(\frac{-i}{(p-k_1)^2}\Bigr)^{q-i-j+1} }~,
\end{array}
\label{3g-a} \end{equation}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(B)\mu\nu\lambda}^{~~~abc}(k_1,~k_2;q;\mu) ~ = ~ -\frac{1}{2}ig_c^{3-2q}
\mu^{2q}[-6ig_c^2C_2({\bf 8})]^q \frac{1}{N_q}\sum_{i=0}^{q}
\int\frac{d^4p}{(2\pi)^4}f^{as_1t_1} } \nonumber \\
{}~~~~~~~~~~~~~ \displaystyle{ G_{\mu\rho_1\sigma_1}(k_1,-p,p-k_1)
[f^{bcr}f^{t_2s_2r}(g_{\nu\sigma_2}g_{\lambda\rho_2}-g_{\nu\rho_2}g_{\lambda\sigma_2}) } \nonumber \\
{}~~~~~~~~~~~~~
\displaystyle{+f^{bt_2r}f^{s_2cr}(g_{\nu\rho_2}g_{\lambda\sigma_2}-g_{\nu\lambda}g_{\rho_2\sigma_2})+f^{bs_2r}f^{ct_2r}(g_{\nu\lambda}g_{\rho_2\sigma_2}-g_{\nu\sigma_2}g_{\lambda\rho_2})] } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{\delta^{s_1s_2} \delta^{t_1t_2} g^{\rho_1\rho_2}
g^{\sigma_1\sigma_2} \Bigl(\frac{-i}{p^2}\Bigr)^{i+1}
\Bigl(\frac{-i}{(p-k_1)^2}\Bigr)^{q-i+1} } \nonumber \\
{}~~~~~~~~~~~~~
+\{(\mu,a,k_1) \leftrightarrow (\nu,b,k_2)\}+\{(\mu,a,k_1)\leftrightarrow
(\lambda,c,k_3)\}~,
\end{array}
\label{3g-b} \end{equation}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(C)\mu\nu\lambda}^{~~~abc}(k_1,~k_2;q;\mu) ~ = ~ -ig_c^{3-2q}
\mu^{2q}[-ig_c^2C_2({\bf 8})]^q Tr({\bf F}^a{\bf F}^b{\bf F}^c)
\frac{1}{N_q}\sum_{i=0}^{q} \sum_{j=0}^{q-i} \int\frac{d^4p}{(2\pi)^4} }
\nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{p_{\mu}(p+k_2)_{\nu}(p-k_1)_{\lambda}
\Bigl(\frac{i}{p^2}\Bigr)^{i+1}\Bigl(\frac{i}{(p+k_2)^2}\Bigr)^{j+1}
\Bigl(\frac{i}{(p-k_1)^2}\Bigr)^{q-i-j+1} } \nonumber \\
{}~~~~~~~~~~~~~
+\{(\nu,b,k_2) \leftrightarrow (\lambda,c,k_3)\}~,
\end{array}
\label{3g-c} \end{equation}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(D)\mu\nu\lambda}^{~~~abc}(k_1,~k_2;q;\mu) ~ = ~ ig_c^{3}\mu^{2q}
(-i\lambda_q)^{2q} Tr({\bf T}^a{\bf T}^b{\bf T}^c) \frac{1}{N_q} \sum_{i=0}^{2q}
\sum_{j=0}^{2q-i} \int\frac{d^4p}{(2\pi)^4} }~~~~~~~~~~ \nonumber\\
{}~~~~~~~~~~~~~ \times
\displaystyle{Tr[ \gamma_{\mu}\Bigl(\frac{i}{\not{p}-\not{k}_1-m}\Bigr)^{i+1}
\gamma_{\lambda}\Bigl(\frac{i}{\not{p}+\not{k}_2-m}\Bigr)^{j+1}
\gamma_{\nu}\Bigl(\frac{i}{\not{p}-m}\Bigr)^{2q-i-j+1}] } \nonumber \\
{}~~~~~~~~~~~~~
+\{(\nu,b,k_2) \leftrightarrow (\lambda,c,k_3)\}~,
\end{array}
\label{3g-d} \end{equation}
\addtocounter{abc}{1}
\renewcommand{\theequation}{\arabic{equation}}
\bigskip
\noindent
3. \hspace{0.3cm}The integral expressions of the regularized diagrams
contributing to $\Gamma^{abcd}_{\mu\nu\lambda\tau}(k_1,~k_2,~k_3)$:
\renewcommand{\theequation}{\arabic{equation}\alph{abc}}
\setcounter{abc}{0}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(A)\mu\nu\lambda\tau}^{~~~~abcd}(k_1,~k_2,~k_3;q;\mu) ~ = ~ g_c^{4-2q}
\mu^{2q}[-6ig_c^2C_2({\bf 8})]^q Tr({\bf F}^a{\bf F}^b{\bf F}^c{\bf F}^d)
\frac{1}{N_q}\sum_{i=0}^{q} \sum_{j=0}^{q-i} \sum_{l=0}^{q-i-j} } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{
\int\frac{d^4p}{(2\pi)^4} G_{\mu\rho_1\sigma_1}(k_1,-p,p-k_1)
G_{\nu\rho_2\sigma_2}(k_2,-k_2-p,p) } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{ G_{\tau\rho_4\sigma_4}(-k_1-k_2-k_3,k_1-p,k_2+k_3+p) }
\nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{ G_{\lambda\rho_3\sigma_3}(k_3,-k_2-k_3-p,k_2+p)
g^{\rho_1\sigma_2} g^{\rho_2\sigma_3} g^{\rho_3\sigma_4} g^{\rho_4\sigma_1} }
\nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{
\Bigl(\frac{-i}{p^2}\Bigr)^{i+1} \Bigl(\frac{-i}{(p+k_2)^2}\Bigr)^{j+1}
\Bigl(\frac{-i}{(p+k_2+k_3)^2}\Bigr)^{l+1}\Bigl(\frac{-i}{(p-k_1)^2}\Bigr)^{q-i-j-l+1} } \nonumber \\
{}~~~~~~~~~~~~~
+\{(\nu,b,k_2) \rightarrow (\lambda,c,k_3),~(\lambda,c,k_3) \rightarrow
(\tau,d,k_4),~(\tau,d,k_4) \rightarrow (\nu,b,k_2)\} \nonumber \\
{}~~~~~~~~~~~~~
+\{(\nu,b,k_2) \rightarrow (\tau,d,k_4)),~(\tau,d,k_4) \rightarrow
(\lambda,c,k_3),~(\lambda,c,k_3) \rightarrow (\nu,b,k_2)\}~,
\end{array}
\label{4g-a} \end{equation}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(B)\mu\nu\lambda\tau}^{~~~~abcd}(k_1,~k_2,~k_3;q;\mu) ~ = ~ -\frac{1}{2}
g_c^{4-2q} \mu^{2q}[-6ig_c^2C_2({\bf 8})]^q \frac{1}{N_q}\sum_{i=0}^{q}
\int\frac{d^4p}{(2\pi)^4} } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{
[f^{adr_1}f^{t_1s_1r_1}(g_{\mu\sigma_1}g_{\tau\rho_1}-g_{\mu\rho_1}g_{\tau\sigma_1})+f^{at_1r_1}f^{s_1dr_1}(g_{\mu\rho_1}g_{\tau\sigma_1}-g_{\mu\tau}g_{\rho_1\sigma_1}) } \nonumber \\
{}~~~~~~~~~~~~~
\displaystyle{
+f^{as_1r_1}f^{dt_1r_1}(g_{\mu\tau}g_{\rho_1\sigma_1}-g_{\mu\sigma_1}g_{\tau\rho_1})][f^{bcr_2}f^{t_2s_2r_2}(g_{\nu\sigma_2} g_{\lambda\rho_2}- g_{\nu\rho_2}g_{\lambda\sigma_2}) } \nonumber \\
{}~~~~~~~~~~~~~
\displaystyle{
+f^{bt_2r_2}f^{s_2cr_2}(g_{\nu\rho_2}g_{\lambda\sigma_2}-g_{\nu\lambda}g_{\rho_2\sigma_2})+f^{bs_2r_2}f^{ct_2r_2}(g_{\nu\lambda}g_{\rho_2\sigma_2}-g_{\nu\sigma_2}g_{\lambda\rho_2})] } \nonumber\\
{}~~~~~~~~~~~~~ \times
\displaystyle{
\delta^{s_1s_2}\delta^{t_1t_2} g^{\rho_1\rho_2} g^{\sigma_1\sigma_2}
\Bigl(\frac{-i}{p^2}\Bigr)^{i+1} \Bigl(\frac{-i}{(p+k_2+k_3)^2}\Bigr)^{q-i+1} }
\nonumber \\
{}~~~~~~~~~~~~~
+\{(\tau,d,k_4) \leftrightarrow (\lambda,c,k_3)\}+\{(\tau,d,k_4)
\leftrightarrow (\nu,b,k_2)\}~,
\end{array}
\label{4g-b} \end{equation}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(C)\mu\nu\lambda\tau}^{~~~~abcd}(k_1,~k_2,~k_3;q;\mu) ~ = ~
-ig_c^{4-2q} \mu^{2q}[-6ig_c^2C_2({\bf 8})]^q \frac{1}{N_q}\sum_{i=0}^{q}
\sum_{j=0}^{q-i} \int\frac{d^4p}{(2\pi)^4} } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{ \delta^{s_1t_2}\delta^{s_2t_3}\delta^{s_3t_1}
[f^{adr}f^{t_1s_1r}(g_{\mu\sigma_1}g_{\tau\rho_1}-g_{\mu\rho_1}g_{\tau\sigma_1}) } \nonumber \\
{}~~~~~~~~~~~~~
\displaystyle{
+f^{at_1r}f^{s_1dr}(g_{\mu\rho_1}g_{\tau\sigma_1}-g_{\mu\tau}g_{\rho_1\sigma_1})
+f^{as_1r}f^{dt_1r}(g_{\mu\tau}g_{\rho_1\sigma_1}-g_{\mu\sigma_1}g_{\tau\rho_1})
] } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{ f^{bs_2t_2}f^{cs_3t_3} G_{\nu\rho_2\sigma_2}(k_2,-k_2-p,p)
G_{\lambda\rho_3\sigma_3}(k_3,-k_2-k_3-p,k_2+p) } \nonumber \\
{}~~~~~~~~~~~~~ \times
\displaystyle{ g^{\rho_1\sigma_2} g^{\rho_2\sigma_3} g^{\rho_3\sigma_1}
\Bigl(\frac{-i}{p^2}\Bigr)^{i+1} \Bigl(\frac{-i}{(p+k_2)^2}\Bigr)^{j+1}
\Bigl(\frac{-i}{(p+k_2+k_3)^2}\Bigr)^{l+1} } \nonumber \\
{}~~~~~~~~~~~~~
+\{(\tau,d,k_4) \leftrightarrow (\lambda,c,k_3)\}+\{(\tau,d,k_4)
\leftrightarrow (\nu,b,k_2)\} \nonumber \\
{}~~~~~~~~~~~~~
+\{(\mu,a,k_1) \leftrightarrow (\lambda,c,k_3)\}+\{(\mu,a,k_1) \leftrightarrow
(\nu,b,k_2)\} \nonumber \\
{}~~~~~~~~~~~~~
+\{(\mu,a,k_1) \leftrightarrow (\nu,b,k_2),~(\tau,d,k_4) \leftrightarrow
(\lambda,c,k_3)\}~,
\end{array}
\label{4g-c} \end{equation}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(D)\mu\nu\lambda\tau}^{~~~~abcd}(k_1,~k_2,~k_3;q;\mu) ~ = ~ -g_c^{4-2q}
\mu^{2q}[-6ig_c^2C_2({\bf 8})]^q Tr({\bf F}^a{\bf F}^b{\bf F}^c{\bf F}^d)
\frac{1}{N_q}\sum_{i=0}^{q} \sum_{j=0}^{q-i} \sum_{l=0}^{q-i-j} } \nonumber \\
{}~~~~~~~~~~~~ \times
\displaystyle{
\int\frac{d^4p}{(2\pi)^4} p_{\mu}(p+k_2)_{\nu}(p+k_2+k_3)_{\lambda}(p-k_1)_{\tau}
\Bigl(\frac{i}{p^2}\Bigr)^{i+1} } \nonumber\\
{}~~~~~~~~~~~~ \times
\displaystyle{
\Bigl(\frac{i}{(p+k_2)^2}\Bigr)^{j+1}
\Bigl(\frac{i}{(p+k_2+k_3)^2}\Bigr)^{l+1}
\Bigl(\frac{i}{(p-k_1)^2}\Bigr)^{q-i-j-l+1} } \nonumber \\
{}~~~~~~~~~~~~
+\{(\nu,b,k_2) \rightarrow (\lambda,c,k_3),~(\lambda,c,k_3) \rightarrow
(\tau,d,k_4),~(\tau,d,k_4) \rightarrow (\nu,b,k_2)\} \nonumber \\
{}~~~~~~~~~~~~
+\{(\nu,b,k_2) \rightarrow (\tau,d,k_4)),~(\tau,d,k_4) \rightarrow
(\lambda,c,k_3),~(\lambda,c,k_3) \rightarrow (\nu,b,k_2)\} \nonumber \\
{}~~~~~~~~~~~~
+\{(\nu,b,k_2) \leftrightarrow (\lambda,c,k_3)\}+\{(\nu,b,k_2) \leftrightarrow
(\tau,d,k_4)\}+\{(\lambda,c,k_3) \leftrightarrow (\tau,d,k_4)\},
\end{array}
\label{4g-d} \end{equation}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{equation}
\begin{array}{l}
\displaystyle{
\Gamma_{(E)\mu\nu\lambda\tau}^{~~~~abcd}(k_1,~k_2,~k_3;q;\mu) ~ = ~
-g_c^{4}\mu^{2q} (-i\lambda_q)^{2q} Tr({\bf T}^a{\bf T}^b{\bf T}^c{\bf T}^d)
\frac{1}{N_q}\sum_{i=0}^{2q} \sum_{j=0}^{2q-i} \sum_{l=0}^{2q-i-j} }
\nonumber\\
{}~~~~~~~~~~~~ \times
\displaystyle{ \int\frac{d^4p}{(2\pi)^4}
Tr[\gamma_{\mu}\Bigl(\frac{i}{\not{p}-\not{k}_1-m}\Bigr)^{i+1}
\gamma_{\tau}\Bigl(\frac{i}{\not{p}+\not{k}_2+\not{k}_3-m}\Bigr)^{j+1} }
\nonumber\\
{}~~~~~~~~~~~~ \times
\displaystyle{
\gamma_{\lambda}\Bigl(\frac{i}{\not{p}+\not{k}_2-m}\Bigr)^{l+1}
\gamma_{\nu}\Bigl(\frac{i}{\not{p}-m}\Bigr)^{2q-i-j-l+1}] } \nonumber \\
{}~~~~~~~~~~~~
+\{(\nu,b,k_2) \rightarrow (\lambda,c,k_3),~(\lambda,c,k_3) \rightarrow
(\tau,d,k_4),~(\tau,d,k_4) \rightarrow (\nu,b,k_2)\} \nonumber \\
{}~~~~~~~~~~~~
+\{(\nu,b,k_2) \rightarrow (\tau,d,k_4)),~(\tau,d,k_4) \rightarrow
(\lambda,c,k_3),~(\lambda,c,k_3) \rightarrow (\nu,b,k_2)\} \nonumber \\
{}~~~~~~~~~~~~
+\{(\nu,b,k_2) \leftrightarrow (\lambda,c,k_3)\}+\{(\nu,b,k_2) \leftrightarrow
(\tau,d,k_4)\}+\{(\lambda,c,k_3) \leftrightarrow (\tau,d,k_4)\}~,
\end{array}
\label{4g-e} \end{equation}
\addtocounter{abc}{1}
\renewcommand{\theequation}{\arabic{equation}}
\bigskip
\noindent
4. \hspace{0.3cm}The integral expressions of the regularized diagram
contributing to $\tilde{\Pi}^{ab}(p)$:
\begin{eqnarray}
\tilde{\Pi}^{ab}(p;q;\mu) & = & -g_c^{2-2q} \mu^{2q}[-ig_c^2C_2({\bf 8})]^q
Tr({\bf F}^a{\bf F}^b) \int\frac{d^4k}{(2\pi)^4} p\cdot k
\Bigl(\frac{i}{k^2}\Bigr)^{q+1} \Bigl(\frac{-i}{(k-p)^2}\Bigr)~,
\label{ghse} \end{eqnarray}
\bigskip
\noindent
5. \hspace{0.3cm}The integral expressions of the regularized diagrams
contributing to $\tilde{\Gamma}^{abc}_{\mu}(p^{\prime},p)$:
\renewcommand{\theequation}{\arabic{equation}\alph{abc}}
\setcounter{abc}{0}
\addtocounter{abc}{1}
\begin{eqnarray}
\tilde{\Gamma}^{~abc}_{(A)\mu}(p^{\prime},p;q;\mu) & = & ig_c^{3-2q}
\mu^{2q}[-ig_c^2C_2({\bf 8})]^q Tr({\bf F}^a{\bf F}^b{\bf F}^c)
\int\frac{d^4k}{(2\pi)^4} k^{\sigma}p^{\prime\rho} \nonumber \\
& & \times
G_{\mu\rho\sigma}(k_1-k_2,p-k_1,k_2-p)
\Bigl(\frac{i}{k^2}\Bigr)^{q+1}
\Bigl(\frac{-i}{(k-p^{\prime})^2}\Bigr)\Bigl(\frac{-i}{(k-p)^2}\Bigr)~,
\label{glgh-a} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{eqnarray}
\tilde{\Gamma}^{~abc}_{(B)\mu}(p^{\prime},p;q;\mu) & = & ig_c^{3-2q}
\mu^{2q}[-6ig_c^2C_2({\bf 8})]^q Tr({\bf F}^a{\bf F}^b{\bf F}^c)
\displaystyle{\int\frac{d^4k}{(2\pi)^4}} (k+p^{\prime})_{\mu}(k+p)\cdot
p^{\prime} \nonumber \\
& & \times
\displaystyle{\Bigl(\frac{-i}{k^2}\Bigr)^{q+1}
\Bigl(\frac{i}{(k+p^{\prime})^2}\Bigr)\Bigl(\frac{i}{(k+p)^2}\Bigr)}~,
\label{glgh-b} \end{eqnarray}
\addtocounter{abc}{1}
\renewcommand{\theequation}{\arabic{equation}}
\bigskip
\noindent
6. \hspace{0.3cm}The integral expressions of the regularized diagram
contributing to $\Sigma(p)$:
\begin{eqnarray}
\Sigma(p;q;\mu) & = & {\bf T}^a{\bf T}^a \Sigma^{QED}(p;q;\mu) \nonumber \\
& = &
-g_c^{2-2q} \mu^{2q}[-6ig_c^2C_2({\bf 8})]^q {\bf T}^a{\bf T}^a
\int \frac{d^4 k}{(2\pi)^4}
(\gamma^{\mu}\frac{i}{ \not{k}-m}\gamma_{\mu})
\Bigl(\frac{-i}{(k-p)^2}\Bigr)^{q+1},
\label{qse} \end{eqnarray}
\bigskip
\noindent
7. \hspace{0.3cm}The integral expressions of the regularized diagrams
contributing to $\Gamma^{a}_{\mu}(p^{\prime},p)$:
\renewcommand{\theequation}{\arabic{equation}\alph{abc}}
\setcounter{abc}{0}
\addtocounter{abc}{1}
\begin{eqnarray}
\Gamma^{~~~a}_{(A)\mu}(p^{\prime},p;q;\mu) & = & -g_c^{3-2q} \mu^{2q}
[-6ig_c^2C_2({\bf 8})]^q {\bf T}^s{\bf T}^t \frac{1}{N_q}\sum_{i=0}^{q}
\int\frac{d^4k}{(2\pi)^4} \nonumber \\
& & \times
f^{ast} G_{\mu\rho\sigma}(p^{\prime}-p,p-k,k-p^{\prime}) \nonumber \\
& & \times
(\gamma^{\rho}\frac{i}{ \not{k}-m}\gamma^{\sigma})
\Bigl(\frac{-i}{(k-p)^2}\Bigr)^{i+1}\Bigl(\frac{-i}{(k-p^{\prime})^2}\Bigr)^{q-i+1},
\label{glq-a} \end{eqnarray}
\addtocounter{equation}{-1}
\addtocounter{abc}{1}
\begin{eqnarray}
\Gamma^{~~~a}_{(B)\mu}(p^{\prime},p;q;\mu) & = & {\bf T}^b{\bf T}^a{\bf T}^b
\Gamma^{QED}(p^{\prime},p;q;\mu) \nonumber \\
& = &
-ig_c^{3-2q} \mu^{2q}[-6ig_c^2C_2({\bf 8})]^q {\bf T}^b{\bf T}^a{\bf T}^b
\int \frac{d^4 k}{(2\pi)^4} \nonumber \\
& & \times
(\gamma^{\rho}\frac{i}{\not{k}+\not{p^{\prime}}-m}
\gamma_{\mu}\frac{1}{\not{k}+\not{p}-m}
\gamma_{\rho}) \Bigl(\frac{-i}{k^2}\Bigr)^{q+1},
\label{glq-b} \end{eqnarray}
\addtocounter{abc}{1}
\renewcommand{\theequation}{\arabic{equation}}
\bigskip
\bigskip
{\raggedright{\Large \bf Acknowledgment}}
\vskip 0.5cm
\noindent
{The work is supported in part by the National Natural Science Foundation of
China. One of the author (YC) is also supported in part by Local Natural
Science Foundation of Xinjiang.}
\newpage
\parindent 0 pt
|
\section{Introduction}
The nature of the glass transition is still poorly understood
\cite{mode,glass}. Under
slow cooling real glasses reach a metastable phase of free energy
larger than that of the crystal phase.
Glasses show a strong slowing down of the
dynamics when the temperature is lowered and the transport coefficients
increase by several orders of magnitude in a narrow range of
temperatures. It is natural to think that the appareance of high
free-energy barriers is the mechanism responsible for the glass
transition. But free-energy barriers are composed of energy barriers
and entropy barriers.
The question about the relevance of both kind of barriers
in real glasses is of
the outmost importance. Activated jumps of energy-barriers
are
strongly dependent on temperature.
The typical time
$\tau$ to overcome an energy barrier $\De E$ is
$\tau\sim\exp(\frac{\De E}{T})$ where $T$ is the temperature. This
typical time diverges when the temperature $T$ goes to zero.
Conversely, relaxation times related to entropy barriers do not depend
directly on the
temperature.
The simplest way to visualize entropy barriers is the following,
Consider a dynamics in which at each time step the system
can reach a new state with uniform probability;
the typical time to decrease the energy of one unity is $\tau\sim
\frac{\Omega_i}{\Omega_f}=\exp(\Delta S)$ where $\Delta S$
is the height of the entropy barrier ($\Omega_i$ stands for the initial
available volume of phase space and $\Omega_f$ stands for the final
volume of phase space with lower energy).
If the effect of energy and entropy barriers is combined one
expects that entropy barriers should affect the temperature activated
relaxation time in its prefactor $\tau\sim
\frac{\Omega_i}{\Omega_f}\exp(\frac{\De E}{T})$.
According to that, the relaxation
time can diverge independently
of the temperature if the phase space volume of lower energy
configurations in the system shrinks to zero during the dynamical
evolution. The idea that
an entropy crisis could be relevant to the glassy transition is very
old \cite{kauzmann,adam}, and it has had interesting
developements in recent times \cite{kirtirwo,par} in the framework
of mean-field theory of disordered systems. However in the models
studied in \cite{kirtirwo,par}
it is very difficult to disentangle
entropic effects from energetic ones.
To this aim a simple model (the backgammon -BG- model) was recently proposed
by one of us \cite{I} (hereafter referred as I), in
which energy barriers are completely absent (a diffusive model
with entropy barriers has also been considered in \cite{BM}).
While the model has no thermodynamic transition, it shows a slow dynamics
with strong hysteresis effects and Arrhenius behavior of the
relaxation time. The off-equilibrium dynamics of this model was
studied subsequently by us
\cite{II}
(hereafter referred
as II) using an adiabatic approximation, obtaining fair
good results concerning the relaxation of the energy.
The same approximation has been recently rederived, and slightly refined,
in a
paper by Bouchaud, Godreche and M\'ezard \cite{BMG}.
In this paper we derive the exact mean-field equation
for the order parameter
for the dynamics of the BG
model, which turns out to be the energy itself.
The techniques
we use are similar to these of \cite{BMG}, however,
the equations we get were not discussed there.
We find that the energy verifies a causal
functional equation with memory. This is at variance with the approximate
treatments where the evolution is described by a Markovian
equation.
In its original formulation, the model does not have any energy
barriers. However, in real systems energy
barriers are present.
The BG model is flexible enough to allow for the introduction (and the
tuning) of energy barriers. This is done by simple modifications
of the Hamiltonian of the system. The formalism developed for the
original model applies in these cases.
In the second section we define the Hamiltonian of the BG model and
the Monte Carlo dynamics we have used to study it. In the third
section we present some exact results for the behavior of the one-time
quantities (for instance, the energy) and for the two-time quantities
like the correlation function. The fourth section is devoted to the
study of the effect of energy barriers in the BG model. Finally we
present the conclusions and a discussion of the results.
\section{The BG model and the dynamics}
Let us take $N$ distinguishable particles which can occupy $M$ different
states and let us denote by $\rho=\frac{N}{M}$ the density, i.e. the
number of particles per state. The BG Hamiltonian is defined by,
\be
H=-\sum_{r=1}^M\,\de_{n_r,0}
\label{eq1}
\ee
where $n_r$ is the occupation level of the state $r=1,...,M$, i.e. the
number of particles which occupy that state. The numbers $n_r$ satisfy
the global constraint,
\be
\sum_{r=1}^M\,n_r= N~~~~.
\label{eq2}
\ee
Eq.(\ref{eq1}) shows that energy is simply given by the number of empty
states (with negative sign). We define the occupation probabilities,
\be
P_k=\frac{1}{M}\,\sum_{r=1}^M\,<\de_{n_r,k}>
\label{eq3}
\ee
which is the probability of finding one state occupied by $k$ particles.
The statics of this model in the canonical
ensemble can be easily solved (see (I) and (II)). In
particular one gets the result,
\be
P_k=\rho \frac{z^{k-1} \exp(\beta \de_{k,0})}{k!\exp(z)}
\label{eq4}
\ee
where $z$ is the fugacity and $\beta$ is the inverse of the temperature
$T$ and they are related by the condition,
\be
\rho(e^{\beta}-1)=(z-\rho)e^z~~~~.
\label{eq5}
\ee
expressing that the density is fixed to $\rho$.
The probabilities $P_k$ satisfy the relation $\sum_{k=0}^{\infty}P_k=1$
and they yield all the static observables, in particular the energy
$U=-P_0$.
Several dynamical rules, thermalizing to the Boltzmann distribution,
can be attached to the model. The simplest choice (I) is the
Metropolis single particle dynamics, in which
at each sweep a particle is chosen at random
and it is proposed a move to a new state.
The move is accepted with probability one
if the energy does not increase and with probability
$\exp(-\beta)$ otherwise.
In the mean-field version of the model, the possible
arrival states of the
particles are chosen at random with uniform probability in all the space.
This random motion of the particles allows a complete analytical treatment
of the problem \footnote{The interesting case of a sequential dynamics is more
complicated.}.
Finite dimensional models, where at each sweep
the particles are only allowed
to move to neighbours on a lattice are currently
under study \cite{rito_folla}.
The model has no energy barriers. Consequently there is no
frustration (in the usual sense) and no metastable states. However it was
shown in (I) that the dynamics is highly non
trivial and a dramatic slowing
down occurs at low temperatures.
This can be qualitatively understood as follows.
Suppose the system is at zero temperature and the dynamics starts
from a random initial configuration of high energy. As the system evolves
towards the equilibrium
more and more states are progressively emptied and the energy decreases.
Because the average number of particles per occupied state
increases with time (the total number of particles is conserved)
the time needed to empty one more state
also increases.
The result is that the energy
goes extremely slowly to its equilibrium value.
The dynamical quantities we are interested in are the time-dependent
occupation number probabilities
\be
P_k(t)=\frac{1}{M}\,\sum_{r=1}^M\,\langle\de_{n_r(t),k}\rangle
\label{eq6}
\ee
($E(t)=-P_0(t)$)
and the two time energy-energy correlation function
\cite{I},
\bea
C_E(t,s)=& &\frac{\frac{1}{M}\sum_r\delta_{n_r(t),0}\delta_{n_r(s),0}
-E(t)E(s)}{-E(s)(1+E(s))}
\nonumber\\
\equiv & &
{P(n_r(t)=0,n_r(s)=0)-P_0(t)P_0(s)\over P_0(s)[1-P_0(s)]}
\,\,\,\,t\ge s
\label{eq7}
\eea
At finite temperature, when $t,s>>t_{eq}\sim \exp(\beta)/\beta^2$ (see
reference (II) and also section III.B) this function is time
translationally invariant. In the regime in
which both times
$t,s$ are much less than $t_{eq}$, and at all times
at zero temperature, the system is off-equilibrium,
time-translation invariance does not hold, and the correlation function
displays aging (see \cite{I}).
\section{Mean-field equations for the dynamics of the BG model}
In this section we derive exact mean-field
equations for the Monte Carlo
dynamics of the BG model. First we
adress the dynamical problem associated to the one-time probability
distributions $P_k(t)$. These probabilities generate an
infinite hierarchy of Markovian equations which can be closed
in terms of the only quantity $P_0(t)$.
Then we will study the
two-time correlation functions in a
similar way. For simplicity, we will
restrict all the future computations to the case $\rho=1$ (i.e $M=N$),
the generalization to an arbitrary density being very simple.
\subsection{Dynamical equations for $P_k(t)$}
The purpose of
this section is to write the dynamical evolution equations
for the probabilities $P_k(t)$ and, in particular, for the internal
energy $E(t)=-P_0(t)$. An elementary
Monte Carlo move consists in a random selection of one particle (hence, the
probability to select a particular departure state $d$
is $n_d/N$ where
$n_d$ is the occupation level of that state) and moving it to a randomly
selected arrival state $a$ with uniform
probability independent of the occupation level $n_a$. One Monte carlo
step (our unity of time) consists of $N$ elementary moves.
In the elementary move there are several processes
which contribute to the variation of $P_k(t)$.
In appendix A we write explicitly the balance equations,
the result is:
\be
{{\rm d} P_k(t)\over {\rm d} t}=
(k+1)(P_{k+1}-P_{k})+P_{k-1}+
P_0(e^{-\beta}-1)(\de_{k,1} - \de_{k,0} - kP_k + (k+1)P_{k+1})
\label{eq10}
\ee
where the time index for the probabilities $P_k$ has been omitted.
The previous equation holds for $k\ge 0$ with $P_{-1}=0$.
In particular for $k=0$ we obtain the equation studied in (II),
\be
\frac{\partial P_0}{\partial t}=P_1(1-P_0) - e^{-\beta} P_0 (1-P_1)
\label{eq11}
\ee
The hierarchy was closed in (II) assuming fast relaxation on the surfaces
of constant energy, and slow variation of the energy itself. In this
condition eq.(\ref{eq11}) was solved assuming
$P_k(t)={\exp[{\beta(t)\delta_{k,0}-z(t)}]}{z(t)^{k-1}\over k!}$ with $\beta(t)$ and
$z(t)$ related at all times by eq.(\ref{eq5}).
Here we study the hyerarchy (\ref{eq11}) with the method of the generating
function, that we define as
\be
G(x,t)=\sum_{k=0}^{\infty}\,x^k\,P_k(t)
\label{eq12}
\ee
A similar approach was also used in \cite{BMG} where the adiabatic
approximation of (II) was rederived and improved\footnote{The
technique of the generating function in the study of the dynamics
has also been applied
to some mean-field spin glass models \cite{BPPR}.}.
{}From the equation (\ref{eq10}) it is easy to check that the $G(x,t)$
satisfies the partial differential equation,
\be
\frac{\partial G(x,t)}{\partial t}=(x-1)[G(x,t) + \la(t)
- (1+\la(t))\frac{\partial
G(x,t)}{\partial x}]
\label{eq14b}
\ee
with $\la(t)=P_0(t) (e^{-\beta}-1)$.
Eq. (\ref{eq14b}) is a non linear partial differential
equation, the nonlinearity
is contained
in the dependence of $\lambda$ on $P_0(t)=G(0,t)$.
The equilibrium solution $G_{eq}(x)$ is easily obtained
from equations (\ref{eq4}) and (\ref{eq12}),
\be
G_{eq}(x)=\frac{e^{\beta}-1+e^{zx}}{z\,e^{z}}
\label{eq15b}
\ee
and one can check that this is consistent with eq.(\ref{eq14b}).
The previous partial differential equation can be implicitly
solved to get $G(x,t)$ as a functional of $\lambda$.
The details are presented in the Appendix B, we give here the result
\be
G(x,t)={\rm e}^{(x-1)D(t,0)}G_0(1+(x-1)B(t,0))+
(x-1)\int_0^t {\rm d} s \lambda(s) B(t,s) \ {\rm e}^{(x-1)D(t,s)}
\label{aa}\ee
where we have written
\begin{eqnarray}
B(t,s)&=&\exp\left({-\int_s^t {\rm d} v \ [1+\lambda(v)]}\right)
\nonumber\\
D(t,s)&=&{\int_s^t {\rm d} v \ B(t,v)}
\end{eqnarray}
and $G_0(x)=G(x,0)$ is the initial condition at time $t=0$.
Setting $x=0$ in (\ref{aa}) we get a closed equation for
$P_0(t)$ which reads
\be
P_0(t)=
{\rm e}^{-D(t,0)}G_0(1-B(t,0))
+(1-e^{-\beta})\int_0^t {\rm d} s P_0(s) B(t,s) \ {\rm e}^{D(t,s)}
\label{bb}
\ee
The previous equation, although
strongly non-markovian is {\it causal}, as the l.h.s. depends on
the values of $P_0(s)$ for $s\le t$. It has a unique solution
that can be found numerically with good
precision, discretizing the time and integrating
it step by step.
The evaluation of the
previous expressions gives the full solution of the BG model as far as the
one-time dynamical quantities are concerned.
The solution of (\ref{aa}) is explicit at infinite temperature ($\beta=0$).
In this case $\la(t)=0$ and the solution of $G(x,t)$ simplifies,
\be
G(x,t)=e^{(1-e^{-t})(x-1)}\,G_0((x-1)e^{-t}+1)
\label{cc}
\ee
It is not surprising that
at infinite temperature the system goes exponentially fast to the
equilibrium (with
relaxation time equal to 1).
At infinite temperature the equilibrium probabilities eq.(\ref{eq4}) are
given by $P_k=\frac{1}{k!e}$, the energy being $E=-P_0=-\frac{1}{e}$. If
we start from the initial condition in which all particles occupy the same
state ($P_0=1,\,P_k=0,k>0$) then we have $G_0(x)=1$. From eq.(\ref{cc})
we obtain the time evolution of the energy,
\be
E(t)=-G(0,t)=-e^{e^{-t}-1}
\label{dd}
\ee
We studied numerically the solution of (\ref{bb}) at $T=0$. In
figure 1 we display the result for the energy, starting from
the initial condition $P_k(0)=1/(e k!)$ at time 0 (i.e. $G_0(x)=e^{x-1}$).
For comparison
we plot the results of the Monte Carlo simulations and of the
adiabatic hypothesis of (II) with the same initial
condition.\footnote{The adiabatic hypothesis gives better results if
the integration is started at latter times.}
\subsection{The Correlation Function $C_E(t,s)$}
In this section we investigate the behavior of the
energy-energy correlation functions (\ref{eq7}).
We proceed in a similar
way as we have done for the occupation probabilities.
We need to study the joint occupation
probability in a given site $r$ at two different times
$t,s$ ($t>s$), $P(n_r(t)=0,n_r(s)=0)=P(n_r(t)=0|n_r(s)=0)P_0(s)$.
The correlation function eq. (\ref{eq7}) can be written as
\be
C_E(t,s)={P(n_r(t)=0|n_r(s)= 0)-P_0(t)\over 1- P_0(s)}.
\ee
We now write a set of equations that allow to study
$P(n_r(t)=0|n_r(s)= 0)$.
Let us define the probabilities
\be
\nu_k(t,s)=P(n_r(t)=k|n_r(s)=0)
\ee
i.e. the occupation number probabilities in the set
$S_s$ of states which are empty at time $s$.
In general, it is possible
to restrict the balance equations that led to (\ref{eq10})
to any subset $S$ of the whole space.
Irrespectively of $S$ the result is:
\begin{eqnarray}
\frac{\partial \nu_k}{\partial t}=
& &\nu_{k-1}-\nu_{k}+[(k+1)\nu_{k+1}-k\nu_k][1-P_0(1-e^{-\beta})]
\nonumber\\
& &-(\de_{k,1}-\delta_{k,0})[\nu_0(1-P_1)+\nu_1 P_0](1-e^{-\beta}).
\label{nuk}
\end{eqnarray}
In particular if the set $S$ is the whole space $\nu_k=P_k$ and
we get back to (\ref{eq10}).
Of course the initial
conditions depend on the set under study.
For the set $S_s$ we are interested to, we must choose
\be
\nu_k(s,s)=\delta_{k,0}.
\ee
In terms of the generating function
\be
\G(x,t,s)=\sum_{k=0}^\infty x^k \nu_k(t,s)
\ee
eq.(\ref{nuk}) reads
\be
\frac{\partial \G}{ \partial t}=
(x-1)[\G-(1-P_0(1-e^{-\beta}))\frac{\partial \G}{ \partial x}
-(\nu_0(1-P_1)+\nu_1 P_0)(1-e^{-\beta})]
\label{gamma}\ee
with condition at time $s$
\be
\G_s(x)\equiv \G(x,s,s)=1.
\ee
Note that if we suppose to know the $P_k(t)$ then the
eq.(\ref{nuk},\ref{gamma})
are linear. Obviously, if one considers the
set $\overline{S}_s$ complementary to $S_s$, and its respective generating
function $\overline{\Gamma}$, it holds the equality:
$P_0(s) \G(x,t,s)+(1-P_0(s))\overline{\G}(x,t,s)=G(x,t,s)$.
Defining
\begin{eqnarray}
& &\mu(t,s)=[\nu_0(t,s)(1-P_1(t))+\nu_1(t,s)P_0(t)](1-{\rm e}^{-\beta})
\nonumber\\
& &B(t,s)=\exp\left({-\int_s^t {\rm d} v \ (1-P_0(t)(1-{\rm e}^{-\beta})) }\right)
\nonumber\\
& &D(t,s)=\int_s^t {\rm d} v \ B(t,v)
\end{eqnarray}
we find
\be
\G(x,t,s)=
{\rm e}^{(x-1)D(t,s)}-
(x-1)\int_s^t {\rm d} u \mu(u,s) B(t,u) \ {\rm e}^{(x-1)D(t,u)}
\label{bbb}\ee
which depends implicitly on $\nu_0$ and $\nu_1$.
In order to find a closed
system we have to consider eq.(\ref{bbb}) and its $x$-derivative in
$x=0$
\begin{eqnarray}
& &\nu_0(t,s)=1+ \int_s^t {\rm d} u \ \left[
-\nu_0(u,s)[1-(1-P_1(u))(1-{\rm e}^{-\beta})]+\nu_1(u,s)\right]
\nonumber\\
& &
\nu_1(t,s)= \int_s^t {\rm d} u \
\left[
\mu(u,s) B(t,u){\rm e}^{-D(t,u)}(D(t,u)-1)
\right]
+{\rm e}^{-D(t,s)}D(t,s)
\label{volt}
\end{eqnarray}
The system (\ref{volt}), if one assumes known the probabilities
$P_k(t)$, consists in a vectorial
linear Volterra equation of second kind for $\nu_0$ and $\nu_1$
which can in all generality be
integrated numerically,
and in some particular case also analytically.
The simplest case is the equilibrium
at finite temperature. In that case, $P_k\equiv P_k^{eq}$ and
the various functions appearing in (\ref{volt}) are time traslation
invariant. In these conditions eq.(\ref{volt}) can be
solved in Laplace transform. Simple algebra, and the
formula (see e.g. \cite{tricomi})
\be
\int_0^\infty d t \ exp(-a \exp(t)-E)=a^{-E} \ \gamma(p,a)
\ee
($\gamma$ is the incomplete gamma function),
shows that $\nu_0(E)$, the Laplace transform of $\nu_0(t-s)$, is given
by
\be
\nu_0(E)={A(E)+{z-1\over z}(1-E A(E))\over 1-\left[{(z-1)e^z\over (z-1)e^z +1}
+{z-1\over z}E\right](1-E A(E))}
\label{lap}
\ee
where we have expressed all the equilibrium quantities in terms
of the fugacity $z$ (see eq. (\ref{eq5})),
$E$ is the Laplace variable conjugated to time, and
\be
A(E)={1\over e^z z^{zE-1}}\int_0^z u^{zE-1}e^u.
\ee
$\nu_0(E)$, as it should, has a pole in $E=0$ with residue $P_0$ coresponding
to $\nu_0(t)\to P_0$ for large time. Poles on the real negative $E$ axis
correspond to exponential relaxation modes. The largest relaxation time
is given by minus the inverse of the value of $E$ in the pole closest
to the origin. This can be obtained explicitely for large $\beta$, where
$z\approx \beta-\log(\beta)$ is large, from the asympotic expantion of
$E A(E)$ for small $E$
\be
E A(E)\approx
e^{-z}+E.
\ee
The result is simply $E_{pole}\approx -e^{-z}$ and correspondingly
$\tau_{max}\approx e^z\approx exp(\beta)/ \beta$.
In the off-equilibrium regime the integration of (\ref{volt}) can be
performed
numerically.
In fig. 2 we show the result of the integration for $T=0$ for different
values of $s$ ( i.e. different waiting times) compared to the
Monte Carlo results.
Although we did not try very sophisticated algorithms, with standard
ones \cite{num_rec}, we were able to reach enough large times to
see the scaling
behavior $C_E(t,s)=f((t-s)/s)$ observed numerically in I.
It would be interesting to see if equation (\ref{bbb}) could be solved
with the aid of some simple approximation as it is the case
for the energy (II and \cite{BMG}).
\section{The effect of energy barriers}
The BG model has no energy barriers and hence there is no finite
temperature thermodynamic phase transition. In real glasses energy
barriers are always present and it can be instructive to understand
their effect when combined with entropy barriers. One can easily
modify the Hamiltonian (\ref{eq1}) to include
energy barriers. In this paper we have focused
on two different ways. In the first, we have
considered interaction
between the different states, introducing an
energy gain when groups of states are
simultaneously empty. This interaction term is enough to make
appear a finite temperature thermodynamic transition, but
metastability and frustration are absent and the system monotonically
reaches the ground state at zero temperature. In the other,
we introduced metastable
configurations in the dynamics. In this case the system fails to reach
the ground state at zero temperature while thermodinamically
there is no finite-temperature phase transition.
\subsection{The $p$-states model}
The simplest way we can introduce interaction among
different states in the model is the following, consider the quantity
\be
M[\{ n_r\}] ={1\over N}\sum_{r=1}^N (\delta_{n_r,0}-1/e).
\ee
Any Hamiltonian of the form
\be
H=N F(M[\{ n_r\}])
\label{hf}
\ee
with $F$ gentle enough, is a good candidate for a mean-field model.
We did not try a systematic study of the form (\ref{hf}) for generic
$F$, but we concentrated to the class of monomials, where
\be
H_p=-\frac{1}{N^{p-1}}\left(\sum_{r=1}^N(\de_{n_r,0}-{1}/{e})\right)^p
\label{23}
\ee
For $p=1$ this model reduces to the BG model.
For larger values of $p$ there is
interaction between different states. The ground
state of this model is the same as the one of
the BG model (all particles
occupying the same state) and there are no energy barriers at zero
temperature.
A careful study of the thermodynamics of this model shows that for any
$p>1$ there is a first order phase transition from a
completely disordered phase with $M=0$ for $T>T_c$ to an `ordered' phase
with $M\ne 0$ for $T<T_c$. This leads to the curious situation that the
completely disordered state is dynamically stable at all temperatures
but at $T=0$. This can be understood by a simple argument.
Suppose to start the dynamics in a random initial condition
and consider a sweep that lead to the filling of an empty
state. The energy variation in this process is
\be
\delta H=-{1\over N^{p-1}} \left[ \left(\sum_r(\delta_{n_r,0}-{1\over e})-1 \right)^p
-\left(\sum_r(\delta_{n_r,0}-{1\over e}) \right)^p \right]\simeq
{1\over N^{p-1}} p \left(\sum_r(\delta_{n_r,0}-{1\over e}) \right)^{p-1}.
\ee
But according to our hypothesis $\sum_r(\delta_{n_r,0}-{1\over e})$
is a random variable of order
$\sqrt{N}$, correspondingly $\delta H\sim N^{-(p-1)/2}$ and the acceptance
rate
\be
{\rm e}^{-\beta \delta H}\sim {\rm e}^{-\beta /N^{(p-1)/2}}
\ee
For finite temperature and large $N$ all the moves are accepted and the
energy in average never decreases. In other words
the
statistic of configurations is not changed by the dynamics.
A crossover is found for $\beta\sim N^{(p-1)/2}$ showing that the relevant
scale of temperature for the dynamics is different
from that of the statics. Right at zero temperature, where
only the sign of the energy change and not the magnitude
matters, the dynamics
of the model coincides for any $p$ with the one of the conventional
case $p=1$.
\subsection{The effect of metastability}
The $p$-states model has no metastability at zero temperature. We
want to study here a simple model where metastability is present but without
interaction. In the BG model the ground state is reached by emptying
progressively more and more states. To empty a given
state at a certain time
$t$ it is necessary
to pass in a configuration where a unique particle occupies that state.
We then consider the following model,
\be
H=\sum_{r=1}^N\,(-\de_{n_r,0}\,+\,g\de_{n_r,1})
\label{30}
\ee
where $g$ is positive constant and we have the usual constraint
eq.(\ref{eq2}). At zero temperature
the transition $n_r=2 \ \to \ n_r=1$ is forbidden, hence
energy barriers are present in the model.
More general models
are obtained including in the Hamiltonian all possible terms of the type
$\de_{n_r,k}$
\be
H=-\sum_{r=1}^N\,\sum_{k=0}^{\infty}g_k\de_{n_r,k}=-\sum_{r=1}^N\,g_{n_r}
\label{31}
\ee
We focus here to the case (\ref{30}).
The statics of this model is easily solved. We obtain the free energy,
\be
\beta f=\log(z)-\log(e^z+e^{\beta}-1+z(e^{-\beta g}-1))
\label{32}
\ee
and the fugacity is related to the temperature $\frac{1}{\beta}$ by the
$g$-independent relation eq.(\ref{eq5}). The equilibrium probabilities
$P_k$ (see eq.(\ref{eq3})) are given by,
\be
P_k=\frac{z^{k-1} \exp(\beta\de_{k,0}-\beta g \de_{k,1})}{k!
(e^z+e^{-\beta g}-1)}
\label{33}
\ee
The dynamics of this model is expected to be substantially different to
that of the BG model at least at very low temperatures. Concretely, at
zero temperature the ground state is the same as for the BG case but
there is a large number of metastable states (for instance, half states
empty and half of the states with two particles). It is easy to show
that for each value of $E$ between $E=-1/2$ and the ground state $E=-1$
there exists a metastable configuration with that energy. Then we expect
the value of the energy extrapolated to infinite time to depend strongly
on the initial configuration. In order to minimize the energy we have to
maximize $P_0$ and minimize $P_1$. While the maximization of $P_0$ is a
process where entropy barriers are dominant (this is the reason why the
BG model defined as $E=-P_0$ is interesting) this is not the case for
minimizing $P_1$ where entropy barriers are absent. Then, independently
of the initial configuration we expect that $P_1$ will go to zero in the
exponentially fast for large times. In these conditions, we do not
expect
that the adiabatic solution of (II) can give a good approximation
of the dynamics. This
approximation was based on the
fact that in the BG the surfaces of constant energy are connected,
a situation which does not hold here.
However the dynamics of this model
can be directly solved as in the BG case.
Skipping all the details we find for the generating function:
\bea
\frac{\partial G(x,t)}{\partial t}=(x-1)\Bigl(-(1+\la(t))\frac{\partial
G(x,t)}{\partial x} + (1+2 P_2(e^{-\beta g}-1))G(x,t)+\\
\la(t)-2 P_2 (e^{-\beta g}-1)(e^x-P_0 (x-1)(1-e^{-\beta
(1+g)})\Bigr )
\label{36}
\eea
where $\la(t)=P_0(t) (e^{-\beta(1+g)}-1)$. Observe
that eq.(\ref{36})
depends only on the probabilities
$P_0(t)$ and $P_2(t)$. The solution is
more complicated to that of eq.(\ref{aa}), however it can be found.
In figure 3 we show the result of the numerical
integration of (\ref{36}) for the energy at $T=0$ compared with the
Monte Carlo simulations, starting from a completely random initial
condition. The energy seems to converge to a value $\lim_{t\to\infty}E(t)=
-.564$, a result that it would be interesting to derive analytically.
It is under current study the finite temperature
dynamics, where we expect the effect of the energy and entropy
barriers to combine to give rise to a dynamics slower than that
of the BG model.
\section{Conclusions}
In this paper we have derived the exact mean-field equations
of the dynamics of the Backgammon model. This has been achieved
through the study of the single site occupation number probability
for which a hierarchical set of equations hold. With the method of the
characteristic function, we have derived a closed functional
equation for the energy. This, although non-Markovian,
has a causal character and can be integrated
step by step discretizing the time. The non-Markovian character of the
evolution equation, suggests that history dependent effects should
be observable in the system. However the analysis of II, where
the evolution of the energy was described by an approximate equation,
shows that even in subtle phenomena as hysteresis cycles
in cooling-heating processes, history dependent effects are very small.
This should reflect in the fact that the memory kernels that
appear in the equation for the energy are short range in time.
The method of the generating function also allowed us to derive also
a system of linear Volterra equations describing
the evolution of the energy autocorrelation function.
The numerical solution of these equation confirmed the aging
behaviour found in I. It would be interesting to derive analytically
the scaling
$C_E(t,t')=f(t'/t)$.
In the last section we have derived the mean-field theory
for a model where entropic and energetic barriers are combined.
We have seen that at temperature zero, starting from a random configuration
the system fails in finding the ground state. For future work
it is left the study of this model for finite temperature.
Non linear equations with memory appear in
phenomenological glass theory
under the name of Mode Coupling Theory \cite{mode}.
Mode coupling equations appear naturally in the
mean-field treatment of the
dynamics of disordered \cite{cuku,frame}
or quasi-disorderd systems \cite{FH}, in off-equilibrium
situation they involve a set of coupled
integral equation for the two time auto-correlation function
and its conjugated response function. The most striking
manifestation of the importance of memory effects in off-equilibrium
mode coupling theory is in the
aging behaviour of the response function \cite{cuku,frame}.
Structural glasses are generally
classified as strong glasses (Arrhenius behavior of the relaxation time)
or fragile glasses (Vogel-Tamman-Fulcher behavior of the relaxation time).
In this classification the BG model is a strong glass. Polymer glasses are
fragile glasses which show strong aging effects in its physical properties
\cite{glass}.
It would be desiderable to know from experiments if there is a correlation
between the fragility of glasses and its aging properties. This could
shed light on the role of energy barriers in the mechanism of the
glass transition. We believe that only entropy barriers cannot
yield aging effects in the response function. In this framework a more
detailed
study of the BG model with metastability (as presented in the last section)
at finite temperature could be instructive. In particular, the study
of the relaxation time as a function of the temperature
and the existence of aging due to the presence
of energy barriers.
\begin{center}
{\bf Acknowledgements}
\end{center}
We would like to thank J.P. Bouchaud, E.Follana and M. M\'ezard for
interesting discussions. (F.R.) acknowledges Ministerio de Educacion
y Ciencia of Spain for financial support.
\section{Appendix A}
In this appendix we derive the evolution equation for
the probability $P_k(t)$.
Define as $N_k(t)$ the number of states occupied by
$k$ particles ($P_k(t)=N_k(t)/N$).
The processes leading to a variation of $N_k$ can
be classified in this way:
\begin{itemize}
\item{{\it Process A$+$}: arrival of a particle in a state
with $k-1$ particles.}
\item{{\it Process A$-$}: departure of a particle from a state
with $k$ particles.}
\item{{\it Process B$+$}: departure of a particle from a state
with $k+1$ particles.}
\item{{\it Process B$-$}: arrival of a particle in a state
with $k$ particles.}
\end{itemize}
Note that in a sweep the processes above are not mutually
exclusive, so, for example the contemporary
occurency of $A+$ and $B-$ lead to no variation in $N_k$.
At each Monte Carlo sweep three independent random variables are extracted:
a departure state $d$ with probability $n_d/N$, an arrival state $a$
with probability $1/N$ and an acceptance variable
\be
x=
\left\{
\begin{array}{lll}
1 & {\rm with \;\;\; prob.} & {\rm e}^{-\beta}
\\
0 &{\rm with \;\;\; prob.}&1- {\rm e}^{-\beta}
\end{array}
\right.
\ee
In terms of these
variables the variation in $N_k$ in each process is given by:
\begin{itemize}
\item{{\it Process A$+$}: $\ \ \de_{n_a,k-1}
[1-\de_{k,1}(1-\de_{n_d,1})\delta_{x,0}]$}
\item{{\it Process A$-$}: $\ \ -\de_{n_d,k}[1-\de_{n_a,0}\delta_{x,0}
+ \de_{k,1}+\de_{k,1}\de_{n_a,0}\delta_{x,0}]$}
\item{{\it Process B$+$}: $\ \ \de_{n_d,k+1}[1-\de_{n_a,0}\delta_{x,0}
+\de_{k,0}\de_{n_a,0}\delta_{x,0}]$}
\item{{\it Process B$-$}: $\ \ -\de_{n_a,k}
[1-\de_{k,0}(1-\delta_{n_d,1})\delta_{x,0}] $}
\end{itemize}
The contribution in the different processes can be easily understood,
for example in process A$+$ we must have $k-1$ particles in the
arrival state. If $k=1$ and $n_d>1$ the move imply an energy cost, and
is accepted only if $x=1$.
Summing all the contribution
and averaging over $p$, $a$ and $x$ we find:
\bea
& &
\lan
N_k(t+\delta t)-N_k(t)\ran =
N[P_k(t+\delta t)-P_k(t)]\equiv {{\rm d} P_k(t)\over {\rm d} t}=
\nonumber\\
& &(k+1)(P_{k+1}-P_{k})+P_{k-1}+
P_0(e^{-\beta}-1)(\de_{k,1} - \de_{k,0} - kP_k + (k+1)P_{k+1})
\label{A10}
\eea
Very similar considerations lead to (\ref{nuk})
if one restricts the balance equation to a subset
of the whole
space.
\section{Appendix B}
In this Appendix we obtain the solution of eq.(\ref{eq14b}).
We perform the change
of variables $(x,t)\to(u,t)$ where $x-1=e^{u+\int_0^t{\rm d} s (1+\lambda(s))}
=e^u B(t,0)$.
In terms of the new
variables eq.(\ref{eq14b}) reads,
\be
\frac{\partial \hat{G}}{\partial t}=
e^u B(t,0)(\hat{G}
+\la)
\label{ap2}
\ee
where $\hat{G}(u,t)=G(x(u,t),t)$
This is a linear differential equation which can be readily solved
\be
\hat{G}(u,t)=e^{e^u\int_0^t\, {\rm d} s B(s,0)} F(u)
+
e^u
\int_0^t {\rm d} s \ \lambda(s)B(s,0)e^{e^u\int_s^t\, {\rm d} v B(v,0)}
\label{ap3}\ee
where $F$ is an arbitrary function.
Going back to $(x,t)$ and imposing
the initial condition we get eq.(\ref{aa}).
|
\section{Introduction}
Reduced-dimensionality structures are currently attracting much
attention.\cite{P5,M1,K2,P6} Modern technology made it possible to
create nanoscale 0D structures where motion in all directions is
quantized (quantum dots). Often a system represents a large array of
independent quantum dots. Because of extremely small dimensions,
the fluctuations of parameters of individual quantum dots become an
important factor since they determine the properties of the whole
array.
This paper was initiated by the experimental work Ref.\
\onlinecite{P5}, where nice photoluminescence (PL) and
photoluminescence excitation (PLE) experiments were performed with an
array of self-assembling quantum dots. The conventional nonselective
(with above-barrier excitation) PL reveals a broad peak of about
$\sim$ 50 meV in width. This width originates from a wide spread of
energies of different quantum dots in the array. Both PLE and
selectively-excited PL spectra (when only the quantum dots that are
in resonance with incident light are excited) show sets of broadened
peaks with 2-3 times smaller widths. These peaks correspond to the
distribution of energy levels in the subset of quantum dots that are
resonantly excited. It was observed that the first PLE peak is
strongly shifted from the detection energy. The origin of this shift
remained unclear. Moreover, no measurable Stokes shift was observed
in a recent paper Ref. \onlinecite{K2}, where the PL has been
measured together with the direct absorption by the layer of quantum
dots.
In this paper a simple theory of the PL from an array of quantum dots
is developed. We suggest that the process of photoexcitation of a dot
into its lowest optically-excited state does not contribute to the PL
signal. The proposed interpretation of the experimental data explains
large Stokes-like shift between the PL and PLE peaks as the average
distance between the two lowest optically-excited states. In what
follows we use the term ``ground state'' to denote the lowest
optically-excited state because it corresponds to the ground state of
the photoexcited carriers in the dot.
It is suggested that the PL and PLE lineshapes are completely
determined by the statistical distribution of the energy levels of
different quantum dots in the array. More precisely, it is determined
by the distribution of {\em pairs} of energy levels. It is shown that
such distribution is essentially two-variable. That is, there exists
a correlation in the positions of different energy levels in each
quantum dot throughout the array, however, such correlation is not
100\%. We show that this feature causes the difference in the
positions of the maxima of the PL and PLE spectra.
The fluctuations of energy levels due to the random potential caused
by the alloy composition fluctuations are studied in Section IV. We
show that the major part of the observed linewidth can be accounted
for by this mechanism. We also suggest that the first two excited
states observed in Ref.\ \onlinecite{P5} originate from the twofold
degenerate first excited level, when the degeneracy is lifted by the
random potential. The density of states and the two-level
distribution function, which accounts for the correlation in
energy-level positions, are calculated. The spectra determined by
these functions describe most of the features of the experimental
spectra.
Finally, the effects caused by fluctuations of the {\em shape} of the
quantum dots are discussed. It is suggested that the fluctuations of
the shape of the dots should increase the energy-level correlation.
\section{Origin of the Stokes-like Shift}
In the present paper we consider the case when both the electron and
the hole are confined in the quantum dot. Energy of the quantum dot
is measured from the unexcited state of the dot with no electrons and
holes. The term ``ground state'' refers to the ground state of the
quantum dot with an excited electron-hole pair, i.e.\ to the lowest
optically-excited state.
The key feature of an array of quantum dots that differs it from a
quantum well or bulk material is that there is no charge transfer
between the dots, or at least such transfer is strongly suppressed.
Different quantum dots therefore give independent additive
contributions in any optical experiment.\cite{C1}
The density of states for each quantum dot represents a set of
delta-function-like peaks, while the total density of states of the
whole system can be spread in a wide energy range due to the
inhomogeneous broadening. Several experiments have been reported
recently where the contributions to the luminescence from {\em
single} quantum dots were found.\cite{M1,K2,P6,A1} Single quantum
dots give extremely narrow sub-meV spectral lines. When, however, the
number of the excited quantum dots is large, a broad luminescence
peak is observed.\cite{P5,K2,P6}
We suggest that such a feature makes it difficult to probe optically
the ground states of the dots and thus causes an apparent Stokes-like
shift observed in Ref.\ \onlinecite{P5}. Probing of the ground states
of the dots may require special technique like high resolution or
time-resolved PL.
Indeed, let's consider the contribution to the PL from the process
when the photon is absorbed into the ground state of a dot and then
re-emitted. If the optical process is not phonon-assisted, the
energies of the incoming and outgoing photons are exactly equal. The
PL response can hardly be observed since it is hidden by the incident
beam scattered by other elements of the experimental environment.
This situation is quite different from that in quantum wells or bulk
semiconductors, where an electron excited to the conduction band has
always quantum states with smaller energies. It can loose some
energy before it recombines in an {\em independent} optical process.
In a quantum dot, the emitted photon may have its energy below (or
above) that of the exciting light if the PL is phonon-assisted.
However, the phonon must be emitted (absorbed) together with photon
{\em in the same quantum process}. The probability of such process is
determined in higher-order perturbation theory in the electron-phonon
coupling constant and is therefore much smaller than the probability
of the direct transition.
In order to give a substantial contribution to the PL signal, the dot
must be pumped into one of its {\em excited} states. Therefore, the
minimum distance between the excitation and detection energies seen
in the spectra is equal to the distance between the ground and first
excited states of the dot. In fact, no Stokes shift was found between
the PL and the direct absorption by a layer of quantum dots measured
in Ref.\ \onlinecite{K2}.
\section{Positions of peaks in the PL and PLE spectra}
In this section the shape of the PL and PLE spectra is described
qualitatively. Different peaks observed in the spectra are assigned.
It is shown that existence of a random spread of interlevel
distances in quantum dots causes substantial deviations of the
positions of peaks seen in PLE and selectively-excited PL spectra.
We shall assume for simplicity that the energy relaxation in a
quantum dot occurs faster than recombination, so that the light is
always emitted from the ground state of the dot.
Let us first consider a simple model when the distances between
different energy levels are the same for all quantum dots, however,
there is a wide distribution of energy levels in the array. In other
words, let the picture of energy levels be the same in each quantum
dot but shifted randomly as a whole. This implies a 100\% correlation
in the positions of different energy levels in each dot.
If the conventional PL technique is used so that the pumping is
performed with energies well above the barriers between the dots,
all the dots are excited and emit light at their ground-state
energies. The emission then represents a broad peak due to the wide
distribution of the ground-state energies.
Let us now consider PLE and selectively-excited PL measurements. Here
only those dots are excited, which have one of their energy levels in
resonance with the exciting light. As suggested above, the dots
pumped into their ground states do not contribute to the spectra.
When the interlevel distances are the same in all dots, both PLE and
selectively-excited PL spectra would show a set of
delta-function-like peaks as sketched in the Fig.\ 1.
It is more convenient to start with the PLE (left). By fixing the
detection energy one selects the subset of all quantum dots in the
array with the ground-state energy $E_0=E_{detector}$ (we have
assumed that the light is always emitted from the ground state of the
dots). The PLE signal appears when the excitation energy matches the
energy of an excited state ($E_{laser}=E_1$, $E_2$, etc.) in the
selected subset. Thus the first observed (lowest in energy) PLE peak
corresponds to the {\em first excited energy levels of such dots}
that have their ground-state level at the detection energy.
The analysis of the selectively-excited PL (right) is somewhat
more complicated but similar. The fixed energy of excitation selects
{\em several} subsets of all quantum dots such that $E_1=E_{laser}$,
$E_2=E_{laser}$, etc. A nonzero PL signal appears when the detection
energy matches the ground-state energy of one of the subsets. The
position of the first observed PL peak (highest in energy) thus
corresponds to the {\em ground state energy of such dots} that have
their {\em first excited} energy level at the excitation energy. The
distance between the first two PL peaks is equal to $E_{21}$, the
distance between {\em the first and the second} excited states of the
dots.
When interlevel distances are the same in all dots, all distances
between different peaks in the PL (PLE) series also remain the same,
while the whole picture shifts with the shift of the laser (detector)
energies. It means that positions of these peaks being plotted
against the laser (detector) energies must form a set of straight
lines with the unit slope.
The slopes of such lines obtained from the PLE and
selectively-excited PL spectra obtained in Ref.\ \onlinecite{P5} are
all less then 1. For the first PL and PLE peaks, e.g.,
$dE_{max,1}^{PL}/dE_{laser}=0.91$, $dE_{max,1}^{PLE}/
dE_{detector}=0.77$.\cite{P5}
It is easy to see that such deviation cannot be attributed to the
dependence of the interlevel distances on the ground-state energy.
Indeed, in this case the slopes obtained from the PL and PLE spectra
should be inverse of each other. We see, however, that
\begin{equation}
\frac{dE_{max,m}^{PL}}{dE_{laser}}\neq
\left(\frac{dE_{max,m}^{PLE}}{dE_{detector}}\right)^{-1}.
\end{equation}
We show below that existence of a spread of interlevel distances in
the array causes deviation of the positions of peaks
seen in the PL and PLE spectra in such a way that both derivatives in
Eq.\ (1) become less then 1.
To demonstrate this, let us assume that the interlevel distances in
the dots fluctuate randomly in some (narrow) energy interval. In this
case selection of a subset of quantum dots by fixing one of the
energy levels does not yet determine the positions of other energy
levels. Then one should observe a sequence of broadened peaks in both
PL and PLE. The scale of broadening is determined by the
distribution of interlevel distances only, therefore it can be
narrower than the distribution of the absolute energies. As shown
below such behavior is natural if the spread is caused by a random
potential.
To understand how the {\em positions of maxima of the broadened
peaks} are shifted, it is convenient to draw a three-dimensional
picture shown in Fig.\ 2. Here the intensity measured by the detector
is plotted as a function of two variables --- the excitation and the
emission energies. Clearly, such a plot contains all information,
which both PL and PLE can provide. To get the shapes of the PL or PLE
spectra one simply has to slice the three-dimensional plot along the
$E_{laser}$ or $E_{detector}$ axis.
First, the intensity is zero in the half-plane
$E_{laser}<E_{detector}$. The intensity is nonzero only when
$E_{laser}$ and $E_{detector}$ are in resonance with the excited and
ground states of some dot {\em simultaneously}. If the spread of
interlevel distances is small, the three-dimensional plot has a shape
of a set of narrow {\em ridges}, elongated in the direction parallel
to the line $E_{laser}=E_{detector}$. Each ridge corresponds to the
optical process for which $E_{detector}=E_0$ and $E_{laser}=E_m$,
$m$=1, 2, etc. The width of each ridge is determined by the spread of
the corresponding interlevel distances, while the length is larger
and represents the large spread of the ground-state energies. The
distance between the ridge and the line $E_{laser}=E_{detector}$ is
determined by the average distance between the ground and the
corresponding excited states of the dots.
The inset in Fig.\ 2 shows a fragment of the same plot in the isoline
projection (dashed lines). Crosses and pluses show the positions of
the maxima in the PL and PLE crossections respectively. The ``PL and
PLE lines'' are the loci of the maxima observed in the
PL and PLE spectra.
As one can see, the positions of the maxima deviate from the major
axis of the ridge and from each other. Indeed, these positions lie in
such points where the crossection line (parallel to one of the axes)
is tangential to the lines of equal intensity. Note that the
deviation of the positions of the maxima is such that both
derivatives in the Eq.\ (1) are less than 1.
Both lines intersect exactly at the maximum of the ridge, which
coordinates give the average positions of the ground and the excited
states of all the dots in the array.
In general, if the oscillator strengths are the same for all allowed
transitions, the intensity as a function of the excitation and
detection energies is proportional to the mutual level-level
distribution function: $I(E_{detector},E_{laser})\propto\sum_m
P(E_{detector}= E_0,E_{laser}=E_m)$, where $E_m$ are the energy-level
positions, $E_0$ being the position of the ground state. In the
following section we derive the shape of the distribution function
$P(E_0,E_m)$ when the spread of the energy-level positions is
caused by the composition fluctuations. We show that a quite
complicated PL and PLE lineshape observed in Ref.\ \onlinecite{P5}
can be sufficiently well described in this way.
\section{fluctuations of energy levels in quantum dots}
In this section we study the properties of the statistical
distribution of the energy levels in quantum dots caused by a
white-noise random potential. We present a strong evidence that the
major part of the observed spread of the PL and PLE peaks is
caused by universal composition fluctuations in the dots. These
fluctuations are a generic property of semiconductor alloys and
produce a theoretical limit for unification of quantum dots in an
array.
We also suggest that the first two excited levels that reveal
themselves in the PLE and selectively-excited PL experiments
originate {\em from a single doubly degenerate level} when the
degeneracy is lifted by the random potential.
In semiconductor alloys the lattice sites are occupied randomly with
two types of substitutional atoms. We ignore here the correlation
between occupation of different sites. The composition $x$ averaged
over a small volume always fluctuates. The order of magnitude of the
fluctuations is inversely proportional to the square root of the
volume over which the averaging is performed. Though usually small,
this effect can be important in extremely small structures.
In small quantum dots such composition fluctuations cause shifts of
energy levels from the positions determined by the average
composition $x_0$. The ``local'' composition $x$ varies from dot to
dot, and also {\em inside} the dot. According to that, the energy
levels fluctuate from dot to dot. Shifts of the energy levels in
different quantum dots are independent from each other. Inside a
single quantum dot the shifts of different energy levels are
correlated. Note, however, that such correlation is not 100\% (as it
would be if there was only one fluctuating parameter like the
diameter of a dot).
In order to calculate the statistical distribution of energy levels
caused by composition fluctuations we use the method developed by
Efros and Raikh (for a review see Ref.\ \onlinecite{RE}). This method
is applicable if the size of the wave function is much larger than
the lattice constant. In this case one needs to know only the shape
of the wave function in a quantum dot and the slope of the dependence
of the gap on composition, $dE_g/dx$, at average composition $x_0$.
The result depends on whether the unperturbed energy levels are
degenerate or not. Without degeneracy the distribution of energy
levels $E_m$ is Gaussian with the standard deviation given by
\begin{equation}
\sigma_m^2=\overline{\epsilon_m^2}=
\gamma\int d^3r\ \psi_m^4({\bf r}),
\end{equation}
where $\epsilon_m=E_m-\overline E_m$, $\psi_m$ is the wave function
(real) corresponding to the energy level $E_m$, and $\gamma$ is given
by
\begin{equation}
\gamma=\left(\frac{dE_g}{dx}\right)^2\frac{x_0(1-x_0)}{N},
\end{equation}
where $N$ is the number of lattice sites per unit volume.
The {\em covariance} between $\epsilon_m$ and $\epsilon_n$ can be
obtained in a similar way:
\begin{equation}
\overline{\epsilon_m\epsilon_n}=\rho\sigma_m\sigma_n
=\gamma\int d^3r\ \psi_m^2({\bf r})\psi_n^2({\bf r}).
\end{equation}
The coefficient $\rho$, $\rho\leq1$, is called the coefficient of
correlation between $E_m$ and $E_n$.
To find the shape of the PLE and selectively-excited PL spectra we
need the mutual distribution function for $E_0$ and $E_m$. The most
general form of the two-variable Gaussian distribution is given by
\begin{eqnarray}
&G&_2(\epsilon_0,\epsilon_m;\sigma_0,\sigma_m,\rho)=
\frac{1}{2\pi\sigma_0\sigma_m\sqrt{1-\rho^2}}\nonumber\\
&&\times\exp\left\{-\frac{1}{2(1-\rho^2)}
\left[\frac{\epsilon_0^2}{\sigma_0^2}
-2\rho\frac{\epsilon_0\epsilon_m}{\sigma_0\sigma_m}
+\frac{\epsilon_m^2}{\sigma_m^2}\right]\right\},
\end{eqnarray}
This is just an analytical expression for the shape of the
ridge, discussed in the previous section. The ridge is strongly
elongated when the parameter $\rho$ is close to 1. It is equal to 1 in
the limiting case when an exact relation between the energy-level
positions exists, so that the two-variable statistical distribution
becomes effectively one-variable. The ratio $\sigma_m/\sigma_0$
determines the orientation of the ridge. The ridge is parallel to the
line $E_{laser}=E_{detector}$ when $\sigma_m=\sigma_0$.
The situation is, however, more complicated if a degeneracy exists.
The random potential shifts the degenerate energy level and lifts the
degeneracy. The distribution of energies in the vicinity of an
unperturbed degenerate level appears to be not Gaussian. The mutual
two-variable distribution function for a transition between $E_0$ and
$E_m$ is not Gaussian either. Instead, it has a shape of two close
parallel ridges corresponding to each of the split-off energy levels.
For simplicity we restrict ourselves to the axially symmetric quantum
dots, where each energy level except the ground state is doubly
degenerate. In this case it appears to be possible to derive the
general form of the distribution function without knowledge of the
shape of the wave functions in the quantum dot.
For an axially symmetric system two degenerate wave functions with
the angular momentum $|m|$ have the form $\psi_{m\pm}({\bf
r})=\psi_m(r) e^{\pm im\phi}$, where $\psi_m$ can be made real. The
positions of energy levels $E_{m\pm}$, split and shifted by random
potential of the particular configuration, can be obtained as the
eigenvalues of the secular matrix
\begin{equation}
\delta H=\left[
\begin{tabular}{cc}
$u$ & $x+iy$\\
$x-iy$ & $u$
\end{tabular}
\right],
\end{equation}
where the matrix elements $u$ and $x+iy$ take random values in each
quantum dot and are given by:
\begin{mathletters}
\begin{eqnarray}
u&=&\sqrt{\gamma}\int d^3r\ V({\bf r})\psi_m^2({\bf r}),\\
x+iy&=&\sqrt{\gamma}\int d^3r\ V({\bf r})\psi_m^2({\bf r})
e^{2im\phi}.
\end{eqnarray}
\end{mathletters}
Here $V({\bf r})$ is the white-noise random potential with correlator
$<V({\bf r})V({\bf r}')>=\gamma\delta({\bf r}-{\bf r}')$.
Eigenvalues of $\delta H$ are $\epsilon_\pm=u\pm\sqrt{x^2+y^2}$
(the energy is measured from the unperturbed energy level). It is
easy to see that $u$, $x$, and $y$ are {\em independent} Gaussian
random variables with $\overline{u^2}=\sigma_m^2$,
$\overline{x^2}=\overline{y^2}=\sigma_m^2/2$, $\sigma_m^2$ being
given by Eq.\ (2).
The density of states in the vicinity of $E_m$ (the distribution
function for $\epsilon_m=E_m-\overline E_m$) is given by
\begin{eqnarray}
P(\epsilon_m)&=&\sum_\pm\int\!\int\!\int
dudxdy\ \delta(\epsilon_m-u\mp\sqrt{x^2+y^2})\nonumber\\
&&\times G_1(u;\sigma_m)
G_1(x;\frac{\sigma_m}{\sqrt{2}})G_1(y;\frac{\sigma_m}{\sqrt{2}})
\nonumber\\
&=&D_0(\epsilon_m;\sigma_m),
\end{eqnarray}
where $G_1$ is the standard one-variable Gaussian distribution,
$G_1(\epsilon;\sigma)\equiv\sigma^{-1}G_1(\epsilon/\sigma;1)$,
$G_1(z;1)=(2\pi)^{-1/2}\times$ $\exp(-z^2/2)$; and we have
introduced the notation for the distribution function
$D_0(\epsilon;\sigma)\equiv\sigma^{-1}D_0(\epsilon/\sigma;1)$,
\begin{equation}
D_0(z;1)=\frac{2}{3}\sqrt{\frac{2}{\pi}}e^{-z^2/2}+
\frac{2z}{3\sqrt{3}}e^{-z^2/3}\text{erf}(\frac{z}{\sqrt{6}}).
\end{equation}
The meaning of the index 0 in $D_0(\epsilon;\sigma)$ will soon become
clear.
The Eq.\ (9) describes a bell-shaped curve, which determines the
density of states and, hence, the {\em absorption spectrum} in the
vicinity of the excited level $E_m$. It's shape can be fairly well
approximated by a Gaussian with the effective dispersion
$\sigma_{eff}=\sigma_m\sqrt{2}$, as shown in Fig.\ 3.
In order to find the mutual two-level distribution function that
determines the shape of the PL and PLE spectra it is important to
account for the correlation between matrix elements of $\delta H$ and
the shift of the ground-state $\epsilon_0$.
It is useful to note that the expression for the matrix element $u$ is
the same as the expression for the shift of a non-degenerate level
with the wave function $\psi_m(r)$. The value of $u$ is, therefore,
{\em correlated} with $\epsilon_0$ via the same two-variable Gaussian
distribution as in Eq.\ (5):
$P(\epsilon_0,u)=G_2(\epsilon_0,u;\sigma_0,\sigma_m,\rho)$, with the
parameters $\sigma_0$, $\sigma_m$, and $\rho$ given by Eqs.\ (2),
(4). The parameters $x$ and $y$ are statistically independent of
$\epsilon_0$ and $u$, because the corresponding covariances become
zero when the integration over the azimuthal angle $\phi$ is
performed. Thus, we obtain:
\begin{eqnarray}
P(\epsilon_0,\epsilon_m)&=&\sum_\pm\int\!\int\!\int
dudxdy\ \delta(\epsilon_m-u\mp\sqrt{x^2+y^2})\nonumber\\
&&\times G_2(\epsilon_0,u;\sigma_0,\sigma_m,\rho)
G_1(x;\frac{\sigma_m}{\sqrt{2}})G_1(y;\frac{\sigma_m}{\sqrt{2}})
\nonumber\\
&=&G_1(\epsilon_0;\sigma_0)
D_\rho(\epsilon_m-\epsilon_0\rho\frac{\sigma_m}{\sigma_0};\sigma_m),
\end{eqnarray}
where the function
$D_\rho(\epsilon;\sigma)\equiv\sigma^{-1}D_\rho(\epsilon/\sigma;1)$,
and
\begin{eqnarray}
D_\rho(z;1)&=&\frac{2\sqrt{\mu-1}}{\sqrt{\pi}\mu}
\exp\left(-\frac{z^2}{\mu-1}\right)\nonumber\\
&+&\frac{2z}{\mu^{3/2}}\exp\left(-\frac{z^2}{\mu}\right)
\text{erf}\left[\frac{z}{\sqrt{\mu(\mu-1)}}\right].
\end{eqnarray}
Here $\mu=3-2\rho^2$. The function $D_0(\epsilon;\sigma)$ defined
by Eq.\ (9) is a particular case of (11) with $\rho=0$.
The function $D_\rho(\epsilon_m;\sigma_m)$ gives the distribution of
the energy sublevels in the vicinity of $\overline E_m$, when {\em
the position of the ground state is fixed}. It is normalized in such
a way that $\int d\epsilon\ D_\rho(\epsilon;\sigma)=2$ (according to
the twofold degeneracy), and $\int d\epsilon\ \epsilon^2
D_\rho(\epsilon;\sigma)=2\sigma^2(2-\rho^2)$. The function
$D_\rho(\epsilon;\sigma)$ is symmetric in $\epsilon$. It has one
maximum when $\rho\leq1/\sqrt{2}$ and two maxima, when
$\rho>1/\sqrt{2}$. The maxima become more pronounced when $\rho$
tends to 1.
The function $P(\epsilon_0,\epsilon_m)$ determined by Eq.\ (10)
gives the probability density for a quantum dot to have the ground
state at the energy $\overline E_0+\epsilon_0$ and an excited state
at the energy $\overline E_m+\epsilon_m$. It is proportional to the
intensity measured by the detector at fixed excitation and detection
energies. Hence, it determines the shape of both PL and PLE spectra
when the proper argument is fixed.
The function $P(\epsilon_0,\epsilon_m)$ appears to be not very
sensitive to the ratio $\sigma_m/\sigma_0$. It is, however, quite
sensitive to the value of the correlation coefficient $\rho$.
This function is plotted in Fig.\ 4 for $\sigma_m=\sigma_0$ and two
different values of $\rho$. As shown in the next section, it is
natural for the coefficient $\rho$ to be close to 1. The value
$\rho=0.94$ [Fig.\ 4(a)] gives the best fit to the experimental data.
When $\rho$ tends to 1 [Fig.\ 4(b)] the intensity in the dip between
two maxima approaches zero. The lineshapes of the spectra described
by Eq.\ (10) are discussed in detail in the next section.
\section{discussion}
We replot the positions of the PL and PLE maxima observed in Ref.\
\onlinecite{P5} on the combined plot in Fig.\ 5 to illustrate that
the experimental behavior is in agreement with our consideration
(compare with the inset in Fig.\ 2). The dotted line shows unit
slope, $E_{laser}=E_{detector}$. Three lines correspond to the
positions of two maxima observed in each PL curve (crosses) and one
--- in each PLE curve (pluses). Positions of only one PLE maximum for
each $E_{detector}$ are shown because the second PLE maximum is seen
not clear enough. The constant value of 1290 meV has been subtracted
from all energies.
Three lines shown in Fig.\ 5 are consistent with the qualitative
picture of two close parallel ridges. As suggested above, these
ridges correspond to the two optical processes where
$E_{detector}=E_0$ and $E_{laser}=E_{1\pm}$; $E_0$ being the
ground-state energy and $E_{1\pm}$ --- the energies of the first and
second excited states.
We use the notation $E_{1\pm}$ instead of $E_{1,2}$ for the two
excited states according to the idea that they originate from the
twofold-degenerate first excited energy level $E_1$ when the
degeneracy is lifted by random potential. This idea is supported by
the fact that the ratio of interlevel distances appears to be
$(\overline E_{1+}-\overline E_{1-})/(\overline E_{1-}-\overline
E_0)\approx 0.6$. For a cylindrical quantum well with infinite
potential walls the corresponding ratio is about $(\overline
E_{21}/\overline E_{10})=1.3$.
The point of intersection of two lines gives the position of the
maximum of the first ridge. The difference of 55 meV in its
coordinates is nothing but the observed Stokes-like shift between
emission and absorption energies. The distance between two ridges
measured along the $E_{laser}$ axis gives $\overline E_{1+}-\overline
E_{1-}$, the mean splitting of the excited state $E_1$.
Using Fig.\ 5 we may conclude that the average distance between the
ground and (split) first excited state is about 75 meV, while the
average splitting of the excited state is approximately 40 meV.
Fig.\ 6 shows the comparison between the presented theory and the
experiment. The experimental spectra from Ref.\ \onlinecite{P5} are
replotted in Fig.\ 6(a). The theoretical PLE and selectively-excited PL
spectra [Fig.\ 6(b)] are obtained with the use of Eq.\ (10) by fixing
$\epsilon_0$ or $\epsilon_m$ respectively. Experimental values for
the average positions of the energy levels $\overline E_0=1276$ meV
and $\overline E_1=1351$ meV are used.
The Fig.\ 6(b) is essentially the Fig.\ 4(a), replotted in the wavelength
scale for better comparison. The value $\rho=0.94$ gives the best fit
for the data from Ref.\ \onlinecite{P5}. It is seen that the curves
obtained reproduce all the features of both PL and PLE spectra.
The structure of the expression (10) is such that if $\epsilon_0$
is fixed, it describes a curve symmetric around the point
$\epsilon_m=\epsilon_0\rho\sigma_m/\sigma_0$. Thus (if
$\rho>1/\sqrt{2}$) the PLE line should reveal two {\em symmetric}
peaks at any detection energy. This is exactly the behavior seen in
the experiment. When the detection energy is changed, the whole
spectrum must shift linearly with it. Indeed, Fig.\ 5 shows that the
positions of PLE maxima depend linearly on the detection energy.
The slope of the PLE curve in Fig.\ 5 gives the ratio
$\sigma_1/\sigma_0$, which appears to be close to 1.
If the position of the excited level, $\epsilon_m$, is fixed, the
lineshape is asymmetric. When $\rho$ is close to 1, the PL lineshape
also shows two maxima, however, there is a peculiar interplay in their
magnitudes when the excitation energy is changed. This also matches
the experimental data quite well.
Such interplay can be easily understood. First (second) PL maximum
corresponds to the distribution of ground-state energies of such
subset of quantum dots, for which the lowest (highest) split-off
level coincides with $E_{laser}$. When $E_{laser}$ is below $\overline
E_m$, the amounts of first type of dots is larger than that of the
second type. When $E_{laser}$ is larger than $\overline E_m$, the dots
of the second type win.
The interplay is absent in PLE, because for each fixed ground-state
energy there is always equal amount of lowest and highest split-off
levels.
The magnitudes of $\sigma_1$ and $\sigma_0$ can be obtained
independently as follows. For $\rho=0.94$, the maximum of
$D_\rho(\epsilon;\sigma)$ lies at $\epsilon=0.765\sigma$. Then, from
the position of the PLE maximum, $\sigma_1$=14.5 meV/0.765=19 meV.
The spread of the ground states $\sigma_0$ can be determined
independently from the nonselective PL [curve 9 in Fig.\ 6(a)]. It
gives $\sigma_0$=18.2 meV.
If we know the shape of the wave function, we may find the values for
the parameters $\sigma_{0,1}$ and $\rho$, using Eqs.\ (2,4). As a
guess, we may try the wave functions for the cylindrical quantum well
with infinite walls:
\begin{equation}
\psi_m({\bf r})\propto\cos(\frac{\pi r_\bot}{h})
J_m(\frac{\nu_m r_\|}{R}),
\end{equation}
where $h$ and $R$ are correspondingly the thickness and the radius of
the quantum dot, and $\nu_m$ is the root of the Bessel function
$J_m$. The integrals of interest, $J_{mn}=\int d^3r\
\psi_m^2\psi_n^2$, are equal to: $J_{00}=2.098$, $J_{11}=1.552$, and
$J_{01}=1.435$.
It is easy to find all parameters in this approximation. First, let
us estimate the magnitude of the spread. To find $\sigma_0$ one has
to know the volume of the quantum dot. Taking it to be the volume of
the cylinder with the thickness 2.5 nm and diameter 25 nm,\cite{P5}
and using the parameters of InGaAs: $x_0=0.5$, lattice constant
$a=0.585$ nm and $(dE_g/dx)=1.16$ eV (both at $x=0.5$),\cite{LB} we
obtain $\sigma_0=13$ meV. This value is less than the experimental
value of 18.2 meV. It, however, shows that at least a significant
part of the spread is caused by the composition fluctuations.
There are, of course, some other reasons for spreading. The total
spread, however, cannot be less than the calculated value.
Two remaining dimensionless parameters are: $\rho=0.795$, and
$(\sigma_0/\sigma_1)=1.16$. Though the ratio $\sigma_0/\sigma_1$ is
in a reasonable agreement with the experiment, the value of the
correlation coefficient $\rho$ is significantly less than the
experimental value of 0.94. Note that the dimensionless parameters
depend only on the form of the wave functions. Though it is possible
to relate the discrepancy in $\rho$ to the unknown shape of the real
wave function, there is a more serious reason for this coefficient to
be closer to 1.
Among the other causes of spreading of energy levels, which are
not taken into account in the presented theory, there are {\em
fluctuations of the shape} of quantum dots. The distortion of the
shape of a quantum dot, even when small, cannot be represented as a
potential perturbation in the Schr\"odinger equation. It's effect on
the positions and splitting of the energy level $E_m$ can, however,
be described by an effective secular matrix $\delta H$ of the same
form as Eq.\ (6). The only difference is that the matrix elements $u$
and $x+iy$ are not described by the Eqs.\ (7) any more. Instead, they
are determined by the integrals of the derivative of the unperturbed
wave function at the boundary of the quantum dot.\cite{ED}
Thus, the effect of the shape fluctuations is only in renormalizing
the parameters $\sigma_0$, $\sigma_m$, and $\rho$. It is important,
however, that for the case of pure shape fluctuations, the parameter
$\rho$, defined as a correlation coefficient between $u$ and
$\epsilon_0$, is exactly equal to 1. The reason is that in an
axially-symmetic quantum dot, the normal derivative of the wave
function at the boundary is just a number rather than a function of
coordinate. Therefore, there should be an exact relation between
$u$ and $\epsilon_0$ {\em for each particular shape distortion}.
It is natural to assume that the effective value of $\rho$ is closer
to 1 when both mechanisms are involved, than it is in the case of
pure composition fluctuations.
\section{Conclusion}
In the present paper a simple theory is developed, which allows to
describe selective photoluminescence data from an array of quantum
dots with random parameters. The theory explains large apparent
Stokes-like shift between emission and absorption energies as the
average distance between the ground and first excited energy levels
in the dots.
It is shown that existence of a random spread of the interlevel
distances in the dots causes deviation of the positions of the maxima
of the peaks seen in the PL and PLE spectra. Such deviation can make
it difficult to determine the properties of the statistical
distribution of energy levels in the array. It is suggested how the
proper parameters of the statistical distribution may be obtained
from the experimental data.
The random shifts and splittings of energy levels caused by a
white-noise random potential in the dots are studied. The density of
states and mutual level-level distribution function are obtained for
the case of axially symmetric quantum dot. The energy-level
distribution and the resulting PLE and selectively-excited PL spectra
appear to be close to that observed in the experiment (see Fig.\ 6).
It is shown that the major part of the spread of energies observed in
the experiment originates from the random potential caused by the
composition fluctuations. It is also suggested that the random
fluctuations of the shape of the dots also contribute to the spread.
\section*{Acknowledgments}
I am grateful to A. L. Efros for formulating the problem and for
numerous illuminating discussions. I would like to thank P. M.
Petroff for providing the experimental data prior to the publication.
I appreciate important comment by M. E. Raikh and useful discussions
with P. M. Petroff, E. I. Rashba, J. M. Worlock, and F. G. Pikus.
This work was supported by the Center for Quantized Electronic
Structures (QUEST) of UCSB under subagreement KK3017.
|
\section{Introduction}
More than twenty years ago, Brezin, Le Guillou and Zinn-Justin (BGZ)
studied the
phase transition of a cubic anisotropic system by means of renormalization
group equations \cite{BreGuZ1}.
Within a $(4-\varepsilon)$-expansion, they found that to lowest
nontrivial order in $\varepsilon$, the only stable
fixed point for $N < N_c = 4$ is
the ${\rm O}(N)$-symmetric one, where $N$ is the number of field components
appearing in the cubic anisotropic model.
They interpreted this as an indication
that the anisotropy is irrelevant as long as $N$ is smaller than four.
For $N > 4$, the isotropic fixed point
destabilized and the trajectories crossed over to the cubic fixed point.
Recently, our knowledge of perturbation coefficients
of the renormalization group functions of the anisotropic system was extended
up to the five-loop level by Kleinert and Schulte-Frohlinde \cite{KleSchu2}.
Since the perturbation expansions are badly divergent, they
do not directly yield improved estimates for the crossover
value $N_c$ where the isotropic fixed point destabilizes in favor of the
cubic one. An estimate using Pad\'{e} approximants \cite{KleSchu2}
indicates $N_c$ to lie below $3$, thus permitting real
crystals to exhibit critical exponents of the cubic universality class.
For a simple $\phi^4$-theory, the Pad\'{e} approximation is known to be
inaccurate.
In fact, the most accurate renormalization group functions for that theory
have been obtained by combining perturbation expansions
with large-order estimates and using a resummation procedure based on
Borel-transformations \cite{KleJan}--\cite{GuZin5}.
It is the purpose of this paper to derive the large-order behavior of the
renormalization group functions for the anisotropic $g {\phi}^4$-theory.
In a forthcoming paper we will combine these results with the five-loop
perturbation expansion of Kleinert and Schulte-Frohlinde to derive the
precise value for the crossover value $N_c$.
For the simple $g {\phi}^4$-theory,
the large-order behavior of perturbation coefficients
has been derived by Lipatov \cite{Lip1,Lip2},
BGZ \cite{BreGuZ2} and others \cite{Bog}--\cite{Par2}
in a number of papers. The generalization to the ${\rm O}(N)$-symmetric case
was given in \cite{BreGuZ2}.
An equivalent method for calculating the large-order behavior
is based on the observation that
for a negative coupling constant Green functions possess an
exponentially small imaginary part due to the fact that the
ground state is unstable \cite{Lang,CollSop}.
The imaginary part is associated with the tunneling decay rate
of the ground state. It determines directly the large-order behavior of
the perturbation coefficients via a dispersion relation in the
complex coupling constant plane.
In the semiclassical limit,
the imaginary part of all Green functions can be calculated with the
help of classical solutions called {\em instantons}.
For a massless $g {\phi}^4$-theory in $d = 4$ space dimensions, these
instantons can be found analytically.
The imaginary part is a consequence of a negative frequency mode in the
spectrum of the fluctuation operator, whose determinant enters the
one-loop correction to the instanton contribution.
McKane, Wallace and de Alcantara Bonfim \cite{WalBon} found a way to
continue the results of the
$g {\phi}^4$-theory in $d = 4$ to a field theory in $d = 4-\varepsilon$
dimensions. They proposed an extended dimensional regularization scheme
for nonperturbative renormalizing the imaginary parts of vertex functions.
In the present work we have to extend this scheme to the case of
a ${\phi}^4$-theory with cubic anisotropy, where the energy functional
has the following form:
\begin{equation}
\label{HN}
H(\vec{\phi})=\displaystyle{\int}d^dx
\left[\frac{1}{2} \sum_{i=1}^N {\phi}_i (-{\nabla}^2) {\phi}_i +
\frac{u}{4} (\sum_{i=1}^N {\phi}_i^2)^2+
\frac{v}{4} \sum_{i=1}^N \phi_i^4 \right] \,\, .
\end{equation}
For $N = 2$, the corresponding model in quantum mechanics
was first studied by Banks, Bender and Wu \cite{BaBeWu,BaBe}
who used multidimensional WKB-techniques
to derive the large-order behavior of the perturbation
expansion for the ground state energy.
In $1990$ Janke \cite{Jan} presented a more efficient calculation
using a path integral approach. In the present work, this approach will be
generalized to quantum field theory and extended by a careful discussion
of the region near the isotropic limit $v \rightarrow 0$.
This is important, since the infrared-stable cubic fixed point
is expected to appear very close to the ${\rm O}(N)$-symmetric one.
In fact, it will be sufficient to give the quantum-field
theoretical generalization of \cite{Jan} in terms of an expansion
about the isotropic case in powers of $v$.
The paper is organized as follows.
The method is developed by treating first the case $N=2$.
In \mbox{Section $2$} we derive the Feynman rules
for the power series expansion of all Green functions
around the isotropic limit. In \mbox{Section $3$} we calculate
the small-oscillation determinants for the transversal and longitudinal
fluctuations. In \mbox{Section $4$} we use the
extended renormalization scheme of \cite{WalBon} to find the
full (real and imaginary) vertex functions, and derive the renormalization
constants to one loop. In \mbox{Section $5$} we calculate
the imaginary parts of the renormalization-group functions and thus the
large-order behavior of the perturbation coefficients.
In \mbox{Section $6$}, finally, we extend the results to the
physically relevant case $N=3$.
\section{Fluctuations around the isotropic instanton}
For positive coupling constants $u$ and $v$, the
system defined by the energy functional (\ref{HN}) is
stable and the Green functions are
real.
On the other hand, if both coupling constants are negative
the system is unstable and the Green functions acquire an imaginary part.
The corresponding functional integrals can be calculated by an analytical
continuation from positive to negative coupling constants, keeping the factor
$\exp{[\int (-\frac{u}{4}
(\sum_{i=1}^N \phi_i^2)^2 -\frac{v}{4}
\sum_{i=1}^N \phi_i^4)\,d^4x]}$
real.
We perform this analytical continuation by means of a joint rotation in the
two complex planes, substituting
$u \rightarrow u \exp{(i\theta)}$ and
$v \rightarrow v \exp{(i\theta)}$,
and rotating the azimuthal angle $\theta$
from $0$ to $\pi$. At the same time, we rotate the
contour of integration in the field space. The convergence of the
functional integrals is maintained by the field rotation
$\phi \rightarrow \phi \exp{(-i{\theta}/4)}$.
A natural parameter for the anisotropy of the system is the ratio
$\delta = v /(u+v)$. The isotropic limit corresponds to $\delta = 0$.
During the joint rotation of $u$ and
$v$, the parameter $\delta$ remains constant.
Thus, $\delta$ is a good parameter for the anisotropy at both
positive and negative couplings $u$ and $v$.
We shall use the coupling constants
$g=u+v$ and $\delta$ for a calculation of the
Green functions. These are given by the functional integrals
\begin{eqnarray}
\label{G2M}
\lefteqn{G^{(2M)} (x_1,x_2,\ldots,x_{2M})_{i_1,i_2,\ldots,i_{2M}} = }
\nonumber \\
\nonumber \\
& &\frac{\int D \phi \,\,
\phi_{i_1} (x_1) \phi_{i_2} (x_2) \cdot \ldots \cdot
\phi_{i_{2M}} (x_{2M})
\exp\{-H[\phi]\} }{\int D \phi \exp\{
-H[\phi]\} } \, ,
\nonumber \\
\end{eqnarray}
where the subscripts $i_k$ run through the $N$ components of the field
$\phi$. As explained in the introduction, we shall first study the case $N=2$
with the free energy functional
\begin{eqnarray}
\label{Hn2}
\lefteqn{H[\phi_1,\phi_2] = }
\nonumber \\
\nonumber \\
& &\!\!\!\!\!\!\!\!
\int d^dx \left\{\frac{1}{2} \left[\phi_1 (-{\nabla}^2) \phi_1+
\phi_2 (-{\nabla}^2) \phi_2 \right]+\frac{g}{4} \left[\phi_1^4+
2(1-\delta) \phi_1^2 \phi_2^2 +
\phi_2^4 \right] \right\}.
\end{eqnarray}
When expressed in terms of the coupling constants $g$ and $\delta$, the
Green functions possess an imaginary part for $g < 0$.
For the reasons explained above, we shall derive an expansion of
the imaginary parts of the Green functions (\ref{G2M})
around the isotropic case,
i.~e., in powers of $\delta$:
\begin{equation}
\label{IMG}
{\rm Im} \, G =
\sum_{n=0}^{\infty} a_n \, {\delta}^n \left(\frac{A}{-g}\right)^{p(n)}
\exp \left(\frac{A}{g}\right) [1+{\cal O}(g)]\, .
\end{equation}
Each power ${\delta}^n$ has its own $n$-dependent imaginary
part.
Given such an expansion, the large-order estimates for the coefficients of the
powers $g^k$ follows from a dispersion relation in $g$:
\begin{equation}
\label{DR}
G=\frac{1}{\pi}\int\limits_{-\infty}^0 d\bar{g}\,
\frac{{\rm Im}\, G(\bar{g}+i0)}{\bar{g}-g}\, .
\end{equation}
For a general discussion of the relationship between imaginary parts
and large-order behavior see, for example, ch.\ $17$ of \cite{book}.
If the power-series expansion of $G$ is
\begin{equation}
\label{LO1}
G=\sum_{k,n=0}^{\infty}G_{kn}\, g^k {\delta}^n \, ,
\end{equation}
the coefficients $G_{kn}$ have the asymptotic behavior
\begin{equation}
\label{LO2}
G_{kn}\stackrel{k \rightarrow \infty}{\longrightarrow}\,
-\frac{a_n}{\pi}\left(-1\right)^k \left(\frac{1}{A}\right)^k k!\, k^{p(n)-1}
\left[1+{\cal O} \left(1/k \right)\right].
\end{equation}
These estimates apply to perturbation coefficients,
in which the maximal power of $g$ is much larger than the power of $\delta$.
Being interested in the region close to the isotropic limit,
this restricted large-order estimation will be sufficient.
An expansion of the Green function (\ref{G2M}) around the instanton yields
exponentially small imaginary parts in both numerator and denominator.
In order to isolate the imaginary part of the numerator, we simplify
the denominator in (\ref{G2M}) and calculate first the imaginary part
of the approximate Green functions
\begin{eqnarray}
\label{TilG}
\lefteqn{\tilde{G}^{(2M)} (x_1,x_2,\ldots,x_{2M})_{i_1,i_2,\ldots,i_{2M}} = }
\nonumber \\
\nonumber \\
\!\!\!\!\!& &\frac{\int D \phi \,\,
\phi_{i_1} (x_1) \phi_{i_2} (x_2) \cdot \ldots \cdot
\phi_{i_{2M}} (x_{2M})
\exp\{-H[\phi]\} }{\int D \phi \exp\{
-H_0 [\phi]\} }\, ,
\nonumber \\
\end{eqnarray}
where the denominator contains only the free energy functional $H_0$.
With the aim of calculating (\ref{IMG}),
we expand the fields around the classical solution of the
isotropic limit $\delta = 0$ for the space dimension $d = 4$. Thus we write:
\begin{equation}
\label{exp}
\vec{\phi}=\vec{u}_L \phi_c+\vec{u}_L \xi+\vec{u}_T \eta=
\left( \begin{array}{r}
\cos{\vartheta} \\ \sin{\vartheta}
\end{array} \right) \phi_c+
\left( \begin{array}{r}
\cos{\vartheta} \\ \sin{\vartheta}
\end{array} \right) \xi+
\left( \begin{array}{r}
-\sin{\vartheta} \\ \cos{\vartheta}
\end{array} \right) \eta \, ,
\end{equation}
where $\phi_c$ is the well-known $g {\phi}^4$-instanton in four dimensions
\begin{equation}
\label{fcl}
\phi_c=\left(\frac{8}{-g}\right)^{1/2}
\frac{\lambda}{1+{\lambda}^2(x-x_0)^2} \, ,
\end{equation}
and $\vartheta$ is the rotation angle of the isotropic instanton in the
$(\phi_1,\phi_2)$-plane. The fields $\xi$ and $\eta$ correspond to the degrees
of freedom orthogonal to the rotation of that instanton.
The parameters $x_0$ and $\lambda$ are position
and scale size of the instanton, respectively. Inserting the
expansion (\ref{exp}) in $H(\phi_1,\phi_2)$, we obtain the expression:
\begin{equation}
\label{Hall}
H(\phi_1,\phi_2)=H(\phi_{1c},\phi_{2c})+H_1+H_2+H_3+H_4 \, ,
\end{equation}
where $H_i$ collects all terms in $\xi$ and $\eta$ of order $i$.
They are given by: \\
(1) Linear terms:
\begin{eqnarray}
\label{H1}
H_1(\xi,\eta)
&=&-\frac{\varepsilon 4(2)^{1/2}}{(-g)^{1/2}}
\int d^dx\frac{\lambda^3}{[1+\lambda^2(x-x_0)^2]^2} \xi
\nonumber \\
& &+\frac{\delta}{(-g)^{1/2}}\, \sin^2(2\vartheta)\,8(2)^{1/2}
\int d^dx \frac{\lambda^3}{[1+\lambda^2(x-x_0)^2]^3} \xi
\nonumber \\
& &+\frac{\delta}{(-g)^{1/2}}\, \sin(4\vartheta) \,4(2)^{1/2}
\int d^dx \frac{\lambda^3}{[1+\lambda^2(x-x_0)^2]^3} \eta
\end{eqnarray}
(2) Quadratic terms:
\begin{eqnarray}
\label{H2}
H_2(\xi,\eta)
&=&\frac{1}{2} \int d^dx \xi M_L \xi+
\frac{1}{2} \int d^dx \eta M_T \eta
\nonumber \\
& &+\delta \, \sin^2(2\vartheta) \, 6 \int d^dx
\frac{\lambda^2}{[1+\lambda^2(x-x_0)^2]^2}(\xi^2-\eta^2)
\nonumber \\
& &+\delta \, \sin(4\vartheta)\, 6 \int d^dx
\frac{\lambda^2}{[1+\lambda^2(x-x_0)^2]^2} \xi \eta \, ,
\end{eqnarray}
\hspace*{3ex} where $M_L$ and $M_T$ are the operators
\begin{equation}
\label{MLT}
M_L = -{\nabla}^2-\frac{24 \lambda^2}{[1+\lambda^2 (x-x_0)^2]^2}\, , \,\,
M_T = -{\nabla}^2-
\frac{ 8(1-\delta) \lambda^2}{[1+\lambda^2 (x-x_0)^2]^2}
\end{equation}
(3) Cubic terms:
\begin{eqnarray}
\label{H3}
H_3(\xi,\eta)
&=&-(-8g)^{1/2} \int d^dx \frac{\lambda}{1+\lambda^2(x-x_0)^2}
\left[\xi^3+(1-\delta) \xi \eta^2 \right]
\nonumber \\
& &-\delta \, \frac{\sin^2(2\vartheta)}{2} \, (-8g)^{1/2}
\int d^dx \frac{\lambda}{1+\lambda^2(x-x_0)^2}
(3 \eta^2 \xi-\xi^3)
\nonumber \\
& &-\delta \, \frac{\sin(4\vartheta)}{4} \, (-8g)^{1/2}
\int d^dx \frac{\lambda}{1+\lambda^2(x-x_0)^2}
(\eta^3-3 \eta \xi^2)
\end{eqnarray}
(4) Quartic terms:
\begin{eqnarray}
\label{H4}
H_4(\xi,\eta)
&=& +\frac{g}{4}\int d^dx \left[\xi^4+
\eta^4+2(1-\delta) \xi^2 \eta^2 \right]
\nonumber \\
& &-\delta \, g \, \frac{\sin^2(2\vartheta)}{8} \int d^dx
(\xi^4+\eta^4-6\xi^2 \eta^2)
\nonumber \\
& &+\delta \, g \, \frac{\sin(4\vartheta)}{4} \int d^dx
(\xi \eta^3-\eta \xi^3)
\end{eqnarray}
The linear terms (\ref{H1}) contain factors $\varepsilon$ or $\delta$,
since the expansion of the fields around the isotropic instanton is
extremal only in the four-dimensional isotropic limit.
The classical contribution of the instanton is
\begin{eqnarray}
\label{Hclas}
\lefteqn{H(\phi_{1c},\phi_{2c}) = }
\\
\nonumber \\
& &-\frac{\lambda^{\varepsilon}}{g} \frac{8\pi^2}{3}
\left[1-\frac{\varepsilon}{2}\left(2+\ln \pi+\gamma \right) \right]
-\frac{\lambda^{\varepsilon} \delta }{g} \, \frac{8\pi^2}{3}
\frac{\sin^2(2\vartheta)}{2}+
{\cal O} \left(\frac{ \delta }{g}\, \varepsilon
\right) \! .
\nonumber
\end{eqnarray}
The first term in $H(\phi_{1c},\phi_{2c})$ is evaluated up to the first
order in $\varepsilon$, because the one-loop renormalization will require
replacing $1/g \rightarrow 1/g_{\rm r}
+ {\cal O}[f({\delta}_{\rm r})/{\varepsilon}]$,
where $g_r$ is the renormalized version of the coupling constant $g$.
A contribution of order $\varepsilon$ is not needed in the second,
$\delta$-dependent term, where it would produce a further factor
$\delta$, and thus be part of the neglected terms
${\cal O}(g)$ in (\ref{IMG}).
Due to the anisotropy of the action, the fluctuation expansion
(\ref{H1})--(\ref{Hclas})
contains $\vartheta$-dependent parts. However, all these terms are of order
$\delta$ and can therefore be handled by straightforward perturbation theory.
Note that the angle $\vartheta$ disappears when the isotropic instanton is
directed along the coordinate axis. Then the isotropic instanton
coincides with the exact solution in $d = 4$ for $\delta > 0$. This is
also seen by inspecting the potential in (\ref{Hn2}).
For $\delta > 0$, the term ${\phi}_1^4 + {\phi}_2^4$ is larger than
$2 (1-\delta) {\phi}_1^2 {\phi}_2^2$,
so that the ``tunneling-paths'' of
extremal action are obviously straight lines along the coordinate axis.
For the region $\delta < 0$, the exact instanton follows from the known fact
that the action (\ref{Hn2})
is invariant under the orthogonal transformation:
$${\phi}_1 =({\tilde{\phi_1}} + {\tilde{\phi_2}})/{\sqrt{2}} \, , \qquad
{\phi}_2 =({\tilde{\phi_1}} - {\tilde{\phi_2}})/{\sqrt{2}} \, ,$$
with the new coupling constants:
$$g=(2-{\tilde{\delta}}) {\tilde{g}}/2 \, , \qquad
\delta = -\frac{2 \, \tilde{\delta}}{2-\tilde{\delta}} \, ,$$
satisfying $\delta < 0$ for $\tilde{\delta} > 0$.
In contrast to the method in \cite{Jan}, the treatment of the fluctuations
in a power series in $\delta$ does not allow us to deal with all modes
perturbatively.
Near the ${\rm O}(2)$-symmetric case, the Gaussian approximation
for the rotation of the instanton becomes invalid, and the Gaussian integral
must be replaced by an exact angle integration.
The separation of the rotation mode must be done with the help
of collective coordinates.
The Jacobian of the relevant transformation can be deduced from the
isotropic system.
The field $\eta$ of small oscillations must not contain modes of
$M_T$ with eigenvalues of the order $\delta$, since these would vanish for
$\delta = 0$. The discussion of the
longitudinal part is given in \cite{Wal}.
In dimensional regularization, all eigenvalues of $M_L$ which would be zero
for $d = 4$ are of order $\varepsilon$ in $4-\varepsilon$ dimensions. To avoid
eigenvalues of the order $1/{\varepsilon}$ in the
propagator for the longitudinal fluctuations, all collective
coordinates of the four-dimensional case must be retained in
$d$ dimensions. Therefore the field $\xi$ in (\ref{exp})
contains no modes of $M_L$ with eigenvalues proportional to $\varepsilon$.
In order to calculate the fluctuation factor and to separate the
almost-zero modes from $\det M_L$ and $\det M_T$, it is convenient to do a
conformal transformation onto a sphere in $d+1$ dimensions leading to
a discrete spectrum for the transformed
differential operators $M_L$ and $M_T$ \cite{Drum,DrumSho}.
Their products of eigenvalues can be given in terms of the Riemann
$\zeta$-function which we easily expanded near $\varepsilon=0$.
Simple $1/{\varepsilon}$-poles, which are characteristic
of dimensional regularization, appear directly from the known singularity
of that function. In place of the fields $\xi$ and $\eta$, we define the fields
${\Phi}_1 (\rho)$ and ${\Phi}_2 (\rho)$ on the unit sphere in $d+1$
dimensions:
\begin{equation}
\Phi_1={\sigma}^{1-d/2} \xi \, \quad \Phi_2={\sigma}^{1-d/2} \eta \, ,
\end{equation}
where $\sigma = 2/(1+x^2)$.
The instanton is supposed to be centered at the origin and
rescaled by $\lambda$. The spherical operator corresponding to the
differential operator ${\nabla}^2$ is
\begin{equation}
V_0=\frac{1}{2} L^2- \frac{1}{4} d(d-2)\, ,
\end{equation}
where $L^2 =\sum_{a,b} ({\rho}_a {\partial}_b-{\rho}_b {\partial}_a)^2$ is
the total angular momentum operator on the $(d+1)$-dimensional sphere.
The eigenfunctions of $V_0$ are the spherical
harmonics $Y_m^l(\rho)$ in $d+1$ dimensions \cite{Harm}.
They satisfy
\begin{equation}
V_0\, Y^l_m (\rho)=-(l+\frac{1}{2} d-1)(l+\frac{1}{2} d)\, Y^l_m(\rho)\, ,
\end{equation}
and have the degeneracy
\begin{equation}
\nu_l(d+1)=\frac{(2l+d-1) \Gamma(l+d-1)}{\Gamma(d) \Gamma(l+1)} \, .
\end{equation}
After the transformation,
the angle-independent quadratic part of $H_2$ in (\ref{H2}) becomes
\begin{equation}
\frac{1}{2} \int d \Omega \Phi_1 (-V_0-6) \Phi_1 +
\frac{1}{2} \int d \Omega \Phi_2 \left[-V_0-2 (1-\delta) \right] \Phi_2 \, ,
\end{equation}
where $d{\Omega}$ is the surface element of the unit sphere in $d+1$
dimensions.
After rewriting the entire energy functional(\ref{Hall}) in terms of the new
fields,
we can summarize the Feynman rules for the diagrammatic evaluation of the
functional integrals as follows: \\
$(1)$ Propagators (from the $\vartheta$-independent part of $H_2$):
\\
\\
\hspace*{2em}~longitudinal propagator:
\begin{equation}
\epsfxsize=2cm
\epsfbox{diaa1.eps}\,\, =\, \left[\left(-V_0-6\right)^{'}\right]^{-1}
\end{equation}
\hspace*{2em}~transversal propagator:
\begin{equation}
\hspace*{6ex}
\epsfxsize=2cm
\epsfbox{diaa2.eps}\, \, =\, \left\{\left[-V_0-2(1-\delta)
\right]^{'}\right\}^{-1}
\end{equation}
The prime indicates that the almost-zero modes are omitted when forming
the inverse.\\
$(2)$ Tadpoles $(H_1)$:
\begin{eqnarray}
\epsfxsize=1.5cm
\epsfbox{diab1.eps} &\raisebox{1ex}{=}&\!\!
\raisebox{1ex}{$\displaystyle{
-\varepsilon \left(\frac{2}{-g}\right)^{1/2}
\sigma^{-1+{\varepsilon}/2} \Phi_1}$}
\nonumber \\
\nonumber \\
\epsfxsize=1.5cm
\epsfbox{diab2.eps}
&\raisebox{1ex}{=}&\!\!
\raisebox{1ex}{$\displaystyle{
\left(\frac{2}{-g}\right)^{1/2} \delta \, \sin^2(2\vartheta)
\sigma^{{\varepsilon}/2} \Phi_1}$}
\nonumber \\
\nonumber \\
\epsfxsize=1.5cm
\epsfbox{diab3.eps} &\raisebox{1ex}{=}&\!\!
\raisebox{1ex}{$\displaystyle{
\left(\frac{1}{-2g}\right)^{1/2} \delta \, \sin(4\vartheta)
\sigma^{{\varepsilon}/2} \Phi_2}$}
\nonumber \\
\end{eqnarray}
The bold dot stands for the $\vartheta$-dependence of the vertex. \\
$(3)$ Two-point vertices ($\vartheta$-dependent part of $H_2$): \\
\begin{eqnarray}
\epsfxsize=2.5cm
\epsfbox{diac1.eps}\,\, \raisebox{1ex}{+} \,\,
\epsfxsize=2.5cm
\epsfbox{diac2.eps} &\raisebox{1ex}{=}&\!\!
\raisebox{1ex}{$\displaystyle{
\frac{3}{2}\sin^2(2\vartheta) \, \delta \, \left(\Phi_1^2-\Phi_2^2 \right)}$}
\nonumber \\
\nonumber \\
\epsfxsize=2.5cm
\epsfbox{diac3.eps} &\raisebox{1ex}{=}&\!\!
\raisebox{1ex}{$\displaystyle{
\frac{3}{2} \sin(4\vartheta) \, \delta \, \Phi_1 \Phi_2}$}
\nonumber \\
\end{eqnarray}
$(4)$ Cubic vertices $(H_3)$:
\begin{eqnarray}
\epsfxsize=1cm
\epsfbox{diad1.eps} &\raisebox{3ex}{=}& \!\!
\raisebox{3ex}{$\displaystyle{
-(-2g)^{1/2}\sigma^{-{\varepsilon}/2} \Phi_1^3}$} \nonumber \\
\epsfxsize=1cm
\epsfbox{diad2.eps} &\raisebox{3ex}{=}&\!\!
\raisebox{3ex}{$\displaystyle{
-(-2g)^{1/2} (1-\delta)\sigma^{-{\varepsilon}/2}
\Phi_1 \Phi_2^2}$} \nonumber \\
\epsfxsize=1cm
\epsfbox{diad3.eps}\,\, \raisebox{3ex}{+} \,\,
\epsfxsize=1cm
\epsfbox{diad4.eps} &\raisebox{3ex}{=}& \!\!
\raisebox{3ex}{$\displaystyle{
-(-2g)^{1/2} \, \delta \, \frac{\sin^2(2\vartheta)}{2}
\sigma^{-{\varepsilon}/2}\left(3\Phi_2^2 \Phi_1-\Phi_1^3 \right)}$}
\nonumber \\
\epsfxsize=1cm
\epsfbox{diad5.eps}\,\, \raisebox{3ex}{+} \,\,
\epsfxsize=1cm
\epsfbox{diad6.eps} &\raisebox{3ex}{=}&\!\!
\raisebox{3ex}{$\displaystyle{
-(-2g)^{1/2} \, \delta \, \frac{\sin(4\vartheta)}{4}
\sigma^{-{\varepsilon}/2} \left(\Phi_2^3-3\Phi_2 \Phi_1^2 \right)}$}
\nonumber \\
\end{eqnarray}
$(5)$ Quartic vertices $(H_4)$:
\begin{eqnarray}
\epsfxsize=1cm
\epsfbox{diae1.eps}\,\, \raisebox{2.5ex}{+} \,\,
\epsfxsize=1cm
\epsfbox{diae2.eps}\,\, \raisebox{2.5ex}{+} \,\,
\epsfxsize=1cm
\epsfbox{diae3.eps} &\raisebox{2.5ex}{=}& \!\!
\raisebox{2.5ex}{$\displaystyle{
\frac{g}{4} \sigma^{-\varepsilon} \left(\Phi_1^4+\Phi_2^4+
2(1-\delta) \Phi_1^2 \Phi_2^2 \right)}$}
\nonumber \\
\nonumber \\
\epsfxsize=1cm
\epsfbox{diae4.eps}\,\, \raisebox{2.5ex}{+} \,\,
\epsfxsize=1cm
\epsfbox{diae5.eps} &\raisebox{2.5ex}{=}& \!\!
\raisebox{2.5ex}{$\displaystyle{
\frac{g}{4}\, \delta \,\sin(4\vartheta)
\sigma^{-\varepsilon} \left(\Phi_1 \Phi_2^3-
\Phi_2 \Phi_1^3 \right)}$}
\nonumber \\
\nonumber \\
\epsfxsize=1cm
\epsfbox{diae6.eps}\,\, \raisebox{2.5ex}{+} \,\,
\epsfxsize=1cm
\epsfbox{diae7.eps}\,\, \raisebox{2.5ex}{+} \,\,
\epsfxsize=1cm
\epsfbox{diae8.eps} &\raisebox{2.5ex}{=}&\!\!
\raisebox{2.5ex}{$\displaystyle{
-\frac{g}{4} \, \delta \, \frac{\sin^2(2\vartheta)}{2}
\sigma^{-\varepsilon} \left(\Phi_1^4 +\Phi_2^4-6\Phi_1^2 \Phi_2^2 \right)}$}
\nonumber \\
\end{eqnarray}
In order to obtain the leading
corrections we must consider all connected diagrams of order ${\cal O}(1/g)$,
${\cal O}({\delta}/g)$, and ${\cal O}(1)$.
The contributions of these diagrams are
added to $H_c$ in (\ref{Hclas}). We consider
first the contributions to ${\cal O}(1/g)$, which
arise from the connected tree diagrams
\begin{equation}
\epsfxsize=2cm \epsfbox{diaf1.eps}\, , \qquad
\epsfxsize=1.5cm \epsfbox{diaf2.eps}\, , \qquad
\epsfxsize=1.5cm \epsfbox{diaf3.eps}
\end{equation}
and their ${\Phi}_1$ insertions of zeroth order in $g$:
\begin{equation}
\epsfxsize=1.5cm \epsfbox{diag1.eps}\, ,\quad
\epsfxsize=2cm \epsfbox{diag2.eps}\, , \quad
\epsfxsize=2.5cm \epsfbox{diag3.eps}\, \quad \ldots
\end{equation}
Since each ${\Phi}_1$-vertex produces an $\varepsilon$-factor, the smallest
order in $\varepsilon$ is two. However, terms of order ${\varepsilon}^2/g$
are negligible for our calculation even after one-loop renormalization, since:
$1/g \rightarrow 1/g_{\rm r} + {\cal O}[f({\delta}_{\rm r})/{\varepsilon}]$ and
$\varepsilon \rightarrow 0$. The diagrams of ${\cal O}({\delta}/g)$ can be
generated from those of ${\cal O}(1/g)$ by one of the substitutions:
\begin{equation}
\epsfysize=1cm \epsfbox{diah1.eps} \quad
\raisebox{2ex} {$\Longrightarrow$}
\quad \epsfysize=1cm \epsfbox{diah2.eps} \raisebox{2ex}{$\, \, .$}
\end{equation}
By inspection, we see that the only diagram with an $\varepsilon$-power
smaller than two is given by
\begin{equation}
\epsfxsize=2cm \epsfbox{diai1.eps}
\end{equation}
which is of order $\varepsilon \, {\delta}/g$.
After the renormalization, a term of order ${\delta}_{\rm r}$ appears.
But this term can be neglected in comparison
with the ${\delta}_{\rm r}/g_{\rm r}$-term
from $H_c$ in expression (\ref{Hclas}).
Hence all the tree diagrams would enter only in a higher-order calculation.
Contributions to order $g^0$ can appear
from one-loop diagrams. The only possible
candidates are
\begin{equation}
\epsfxsize=7cm \epsfbox{diaj3.eps} \raisebox{3ex}{$\, \, ,$}
\end{equation}
where the $1/{\varepsilon}$-pole from the loop integration and
the $\varepsilon$-factor of the ${\Phi}_1$-vertex cancel.
Following the method of \cite{Wal}, we have found that these diagrams
contribute to the coefficient of the imaginary part of $\tilde{G}^{(2M)}$
in (\ref{TilG}) a factor
\begin{equation}
c_1^L c_1^T \, =\exp(-5+\delta-{\delta}^2/2) \, .
\end{equation}
\section{Quadratic Fluctuation Determinants}
The angle-independent quadratic form is decomposed into a longitudinal and a
transverse part. For the longitudinal fluctuations we obtain the
determinant
\begin{equation}
\label{Lprod}
\left(\frac{\det V_L}{\det V_{0L}} \right)^{-\frac{1}{2}} \!\!=
\left[\frac{\det(-V_0-6)}{\det(-V_0)} \right]^{-\frac{1}{2}} \!\!=
\prod^{\infty}_{l=0} \left[\frac{(l+\frac{1}{2} d-3)(l+\frac{1}{2} d+2)}{
(l+\frac{1}{2} d-1)(l+\frac{1}{2} d)} \right]^{-\frac{1}{2} \nu_l(d+1)}_{,}
\end{equation}
which coincides with that appearing in the one-component $g{\phi}^4$ theory,
and is therefore known \cite{Wal}. It contains
a bound state at $l=0$, this being
responsible for the expected imaginary part of $\tilde{G}^{(2M)}$, and $d+1$
almost-zero eigenmodes of order $\varepsilon$ associated with the
dilatation and translation degrees of freedom of the instanton in
$4-\varepsilon$ dimensions. Extracting these modes from the
product (\ref{Lprod})
in the framework of collective coordinates \cite{ZittLan,GeSa},
we obtain the well-known formal
replacement rule for the determinant
\begin{equation}
\left(\frac{\det V_L}{\det V_{0L}} \right)^{-\frac{1}{2}} \Longrightarrow
J^{V_L} (2\pi)^{-\frac{(d+1)}{2}} c_2^L
\end{equation}
with
\begin{equation}
c_2^L=(2\pi)^{-1/2} 5^{5/2} \exp \left[ \frac{3}{\varepsilon}+\frac{3}{4}-
\frac{7}{2} \gamma+\frac{3}{\pi^2} \zeta^{'}(2) \right] \, ,
\end{equation}
and the Jacobian of the collective coordinates transformation:
\begin{equation}
J^{V_L}=\lambda^{(d-1)} \left(-\frac{16\pi^2 \lambda^{\varepsilon}}{
15g} \right)^{\frac{d+1}{2}} [1+{\cal O}(\varepsilon,g)] \, .
\end{equation}
The expression for $c_2^L$ contains the Euler constant $\gamma$, and the
derivative of the Riemann-zeta function $\zeta(x)$ at the value
$x=2$. Note the simple pole in $\varepsilon$, which is related to the
ultraviolet divergence on the one-loop level.
The transverse fluctuations
contain neither negative nor zero modes for $\delta$ larger than zero.
The corresponding fluctuation determinant is given by
\begin{eqnarray}
\label{Tprod}
\left(\frac{\det V_T}{\det V_{0T}}
\right)^{-\frac{1}{2}} &=& \left\{\frac{\det \left[-V_0-2(1-\delta) \right]}
{\det(-V_0)} \right\}^{-\frac{1}{2}}
\nonumber \\
\nonumber \\
&=& \prod^{\infty}_{l=0} \left[\frac{(l+
\frac{1}{2} d-1)(l+\frac{1}{2} d)-2(1-\delta)}
{(l+\frac{1}{2} d-1)(l+\frac{1}{2} d)}
\right]^{-\frac{1}{2} \nu_l(d+1)}.
\end{eqnarray}
Just one zero eigenmode appears for $l=0$ in the isotropic limit
$\delta \rightarrow 0$, due to the rotational invariance in that case. For
$\delta > 0$, the numerator in (\ref{Tprod}) contributes a factor
$1/{\sqrt{2 \delta}}$. To avoid this artificial
zero-mode divergence for $\delta \rightarrow 0$ we separate,
as announced earlier, the angular
degree of freedom from the integral measure via the collective-coordinates
method. Similar to the longitudinal case, we make the
formal substitution
\begin{equation}
\left(\frac{\det V_T}{\det V_{0T}} \right)^{-\frac{1}{2}} \Longrightarrow
J^{V_T} (2\pi)^{-\frac{1}{2}} c_2^T
\end{equation}
with
\begin{equation}
J^{V_T}=\left(-\frac{16\pi^2 \lambda^{\varepsilon}}{
3g} \right)^{\frac{1}{2}} \, ,
\end{equation}
which coincides with the Jacobian of the isotropic model.
Since the integration interval for the angle $\vartheta$ is compact,
no singularity appears in the limit $\delta \rightarrow 0$.
It is useful to illustrate the appearance of a divergent factor
$1/{\sqrt{2 \delta}}$ in a careless use of Gaussian integral.
Assuming $\delta > 0$, we expand the angle-dependent
classical action around $\vartheta = 0$ up to quadratic order.
The ensuing Gaussian integral can be evaluated after replacing the finite
integration interval $\vartheta \in [-\pi,\pi]$ by the noncompact one
$\vartheta \in [-\infty,\infty]$.
In this way, the integral
\begin{displaymath}
J^{V_T}(2\pi)^{-\frac{1}{2}}
\int\limits_{-\pi}^{+\pi} d \vartheta \exp
\left[-\lambda^{\varepsilon} \frac{8\pi^2}{3} \, \frac{\delta}{|g|} \,
\frac{\sin^2(2\vartheta)}{2} \right]
\end{displaymath}
is evaluated to
\begin{displaymath}
\left(\frac{8\pi \lambda^{\varepsilon}}{3|g|} \right)^{\frac{1}{2}}
\int\limits_{-\infty}^{+\infty} d \vartheta \exp \left[-\frac{1}{2} \left(
\frac{16\pi^2 \lambda^{\varepsilon}}{3|g|} \, 2\delta \right) \vartheta^2
\right] =\frac{1}{\sqrt{2 \delta}}
\end{displaymath}
showing the spurious would-be zero-mode singularity, as the consequence of
false rotation-mode treatment.
Excluding the $l=0$~-mode in the numerator on the right-hand
side of (\ref{Tprod}),
we obtain for $c_2^T$:
\begin{eqnarray}
& \displaystyle{c_2^T}= & \!\!\! 2^{1/2} \exp \left\{-\frac{1}{2}
\sum_{l=1}^{\infty}
\frac{(2l+d-1) \Gamma(l+d-1)}{\Gamma(d)
\Gamma(l+1)} \ln \left[
\frac{(l+\frac{d}{2}+1)(l+\frac{d}{2}-2)}
{(l+\frac{d}{2}-1)(l+\frac{d}{2})}
\right] \nonumber \right. \\
\nonumber \\
& & \left.
-\frac{1}{2} \sum_{l=1}^{\infty}
\frac{(2l+d-1) \Gamma(l+d-1)}{
\Gamma(d) \Gamma(l+1)} \ln \left[
1+\frac{2\, \delta}{(l+\frac{d}{2}+1)
(l+\frac{d}{2}-2)} \right] \right\}.
\nonumber \\
\end{eqnarray}
The first sum is the well-known contribution from the isotropic limit. The
second sum is to be expanded in powers of $\delta$. For large $l$, the sum
diverges as $\varepsilon$ tends to zero.
In order to separate off the divergence
as a simple $1/{\varepsilon}$-pole, we use the
zeta-function regularization method described in \cite{Wal}.
A somewhat tedious but straightforward calculation leads to
\begin{equation}
c_2^T=(2\pi)^{-1/6} 3^{1/2} \exp \Bigg[\frac{1}{3\varepsilon}(1-\delta)^2
+\frac{1}{4}-\frac{1}{2}\, \gamma +
\frac{\zeta^{'} (2)}{\pi^2}-\frac{1}{9} \delta+
\frac{37}{81} {\delta}^2 +{\cal O} \left({\delta}^3 \right) \Bigg]\, .
\end{equation}
The series in front of the $\varepsilon$-singularity ends after the
quadratic power of $\delta$, thereby leading to the exact
$(1-\delta)^2/3$-coefficient of the simple $\varepsilon$-pole.
Collecting all contributions to the one-loop
expression of ${\rm Im} \, \tilde{G}^{(2M)}$,
we obtain the imaginary parts
\begin{eqnarray}
\lefteqn{{\rm Im} \, \tilde{G}^{(2M)} (x_1,x_2,
\ldots,x_{2M})_{i_1,i_2,\ldots,i_{2M}} = }
\nonumber \\
\nonumber \\ \
& & \!\!\!\!\!\!\!\!
-\int d \lambda \, d^dx_0 \left\{ c_1^L c_1^T c_2^L c_2^T
(2\pi)^{-\frac{(d+1)}{2}} (2\pi)^{-1/2} J^{V_L} J^{V_T}
\exp \left(\frac{\lambda^{\varepsilon}A}{g} \right)
\nonumber \right. \\
\nonumber \\
& & \left. \mbox{} \!\!\!\!\!\!\!\!
\times \prod^{2M}_{\nu=1} \phi_c(x_{\nu}) \,\,
\frac{1}{2} \int\limits_0^{2\pi} d \vartheta u_{L\,i_1}(\vartheta)
u_{L\,i_2}(\vartheta) \cdots u_{L\,i_{2M}} (\vartheta)
\exp[a\sin^2(2\vartheta)] \right\}
\end{eqnarray}
with
\begin{equation}
\label{aA}
a=\frac{\lambda^{\varepsilon} 8\pi^2}{3g} \, \frac{\delta}{2} \, , \quad
A=\frac{8\pi^2}{3} \left[1-\frac{\varepsilon}{2}(2+\ln \pi+\gamma) \right]
\end{equation}
and $i_k=1,2$. The expression contains a standard factor $1/2$,
since for symmetry reasons each saddle point contributes
only one half of the Gaussian integral.
It remains to perform the collective coordinates
integration over the dilatation
$(\lambda)$, translation $(x_0)$, and the rotation
$(\vartheta)$ degree of freedom.
Having obtained the imaginary parts of the Green functions, we go over
to the bare imaginary part of the vertex functions.
Taking the Fourier transform and excluding the
$(2\pi)^d\, \delta(\sum_i q_i)$-factors, an amputation of the external legs
of $\tilde{G}^{(2M)}(x_1,x_2,\ldots,x_{2M})_{i_1,i_2,\ldots,i_{2M}}$
leads to:
\begin{eqnarray}
\label{V2M}
\lefteqn{{\rm Im} \,\Gamma_{\rm b}^{(2M)} (q_m)_{i_1,i_2,\ldots,i_{2M}} = }
\nonumber \\
\nonumber \\
& &\!\!\!\!\!
\displaystyle{-c_{\rm b}} \int\limits_0^{\infty}
\frac{d\lambda}{\lambda} \left\{
\lambda^{d-M(d-2)} \left(-\frac{\lambda^{\varepsilon}8\pi^2}{3g}
\right)^{(d+2+2M)/2} \!\!\! \exp \left(\frac{\lambda^{\varepsilon}A}{g}
\right) \prod^{2M}_{\nu=1}
\left(\frac{q_{\nu}}{\lambda}\right)^2 \!
\tilde{\phi} \left( \frac{q_{\nu}}{\lambda} \right)
\nonumber \right. \\
\nonumber \\
& & \left. \mbox{} \!\!\!\!\!
\times \frac{1}{2} \int\limits_0^{2\pi}
d \vartheta u_{L\,i_1}(\vartheta)
u_{L\,i_2}(\vartheta) \cdots u_{L\,i_{2M}} (\vartheta)
\exp[a\sin^2(2\vartheta)] \right\} [1+{\cal O}(\varepsilon,g)]
\end{eqnarray}
with
\begin{equation}
\tilde{\phi}(q)=2^{d/2} \pi^{\frac{(d-2)}{2}} 3^{1/2} \left|q\right|^{1-(d/2)}
K_{\frac{d}{2}-1} (\left|q\right|) \, ,
\end{equation}
where $K_\mu(z)$ is the modified Bessel-function.
The coefficient $c_{\rm b}$ is given by
\begin{eqnarray}
&c_{\rm b}=2^{-2/3} 3^{1/2} \pi^{-11/3} \exp &\!\!\!
\Bigg[
\frac{1}{3\varepsilon}
\left(10-2\delta+{\delta}^2\right)+
\frac{4 \zeta^{'} (2)}{\pi^2}-
4\gamma -4+\frac{8}{9} \, \delta
\nonumber \\
\nonumber \\
& &\mbox{}
-\frac{7}{162} \, {\delta}^2
+{\cal O} \left( {\delta}^3 \right) \Bigg].
\end{eqnarray}
An $\varepsilon$-singularity results from the ultraviolet
divergence of one-loop diagrams. As discussed in the following section, this
simple pole in $\varepsilon$ is removed by a conventional coupling constant
renormalization.
The contribution from the imaginary part of the denominator in
equation (\ref{G2M}) follows for $M=0$ in (\ref{V2M}). We observe that it is
of higher order in $g$ in comparison with the imaginary part of the numerator,
and can therefore be neglected. So the imaginary part of (\ref{TilG})
supplies the desired imaginary part of (\ref{G2M}).
\section{Renormalization}
In the preceding section we have calculated the contribution of the quadratic
fluctuations around the isotropic instanton. The almost-zero eigenvalues of
translation, dilatation and rotation have been extracted and treated
separately. Of course, the resulting expressions
are useless, if we are not able
to renormalize the theory.
This is the most difficult additional problem which
arises in the transition from quantum mechanics to higher-dimensional field
theories.
A systematic scheme to renormalize both the real and imaginary part
of vertex functions for a $g{\phi}^4$ theory in $4-\varepsilon$ dimensions
was introduced by McKane, Wallace and de Alcantara Bonfim.
They calculated the full (real and imaginary)
renormalization-group constants using an extended minimal
subtraction scheme to one loop (the conventional
one is given in \cite{HoVe,Hooft}).
We have extended their method to
the case of two coupling constants in view of applications to critical
phenomena.
For an illustration we consider first the four-point vertex function.
The result (\ref{V2M}) of the one-loop calculation about the instanton reads:
\begin{eqnarray}
\label{pol}
\lefteqn{{\rm Im} \, \Gamma_{\rm b}^{(4)}(q_m)_{i,j,k,l}= } \nonumber \\
\nonumber \\
& &\displaystyle{\sum_{n=0}^{\infty}}
\left\{ \frac{(-{\delta}/2)^n}{n!}
\frac{1}{2} \int\limits_0^{2\pi} d \vartheta u_{L\,i}(\vartheta)
u_{L\,j}(\vartheta) u_{L\,k}(\vartheta) u_{L\,l}(\vartheta)
[\sin(2\vartheta)]^{2n}
\nonumber \right. \\
\nonumber \\
& &\left. \mbox{} \times \left[
-c_{\rm b} \int\limits_0^{\infty} \frac{d\lambda}{\lambda}
\lambda^{\varepsilon}
\left(-\frac{\lambda^{\varepsilon}A}{g}\right)^{\frac{d+6+2n}{2}}
\!\!\! \exp \left(\frac{\lambda^{\varepsilon}A}{g}
\right) \prod^{4}_{\nu=1}
\left(\frac{q_{\nu}}{\lambda}\right)^2 \!
\tilde{\phi}
\left( \frac{q_{\nu}}{\lambda} \right) \right] \right\}
\end{eqnarray}
with $A=\frac{8{\pi}^2}{3} +{\cal O}(\varepsilon)$ and $i,j,k,l=1,2$.
It remains to evaluate the integral over the parameter $\lambda$. This is a
relict of the introduction of collective coordinates in the instanton
calculation. The integral converges for small $\lambda$ due to the
exponentially decreasing of modified Bessel-function $K_{d/2-1}$.
For large $\lambda$, the product
$(q_{\nu}/{\lambda})^2 \tilde{\phi}(q_{\nu}/{\lambda})$ behaves like
\begin{eqnarray}
\left(\frac{q_{\nu}}{\lambda}\right)^2 \!
\tilde{\phi}\left(\frac{q_{\nu}}{\lambda}\right)
\!\!&=&\!\!2^2 3^{1/2} \pi \Bigg\{
\left(\frac{q_{\nu}}{2\lambda}\right)^{\varepsilon}
\left[1+\frac{1}{2}\varepsilon \left(\gamma-\ln \pi \right)+
{\cal O}\left({\varepsilon}^2\right) \right]
\nonumber \\
& &+\left(\frac{q_{\nu}}{2\lambda}\right)^2 \Bigg[
\left(\frac{q_{\nu}}{2\lambda}\right)^{\varepsilon}
\left(\frac{2}{\varepsilon}+\gamma-\ln \pi+{\cal O}(\varepsilon)\right)
\nonumber \\
& &\quad
-\left(\frac{2}{\varepsilon}+1-\gamma-\ln \pi
+{\cal O}(\varepsilon) \right) \Bigg]
+{\cal O}\left(\frac{q_{\nu}}{2\lambda}\right)^4 \Bigg\} \, ,
\end{eqnarray}
so that the integral diverges logarithmically as $\varepsilon$ goes to zero.
This divergence causes a $1/{\varepsilon}$-pole in the imaginary part.
Since after the above approximation of
$\tilde{\phi}\left(q_{\nu}/{\lambda}\right)$ the integral diverges for
$\lambda \rightarrow 0$, a small $\lambda$ cutoff $\mu$ has to be
introduced.
In this way, the $\lambda$-integral in (\ref{pol}) takes the form
\begin{equation}
\label{scal}
-c_{\rm b} \int\limits_{\mu}^{\infty} \frac{d\lambda}{\lambda}
\lambda^{\varepsilon}
\left(-\frac{\lambda^{\varepsilon}A}{g}\right)^{\frac{d+6+2n}{2}}
\!\!\! \exp \left(\frac{\lambda^{\varepsilon}A}{g} \right)
(2^2 3^{1/2} \pi)^4 \prod^{4}_{\nu=1} \left(\frac{q_{\nu}}{2\lambda}
\right)^{\varepsilon}
\left[1+{\cal O}\left(\frac{q_{\nu}}{2\lambda} \right) \right] \, .
\end{equation}
The coupling constant $g$ is chosen to lie on top of the tip of the left-hand
cut in the complex $g$-plane. Being integrated only near
$g \rightarrow 0^{-}$, the integral can be evaluated perturbatively in $g$.
Using
\begin{equation}
\label{lambda}
\int\limits_{\mu}^{\infty} \frac{d\lambda}{\lambda} \lambda^{a\varepsilon}
\lambda^{\varepsilon}
\exp\left[-\left(\frac{\lambda^{\varepsilon} A}{|g|}\right)\right]=
\frac{1}{\varepsilon}\frac{|g|}{A} \mu^{a\varepsilon}
\exp\left[-\left(\frac{\mu^{\varepsilon} A}{|g|}\right)\right]
[1+{\cal O}(g)]\, ,
\end{equation}
we get for (\ref{scal}) the expression
\begin{equation}
2^8 3^2 \pi^4 \,c_{\rm b}\,\, \frac{1}{\varepsilon}\,\, \frac{g}{A}\,
\left(-\frac{\mu^{\varepsilon}A}{g}\right)^{\frac{d+6+2n}{2}}
\!\!\! \exp \left(\frac{\mu^{\varepsilon}A}{g} \right)
\prod^{4}_{\nu=1} \left(\frac{q_{\nu}}{\mu}
\right)^{\varepsilon}.
\end{equation}
After expanding the factor $(q_{\nu}/{\mu})^{\varepsilon}$ in powers of
$\varepsilon$, we obtain
the typical finite contribution $\ln(q_{\nu}/{\mu})$. Since this term can be
omitted by a minimal subtraction, we are left with
\begin{eqnarray}
\label{gim}
\lefteqn{{\rm Im} \, \Gamma_{\rm b}^{(4)}(q_m)_{i,j,k,l}= } \nonumber \\
\nonumber \\
& &2^8 3^2 \pi^4 \,c_{\rm b}\,\, \frac{1}{\varepsilon}
\sum_{n=0}^{\infty}
\left\{ \frac{(-2\, \delta)^n}{n!\Gamma(2n+3)}
\left(\frac{g}{A}\right)
\left(-\frac{\mu^{\varepsilon}A}{g}\right)^{\frac{d+6+2n}{2}}
\!\!\! \exp \left(\frac{\mu^{\varepsilon}A}{g} \right)
\nonumber \right. \\
\nonumber \\
& &\left. \mbox{} \times \left[
3\Gamma\left(n+\frac{3}{2}\right)^2 \,\,(S_{ijkl}-\delta_{ijkl})+
\Gamma\left(n+\frac{1}{2}\right) \Gamma\left(n+\frac{5}{2}\right)
\,\, \delta_{ijkl}
\right] \right\}\!,
\end{eqnarray}
where
\begin{displaymath}
S_{ijkl}=\frac{1}{3} (\delta_{ij} \delta_{kl}+\delta_{ik} \delta_{jl}+
\delta_{il} \delta_{jk})
\end{displaymath}
and
\begin{displaymath}
\delta_{ijkl} = \left\{ \begin{array}{l}
1\, , \quad i=j=k=l \\ 0\, , \quad {\rm otherwise}\, .
\end{array} \right. \qquad (i,j,k,l=1,2)
\end{displaymath}
The tensor structure arises explicitly upon doing the integrals
\begin{displaymath}
\int\limits_0^{{\pi}/2}
\sin^{2\alpha+1} (\vartheta) \cos^{2\beta+1} (\vartheta)
\,d\vartheta = \frac{\Gamma(\alpha+1) \Gamma(\beta+1)}{2\,\,
\Gamma(\alpha+\beta+2)}=\frac{1}{2} B(\alpha+1,\beta+1)
\end{displaymath}
for all combinations of indices in ${\rm Im} \,{\Gamma}_{ijkl}^{(4)}$.
The first term in (\ref{gim}) contains only mixed index combinations,
whereas the second term has only a contribution from equal indices.
Now we can proceed to
renormalize the bare results.
The real part of ${\Gamma}_{ijkl}^{(4)}(q)$ is the
perturbative one, and it is easily calculated to one loop. In terms of our
coupling constants,
the real parts ${\rm Re} \,{\Gamma}_{ijkl}^{(4)}(q)$ take the form
\begin{eqnarray}
\label{greal}
\lefteqn{{\rm Re} \, \Gamma_{\rm b}^{(4)}(q_m)_{i,j,k,l}= } \nonumber \\
\nonumber \\
& &\!\!\!\!\!
\left\{-6(1-\delta) g+\left[\frac{3}{4\pi^2\varepsilon}
\left(10-14 \delta+4 {\delta}^2\right)
+{\cal O}({\varepsilon}^0) \right]
g^2 \mu^{-\varepsilon} \right\} \{
S_{ijkl}-\delta_{ijkl} \}
\nonumber \\
\nonumber \\
& &\!\!\!\!\!
+\left\{-6g+\left[\frac{3}{4\pi^2\varepsilon}
\left(10-2 \delta+{\delta}^2\right)
+{\cal O}({\varepsilon}^0) \right]
g^2 \mu^{-\varepsilon} \right\}
\, \delta_{ijkl}\, ,
\end{eqnarray}
where $\mu$ is the arbitrary momentum scale introduced above in (\ref{scal}).
In the absence of an instanton, the wave function renormalization has no
one-loop contribution. In the consequence,
the expression (\ref{greal})
is rendered finite by a coupling constant renormalization
only. The subscript r of the coupling constants indicates
renormalized expressions in the conventional perturbative manner, i.~e.,
\begin{eqnarray}
g_{\rm r}\!\!\!&=&\!\!\!g\mu^{-\varepsilon}-\frac{1}{8\pi^2\varepsilon}
\left(10-2\delta+{\delta}^2 \right) g^2 \mu^{-2\varepsilon}\, ,
\nonumber\\
\nonumber\\
\delta_{\rm r}\!\!&=&\!\!\delta+\frac{1}{8\pi^2\varepsilon}
\left(-2\delta+{\delta}^2+{\delta}^3 \right) g \mu^{-\varepsilon}
\nonumber \\
\end{eqnarray}
Inserting this into (\ref{gim}) and (\ref{greal}) we find
the perturbatively renormalized vertex functions
\begin{equation}
\label{gren}
\Gamma_{\rm r}^{(4)}(q_m)_{i,j,k,l}=-6\mu^{\varepsilon}
\left[F_1(\delta_{\rm r},g_{\rm r})
\left(S_{ijkl}-\delta_{ijkl}\right)+F_2(\delta_{\rm r},g_{\rm r})\delta_{ijkl}
\right]
\end{equation}
with
\begin{eqnarray}
\lefteqn{F_1({\delta}_{\rm r},g_{\rm r})= } \nonumber \\
& & \!\!\!\!\!(1-{\delta}_{\rm r}) g_{\rm r}+i 2^7 3^2 \pi^4 \,
c_{\rm r} \,\, \frac{1}{\varepsilon}
\sum_{n=0}^{\infty} \frac{(-2{\delta}_{\rm r})^n}{n!}
\frac{\Gamma\left(n+\frac{3}{2} \right)^2}{\Gamma(2n+3)}
\left(-\frac{A}{g_{\rm r}}\right)^{\frac{d+4+2n}{2}}
\!\!\! \exp \left(\frac{A}{g_{\rm r}} \right),
\nonumber \\
\end{eqnarray}
and
\begin{eqnarray}
&F_2&\!\!\!\!({\delta}_{\rm r},g_{\rm r})= \nonumber \\
&g_{\rm r}&\!\!\!\!+i 2^7 3 \pi^4 \,
c_{\rm r} \, \frac{1}{\varepsilon}
\sum_{n=0}^{\infty} \frac{(-2{\delta}_{\rm r})^n}{n!}
\frac{\Gamma\left(n+\frac{1}{2} \right) \Gamma\left(n+\frac{5}{2} \right)}
{\Gamma(2n+3)}
\left(-\frac{A}{g_{\rm r}}\right)^{\frac{d+4+2n}{2}}
\!\!\! \exp \left(\frac{A}{g_{\rm r}} \right).
\nonumber \\
\end{eqnarray}
The coefficient $c_{\rm r}$ is given by
\begin{eqnarray}
\label{cr}
c_{\rm r}\!\!\! &=&\!\!\! c_{\rm b}
\exp \left[ -\frac{A}{8 \pi^2\varepsilon}
\left(10-2 {\delta}_{\rm r}
+{\delta}_{\rm r}^2 \right) \right]
\nonumber \\
\nonumber \\
&=&\!\!\! 3^{1/2} 2^{-2/3} \pi^{-2} \exp \bigg[\frac{1}{3\varepsilon}
\left(10-2 {\delta}_{\rm r}+{\delta}_{\rm r}^2 \right)-
\frac{1}{3\varepsilon}
\left(10-2 {\delta}_{\rm r}+{\delta}_{\rm r}^2 \right)
\nonumber \\
\nonumber \\
& & \qquad \qquad \qquad \quad
+\frac{4\zeta^{'}(2)}{\pi^2}-\frac{7}{3}\gamma-\frac{2}{3}
+{\cal O}({\delta}_{\rm r})
+{\cal O}({\delta}_{\rm r} \, g_{\rm r}) \bigg].
\end{eqnarray}
The $1/{\varepsilon}$-pole terms cancel.
Thus, $c_{\rm r}$ remains finite for $\varepsilon \rightarrow 0$.
For the leading expansion (\ref{IMG}) we can
take the perturbative renormalized
expression (\ref{cr}) at the position ${\delta}_{\rm r} = 0$. The only
remaining singularity
is the $1/{\varepsilon}$-factor in
${\rm Im} \,{\Gamma}^{(4)}_{\rm r}(q)_{ijkl}$.
It requires a further renormalization.
In our special choice of coupling constants, the disastrous
$k!$-divergence of perturbation series appears in
the expansions coefficients of $g^k$.
Since $\delta$ is a
dimensionless anisotropy measure for positive as well as negative $g$,
we have expanded the imaginary part of vertex functions around the isotropic
case in a well defined manner. For the derivation of nonperturbative
corrections to the renormalization-group functions and for later
applications, however, it is convenient
to avoid ratios of coupling constants. Therefore we proceed
with $v_{\rm r} = g_{\rm r} \, {\delta}_{\rm r}$ and
$g_{\rm r}$ instead of ${\delta}_{\rm r}$ and $g_{\rm r}$.
Then $v_{\rm r}$ receive the rule of ${\delta}_{\rm r}$.
Now we apply the extended minimal subtraction and perform a second
(nonperturbative) renormalization step to eliminate
the $1/{\varepsilon}$-pole term:
\begin{eqnarray}
\label{REN}
g_{\rm R}\!\!&=&\!\!g_{\rm r}+i 2^7 3 \pi^4 c_{\rm r} \frac{1}{\varepsilon}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
B \left( n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
& & \qquad \qquad \qquad \quad
\times \left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+4+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg]\, , \nonumber \\
\nonumber \\
\nonumber \\
v_{\rm R}\!\!&=&\!\!v_{\rm r}-i 2^7 3 \pi^4 c_{\rm r} \frac{1}{\varepsilon}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
\left(\frac{4n}{2n+3}\right)
B \left(n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
& & \qquad \qquad \qquad \quad
\times \left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+4+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg]\, .
\end{eqnarray}
The subscript R denotes the fully renormalized coupling constants. The
required wave function renormalization in nonperturbative terms give no
leading contributions to the imaginary part (see below). Therefore the
$1/{\varepsilon}$-singularity in (\ref{gren}) is removed by a
nonperturbative coupling-constant renormalization only.
We now proceed to investigate the two-step renormalization for the
wave function. There are two additional complications for the
${\Gamma}^{(2)}$-functions. First, ${\rm Im} \,{\Gamma}^{(2)}$ is quadratically
divergent. Second, there is a momentum-dependent divergence in the imaginary
part of ${\Gamma}^{(2)}$. For the $g{\phi}^4$-theory it was shown in
\cite{WalBon} that the undesirable momentum-dependence disappears
during the nonperturbative renormalization process. This is comparable to the
situation in conventional two-loop perturbation expansion. We continue to
show that this non-obvious cancellation still works in the more complex case
of two coupling constants and more than one field components.
According to (\ref{V2M}) the imaginary part of
the unrenormalized two-point vertex function reads
\begin{eqnarray}
{\rm Im} \, \Gamma_{\rm b}^{(2)}(q)_{i,j}\!\!&=&\!\!
c_{\rm b} \int\limits_0^{\infty} \frac{d\lambda}{\lambda}
\left\{ \lambda^2
\left(-\frac{\lambda^{\varepsilon}A}{g}\right)^{\frac{d+4}{2}}
\!\!\! \exp \left(\frac{\lambda^{\varepsilon}A}{g} \right)
\left[ \left(\frac{q}{\lambda}\right)^2 \!
\tilde{\phi} \left( \frac{q}{\lambda} \right) \right]^2
\nonumber \right.\\
\nonumber \\
& &\left. \qquad \qquad \times
\frac{1}{2} \int\limits_0^{2\pi} d \vartheta u_{L\,i}(\vartheta)
u_{L\,j}(\vartheta)
\exp\left[a \sin^2(2\vartheta)\right] \right\}.
\end{eqnarray}
The integration over the angle $\vartheta$ can be done:
\begin{eqnarray}
\lefteqn{\frac{1}{2} \int\limits_0^{2\pi}
d\vartheta u_{L\,i}(\vartheta)
u_{L\,j}(\vartheta) \exp\left[a \sin^2(2\vartheta)\right]}
\nonumber \\
\nonumber \\
& &\!\!\!\!\!\!\!
=\delta_{ij} \sum_{n=0}^{\infty} \frac{a^n}{n!}\, 4^n \frac{\Gamma\left(n+
\frac{1}{2} \right) \Gamma\left(n+\frac{3}{2} \right) }{\Gamma(2n+2)}
=\delta_{ij} \sum_{n=0}^{\infty} \frac{a^n}{n!}\, 4^n B \left(n+
\frac{1}{2},n+\frac{3}{2} \right).
\nonumber \\
\end{eqnarray}
Thus, we get the usual form:
${\rm Im} \,{\Gamma}^{(2)}_{b\,\,ij}
={\delta}_{ij}\,{\rm Im} \,{\Gamma}^{(2)}_{\rm b}$.
The wave function renormalization is contained in
$\frac{\partial}{\partial q^2} {\Gamma}^{(2)}(q)$.
Following a similar procedure
as in the case of ${\Gamma}^{(4)}$, we obtain the imaginary part
\begin{eqnarray}
\label{WFim}
\frac{\partial}{\partial q^2}\, {\rm Im} \,\Gamma_{\rm b}^{(2)}(q)\!\!&=&\!\!
2^4 \pi^2 3 \,c_{\rm b}\,
\frac{1}{\varepsilon} \sum_{n=0}^{\infty} \left\{\frac{(-2\, \delta)^n}{n!}
B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\nonumber \right. \\
\nonumber \\
& &\left. \mbox{} \times
\left[\gamma-\frac{1}{2}+\ln \left(\frac{q}{2\mu}\right) \right]
\left(-\frac{\mu^{\varepsilon}A}{g} \right)^{\frac{d+2+2n}{2}} \exp\left(
\frac{\mu^{\varepsilon}A}{g}\right) \right\} \, .
\nonumber \\
\end{eqnarray}
Apart from the appearance of the $(1/{\varepsilon}) \ln(q/{\mu})$ term,
a very similar expression is found for
${\rm Im} \,{\Gamma}^{(4)}_{\rm b}(q)$. In order to
show the cancellation of this term during the second stage of renormalization
we need the two-loop expression for the real part of
$\frac{\partial}{\partial q^2} {\Gamma}^{(2)}(q)$
\begin{equation}
\label{WFreal}
\frac{\partial}{\partial q^2}\,{\rm Re}\,\Gamma_{\rm b}^{(2)}(q)=
1+q^{-2\varepsilon}
\left(\frac{4}{3}-\frac{2}{3} \delta+\frac{1}{3} {\delta}^2 \right) g^2
\left[\frac{3}{4(8\pi^2)^2\varepsilon}
+{\rm P}+{\cal O}(\varepsilon) \right] \, ,
\end{equation}
where {\rm P} denotes some number which is the contribution of
order ${\varepsilon}^0$.
At $q={\mu}$, (\ref{WFreal}) is commonly defined as the inverse wave
function renormalization constant $(Z^{\phi}_{\rm p})^{-1}$. In terms of the
perturbative renormalized coupling constants $Z^{\phi}_{\rm p}$ is given by
\begin{equation}
\label{zp}
Z_{\rm p}^{\phi}=1-\frac{1}{4(8\pi^2)^2\varepsilon}
\left(4-2 {\delta}_{\rm r}
+{\delta}_{\rm r}^2 \right) g_{\rm r}^2 \, ,
\end{equation}
and together with (\ref{WFim}), we find after the first step of renormalization
\begin{eqnarray}
\frac{\partial}{\partial q^2}\,\Gamma_{\rm r}^{(2)}(q)\!\!&=&\!\!
1+\left(\frac{4}{3} g_{\rm r}^2-\frac{2}{3} v_{\rm r} g_{\rm r}
+\frac{1}{3} v_{\rm r}^2 \right)
\left[{\rm P}-\frac{3}{2(8\pi^2)^2} \ln \left(\frac{q}{\mu}\right)
+{\cal O}(\varepsilon)
\right]
\nonumber \\
\nonumber \\
&+&\!\!i\,2^4 \pi^2 3 \,c_{\rm r}\,
\frac{1}{\varepsilon} \sum_{n=0}^{\infty} \left\{\frac{v_{\rm r}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n
B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\nonumber \right. \\
\nonumber \\
& &\left. \mbox{} \times
\left[\gamma-\frac{1}{2}+\ln \left(\frac{q}{2\mu}\right) \right]
\left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+2+4n}{2}} \exp\left(
\frac{A}{g_{\rm r}}\right) \right\} \, .
\end{eqnarray}
We have again introduced the coupling constants $v_{\rm r}$ and $g_{\rm r}$
which are more convenient for practical applications.
This expression reads in the fully renormalized
couplings (\ref{REN})
\begin{eqnarray}
\lefteqn{\frac{\partial}{\partial q^2}\,\Gamma_{\rm r}^{(2)}(q) =
1+\left(\frac{4}{3} g_{\rm R}^2-\frac{2}{3} v_{\rm R} g_{\rm R} +
\frac{1}{3} v_{\rm R}^2 \right)
\left({\rm P}-\frac{3}{2(8\pi^2)^2} \ln \left(\frac{q}{\mu}\right)
\right)} \nonumber \\
\nonumber \\
& & \mbox{}+i\,2^4 \pi^2 3 \,c_{\rm r}\,
\frac{1}{\varepsilon} \sum_{n=0}^{\infty} \Bigg\{\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n
\nonumber \\
\nonumber \\
& & \qquad
\times \Bigg[2^4 \pi^2 A \,\,
B \left(n+\frac{1}{2},n+\frac{5}{2} \right)\, \frac{4(n+1)}{2n+3}
\left({\rm P}-\frac{3}{2(8\pi^2)^2} \ln \left(\frac{q}{\mu}\right) \right)
\nonumber \\
\nonumber \\
& & \qquad \quad
+B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\left(\gamma-\frac{1}{2}-\ln2+\ln \left(\frac{q}{\mu}\right) \right) \Bigg]
\nonumber \\
\nonumber \\
& &
\times \left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+2+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg\} \, .
\end{eqnarray}
By using the definition of the $B$-function in terms of Gamma-functions one
can easily read off the cancellation of $(1/{\varepsilon}) \ln(q/{\mu})$
singularities for every power $n$ of the coupling constant $v_{\rm R}$.
Therefore we
are left with
\begin{eqnarray}
\frac{\partial}{\partial q^2}\,\Gamma_{\rm r}^{(2)}(q)\!\!&=&\!\!
1+\left(\frac{4}{3} g_{\rm R}^2-\frac{2}{3} v_{\rm R} g_{\rm R}
+\frac{1}{3} v_{\rm R}^2 \right)
\left[{\rm P}-\frac{3}{2(8\pi^2)^2} \ln \left(\frac{q}{\mu}\right)
\right]
\nonumber \\
\nonumber \\
&+&\!\! i\,2^4 \pi^2 3 \,c_{\rm r}\,
\frac{1}{\varepsilon} \left(2^4 \pi^2 A\,\,{\rm P}+\gamma-\frac{1}{2}-\ln2
\right)
\nonumber \\
\nonumber \\
&\displaystyle{\times}&\!\!\!
\sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+2+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg] \, .
\nonumber \\
\end{eqnarray}
Now we can define the nonperturbative minimally subtracted wave function
renormalization to remove the $1/{\varepsilon}$-pole:
\begin{eqnarray}
\label{znp}
Z_{{\rm np}}^{\phi}\!\!&=&\!\!1-
i\,2^4 \pi^2 3 \,c_{\rm r}\,
\frac{1}{\varepsilon} \left(2^4 \pi^2 A\,\,{\rm P}+\gamma-\frac{1}{2}-\ln2
\right)
\nonumber \\
\nonumber \\
& &\qquad \times \sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\nonumber \\
\nonumber \\
& &\qquad \qquad \quad
\times \left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+2+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg] \, .
\end{eqnarray}
The full renormalization constant is defined by
$Z^{\phi}:=Z^{\phi}_{\rm p}\,Z^{\phi}_{{\rm np}}$ and follows
from the product of (\ref{zp}) and (\ref{znp}).
A further important Green function is associated with the composite
field $(1/2) {\phi}^2$:
\begin{equation}
G^{(2,1)}(x_1,x_2;x_3)_{i,j}=\left\langle {\phi}_i(x_1) {\phi}_j(x_2)
\frac{1}{2} {\phi}_k^2(x_3) \right\rangle \, .
\end{equation}
By a straightforward application of the techniques discussed in the
preceding sections we obtain for the corresponding vertex function
\begin{eqnarray}
\lefteqn{ {\rm Im} \, \Gamma_{\rm b}^{(2,1)}(q_1,q_2;q_3)_{i,j}=
-\frac{1}{2} c_{\rm b} \int\limits_0^{\infty} \frac{d\lambda}{\lambda}
\Bigg\{
\left(-\frac{\lambda^{\varepsilon}A}{g}\right)^{\frac{d+6}{2}}
\!\!\! \exp \left(\frac{\lambda^{\varepsilon}A}{g} \right)}
\nonumber \\
\nonumber \\
& &\!\!\!\!\! \times
\prod_{\nu=1}^2 \left[ \left(\frac{q_\nu}{\lambda}\right)^2 \!
\tilde{\phi} \left( \frac{q_\nu}{\lambda} \right) \right]
\hat{\phi} \left(\frac{q_3}{\lambda}\right)
\frac{1}{2} \int\limits_0^{2\pi} d \vartheta u_{L\,i}(\vartheta)
u_{L\,j}(\vartheta)
\exp\left[a \sin^2(2\vartheta)\right] \Bigg\} \, ,
\nonumber \\
\end{eqnarray}
where
\begin{equation}
\hat{\phi}(q)=2^{(d-2)/2}\pi^{(d-4)/2} 3
\left| q \right|^{(4-d)/2} K_{(d-4)/2}\left(\left|q \right|\right) \, .
\end{equation}
Now we can proceed as in the case of wave function renormalization.
Performing the collective coordinates integration over $\lambda$ and
$\vartheta$ we obtain
\begin{eqnarray}
{\rm Im} \,\Gamma_{\rm b}^{(2,1)}(q_1,q_2;q_3)_{i,j}\!\!&=&\!\!
{\delta}_{ij} \, 2^4 \pi^2 3^2 \,c_{\rm b}\,
\frac{1}{\varepsilon} \sum_{n=0}^{\infty} \left\{\frac{(-2\, \delta)^n}{n!}
B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\nonumber \right. \\
\nonumber \\
& &\left. \!\!\! \times
\left[\gamma+\ln \left(\frac{q_3}{2\mu}\right)\right]
\left(-\frac{\mu^{\varepsilon}A}{g} \right)^{\frac{d+4+2n}{2}} \exp\left(
\frac{\mu^{\varepsilon}A}{g}\right) \right\},
\nonumber \\
\end{eqnarray}
which agrees with the usual form ${\Gamma}^{(2,1)}_{{\rm b}\, ij}=
\delta_{ij}\, {\Gamma}^{(2,1)}_{{\rm b}}$.
At $q_3=0$, $q_1^2=q_2^2={\mu}^2$, the vertex function
${\Gamma}^{(2,1)}(q_1,q_2;q_3)$ defines a third renormalization constant,
called $Z^{\phi^2}$, via
\begin{equation}
{\Gamma}^{(2,1)} \equiv \left(Z^{\phi} Z^{\phi^2} \right)^{-1} \, .
\end{equation}
The perturbative $Z^{\phi^2}$ follows from the real part of
${\Gamma}^{(2,1)}$:
\begin{equation}
{\rm Re} \Gamma^{(2,1)}_{\rm b}=1+
g q_3^{-\varepsilon} \left(\frac{4}{3}-\frac{1}{3} \delta \right)
\left(-\frac{3}{8\pi^2\varepsilon}+{\rm Q}+{\cal O}(\varepsilon) \right)\, ,
\end{equation}
where ${\rm Q}$ is a number of order ${\varepsilon}^0$.
In terms of the perturbative renormalized coupling constants, the one-loop
expression of $Z_p^{\phi^2}$ is given by
\begin{equation}
\label{pPP}
Z_{\rm p}^{\phi^2}=1+\frac{1}{8\pi^2\varepsilon}
\left(4-\delta_{\rm r}\right) g_{\rm r} \, .
\end{equation}
Including ${\rm Re} \Gamma^{(2,1)}_{\rm b}$
and ${\rm Im} \Gamma^{(2,1)}_{\rm b}$
we obtain after the perturbative step of renormalization
\begin{eqnarray}
\Gamma_{\rm r}^{(2,1)}(q_1,q_2;q_3)\!\!&=&\!\!
1+\left(\frac{4}{3} g_{\rm r}-\frac{1}{3} v_{\rm r}\right)
\left[{\rm Q}+\frac{3}{8\pi^2} \ln \left(\frac{q_3}{\mu}\right)
+{\cal O}(\varepsilon)
\right]
\nonumber \\
\nonumber \\
&+&\!\!i\,2^4 \pi^2 3^2 \,c_{\rm r}\,
\frac{1}{\varepsilon} \sum_{n=0}^{\infty} \left\{\frac{v_{\rm r}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n
B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\nonumber \right. \\
\nonumber \\
& &\left. \mbox{} \times
\left[\gamma+\ln \left(\frac{q_3}{2\mu}\right) \right]
\left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+4+4n}{2}} \exp\left(
\frac{A}{g_{\rm r}}\right) \right\} \, .
\end{eqnarray}
In the second step of renormalization the $(1/{\varepsilon})\ln (q/{\mu})$
singularities cancel, and we find
\begin{eqnarray}
&\Gamma&\!\!\!\!\!_{\rm r}^{(2,1)}(q_1,q_2;q_3)=
1+\left(\frac{4}{3} g_{\rm R}-\frac{1}{3} v_{\rm R} \right)
\left[{\rm Q}+\frac{3}{8\pi^2} \ln \left(\frac{q_3}{\mu}\right)
\right]
\nonumber \\
\nonumber \\
& &\quad -i\,2^4 \pi^2 3^2 \,c_{\rm r}\,
\frac{1}{\varepsilon} \left(\frac{8\pi^2}{3} {\rm Q}-\gamma+\ln2 \right)
\nonumber \\
\nonumber \\
& &\quad \displaystyle{\times}
\sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+4+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg] \, .
\nonumber \\
\end{eqnarray}
Hence, the nonperturbative $Z^{\phi^2}$ is
\begin{eqnarray}
\label{npPP}
\lefteqn{Z_{{\rm np}}^{\phi} Z_{{\rm np}}^{\phi^2}=1+
i\,2^4 \pi^2 3^2 \,c_{\rm r}\,
\frac{1}{\varepsilon} \left(\frac{8\pi^2}{3} {\rm Q}-\gamma+\ln2 \right)}
\nonumber \\
\nonumber \\
& &\times \sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+4+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg].
\end{eqnarray}
With respect to the leading expansion (\ref{IMG}) this is equal to
$Z_{{\rm np}}^{\phi^2}$.
\section{Renormalization group functions}
After having calculated the imaginary parts of the renormalized couplings
and of the renormalization constants $Z^{\phi}$ and $Z^{\phi^2}$,
we are in the position to derive the nonperturbative renormalization
group functions $\beta(g_{\rm R},v_{\rm R})$, $\gamma(g_{\rm R},v_{\rm R})$
and $\gamma_{\phi^2}(g_{\rm R},v_{\rm R})$ in the usual way. First we derive
the $\beta$-functions from the expressions (\ref{REN}) for $g_{\rm R}(\mu)$
and $v_{\rm R}(\mu)$, respectively, and obtain:
\begin{eqnarray}
\lefteqn{\beta_g (g_{\rm R},v_{\rm R})=} \nonumber \\
& &\mu \frac{\partial}{\partial \mu} g_{\rm R}
=\mu \frac{\partial}{\partial \mu} g_{\rm r}+
i 2^7 3 \pi^4 c_{\rm r} \frac{1}{\varepsilon}
\displaystyle{\sum_{n=0}^{\infty}} \mu \frac{\partial}{\partial \mu}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
B \left( n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
\nonumber \\
& & \qquad \qquad \qquad \qquad \qquad
\times \left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+4+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg]\, ,
\nonumber \\
\lefteqn{\beta_v(g_{\rm R},v_{\rm R})=} \nonumber \\
& &\mu \frac{\partial}{\partial \mu} v_{\rm R}
=\mu \frac{\partial}{\partial \mu} v_{\rm r}-
i 2^7 3 \pi^4 c_{\rm r} \frac{1}{\varepsilon}
\displaystyle{\sum_{n=0}^{\infty}} \mu \frac{\partial}{\partial \mu}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
B \left(n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
\nonumber \\
& & \qquad \qquad \qquad \qquad \qquad
\times \left(\frac{4n}{2n+3}\right)
\left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+4+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg]\, .
\end{eqnarray}
The derivatives are taken at fixed bare couplings $g$ and $v$. These
functions will be finite as $\varepsilon \rightarrow 0$. Furthermore, they
will have a left-hand cut in the complex $g$-plane. On top of the tip of
this cut we find
\begin{eqnarray}
\label{bet1}
\beta_g (g_{\rm R},v_{\rm R})=
\beta_g^{\rm p}(g_{\rm r},v_{\rm r})
\!\!\!\!&-&\!\!\!\!i 2^7 3 \pi^4 c_{\rm r}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
B \left( n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
\nonumber \\
& &
\times \left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+6+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg] \, ,
\nonumber \\
\nonumber \\
\beta_v(g_{\rm R},v_{\rm R})=
\beta_v^{\rm p}(g_{\rm r},v_{\rm r})
\!\!\!\!&+&\!\!\!\!i 2^7 3 \pi^4 c_{\rm r}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
B \left(n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
\nonumber \\
& &
\times \left(\frac{4n}{2n+3}\right)
\left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+6+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg] \, .
\end{eqnarray}
The real parts are the familiar perturbative $\beta$-functions
\begin{eqnarray}
\beta_g^{\rm p}(g_{\rm r},v_{\rm r})\!\!&=&\!\!
-\varepsilon g_{\rm r}+\frac{1}{8\pi^2}
\left(-2 g_{\rm r} v_{\rm r}+v_{\rm r}^2+10 g_{\rm r}^2 \right)\, ,
\nonumber \\
\beta_v^{\rm p}(g_{\rm r},v_{\rm r})\!\!&=&\!\!
-\varepsilon v_{\rm r}+\frac{3}{8\pi^2} \left(
4g_{\rm r} v_{\rm r}-v_{\rm r}^2 \right)\, .
\end{eqnarray}
In the leading contributions the pair of
coupling constants $(g_{\rm r},v_{\rm r})$ may be replaced
by $(g_{\rm R},v_{\rm R})$ in (\ref{bet1}).
This can be proven using (\ref{REN}). The result is
\begin{eqnarray}
\label{betaR}
\beta_g (g_{\rm R},v_{\rm R})=
\beta_g^{\rm p}(g_{\rm R},v_{\rm R})
\!\!\!\!&-&\!\!\!\!i 2^7 3 \pi^4 c_{\rm r}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm R}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
B \left( n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
\nonumber \\
& &
\times \left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+6+4 n}{2}}
\exp \left(\frac{A}{g_{\rm R}} \right) \Bigg]\, ,
\nonumber \\
\nonumber \\
\beta_v(g_{\rm R},v_{\rm R})=
\beta_v^{\rm p}(g_{\rm R},v_{\rm R})
\!\!\!\!&+&\!\!\!\!i 2^7 3 \pi^4 c_{\rm r}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm R}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n
B \left(n+\frac{1}{2},n+\frac{5}{2} \right)
\nonumber \\
\nonumber \\
& &
\times \left(\frac{4n}{2n+3}\right)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+6+4 n}{2}}
\exp \left(\frac{A}{g_{\rm R}} \right) \Bigg] \, .
\end{eqnarray}
Note that the $\beta$-functions have a form expected within a minimal
subtracted scheme:~${\beta}_g(g_{\rm R},v_{\rm R})=
-\varepsilon g_{\rm R} +{\beta}^{\, 4}_g(g_{\rm R},v_{\rm R})$
and ${\beta}_v(g_{\rm R},v_{\rm R})=
-\varepsilon v_{\rm R}+{\beta}^{\, 4}_v(g_{\rm R},v_{\rm R})$, where the
${\beta}^{\, 4}$ denotes the $\beta$-functions in four dimensions.
The critical
large-distance behavior of the correlation functions is defined by the
anomalous dimension of the field $\phi$, which is defined by
\begin{equation}
\gamma(g_{\rm R},v_{\rm R})=\mu \frac{\partial}{\partial \mu} \ln Z^{\phi}=
\mu \frac{\partial}{\partial \mu} \ln Z_{\rm p}^{\phi}+
\mu \frac{\partial}{\partial \mu} \ln Z_{{\rm np}}^{\phi} \, .
\end{equation}
With (\ref{zp}) for $Z^{\phi}_{\rm p}$ and (\ref{znp})
for $Z^{\phi}_{{\rm np}}$ we get the
$\varepsilon$-independent expression
\begin{eqnarray}
\label{gamR}
\lefteqn{\gamma(g_{\rm R},v_{\rm R}) =}
\nonumber \\
\nonumber \\
& &\!\!\!\!\!\frac{1}{2(8\pi^2)^2}
\left(4 g_{\rm R}^2-2 v_{\rm R} g_{\rm R} +v_{\rm R}^2 \right)
+i\,2^4 \pi^2 3 \,c_{\rm r}\,
\left(2^7 \pi^4 3^{-1}\,{\rm P}+\gamma-\frac{1}{2}-\ln2 \right)
\nonumber \\
\nonumber \\
& & \quad \times \sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+4+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg] \, .
\end{eqnarray}
The divergence of the length scale at critical temperature is governed by
the anomalous dimension of the composite field $\phi^2$:
\begin{equation}
\gamma_{\phi^2}(g_{\rm R},v_{\rm R})=
-\mu \frac{\partial}{\partial \mu} \ln Z^{\phi^2}=
-\mu \frac{\partial}{\partial \mu} \ln Z_{\rm p}^{\phi^2}
-\mu \frac{\partial}{\partial \mu} \ln Z_{{\rm np}}^{\phi^2} \, .
\end{equation}
Using (\ref{pPP}) and (\ref{npPP}), we obtain
\begin{eqnarray}
\label{Npp2}
\lefteqn{\gamma_{\phi^2}(g_{\rm R},v_{\rm R})=\frac{1}{8\pi^2}
\left(4 g_{\rm R}-v_{\rm R} \right)
+i\,2^4 \pi^2 3^2 \,c_{\rm r}\,
\left(\frac{8\pi^2}{3} {\rm Q}-\gamma+\ln2 \right)}
\nonumber \\
\nonumber \\
& & \quad \times \sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n B \left(n+\frac{1}{2},n+\frac{3}{2} \right)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+6+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg].
\nonumber \\
\end{eqnarray}
For the leading contributions to the imaginary part, it was possible to replace
$(g_{\rm r},v_{\rm r})$ by $(g_{\rm R},v_{\rm R})$.
We remark that the expansions (\ref{betaR}),(\ref{gamR}) and (\ref{Npp2})
have the correct isotropic limit for $v_{\rm R} \rightarrow 0$.
\section{Generalization to the case $N=3$}
Generalizing (\ref{exp}) to $N=3$ we write:
\begin{eqnarray}
\vec{\phi}\!\!\!&=&\!\!\!\vec{u}_L \phi_c+\vec{u}_L \xi+\vec{u}_{T_1} \eta+
\vec{u}_{T_2} \chi
\nonumber \\
\nonumber \\
\!\!\!&=&\!\!\!
\left( \begin{array}{c}
\sin{\theta} \cos{\vartheta} \\
\sin{\theta} \sin{\vartheta} \\ \cos{\theta}
\end{array} \right) (\phi_c+\xi)+
\left( \begin{array}{c}
-\sin{\vartheta} \\ \cos{\vartheta} \\ 0
\end{array} \right) \eta+
\left( \begin{array}{c}
\cos{\theta} \cos{\vartheta} \\ \cos{\theta} \sin{\vartheta}
\\ -\sin{\theta}
\end{array} \right) \chi \, ,
\end{eqnarray}
where $\theta$ and $\vartheta$ are the rotation angles of the isotropic
instanton.
This change of variables yields the
following expansion of the energy functional:
\begin{eqnarray}
\label{Hex3}
H(\vec{\phi})=&H_c&\!\!\!+\frac{1}{2} \int d^dx \xi M_L \xi+
\frac{1}{2} \int d^dx \eta M_{T_1} \eta+
\frac{1}{2} \int d^dx \chi M_{T_2} \chi
\nonumber \\
\nonumber \\
&-&\!\!\!\frac{\varepsilon 4(2)^{1/2}}{(-g)^{1/2}}
\int d^dx\frac{\lambda^3}{[1+\lambda^2(x-x_0)^2]^2} \xi
\nonumber \\
\nonumber \\
&-&\!\!\!
(-8g)^{1/2} \int d^dx \frac{\lambda}{1+\lambda^2(x-x_0)^2}
\left[\xi^3+(1-\delta)(\xi \eta^2+\xi \chi^2) \right]
\nonumber \\
\nonumber \\
&+&\!\!\!
\delta \,\, F(g,\vartheta,\theta,\xi,\eta,\chi)
\end{eqnarray}
with $M_L$ and $M_{T_1}=M_{T_2}=M_T$ of equation (\ref{MLT}).
For brevity, we have written down explicitly
only the terms responsible for the
leading contributions in the expansion (\ref{IMG}).
The remaining terms are denoted collectively by
$\delta \,\, F(g,\vartheta,\theta,\xi,\eta,\chi)$.
The classical contribution to the energy functional is
\begin{eqnarray}
H(\phi_{1c},\phi_{2c},\phi_{3c})\!\!\!&=&\!\!\!
-\frac{\lambda^{\varepsilon}}{g} \frac{8\pi^2}{3}
\left[1-\frac{\varepsilon}{2}\left(2+\ln \pi+\gamma \right) \right]
\nonumber \\
\nonumber \\
& &\!\!\!\!\!
-\frac{\lambda^{\varepsilon} \delta }{2g} \, \frac{8\pi^2}{3}
\left[\sin^4{\theta} \sin^2(2\vartheta)+\sin^2(2\theta) \right]+
{\cal O}\left(\frac{ \delta }{g}\, \varepsilon \right).
\end{eqnarray}
With similar steps as for $N=2$ we obtain
the bare imaginary parts of the vertex functions:
\begin{eqnarray}
\lefteqn{{\rm Im} \,\Gamma_{\rm b}^{(2M)} (q_m)_{i_1,i_2,\ldots,i_{2M}} = }
\nonumber \\
\nonumber \\
& &\!\!\!\!\!
\displaystyle{-c_{\rm b}^{(3)}} \int\limits_0^{\infty}
\frac{d\lambda}{\lambda} \Bigg\{
\lambda^{d-M(d-2)} \left(-\frac{\lambda^{\varepsilon}8\pi^2}{3g}
\right)^{(d+3+2M)/2} \!\!\! \exp \left(\frac{\lambda^{\varepsilon}A}{g}
\right) \prod^{2M}_{\nu=1}
\left(\frac{q_{\nu}}{\lambda}\right)^2 \!
\tilde{\phi} \left( \frac{q_{\nu}}{\lambda} \right)
\nonumber \\
\nonumber \\
& & \!\!\!\!\!
\times \frac{1}{2} \int\limits_0^{2\pi} \int\limits_0^{\pi}
\sin{\theta} d \theta d \vartheta \, \, u_{L\,i_1}(\vartheta,\theta)
u_{L\,i_2}(\vartheta,\theta) \cdots u_{L\,i_{2M}} (\vartheta,\theta)
\nonumber \\
\nonumber \\
& & \qquad \times
\exp \left[a \sin^4{\theta} \sin^2(2\vartheta)+a \sin^2(2\theta) \right]
\Bigg\} \left[1+{\cal O}(\varepsilon,g) \right] \, ,
\end{eqnarray}
where
\begin{eqnarray}
c_{\rm b}^{(3)}\!\!\!&=&\!\!\!
2^{-5/6} 3 \pi^{-13/3} \exp
\Bigg[
\frac{1}{3\varepsilon}
\left(11-4\delta+2{\delta}^2\right)+
\frac{5 \zeta^{'} (2)}{\pi^2}-
\frac{9}{2} \, \gamma -\frac{17}{4}
\nonumber \\
\nonumber \\
& &\qquad \qquad
+\frac{16}{9} \, \delta -\frac{7}{81} \, {\delta}^2
+{\cal O} \left( {\delta}^3 \right) \Bigg] \, ,
\quad (i_k\, =\, 1,2,3)\, ,
\end{eqnarray}
with $a$, $A$ as in (\ref{aA}).
After evaluating the integrals over the collective coordinates
$\lambda$, $\theta$, and $\vartheta$ we obtain
for the imaginary part of the four-point vertex function
\begin{eqnarray}
\lefteqn{{\rm Im} \, \Gamma_{\rm b}^{(4)}(q_m)_{i,j,k,l}= } \nonumber \\
\nonumber \\
& &2^8 3^2 \pi^4 \,c_{\rm b}^{(3)}\,\, \frac{1}{\varepsilon}
\sum_{n=0}^{\infty}
\Bigg\{ \frac{(-2\, \delta)^n}{\Gamma(2n+7/2)}
\left(\frac{g}{A}\right)
\left(-\frac{\mu^{\varepsilon}A}{g}\right)^{\frac{d+7+2n}{2}}
\!\!\! \exp \left(\frac{\mu^{\varepsilon}A}{g} \right)
\nonumber \\
\nonumber \\
& &\!\!\!\times
\sum_{p=0}^n \frac{\Gamma(3+2n-p)\,\Gamma\left(p+\frac{1}{2} \right)}
{(n-p)!\, p! \, \Gamma\left(2(n-p)+3\right)}
\Bigg[
3\Gamma\left(n-p+\frac{3}{2}\right)^2 \,\,(S_{ijkl}-\delta_{ijkl})
\nonumber \\
\nonumber \\
& & \qquad
+\Gamma\left(n-p+\frac{1}{2}\right) \Gamma\left(n-p+\frac{5}{2}\right)
\,\, \delta_{ijkl}
\Bigg] \Bigg\}\, .
\end{eqnarray}
Similarly the bare imaginary parts of
$\frac{\partial}{\partial q^2}\Gamma^{(2)}(q)$
and $\Gamma^{(2,1)}(q_1,q_2;q_3)$ are found to be
\begin{eqnarray}
\lefteqn{\frac{\partial}{\partial q^2}
{\rm Im} \,\Gamma_{\rm b}^{(2)}(q)_{i,j}=
\delta_{ij} \, 2^4 \pi^2 3 \,c_{\rm b}^{(3)}
\left[\gamma-\frac{1}{2}+\ln \left(\frac{q}{2\mu}\right) \right]}
\nonumber \\
\nonumber \\
& & \times
\frac{1}{\varepsilon} \sum_{n=0}^{\infty} \Bigg[\frac{(-2\, \delta)^n}{n!}
\, I(n) \,
\left(-\frac{\mu^{\varepsilon}A}{g} \right)^{\frac{d+3+2n}{2}} \exp\left(
\frac{\mu^{\varepsilon}A}{g}\right)
\Bigg] \, ,
\end{eqnarray}
and
\begin{eqnarray}
{\rm Im} \,\Gamma_{\rm b}^{(2,1)}(q_1,q_2;q_3)_{i,j}\!\!&=&\!\!
\delta_{ij} \, 2^4 \pi^2 3^2 \,c_{\rm b}^{(3)}\,
\frac{1}{\varepsilon} \sum_{n=0}^{\infty} \left\{
\frac{(-2\, \delta)^n}{n!} I(n)
\nonumber \right. \\
\nonumber \\
& &\left. \!\!\!\!\!\!\!\! \times
\left[\gamma+\ln \left(\frac{q_3}{2\mu}\right)\right]
\left(-\frac{\mu^{\varepsilon}A}{g} \right)^{\frac{d+5+2n}{2}} \exp\left(
\frac{\mu^{\varepsilon}A}{g}\right) \right\},
\nonumber \\
\end{eqnarray}
with
\begin{equation}
I(n) \equiv
\sum_{p=0}^n
\left( \begin{array}{c}
n \\ p
\end{array} \right)
B \left(2+2n-p,p+\frac{1}{2}\right)
B \left(n-p+\frac{1}{2},n-p+\frac{3}{2} \right).
\end{equation}
Now we apply the extended renormalization scheme. Using the coupling
constants $v$ and $g$, we obtain for the nonperturbative renormalized
couplings
\begin{eqnarray}
\label{c3R}
g_{\rm R}\!\!\!&=&\!\!\!g_{\rm r}
+i 2^7 3 \pi^4 c_{\rm r}^{(3)} \frac{1}{\varepsilon}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n J_g(n)
\left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+5+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg] \, ,
\nonumber \\
\nonumber \\
v_{\rm R}\!\!\!&=&\!\!\!v_{\rm r}
-i 2^7 3 \pi^4 c_{\rm r}^{(3)} \frac{1}{\varepsilon}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm r}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n J_v(n)
\left(-\frac{A}{g_{\rm r}} \right)^{\frac{d+5+4 n}{2}}
\exp \left(\frac{A}{g_{\rm r}} \right) \Bigg]\, ,
\nonumber \\
\end{eqnarray}
with
\begin{eqnarray}
\label{J}
\displaystyle{ J_g(n)}
\!\!\!&\equiv&\!\!\!
\sum_{p=0}^n
\left( \begin{array}{c}
n \\ p
\end{array} \right)
B \left(3+2n-p,p+\frac{1}{2}\right)
B \left(n-p+\frac{1}{2},n-p+\frac{5}{2} \right)\, ,
\nonumber \\
\nonumber \\
\displaystyle{ J_v(n)}
\!\!\!&\equiv&\!\!\!
\sum_{p=0}^n \Bigg\{
\left( \begin{array}{c}
n \\ p
\end{array} \right)
\, \left[ \frac{4(n-p)}{2(n-p)+3} \right] \,
B \left(3+2n-p,p+\frac{1}{2}\right)
\nonumber \\
\nonumber \\
& & \qquad \qquad \qquad \qquad \quad \times
B \left(n-p+\frac{1}{2},n-p+\frac{5}{2} \right) \Bigg\} \, ,
\end{eqnarray}
and
\begin{equation}
\label{cc33}
c_{\rm r}^{(3)}=3\,\, 2^{-5/6} \pi^{-5/2} \exp \bigg[
\frac{5\zeta^{'}(2)}{\pi^2}-\frac{8}{3} \, \gamma-\frac{7}{12}
+{\cal O}({\delta}_{\rm r})+{\cal O}({\delta}_{\rm r} \, g_{\rm r}) \bigg]\, .
\end{equation}
For the nonperturbative renormalization constants $Z^{\phi}$ and $Z^{\phi^2}$
we find
\begin{eqnarray}
Z_{{\rm np}}^{\phi}\!\!&=&\!\!1-
i\,2^4 \pi^2 3 \,c_{\rm r}^{(3)}\,
\frac{1}{\varepsilon} \left(2^4 \pi^2 A\,\,{\rm P}+\gamma-\frac{1}{2}-\ln2
\right)
\nonumber \\
\nonumber \\
& & \mbox{} \times \sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n \, I(n) \,
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+3+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg] \,
\end{eqnarray}
and
\begin{eqnarray}
\lefteqn{Z_{{\rm np}}^{\phi} Z_{{\rm np}}^{\phi^2}=1+
i\,2^4 \pi^2 3^2 \,c_{\rm r}^{(3)}\,
\frac{1}{\varepsilon} \left(\frac{8\pi^2}{3} {\rm Q}-\gamma+\ln2 \right)}
\nonumber \\
\nonumber \\
& &\times \sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n I(n)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+5+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg].
\end{eqnarray}
To calculate the nonperturbative renormalization
group functions $\beta(g_{\rm R},v_{\rm R})$, $\gamma(g_{\rm R},v_{\rm R})$
and $\gamma_{\phi^2}(g_{\rm R},v_{\rm R})$,
we simply repeat the calculations of the case $N=2$.
Using the expressions (\ref{c3R}) for $g_{\rm R}$ and $v_{\rm R}$ yields the
$\beta$-functions
\begin{eqnarray}
\label{B3}
\lefteqn{\beta_g (g_{\rm R},v_{\rm R})=}
\nonumber \\
& &\beta_g^{\rm p}(g_{\rm R},v_{\rm R})
-i 2^7 3 \pi^4 c_{\rm r}^{(3)}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm R}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n J_g(n)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+7+4 n}{2}}
\exp \left(\frac{A}{g_{\rm R}} \right) \Bigg]\, ,
\nonumber \\
\nonumber \\
\lefteqn{\beta_v(g_{\rm R},v_{\rm R})=}
\nonumber \\
& &\beta_v^{\rm p}(g_{\rm R},v_{\rm R})
+i 2^7 3 \pi^4 c_{\rm r}^{(3)}
\displaystyle{\sum_{n=0}^{\infty}}
\Bigg[ \frac{v_{\rm R}^n}{n!} \left(\frac{3}{4\pi^2}\right)^n J_v(n)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+7+4 n}{2}}
\exp \left(\frac{A}{g_{\rm R}} \right) \Bigg]\, .
\nonumber \\
\nonumber \\
\end{eqnarray}
The expressions for the nonperturbative anomalous dimension
of the field $\phi$ and the composite field $\phi^2$ are given by
\begin{eqnarray}
\label{G3}
\gamma(g_{\rm R},v_{\rm R})\!\!\!\!&=&\!\!\!\!
\gamma^{\rm p}(g_{\rm R},v_{\rm R})
+i\,2^4 \pi^2 3 \,c_{\rm r}^{(3)}\,
\left(2^7 \pi^4 3^{-1}\,{\rm P}+\gamma-\frac{1}{2}-\ln2 \right)
\nonumber \\
\nonumber \\
& & \quad \times
\sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n I(n)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+5+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg] \, ,
\end{eqnarray}
and
\begin{eqnarray}
\label{Npp3}
\gamma_{\phi^2}(g_{\rm R},v_{\rm R})\!\!&=&\!\!
\gamma_{\phi^2}^{\rm p}(g_{\rm R},v_{\rm R})
+i\,2^4 \pi^2 3^2 \,c_{\rm r}^{(3)}\,
\left(\frac{8\pi^2}{3} {\rm Q}-\gamma+\ln2 \right)
\nonumber \\
\nonumber \\
\!\!&\times&\!\! \sum_{n=0}^{\infty} \Bigg[\frac{v_{\rm R}^n}{n!}
\left(\frac{3}{4\pi^2}\right)^n I(n)
\left(-\frac{A}{g_{\rm R}} \right)^{\frac{d+7+4n}{2}} \exp\left(
\frac{A}{g_{\rm R}}\right) \Bigg] \, ,
\end{eqnarray}
respectively.
\section{Discussion and Conclusions}
Our main results are the real and imaginary parts (\ref{B3})--(\ref{Npp3})
of the renormalization-group functions $\beta$, $\gamma$ and $\gamma_{\phi^2}$
in terms of the nonperturbative renormalized
couplings $g_{\rm R}$ and $v_{\rm R}$. Via the dispersion relation (\ref{DR}),
we obtain the large-order behavior (\ref{LO2}).
It is useful to check the expansion terms proportional to powers
of $1/|g|$ accompanying the $v^n$ in the imaginary part of
each renormalization-group function.
If the power of $1/|g|$ is denoted by $p(n)$, we note that $p(n)$
is the same as for the corresponding vertex function
before the integration over the dilatation parameter $\lambda$.
This integration with the measure $d\ln \lambda$
reduces $p(n)$ by $1$ [see (\ref{lambda})].
However this effect is compensated by the differentiation with respect
to the logarithm of the scale parameter $\mu$ in the definition of the
renormalization group functions.
The power $p(n)$ in the vertex functions can simply be explained:
For a $2M$-point vertex function with $k$ $\phi^2$-insertions
there is first a power $(1/|g|)^{(M+k)}$ from the fields.
This is multiplied by a factor $(1/|g|)^{(d+(N-1)+1)/2}$ from the $d$
translational, $(N-1)$ rotational, and one dilatational would-be zero modes.
The classical contribution to the energy functional yields
a factor $(1/|g|)^2$ for each power of the anisotropic constant $v$.
Hence, for the renormalization
group function which follows at the one-loop level from the vertex
function $\Gamma^{(2M,k)}$, the value of $p(n)$ is given by
\begin{equation}
p(n)=\frac{d+N+2(M+k)+4n}{2} \, .
\end{equation}
Inserting the imaginary parts into a dispersion relation (\ref{DR}),
we can estimate the large-order coefficients of the corresponding
series expansions
\begin{equation}
f(g,v)=\sum_{k,n=0}^{\infty} f_{kn}\, g^k \, v^n \, ,
\end{equation}
where $n \ll k$.
The function $f$ stands for
${\beta}_g$, ${\beta}_v$, $\gamma$ and $\gamma_{\phi^2}$, respectively.
The rules how to go from the imaginary parts of $f$ to $f_{kn}$ are
known from (\ref{LO2}).
For instance, the large-order coefficients of the $\beta$-functions for $N=3$
are given by
\begin{eqnarray}
\label{concl}
\beta_{g, \,kn}\!\!&=&\!\!c_{\rm r}^{(3)}2^7 3 \pi^3
\left(\frac{3}{4\pi^2}\right)^n
\frac{ J_g(n) }{n!}
\nonumber \\
\nonumber \\
& &\qquad \times
(-1)^k \left(\frac{3}{8\pi^2}\right)^k k! \, k^{2n+(d+5)/2}
\left[1+{\cal O}(1/k)\right] \, ,
\nonumber \\
\nonumber \\
\beta_{v, \,kn}\!\!&=&\!\!-c_{\rm r}^{(3)}2^7 3 \pi^3
\left(\frac{3}{4\pi^2}\right)^n
\frac{ J_v(n) }{n!}
\nonumber \\
\nonumber \\
& &\qquad \times
(-1)^k \left(\frac{3}{8\pi^2}\right)^k k! \, k^{2n+(d+5)/2}
\left[1+{\cal O}(1/k)\right] \, ,
\end{eqnarray}
with $J$, $c_{\rm r}^{(3)}$ from (\ref{J}) and (\ref{cc33}), respectively.
The leading large-$k$ behavior can be checked by means of a simple
combinatorial analysis. We start with a series in terms of the standard
couplings $u$ and $v$ [see (\ref{HN})]. For illustration,
we consider the contribution of the pure $u$-powers.
It is known from a theory with only one coupling constant, that the
coefficients have the large-order behavior
\begin{equation}
\label{lao}
f_l \, u^l \longrightarrow \gamma \, (-\alpha)^l \,
\Gamma(l+b+1)\, u^l \qquad (l \gg 1) \, .
\end{equation}
Changing over to our couplings ($u = g-v$),
the right-hand side of (\ref{lao}) contributes
\begin{equation}
\gamma \, (-1)^{l-k}\, (-\alpha)^l \, \Gamma(l+b+1) \,
\left(\begin{array}{c}
l \\ k
\end{array} \right) \, g^k \, v^{l-k} \, ,
\end{equation}
with $$\left(\begin{array}{c}
l \\ k
\end{array} \right)=\frac{\Gamma(l+1)}{k!\, (l-k)!} \, .$$
Replacing $l$ by $l=n+k$, the contribution of $g^k v^n$ has the form
\begin{equation}
f_{kn}\, g^k \, v^n \longrightarrow \gamma \, (\alpha)^n \, (-\alpha)^k \,
\frac{\Gamma(k+n+b+1)\, \Gamma(k+n+1)}{k! \, n!} \, g^k \, v^n \, .
\end{equation}
Now, for large $k$, the $\Gamma$-functions can be approximated by
\begin{equation}
\Gamma(k+\delta+1) \longrightarrow
k! \, k^{\delta} \, \left(1+{\cal O}(1/k) \right) \, .
\end{equation}
Thus in the region $n \ll k$ we obtain
\begin{equation}
f_{kn} \sim (-\alpha)^k \, k! \, k^{2n+b} \, ,
\end{equation}
in agreement with (\ref{concl}).
As stated in the introduction, the stable cubic fixed point is expected to lie
in the vicinity of the isotropic fixed point.
Since $v$ becomes very small in this region, reasonable results should
be obtained by resumming the $g$-series accompanying each power $v^n$.
This will be done in a forthcoming publication.
\newpage
|
\section{Introduction}
\label{sec:egbrintroduction}
\setcounter{equation}{0}
Quantum cosmology studies the relation between the observed universe and
its boundary conditions in the hope that a natural {\it theory} of the
boundary condition might emerge (see \cite{Halliwell} for an outstanding
review of this enterprise.) Assessment of a particular theory requires an
understanding of its implications for the present day. To that end, this
paper elaborates on work of Gell-Mann and Hartle \cite{TSA} and Davies and
Twamley \cite{DT} by examining the observable consequences for the diffuse
extragalactic background radiation (EGBR) of one possible class of boundary
conditions, those that are imposed time symmetrically at the beginning
and end of a closed universe \cite{TSA}, and sketches some of the
considerable difficulties in rendering this kind of model credible.
Assuming such difficulties do not vitiate the consistency of
time symmetric boundary conditions as a description of our universe,
the principal conclusion is that these boundary conditions imply that the
bath of diffuse optical radiation from extragalactic sources be at least
twice that due only to the galaxies to our past, and possibly much more.
In this sense, observations of the EGBR are observations of the final
boundary condition.
This conclusion will be seen to follow (section \ref{sec:limit}) because
radiation
from the present epoch can propagate largely unabsorbed until the universe
begins to recollapse \linebreak
(\cite{DT}, and section \ref{sec:opacity}), even if the
lifetime of the universe is
very great. By time symmetry, light correlated with the thermodynamically
reversed galaxies of the recollapsing phase must exist at the present
epoch. The {\it minimal} predicted ``excess" EGBR in a universe with
time symmetric boundary conditions turns out to be
consistent with present observations (section \ref{sec:observations}),
but improved observations and modeling of galactic evolution will
soon constrain this minimal prediction very tightly.
In addition, many physical
complications with the ansatz that time symmetric boundary conditions
provide a reasonable and consistent description of the observed universe
will become apparent. Thus this work may be viewed as outlining some
reasons why even if very long-lived, our universe is probably
not time symmetric.
The plan of the paper is as follows. Section \ref{sec:motivations}\
discusses a model universe that will define the terms of the investigation.
Section \ref{sec:tsbc}\ provides some perspective on doing physics with
boundary conditions at two times with an eye toward section
\ref{sec:difficulties}, where some aspects of the
reasonableness of two time boundary conditions not immediately related to
the extragalactic background radiation are discussed. Section
\ref{sec:opacity}\ generalizes and confirms
work of Davies and Twamley \cite{DT} in showing that for processes of
practical interest, our future light cone (FLC) is transparent all the
way to the recollapsing era over a wide range of frequencies, even if the
universe is arbitrarily long-lived. Section \ref{sec:limit}\ explains
why this fact
implies a contribution to the optical extragalactic background radiation
in a universe with time symmetric boundary conditions in excess of
that expected without time symmetry. In the course of this explanation,
some rather serious difficulties will emerge in the attempt to reconcile
time symmetric boundary conditions, and a transparent future light cone,
into a consistent model of the universe which resembles the one in which
we live. Section \ref{sec:observations}\ compares the predictions of
section \ref{sec:limit}\ for the optical EGBR to models of the
extragalactic background light and observations of it.
Section \ref{sec:summation}\ is reserved for summation and conclusions.
\subsection{Motivations and A Model}
\label{sec:motivations}
The possibility that the universe may be time symmetric has been raised
by a number of authors
\cite{Gold,Wheeler1,Wheeler2,Hawking1,Zeh1,Zeh2,Zeh3,KZ}.
Of course, what is meant is not {\it exact} time symmetry, in the sense that
a long time from now there will be another Earth where everything happens
backwards. Rather, the idea is that the various observed ``arrows of time"
are directly correlated with the expansion of the universe, consequently
reversing themselves during a recontracting phase if the universe is closed.
Of central interest is the thermodynamic arrow of entropy increase,
from which other time arrows, such as the psychological arrow of
perceived time or the arrow defined by the retardation of radiation,
are thought to flow \cite[are some reviews]{PTA,Zeh2,HPM}.
However, the mere reversal of the universal expansion is insufficient to
reverse the direction in which entropy increases
\cite{Schulman1,PTA,Wheeler1,Wheeler2,Penrose1,Page1,Hawking2,HLL}.
In order to construct a quantum physics for matter in a
recollapsing universe in which the thermodynamic arrow naturally reverses
itself, it appears necessary to employ something like the time neutral
generalization of quantum mechanics \cite{Cocke,ABL,griffiths84,Unruh,TSA}
in which boundary conditions are imposed near both the
big bang {\it and} the big crunch. These boundary conditions take the form
of ``initial" and ``final" density operators\footnote{I retain this
terminology even though these operators may not be of trace class, for
example in the familiar case where the final ``boundary condition" is
merely the identity operator on an infinite dimensional Hilbert space.}
which, when CPT-reverses of one another, define what is meant here by a
time symmetric universe.\footnote{In other words, the effective
decoherence functional for matter in a time symmetric universe
is a canonical decoherence functional $d(h,h') =
{\rm tr}[\rho_{\omega} h^{\dagger}\rho_{\alpha} h']/({\rm tr}\rho_{\alpha}\rho_{\omega})$,
with $\rho_{\omega} = ({\cal CPT})^{-1}\rho_{\alpha} ({\cal CPT})$ \cite{TSA}.}
In such a model the collection of quantum mechanical
histories is time symmetric in the sense that each history in a
decohering set ({\it i.e.\ }a set of histories in which relative
probabilities may be consistently assigned) occurs with the same
probability as its CPT-reverse \cite{TSA,Page2}. (As CP violation
is small, there is for many purposes no difference between CPT- and
T-symmetry with a T invariant Hamiltonian.) I do not describe these ideas
in further detail because very little of the {\it formalism} of generalized
quantum theory will be directly applied in this paper,
but it is worth mentioning that in order for the resulting time
symmetric quantum theory to have non-trivial predictions,
the initial and final density operators must not represent pure
states.\footnote{As the first in an occasional series of
comments directed to those familiar with the ideas of generalized quantum
theory, this is because if the initial and final density operators are
pure, at most two histories can simultaneously be assigned probabilities
\cite{TSA,dhgqm,craig95}, {\it i.e.\ }the maximum number of histories
in any weakly decohering set is two!
Complete information about the history of the universe must be encoded
in the boundary conditions. Details of the formalism of generalized
quantum theory may be found in, {\it e.g.,} \cite{LesH}. Generalized
quantum mechanics with boundary conditions at two times is discussed
more extensively in \cite{TSA} and \cite{dhgqm,craig95}.}
Interesting theories therefore have boundary conditions
which are quantum statistical ensembles.
The interest in applying this class of quantum theories (namely,
theories with CPT-related boundary conditions) in cosmology
lies in the idea that the manifest arrow of time we observe is an
emergent property of the universe, and not built directly into its
structure by asymmetric dynamical laws or an asymmetric choice of
boundary conditions. Dynamical laws are believed to be
CPT-symmetric, so an asymmetric choice of boundary conditions is usually
cited as the explanation for the existence of a definite arrow of
time which flows in the same direction throughout the observable part of
spacetime \cite{PTA,Zeh2,HPM}. However, it is worth investigating whether
this assumption is {\it required} of us by making the alternate, apparently
natural ansatz that the boundary conditions on a closed universe are (in a
relevant sense) equivalent at the beginning and end of time, instead of the
more usual assumption that the initial condition is somehow special and the
final condition one of ``indifference," {\it i.e.\ }determined entirely by
the past. Another point of view is that, as noted in \cite{TSA}, these
alternative choices are in some sense opposite extremes. It is therefore
of interest to determine whether they are distinguishable on experimental
grounds, employing time symmetric boundary conditions as a laboratory for
testing the sensitivity of physical predictions to the presence of a final
boundary condition.
For the benefit of those eager to proceed to the definite physical
predictions of sections \ref{sec:opacity}\ and \ref{sec:egbr}, I now
specify a model in which they might be expected to arise.
(The cautiousness of this statement is explained in the sequel.)
Sections \ref{sec:tsbc}, \ref{sec:amplifications},
and \ref{sec:difficulties}\ elaborate on the physics expected
in a universe with
two-time, time symmetric boundary conditions; here I merely summarize
what is required from those sections for a complete statement of the
assumptions.
The model of the universe considered here consists in:
\begin{itemize}
\item A fixed closed, homogeneous and isotropic background spacetime,
{\it viz.} a $k=+1$ Friedmann-Robertson-Walker (FRW) universe.
The evolution
of the scale factor is determined from Einstein's equations by the
averaged matter content of the universe. (Inhomogeneities in the matter
content can be treated as additional matter fields on this background.)
\item Boundary conditions imposed on the matter content
through a canonical decoherence functional $d_{\alpha\omega}$ \cite{dhgqm}
with CPT-related density operators $\rho_{\alpha}, \rho_{\omega}$
that describe the
state of matter at some small fiducial scale factor, near what would in
the absence of quantum gravitational effects be the big bang and big
crunch, but outside of the quantum gravity regime. The matter state
described by one of
these density operators reflects the presumed state of the early universe,
namely matter fields in apparent thermal equilibrium at the temperature
appropriate to the fiducial scale factor and the total amount of matter
in the universe. Spatial fluctuations should be consistent with present
day large scale structure, say being approximately scale invariant and
leading to an amplitude of order $10^{-5}$ at decoupling in order to be
consistent with recent COBE results \cite{COBE}. Possible further
conditions on $\rho_{\alpha}, \rho_{\omega}$ are discussed at the end
of this subsection.
\end{itemize}
The essence that a choice of boundary conditions with these apparently
natural characterisics intends to capture is that of a universe in which
the cosmological principle holds, which is smooth (and in apparent thermal
equilibrium) whenever it is small, and which displays more or less familiar
behaviour when larger. Most of the conclusions of this investigation
really only
rely on these general properties, but for the sake of definiteness a
fairly specific model which has the right general physical characteristics
is offered.
However, as will be repeatedly emphasized in the sequel, in models with
boundary conditions at two times, not only does the past have implications
for the future, but the future has implications for the past. Therefore,
in any attempt to model the {\it observed} universe with time symmetric
boundary conditions, we need to make sure it is possible to integrate them
into a self-consistent picture of the universe as we see it today.
This means in particular that we must be prepared
for the possibility that the early universe in a time symmetric universe
may have properties different from those expected in a universe with an
initial condition only. The model boundary conditions sketched above are
not intended to be so restrictive as to rule out such differences, and
hence must be taken with a grain of salt or two. The issue is then
whether or not these properties are consistent with observation.
Indeed, the prediction of an ``excess" optical EGBR (to be discussed
in section \ref{sec:egbr}) in a universe with time symmetric boundary
conditions is precisely of this character. (Of course, the most
extreme possibility is that a universe
burdened with these boundary conditions would look {\it nothing at all}
like the universe in which we actually live.) Variations on this theme
will recur frequently in the following sections.
With these boundary conditions \cite{TSA}, the time neutral
generalized quantum mechanics of Gell-Mann and Hartle defines
an effective quantum theory for
matter in the universe which may be imagined to arise from some other,
more fundamental, quantum cosmological theory of the boundary conditions.
(Presumably, the fundamental quantum cosmological decoherence functional
incorporates the
quantum mechanics of the gravitational field as well,
a complication that is not addressed in this paper.)
In this connection it is perhaps worth mentioning that it was once
claimed \cite{Hawking1} that the no-boundary proposal for the initial
condition\cite[\cite{Halliwell} is an excellent review]{HH} implied just
this sort of effective theory in that it appeared to require that the
universe be smooth whenever it was small. Thus a fundamental theory of
an initial condition {\it only} apparently could be decribed by an
effective theory with {\it two-time} boundary conditions. However, this
claim has since been recanted \cite{Hawking2} due to a mathematical
oversight. More generally, the no-boundary wave function does not
appear to be a good candidate for a boundary condition imposed time
symmetrically at both ends of a closed universe because it is a pure
state \cite{TSA}, which as noted above yield quantum theories in which
essentially all physical information is encoded in the boundary
conditions alone. For contrast, see \cite{KZ}, in which it is asserted
that the only sensible way to interpret the Wheeler-DeWitt equation
necessitates a boundary condition requiring the universe to be
smooth whenever it is small.
In order to allow definite predictions to be made, the final key assumption
of the model is that, for a suitable class
of physically interesting coarse grainings ({\it e.g.\ }coarse grainings
defining the domain of classical experience, or local quantum mechanics
experiments), the probabilities for such coarse grained physical histories
unfold near either boundary condition (relative to the total lifetime of the
universe) in a fashion insensitive to the presence of the boundary
condition at the other end, {\it i.e.\ }as if the other boundary condition
were merely the identity. As discussed in section \ref{sec:tsbc},
simple stochastic models \cite{Cocke,TSA,Schulman2,Schulman3,Schulman4}
suggest that this holds for
processes for which the ``relaxation time"
of the process to equilibrium is short compared to the total time between
the imposition of the boundary conditions (see section \ref{sec:tsbc}),
which expectation is
rigorously supported in the case of Markov processes \cite{Schulman5}.
Therefore, for the sake of brevity this predictive ansatz shall often be
referred to as the ``Relaxation Time Hypothesis" (RTH). A more careful
statement of the RTH would identify the classes of coarse grained histories
(presumably at least those decribing short relaxation time processes at
times sufficiently close to, for instance, the big bang)
for which conditional probabilities (when defined) are, in a universe with
boundary conditions $\rho_{\alpha}$, $\rho_{\omega}$, supposed to be close
to those of a universe with $\rho_{\omega}=1$.\footnote{Some subtleties
regarding
equivalences between boundary conditions at one and two times are being
concealed here, for which see \cite{craig95}.}
It would also define ``close" and ``short relaxation time process" more
precisely. (Thus, a rigorous statement of what is meant by the RTH
requires a definite mathematical model. Because the concerns of this
investigation
encompass a variety of complicated physical processes, with the entire
universe considered as a single physical system, I do not attempt that
here. In specific cases the intuitive content of the ansatz ought
to be clear enough.) In order to exploit the RTH to its fullest, I shall,
when convenient, assume also that the universe is close to the critical
density, so that its lifetime is very long. This plausibly realistic
assumption maximizes the possibility that a model of the kind given
above accurately describes our universe.
With these assumptions, this model universe might be expected to closely
resemble the universe as it is observed today {\it if} most familiar and
important physical processes are examples of such ``short relaxation
time processes.'' In particular, under the assumption that they are,
predictions we expect of a single initial condition $\rho_{\alpha}$ only
({\it i.e.\ }$\rho_{\omega}= 1$) can (by the RTH)
be assumed to hold near the initial condition in the model with
CPT-related initial and final conditions. Such predictions should
include those regarding
inflation, relative particle abundances, and the formation of large (and
small) scale structure. Because of the CPT-related boundary condition
at the big crunch, a qualitatively similar state of affairs is then
expected in the recollapsing era, but CPT reversed. As the
thermodynamic arrow is caused fundamentally by gravitational collapse
driving initially smooth matter away from
equilibrium \cite[for example]{Penrose1,Zeh2}, the arrow of
entropy increase near the big crunch will run in the opposite direction
to that near the big bang.\footnote{Some discussion of the state of the
universe when it is large, which somehow must interpolate between these
opposed thermodynamic arrows, is provided in sections \ref{sec:tsbc},
\ref{sec:amplifications}, and
\ref{sec:difficulties}.} Observers on planets in the recollapsing phase
will find their situation indistinguishable from our own, with all time
arrows aligned in the direction of increasing volume of the universe.
It is this interesting (if unconventional) state of affairs which leads
to the conclusion that observations of the EGBR can reveal the presence
of a final boundary condition that is CPT-related to the initial, even
if the lifetime of the universe is very great. How this comes about is
the topic of section \ref{sec:egbr}.
Before proceeding, some points made in the preceeding paragraph require
qualification. (Readers whose only interest is the extragalactic
background
radiation should procede directly to section \ref{sec:opacity}. This
discussion, and that of the next subsection, are positioned here for
the sake of unity and perspective.) First, the ``relaxation times" for
many important physical processes in a recollapsing universe
{\it do not} appear to be short compared to the lifetime of the universe,
even if that lifetime is arbitrarily long. The result described in section
\ref{sec:opacity}, that light from the present epoch can propagate
unabsorbed into the recollapsing phase, is an excellent example of this
\cite{DT,TSA}. Some physical difficulties with the consistency of time
symmetry to which this gives rise will become apparent in section
\ref{sec:amplifications}. Further examples of such physical
difficulties relating to issues such as gravitational collapse, the
consequent
emergence of a familiar thermodynamic arrow, and baryon decay in a universe
with time symmetric boundary conditions are discussed briefly in section
\ref{sec:difficulties}.
The self-consistency of such models is thus in doubt, with these kinds
of complications constituting arguments against the possibility that our
universe possesses time symmetric boundary conditions. That is, it appears
that time symmetry is {\it not} consistent with the central predictive
ansatz that physics near either boundary condition is practically
insensitive
to the presence of the other, and it appears likely that a universe with
time symmetric, low (matter) entropy boundary conditions would look
{\it nothing like} the universe in which we actually live.
Nevertheless, the strategy of sections \ref{sec:opacity}\
and \ref{sec:egbr}\
is to assume that the time symmetric picture is consistent with the gross
characteristics of the observed universe and see what it predicts.
Because of the prediction of a diffuse optical extragalactic background
radiation that is testably different from that in a universe which is not
time symmetric, we are provided with a two pronged attack on the
hypothesis of time symmetric
boundary conditions as a realistic description of our universe:
the lack of a self-consistent picture of the observed universe,
and observations of the EGBR.
Second, for completeness it should be mentioned that in the context of
the time neutral generalized quantum mechanics assumed here, there
are further restrictions on the viability of time symmetric models
as a realistic description of our universe which arise from the
requirements of decoherence and the emergence of approximately classical
behaviour.\footnote{It is not obvious that boundary conditions of the
character noted above satisfy these restrictions, but if sufficient
conditions obtain for the RTH to hold then it is at least plausible
that they do. This is because such requirements might be expected
to be more severe the more strongly correlated are the detailed states
of the expanding and recollapsing eras. This should become clearer
in section \ref{sec:tsbc}.}
These topics are discussed by Gell-Mann and Hartle \cite{TSA}.
Third, it is important to note that the conclusions of
section \ref{sec:egbr}\ regarding the EGBR depend essentially
on the assumed {\it global} homogeneity
and isotropy, {\it i.e.\ }the ``cosmological principle."
Thus, while inflation (apparently a generic consequence of
quantum field theory in a small universe)
can be taken to be a prediction of the assumed boundary conditions if
they are imposed when the universe is sufficiently small, the
popular point of view (somewhat suspect anyway) that inflation provides
an {\it explanation} of homogeneity and isotropy inside the horizon is
not a prediction of the effective theory of the universe considered here;
it is an assumption. However, the other good things inflation does for us
would still qualify as predictions.
One further point requires mention. In the conventional picture of
inflation, the matter in the universe is in a pure state (say, some vacuum
state) when small. After inflation and reheating the matter fields
appear to be, according to local coarse grainings, in thermal equilibrium,
the states of different fields being highly entangled. Furthermore, the
correlations required to infer that the complete state is actually pure
have been inflated away. Nevertheless, the quantum state is of course
still pure \cite{CC}. As noted previously, pure states are not viable
candidates for the initial and final conditions. Therefore,
in order to fit the conventional picture of inflation into an effective
theory of the kind considered in this paper,
the $\rho_{\alpha}$ and $\rho_{\omega}$ defined above must be a local
description in the sense that they not contain the information required to
know that the state from which they were inferred was actually pure,
{\it i.e.\ }they must coarse-grain this information away. (The
alternative is to employ an inflationary scenario in which the universe
is not required to be in a pure state.) The unpleasantness of this
restriction could be taken as an argument against the use of CPT-related
boundary conditions in a fundamental quantum cosmological theory.
Finally, for the cognoscenti of generalized quantum
mechanics \cite[for example]{LesH}, there is a related
observation that is even more interesting. Field
theoretic Hamiltonians are CPT-invariant, so that if the CPT-related
$\rho_{\alpha}$ and $\rho_{\omega}$ depend only on the Hamiltonian (as for
instance an exactly thermal density operator does), $\rho_{\alpha}$ and
$\rho_{\omega}$ are actually {\it equal}. (The same observation holds for
merely T-related boundary conditions if the Hamiltonian is T invariant.
However, CPT seems the more relevant symmetry if the Hamiltonian and
boundary conditions of the considered effective theory of the universe
arise from some more fundamental theory.) This is a potential difficulty
for the sample model
given above. A simple extension of a result of Gell-Mann and
Hartle \cite[section 22.6.2]{TSA} to the case where the background
spacetime is an expanding universe shows that the only dynamics allowed
with identical initial
and final boundary conditions is trivial: for alternatives allowed in
sets of histories in which probabilities may be assigned at all, the time
dependence of probabilities for alternative outcomes is essentially
{\it independent of the Hamiltonian},
and is due only to the expansion of the universe.\footnote{To be more
precise, consider the projections which appear in sets of histories that
decohere for some $\rho_{\alpha}=\rho_{\omega}.$ For time independent
Hamiltonians, time dependence of probabilities for such projections
arises only if the projections are time-dependent in the Schr\"{o}dinger
picture. Due to the expansion, this will be the case for many projections
onto (otherwise time-independent)
quantities of physical interest in an expanding universe.}
In the present case this may be interpreted as a prediction that
a universe which is required to be in thermal equilibrium whenever
it is small {\it must remain so when large,} at least in the context
of the theoretical structure of Gell-Mann and Hartle. (Classically the
expectation is the same, so this result can hardly be written off as an
artifact of the formalism.)
Now, the evidence suggests that matter in the the early universe
{\it was} in local thermal equilibrium. However, the inhomogeneities
in the matter required to generate large scale structure must also be
described by the boundary conditions, and, even in the usual case where
there is only an initial condition, it is after all gravitational
condensation which drives the
appearance of a thermodynamic arrow.\footnote{More precisely, smoothly
distributed matter in thermal equilibrium is {\it not} an equilibrium
state when the gravitational field is included. Equilibrium states in
the presence of gravitation have clumpy matter in them.}
In the present case (the model with the two-time boundary conditions
characterized above), the meaning of these observations is that one way
to break the CPT invariance of the boundary conditions is for the
specification
of the deviations from perfect homogeneity and isotropy to be CPT
non-invariant, in order that the theory admit interesting dynamics.
As a specific example,
I briefly consider the common circumstance where cosmological matter
is given a hydrodynamic description. Scalar-type adiabatic pertubations
may be completely
specified by, {\it e.g.,} the values of the energy density perturbation
and its (conformal-) time derivative over a surface of constant conformal
time \cite{MFB}. (Scalar-type pertubations are the ones of interest, as
they are the ones which may exhibit instability to collapse.)
As it is intended here that gravity is treated classically, such a
specification might come in the form of a probability distribution for
approximately scale invariant
metric and energy density fluctuations averaged over macroscopic scales.
If this probability distribution reflects the underlying FRW homogeneity
and isotropy (and is thus in particular P-invariant), in order to break
T-invariance the distribution must distinguish between opposing signs of
the (conformal-) time derivatives of the energy density perturbations.
An alternative way to break the equality of CPT-related boundary conditions
is for the locally observed matter-antimatter asymmetry to extend across
the
entire surface of constant universal time.
(The recollapsing era will then be
antimatter dominated in a CPT symmetric universe, consequently requiring
that there be some mechanism to permit baryon decay.) This possibility may
appear more attractive, but it is not without significant complications.
These are addressed briefly at the end of section \ref{sec:amplifications}.
\subsection{Physics with Time Symmetric Boundary Conditions}
\label{sec:tsbc}
As noted above, the time neutral generalized quantum mechanics of Gell-Mann
and Hartle \cite{TSA} with CPT-related initial and final density operators
yields a CPT symmetric ensemble of quantum mechanical histories, in the
sense that each history in the collection is accompanied with equal
probability by its CPT-reverse. The {\it collection} of histories (in
each decohering set) is therefore statistically time symmetric (STS).
Now, a set of histories can be statistically time symmetric without any
individual history possessing qualitatively similar (if time-reversed)
characteristics near both ends. Completely time asymmetric histories can
constitute a statistically time symmetric set
if each history in the collection is accompanied by its
time reverse \cite{TSA,Page2}. With two-time boundary
conditions, however, more is required. That is, there is a
distinction between a CPT invariant {\it set} of histories, in which
each history in a decohering set is accompanied, with equal
probability, by its CPT reverse, and the considerably more restrictive
notion of a statistically time symmetric universe studied here. In
virtue of the boundary conditions at both the beginning and the end of
time, probable histories (in a set which decoheres with these boundary
conditions) will in general have both initial and final states which to
some extent resemble the boundary
conditions $\rho_{\alpha}, \rho_{\omega}$ \cite{craig95}.
Perhaps the clearest way to understand this is to consider the
construction
of a classical statistical ensemble with boundary conditions at
two times. The probability for each history in the ensemble may be found
in the following way \cite{Cocke,TSA}. The boundary conditions are given
in the form of phase space probability densities for the initial and final
states of the system. Pick some initial state. Evolve it forward to the
final time. Weight this history by the product of the probabilities that
it meets the initial and final conditions, and divide by a normalizing
constant so that all the probabilities sum to one. Thus, roughly speaking,
in order for a
history to be probable in this two-time ensemble both its initial and final
state must be probable according to the initial and final probability
densities, respectively. (For a deterministic classical system, two-time
statistical boundary conditions are equivalent to a boundary condition at
one time constructed in the obvious way. In quantum mechanics this is
no longer true due to the non-commutability of operators at different
times. Thus, while
the described algorithm is merely a useful heuristic for understanding the
implications of two-time boundary conditions {\it classically}, it is more
essential quantum-mechanically.) What is not so clear is that the
resulting probable histories look anything like the probable histories
in the
ensemble with just, say, the initial condition. In general they will not.
Classically, this is merely the statement that the probability measure on a
space of classical deterministic solutions defined by two-time boundary
conditions will in general be quite different than the measure defined
by the initial probability density only.
For classical systems with stochastic dynamics or for quantum statistical
systems, the simple models studied by Cocke \cite{Cocke},
Schulman \cite{Schulman2,Schulman3,Schulman4,Schulman5},
and others yield some insight into what is required for evolution near
either boundary condition to be influenced only by that boundary
condition. Here I
merely summarize the intuitively transparent results of this work in the
context of time symmetric boundary conditions. In particular, the
emphasis will be on boundary conditions which represent ``low entropy"
initial and final states.
In the absence of a final condition, an initially low entropy state
generally implies that evolution away from the initial state displays
a ``thermodynamic arrow" of entropy increase
(relative to the coarse graining defining the
relevant notion of entropy; see \cite{Zeh2} for a pertinent overview.)
In an ensemble with two-time, time symmetric, low entropy boundary
conditions, under what conditions will a familiar ``thermodynamic arrow"
appear near either boundary condition? A pertinent observation is that,
roughly speaking, in equilibrium all arrows of time disappear. Therefore,
if the total time between the imposition of the boundary conditions is much
longer than the ``relaxation time" of the system to equilibrium (defined as
the characteristic time a similar system with an initial condition only
takes to relax), it might be expected that entropy will increase in the
familiar fashion away from either boundary condition. That is,
following the coarse-grained evolution of the system forward from the
initial condition, entropy increases at the expected rate until the
system achieves equilibrium. The system languishes in equilibrium for a
time, and then entropy begins to decrease again until the entropy
reaches the low value demanded by the final condition. Thus, this is a
system in which the ``thermodynamic arrow of time" reverses itself, which
reversal is enforced by the time symmetric, low entropy boundary
conditions. There is no cause to worry about the coexistence of
opposed thermodynamic arrows. Histories for which entropy increases away
from either boundary condition are readily compatible with both boundary
conditions, and the probable evolutions are those in which, {\it near either
boundary condition}, there is a familiar thermodynamic arrow of entropy
increase away from the nearest boundary condition. There, the other
boundary condition is effectively invisible. Essentially this is
because in equilibrium all states compatible with constraints are
equally probable; the system ``forgets" its boundary conditions.
On the other hand, if there is not time enough for equilibrium to be
reached there must be some reconciliation between the differing arrows of
increasing entropy \cite{Cocke,PTA,Penrose1,Penrose2,TSA,Schulman5}.
As only histories which satisfy the required boundary conditions are
allowed,
it can be anticipated that the statistics of physical processes that would
ordinarily (in the absence of the final boundary condition) lead to
equilibration would be affected because histories which were probable with
an initial condition only are no longer compatible with the final condition.
The rate of entropy increase is slowed, and in fact the entropy may never
achieve its maximum value. In other words, the
approach to equilibrium is suppressed by the necessity of complying with
the low entropy boundary condition at the other end. More probable in
this two-time ensemble is that the state {\it will continue to resemble the
low entropy boundary conditions.} The entire evolution is sensitive to the
presence of {\it both} boundary conditions.
In fact, as noted already, simple models of statistical systems with a
boundary condition of low entropy at two times bear out these
intuitively transparent
expectations \cite{Cocke,Schulman2,Schulman3,Schulman4}.
Moreover, in the case
of Markov processes Schulman has demonstrated the described behaviour
{\it analytically} \cite{Schulman5}. This confluence of intuitive
clarity, and, for simple systems, analytic and (computerized)
experimental evidence will be taken as suggestive that the significantly
more complicated physical system considered in this work (matter fields
in a dynamic universe) behaves in a qualitatively similar fashion when
burdened with time symmetric boundary conditions.\footnote{Of course,
the physics must at the completely fine-grained level be consistent
with whatever dynamics the system obeys, including such
constraints as conservation laws. As a moment's reflection on classical
Hamiltonian systems with two-time phase space boundary conditions reveals,
this may be a severe constraint! A purely stochastic dynamics (no
dynamical conservation laws) allows systems great freedom to respect
the RTH, and conclusions drawn from
the behaviour of such systems may therefore be misleading. The
restrictions on sensible boundary conditions in time neutral generalized
quantum mechanics noted in the last paragraphs of section
\ref{sec:motivations}\ (which arise essentially as a result of the
requirements of decoherence) are examples of phenomena with no counterpart
in stochastic systems. See also the discussion of gravitational collapse
in section \ref{sec:collapse}.}
The lesson of this section is that the place to look for signs of
statistical time symmetry in our universe is in physical processes
with long ``relaxation times," or more generally in any process
which might couple the expanding and recollapsing eras \cite{TSA}.
Such processes might be constrained by the presence of a final boundary
condition. Obvious candidates include decays of long-lived
metastable states \cite{Wheeler1,Wheeler2,Schulman3,Schulman5,TSA},
gravitational collapse
\cite[see also \cite{Schulman3}]{Penrose1,Penrose2,Laflamme}
and radiations \cite{DT,TSA} with great penetrating power such as
neutrinos, gravitational waves, and possibly
electromagnetic radiation. Thus, there are a variety of tests a
cosmological model with both an initial and a final condition must pass
in order to provide a plausible description of our universe.
The next two sections discuss the extragalactic background radiation as an
example of a physical prediction that is expected to be sensitive to the
presence of a final boundary condition (time symmetrically related to the
initial.) I focus on electromagnetic radiation because it is the most
within our present observational and theoretical grasp, but the essence
of the discussion is relevant to any similar wave phenomenon. Finally,
in section \ref{sec:difficulties}\ I
offer a few comments on some other issues that need to be addressed
in any attempt to describe our universe by a model such as the one
sketched in section \ref{sec:motivations}.
\section{The Opacity of the Future Light Cone}
\label{sec:opacity}
\setcounter{equation}{0}
The aim of this section is to extend arguments of Davies and Twamley
\cite{DT} showing that a photon propagating in intergalactic space is
likely to survive until the epoch of maximum expansion (assuming the
universe to be closed), no matter how long the total lifetime of the
universe. That is, the future light cone is essentially transparent
over a wide range of frequencies for extinction processes relevant
in the intergalactic medium (IGM), with an optical depth
of at most $\tau \sim .01$ at optical frequencies. The physics in this
result is that, in cases of physical interest, the dilution of
scatterers due to the expansion of the universe wins out over the
extremely
long path length the photon must traverse. As a consequence, the
integrated background of light from galaxies in the expanding phase will
still be present in the recollapsing phase. (As explained in section
\ref{sec:egbr}, it
is this fact which implies that there is an ``excess" EGBR in a time
symmetric universe that is not associated with galaxies to our past.)
To show
this, I compute, for a fairly general class of absorption coefficients, the
optical depth between the present epoch and the moment of maximum expansion.
For realistic intergalactic extinction processes this optical depth turns
out to be small. Indeed, the opacity of the future light cone turns out
to be dominated by ``collisions" of intergalactic photons with other
galaxies, if they are regarded as completely opaque hard spheres. Even in
the limit that the lifetime of the universe $T$ becomes infinite
($\Omega \rightarrow 1$ from above) all of the relevant processes yield
finite optical depths. As this is essentially the limit of a flat
universe, it is no surprise that extremely simple expressions result.
The results of this section are modest extensions of the work of Davies
and Twamley \cite{DT}. For the intergalactic extinction mechanisms and
at the frequencies they consider, the formulae for the opacity derived
here give numbers in agreement
with their results (using the same data for the IGM, of course.)
The present work is of slightly broader applicability
in that the opacity is evaluated for
a fairly general class of frequency dependent extinction coefficients
(not just for a few specific processes), and its behaviour as the lifetime
of the
universe becomes arbitrarily long is determined. It turns out that quite
generally, the asymptotic limit is in fact of the order of magnitude of
the {\it upper} limit to the opacity in a closed universe. These results
are less general than that of \cite{DT} in that I do not include the
effects of a cosmological constant (which makes the perturbative analysis
below significantly more awkward.) However, a small cosmological constant
does not effect the qualitative nature of the conclusions of this section.
\subsection{The Future Light Cone Can Be Transparent}
\label{sec:transparent}
Optical depth $\tau$ is defined by
\begin{equation}
d\tau = \Sigma \, dl,
\end{equation}
where $\Sigma$, the linear extinction coefficient, is the fractional
loss of flux per unit (proper) length $l$, and is given microscopically by
\begin{equation}
\Sigma = \sigma n \label{eq:egbrsigmadef}
\end{equation}
for incoherent scattering from targets with cross section $\sigma$ and
proper number density $n$ (this neglects stimulated emission and scattering
into the line of sight.) Given $\tau$, the flux density along the line
of sight thus obeys
\begin{equation}
i(l) = i_{0}e^{-\tau (l)}.
\end{equation}
Put another way, the probability a photon will propagate a distance $l$
without being absorbed is $ e^{-\tau (l)} $. For further details, see
for example \cite{RL}.
In order to compute $\tau$ we need $\Sigma (l)$. In the approximation
(appropriate to the calculation of optical depths between the present and
the moment of maximum expansion) that the universe is exactly described by
closed, dust-filled Friedmann-Robertson-Walker, it turns out to be helpful
to trade in the dependence on proper length for time. The metric is
\begin{equation}
ds^{2} = a^{2}\, [ -d\eta^{2} + d\Omega_{3}^{2} ] ,
\end{equation}
where
\begin{equation}
d\Omega_{3}^{2} = d\chi^{2} + \sin^{2}\chi \, d\Omega_{2}^{2}
\end{equation}
is the metric on the unit 3-sphere, and
\begin{equation}
dt = a \, d\eta
\end{equation}
relates the cosmological time (proper time in the cosmological rest
frame) to the conformal time $\eta $. For dust the time dependence
of the scale factor $a$ can be expressed parametrically as
\begin{equation}
a(\eta) = M \, (1-\cos \eta ), \label{eq:a=}
\end{equation}
so that
\begin{equation}
t(\eta ) = M \, (\eta - \sin \eta ). \label{eq:t=}
\end{equation}
The lifetime of the universe is then $ T = 2\pi M$. Here $ M =
\frac{4\pi}{3}\rho a^{3} $ is a constant as the universe expands,
$\rho $ being the mass density of the dust. $M$ is related to
more familiar cosmological parameters by
\begin{equation}
M = \frac{1}{H_{0}}\frac{q_{0}}{(2q_{0}-1)^{\frac{3}{2}}}.
\label{eq:M}
\end{equation}
Employing the symmetry of the model to take a photon's path as radial,
$ds^{2} = 0$ gives $dl = dt = a \, d\eta $, from which
\begin{equation}
d\tau = \Sigma a \, d\eta. \label{eq:dtau=}
\end{equation}
The high symmetry of Friedmann-Roberston-Walker also means that nearly
all the relevant physical quantities simply scale as a power of $a$.
Thus, as will be seen explicitly in the next subsection, it is necessary
to consider only extinction coefficients of the form
\begin{equation}
\Sigma = \Sigma_{0} \, {\left( \frac{a_{0}}{a}\right)}^{p+1}
, \label{eq:sigma=}
\end{equation}
where $a_{0}$ is some fiducial scale factor (conventionally the present
one) and $p$ is a number.
The goal is to compute the optical depth between now and the moment of
maximum expansion. ``Now" will be taken to be the time $t_{0}$ from the
big bang to the present. For absorption coefficients of the form
(\ref{eq:sigma=}), the optical depth of the future light cone is
\begin{eqnarray}
\tau & = & \int_{\tau(t_{0})}^{\tau(T/2)}\! d\tau \nonumber \\
& = & \int_{\eta_{0}}^{\pi }\! \Sigma \, a \, d\eta \nonumber \\
& = & \Sigma_{0} \, t_0 \, g_{p}(\eta_{0}) \label{eq:tau=}
\end{eqnarray}
using (\ref{eq:a=}) and (\ref{eq:dtau=}). Here $\eta_{0}$ is
the conformal time of the present epoch, and $g_{p}(\eta_{0})$
is the dimensionless function
\begin{eqnarray}
g_{p}(\eta_{0})& \equiv & \left( \frac{M}{t_{0}}\right)
{\left( \frac{a_{0}}{M}\right)}^{p+1}
\int_{\eta_{0}}^{\pi} \frac{d\eta }{(1-\cos \eta )^{p}} \nonumber \\
& = & \frac{(1-\cos \eta_{0} )^{p+1}}{\eta_{0}-\sin\eta_{0}}
\int_{\eta_{0}}^{\pi} \frac{d\eta}{(1-\cos \eta )^{p}}
. \label{eq:g=}
\end{eqnarray}
For integral and half-integral $p$ explicit evaluation of $g_{p}(\eta_{0})$
is possible, but not terribly illuminating. In the limit that the total
lifetime of the universe $T$ is very long compared to $t_{0}$, however,
simple expressions for any $p$ result. (This is no surprise as the
results must approach those of a flat universe.) One straightforward
procedure involves inverting (\ref{eq:t=}) to get a power series in
${\left(\frac{t_{0}}{M}\right)}^{\frac{1}{3}}$ for $\eta_{0}$, and using
this
to evaluate the asymptotic behaviour of $g_{p}(\eta_{0})$ as $M$ becomes
large relative to $t_{0}$, which is held fixed. ($T=2\pi\, M \gg t_{0}$
corresponds to $\eta_{0} \ll 1$.) It is then tedious but straightforward
to show that
\begin{equation}
g_{p} \sim
\left\{ \begin{array}[c]{ll}
\frac{3}{2p-1} \left[ 1+\frac{p+1}{10(2p-3)}
{\left( \frac{6t_{0}}{M} \right) }^{\frac{2}{3}} \right]
& \mbox{$ p > \frac{1}{2} \ ; \ p \not= \frac{3}{2} $} \\
\frac{3}{2}
\left[ 1-\frac{1}{12}{\left( \frac{6t_{0}}{M} \right)}^{\frac{2}{3}}
\ln\left(\frac{6t_{0}}{M}\right) \right]
& \mbox{$p = \frac{3}{2}$}\\
\ln\left(\frac{M}{t_{0}}\right)
\left[ 1-\frac{3}{40}{\left(\frac{6t_{0}}{M}\right) }^{\frac{2}{3}}
\right]
& \mbox{$p=\frac{1}{2} $}\\
3\pi \frac{\Gamma (1-2p)}{\Gamma^{2} (1-p)}
{\left( \frac{M}{6t_{0}}\right) }^{\frac{1-2p}{3}} - \frac{3}{1-2p}
& \mbox{$-\frac{1}{2} < p < \frac{1}{2}$} \\
12{\left( \frac{M}{6t_{0}}\right) }^{\frac{2}{3}} - \frac{9}{5}
& \mbox{$ p = -\frac{1}{2}$} \\
3\pi \frac{\Gamma (1-2p)}{\Gamma^{2} (1-p)}
{\left(\frac{M}{6t_{0}}\right) }^{\frac{1-2p}{3}}
\left[ 1-\frac{p+1}{20}{\left( \frac{6t_{0}}{M}\right) }^{\frac{2}{3}}
\right]
& \mbox{$ p < -\frac{1}{2}$}
\end{array} \right.
\label{eq:g}
\end{equation}
In each case only the leading order correction in $\frac{t_{0}}{M}$
has been retained. The most important thing to notice is that
for $p > \frac{1}{2}$,
$g_{p}$ is perfectly finite even as the lifetime of the universe becomes
arbitrarily big, and as $\frac{t_{0}}{M}$ becomes very small
the opacity converges to the value it would have in a flat universe.
It is clear that for $p > \frac{3}{2}$ the $\Omega_{0} = 1$ result is a
local lower limit on the opacity of the future light cone, as may be
verified directly also from the available exact results. However, for
reasonable $p$ the corrections to the flat universe result are only a
factor of order one. (In fact, examination of the exact results reveals
that the {\it maximum} of $g_{p}$ as one varies $\frac{t_{0}}{M}$ is at
most $20\%$ larger than the flat universe result for $p \sim {\rm few}$.
This is a good thing, because unless $\Omega$ is fairly close to one,
$\frac{t_{0}}{M}$ is not a particularly small parameter!
In terms of familiar cosmological parameters,
\begin{equation}
\frac{t_{0}}{M}= \left[ \cos^{-1}({q_{0}}^{-1}-1)
-{q_{0}}^{-1}(2q_{0}-1)^{\frac{1}{2}}\right].)
\label{eq:t/M}
\end{equation}
Physically, what's at work is the competition between the slower
expansion rates of universes with larger $\Omega_{0}$'s, which tends to
increase the opacity because the scattering medium isn't diluted as
rapidly, and the decrease in the opacity due to the shortened time
between the present epoch and the moment of maximum expansion.\footnote{It
will be noticed by combining (\ref{eq:M}) and (\ref{eq:t/M}) that taking
the limit $\Omega_{0} \rightarrow 1$ holding $t_{0}$ fixed requires $H_{0}$
to vary as well, converging to the flat universe relation $H_{0} =
\frac{2}{3t_{0}}$. It is possible to repeat the entire analysis holding the
observable quantity $H_{0}$ fixed instead of $t_{0}$ (for this purpose
the more standard redshift representation is more useful than that in
terms of conformal time used above), but unsurprisingly the conclusions
are the same: the opacity is always finite for $p > \frac{1}{2}$;
as $\Omega_{0}$ approaches one, the opacity approaches the flat universe
result; and the {\it maximum} opacity for reasonable $p$ is only a factor
$ \ltwid 1.2$
times the flat universe result. The resultant opacities are of course
related in these limits via $t_{0} = \frac{2}{3H_{0}}$. Similarly,
it is possible to perform a related analysis of more complicated
extinction coefficients than
(\ref{eq:sigma=}), for example incorporating the exponential behaviour
encountered in free-free absorption (see (\ref{eq:Sff})) or in modeling
evolving populations of scatterers with, for instance, a Schecter function
type profile. However, these embellishments are not required in the sequel,
and the techniques are tedious and fairly ordinary,
so space will not be taken to describe them here.}
To summarize, all of the processes relevant to extinction
in the intergalactic medium have extinction coefficients that
can be bounded above by a coefficient
of the form (\ref{eq:sigma=}) with $p > \frac{1}{2}$.
Using the limiting relationship
$t_{0} = \frac{2}{3H_{0}}$, we have from (\ref{eq:tau=}) and (\ref{eq:g})
the simple result that for these processes, the upper limit to the opacity
between the present epoch and the moment of maximum expansion, no matter
how long the total lifetime of the universe, is of order
\begin{equation}
\tau = \frac{2}{2p-1}\frac{\Sigma_{0}\, c}{H_{0}}.
\label{eq:tau}
\end{equation}
(I have returned to conventional units in this formula.)
\subsection{The Opacity of the Future Light Cone}
\label{sec:flc}
In this section I apply the asymptotic formula (\ref{eq:tau}) for the upper
limit to the optical depth of the FLC in a
long-lived universe to show that if our universe is closed, photons escaping
from the galaxy are (depending on their frequency) likely to survive into
the recollapsing era. That is, the finite optical depths computed in
the previous section are actually small for processes of interest in the
intergalactic medium (IGM). For simplicity, I focus on photons
softer than the ultraviolet at the present epoch; the cosmological
redshift makes it necessary to consider absorption down to very low
frequencies.
It is important to note that in employing standard techniques for
computing opacities the effects of the assumed statistical time symmetry
of the universe are being neglected. As discussed in section
\ref{sec:tsbc}\ and in section \ref{sec:difficulties}, when
the universe is very large the thermodynamic and gravitational behaviour of
matter will begin to deviate from that expected were the universe not
time symmetric. Due to the manifold uncertainties involved here it is
difficult to approach the effects of time symmetric boundary conditions
on the opacity of the
future light cone with clarity.\footnote{For example, how is scattering of
light by a ``thermodynamically reversed" medium to be treated, as when
light from the expanding era reaches the intergalactic medium in the
recollapsing phase? The standard account assumes incoherent scattering.
Thus a laser beam shone on a plasma is diffused. Time-reversing this
description yields {\it extremely} coherent scattering from the plasma
which reduces its entropy. Thus scattering or absorption of light
correlated
with sources (such as galaxies) in the expanding phase by material in the
recollapsing phase appears to require entropy reducing (according to the
observers of the recollapsing era) correlations in the matter there, in
contradiction with the presumed local thermodynamic arrow (and
with the RTH), in order to yield what there appears as emission.
This is just the sort of detailed connection between the expanding and
recollapsing eras which would lead one to expect physical predictions in
a time symmetric model, even very near one of the boundary conditions,
to be very different than those in a model with an initial condition
only (section \ref{sec:tsbc}). This complication is closely related
to the difficulty, mentioned in section \ref{sec:retardation}, in
deriving the retardation of radiation in a universe
which is time symmetric and in which the future light cone is transparent.}
I shall assume they are not such as to increase it.
This is reasonable as the dominant contribution to the opacity comes
when the universe is smallest, where in spite of the noted complications
the RTH is assumed to hold.
What are the processes relevant to extinction of photons in the
intergalactic medium? Because the IGM appears to consist in hot,
diffuse electrons, and perhaps a little dust \cite{BFR}, extinction
processes to include are Thomson scattering, inverse bremsstrahlung
(free-free absorption), and absorption by dust. In addition, absorption
by material in galaxies (treated as completely black in order to gauge
an upper limit) is important. These processes will treated in turn.
(A useful general
reference on all these matters is \cite{RL}.) The conclusion will be
that while absorption by galaxies and Thomson scattering are most
significant above the radio, none of these processes pose a serious
threat to a photon that escapes from our galaxy. This confirms the
results of Davies and Twamley \cite{DT}, who however did not consider the
possibly significant interactions with galaxies.
Consequently I will be brief. Some results of Davies and Twamley
regarding absorption mechanisms which may be important when the
universe is very large and baryons have had time to decay are
quoted at the end of this section. These do not appear to be
significant either.
(For high energy photons Compton scattering, pair production,
photoelectric absorption by the apparently very small amounts of neutral
intergalactic hydrogen, and interactions with CMBR photons will be
important, but as none are significant below the ultraviolet I do not
discuss them here. All can be treated by the same methods as the
lower energy processes.)
To begin, following Davies and Twamley \cite{DT}, I quote
Barcons {\it et al.\ }\cite{BFR} on current beliefs regarding the state
of the IGM in the form
\begin{eqnarray}
n_{{\rm H_{II}}}
& = & \delta \, n_{0} \, {\left( \frac{a_{0}}{a} \right) }^{3}
\nonumber \\
T_{{\rm H_{II}}}
& = & \epsilon \, T_{0} \, {\left( \frac{a_{0}}{a} \right) }^{2}
\label{eq:IGM}
\end{eqnarray}
where
\begin{eqnarray}
n_{0} & = & 1.12 \, h^{2} \, 10^{-7} \, {\rm\,cm}^{-3}, \ \
\delta \in (1,10) \nonumber \\
T_{0} & = & 10^{4} \, {\rm\,K} , \ \
\epsilon \in (1,10^{3})
\label{eq:IGMparams}
\end{eqnarray}
with the values $\delta = \epsilon = 1$ somewhat preferred by the
authors. In addition, the present upper limit on a smoothly distributed
component of neutral hydrogen is about
$n_{{\rm H_{I}}} < 10^{-12} \, {\rm\,cm}^{-3}$.
Thus, the intergalactic medium consists in hot (but
non-relativistic) electrons, protons, and essentially no neutral
hydrogen. The lack of distortions in the microwave background indicates
its relative uniformity, at least to our past. From now on, $n$ and $T$
simply will be used to refer to the number density and temperature of
intergalactic electrons.
Finally, very little is known about a possible diffuse component of
intergalactic dust \cite{BFR,WR}, except that there is probably very
little of it. Most dust seems to be clumped around galaxies. Therefore
I will ignore possible extinction due to it, subsuming it into the ``black
galaxy" opacity. Davies and Twamley \cite{DT} make some
estimates for one model for the dust, finding its contribution to the
opacity insignificant. At any rate, models for the absorption
coefficient due to dust \cite{Peebles,HW} all give a cross section $\sigma$
that falls with increasing wavelength, $\sigma \sim 1/{{\lambda}^{q}}$ with
$1 \leq q \leq 4$, so that
$\Sigma = \sigma \, n \propto {\left( \frac{a_{0}}{a}\right)}^{q+3}$
(neglecting of course a clumping factor expressing the fact that
clumping decreases the opacity.) Thus
$p_{{\rm dust}} = q + 2 > \frac{1}{2}$,
the dust opacity is bound to be finite, and with a small present
density of diffuse dust it is not surprising to find its contribution
to be small.
Before considering the optical depth due to interactions with
intergalactic electrons, I will show that it is reasonable to
approximate that most photons escaping our galaxy will travel freely
through intergalactic space. That is, few photons will end up running
into another galaxy.
\subsubsection*{Collisions with Galaxies}
\label{sec:collisions}
Drastically overestimating the opacity due to galaxies by pretending
that any photon which enters a galaxy or its halo will be absorbed by
it (the ``black galaxy" approximation), and taking the number of
galaxies to be constant,
\begin{eqnarray}
\label{eq:hardspheresigma}
\Sigma_{{\rm gal}} & = & \sigma \, n \nonumber \\
& = & \sigma \, n_{0} \, {\left( \frac{a_{0}}{a} \right) }^{3},
\end{eqnarray}
where $\sigma$ is the cross-sectional area of a typical galaxy and
$n_{0}$ is their present number density. Thus from (\ref{eq:tau}), the
upper limit on the opacity due to collisions with galaxies is
\begin{equation}
\tau = \frac{2}{3}\, \frac{\sigma \, n_{0} \, c}{H_{0}}.
\end{equation}
As noted above, this is finite (even as the lifetime of the universe
becomes very large) because the dilution of targets due to
the expansion of the universe is more important than the length of the
path the photon must traverse.
Notice that assuming target galaxies to be perfectly homogeneously
distributed only overestimates their ``black galaxy" opacity. Volume
increases faster than cross-sectional area, so clustering reduces the
target area for a given density of material. As galaxy clustering is
not insignificant today and will only increase up to the epoch of
maximum expansion even in a time symmetric universe, the degree of
overestimation is likely to be significant.
Taking
$n_{0}\sim .02\, h^{3}\, {\rm\,Mpc}^{-3}$,
$\sigma = \pi\, r_{{\rm gal}}^{2}$
(where $r_{{\rm gal}} \sim 10^{4}\, h^{-1}\, {\rm\,pc}$), and
$H_{0} \sim \frac{1}{3} \cdot 10^{-17}\, h\, {\rm\,s}^{-1}$ (here
$.4 < h < 1 $ captures as usual the uncertainty in the Hubble constant)
gives the upper limit
\begin{equation}
\tau \sim .01. \label{eq:HS}
\end{equation}
This can be interpreted as saying that at most about one percent of the
lines of sight from our galaxy terminate on another galaxy
before reaching the recollapsing era. By time symmetry, neither do
most lines of sight connecting the present epoch to its time-reverse.
\subsubsection*{Thomson Scattering}
\label{sec:thomson}
Use of the Thomson scattering cross section
$ \sigma_{{\rm T}} = \frac{8\pi}{3}r_{0}^{2} = 6.65 \cdot 10^{-25}\,
{\rm\,cm}^{-2}$
is acceptable for scattering from non-relativistic electrons for any
photon softer than a hard X-ray ($\hbar \omega \ll mc^{2}$).
Thus, for the frequencies I will consider,
$\Sigma_{{\rm T}} = \sigma_{{\rm T}} n$
will suffice, giving
\begin{eqnarray}
\tau_{{\rm T}} & = &
\frac{2}{3}\, \frac{\delta \sigma_{{\rm T}} n_{0} c}{H_{0}}
\nonumber \\
& = & 4.7 (\delta h) 10^{-4}. \label{eq:T}
\end{eqnarray}
Recalling that $\delta$ is at worst one order of magnitude, it is clear
that Thomson scattering is not signficant for intergalactic
photons \cite{DT}. It is perhaps worth mentioning that quantum and
relativistic effects only tend to decrease the cross section at higher
energies. More significant for the purposes of this investigation is the
observation that, at the considered range of frequencies,
Thomson scattering
does not change a photon's frequency, merely its direction. Thus
Thomson scattering of a homogeneous and isotropic bath of radiation by a
homogeneous and isotropic soup of electrons has {\it no effect} as
regards the predictions of section \ref{sec:egbr}.\footnote{Were it
significant, it would however be a means of hiding the
{\it information} contained in the background.}
\subsubsection*{Inverse Bremsstrahlung}
\label{sec:bremsstrahlung}
Even less significant than Thomson scattering for frequencies of
interest is free-free absorption by the IGM \cite{DT}. From,
{\it e.g.\ }\cite{RL}, the linear absorption coefficient for scattering
from a thermal bath of ionized hydrogen is
\begin{eqnarray}
\Sigma_{{\rm ff}} & = & \frac{2e^{6}}{3m\hbar c} {\left( \frac{2}{3\pi km}
\right) }^{\frac{1}{2}} n^{2}T^{-\frac{1}{2}} \nu^{-3} \overline{g}(b)
(1-e^{-b}) \nonumber \\
&=&3.7\cdot 10^{8} n^{2} T^{-\frac{1}{2}} \nu^{-3}
\overline{g}(b)(1-e^{-b}) {\rm\,cm}^{-1}.
\label{eq:Sff}
\end{eqnarray}
in cgs units.
Here $b \equiv \frac{h\nu}{kT}$, the factor $e^{-b}$ contains the effect
of stimulated emission, and $\overline{g}(b)$ is a ``Gaunt factor"
expressing quantum deviations from classical results. It is a
monotonically decreasing function of $b$ which is of order one in the
optical ({\it cf.\ }\cite{RL} for a general discussion and
some references.) As
\begin{eqnarray}
b &=& \frac{h\nu}{kT} \nonumber \\
&=& \frac{h\nu_{0}}{kT_{0}} \left( \frac{a}{a_{0}}\right) \nonumber
\end{eqnarray}
increases as the universe expands, taking $\overline{g}(b) = g_{0}$, a
constant of order one, will only overestimate the opacity. Similarly,
following \cite{DT} in dropping the stimulated emission term will yield
an upper limit to the free-free opacity. With $\epsilon = 1$,
$\frac{h\nu_{0}}{kT_{0}} = 1$ when $\nu_{0} \sim 10^{14}\, {\rm\,s}^{-1}$,
so stimulated emission will only lead to a noticeable reduction in
$\Sigma_{{\rm ff}}$
well below the optical. (Actually, methods similar to that employed in
section \ref{sec:transparent}\ can be employed to calculate this term,
but as $\tau_{{\rm ff}}$ will turn out to be insignificant even
neglecting it there is no need to go into that here.)
With these approximations,
$$
\Sigma_{{\rm ff}} \approx
(4.6 \cdot 10^{-8}) g_{0}\delta^{2}\epsilon^{-\frac{1}{2}}
h^{4} \nu_{0}^{-3} {\left( \frac{a_{0}}{a} \right) }^{2},
$$
and thus
\begin{eqnarray}
\tau_{{\rm ff}} & = & 2 \frac{\Sigma_{0} c}{H_{0}} \nonumber \\
&=&8.6\cdot 10^{20}h^{3} g_{0}\delta^{2}\epsilon^{-\frac{1}{2}}
\nu_{0}^{-3}. \label{eq:ff}
\end{eqnarray}
Recalling that $\delta = \epsilon = 1 $ seem likely physical values, and
noting that $\delta^{2}\epsilon^{-\frac{1}{2}} \ltwid 10^{2}$ at worst,
taking $h^{3} g_{0}\delta^{2}\epsilon^{-\frac{1}{2}} = 1$ is not
unreasonable for an
order of magnitude estimate. Thus $\nu_{0} \sim 10^{7} {\rm s}^{-1}$
(long radio) is required to get $\tau_{{\rm ff}} \sim 1$. Since
$\tau_{{\rm ff}} \propto \nu_{0}^{-3}$ it drops sharply for photons
with present frequency above that. For instance, at 5000\AA
$$ \tau_{{\rm ff}} = g_{0}\delta^{2}\epsilon^{-\frac{1}{2}} 10^{-24}, $$
and inverse bremsstrahlung is completely negligible.
\subsubsection*{The Far Future}
\label{sec:future}
Finally, I mention that Davies and Twamley \cite{DT} consider
what happens if
baryons decay in a long lived universe. Following the considerations
of \cite{PM}, they conclude that the positronium ``atoms" which will
form far in the future (when the universe is large) remain transparent
to photons with
present frequencies in the optical. This is because the redshifted photons
haven't enough energy to cause transitions between adjacent Ps energy
levels. Similarly, if in the nearer future the electrons and protons in
the IGM recombine to form more neutral hydrogen, this will also be
transparent at the considered frequencies.
\section[The EGBR in Time Symmetric Universe]{Extragalactic
Background Radiation in a \protect\\
Statistically Time Symmetric Universe}
\label{sec:egbr}
\setcounter{equation}{0}
\subsection{Lower Limit to the Excess Optical EGBR}
\label{sec:limit}
The goal of this section is to explain why, in a
statistically time symmetric
universe (such as one with the CPT-related boundary conditions discussed
in section \ref{sec:motivations}), the optical extragalactic background
radiation should be at least twice
that expected in a universe which is not time symmetric, and possibly
considerably more. Thus, assuming consistency with the
RTH ({\it i.e.\ }the
predictive assumption that physics near either boundary condition is
practically insensitive to the presence of the other boundary condition,
{\it cf.\ }section \ref{sec:tsbc}), it is possible to discover
{\it experimentally}
whether our universe is time symmetric. Section \ref{sec:observations}\
compares this prediction with present observations, concluding that the
minimal
prediction is consistent with upper limits on the observed optical EGBR.
However, better observations and modeling may soon challenge even this
minimal prediction.
At optical wavelengths, the isotropic bath of radiation from sources outside
our galaxy is believed to be due almost exclusively to galaxies on our past
light cone \cite[are some good general references]{Peebles,Deep,IAU,EGBR}.
There is no other physically plausible source for this radiation.
In a model universe with time symmetric boundary conditions, however, there
must in addition be a significant quantity of radiation correlated with the
time-reversed galaxies which will exist in the recollapsing era, far to our
future \cite{DT,TSA}. The reason for this is that light from our galaxies
can propagate largely unabsorbed into the recollapsing phase no matter how
close to open the universe is, as shown in \cite{DT} and in section
\ref{sec:opacity}.
This light will eventually arrive on galaxies in the recollapsing phase,
or, depending on its frequency, be absorbed in the time-reversed equivalent
of one of the many high column density clouds (Lyman-limit clouds and damped
Lyman-$\alpha$ systems) present in our
early universe \cite{Peebles,Deep,QSO},
in the intergalactic medium, or failing that, at the time-reversed
equivalent of the surface of last scattering. This will appear to observers
in the recontracting phase as emission by one of those sources sometime in
their galaxy forming era. Since future galaxies, up to high time-reversed
redshift,
occupy only a small part of the sky seen by today's (on average)
isotropically
emitting galaxies, much of the light from the galaxies of the expanding
phase will proceed past the recontracting era's galaxies. Thus most of
this
light will be absorbed in one of the other listed media.
Because of the assumption of global homogeneity and isotropy, the light
from the entire history of galaxies in the
expanding phase will constitute an isotropic bath of radiation to
observers at the time-reverse of the present epoch that
is {\it in addition} to the light from the galaxies to {\it their} past.
By time symmetry, there will be a similar contribution to our EGBR
correlated with galaxies which will
live in the recollapsing phase, over and above that due to galaxies on
our past light cone. To us this radiation will appear to arise in
isotropically distributed sources {\it other} than galaxies. This
picture of a transparent, time symmetric universe is illustrated in
figure 1.
\begin{figure}[ht]
\label{fig:egbr}
\begin{center}
\epsfig{file=egbrfig.eps
\end{center}
\caption{Schematic representation of the
origin of the ``excess'' extragalactic background light correlated
with the thermodynamically reversed galaxies of the recollapsing era.
The model is a Friedmann-Robertson-Walker universe equipped with time
symmetric boundary conditions requiring the universe to be smooth
and in local thermodynamic equilibrium whenever it is small.}
\end{figure}
A lower limit to this excess background can be obtained by considering how
much light galaxies to our past have emitted already ({\it cf.\ }section
\ref{sec:observations}).
According to observers at the time reverse of the present epoch,
this background will (in the absence of interactions) retain its
frequency spectrum and energy density because the size of the universe is
the same. Thus, by time symmetry, at a {\it minimum} the predicted optical
EGBR in a universe with time symmetric boundary conditions is twice that
expected in a universe in which the thermodynamic arrow does not reverse.
If much of the luminous matter in galaxies today will eventually be burned
into radiation by processing in stars or galactic black holes, the total
background radiation correlated with galaxies in the expanding phase could
be several orders of magnitude larger, a precise prediction
requiring a detailed understanding of the future course of galactic
evolution \cite{DT}. Several other processes may also contribute
significant excess backgrounds. These topics are discussed further below.
\subsubsection*{The Prediction}
A number of points in the summary argument above require amplification.
First, however, I summarize the {\it minimal}\footnote{By ``minimal" I mean
the lower limit in each band provided by taking the integrated background of
light from galaxies in the expanding era to be only that which has
{\it already}
been emitted up to the present epoch. In the absence of absorption, by time
symmetry this is the minimum background at the present epoch that must be
correlated with galaxies in the recontracting phase, as in the previous
paragraph.} predictions for the ``excess" extragalactic background
({\it i.e.\ }radiation from non-galactic sources to our past
that is correlated with time-reversed galaxies) in bands
for which the future light cone is transparent:
\begin{itemize}
\item isotropy: the ``future starlight" should appear in the comoving frame
as an approximately isotropic background. This conclusion depends
crucially on the assumed global validity of the cosmological principle.
\item energy density: comparable to the present energy density in starlight
due to the galaxies on our past light cone. This assumes the future light
cone (FLC) is totally transparent.
\item spectrum: similar to the present spectrum of the background starlight
due to galaxies on our past light cone. Again, neglect of further emissions
in the expanding phase makes this, by time symmetry, a lower limit in
each band. This conclusion relies on the assumption of a transparent FLC
in part to the extent that this implies a paucity of standard astrophysical
mechanisms for distorting spectra.
\end{itemize}
Thus, at for instance optical frequencies, time symmetry requires an
isotropic extragalactic background at least twice that due to galaxies on
our past light cone alone.\footnote{To be totally accurate, the quantity
of radiation absorbed or scattered into another band between the present
epoch and its time reverse should be subtracted. However, the upper limit
to the total FLC opacity (due to anticipated processes) computed in section
\ref{sec:opacity}\ was of order $10^{-2}$, mostly due to a liberal
(``black galaxy") assessment of the rate of
interception of photons by galaxies, and I will therefore neglect
such losses. Further, it is worth remembering that processes like
Thomson scattering do not destroy photons or change their frequency,
but only scatter them. Thus mere
scattering processes may introduce isotropically distributed (via the
cosmological principle) fluctuations in the background, but not change its
total energy. Similarly, line or dust absorption usually result in
re-radiation of the absorbed photons, conserving the total energy in an
isotropic background (if the size of the universe doesn't change much before
the photons are re-radiated), if not the number of photons with a given
energy.}
The potentially far greater background predicted (by time symmetry) if
further emissions in the expanding phase are accounted for is a subject
taken up in the sequel.
\subsubsection*{Consistency with the RTH?}
Before proceeding, a comment on the consistency of this picture is
in order.
As the ``excess" radiation is correlated with the detailed histories of
future galaxies, the transparency
of the future light cone does not appear consistent with the predictive
assumption (the RTH) that physics in the expanding era should be
essentially
independent of the specifics of what happens in the recontracting phase.
At a minimum, if the model is to be at all believable it is legitimate to
demand that the required radiation appears to us to arise in sources in a
fashion consistent
with known, or at least plausible, astrophysics. Thus it may be
that given a transparent FLC, the only viable picture of a time symmetric
universe is one in which the radiation correlated with future galaxies {\it
``should" be there anyway, i.e.\ }be predicted also in some reasonable model
of our universe which is {\it not} time symmetric, and consequently not
be ``excess" radiation at all, but merely optical radiation arising in
non-galactic sources during (or before) the galaxy forming era.
On the basis of present knowledge this does not describe our universe.
The presence of the radiation required by time symmetry and the
transparency of the FLC appears to be in significant disagreement
with what is known about our galaxy
forming era, as will become apparent below.
Were it the case for our universe that non-galactic sources provided a
significant component of the optical EGBR, the difficulties with
time symmetric
boundary conditions would from a {\it practical} point of view be less
severe. It is true that the non-galactic sources emitting the
additional isotropic background would have to do so in just such a way
that the radiation contain the correct spatial and spectral correlations
to converge on
future galaxies at the appropriate rate. This implies a distressingly
detailed connection between the expanding and recontracting phases.
However, if the emission rate and spectrum were close to
that expected on the basis of conventional considerations these
correlations (enforced by the time symmetric boundary conditions)
would likely be wholly
unobservable in practice, existing over regions that are not causally
connected until radiation from them converges onto a future
galaxy \cite{DT},
and thus not visible to local coarse grainings (observers) in the expanding
era. In any event, the meaning of the transparency of the FLC is that
starlight is by no means a ``short relaxation time process."
Of course, on the basis of the models discussed in section \ref{sec:tsbc},
perhaps the conclusion to draw from this apparent inconsistency between the
RTH and the transparency of the FLC
should rather be, that physical histories would unfold in a fashion
quite different from that in a universe in which $\rho_{\omega}=1$, namely,
in such a way that such detailed correlations would never be required in
the first place. The very formation of stars might be suppressed
({\it cf.\ }section \ref{sec:difficulties}).
Be that as it may, to the extent that the universe to our
past is well understood, there are {\it no} sources that could plausibly
be responsible for an isotropic optical background comparable to that
produced by galaxies. Such an additional background, required in a
transparent, time symmetric universe, requires significant, observable
deviations from established astrophysics. This is in direct contradiction
with the RTH. Thus the entire structure in which an
additional background is predicted appears to be both
internally inconsistent (in that it is inconsistent with the postulate
which allows predictions to be made in the first place), and,
completely apart from observations of the EGBR
discussed in section \ref{sec:observations},
inconsistent with what is known about our galaxy forming
era. This should be taken as a strong argument that our universe does not
possess time symmetric boundary conditions. Nevertheless, in order to
arrive at this conclusion it is necessary to pursue the consequences of
assuming the consistency of the model. The situation may be stated thus:
{\it either} our universe is not time symmetric, {\it or} there is an
unexpected contribution to the optical EGBR due to non-galactic sources to
our past (and there are indeed detailed correlations between the expanding
and recontracting
eras), {\it or} perhaps our future light cone is not transparent after all.
This latter possibility, perhaps related to the considerable uncertainty
regarding the state of the universe when it is large,
seems the last resort for a consistent time symmetric model of our universe.
\subsection{Amplifications}
\label{sec:amplifications}
It is now appropriate to justify further some of the points made in
arriving at the prediction of an excess contribution to the EGBR.
Claims requiring elaboration include: i) most of the light from the
expanding era's galaxies won't be absorbed by the galaxies of the
recontracting phase, and {\it vice-versa}; ii) it will therefore be
absorbed by something else, and this is inconsistent with the early
universe as presently understood; iii) a detailed understanding of the
old age and death of galaxies, as well as other processes when the
universe is large, may lead to a predicted EGBR in a time symmetric
universe that is orders of magnitude larger than the minimal prediction
outlined above. I deal with these questions in turn.
\subsubsection*{The ``Excess" EGBR is Not Associated with
Galaxies to Our Past}
A photon escaping our galaxy is unlikely to encounter another galaxy
before it reaches the time reverse of the present epoch. In fact, as
shown in section \ref{sec:opacity}, galaxies between the present epoch
and its time reversed
equivalent subtend, at most, roughly a mere $ 2 \times .01 = 2\% $
of the sky (\ref{eq:HS}) (neglecting curvature and clumping.)
In light of the present lack of detailed information about our galaxy
forming era (and via time symmetry its time-reverse), a photon's fate
after that is more difficult to determine. A straightforward extrapolation
of the results of
section \ref{sec:opacity}\ (or {\it cf. e.g.\ }section 13 of Peebles
\cite{Peebles}) shows that the optical
depth for encounters with galaxies of the same size and numbers as today is
$\tau \sim .01 (1+z)^{\frac{3}{2}}$ between a redshift of $z$ and today.
Again this neglects curvature (hence overestimating $\tau$) and clumping
(which now underestimates $\tau$.) Assuming the bright parts of galaxies
form at $z \sim 5$, this gives only $\tau \sim .14$, and the sky isn't
covered with them until $z \sim 20$. This however is roughly at the upper
limit on how old galaxies are thought to be.
On the other hand, examination of quasar spectra
(out to $z \sim 5$) show that most lines of sight pass through many clouds
of high column densities of hydrogen called Lyman-$\alpha$ forest clouds
and,
at higher densities, damped Lyman-$\alpha$ systems.
(Peebles \cite{Peebles}
is a useful entry point on all of these matters, as is \cite{Deep}.
\cite{QSO} are the proceedings of a recent conference concerned with
these Lyman systems.) The highest density clouds
may be young galaxies, but if so galaxies were more diffuse in the past
as the observed rate of interception of an arbitrary line of sight with
these clouds is a factor of a few or more greater than that based
on the assumption that
galaxy sizes are constant. (Obviously, this would not be too surprising.)
For instance, for the densest clouds Peebles \cite[section 23]{Peebles}
relates the approximate formula
$$ \frac{dN}{dz} = 0.3 \, {\Sigma_{20}}^{-0.46} $$
for the observed interception rate per unit redshift of a line of
sight with a cloud of column density greater than or equal to
$\Sigma_{20}$ (in units of $10^{20} {\rm\,cm}^{-2}$),
in a range of redshifts about $z=3$. For Lyman-$\alpha$ forest clouds,
$\Sigma \gtwid 10^{14} {\rm\,cm}^{-2}$,
the interception rate is considerably higher.
(For some models see \cite{Tytler,Sargent}.)
Thus an arbitrary line of sight arriving on our galaxy from a redshift
of five, say, is likely to have passed through at least one cloud of column
density comparable to a galaxy, and certainly many clouds of lower density.
What might this mean for time symmetry?
(For specificity I shall concentrate on photons which are optical today,
say around 5000\AA. This band was chosen because at these wavelengths
we have the luxury of the coincidence of decent observations, relatively
well understood theoretical predictions for the background due to galaxies,
the absence of other plausible sources for significant contributions,
and a respectable understanding of the intergalactic opacity, including
in particular some confidence that the future light cone is transparent.)
Photons at 5000\AA\ today are at the Lyman limit (912\AA) at
$z \approx 4.5$, and so are ionizing before that. At these redshifts the
bounds on the amount of smoothly distributed neutral hydrogen
(determined by independent measures such as the Gunn-Peterson
test \cite{Peebles}) are very low ({\it cf.\ }section \ref{sec:flc}),
presumably because that part of the hydrogen formed at recombination
which had not been swept into forming galaxies was ionized by their
radiation. Before the galactic engines condensed and heated up, however,
this neutral hydrogen would have been very opaque to ionizing radiation.
Similarly, near $z \sim 4$ Lyman-limit clouds with $\Sigma \gtwid
10^{17} {\rm\,cm}^{-2}$ are opaque at these frequencies.
The upshot is that most photons from our galaxies which are optical today
will make it well past the time reverse of the present epoch, likely
ending up in the (time-reversed)
L-$\alpha$ forest or in a young (to time reversed observers!) galaxy by
$\tilde{z} \sim 4$. (Here $\tilde{z}$ is the epoch corresponding to the
time-reverse of redshift $z$.) The very few that survive longer must be
absorbed in the sea of neutral hydrogen between $\tilde{z}=0$ and (their)
recombination epoch, $\tilde{z} \sim 1000$.
Now, the important point is that on average, galaxies radiate
isotropically into the full $4\pi$ of sky available to them. The lesson
of the previous paragraph is that most lines of sight from galaxies in
the expanding phase will not encounter a high column density cloud until a
fairly high time-reversed redshift, $\tilde{z} \sim {\rm few}$,
at which point many lines of sight probably {\it will} intersect
one of these proto-galaxies or their more diffuse halos.
If most photons from our galaxies have not been absorbed
by this point, {\it this is not consistent with time symmetry}: the
rate of emission of (what is today) optical radiation by stars in galaxies
could not be time symmetric if the light of the entire history of galaxies
in the expanding phase ends up only on the galaxies of the recollapsing
phase at high $\tilde{z}$ (due consideration of redshifting effects
is implied, of course.) Put another way, time symmetry requires the
specific energy
density in the backround radiation to be time symmetric. Thus the
emission rate in the expanding era must equal (what we would call) the
absorption rate in the recontracting phase. If stars in galaxies were
exclusively both the sources (in the expanding phase) and sinks (as we
would call them in the recontracting phase) of this radiation, galactic
luminosities in the expanding phase would
have to track
the falling rate of absorption due to photon ``collisions" with galaxies.
This is absurd. At the present epoch, for example, {\it at all frequencies}
galaxies would (by time symmetry) have to be absorbing the diffuse EGBR (a
rate for which the upper limit is determined entirely by geometry
in the ``black galaxy" approximation) at the same rate as their stars were
radiating (a rate that, in a time symmetric universe which resembles our own,
one expects to be mostly determined by conventional physics.)\footnote{This
is illustrated in the appendix with a simplified model.
Related considerations may be used to put detailed constraints on the
self-consistency of time symmetry, but I do not address that any further
beyond the appendix. The essential point has already been made.}
That is, stars would be in radiative
equilibrium with the sky! This may be called the ``no Olber's Paradox"
argument against the notion that a single class of localized objects could
be exclusively responsible for the EGBR in a transparent universe equipped
with time symmetric boundary conditions. (It might be
thought that this problem would be solved if the time symmetric boundary
conditions lead galaxies to radiate preferentially in those directions
in which future galaxies lie. This is not a viable solution, because as
noted above, only a small fraction of the sky is subtended by future
galaxies up to high time reversed redshift. The deviations from
isotropic emission would be dramatic.)
Thus, the option consistent with time symmetry is that most galactic photons
which are optical today will ultimately be absorbed in the many (by time
symmetry) time-reversed Lyman-$\alpha$ forest clouds or Lyman-limit clouds
believed to dwell between galaxies, and not in the stars of the time-reversed
galaxies themselves.\footnote{This may be disappointing. A nice picture of a
time symmetric universe might have photons from our galaxies arriving at
time-reversed galaxies in the recollapsing era, appearing as their emissions.
Even ignoring the highly detailed correlations between the expanding and
recollapsing phases this would imply, the scheme could only work if radiation
could be removed by galaxies in the recollapsing phase at the same rate it is
emitted in the expanding. As noted, for isotropically emitting sources this
is forbidden by time symmetry of the emission rate and geometry.} Fortunately
for the notion of time symmetry this indeed appears to be the case. Careful
studies of the opacity associated with Lyman systems \cite{Zuo,Madau},
indicates, within the bounds of our rather limited knowledge, that the light
cone between $z=4.5$ and $z=0$ is essentially totally opaque to radiation that
is 5000\AA\ at $z=0$, and that this is due largely to Lyman clouds
near $z \sim 4$ and in the middle range of observed column densities,
$\Sigma \sim 10^{16-17} {\rm\,cm}^{-2}$ or so. (To be honest, it must be
admitted that hard data on just such clouds is very limited \cite{MM,Madau}.)
We have now arrived at a terrible conundrum for the notion of time
symmetry. Even if one is willing to accept the amazingly detailed
correlations between the expanding and recontracting eras that reconciling a
transparent future light cone with time symmetry requires, and even if
the ``excess" radiation correlated with the galaxies of the recollapsing
era were to be observed, this picture is
incompatible with what little is known about the physical properties of
the Lyman-$\alpha$ forest. Recalling the minimal prediction above for the
excess background required by time symmetry, the prediction is that the
Lyman-$\alpha$ forest has produced an amount of radiation at least comparable
to that produced by the galaxies to our past. There is no mechanism by which
this is reasonable. {\it There is no energy source to provide this amount
of radiation.} More prosaically, the hydrogen plasma in which the clouds
largely consist is observed (via determination of the line shape,
for example) to be at kinetic temperatures of order $10^{4-5}K$, heated by
quasars and young galaxies \cite{Carswell,Giallongo}. Thermal bremsstrahlung
is notoriously inefficient, and line radiation at these temperatures is
certainly insufficient to compete with nuclear star burning in galaxies!
At for instance 5000\AA\
today, essentially {\it no} radiation is expected from forest clouds at all,
let alone an amount comparable to that generated by galaxies. Remembering
that by redshifts of 4.5 the Lyman forest is essentially totally opaque
shortward of 5000\AA\ (observed) \cite{Madau}, it might have been imagined
that
an early generation of galaxies veiled by the forest heated up the clouds
sufficiently for them to re-radiate the isotropic background radiation
required by time symmetry. While it is true that quasars and such are
likely sources of heat for these clouds \cite[for example]{Giallongo},
aside from the considerable difficulties in getting the re-radiated spectrum
to resemble that of galaxies, the observed temperatures of the clouds are
entirely too low to be compatible with the {\it minimum} amount of energy
emission in the bands required.
(A related restriction arises from present day observations of cosmological
metallicities, which constrain the amount of star burning allowed to our past.
If observed discrete sources came close to accounting for the required
quantity of heavy elements, the contribution of a class of objects veiled
completely by the Lyman forest would be
constrained irrespective of observations of
the EGBR. However, at present direct galaxy counts only provide
about 10\% of
the current upper limits on the extragalactic background light \cite{Tyson}
({\it cf.\ }section \ref{sec:observations}), the
rest conventionally thought to arise in unresolvable galactic sources.
Consequently, correlating formation of the heavy elements with
observed discrete sources does not at present provide a good test of
time symmetry. At any rate, such a test is likely to be a
less definitive constraint on time symmetry because it is possible that a
portion of the radiation lighting the Lyman forest from
high redshift is not due to star burning, but to accretion onto
supermassive black holes at the centers of primordial galaxies. Thus
the best observational test is the most direct one, comparison of the
observed EGBR with the contribution expected from galaxies.)
The possibility that somehow the excess radiation does {\it not} come
from the Lyman-$\alpha$ forest, but somehow shines through from other
isotropically distributed sources even further in the past, is
hardly more appealing. Familiar physics
tells us that the forest is totally opaque to radiation that is 5000\AA\
at $z=0$. The conclusion had better be that the universe is not time
symmetric, rather than that time symmetry engineers a clear path only
for those photons correlated with galaxies in the recollapsing epoch
(and not, say, the light from quasars.)
Moreover, even if that were the case, analagous difficulties apply
to the vast sea of neutral hydrogen that existed after recombination,
totally opaque to ionizing radiation, and again to the highly opaque
plasma which constituted the universe {\it before} recombination.
It is possible to conjure progressively more exotic scenarios which
save time symmetry by placing the onus on very special boundary conditions
which engineer such rescues, but this is not the way to do physics.
The only {\it reasonable} way time symmetry could be rescued would be if
it were discovered that for reasons unanticipated here, the future light
cone were not transparent after all, thus obviating the need for an
excess background radiation with all its attendant difficulties.
Otherwise, it is more reasonable to conclude that a universe with time
symmetric boundary conditions would not resemble the one in which we
actually live.
\subsubsection*{Beyond the Minimal Prediction}
Now that we have seen what kind of trouble time symmetry can get into
with only the {\it minimal} required excess background radiation, it is
time to make the problems worse. The background radiation
correlated with the galaxies of the recollapsing era was bounded from
below, via time symmetry, by including only the radiation that has been
emitted by the galaxies to our past. But as our stars continue to burn,
if the future light cone is indeed transparent it is possible a great
deal more radiation will survive into the recollapsing era \cite{DT,TSA}.
How much more? To get an idea of what's possible it is necessary to
know both what fraction of the baryons left in galaxies will be eventually
be burned into starlight, and when. For a rough upper bound, assume that
{\it all} of the matter in galaxies today, including the apparently
substantial dark halos (determined by dynamical methods to contribute
roughly
$\Omega_{{\rm gal}} \sim .1 $), will eventually be burned into radiation.
To get a rough lower bound, assume that only the observed luminous matter
($\Omega_{{\rm lum}} \sim .004$) will participate significantly, and that
only a characteristic fraction of about 4\% of {\it that} will not end up
in remnants (Jupiters, neutron stars, white dwarfs, brown dwarfs,
black holes, {\it etc.})
To overestimate the energy density of this background at
$\tilde{z} = 0$,
assume that all of this energy is released in a sudden burst
at some redshift
$z_{e} (< 0)$. Then by time symmetry, further star burning will yield a
background of radiation correlated with time-reversed galaxies (expressed
as a fraction of the critical density and scaled to $z=0$)
somewhere in the range
$$ (1+z_{e})^{-1} 10^{-6} \ltwid \Omega_{{\rm burn}}
\ltwid (1+z_{e})^{-1} 10^{-3} \label{Oburn}.$$
(Here I have used the fact that the mass fraction released in nuclear
burning as electromagnetic radiation is .007.)
When $(1+z_{e})^{-1} \sim 1$ the upper limit is two orders of magnitude
more than is in the CMBR today and three orders of magnitude more than
present observational upper limits on a
diffuse optical extragalactic background
({\it cf.\ }section \ref{sec:observations}). The lower bound,
however, is comparable to the amount of radiation that has already been
emitted by galaxies. Thus if the lower bound obtains, the prediction for
the optical EGBR in a time symmetric universe is only of order three times
that due to the galaxies to our past (if the excess background inferred
from continued star burning is not distributed over many decades
in frequency,
and if most of this burning occurs near $z(\tilde{z})=0$.)
As will be seen in section \ref{sec:observations},
this may still be consistent with present
observational upper limits. On the other hand, if something closer to the
upper limit obtains this is a clear death blow to time symmetry.
A more precise prediction is clearly of interest. This would entail
acquiring a detailed understanding of further galactic evolution,
integrating over future emissions with due attention to the epoch
at which radiation of a
given frequency is emitted. (Naturally, this is the same exercise one
performs in estimating the EGBR due to galaxies to our past \cite{Deep}.)
Some idea of the possible blueshift ($(1+z_{e})^{-1}$) involved comes from
estimating how long it will take our galaxies to burn out. This should
not be more than a factor of a few greater than the lifetime of the
longest lived stars, so a reasonable ballpark figure is to assume that
galaxies will live for
only another ten billion years or so. For convenience, assume that
galaxies will become dark by $ t = n t_{0}$ for some $n$, where $t_{0}$
is the
present age of the universe. To overestimate the blueshift at this time,
assume the universe is flat, so that
$$ (1+z_{e})^{-1} = (t/t_{0})^{\frac{2}{3}} = n^{\frac{2}{3}} .$$
For reasonable $n$'s this does not amount to a large (in order of magnitude)
transfer of energy to the radiation from cosmological recontraction.
\subsubsection*{Additional Sources of ``Excess" EGBR and the Far Future}
In a similar fashion to continued burning of our stars, any isotropic
background produced to our future might by time symmetry be expected
to imply
an additional contribution to the EGBR in an appropriately blueshifted
band. For instance, even if continued star burning does not
(by time symmetry)
yield a background in contradiction with observations of the EGBR,
it is possible that accretion onto the supermassive black holes likely
to form at the centers of many galaxies could ultimately yield a
quantity of radiation dramatically in excess of that from star burning
alone.\footnote{I owe this suggestion to R. Antonucci.}
In the absence of detailed information about such possibilities it is
perhaps sufficient to note that ignoring possible additional contributions
leads to a lower limit on the EGBR correlated with sources in the
recontracting era, and I will therefore not consider them.
There is one worrying aspect, however. As discussed in some detail by
Page and McKee \cite{PM} for an approximately $k=0$ universe, and commented
on in a related context in section \ref{sec:collapse}, if baryons
decay then considerable photons
may be produced by for instance the pair annihilation of the resulting
electrons and positrons. Should not this, by time symmetry, yield a further
contribution to the EGBR? The answer may well be yes, but there is a
possible mechanism which avoids this conclusion. Somehow, with CPT symmetric
boundary conditions, the density of baryons must be CPT symmetric. Therefore
either baryons do not decay, or they are re-created\footnote{Note that in the
former case CPT-symmetry requires that the observed
matter-antimatter asymmetry inside the horizon does not persist
at larger scales. In the latter case, {\it if} matter dominates
homogeneously in the expanding era, then antimatter must dominate
homogeneously in the recollapsing phase. Further discussion of
CP-violation in T- and CPT-symmetric universes may be found in \cite{TSA}.}
in precisely correlated collisions. (In the absence of a final boundary
condition, the interaction rate would be too low for (anti-)baryon
recombination to occur naturally.) The latter (boundary condition
enforced) possibility appears extraordinary, but if baryon decay occurs
in a universe with CPT symmetric boundary conditions,
it could be argued that the best electrons and
photons for the job would be just those created during baryon decay in the
expanding phase, thereby removing this photon background. The ``no
Olber's Paradox" argument, that most of an isotropically emitting source's
light must end up in some homogeneous medium, and not, if time symmetry
is to be preserved, equivalent time-reversed point sources, may not apply
here if matter is relatively homogeneously distributed when the universe
is large.
Baryon decay {\it might} smooth out
inhomogeneities somewhat before the resulting electrons and positrons
annihilate. (This requires the kinetic energies of the decay products to be
comparable to the gravitational binding energy of the relevant inhomogeneity.)
Then the picture is no longer necessarily of localized sources emitting
into $4\pi$, but of a more homogeneous photon-producing background
that might cover
enough of the sky to more reasonably secrete the required correlations for
reconstruction of more massive particles in the recontracting phase.
Nevertheless the extreme awkwardness of this scenario is not encouraging.
The former possibility, clearly more palatable, is that baryons do not
decay significantly either because $\Omega$ is not so near one after all
that they have time enough to do so, or because the presence of the final
boundary condition suppresses it. Either way, in this (possibly desperate)
picture there is no additional background due to decaying baryons.
A very similar question relates to the enormous number of particles
produced in the last stages of black hole evaporation. This
time, however, the objection that our black holes cover only a small
portion of the recontracting era's sky, and consequently their isotropic
emissions could not do the job of forming the white holes of the
recontracting
era (black holes to observers there) time symmetrically,
would seem to be forceful. Thus if the universe is indeed very
long-lived, black hole evaporation may well require yet an additional
observable background. This may not be such a serious difficulty if
$\Omega $ is not very close to one, however, as the time scales for the
evaporation of galactic-scale black holes are quite immense. Further
discussion of black holes in time symmetric cosmologies may be had in
\cite{Zeh3,Penrose1,Penrose2}.
One last point regarding the predictions described in this section needs
to be made. Clearly, a loose end which could dramatically change the
conclusions is the condition of matter in the universe
when it is very large.
This is uncertain territory, not the least because that is the era in a
statistically time symmetric universe when the thermodynamic arrow
must begin rolling over. Neglecting this confusing complication
(reasonable for some purposes as many interactions are most significant
when the universe is small), there is not a great deal known about
what the far future should look
like \cite{Rees,Davies,Islam,BT,Dyson,PM}. The study of Page and
McKee \cite{PM} gives the most detailed picture in the case of a
flat universe.
As mentioned at the end of section \ref{sec:flc}, Davies and Twamley
\cite{DT} find from this
work that interactions of optical (at $z=0$) backgrounds with the electrons
produced by baryon decay do not appear to be significant, primarily due to
their diffuseness. On the other hand, if supermassive black holes (or any
large gravitational inhomogeneities) appear,
interactions with them may induce
anisotropies in the future starlight. However, clumping only decreases
the probability a line of sight intersects a matter distribution.
Therefore large overdensities probably never subtend enough
solid angle to interfere with most lines of sight to the recollapsing
era unless gravitational collapse proceeds
to the point where it dramatically alters homogeneity and isotropy on
the largest scales. Because collapse is rather strongly constrained
by time symmetry ({\it cf.\ }section \ref{sec:collapse}) I will not
consider this possibility.
Thus, insofar as the prediction of an ``excess" background radiation
correlated with galaxies in the recollapsing era is concerned,
the state of the universe when it is large would does not obviously
play a substantial role. Nevertheless, given the manifold difficulties
cited, the sentiment expressed
above is that the best hope a time symmetric model has of providing a
realistic description of our universe is that some unforseen mechanism
makes the future light cone opaque after all.
\subsection*{Summary}
\label{sec:opacitysummary}
To summarize, because our future light cone is transparent, neglecting
starburning to our future and considering only the contribution to the
EGBR
from stars in our past provides an estimate of the total EGBR correlated
with galaxies in the recollapsing era that is actually a lower limit on it.
As mechanisms for distorting the spectrum generically become less important
as the universe expands (barring unforseen effects in the far future), it is
reasonable to take models for the present EGBR due to stars in our past as a
minimal estimate of the isotropic background of starlight that will make its
way to the recollapsing era. By time symmetry we can expect that at the same
scale factor in the recollapsing era similar (but time-reversed) conditions
obtain. As argued above, by time symmetry and geometry this
``future starlight" must appear to us as an
additional background emanating from homogeneously distributed sources
to our past {\it other} than galaxies. Therefore, if the universe has
time symmetric
boundary conditions which (more or less) reproduce familiar physics when the
universe is small, and our future light cone is transparent, the optical
extragalactic background radiation should be at least twice that expected to
be due to stars in our past alone, and possesses a similar spectrum. If a
considerable portion of the matter presently in galaxies will be burned into
radiation in our future, by time symmetry the expected background is
potentially much larger, and observations of the EGBR may already be flatly
incompatible with observations. Nevertheless, in the
next section I shall be conservative and stick with the minimal prediction
in order to see how it jibes with observations.
\subsection{Models and Observations of the Optical EGBR}
\label{sec:observations}
At optical wavelengths, it is generally believed that the isotropic
background of radiation from extragalactic sources is due entirely to
the galaxies on our past light cone \cite{Deep,IAU,EGBR}. As shown
in the previous section, if our universe is time symmetric there must in
addition be a significant contribution correlated with the galaxies of the
recollapsing era which arises, not in galaxies, but in some homogeneously
distributed medium, say for instance the Lyman clouds.
The apparent inconsistency of this prediction with what is known about
the forest clouds has been discussed above,
and may be taken as an argument that our universe is not time symmetric.
In this section judgement will be suspended, and the prediction of an
``excess" EGBR at least comparable to, but over and above, that due to
galaxies to our past will be compared with experiment. The conclusion
will be that current data are still consistent with time symmetry {\it if}
our galaxies will not, in the time left before they die, emit a quantity
of radiation that is considerably greater than that which they already have.
A useful resource on both the topics of this section is \cite{Deep}.
Tyson \cite{Tyson} has computed how much of the optical extragalactic
background is accounted for by resolvable galaxies, concluding that
known discrete sources contribute
\begin{equation}
\label{eq:tyson}
\nu i_{\nu} \sim 3 \cdot 10^{-6} {\rm\,erg}\, {\rm\,cm}^{-2}\, {\rm\,s}^{-1}\,
{\rm\,ster}^{-1}
\end{equation}
at 4500\AA. However, because very distant galaxies contribute most of
the background radiation it is believed that unresolvable sources provide
a significant portion of the EGBR. At present it is not possible to
directly identify this radiation as galactic in origin. However, as
understanding of galactic evolution grows so does the ability to model
the optical extragalactic background due to galaxies. These predictions
naturally depend on the adopted evolutionary models, what classes of
objects are considered, the cosmological model, and so on. As
representative samples I quote the results of Code and Welch \cite{CW}
for a flat universe in which all galaxies evolve,
\begin{equation}
\label{eq:CW}
\nu i_{\nu}\sim 8\cdot 10^{-6} {\rm\,erg}\, {\rm\,cm}^{-2}\, {\rm\,s}^{-1}\,
{\rm\,ster}^{-1}
\end{equation}
at 5000\AA, and of Cole {\it et al.\ }\cite{CTS} for a similar scenario,
\begin{equation}
\label{eq:Cole}
\nu i_{\nu}\sim 3\cdot 10^{-6} {\rm\,erg}\, {\rm\,cm}^{-2}\, {\rm\,s}^{-1}\,
{\rm\,ster}^{-1},
\end{equation}
also at 5000\AA. These figures are to be compared with the results of
(extraordinarily difficult) observations. As surveyed by
Mattila \cite{Mattila}, they give at 5000\AA\ an upper limit of
\begin{equation}
\label{eq:mattila}
\nu i_{\nu}\ltwid 2\cdot 10^{-5}{\rm\,erg}\, {\rm\,cm}^{-2}\, {\rm\,s}^{-1}\,
{\rm\,ster}^{-1}.
\end{equation}
As far as I am aware, there has been no direct detection of an optical
radiation background of extragalactic origin. This upper limit represents
what is left after what can be accounted for in local sources is removed.
Comparing these results, it is clear that if current models of galactic
evolution are reliable, present observations of the extragalactic background
radiation leave room for a contribution from non-galactic sources that
is comparable to the galactic contribution, but not a great deal more.
These observations therefore constrain the possibility that our
universe is time symmetric. If believable models indicate that further
galactic emissions compete with what has been emitted so far, time
symmetry could already be incompatible with experiment. A direct detection
of the extragalactic background radiation, or even just a better upper limit,
could rule out time symmetry on {\it experimental} grounds
soon.\footnote{To be more careful, if direct HST galaxy counts,
or reliable models of the (extra-)galactic
contribution closely agree with the observations, the justifiable
conclusion is that {\it if} our universe is time symmetric, then for some
unknown reason our future light cone is not transparent.} (The ideal
observational situation would be convergence of all-sky photometry and
direct HST galaxy counts, allowing one to dispense with models
completely.)
\section{Further Difficulties with Time Symmetry}
\label{sec:difficulties}
\setcounter{equation}{0}
In this section I comment on some issues of a more theoretical nature which
must be faced in any attempt to construct a believable model of a time
symmetric universe. Among other questions, in such a universe the
difficulties with incorporating realistic gravitational collapse, and in
deriving the fact that radiation is retarded, appear to be considerable.
\subsection{Gravitational Collapse}
\label{sec:collapse}
A careful account of the growth of gravitational inhomogeneities from the
very smooth conditions when the universe was small is clearly of fundamental
importance in any model of the universe, time symmetric or not, not the least
because it appears to be the essential origin of the thermodynamic arrow of
time\footnote{For examples of concrete calculations connecting the growth
of gravitational inhomogeneities with the emergence of a thermodynamic arrow
see \cite{HaH,Hawking1,HLL}.}
from which the other apparent time arrows are thought to flow \cite{PTA,Zeh2}.
However, if matter in a closed universe is to be smooth at
both epochs of small
scale factor then it is incumbent to demonstrate that Einstein's equations
admit solutions in which an initially smooth universe can develop interesting
(non-linear) inhomogeneities such as galaxies which eventually
smooth out again
as the universe recollapses. This is because the universe is certainly in
a quasiclassical domain now, and if it is assumed to remain so
whenever the scale factor is large classical general relativity must
apply. Laflamme \cite{Laflamme} has shown that in the linear regime there
are
essentially no perturbations which are regular, small, and growing away
from both ends of a closed FRW universe, so that in order to be small at
both
ends a linear perturbation must be too small to ever become non-linear.
But that is not really the point. Interesting perturbations must go
non-linear, and there is still no proof of which I am aware that
perturbations which go non-linear cannot be matched in the non-linear
regime so as to allow solutions which are small near both singularities.
Put differently, what is required is something like a proof that
Weyl curvature
must increase ({\it cf.\ }\cite{Penrose1,Penrose2}), {\it i.e.\ }that given
some suitable energy conditions the evolution of gravitational inhomogeneity
must be monotonic even in the absence of trapped surfaces,
except possibly on a set of initial data of measure zero.
While perhaps plausible given the attractive nature of gravity, proof has
not been forthcoming.
(It is important for present purposes that genericity conditions
on the initial data are not a central part of such a proof, for physically
realistic solutions which meet the time symmetric boundary conditions
must describe processes in the recollapsing era which from our point of
view would look like galaxies disassembling themselves. Reducing such a
solution to data on one spacelike hypersurface at, say, maximum
expansion, said data will be highly specialized relative to solutions
with galaxies which do not behave so unfamiliarly. If such solutions
with physically interesting inhomogeneities do exist, the real question
here is whether they are generic {\it according to the measure defined
by the two-time boundary conditions.} Since it ought to be possible to
treat this problem classically, in principle this measure is straightforward
to construct. The practical difficulty arises in evaluating the generic
behaviour of solutions once they go non-linear. My own view is that it is
highly likely that such solutions remain exceedingly improbable according to
a measure defined by a {\it generic} set of (statistical) boundary conditions
requiring the universe to be smooth when small. As noted, Laflamme
\cite{Laflamme} has already shown that when the initial and final
states are required to be smooth,
the growth of inhomogeneity is suppressed if only linear perturbations
are considered. Unless boundary conditions with very special correlations
built in are imposed, probable solutions should resemble their smooth initial
and final states throughout the course of their evolution, never developing
physically interesting inhomogeneities.\footnote{One possible out
is Schulman's observation that systems which exhibit chaos, as
general relativity does, may be less restricted in the varieties of
their behaviour by boundary conditions at two times than are systems
with linear dynamics \cite{Schulman3}. This substantially unstudied
possibility would obviously never emerge from Laflamme's linear
analysis.}
Note the concern here is not with occurrences which would be deemed
unlikely in a universe with an initial condition only, but occur in a
time symmetric universe because of the ``fate" represented by the final
boundary condition, but with occurrences which are unlikely {\it even in a
universe with (generic) time symmetric boundary conditions.})
A related question concerns collapsed objects in a time
symmetric universe. Page and McKee \cite{PM}
have studied the far future of a $k=0$ FRW universe under the
assumption that baryons decay but electrons are stable. Assuming
insensitivity to a final boundary condition their conclusions should
have some relevance to the period before maximum expansion in the only
slightly over-closed (and hence very long-lived) model universes that
have mostly been considered here. As discussed in section
\ref{sec:amplifications}, if the universe is very long lived it might be
imagined that the decay of baryons and subsequent annihilation of the
produced electrons and positrons could smooth out inhomogeneities, and
also tend to destroy detailed information about the gravitational history of
the expanding phase (by eliminating compact objects such as neutron stars,
for instance.) Thus, even though interactions are unlikely to thermalize
matter and radiation when the universe is very large \cite{PM}
({\it cf.\ }section \ref{sec:opacity}), there may be an analagous
information loss via
the quantum decay of baryons which could serve a similar function.
(For a completely different idea about why quantum mechanics may effectively
decouple the expanding and recollapsing eras, see \cite{Zeh1,Zeh2,Zeh3,KZ}.)
In any case, if there is no mechanism to eliminate collapsed objects
before the time of maximum expansion, then the collapsed objects of the
expanding phase are the same as the collapsed objects of the
recontracting phase, implying detailed correlations between the
expanding and recontracting era histories of these objects which might
lead to difficulties of consistency with the RTH.
A particular complication is that it is fairly certain that black
holes exist, and that more will form as inhomogeneity grows. The only
way to eliminate a black hole is to allow it to evaporate, yet the
estimates of Page and Mckee indicate that it is more likely for black
holes to coalesce into ever bigger holes unless for some reason (a final
boundary condition?) there is a maximum mass to the black holes which
form, in which case they may have time enough to evaporate (though this
requires $\Omega$ to be {\it exceedingly} close to one.)
In fact, it may be imperative for a time symmetric scenario that black holes
evaporate, else somehow they would have to turn into the white holes of the
recollapsing era (black holes to observers there.) This is because we do
not observe white holes today \cite{Penrose1,Penrose2}. (A related
observation is that in order for the universe to be smooth whenever it
is small, black/white hole singularities cannot arise \cite{Zeh3}.)
Here the evaporation of black holes before
maximum expansion would be enforced by the time symmetric boundary
conditions selecting out only those histories for which there are no
white holes in the expanding era and {\it mutatis mutandis} for the
recollapsing era.\footnote{For more on black holes in time symmetric
universes, see the discussion of Zeh \cite{Zeh3}.}
On the other hand, if evaporating black holes leave remnants they too
must be worked into the picture. Again, if the results of
the stochastic models with two-time low entropy boundary conditions
discussed above are to be taken seriously, the conclusion should probably
be that boundary conditions requiring homogeneity when the universe is
small suppress histories in which significant gravitational collapse
occurs by assigning low probabilities to histories with fluctuations
that will go non-linear. It hardly needs emphasizing that all of these
considerations are tentative, and greater clarity would be welcome.
\subsection{The Retardation of Radiation}
\label{sec:retardation}
Besides gravitational considerations, radiation which connects the
expanding and recollapsing eras provides another example of a physical
process which samples conditions near both boundary conditions \cite{DT}.
While gravitational radiation and neutrinos are highly penetrating and are
likely to provide such a bridge, in neither case are we yet capable of both
effectively observing, and accurately predicting, what is expected
from sources to our past. Therefore the
primary focus of this investigation has been electromagnetic radiation.
Above it was confirmed that modulo the obviously substantial uncertainty
regarding the condition of the universe when it is large, even
electromagnetic radiation is likely to penetrate to the recollapsing
era. Section \ref{sec:egbr}\ was concerned with one relatively prosaic
consequence of this prediction if our universe possesses time symmetric
boundary conditions. Here I comment briefly on another.
Maxwell's equations, the dynamical laws governing electromagnetic
radiation, are time symmetric. It is generally believed that the
manifest asymmetry in time of radiation {\it phenomena}, that is, that
(in the absence of source-free fields) observations are described by
retarded solutions rather than advanced, is ascribable fundamentally
to the thermodynamic arrow of time without additional hypotheses.
(For a contemporary review see \cite{Zeh2}.)
However, if our universe possesses time symmetric boundary conditions
then near the big bang the thermodynamic arrow of entropy increase runs
oppositely to that near the big crunch. Since radiation can connect the
expanding and recollapsing eras, the past light cone of an accelerating
charge in the expanding era ends up in matter for which the entropy is
increasing, while its future light cone terminates in matter for which
entropy is supposed to be {\it decreasing}. If the charge radiates into
its future light cone this implies detailed correlations in this matter
with the motion of the charge which are incompatible with the supposed
entropy decrease there (entropy increase to time reversed observers),
although it is true that these correlations are causally disconnected to
time-reversed observers, and consequently invisible to {\it local}
coarse grainings defining a local notion of entropy for such observers.
This state of affairs makes it difficult to decide whether radiation
from an accelerating charge (if its radiation can escape into intergalactic
space) should be retarded
(from the perspective of observers in the expanding era)
or advanced or some mixture of the two. The conditions under which the
radiation arrow is usually derived from the
thermodynamic arrow of surrounding
matter do not hold. (Notice how this situation is reminiscent of the
requirements necessary to derive retardation of radiation in the
Wheeler-Feynman ``absorber theory" of electrodynamic phenomena \cite{WF}.)
Hence the ability of radiation to connect the expanding and recollapsing
epochs brings into question the self-consistency of assuming
time symmetric boundary conditions on our universe together with
``physics as usual" (here meaning radiation which would be described as
retarded by observers in both the expanding and recollapsing eras) near
either end. The retardation of radiation is another important example
of a physical prediction which would be expected to be
very different in a universe with time symmetric boundary conditions
than in one without. Once again, if the results of the simple stochastic
models are generally applicable, the retardation of radiation should
no longer be a prediction in such a universe.\footnote{Analytical
elucidation
of this idea is hoped to be the subject of a (far) future paper.}
To summarize this section, consideration of gravitational collapse and
radiation phenomena reveals that construction of a model universe with
time symmetric boundary conditions which resembles our own may be a
difficult task indeed. There are strong suggestions that a
model with time symmetric boundary conditions which mimic our own
early universe would behave nothing like the universe in which we live.
Such a model would most likely predict a universe which remained smooth
throughout the course of its evolution, with coupled matter components
consequently remaining in the quasi-static ``equilibrium" appropriate to a
dynamic universe.
\section{Summation}
\label{sec:summation}
\setcounter{equation}{0}
In spite of the oft-expressed intuitive misgivings regarding the
possibility that our universe might be time symmetric \cite[for example]
{Weinberg,PTA,Penrose1}, it has generally been felt that if sufficiently
long-lived, there might be no way to tell the difference between a time
symmetric and time asymmetric universe. Building on suggestions of
Cocke, Schulman, Davies, and Gell-Mann and Hartle (among others), this
work has explored in some detail one physical process which, happily,
belies this feeling: no matter how long our universe will live, the
time symmetry of the universe implies that the extragalactic background
radiation be {\it at least} twice that due to the galaxies to our past.
This is essentially due to the fact that light can propagate unabsorbed
from the present epoch all the way to the recollapsing era.
Moreover, geometry and time symmetry requires this ``excess" EGBR to be
associated with sources {\it other} than the stars in those galaxies,
sources which, according to present knowledge about the era during which
galaxies formed, are not capable of producing this radiation! Thus the
time symmetry of a closed universe is a property which is {\it directly
accessible to experiment} (present observations are nearly capable of
performing this test), as well as extremely difficult to model convincingly
on the basis of known astrophysics. In addition, the other theoretical
obstacles remarked upon briefly in sections \ref{sec:amplifications}\
and \ref{sec:difficulties}\ make it difficult to see how a
plausible time symmetric model for the observed universe
might be constructed.
In particular, any such attempt must demonstrate that in a universe that is
smooth whenever it is small, gravitational collapse can procede to an
interesting degree of inhomogeneity when the universe is larger
\cite{Laflamme}.
This is necessary in order that the universe display a thermodynamic arrow
(consistently defined across spacelike slices)
which naturally reverses itself as the universe begins to recollapse.
Furthermore, it appears unlikely that the usual derivation of the
retardation of radiation will follow through in a time symmetric
universe in which radiation can connect regions displaying opposed
thermodynamic arrows. Finally, in the context of the time neutral
generalized quantum mechanics employed as the framework for this
discussion, unless the locally observed matter-antimatter asymmetry
extends globally across the present universe,\footnote{Recall that
this requires that the universe live long enough for nearly all baryons
to decay, and reform into the antibaryons of the recollapsing era.
This presents serious additional difficulties,
{\it cf.\ }section \ref{sec:amplifications}.}
natural choices of CPT-related boundary conditions yield a theory with
trivial dynamics if the deviations from exact homogeneity and isotropy
are specified in a CPT invariant fashion (see the last paragraph of section
\ref{sec:motivations}).
In sum, were the ``excess" EGBR which has been the primary concern of this
investigation to be observed, it would appear necessary to place the onus of
explanation of the fact that the final boundary condition is otherwise
practically invisible upon very specially chosen boundary conditions which
encode the details of physics in our universe. This would make it difficult
to understand these boundary conditions in a natural way.
On the dual grounds
of theory and experiment, it therefore appears unlikely that we live in a
time symmetric universe. (A definitive expurgation must await more thorough
investigation of at least some of the aforementioned difficulties.)
|
\section{--~ Introduction}
The CHARM2000 workshop\cite{CHARM2000} suggested that the goal
for a future experiment be a factor
$\sim10^2$ increase in statistics over the coming round of fixed-target
charm experiments at Fermilab (E781 and E831).
We consider the physics goals of such an ultrahigh-statistics
charm experiment ($ \sim $ 100 million reconstuctable $ D $'s ).
Some measurements will
test the Standard Model, some will measure its parameters,
and some will elucidate heavy-quark
phenomenology. We can outline the major goals as follows:
\begin{enumerate}
\item
Measurements which search for new physics\cite{Pakvasa}
\begin{itemize}
\parskip=0pt
\item $ D^0 - \overline D {}^0$ mixing
\item explicit flavor-changing neutral currents
\item direct {\it CP} violation.
\end{itemize}
\item
Measurements which test the heavy-quark symmetry of QCD in the charm sector
\begin{itemize}
\parskip=0pt
\item form factors of semileptonic decays of charmed mesons
\item masses and widths of orbitally-excited charmed mesons.
\end{itemize}
These measurements are key to extracting
fundamental parameters from future beauty experiments.
\item
Measurements which probe aspects of perturbative and
nonperturbative QCD (including higher-twist effects)
\begin{itemize}
\parskip=0pt
\item nonleptonic singly and doubly Cabibbo-suppressed decay rates
\item dynamics of charm hadroproduction.
\end{itemize}
\end{enumerate}
\noindent
We next discuss these topics in more detail.
\section{--~ Searches for New Physics}
\subsection{\underline {$ D^0 - \overline D {}^0 $ mixing}}
$ D^0 - \overline D {}^0 $ mixing is one of the most interesting places to look
for physics beyond the Standard Model.
While Wolfenstein\cite{Wolfenstein}
once suggested large long-distance or dispersive
contributions to $ \Delta m_D $ within the Standard Model, more detailed
calculations by Donoghue {\it et
al.}\cite{Donoghue} give $ | \Delta m_D | \approx 10^{-6} $\,eV.
Recent analyses\cite{Georgi,Ohl} based on heavy-quark effective theory (HQET)
suggest
that
cancellations lead to $ | \Delta m_D | < 3.5 \times 10^{-8} $\,eV.
Values of $ | \Delta m_D | $ as large as $ 10^{-4} $\,eV are
possible in many models beyond the Standard
Model\cite{EHLQ}~$^-$\cite{HW}.
Thus there is a large window for observing new physics via $ D^0 -
\overline D{}^0 $ mixing.
A particularly intriguing example is discussed in a recent paper by Hall and
Weinberg\cite{HW}, which emphasizes that electroweak theories with several
scalar doublets are consistent with all known physics (especially with the
approximate {\it CP} symmetry in the neutral-kaon sector), provide an
alternative mechanism for {\it CP} violation,
and have various interesting phenomenological features.
In such models, tree-level scalar-exchange contributions to
neutral $ K $- and $ B $-meson mass mixing are at about the level
observed by experiments if Higgs bosons have masses in the 700\,GeV range.
Hall and Weinberg say that ``although this means that little can be
learned about the CKM matrix from $ \Delta m_K $ and $ \Delta m_B $, the
case of $ D - \overline D $ [mixing] presents different opportunities.
$ \ldots $
If we take the typical Higgs boson mass as near 1\,TeV to account for the
observed values of $ | \Delta m_K | $ and $ | \Delta m_B | $, then the
predicted value of $ | \Delta m_D | $ is close to the current experimental
limit, $ | \Delta m_D | < 1.3 \times 10^{-4} $\,eV."
CLEO II has reported a $ D^0 \rightarrow K^+ \pi^- $ signal with a
branching ratio about $ 2 \times \tan^4 \theta_C \times B ( D^0
\rightarrow K^- \pi^+) $\cite{Cinabro}.
This ``wrong-sign" kaon is a signature either of mixing or of doubly
Cabibbo-suppressed decay (DCSD). The two are
most easily separated in a fixed-target experiment, in which the lifetime
of the $ D $ can be directly measured: a DCSD signal decays exponentially,
while a mixing
signal has an additional $ t^2 $ dependence, so that it peaks at $2\tau_D $.
If the CLEO II signal is primarily a DCSD signal, then it presents an
inescapable background for mixing studies using hadronic final states.
Assuming this is the case, an experiment such as we consider should be
sensitive to a mixing signal on top of a DCSD
signal with $ | \Delta m_D | \approx2 \times 10^{-5} $\,eV.
Morrison\cite{Morrison} has pointed out that similar sensitivity might be
achievable also in semileptonic decays, which are free of the confounding
effects of DCSD. Liu's thorough treatment\cite{Liu} includes the intriguing
suggestion that for mixing arising from the decay-rate difference between the
{\em CP} eigenstates $D_1$ and $D_2$, sensitivity an order of magnitude better
might be achievable in singly Cabibbo-suppressed modes, by using the
interference between mixing and DCSD to enhance the mixing signal.
\subsection{\underline {Charm-changing neutral currents}}
Some of
the models cited above\cite{Buchmuller,Joshipura} also allow the possibility of
explicit flavor-changing neutral currents (FCNC) in $ D $ decays.
E791 has reported\cite{Nguyen}
the best 90\%-confidence-level
upper limit for the branching ratio of $ D^+ \rightarrow \pi^+
\mu^+
\mu^- $ of $ 1.3 \times 10^{-5}$.
An ultrahigh-statistics experiment with good lepton identification
would have a sensitivity to this and other FCNC decays
(and also to lepton-number-violating decays)
one to two orders of magnitude lower.
\subsection{\underline {$C\!P$ violation}}
The Standard Model predicts that direct {\it CP} violation (observed as the
fractional difference between decay rates of particle and antiparticle to
charge-conjugate final states) will be of order $10^{-3}$ or less
in singly Cabibbo-suppressed $ D $ decays\cite{Bucella}.
(In the Standard Model, {\it CP} should be an exact symmetry for
Cabibbo-allowed and DCSD decays.)
Physics beyond the Standard Model might contribute {\it CP}-violating
amplitudes to decay rates, and there is a large window for observing new
physics. At the level of
statistics we consider here,
sensitivity to {\it CP} asymmetries at the fraction-of-a-percent level in
singly Cabibbo-suppressed decays and at the few-percent level in doubly
Cabibbo-suppressed decays may be possible.
Holding systematic uncertainties to the percent level will be challenging,
and experimenters planning to make such measurements must consider
carefully how systematic errors will be minimized and
how they will be measured.
\section{--~ HQET and Semileptonic Decays}
\subsection{\underline {Testing HQET via orbitally-excited charmed mesons}}
Within the Standard Model, it is generally agreed that heavy-quark symmetry
can be used to predict many nonperturbative properties of hadrons containing a
single heavy quark, and the most important of these predictions are for
exclusive semileptonic $ B $-meson decays\cite{Wisgur1}.
These nonperturbative effects will be
important in extracting $ V_{ub} $ and $ V_{cb} $ from measured decay rates.
Heavy-quark symmetry also relates the masses and widths of the
orbitally-excited
$ D^{**} $ mesons (including the $ D_s^{**} $ mesons), as has been discussed
recently in papers by Isgur and Wise\cite{Wisgur2},
Ming-Lu, Isgur, and Wise\cite{Wisgur3}, and
Eichten, Hill, and Quigg\cite{EHQ}.
While some authors argue that the charm-quark
mass is sufficiently large for the limit $ m_c \rightarrow \infty $ to be a
good approximation\cite{EHQ},
others\cite{Godfrey} have argued that even for $ B $ mesons
the $ m_Q \rightarrow \infty $ limit has not been reached.
The experiment we consider will
measure the masses and widths of the orbitally-excited $ D^{**} $
mesons with sufficient precision to confront theoretical models
quantitatively.
Where E691\cite{E691}, ARGUS\cite{ARGUS}, CLEO II\cite{CLEO}, and
E687\cite{E687} have measured the $
D_2^{*0} $, $ D_2^{*+} $, and $ D_1^0 $ widths with 50\% fractional errors,
such an experiment
should be able to achieve few-percent fractional errors.
To untangle the states and reflections which lie on top of each other,
it will also be necessary to measure $ \pi^0 $'s and (perhaps) single
photons well.
However, the benefit of making these measurements is that they
will establish how well heavy-quark symmetry works for charm and
give theorists the numbers they need to develop a more complete
phenomenology of $ B $ physics.
\subsection{\underline {Semileptonic form factors}}
High-statistics charm experiments will also contribute to our understanding of
the
form factors and helicity amplitudes of the vector mesons which can appear
as decay products in both $ D $ and $ B $ decays.
Extracting {\it CP}-violation parameters from measurements of
branching ratios for decays such as
$ B_d \rightarrow \rho^0 \psi $ and
$ B_d \rightarrow K^* \psi $, which Dunietz\cite{Dunietz}
advocates as the best place to
measure the unitarity-triangle angle $ \gamma $,
requires the best possible measurement of the
$ \rho^0$ and $ K^* $ helicity amplitudes and form factors in the
$ D $ semileptonic decays
$ D^+ \rightarrow \rho^0 l \nu $ and $ D^+ \rightarrow K^{*0} l \nu $,
as they should be the same in $ D $ as in $B $ decay.
Assuming single-pole forms for the form factors, the mass of the pole
should be measurable with better than 1\% precision.
In $ D^+
\rightarrow K^{*0} l \nu $,
it should be possible to
measure the polarization of the $ K^* $,
\begin{equation}
{\Gamma_L /
\Gamma_T} =
{{ \int P^*_V t | H_0(t) | ^2 dt}
\over
{\int P^*_Vt \left [ | H_+(t) |^2 + |H_- (t) |^2 \right ] dt} }
\end{equation}
(the ratio of
longitudinal to transverse form-factors),
with percent statistical and systematic uncertainties.
It should be possible to measure the polarization of
the $ \rho^0 $ in the Cabibbo-suppressed decay with few-percent statistical
accuracy.
$ D_S \rightarrow \phi l \nu $ should be measured with similar precision,
which will provide another test of the applicability of heavy-quark symmetry to
the study of semileptonic decays.
\subsection{\underline {Studying the CKM matrix with semileptonic decays}}
Studying semileptonic decays also contributes directly to our knowledge of
the CKM matrix. High-statistics charm experiments are able to measure the
magnitudes of $ V_{cs} $ and $ V_{cd} $ from the semileptonic decays of the $ D
$ mesons. The absolute decay rates depend on various well-measured constants
(such as the $ D $ masses and lifetimes), the CKM matrix elements, and the
form factors of the hadrons produced along with the leptons.
Currently, $| V_{cd}| $ and $ | V_{cs} |$ are known with $ \pm 8 \% $ and $ \pm
20 \% $ precision\cite{PDG,Rosner}.
{}From the branching ratios for the semileptonic decays
$ D^0 \rightarrow \pi^- l^+ \nu_{l} $
and $ D^0 \rightarrow K^- l^+ \nu_{l} $,
the ratio $ |
V_{cd} | / | V_{cs} | $ should be determined
with a statistical accuracy of $\sim10^{-3}$.
In addition to testing the unitarity of the CKM matrix in the charm sector,
this ratio is
explicitly required to extract the unitarity angle $ \gamma $ from the ratio
$B( B_d \rightarrow \rho^0 \psi )/B( B_d \rightarrow K^* \psi )$
discussed earlier.
\section{--~ Testing QCD with Charm Hadroproduction}
At the parton level, $ c \bar c $ production is supposed to be described by
perturbative QCD\cite{NDE}.
At the hadron level, the situation becomes more complicated.
Several experiments have reported large leading-particle effects at high
$x_F$\cite{E769}.
Leading-twist factorization in perturbative QCD predicts that
the charm quark's fragmentation is independent of the structure of the
projectile, while the data indicate that the produced charm quark coalesces or
recombines with the projectile spectator.
To test models of higher-twist effects\cite{Brodsky},
one wants to look at the observed production
asymmetries as functions of $ p_T $ and $ x_F $ jointly.
Measuring these asymmetries for different target nuclei
({\it i.e.} measuring the $ A $-dependence of these asymmetries)
will provide an extra handle on how quarks evolve into hadrons.
\section{--~ Experimental Issues}
Building an ultrahigh-statistics charm experiment will be a challenge.
The next-generation fixed-target experiments at Fermilab each project
reconstructed charm samples of order $ 10^6 $ events.
A Tau-Charm Factory operating at a luminosity of
$ 10^{33} \,$cm${}^{-2}$\,sec${}^{-1}$,
such as that proposed for SLAC\cite{slac343},
would reconstruct about $ 5 \times 10^6 $ charm per year.
The $ B $ factories planned for KEK and SLAC will produce of order $ 10^8 $
$ b \overline b $ events per year at design luminosity.
However, the number of reconstructed charm will be
similar to that projected for the Tau-Charm Factory.
HERA-$B$\cite{albrecht94} will produce a sufficient number of
$ D $'s in $ p p $ collisions to imagine an ultrahigh-statistics experiment,
but the triggering requirements for charm physics
differ substantially from those for $ B $ physics,
and the data acquisition system is currently designed to operate at 10 Hz.
In addition, the current design for the HERA-$B$ vertex detector entails much
more multiple scattering and much poorer vertex resolution
than are desirable for a charm experiment.
There is no clear route to higher luminosities for $ e^+ e^- $ machines or
photon beams, so we are left with the problem of working with a relatively
small cross-section in a hadronic environment.
Whether it is a fixed-target experiment or a collider experiment that
we consider,
triggering selectively and efficiently will be the first major problem.
Building a detector which minimizes backgrounds will be another problem.
If we are looking for physics beyond the Standard Model, or looking for
relatively rare decays expected within the Standard Model, reducing backgrounds
will be at least as important as maintaining high efficiency for the
interesting signals. Two examples should suffice:
\begin{description}
\item [1)~ ]
The FCNC decay $ D^+ \rightarrow \pi^+ \mu^+ \mu^- $ is expected to have
a branching ratio less than $ 10^{-8} $ in the Standard Model\cite{schwartz93}
(except for the decay $ D^+ \rightarrow \phi \pi^+ $ followed by
$\phi \rightarrow \mu^+ \mu^- $, which populates a limited region of the
Dalitz plot).
E791\cite{Nguyen} finds that its sensitivity is greatest when the expected
number of background events is between 5 and 10 in the signal region.
If one were to scale up from this experiment simply, sensitivity would
increase only as the square root of the number
of reconstructed charm, since the background would grow linearly
with the signal. To increase sensitivity here, it will be important to reduce
backgrounds without
substantially reducing efficiency for detecting muons.
This can be achieved by adding redundancy, {\it e.g.} a second
view in the muon detector or a redundant muon-momentum measurement,
so that the double muon-misidentification probability becomes
approximately the square of the single muon-misidentification probability.
\item [2)~ ]
To measure the ratio of CKM matrix elements by comparing the decay
rates for $ D^+ \rightarrow K^* \ell \nu $ and $ D^+ \rightarrow \rho^0 \ell
\nu $, it will be critical to separate pions from kaons with a very high
degree of confidence;
the reflections of these signals feed into each other directly.
A fast RICH technology may suffice, but this is another area where
redundancy seems necessary to reduce the confusions which lead to
systematic errors.
\end{description}
Finally, it seems obvious that silicon pixel devices will
be necessary to provide both the
spatial resolution and the segmentation that are required for unambiguous
vertexing in the high-rate small-angle region.
\section{--~ Summary}
Charm physics provides a window into the Standard Model, and possibly beyond,
that complements those provided by other types of experiments.
In searches for $D-{\overline D}$ mixing,
FCNC, lepton-number-violating decays, or {\it CP}-violating
amplitudes, we are probing physics at the TeV level which may not
be accessible to other experiments: down-sector and up-sector
quarks can couple differently to new physics, and the charm quark is the only
up-sector quark for which such studies are possible.
Within the Standard Model, charm is probably the best place to test heavy-quark
symmetry quantitatively, and it is the best place to measure some of the
CKM matrix elements.
While ultrahigh-statistics experiments will be extremely difficult,
we can reasonably imagine that the technology will exist in the
next decade to reconstruct 100 million charm.
Getting from here to there will require a substantial R\&D effort,
and developing the expertise to design and build such an experiment
will require commitment from the individuals who will contribute directly,
from the laboratory at which it will be done, and from
the community as a whole.
We thank the organizers of this Workshop for offering us
this opportunity to discuss heavy-quark physics in such attractive
surroundings.
We also thank our colleagues from Fermilab proposal
P829\cite{p829}
for their contributions.
This work was supported by NSF grant PHY 92-04239 and
DOE grant DE-FG02-94ER40840.
|
\section{Introduction}
Perturbation theory has been an important calculational method in
quantum field theory for several decades. However the practical usefulness of
perturbation theory is limited mainly by the rapid rise in the amount of
labour required to improve the order of approximation. For instance, the
renormalization group functions can be calculated analytically to three
loops in non-abelian gauge theory, which requires the evaluation of 440 three
loop diagrams, and to five loops in $\lambda \phi^{4}$ theory~\cite{tar80,che81}.
The non-numerical nature of such perturbative calculations in quantum field
theory has complicated the effort to automate these calculations.
However, the development of specialised computer algebra systems~\cite{str74}
and improvements in general purpose computer algebra systems have greatly
facilitated this task~\cite{kub90,hy92,ber92,ft92,jam93,west93}.
In devising a scheme for the perturbative calculation of amplitudes in quantum
field theory the main areas which need to be developed are:
\begin{itemize}
\item The perturbation expansion itself,
\item Lorentz tensor, Dirac and symmetry group algebra, and
\item Feynman integration.
\end{itemize}
The perturbation expansion is either a Dyson-Wick expansion in the canonical
formalism or equivalently an expansion of the generating functional in
the path integral formalism. The perturbation expansion will not concern us
here, nor will we be concerned with automating symmetry group algebra which
is simple for all cases of interest to us. Our focus will be on Lorentz
tensor and Dirac algebra, and on Feynman integration.
There are a number of existing packages that tackle
the problem of Dirac algebra computations~\cite{kub90,hy92,ber92,jam93}.
The symbolic evaluation of Feynman diagrams at
tree level~\cite{ber92} and one loop~\cite{west93} have been
approached using Wolfram's computer algebra package Mathematica~\cite{wolf88}.
Fleischer and Tarasov~\cite{ft92} present a package for the evaluation of
certain two loop Feynman integrals written in the computer algebra language
FORM~\cite{ver90},
while symbolic three loop calculations~\cite{lar93}
have also been carried using FORM.
The current paper presents a package that combines a method for simplifying
the Dirac algebra with procedures for evaluating massless scalar or tensor
Feynman integrals at one or two loops,
and is implemented in Mathematica.
The following sections discuss our scheme for representing Feynman diagrams.
The scheme is essentially a definition of a notation that can be readily
expressed in Mathematica.
Section two introduces the notational scheme,
section three will discuss some aspects
of rules which have been encoded to simplify and integrate expressions,
and section four demonstrates the evaluation of some specific diagrams.
Usage messages for functions defined by us can be found in the appendix.
\section{Notational Scheme}
In this section it will be shown how one can use the symbolic
algebra capabilities of Mathematica to represent and simplify the
various types of tensorial and matrix expressions that arise in
amplitudes of gauge theories. This is based upon a suitable notation for
dimensionally continued tensors and Dirac gamma matrices. The aim is to
establish a notation that can be readily accommodated by the symbolic
capabilities of Mathematica. Dirac algebra can be performed without
reference to any representation of the Dirac matrices, by using the
commutation relations which define the algebra and identities derived
from those relations, along with the rules of matrix multiplication.
Tensor manipulations are also readily implemented symbolically. As we are
interested in evaluating amputated one particle irreducible amplitudes
which have been regularised using dimensional regularisation~\cite{tv72},
no explicit representation of Dirac gamma matrices is necessary.
The first step is to separate symbols which will represent
four-vectors from symbols which will represent tensor indices. We
generally use one or more lower case letters and numerals which are not
otherwise defined, such as {\tt a,k,p1 \ldots } for four-vectors and indices.
The two types of symbols can be distinguished by declaring a list of symbols
which will represent four-vectors. Any symbols which
appear that are not in this list and are not otherwise defined shall
be understood to represent a Lorentz index, and need not be declared.
These symbols will appear only as the arguments of Mathematica
expressions. The expressions
themselves will represent Dirac matrices, Lorentz scalars,
four-vectors and tensors depending on the symbols that appear in their
arguments. We give the list of declared four vectors the name {\tt momenta}.
To illustrate this scheme assume {\tt momenta = \{k,p,q\} } throughout
the rest of this section. This list may be enlarged
at any time provided symbols that have already been used as Lorentz
indices are not included. Let the function {\tt g,} a
symmetric function of two arguments, denote the object that will
carry the properties of the metric tensor. What {\tt g} actually represents
will depend on its arguments as follows:
\begin{enumerate}
\item {\tt g[a,b]} represents the metric tensor with Lorentz
indices {\tt a} and {\tt b,} because {\tt a} and {\tt b} are not in the list
{\tt \{k,p,q\},}
\item {\tt g[a,k]} represents the four-vector {\tt k} with Lorentz
index {\tt a,} because only {\tt k} is in the list {\tt \{k,p,q\},} and
\item {\tt g[k,p]} represents the scalar product of the
four-vectors {\tt k} and {\tt p} which are both in {\tt \{k,p,q\}.}
\end{enumerate}
Products of gamma matrices are represented by a function
{\tt d} of any non-zero number of arguments. Again, what {\tt d} represents
depends
on the arguments. Wherever there appears an argument that is a member of
{\tt momenta} the corresponding matrix is the contraction of the gamma
matrix in that position with the four-vector. Otherwise the matrix is simply a
gamma matrix with the symbol denoting a Lorentz index. For example,
the object {\tt d[a,b,k,p,b,d,k,q]} denotes
\begin{equation}
\gamma^{\alpha} \gamma^{\beta} \! \not \! k \not \! p \:
\gamma_{\beta} \: \gamma^{\delta} \! \not \! k \not \! q .
\end{equation}
Free Lorentz indices appearing as arguments of {\tt g} and {\tt d}
can represent
contravariant or covariant indices. In renormalizing quantum field theories
we consider only amputated amplitudes. Information about components of
four-vectors and matrices is not required.
Thus the distinction between contravariant and
covariant components is not important here. One simply takes
the rank structure of the final expression to be that of the
input expression. If one wishes to consider expressions which included
external lines then one would need to distinguish between contravariant
and covariant indices. This could be done by declaring all covariant
indices in a list which one may call {\tt covariant.} Then symbols which
appear in neither of {\tt momenta} or {\tt covariant} and were not
otherwise defined would implicitly represent contravariant indices.
Traces of products of gamma matrices can be denoted by a function {\tt trace}
whose argument is a linear combination of products of gamma matrices.
This function will automatically convert the trace of a linear combination
to the linear combination of traces before any traces are evaluated.
Other tensors may be represented by other functions with the rank
structure given implicitly by the arguments. For example one may wish
to manipulate the Levi-Civita tensor in four dimensions. One would
define a function {\tt e} depending on four arguments.
Then, for example {\tt e[a,b,u,v]} would represent
\begin{equation}
\epsilon^{\alpha \beta \mu \nu}
\end{equation}
and {\tt e[a,b,k,p]} would represent
\begin{equation}
\epsilon^{\alpha \beta \mu \nu} k_{\mu} p_{\nu}.
\end{equation}
The mass of a particle is represented by the function {\tt M[x]}
where {\tt x} is a label for the convenience of the user
and is not used by the program.
Similarly, other scalars are represented by the function {\tt scalar[x].}
The left and right helicity projection operators are represented
by {\tt L} and {\tt R,} respectively. The matrix
$\gamma_{5}$ has not been explicitly represented,
nor have traces involving $\gamma_{5}$ been implemented. However the
definitions could be easily extended if traces involving $\gamma_{5}$ were
required (see for example~\cite{west93}).
Manipulation and simplification of expressions can be performed by
machine using pattern recognition, procedural programming, rule based
programming and functional programming all of which are supported by
Mathematica.
One can perform basic manipulations such as
\begin{itemize}
\item contracting repeated indices in tensor expressions,
\item simplifying an expression of the form
$ \gamma^{\mu} \gamma^{\alpha_{1}} \ldots \gamma^{\alpha_{j}}
\gamma_{\mu} $
\item commuting a gamma matrix through one or more gamma matrices,
\item the evaluation of the trace of any number of gamma matrices, and
\item evaluation of Feynman integrals
\end{itemize}
and any other lengthy algebraic manipulation that would be
prohibitively tedious to perform manually. Some of the rule definitions
which are useful for the manipulations described are
illustrated in the next section. Note that the system function {\tt Dot}
has been used in place of {\tt d} for products of gamma matrices. This
function will perform matrix multiplication when explicit matrices in
component form are placed in its arguments. Since we require no
representation for the gamma matrices and thus only symbols appear as
its arguments, the function {\tt Dot} will simply represent non-commutative
multiplication and one can assign rules required to invoke the
various properties of Dirac algebra which are required. We have chosen to
use {\tt Dot} for this purpose because a convenient input notation
\begin{equation}
{\tt Dot[a1, a2, \ldots] \; \equiv \; a1.a2. \, \ldots}
\end{equation}
is available. Since {\tt Dot} has the attribute {\tt OneIdentity} the
expression
{\tt Dot[a]} is equivalent to {\tt a} where {\tt Head[a] = Symbol.} Hence a
solitary
symbol which is otherwise undefined represents a single gamma matrix.
Together with symbols {\tt i} for the imaginary unit, {\tt n} for the
number of space-time dimensions, and {\tt eps = 4-n,} this scheme is
sufficient to represent any amputated diagram in a gauge theory. In
the next section we will discuss how to manipulate, simplify, and
integrate expressions within this scheme.
\section{Implementation of Symbolic Algebra}
We will begin by discussing expressions representing Lorentz tensors.
This will be followed with Dirac algebra and expressions involving
both gamma matrices and tensors, and traces of gamma matrices.
Finally the evaluation of two-loop Feynman integrals will be discussed.
In evaluating Lagrangian counterterms, which in turn give renormalization
group functions, we are interested in the pole part of amputated
diagrams. This necessitates the evaluation of Laurent expansions
and some points relating to this will be discussed briefly.
The metric of $n$-dimensional Minkowski space is represented by the
symmetric function {\tt g} with two arguments,
\begin{eqnarray}
\lefteqn{{\tt Attributes[g] \,=\, \{Orderless\} }} & & \\
\lefteqn{{\tt g[x\_,x\_] \, := \, n}} & {\tt /; \,FreeQ[momenta,x]} & \\
\lefteqn{{\tt g[a\_,b\_]\hat{\:}2 \, := \, n}} & & \nonumber \\
& \hspace{1.4cm}{\tt /; FreeQ[momenta,a] \: \&\& \: FreeQ[momenta,b]} & \\
\lefteqn{{\tt g[a\_,b\_]\hat{\:}2 \, := \, g[b,b]}} & & \nonumber \\
& \hspace{1.4cm}{\tt /; FreeQ[momenta,a] \: \&\& \: MemberQ[momenta,b] } &
\hspace{2.0cm}
\end{eqnarray}
Note that the action of the rules is conditional upon whether or not one or
both arguments are in the list {\tt momenta.} In conventional notation the
rules stated are
\begin{eqnarray}
&& g^{\alpha}_{\alpha} = n \\
&& g^{\alpha \beta} g_{\alpha \beta} = n \\
&& b^{\alpha}b_{\alpha} = b^{2}
\end{eqnarray}
where b is a vector. Rules can be applied at the discretion of the
user by defining new functions which act on tensors. For example
the relations
\begin{eqnarray}
&& g^{\alpha \beta} g_{\beta \lambda} = \delta^{\alpha}_{\lambda} \label{eq:rmind} \\
&& b^{\alpha} g_{\alpha \lambda} = b_{\lambda} \label{eq:lower}
\end{eqnarray}
can be implemented by
\begin{eqnarray}
\lefteqn{ {\tt cc[x\_,a\_,b\_,s\_] \: := \: x \, /. \,g[a,s]\, g[s,b]\:
\rightarrow \: g[a,b]} } \nonumber \\
&& \hspace{6.0cm} /; \, {\tt FreeQ[momenta,s]} \hspace{3.0cm}
\end{eqnarray}
Note that in this notation both (\ref{eq:rmind}) and (\ref{eq:lower}) are
embodied in one rule.
Functions can appear in the definitions of functions. For instance, a
function which contracts all occurrences of a repeated index is
valuable. Suppose all the arguments of {\tt g} in an expression {\tt x}
are placed in a list called {\tt args} except for one repeated index {\tt s.}
We have defined a function, {\tt contract[x, args, s],} which
will perform all contractions of the repeated index {\tt s} in {\tt x.}
The definition of contract involves two other functions, {\tt cc} which
controls the actual replacement that is made, and {\tt ll} which takes a list
of symbols and returns a list of all pairs of those symbols. Extending
this further, we have defined a function,
{\tt contractall[x, repeated, other],} which takes an expression {\tt x,} a
list of all repeated indices placed in the second argument, and a list of
all remaining arguments of {\tt g} in {\tt x} in the third argument and
performs the contraction of all repeated indices. The
definition of this function is in terms of {\tt contract.}
As for {\tt g}, rules are assigned to {\tt Dot} to automatically implement
properties of matrix algebra. Note that since {\tt Dot} is a system function
it must be unprotected before rules can be added to its definition.
The rules we define for {\tt Dot} are applied automatically to any input
expression until no further changes occur.
These rules correspond to properties such as the distributive law over
addition and scalar multiplication. We have chosen to use the symbol
{\tt J} to denote the identity matrix.
Expressions which represent a combination of tensors and gamma matrices will
involve both {\tt g} and {\tt Dot.} In cases where a Lorentz index appears in
an argument of both {\tt g} and {\tt Dot} the index may be contracted out of
the
expression. A function {\tt slash} is defined in the appendix which performs
this task.
Identities, such as commutation relations can be applied at the
discretion of the user by defining functions which take an expression,
search for a specified pattern and replace the pattern by an equivalent
expression. For example, the commutation relations
\begin{eqnarray}
&& \gamma_{\alpha} \gamma_{\beta} = - \gamma_{\beta} \gamma_{\alpha} + 2 g_{\alpha \beta} \\
&& \gamma_{\alpha} \gamma_{\beta} \gamma_{\lambda} = \gamma_{\beta} \gamma_{\lambda} \gamma_{\alpha} +
2 \gamma_{\lambda} g_{\alpha \beta} - 2 \gamma_{\beta} g_{\alpha \lambda}
\end{eqnarray}
are applied by the function {\tt comm}, defined as
\begin{eqnarray}
\lefteqn{ {\tt comm[x\_,a\_,b\_] \: := \: x \, /. \,a.b \rightarrow
b.a \: + \: 2J \, g[a,b]}} & & \\
\lefteqn{{\tt comm[x\_,a\_,b\_,c\_] \: := \: x \, /. \, a.b.c \rightarrow
b.c.a \: + \: 2c\, g[a,b] \: - \: 2b\, g[a,c]}} & & \hspace{12.0cm}
\end{eqnarray}
Which of the two rules is applied depends on how many arguments are
passed to {\tt comm}. Note that the rules are valid whether or not
any of {\tt a, b} and {\tt c} appear in the list {\tt momenta.} Further rules
have been defined for commuting through more than two gamma matrices, and
another function called {\tt rcomm} for commuting in the reverse order.
Rules which apply identities for expressions of the form
\begin{eqnarray}
&&\gamma_{\mu} \gamma_{\alpha_{1}} \ldots \gamma_{\alpha_{j}} \gamma^{\mu} \\
&& k\!\!\!/ \gamma_{\alpha_{1}} \ldots \gamma_{\alpha_{j}} k\!\!\!/ \; , \;j=1,2,3,\ldots
\end{eqnarray}
have been similarly defined. For instance
\begin{eqnarray}
{\tt con[x\_,a\_] \: := } & {\tt If \left[ \:MemberQ[momenta,a] \, ,
\right. }\hspace{5.5cm} & \nonumber \\
& {\tt x\, /. \, a.a \rightarrow g[a,a]\, J,}\hspace{2.0cm} & \nonumber \\
& {\tt x\, /. \, a.a \rightarrow n\, J }\hspace{3.0cm} & \nonumber \\
& \lefteqn{\left. \right]} \hspace{8.5cm} &
\end{eqnarray}
\begin{eqnarray}
{\tt con[x\_,a\_,b\_]\: := } & {\tt If \left[ MemberQ[momenta,a] \, ,
\right. } \hspace{5.5cm} & \nonumber \\
& {\tt x\, /. \, a.b.a \rightarrow - b\, g[a,a]\: +\: 2a\,
g[a,b],} & \nonumber \\
& {\tt x\, /. \, a.b.a \rightarrow (2-n)\, b}\hspace{2.5cm} & \nonumber \\
& \lefteqn{\left. \right]} \hspace{8.5cm} &
\end{eqnarray}
The answer is dependent on whether or not the repeated symbol {\tt a}
appears in the list {\tt momenta.} Rules for $j>1$ can be included as needed.
To simplify expressions involving helicity projection operators,
the functions {\tt movel} and {\tt mover} have been defined.
They, respectively, move the helicity projection operators to the left or
to the right in each term in an expression.
The trace of an arbitrary linear combination of products of gamma matrices
can be evaluated using the linearity property of traces and the
recursive relation
\begin{eqnarray}
Tr(\gamma^{\alpha_{1}} \gamma^{\alpha_{2}}) & = &
4 g^{\alpha_{1} \alpha_{2}}\\
Tr(\gamma^{\alpha_{1}} \ldots \gamma^{\alpha_{2j}}) & = & \sum_{l=o}^{2j-2}
(-1)^{l} g^{\alpha_{1} \alpha_{2j-l}} Tr(\gamma^{\alpha_{2}} \ldots
\hat{\gamma}^{\alpha_{2j-l}} \ldots \gamma^{\alpha_{2j}}) \label{eq:tr} \\
Tr(\gamma^{\alpha_{1}} \ldots \gamma^{\alpha_{2j+1}}) & = & 0
\end{eqnarray}
where the hat over the gamma matrix denotes its absence~\cite{west93}.
The function
{\tt trace} will firstly convert the trace of a linear combination into a
linear combination of traces, and then applies a function {\tt tr}
which makes use of the recursion relation (\ref{eq:tr}) to evaluate
the traces.
Note that {\tt tr} first checks that the length of the product is even and
returns zero if it is not. The case of a product of two gamma matrices
acts as an initial condition.
The scheme described so far can be used to represent any Feynman integral
which can arise in a gauge theory. A means of evaluating the integrals
would complete a scheme for the evaluation of amputated Feynman diagrams.
We will present a simple procedure which can be adapted to all cases.
We begin by considering integrals of the form
\begin{equation}
\mu^{\lambda} \int \frac{d^{n}q}{(2 \pi)^{n}}
\frac{q_{\alpha} q_{\beta} \ldots}{(q^{2})^{j} ((q \pm p)^{2})^{l}}
\label{eq:stdint1}
\end{equation}
and
\begin{equation}
\mu^{\lambda} \int \frac{d^{n}q}{(2 \pi)^{n}}
\frac{q \!\!\!/ q_{\alpha} q_{\beta} \ldots}{(q^{2})^{j} ((q \pm p)^{2})^{l}}
\label{eq:stdint2}
\end{equation}
where $\mu$ is an arbitrary mass scale, $q$ is the integration variable and
the rank of the tensor in the numerator is usually less than or
equal to 3. The factor $\mu^{\lambda}$ is
introduced to ensure that the integral is dimensionless for any $n$ (hence
$\lambda$ depends on $n$). This is sufficient for most
two loop calculations but if rules for the integration of higher rank
expressions are needed, then extra rules can be included.
The main consideration in defining an integration function is to be certain
that there is no residual dependence on the integration variable in the final
expression. For example, a function could be defined which searches for the
pattern
\begin{equation}
{\tt \frac{1}{g[q,q] \: g[q+p,q+p]}}
\end{equation}
and replaces it using a replacement rule. If that function were then
applied to a pattern like
\begin{equation}
{\tt \frac{T[q]}{g[q,q] \:g[q+p,q+p]}}
\end{equation}
the integration function would treat this as a product of two factors
\begin{equation}
{\tt T[q]\; \frac{1}{g[q,q] \:g[q+p,q+p]}}
\end{equation}
and replace only the second factor, leaving {\tt T[q]} in the resulting
expression. Hence it is important to check the final expression for the
presence of the integration variable.
In the integrals (\ref{eq:stdint1}) and (\ref{eq:stdint2}) there are other
considerations. The rank, the symbols used for indices, the integration
variable, the parameter {\tt p,} and the powers {\tt j} and {\tt l} should
all be variables which can be passed to the integrating function. We have
defined functions {\tt inttensor} and {\tt intgamma} which will return the
integrals of (\ref{eq:stdint1}) and (\ref{eq:stdint2}) respectively.
The arguments to be passed to {\tt inttensor} and {\tt intgamma} are
\begin{equation}
{\tt inttensor[x, q, p, a, b, \ldots, j, l] }
\end{equation}
and
\begin{equation}
{\tt intgamma[x, q, p, a, b, \ldots, j, l], }
\end{equation}
where {\tt x} is the expression to be integrated, {\tt q} is the integration
variable, {\tt p} is the four-momentum parameter, {\tt a,b,\ldots} are
Lorentz indices, and {\tt j} and {\tt l} are the powers in the denominator.
Again the definitions of {\tt inttensor} and {\tt intgamma} consist of a
limited series of rules. Which of the rules is applied depends on the number of
arguments passed to {\tt inttensor} and {\tt intgamma.} If the rules do not
cover cases of high rank that may be needed then new rules may be included
as needed. Rank three integrals are usually sufficient for most two-loop
calculations. Integrals involving non-zero masses have not been implemented,
though a function for these could be defined.
Dimensionally regularised integrals can be expanded as a Laurent series in
$\epsilon \; = \; 4-n$. In general an integral over $l$ four vectors will have a
leading pole of order $\epsilon^{-l}$. To evaluate counterterms only pole parts of
the Laurent expansions of integrals need be evaluated.
A table of pole parts of integrals can be prepared with the help of
{\tt inttensor} and {\tt intgamma.} This table can then be used to evaluate
the pole parts of diagrams and hence the counterterms. The integrals
required for this can be classified according to the form of the
denominators of the Feynman integrals. For example, to
calculate the two-loop boson self energy in a $SU(m)$ gauge
theory with massless fermions there is just one class of integrals
as the Feynman integral is always of the form
\begin{equation}
\mu^{\lambda}
\int_{q,k} \frac{T_{\mu \nu}(q,k,p)}{q^{2}(q+p)^{2}k^{2}(k+p)^{2}(k-q)^{2}}
\label{eq;twopt}
\end{equation}
where $p$ is the external momentum, and $T_{\mu \nu}(q,k,p)$ is a rank-2,
dimension 4 tensor constructed from $q$, $k$, and $p$. There are 75
distinct possibilities for $T_{\mu \nu}$ but fortunately many of the
corresponding integrals are related, and the number of independent integrals
is reduced to about 10. If we are interested only in the pole parts then
we can express any integral of the form (\ref{eq;twopt}) in terms of the pole
parts of just four integrals. We choose the four integrals where
$T_{\mu \nu} (q,k,p)$ is either
\begin{equation}
t_{\mu \nu}(p) (q+p)^{2}
\end{equation}
or
\begin{equation}
t_{\mu \nu}(p) (k-q)^{2}
\end{equation}
and $t_{\mu \nu}(p)$ is either $p_{\mu} p_{\nu}$ or $g_{\mu \nu} p^{2}$.
In the integrals discussed we have represented the Euler gamma function
by {\tt gam[x].} For the evaluation of pole parts of one-particle irreducible
amplitudes one must be able to perform the Laurent expansions. This can be
done by either replacing {\tt gam} with the system defined version of the
Euler gamma function and using such functions as {\tt Series,} or by defining
a sufficient set of rules for {\tt gam} to evaluate the expansion.
\section{Examples}
As an illustration of the application of this work to some specific examples,
we will briefly discuss the evaluation of the integral
\begin{equation}
\mu^{2(4-n)} \int \frac{d^{n}q}{(2 \pi)^{n}} \frac{d^{n}k}{(2 \pi)^{n}}
\frac{1}{k^{2} (k+p)^{2} q^{2} (q-k)^{2}} \label{eq;intex}
\end{equation}
the evaluation of the diagram shown in figure~\ref{fig;bosonse},
and the simplification of the diagram in figure~\ref{fig;quarkem}.
Diagram~\ref{fig;bosonse} is a contribution to the gauge boson self energy
in the presence of massless fermions in some representation R of the gauge
group. The diagram~\ref{fig;quarkem} represents a contribution to the
quark-photon vertex in the Weinberg-Salam model.
Before evaluating the integral (\ref{eq;intex}) we must first declare the list
of four-vector symbols as
\begin{equation}
{\tt momenta = \{k, p, q, k+p, q+p, k-q, q-k\} \label{eq;mlist}}
\end{equation}
Then we specify the integrand by
\begin{equation}
{\tt x = \frac{1}{g[k, k]\, g[k+p, k+p]\, g[q, q]\, g[q-k, q-k]}}
\end{equation}
The integral over {\tt q} is performed using inttensor,
\begin{equation}
{\tt x= inttensor[x, q, k, 1, 1] }
\end{equation}
which results in an expression depending on the integration variable {\tt k.}
The integrand of this expression is proportional to
\begin{equation}
{\tt \frac{1}{(g[k,k])^{1+ \frac{eps}{2}} \, g[k+p, k+p]}}
\end{equation}
and can also integrated by using {\tt inttensor} as follows.
\begin{equation}
{\tt x= inttensor[x, k, p, 1+eps/2, 1]}
\end{equation}
The resulting expression is
\begin{equation}
{\tt \frac{-(2^{2 eps} \, mu^{2 eps}\, pi^{-4\,+\,eps}\, gam[1-eps]\, gam[1-
\frac{eps}{2}]\, gam[\frac{eps}{2}]\, gam[eps])}
{256\, (-1)^{eps}\, g[p,p]^{eps}\, gam[2 -\frac{3 eps}{2}]\, gam[2-eps]\,
gam[1+\frac{eps}{2}]}}
\end{equation}
where {\tt mu} is the introduced mass scale $\mu$,
and this can be Laurent expanded to order $\epsilon^{-1}$. We obtain
\begin{equation}
{\tt \frac{-2 \: + \: 2 \, elog \: - \: 5 \, eps}{256\, eps^{2}\, pi^{4}}
+ O( eps)^{0}}
\end{equation}
where
\begin{equation}
{\tt elog = egam - log \left( \frac{4 \, pi \, mu^{2}}{-g[p,p]} \right)}
\end{equation}
and {\tt egam = 0.577\ldots} is Euler's constant. This gives the result for
two of the four integrals required to determine the pole part
of (\ref{eq;twopt}) and the remaining integrals can be evaluated
similarly. Then to integrate the expression (\ref{eq;twopt}) we first
replace the integrand with an expression that has the same
pole part and then replace the integrals with the Laurent expansions.
The functions we have defined to do this are {\tt samepole} and {\tt poleform}
respectively.
To evaluate the diagram of figure~\ref{fig;bosonse} we declare the
momenta list to be~(\ref{eq;mlist}), and define some initial expressions.
In the notation of our scheme they are the coefficient of the integral,
the numerator and denominator of the integrand. The coefficient is
\begin{equation}
{\tt coefficient = \frac{i}{2} \, g^{4} \,C[G] \, T[R] \, delta[a, b]}
\end{equation}
where {\tt a} and {\tt b} are gauge group indices, {\tt delta} is the delta
symbol and {\tt C[G]} and
{\tt T[R]} are gauge group factors. The numerator of the integrand has
two factors
\begin{eqnarray}
{\tt y = (g[p,a]\, -\, g[k,a]) \, g[u,l] \: +\: (2 g[k,u]\, +\,
g[p,u])\, g[l,a] -} & & \nonumber \\
{\tt (2 g[p,l] \, + \, g[k,l])\, g[a,u]}\hspace{4.0cm} & &
\end{eqnarray}
and
\begin{equation}
{\tt z = trace[a.(p+q).v.q.l.(k-q)]}
\end{equation}
and the denominator is
\begin{equation}
{\tt denominator = g[q,q]\, g[q+p,q+p]\, g[k,k]\, g[k+p,k+p]\, g[k-q,k-q]}
\end{equation}
In the numerator there are two repeated indices. We set
${\tt repeated = \{a,l\},}$ ${\tt other=\{u,v,k,p,q\} }$ and
${\tt x = Expand[y\: z]}$ and
remove the repeated indices from {\tt x} using {\tt contractall,}
\begin{equation}
{\tt x=contractall[x,repeated,other] }
\end{equation}
All scalar products are then eliminated using {\tt preps} and {\tt prepd}
which make use of identities such as
\begin{equation}
k.p = \frac{1}{2} ((k+p)^{2} -k^{2} -p^{2}).
\end{equation}
The resulting expression for {\tt x} consists of 69 terms. We then replace
{\tt x} with an expression with the same pole part using {\tt samepole.}
The result is divided by {\tt denominator} and the Laurent expansion is
evaluated by applying {\tt poleform.} Simplifying the result and multiplying by
{\tt coefficient} gives the final result for the diagram
\begin{eqnarray}
& & {\tt -( g^{4}\, i\, C[G]\, T[R]\, TR[J]\, delta[a,b] } \nonumber \\
& & \hspace{0.4cm} {\tt ((-48\: -\: 76 eps\:+\:48 \, elog \: eps)\, g[p,u] \,
g[p,v] \:+ } \nonumber \\
& & \hspace{0.4cm} {\tt (12\: -59\, eps\: -\: 12 \, elog\: eps)\, g[p,p]\,
g[u,v]))/(55296 \, eps^2 \, pi^4) }
\end{eqnarray}
where {\tt TR[J]} is the trace of the spin identity matrix.
We have used our scheme to evaluate the gauge boson, fermion and
ghost anomalous dimensions to two loops. We have also calculated the coupling
constant renormalization by evaluating the gauge boson - ghost vertex,
and evaluated the Callan-Symanzik beta function. Our results are in
complete agreement with~\cite{ego79} for these calculations.
Now consider the diagram in figure~\ref{fig;quarkem},
which is a two loop diagram contributing to the quark-photon vertex.
The one loop, flavour-changing quark self energy, with momentum $p$
can be written as
\begin{equation}
A \, p\!\!\!/ L + B\, p\!\!\!/ R + C\, L + D \, R
\end{equation}
where $A, B, C$ and $D$ are scalars depending upon $p^{2}$ and
the quark masses. After defining
\begin{equation}
{\tt momenta = \{k,p\} }
\end{equation}
the numerator of the diagram in figure~\ref{fig;quarkem} is proportional,
in our notation, to
\begin{eqnarray}
{\tt x} & = & {\tt alpha \, .\, L\, .\, (p+k+M[s])\, .\,
(scalar[A]\, (p\, +\, k)\, .\, L + scalar[B]\, (p\, +\, k)\, .\, R } \nonumber \\
& & \hspace{0.4cm} {\tt + scalar[C]\, L + scalar[D]\, R)\, .\, (p+k+M[d])\,
.\, mu \, .\, (p-k+M[d])\, .\,} \nonumber \\
& & \hspace{0.8cm}{\tt alpha. L }
\end{eqnarray}
The masses are implicitly multiplied by the identity matrix.
To simplify this we first apply
\begin{equation}
{\tt x = movel[x]}
\end{equation}
Then repeated application of the function {\tt con} will do the contractions
over $\gamma_{\alpha}$ and contract the pairs of $k\!\!\!/$'s and $p\!\!\!/$'s.
For example, the following does all possible contractions over
$k\!\!\!/$ and $p\!\!\!/$.
\begin{eqnarray}
& & {\tt x\, =\, con[x,p]\, ;\, x\, =\, con[x, k]\, ;\, x\, =\, con[x, p,mu]\,
; } \nonumber \\
& & \hspace{0.4cm} {\tt \, x\, =\, con[x,k,mu]\, ;\, x\, =\, con[x,p,k]\, ;\,
x\, =\, con[x,k,p]\, ; } \nonumber \\
& & \hspace{0.8cm} {\tt x\, =\, con[x,p,mu]\, ;\, x\, =\, con[x,k,mu] }
\end{eqnarray}
A further seven applications of {\tt con} are necessary to do all the
contractions over $\gamma_{\alpha}$.
In the resulting expression, repeated use of {\tt comm}
\begin{eqnarray}
& & {\tt x\, =\, comm[x,p,k]\, ; \, x\, =\, comm[x,p,mu]\, ;\, x\, =\,
comm[x,k,mu]\, ; } \nonumber \\
& & \hspace{0.4cm} {\tt x\; =\; comm[x,p,k] }
\end{eqnarray}
followed by collecting with respect to {\tt Dot,}
produces a final expression of four terms, proportional to
$R \gamma_{\mu} k\!\!\!/ p\!\!\!/$, $R \gamma_{\mu}$, $R k\!\!\!/$ and $R p\!\!\!/$.
The full expression has seventy-six terms when expanded and is too long to
reproduce here.
\section{Conclusion}
We have presented a scheme for representing amputated Feynman diagrams
beyond tree level in terms of Mathematica expressions and have shown how
the algebra involved in simplifying these expressions and evaluating the
integrals can be automated using the symbolic algebra capacity of
Mathematica. This scheme is based upon the specification of a
sufficiently convenient notation for Lorentz tensors and Dirac algebra.
An automatic procedure for the evaluation of pole parts for dimensionally
regularised massless Feynman integrals has been used to evaluate
two loop counterterms in a non-abelian gauge theory with fermions
and for the simplification of amputated amplitudes in the Weinberg-Salam model.
The methods used in this paper can be readily extended to handle
integrals of higher loop order, and integrals involving masses.
Our approach succeeds in eliminating all tedious hand calculations.
The time required to perform calculations automatically is generally small
compared to the time required to prepare input and organise a calculational
sequence.
\section{Acknowledgements}
We would like to thank A. A. Rawlinson and A. C. Kallionatis for their
useful suggestions and assistance.
\section{Appendix}
This appendix contains the usage messages for functions described in this
article.
The following usage messages are for functions which simplify
products of Dirac gamma matrices.
\begin{verse}
{\tt comm::usage = "comm[x, a, b1, b2, \ldots] commutes a to the} \\
{\tt right through b1.b2 \ldots wherever the pattern a.b1.b2 \ldots } \\
{\tt appears in x (up to 5 b's programmed)."}
\end{verse}
\begin{verse}
{\tt rcomm::usage = "rcomm[x, b1, b2, \ldots, a] commutes a to} \\
{\tt the left through b1.b2 \ldots wherever the pattern b1.b2 \ldots .a} \\
{\tt appears in x (up to 5 b's programmed)."}
\end{verse}
\begin{verse}
{\tt con::usage = "con[x, a, b1, b2, \ldots] evalates a.b1.b2 \ldots .a} \\
{\tt by contracting out the a's. Replacement made wherever the} \\
{\tt pattern appears in x (Up to 6 b's programmed)."}
\end{verse}
\begin{verse}
{\tt trace::usage = "trace[x] evaluates the trace of a linear} \\
{\tt combination of products of gamma matrices that does not} \\
{\tt include gamma\_5."}
\end{verse}
\begin{verse}
{\tt movel::usage = "movel[x] commutes all helicity projection} \\
{\tt operators to the left most position in each term in the} \\
{\tt expression x."}
\end{verse}
\begin{verse}
{\tt mover::usage = "mover[x] commutes all helicity projection} \\
{\tt operators to the right most position in each term in the} \\
{\tt expression x."}
\end{verse}
The following usage messages are for functions which simplify Lorentz tensors.
\begin{verse}
{\tt contract::usage = "contract[x, other, s] will eliminate} \\
{\tt repeated index s from x. A list other, of all other} \\
{\tt symbols appearing in the arguments of g's must be passed} \\
{\tt to the function."}
\end{verse}
\begin{verse}
{\tt contractall::usage = "Given a list of repeated indices,} \\
{\tt repeated, and a list, other, of any indices appearing in} \\
{\tt the arguments of g's that are not in repeated,} \\
{\tt contractall[x, repeated,other] will eliminate all} \\
{\tt of the repeated indices from x."}
\end{verse}
\begin{verse}
{\tt slash::usage = "slash[x, a, b] finds all patterns of the} \\
{\tt form c1.c2. \ldots .ci.a.d1.d2. \ldots .dj*g[a,b] in x and} \\
{\tt replaces them with c1.c2. \ldots .ci.b.d1.d2. \ldots .dj} \\
{\tt where either or both of i and j may be zero."}
\end{verse}
\begin{verse}
{\tt preps::usage = "preps[x, k, p] replaces g[k, p] by} \\
{\tt (g[k+p, k+p] - g[k, k] - g[p, p])/2 everywhere in x."}
\end{verse}
\begin{verse}
{\tt prepd::usage = "prepd[x, k, p] replaces g[k, p] by} \\
{\tt (g[p, p] + g[k, k] -g[k-p, k-p])/2 everywhere in x."}
\end{verse}
The following usage messages are for functions which perform Feynman
integration.
\begin{verse}
{\tt inttensor::usage = "inttensor[x, q, p, a, b, \ldots , j, l]} \\
{\tt integrates tensors (g[q, a] g[q, b] \ldots )/(g[q, q]$\hat{\:}$j
g[q $\pm$} \\
{\tt p, q $\pm$ p]$\hat{\:}$l) of low rank, where the integration variable} \\
{\tt q, external momentum p, powers j and l, and indices} \\
{\tt a,b \ldots are specified in the arguments of inttensor."}
\end{verse}
\begin{verse}
{\tt intgamma::usage = "intgamma[x, q, p, a, b, \ldots , j, l]} \\
{\tt integrates matrices (q\, g[q, a]\, g[q, b] \ldots )/(g[q, q]$\hat{\:}$j
g[q} \\
{\tt $\pm$ p, q $\pm$ p]$\hat{\:}$l) of low rank, where the integration} \\
{\tt variable q, external momentum p, powers j and l, and} \\
{\tt indices a,b \ldots are specified in the arguments of intgamma."}
\end{verse}
\begin{verse}
{\tt samepole::usage = "samepole[x, p] replaces x, the numerator} \\
{\tt of a Feynman integral for a gauge boson self energy diagram} \\
{\tt with a simpler expression that has the same pole} \\
{\tt part (p is the external momentum)."}
\end{verse}
\begin{verse}
{\tt poleform::usage = "poleform[x, p] evaluates the pole part} \\
{\tt of x, a Feynman integral for a gauge boson self energy} \\
{\tt diagram, after samepole has been applied to the numerator."}
\end{verse}
\pagebreak
|
\section{Introduction}
In a magnetic field the electronic density is elongated along the field
direction. There are many intuitive ways of seeing this. For example, in a
path integral description, since the zero field propagator is multiplied by
the phase factor $exp \bigl[ {ie \over {{\hbar}c}}\oint{\bf A}\cdot d\bf
l\bigr]$, each closed path is reweighted by a phase proportional
to the magnetic flux it encloses. In the average over paths which produces
the electronic density, paths parallel to the field, enclosing no flux,
receive a weight of one; while paths perpendicular to the field, enclosing
flux, receive a reduced weight. The resulting average for the electronic
density, thus, has an elongated shape.\
The lowest order statistical model however leads to an isotropic
density[1]. The anisotropy requires gradient corrections [2]. The necessity of
going beyond the lowest order has also spawned a variety of approximate
methods to regain anisotropy [3,4,5,6,7,8].\
In this Letter, we first review the fundamental results of CDFT.
Then, neglecting the current variation, we rederive the gradient corrections
of Tomishima and Shinjo. Finally, taking this variation into account, we
calculate all second order electronic density gradient corrections.
\section{Current density functional formalism}
For systems in a magnetic field, Vignale and Rasolt [9]
show that the density $\rho$ and the paramagnetic current density
${\bf j}_p$, related to the gauge-invariant total current density, by
\begin{eqnarray}
{\bf j({\bf r})} = {\bf j_p({\bf r})} +{e\over
mc}\rho({\bf r})A({\bf r})
\label{current}
\end{eqnarray}
uniquely determine the external scalar, $V_{x}({\bf r})$, and vector
potential $\bf A({\bf r})$.
The ground state energy functional, which may be written as
\begin{eqnarray}
& &E[\rho,{\bf j}_p] = K[\rho,{\bf j}_p] +\int d{\bf r} \, \rho({\bf r}) V_{x}
({\bf r}) + {e^2\over2} \int \int d{\bf r}\, d{\bf r}' {{\rho({\bf
r})\rho({\bf r}')}\over{|{\bf r}-{\bf r}'|}} \nonumber\\
& & \mbox{\ \ \ \ \ \ \ \ \ \ \ \ \ } + {e \over c}\int d{\bf r} \,
{\bf j}_p({\bf r}){\bf A}({\bf r}) + {e^2\over 2mc^2}\int d{\bf r} \,
\rho({\bf r}){\bf A^2({\bf r})} + E_{xc}[\rho,{\bf j}_p]
\label{energie}
\end{eqnarray}
where $K[\rho,{\bf j}_p]$ is the kinetic energy of a free electron gas
in the external potentials $V_{x}({\bf
r})$ and $\bf A({\bf r})$, and $E_{xc}[\rho,{\bf j}_p]$ is the
exchange-correlation energy, is minimized at the ground state energy
by the correct density and current distribution.
In the spirit of Thomas-Fermi approximations we subsequently ignore the
exchange-correlation term and the Hartree mean field term.
Although $K[\rho,{\bf j}_p]$ is not invariant under a gauge transformation:
\begin{eqnarray}
& &\bf A({\bf r})\to \bf A({\bf r})-\bf \nabla\chi \nonumber\\
& &\bf j_p({\bf r})\to \bf j_p({\bf r})+ {e\over mc}\rho({\bf r}) \bf
\nabla\chi
\label{gauge}
\end{eqnarray}
the difference:
\begin{eqnarray}
U[\rho,{\bf j}_p] = K[\rho,{\bf j}_p]-
{m\over2}\int d{\bf r} \, {{\bf j_p\bf
j_p}\over\rho}=E[\rho,{\bf j}_p]-\int d{\bf r} \, \rho({\bf r}) V_{x}({\bf r})
-{m\over2}\int d{\bf r} \, {{\bf j\bf j}\over\rho}
\label{Uen}
\end{eqnarray}
which depends on the energy and physical current, is gauge invariant
and, since unchanged by the addition of the gradient of an arbitrary function,
depends on ${\bf j}_p$ only through
\mbox{$\bf \nu({\bf r})=\bf \nabla\times \bigl(\bf j_p /\rho\bigr)$},
the vorticity introduced by Vignale and Rasolt [9].
In the next two sections we determine the gradient expansion for $U$.
\section{Gradient corrections in strong fields}
We consider an electron gas in an uniform magnetic field ${\bf B}=
B_{0}\hat{{\bf z}}$. As a first step we will neglect the variation of
vorticity. Then the functional U is reduced to a functional of
only one variable, $\rho$.
We postulate an expression for U with gradient
terms:
\begin{eqnarray}
U[\rho] = \int d{\bf r} \, \biggl\{ f(\rho({\bf r}))+g_{\bot}(\rho({\bf
r})){\bigl|} {\bf \nabla}_{\bot}\rho {\bigr|}^2+g_{\|}(\rho({\bf
r})){\bigl|} {\bf \nabla}_{\|}\rho {\bigr|}^2 \biggr\}
\label{U}
\end{eqnarray}
where $\bot$ and $\|$ denote orientations perpendicular and parallel to the
field.
The function $f(\rho({\bf r}))$ can be determined from the Euler equation
\begin{eqnarray}
\begin{array}{c}\underline{\delta E[\rho({\bf r})]}\\\delta\rho({\bf r})
\end{array}=\mu\;.
\end{eqnarray}
For the assumed $U$ this gives
\begin{eqnarray}
& &f'(\rho({\bf r})) -2g_{\bot}(\rho({\bf
r})) {\bf \nabla}_{\bot}^2\rho -2g_{\|}(\rho({\bf
r})) {\bf \nabla}_{\|}^2\rho \nonumber\\
& & \mbox{\ \ \ \ } - g_{\bot}'(\rho({\bf
r})){\bigl|} {\bf \nabla}_{\bot}\rho {\bigr|}^2 -g_{\|}'(\rho({\bf
r})){\bigl|} {\bf \nabla}_{\|}\rho {\bigr|}^2=-V_{x}({\bf r})+\mu
\label{equa}
\end{eqnarray}
which, at uniform density $\rho_0$ and no external potential, is simplified
to:
\begin{eqnarray}
f'[\rho_0]=\mu_0
\label{equa1}
\end{eqnarray}
where $\mu_0$ is the chemical potential of a free electron gas.
$g_{\bot}(\rho({\bf r}))$ and $g_{\|}(\rho({\bf r}))$ will be determined by
comparing the calculated linear density response to a variation in the
external potential, $\delta\rho(\bf q)=\chi(\bf q)\delta{V_x}(\bf q)$, to
the free electron gas results. A second variation of the Euler
equation gives (after a Fourier transform),
\begin{eqnarray}
f''(\rho_0) +2g_{\bot}(\rho_0){\bf k}_{\bot}^2
+2g_{\|}(\rho_0){k}_{\|}^2=-1/\chi(\bf k)
\label{equa3}
\end{eqnarray}
at uniform density $\rho_0$.
As a preliminary example recall the case of no magnetic field.
The chemical potential $\mu=\hbar^{2}k_{F}^{2}/2m=
\hbar^{2}(3\pi^{2}\rho)^{2/3}/2m$ so eqn.~(\ref{equa1}) reproduces
the usual Thomas-Fermi result,
\begin{eqnarray}
f(\rho)={3\over 10}{\hbar^{2}\over m}(3\pi^{2})^{2/3}\rho^{5/3}
\end{eqnarray}
and the Lindhard response formula
\begin{eqnarray}
\chi( k)=-{m k_{F}\over \hbar^{2} \pi^{2}}\left[ {1\over 2}+
\begin{array}{c}\underline{ 2( k_{F}^{2}- k^{2})}\\ k_{F} k
\end{array}
\ln\left|
\begin{array}{c}\underline{2 k_{F}- k}\\2 k_{F}+ k\end{array}
\right| \right]
=-{m k_{F}\over \hbar^{2} \pi^{2}}\left(1-{1\over 3}( k/2 k_{F})^{2}\right)
+O(k^{4})
\end{eqnarray}
or
\begin{eqnarray}
\chi^{-1}( k)=-{\hbar^{2}\pi^{2}\over m k_{F}}\left(1+{1\over 3}( k/
2 k_{F})^{2} \right)+ O( k^{4})\;.
\end{eqnarray}
gives, on comparison with eqn.~(\ref{equa3}), the von Weizsacker gradient
corrections
\begin{eqnarray}
g_{\bot}=g_{\|}={\hbar^{2} \over 3 m}{\pi^{2}\over 8 k_{F}^{3}}=
{1\over 9}{1\over 8 \rho}{\hbar^{2}\over m}
\end{eqnarray}
(with the $1/9$ factor from the long wavelength comparison)[10].
In the case of
very strong fields, where all the electrons are in the lowest Landau
level and spin polarized, the free electron gas chemical potential
$\mu_0=\hbar\omega_{c}/2+ \hbar^{2} k_{F}^{2}/2m$ with
$k_{F}=2\pi^{2}\hbar\rho/m\omega_{c}$ so
\begin{eqnarray}
f(\rho)={{\hbar\omega_c} \over 2}\rho+{{2\pi^4\hbar^4}\over
{3m^3{\omega_c}^2}}{\rho}^3
\label{equa2}
\end{eqnarray}
The free electron gas susceptibility for this case is
well known since Horing [11,12]
\begin{eqnarray}
\chi({\bf k})=-{2\over {4{\pi^2}{\omega_c}l^4 \hbar}}
{{F({\bf k}_{\bot})} \over{{k}_{\|
}}} \mbox{\ ln \ } \Biggl| {{2k_F+{k_\|}} \over {2k_F-{k_\|}}} \Biggr|
\label{equa4}
\end{eqnarray}
where ${F({\bf k})}=e^{- k^2 l^2/2}$ and $l$ is the
associated magnetic length defined by: ${l^2}=\hbar c/eB_0$.
The expansion of $\chi^{-1}$ to second order in $\bf k$
gives:
\begin{eqnarray}
-{\chi^{-1}(\bf k)}={{4{\pi^4}{\hbar^4}\rho}\over {m^3{\omega_c}^2}}+
{{2{\pi^4}{ \hbar^5}\rho}\over {m^4{\omega_c}^3}}{{\bf k}_{\bot}^2}
-{{\hbar^2} \over {12\rho m}}{{k_\|}^2}
\label{equa5}
\end{eqnarray}
so
\begin{eqnarray}
g_{\bot}={\pi^{4}\hbar^{5}\rho\over m^{4}\omega_{c}^{3}}
\end{eqnarray}
and
\begin{eqnarray}
g_{\|}= -{1\over 24 \rho}{\hbar^{2}\over m}
\end{eqnarray}
The resulting energy functional:
\begin{eqnarray}
E[\rho] = \int d{\bf r} \, \biggl\{ {{{\hbar\omega_c} \over
2}}\rho+{{{2\pi^4\hbar^4}\over
{3m^3{\omega_c}^2}}{\rho}^3}+{{{\pi^4}{\hbar^5}}\over
{m^4{\omega_c}^3}}{\rho{\bigl|} {\bf \nabla}_{\bot}\rho {\bigr|}^2}-{{\hbar^2}
\over 24m}{{\bigl|}{{\bf \nabla}_{\|}\rho {\bigr|}^2} \over \rho}
\biggr\} +\int d{\bf r} \, \rho V_{x}
\label{equa6}
\end{eqnarray}
is the same as the one originally derived by Tomishima and Shinjo [6].\\
Note that even though the free electron states are still plane waves along
the field direction the gradient correction for this direction
has the opposite sign from the free field case. This is a consequence of
the 1-dimensional nature of the high field limit. Explicitly, the free
fermion response function in 1-dimension is
\begin{eqnarray}
\chi(k) ={-m\over \hbar^{2} \pi k}
\mbox{\ ln \ } \Biggl| {{2k_F+{k}} \over {2k_F-{k}}} \Biggr|
={-m\over \hbar^{2}\pi k_{F}}\left\{1+{1\over
12}k^{2}/k_{F}^{2}\right\}+O(k^{4})
\end{eqnarray}
or
\begin{eqnarray}
\chi^{-1}(k)={-\hbar^{2}\pi k_{F}\over m}\left\{1-{1\over
12}k^{2}/k_{F}^{2}\right\}
+O(k^{4})
\end{eqnarray}
with $k_{F}=\pi\rho$. Therefore [13]
\begin{eqnarray}
g_{\|}=-{\pi\over 24 k_{F}}{\hbar^{2}\over m}=-{1\over 24\rho}{\hbar^{2}\over
m}
\end{eqnarray}
which is exactly the high magnetic field result derived above.
\section{Current corrections}
The Tomishima and Shinjo approximation is obtained
when the variation of vorticity is omitted.
However, this variation involves the gradient of
$\rho$ and can not a priori be neglected. We therefore generalize our
previous expression for $U$ to include vorticity as follows:
\begin{eqnarray}
U[{\rho},{\bf \nu}] = \int d{\bf r} \, \biggl\{ f(\rho({\bf r}),{\bf
\nu({\bf r})})+g_{\bot}(\rho({\bf r}),{\bf \nu({\bf r})})
{\bigl|} {\bf \nabla}_{\bot}\rho {\bigr|}^2+g_{\|}(\rho({\bf r}),{\bf
\nu({\bf r})})
{\bigl|} {\bf \nabla}_{\|}\rho {\bigr|}^2 \biggr\}
\label{Urho}
\end{eqnarray}
{}From the definition of the vorticity:
\begin{eqnarray}
{\bf \nu({\bf r})}=-{{{eB_0}\over mc}{\bf u}_z}+\bf \nabla\times \bigl({\bf j
\over \rho}\bigr) \;.
\label{vort}
\end{eqnarray}
Consequently, we do not need to add in the $U[{\rho},{\bf
\nu}]$ expansion terms in
${\bigl|} {\bf \nabla\nu} {\bigr|}$ which are higher order.
Variation of the energy with respect to ${\bf j}_p$,
$\delta E/\delta {\bf j}_p=0$, using the chain rule result
\[
{\delta \over \delta {\bf j}_{p}({\bf r})}
={1\over{\rho({\bf r})}}
{\bf \nabla}\times {\delta\over{\delta\nu ( {\bf r})}}
\]
gives:
\begin{eqnarray}
{\bf j}=-{1 \over m}{\bf \nabla\times} {{\partial f}\over{\partial{\bf
\nu}}}
\label{cou5}
\end{eqnarray}
to second order.
The Euler equation, $\delta E/\delta\rho({\bf r})=\mu$, becomes:
\begin{eqnarray}
& &{{\partial f}\over{\partial\rho}}[{\rho},{\nu}]
-2g_{\bot} {\bf \nabla}_{\bot}^2\rho -2g_{\|}
{\bf \nabla}_{\|}^2\rho \nonumber\\
& & \mbox{\ } -{{\partial g_{\bot}}\over {\partial\rho}}
{\bigl|} {\bf \nabla}_{\bot}\rho {\bigr|}^2 -{{\partial g_{\|}}\over
{\partial\rho}}
{\bigl|} {\bf \nabla}_{\|}\rho {\bigr|}^2=-V_{x}({\bf r})+\mu+{m\over2}
{{\bf j\bf j}\over\rho^2}
\label{equa15}
\end{eqnarray}
At uniform density we get:$${{\partial
f}\over{\partial\rho}}[{\rho_0},{\nu_0}]=\mu[{\rho_0},{\nu_0}] \mbox{\ \
\ \ with \ \ \ \ } {{\bf \nu}_0}=\omega_{c}\hat{{\bf z}} $$
so the expression for $f$ is formally the same as (\ref{equa2}) with the
variables $\rho$ and $\bf \nu$:
\begin{eqnarray}
f[{\rho},{\bf \nu}]={{\hbar \over 2}|{\bf \nu}|}\rho+{{2\pi^4\hbar^4}
\over {3m^3{|{\bf \nu}|}^2}}{\rho}^3 \;.
\label{equa9}
\end{eqnarray}
With equation (\ref{cou5}), we can now calculate the total current density,
\begin{eqnarray}
{\bf j} {\simeq}- \biggl\{ {{\hbar \over 2m}-{{4\pi^4\hbar^4}\over{m^4}}{\rho^2
\over{|\nu_0|^3}}} \biggr\}{{\bf \nabla}\rho\times {\bf u}_z}
\label{equa10}
\end{eqnarray}
where we have replaced $|\bf \nu|$ by $|\bf \nu_0|$ in the second member
of (\ref{equa10}) since the $\bf \nu$ variations are higher order.
In the limit
of very strong fields we find the result of Skudlarski and Vignale~[14].
On the other hand equation (\ref{equa3}) remains unchanged at $\rho_0$
since all
the additional terms are proportional to ${\bf \nabla}\rho$. Thus, the
expressions for $g_{\bot}$, $g_{\|}$ are absolutely the same as in section III
with ${\omega_c} \to {|\bf \nu|}$.
In return, E is modified by the
presence of the current, and for a comparison with Tomishima and Shinjo
we have to develop $\bf \nu$, eqn. (\ref{vort}), using eqn. (\ref{equa10}):
\begin{eqnarray}
{|{\bf \nu}|} = {\omega_c} - {\hbar \over 2m}{\biggl[}{{{\bigl|}{{\bf
\nabla}_{\bot} \rho{\bigr|}^2} \over \rho^2} - {{{\bf
\nabla}_{\bot}^2 \rho} \over \rho}}{\biggr]} - {{{4\pi^4}{\hbar^4}}\over
{m^4{\omega_c}^3}}{\biggl[}{{\bigl|}{\bf
\nabla}_{\bot} \rho{\bigr|}^2}+{\rho{\bf
\nabla}_{\bot}^2 \rho} {\biggr]} \;.
\label{equa12}
\end{eqnarray}
Using this expansion in eqn. (\ref{equa9}) for $f[\rho,\nu]$ and with
eqn. (\ref{equa10}) for the current we finally obtain for the energy
functional E:
\begin{eqnarray}
& E[\rho] = \int d{\bf r} \, &\biggl\{ {{{\hbar\omega_c} \over
2}}\rho+{{{2\pi^4\hbar^4}\over
{3m^3{\omega_c}^2}}{\rho}^3}-{{\hbar^2}\over 8m}{{\bigl|}{{\bf
\nabla}_{\bot}\rho
{\bigr|}^2} \over \rho}+{{{3\pi^4}{\hbar^5}}\over
{m^4{\omega_c}^3}}{\rho{\bigl|} {\bf \nabla}_{\bot}\rho
{\bigr|}^2}\nonumber \\
& &-8m\left({{{\pi^4}{\hbar^4}}\over{m^4{\omega_c}^3}}\right)^{2}
\rho^{3}\left|{\bf \nabla}_{\bot}\rho\right|^{2}
-{{\hbar^2}\over 24m}
{{\bigl|}{{\bf \nabla}_{\|}\rho {\bigr|}^2} \over \rho} \biggr\} +\int
d{\bf r} \, \rho V_{x}
\label{equa13}
\end{eqnarray}
\section{Conclusion}
Current corrections do not change anything along the field direction
since this correction is only due of the 1-dimensional nature of the
high field limit. The transverse corrections however, have greatly changed.
First we note an important additional
term in $|{\bf \nabla}_{\bot}\rho |^{2}/ \rho$ which does not
depend on the strong field intensity. Perpendicular gradients are
favored by this term. The next term in ${{\bf
\nabla}_{\bot}\rho}$, which is small compared to the constant term,
is the same as found by Tomishima and Shinjo but with a factor $+3$.
This term and the third term in
${{\bf \nabla}_{\bot}\rho}$ decrease with $B_0$. Therefore,
contrary to Tomishima and Shinjo the total transverse corrections increase with
$B_0$. This result modifies the transverse pinch.
In summary, as expected, the
behavior of the total corrections produces an anisotropic
density profile. This anisotropy is caused by a reduction of dimensionality
due to the high field limit. Our approach, which is technically simple, allows
us
to extend the calculus to finite temperatures and to include some Landau
levels. But since the minimization of eqn.~(\ref{equa13}) is rather
complicated,
we leave its numerical solution for a forthcoming article.
\acknowledgements
One of us, A.M. would like to thank Dr D. Levesque for his kind
hospitality at LPTHE and the CEA for financial support. LPTHE is Unit\'e de
Recherche de L'Universit\'e Paris XI associ\'ee au CNRS.
Work done at the Lawrence Livermore National Laboratory is supported by the
U.S. Department of Energy under Contract W-4705-Eng-48.
|
\section{Introduction}
The Standard Model of electroweak interactions is undeniably one of
the most successful theories in particle physics, accommodating
all observed phenomena to date. For all its glory,
the theory has the undesirable feature that most of its parameters,
although physically well motivated, must be determined empirically.
As an example, the way in which the weak interaction can mix quarks
of different flavor is described by the Cabibbo-Kobayashi-Maskawa matrix\cite{cabibbo,km}.
The nine entries in this $3\times 3$ matrix contain information
about the relative strengths and phases with which the ($u,c,t$)
quarks couple to the ($d,s,b$) quarks, and their values
must be measured experimentally. Our present knowledge of the magnitudes
of these parameters are summarized below\cite{pdg}.
\begin{equation}
{\bf\rm V} =
\left(\matrix{
V_{ud} & V_{us} & V_{ub} \cr
V_{cd} & V_{cs} & V_{cb} \cr
V_{td} & V_{ts} & V_{tb} \cr}\right) =
\left(\matrix{
0.9747\rightarrow 0.9759 & 0.218\rightarrow 0.224 & 0.002\rightarrow 0.005 \cr
0.218\rightarrow 0.224 & 0.9738\rightarrow 0.9752 & 0.032\rightarrow 0.048 \cr
0.004\rightarrow 0.015 & 0.030\rightarrow 0.048 & 0.9988\rightarrow 0.9995\cr}\right)
\label{ckmb}
\end{equation}
\noindent
Note that the extreme diagonal elements, $V_{ub}$ and $V_{td}$, are
the least well determined.
\medskip
In this paper the latest inclusive and exclusive measurements of
the ratio $|V_{ub}/V_{cb}|$~will be discussed. These were both made
by the CLEO collaboration, using data accumulated at the
Cornell Electron Storage Ring (CESR). In the CLEO-II detector, three
concentric tracking devices provide charged-particle momentum resolution of
$\sigma_p/p = 0.005+0.0015p$ ($p$ in GeV/c), and a 7800 crystal CsI
calorimeter provides neutral shower energy resolution of
$\sigma_E/E = 0.019 - 0.001E + 0.0035/E^{0.75}$.
More detailed information is available elsewhere\cite{cleodetector}.
\smallskip
The data sample used in the inclusive analysis consists of
$924~pb^{-1}$ accumulated with the CESR center of mass energy tuned to
the $\Upsilon$(4S) resonance, and $416~pb^{-1}$ accumulated at energies
below the $B\overline{B}$ production threshold. The more recent
exclusive analysis was done using approximately a factor of two more data.
\section{Inclusive Measurements}
All determinations of $V_{ub}$ ~to date have been made by studying charmless
semileptonic decays of $B$ mesons produced in $e^+-e^-$ collisions
with center of mass energy at or near the $\Upsilon$(4S) resonance
\cite{cleo90,argus90,argus91,cleo93,jknelson}.
These are all measurements of the inclusive momentum-dependent rate of
leptons, $dN_\ell/dP_\ell$, from $B\rightarrow X_u\ell\nu$
decays\cite{leptons}.
Since the rate of $b\rightarrow u\ell\nu$ ~is very small, the main experimental
challenge in these analyses is the suppression of backgrounds. For
inclusive $b\rightarrow u\ell\nu$ ~analyses, the three main sources of unwanted leptons are
$b\rightarrow c\ell\nu$~decays, other $B$ meson decays (for example
$B\rightarrow\psi X, \psi\rightarrow\ell^+\ell^-$), and continuum (non
$B\overline{B}$) events. ``Fake'' leptons, for example $\pi$'s that
penetrate the detector iron and are misidentified as $\mu$'s, must also
be considered.
\subsection{$b\rightarrow c\ell\nu$~Suppression}
The elimination of $b\rightarrow c\ell\nu$~decays from the $b\rightarrow u\ell\nu$~sample is achieved
with a simple lepton momentum cut. As can be seen in
Fig.~\ref{leptonspectrum}(a), the kinematic endpoint
momentum of leptons from $B\rightarrow X_c\ell\nu$ is
about 2.4 GeV/c\cite{batrest}. Fig.~\ref{leptonspectrum}(b) shows
theoretical lepton spectra for several models of
$b\rightarrow u\ell\nu$\cite{accmm,isgw,wsb}.
Although the models differ significantly in detail,
they all share the basic kinematic feature that the lepton momentum
endpoint extends to $P_\ell\sim 2.6$ GeV/c.
\smallskip
The approach taken in all inclusive analyses has been to examine
the yield of leptons only in the endpoint region between
$\sim 2.4$ GeV/c and $2.6$ GeV/c\cite{lowerpcut}. This restriction
eliminates most of the $b\rightarrow c\ell\nu$ contamination, but
only at the price of introducing a severe model-dependence in the
procedure used to extract $V_{ub}$~from the data. The reason for this
is clear upon examination of Fig.~\ref{leptonspectrum}(b).
The models shown differ considerably, both in overall
rate (plot area) and lepton momentum dependence (plot shape).
The effect of these differences are discussed in a later section.
\begin{figure}[htb]
\vspace{0.in}
\begin{center}
\unitlength 1.0in
\begin{picture}(6.,2.5)(0,0)
\put(0.,0.)
{\psfig{bbllx=0in,bblly=0in,width=2.8in,file=lepspec.ps}}
\put(3.,0.)
{\psfig{bbllx=0in,bblly=0in,width=2.8in,file=btouspec.ps}}
\end{picture}
\medskip
\caption{(a) Inclusive lepton momentum spectrum showing the
$b\rightarrow c\ell\nu$ ~(dashed line) and $b\to c\to y\ell\nu$ (dot-dashed line) components,
as well as the total fit to CLEO data using the ACCMM model (solid line).
(b) The predicted $b\rightarrow u\ell\nu$ ~lepton momentum spectra for the ACCMM (solid line),
ISGW (dotted line) and WSB (dashed line) models.}
\label{leptonspectrum}
\end{center}
\end{figure}
\subsection{Continuum Suppression}
Leptons from non-$B\overline{B}$ continuum events, for example
$e^+e^-\rightarrow c\overline{c},~ c\rightarrow s\ell\nu$, are not
kinematically excluded from the inclusive $b\rightarrow u\ell\nu$ ~signal
region between 2.4 and 2.6 GeV/c. This is a large source of background,
and is dealt with in two steps.
First, a set of continuum suppression requirements are used to eliminate
most of these events. Second, the remaining background is removed by
subtracting the luminosity scaled lepton spectrum obtained by analyzing
``signal free'' data accumulated at center of mass energies below the
$B\overline{B}$ production threshold.
\smallskip
The most powerful continuum suppression requirements are topological
in nature, designed to select the typically spherical $B\overline{B}$
events while rejecting the more ``jetty'' continuum background.
The shape variable used by CLEO is the Fox-Wolfram parameter
$R_2 = H_2/H_0$\cite{r2}. The missing momentum of an event, $p_{miss}$,
provides additional discrimination against continuum processes.
This quantity should be large for $b\rightarrow u\ell\nu$ events where the
neutrino carries off appreciable momentum. The requirements that $R_2 < 0.2$,
that $p_{miss} > 1~$GeV/c, and that the lepton and missing momentum point into
opposite hemispheres are used. The net effect of these cuts is to
reduce the continuum background by a factor of 70 while maintaining
38\% efficiency for the $b\rightarrow u\ell\nu$ signal.
\smallskip
The statistical uncertainty introduced by the continuum subtraction is
a function of the size of the continuum data sample. The normal operating
mode of CESR/CLEO-II is to accumulate data on the $\Upsilon$(4S) resonance
two-thirds of the time, and just below $B\overline{B}$ threshold
the remaining one-third. The resulting continuum data sample is large enough
that the error introduced by the subtraction are much smaller than
the statistical error on the signal yield.
\subsection{Other Backgrounds}
After continuum suppression and subtraction, the events remaining in the
lepton momentum endpoint region are either true $b\rightarrow u\ell\nu$ ~signal or
non-continuum background. This background is due to leptons from
$b\rightarrow c\rightarrow X\ell\nu$, from $B\rightarrow\psi$ and
$B\rightarrow\psi^\prime$ decays, and from fake leptons. The fraction
of events in the lepton spectrum signal region due to these sources
is less than 10\%, is well understood from studying the data, and is
carefully accounted for when calculating the final $b\rightarrow u\ell\nu$~yield.
\subsection{Signal Extraction and Model Dependence}
Fig.~\ref{endpoint}(a) shows lepton spectra from $\Upsilon$(4S)
and continuum data. The solid line is a fit to the continuum lepton
distribution. Fig.~\ref{endpoint}(b) shows the result of luminosity
scaling and subtracting this fitted line from the $\Upsilon$(4S)
data, as well as the predicted contribution from $b\rightarrow c\ell\nu$ ~decays.
A significant excess of events is seen, indicating the presence of
charmless B decays.
\begin{figure}[htb]
\vspace{0.in}
\begin{center}
\unitlength 1.0in
\begin{picture}(3.,3.)(0,0)
\put(-0.5,0.1)
{\psfig{bbllx=0pt,bblly=0pt,width=3.0in,height=3.0in,file=inc.ps}}
\end{picture}
\bigskip
\caption{CLEO-II inclusive lepton momentum distributions.
Shown in (a) are lepton spectra
from $\Upsilon$(4S) (filled circles) and continuum (hatched histogram) data,
as well as a fit to the continuum lepton distribution (solid line).
The filled circles in (b) show the result of subtracting the fitted and
scaled continuum distribution from the $\Upsilon$(4S) data.
The solid histogram shows the predicted contribution from $b\rightarrow c\ell\nu$ ~decays.}
\label{endpoint}
\end{center}
\end{figure}
Extracting $|V_{ub}/V_{cb}|$~ from the endpoint data involves several model
dependent factors:
\begin{equation}
\left|{V_{ub}\over V_{cb}}\right|={\Delta B_{ub}(p) \over d(p)B_{cb}}
\label{extvub}
\end{equation}
In this expression $B_{cb}$ is the $b\rightarrow c\ell\nu$ ~branching
ratio and $\Delta B_{ub}(p)$ is the partial branching
ratio observed in the endpoint signal region\cite{pdependent}:
\begin{equation}
\Delta B_{ub}(p) = {N_{ub}(p)/\epsilon(p) \over 2N_{B\overline{B}}}
\label{partial}
\end{equation}
where $N_{ub}(p)$ is the number of $b\rightarrow u\ell\nu$~events in the signal region,
$\epsilon(p)$ is the detection efficiency for these events,
and $2N_{B\overline{B}}$ is the number of $B$ mesons produced.
$\Delta B_{ub}(p)$ is somewhat model dependent since the detection efficiency
$\epsilon$ depends weakly on the shape of the lepton momentum spectrum.
\smallskip
The term $d(p)$ is the product of several strongly model-dependent
factors:
\begin{equation}
d(p) = f_u(p){\gamma_u \over \gamma_c}
\label{scalefact}
\end{equation}
where $f_u(p)$ is the fraction of the total $b\rightarrow u\ell\nu$~spectrum in the
endpoint signal region (model dependence on spectrum shape),
and $\gamma_u$ relates the semileptonic
width and $V_{ub}$~via $\Gamma_{b\rightarrow u\ell\nu} =
\gamma_u |V_{ub}|^2$\cite{gammac}, (model dependence on spectrum area).
\smallskip
Table \ref{incresults} shows the factors $d(p)$ and $\epsilon(p)$
as well the final extracted CLEO values of $\Delta B_{ub}(p)$
and $|V_{ub}/V_{cb}|$ for various models. The model-dependence of
$|V_{ub}/V_{cb}|$ is larger than the experimental uncertainty,
and at present is the biggest factor limiting the accuracy of this
measurement.
\begin{table}[htb]
\caption{CLEO-II results for $d(p)$, $\epsilon(p)$, $\Delta B_{ub}(p)$
and $|V_{ub}/V_{cb}|$. The values listed for $d(p)$ and $\epsilon(p)$
were calculated for the momentum range $2.4 < p_\ell({\rm GeV/c}) < 2.6$,
and are shown to illustrate the effect of model dependence. The
values of $\Delta B_{ub}(p)$ and $|V_{ub}/V_{cb}|$ represent the
extended signal region $2.3 < p_\ell({\rm GeV/c}) < 2.6$.}
\label{incresults}
\begin{tabular}{lcccc}
Model & $d(p)$ & $\epsilon(p)$ & $10^6\times\Delta B_{ub}(p)$
& $|V_{ub}/V_{cb}|$ \\ \hline
ACCMM\cite{accmm}
& $0.12$ & $0.16\pm 0.01$ & $154\pm 22\pm 20$ & $0.076\pm 0.008$ \\
ISGW\cite{isgw}
& $0.05$ & $0.21\pm 0.02$ & $121\pm 17\pm 15$ & $0.101\pm 0.010$ \\
WSB\cite{wsb}
& $0.11$ & $0.20\pm 0.02$ & $122\pm 17\pm 16$ & $0.073\pm 0.007$ \\
KS\cite{ks}
& $0.19$ & $0.22\pm 0.02$ & $115\pm 16\pm 15$ & $0.056\pm 0.006$ \\
\end{tabular}
\end{table}
\section{Exclusive Measurements}
The CLEO collaboration has recently observed the exclusive charmless
semileptonic decay mode $B\rightarrow\pi\ell\nu$. Measurement of this and other exclusive
channels will provide a new avenue for determining $|V_{ub}/V_{cb}|$ and
studying the $b\rightarrow u\ell\nu$ ~form factors, and should be a powerful tool for testing
the various available models.
\smallskip
This section will provide some details of the CLEO $B\rightarrow h\ell\nu$
analysis, where $h$ is $\pi^\pm$, $\pi^0$, $\rho^\pm$, $\rho^0$ or $\omega$.
Table \ref{exclpred} lists the theoretical predictions for the partial widths
of $B\rightarrow\pi\ell\nu$~ and $B\rightarrow\rho\ell\nu$, and their ratio. Note again that the predicted
widths show a strong model-dependence.
\begin{table}[htb]
\caption{Predictions for the exclusive partial widths $\Gamma(B^0\rightarrow\pi^-\ell\nu)$ and $\Gamma(B^0\rightarrow\rho^-\ell\nu)$.}
\label{exclpred}
\begin{tabular}{lccc}
Model & $\Gamma(B^0\rightarrow\pi^-\ell\nu)$~[$10^{12}|V_{ub}|^2$ sec$^{-1}$]
& $\Gamma(B^0\rightarrow\rho^-\ell\nu)$~[$10^{12}|V_{ub}|^2$ sec$^{-1}$]
& $\Gamma(B^0\rightarrow\rho^-\ell\nu)/\Gamma(B^0\rightarrow\pi^-\ell\nu)$ \\ \hline
ISGW\cite{isgw} & 2.1 & 8.3 & 4.0 \\
WSB\cite{wsb} & 6.3 -- 10.0 & 18.7 -- 42.5 & 3.0 -- 4.3 \\
KS\cite{ks} & 7.25 & 33.0 & 4.6 \\
ISGW II\cite{isgwii} & 9.6 & 14.2 & 1.5 \\
FGM\cite{faustov} & $3.1\pm 0.6$ & $5.7\pm 1.2$ & $1.8\pm 0.5$ \\
\end{tabular}
\end{table}
\subsection{Neutrino Reconstruction}
The difficulty with exclusive reconstruction of semileptonic decays
is the lack of knowledge of the undetected neutrino's 4-momentum.
Using the large solid-angle coverage of the CLEO-II detector\cite{coverage}
to measure the {\it total} momentum and energy of the {\it rest} of
the event, CLEO is able to infer $(E_\nu,\vec{p}_\nu)$ of the neutrino
from the missing momentum and energy $(E_{miss},\vec{p}_{miss})$
of the event as a whole:
\begin{equation}
\label{eneutrino}
E_\nu\sim E_{miss} = 2E_{beam} - E^{tot}_{meas}~~{\rm and}~~
\vec{p}_\nu\sim \vec{p}_{miss} = - \vec{p}^{~tot}_{meas}
\end{equation}
where $E^{tot}_{meas}$ ($\vec{p}^{~tot}_{meas}$) is the total
measured energy (momentum) of all tracks and showers in the event.
Analysis of Monte Carlo generated $B\rightarrow\pi\ell\nu$~events yields
resolutions of 110 MeV and 260 MeV for $|\vec{p}_\nu|$ and $E_\nu$
respectively.
\smallskip
This method assumes the neutrino from $b\rightarrow u\ell\nu$~is the only
undetected particle, making it crucial to reject events containing
additional unseen particles (neutrinos and/or $K_L$'s). This is
accomplished by demanding that candidate events have only one
identified lepton, have zero net charge, and that the reconstructed
mass of the neutrino ($m_\nu^2=M^2_\nu - P^2_\nu$) be consistent with
zero\cite{zeromasscut}.
\smallskip
Additional constraints are placed on the final state hadrons.
Candidate $\pi^0$'s must have a
$2\gamma$ invariant mass within $2\sigma$ (about 12 MeV) of the nominal
$\pi^0$ mass, and $2\pi$ ($3\pi$) combinations must have invariant
mass within 90 MeV (30 MeV) of the $\rho$ ($\omega$) mass.
Identified leptons are required to have a momentum greater than
1.5 GeV/c (2.0 GeV/c) in the $\pi\ell\nu$ ($\rho/\omega\ell\nu$) modes.
\smallskip
For events passing the above neutrino, meson, and lepton selection
requirements, the reconstructed ``beam constrained'' $B$ mass
$m_B\equiv\sqrt{E^2_{beam}-|\vec{p}_h + \vec{p}_\ell + \vec{p}_\nu|^2}$
and energy difference $\Delta E\equiv E_{beam}-(E_h-E_\ell-|\vec{p}_\nu|)$
are calculated. Real $B\rightarrow h\ell\nu$~events should have $\Delta E$ close to
zero and $m_B$ close to the $B$ meson rest mass. Monte Carlo studies
are used to determine the optimum location of the signal region
in the $\Delta E-m_B$ plane, finding -250 $< \Delta E {\rm (MeV)} <$ 150
and 5.265 $< m_B {\rm (GeV/c^2)} <$ 5.2875.
\smallskip
As in the inclusive analysis, event shape variables\cite{r2} are used to
suppress continuum backgrounds, and a continuum subtraction is done.
The contribution from fake leptons is studied using data, and is
also subtracted. The remaining background is predominantly due to
$b\rightarrow c\ell\nu$~decays containing an additional $\nu$ or $K_L$, and to cross-feed
from other $b\rightarrow u\ell\nu$~modes.
\begin{figure}[htb]
\vspace{0.in}
\begin{center}
\unitlength 1.0in
\begin{picture}(3.,3.)(0,0)
\put(-2.0,0.0)
{\psfig{bbllx=0pt,bblly=0pt,height=2.5in,file=excl.eps}}
\end{picture}
\bigskip
\caption{Beam constrained mass distributions for combined $\pi\ell\nu$ (left)
and $\rho\ell\nu$ (right) modes. The points are continuum and fake
subtracted data, the histograms show the contribution from signal (hollow),
$b\rightarrow c\ell\nu$~(shaded), and $b\rightarrow u\ell\nu$~cross-feed (hatched).}
\label{lawrence}
\end{center}
\end{figure}
\subsection{Signal Extraction and Model Dependence}
After subtracting the contributions from continuum and fake leptons and
selecting events in the $\Delta E$ signal region, the beam-constrained
mass ($m_B$) distributions of the five signal modes are simultaneously
fitted. The shapes of the $b\rightarrow c\ell\nu$~and cross-feed background contributions
are obtained from Monte Carlo. The normalization of the $b\rightarrow c\ell\nu$~background
for each mode are free parameters in the fit. The isospin relations
${1\over 2}\Gamma(B^0\rightarrow\pi^-\ell^+\nu)=
\Gamma(B^+\rightarrow\pi^0\ell^+\nu)$ and
${1\over 2}\Gamma(B^0\rightarrow\rho^-\ell^+\nu)=
\Gamma(B^+\rightarrow\rho^0\ell^+\nu)\approx
\Gamma(B^+\rightarrow\omega\ell^+\nu)$ constrain the relative neutral
and charged meson rates. The total signal and cross-feed background
for all five modes is in this way parameterized by two numbers $N_\pi$
and $N_\rho$, the total yield of $\pi$ events and $\rho$ events
respectively.
\smallskip
The beam-constrained mass distributions for $B\rightarrow\pi\ell\nu$ ~and $B\rightarrow\rho\ell\nu$
{}~are shown in Fig.~\ref{lawrence}. The results of the fit
for the case of $B\rightarrow\pi\ell\nu$~ are summarized in Table \ref{pifit}, where
the signal yield, efficiency, $b\rightarrow c\ell\nu$~ background, and $b\rightarrow u\ell\nu$~ cross-feed
probabilities are shown for both the ISGW and WSB models. Note that
although the yields are very similar for both models, the efficiencies
differ significantly. The contribution from higher mass and non-resonant
$b\rightarrow u\ell\nu$~modes, denoted ``other $u\ell\nu$'', is fixed by inclusive lepton
endpoint spectrum measurements.
\smallskip
Many checks have been performed to verify that the observed signal is real,
including fitting the energy difference ($\Delta E$) distribution rather
than the beam-constrained mass, examining the lepton momentum spectrum and the
distribution of angles between the pion and the lepton in the $W$ rest-frame.
More details of these studies, as well as a discussion of systematic errors,
can be found elsewhere\cite{moriond}.
\medskip
\begin{table}[htb]
\caption{Backgrounds, efficiencies and fit results for the $B\rightarrow\pi\ell\nu$~
analysis.}
\medskip
\label{pifit}
\begin{tabular}{lcccc}
& \multicolumn{2}{c}{$\pi^-\ell\nu$} & \multicolumn{2}{c}{$\pi^0\ell\nu$} \\
& ISGW & WSB & ISGW & WSB \\ \hline
Raw Data & \multicolumn{2}{c}{30} & \multicolumn{2}{c}{15} \\
Continuum Bkg. & \multicolumn{2}{c}{$2.3\pm0.8$} &
\multicolumn{2}{c}{$1.0\pm0.5$} \\
Fake Lepton Bkg. & \multicolumn{2}{c}{$1.2\pm0.3$} &
\multicolumn{2}{c}{$0.7\pm0.2$} \\
other $u\ell\nu$ Bkg. & \multicolumn{2}{c}{0.6} & \multicolumn{2}{c}{0.2} \\
Efficiency & 2.9\% & 2.1\% & 1.9\% & 1.4\% \\
Signal Yield & $15.6\pm5.3$ & $16.3\pm5.3$ & $5.0\pm1.7$ & $5.3\pm1.7$ \\
$b\to c$ Bkg. & $9.8\pm1.1$ & $9.8\pm1.1$ & $1.8\pm0.5$ & $1.7\pm0.5$ \\
$\rho/\omega$ Bkg.&$3.8\pm1.7$ & $3.4\pm1.4$ & $1.8\pm0.8$ & $1.6\pm0.7$ \\
\end{tabular}
\end{table}
Correcting the yield for efficiency results in a preliminary measurement of
the $B\rightarrow\pi\ell\nu$~ branching ratio: ${\cal B}(B^0\rightarrow\pi^-\ell\nu)=
(1.19\pm0.41)\times 10^{-4}$ using ISGW and $(1.70\pm0.55)\times 10^{-4}$
using WSB. The errors shown are statistical only.
\smallskip
For the vector meson modes the fit results are used to calculate an
upper limit rather than a branching ratio since the non-resonant
contribution is uncertain. Using the conservative assumption that there
is no non-resonant $b\rightarrow u\ell\nu$~background present in the $B\rightarrow\rho\ell\nu$~signal
region, the 90\% confidence level upper limits for
${\cal B}(B^0\rightarrow\rho^-\ell\nu)$ is found using the ISGW (WSB)
model to be $<3.1\times 10^{-4}$ ($<4.6\times 10^{-4}$), consistent with
the upper limits previously published by CLEO\cite{cleoexclusive}.
\smallskip
One of the more interesting quantities that can be extracted from the data is
an upper limit on the {\it ratio} of branching ratios
${\cal B}(B^0\rightarrow\rho^-\ell\nu) / {\cal B}(B^0\rightarrow\pi^-\ell\nu)
< 3.4$ at the 90\% confidence level for both WSB and ISGW models. This can be
directly compared to the ratio of partial widths found in Table \ref{exclpred}.
\section{Conclusions and Future Prospects}
At present, the limiting uncertainly in $|V_{ub}/V_{cb}|$ ~is theoretical.
The statistical error of the inclusive measurements is about 10\%, and the
variation between models is at least twice as large. Even within a
single model, variation of parameters within reasonable limits can yield
significant changes\cite{hwang}. CLEO is in the process of repeating its
inclusive analysis with over than a factor of two more data, which will
further decrease the statistical error of the endpoint analysis.
\smallskip
Experiments can do more, however, to provide guidance to the theoretical
community. With sufficient statistics, measuring the $q^2$ distribution
of leptons in the inclusive endpoint region should provide useful
feedback. The model dependence can in principle be reduced, or at the
very least explored, by examining data in different regions of the
$q^2-p_\ell$ plane.
\smallskip
Semileptonic $\Lambda_b$ decays may provide another avenue to study $V_{ub}$.
A measurement of the form factors in $\Lambda_c\to \Lambda\ell\nu$ can be
related to $\Lambda_b\to p\ell\nu$ using SU(3) and HQET, and used to extract
$|V_{ub}/V_{cb}|$\cite{datta}. Several authors have discussed ways of relating the
differential spectra for $b\rightarrow u\ell\nu$~ and $b\to s\gamma$ to reduce the uncertainty
in the endpoint region due to hadronization\cite{falk,neubert,bigi,korchemsky}.
\smallskip
The statistical significance of exclusive measurements is still poor,
but will improve with experimental running. It is unlikely that exclusive
analyses will ever surpass inclusive measurements in terms of raw statistical
accuracy, however they will certainly provide a very powerful tool for testing
various theoretical models. The experimental limit on the ratio
${\cal B}(B^0\rightarrow\rho^-\ell\nu) / {\cal B}(B^0\rightarrow\pi^-\ell\nu)
< 3.4$ at 90\% confidence level is already slightly challenging for some.
Other measurements such as lepton momentum spectra, $q^2$ distribution,
and vector meson polarization will provide further theoretical tests
\cite{faustov2}. Exclusive measurements at low $q^2$ may also prove
valuable\cite{akhoury}. It has recently been shown that measurements
of $B\rightarrow\rho\ell\nu$ ~and $B\to K^*\ell\overline{\ell}$ can be used to extract
$V_{ub}$~ using SU(3) and HQET\cite{sanda}.
\smallskip
The theoretical problem of determining the form factors needed to calculate
$b\rightarrow u\ell\nu$~is also being approached with lattice gauge calculations, and several
groups are making progress\cite{martinelli,ukqcd,ape,lanl}.
\medskip
In conclusion, good theoretical and experimental progress is being made
in the quest to determine $|V_{ub}/V_{cb}|$, and the next few years should yield
significant advances in both.
\bigskip
I would like to thank Jeff Nelson, Ron Poling, Lawrence Gibbons and Ed
Thorndike for information about the latest CLEO analyses. I would also
like to thank John Sloan and Aida El-Khadra for insight regarding lattice
calculations, and finally I would like to acknowledge Tom Browder and
Klaus Honscheid whose review paper ``B Mesons'' provided valuable
information and references.
\smallskip
\noindent I gratefully acknowledge the support of the Department
of Energy and the A. P. Sloan Foundation.
|
\section{Introduction}
\smallskip
\par
Cosmic strings might be responsible for the formation of large-scale structure
which is observed in the Universe today. They are topological
defects which form in a phase transition in the early universe.
Topological defect models of structure formation are an alternative to
inflationary models. In inflationary models one assumes the existence of a
scalar field which drives a phase of very rapid expansion in the early
universe, and quantum fluctuations produced during this phase turn into
classical density perturbations which grow by gravitational instability
into present-day structures. In topological defect models the defects,
which are present at all times including today (and could possibly be
seen directly some day), act as seeds around which structures form.
It is important to study theories from both classes of models in order
to see which predictions can distinguish between different models, and which
predictions are too generic, arising in models based on widely different
assumptions.
Among the topological defect models, cosmic strings were the first to
be investigated in more detail (Brandenberger 1991), recently textures have
also been
subjected to closer scrutiny (Turok 1991). The cosmic string model looks
promising
so far, since normalizations of the parameters occuring in the model
obtained from observation agree with each other (Bennet, Bouchet \& Stebbins
1992, Perivolaropoulos 1993a).
These obervations include galaxy redshift surveys and measurements of
the temperature anisotropies in the cosmic microwave background radiation
(CMB). Moreover, cosmic strings naturally produce filaments and planar
structures in the matter distribution (Silk \& Vilenkin 1984, Vachaspati 1986,
Stebbins et al. 1987), in encouraging agreement with recent
galaxy redshift surveys. Another nice feature of the cosmic string model
is that it works well in the context of hot dark matter (Brandenberger 1987),
for which there are
candidates known to exist (the $\mu$ and $\tau$ neutrinos), although they may
be
massless. In contrast, none of the candidates for cold dark matter
required in inflationary models has been detected.
Topological defect models have not been as closely investigated as
inflationary models, and more work is necessary to quantify the
predictions of the cosmic string model. It is especially important to
find features accessible to observations in which cosmic strings differ
from the class of inflationary models, and statistics sensitive to
these features. Recently, the probability distribution of peculiar
velocities has been put forward as a statistic which might be capable of
discriminating
between cosmic strings and inflation (Kofman et al. 1994). Here we show that
according
to the calculations in a simple string toy model,
this statistic cannot be used to differentiate between cosmic strings and
inflation.
The IRAS 1.9 Jy redshift survey was analyzed (Strauss et al. 1990 \& 1992,
Yahil 1991) to obtain a uniform galaxy-density map. The peculiar velocities
were reconstructed from it via a self-consistent iterative scheme assuming
linear biasing between the density fluctuations of galaxies and mass (Nusser et
al. 1991). Using these results, Kofman et al. (1994) determined the probability
distribution of a single component of peculiar velocities of regions smoothed
over several $h^{-1}$ Mpc, with the result that it is consistent with the
underlying distribution $p(v_{x})$ being a normal distribution. Their result for the
Gaussianity of the velocity distribution is still tentative because of the
limited volume
sampled, and because velocities were not measured directly but deduced from a
redshift survey.
Inflationary models predict a normal distribution for $p(v_{x})$, whereas some
deviation from gaussianity
is expected in the cosmic string model since individual strings impart
coherent velocity perturbations over extended regions, as we will see in the
following section. The question is, however, just how large this deviation
from gaussianity is and whether it is big enough to be detected by current
observations. For a certain region receives velocity perturbations from
many strings, and by the central limit theorem many nongaussian
perturbations can add up to a gaussian signal.
Scherrer (1992) has shown that for seed models the velocity field can be
very nearly Gaussian even if the density field is nongaussian. In a model
where randomly distributed point masses, all of the same mass, accrete matter
gravitationally in a universe dominated by hot dark matter, he found that a
very low seed density of less than $10^{-2} {\rm{Mpc}}^{-3}$ is required in
order
for $p(v_{x})$ to be nongaussian.
The main aim of this paper is to calculate $p(v_{x})$ within the cosmic
string model in order to quantify its departure from gaussianity, and
to find out if the IRAS results are also in agreement with this stringy
probability distribution $p(v_{x})$.
Previously, 3 dimensional rms velocities have been obtained analytically within
the cosmic string model (Vachaspati 1992, Perivolaropoulos \& Vachaspati 1994),
and numerical simulations have been
performed (Vachaspati 1992, Hara, M\"{a}h\"{o}nen \& Miyoshi 1993) to obtain
the probability distribution of peculiar velocities, and
it was found that the distribution is quite gaussian.
Here we present the first analytical calculation of $p(v_{x})$ , within a model for
the string network which has been previously employed to make predictions
about temperature anisotropies in the CMB (Perivolaropoulos 1993a \& 1993b,
Moessner, Perivolaropoulos \& Brandenberger 1994) and the magnitude of
peculiar velocities (Vachaspati 1992, Perivolaropoulos \& Vachaspati 1994). The
dependence of the deviation from
gaussianity on one of the parameters of the model, namely the number
$\nu$ of strings per Hubble volume in the strings' scaling solution,
is shown explicitly.
Our conclusion is that on scales of several $h^{-1}$ Mpc, $p(v_{x})$ from
strings deviates only slightly from a normal distribution. On the smallest
scales and for the smallest number of strings per Hubble volume, the largest
deviation from
gaussianity is expected. Performing a $\chi^{2}$ -test of the stringy $p(v_{x})$
for the smallest scale of $6h^{-1}$ Mpc given in Kofman et al. (1994), with
imaginary data binned in the same way as in that paper, but drawn
from an underlying gaussian distribution (worst case),
yields the result that the stringy probability distribution is in
agreement with these data. This agreement holds for all values of $\nu$ , the
number of long strings per Hubble volume in the strings' scaling solution,
including $\nu = 1$.
This shows that the hope expressed in Kofman et al. (1994) that the
scenario for the formation of large scale structure, where widely
separated strings accrete matter in wakes behind them, can be ruled out
using the statistic $p(v_{x})$ is not realized.
In the next section we will describe the cosmic string model and the
mechanism for the production of velocity perturbations. In the third
section we will describe the analytic model for the string network and
calculate the moment generating function for a single component of the
peculiar velocities averaged over several $h^{-1}$ Mpc. Then we will obtain
the probability distribution $p(v_{x})$ from this moment generating function,
and compare it with observations.
\section{Cosmic Strings and Structure Formation}
Cosmic strings are linear topological defects formed at a phase transition
in the very early universe (Vilenkin 1985).
Those originating in the symmetry braking of a Grand Unified Theory possess an
enormous mass per unit length $\mu$ ($G \mu \approx 10^{-6}$ , where $G$ is
Newton's constant), and they
can be responsible for the structures observed today. Strings can have no ends,
so they are formed as
either infinitely long strings or closed loops. After formation the
network of strings quickly evolves towards a scaling solution where the
energy density in long strings remains a constant fraction of the total
background energy density. This is achieved by intercommutations and
self-intersections
of the strings leading to the production of small loops, which then
decay by emitting gravitational radiation. In this way, some of the energy
input into the string network coming from the stretching of the strings due to
the expansion of the universe is transferred to the background. Long strings
are straight over distances of the order of the horizon, so
that the scaling solution can be pictured as having a fixed number $\nu$
of long strings per Hubble volume at any given time.
Initially loops were thought to make the dominant contribution to
structure formation. At distances much larger than their radius they act
as point masses and accrete surrounding matter (Turok \& Brandenberger 1986,
Stebbins 1986, Sato 1986). Improved cosmic string evolution simulations (Bennet
\& Bouchet 1988, Albrecht \& Turok 1989, Allen \& Shellard 1990) showed that
more of the energy density is in long strings, and they are therefore more
important, accreting matter in the form of wakes
behind them as they move through space (Brandenberger, Perivolaropoulos \&
Stebbins 1990, Perivolaropoulos, Brandenberger \& Stebbins 1990).
Spacetime around a long straight cosmic string can be pictured as locally flat,
but with a deficit angle of $8\pi G \mu$ (Vilenkin 1985). Therefore a string
moving relativistically with velocity $v_{s}$ imparts velocity perturbations
to surrounding matter towards the
plane swept out by the string.
If small-scale structure is present on the string, there is in addition a
Newtonian force towards the string.
The magnitude of this velocity perturbation is
given by (Vachaspati \& Vilenkin 1991, Vollick 1992)
\begin{equation}
u=4 \pi G \mu \gamma_{s} v_{s} f ,
\;\;\;\;f=1+\frac{1-\rm{T}/\mu}{2(\gamma_{s}v_{s})^{2}}
\end{equation}
In the absence of small-scale structure on the string,the tension T of the
string is equal to its mass per unit length $\mu$, and $f=1$. If small-scale
structure is present, $\mu$ denotes the mass per unit length obtained after
averaging over the small scale structure, $\rm{T} \ne \mu$, and $f > 1$.
We consider the perturbations caused by strings after $t_{eq}$, the time of
equal matter and radiation, in a universe filled with hot or cold dark matter.
By the present time, the initial velocity perturbation imparted to the dark
matter at time $t_{i}$ has grown to (Brandenberger 1987, Stebbins 1987, Hara \&
Miyoshi 1990)
\begin{equation}
u_{i}\approx 0.4u\sqrt{z(t_{i})}
\end{equation}
Due to compensation (Traschen, Turok \& Brandenberger 1986, Veeraraghavan \&
Stebbins 1992), the deficit angle of strings which are straight over a horizon
distance, extends out only to a distance of one Hubble radius $H^{-1}$ from the
string, so that matter which is farther away does not receive any velocity
perturbations.
The velocity perturbation given in eq.(1) is independent of distance from
the string (up to the Hubble radius), so that cosmic strings impart
coherent perturbations over regions of the size of half a Hubble volume.
\section{Moment Generating Function}
The moment generating function (mgf) $M_{X}(t)$ of a random variable $X$ is
defined by
\begin{equation}
M_{X}(t)=\langle \exp^{tX} \rangle
\end{equation}
where the brackets denote the ensemble average, and it contains complete
information about $X$. In the following, $X = v_{x}$ denotes the random
variable for the component of the peculiar velocities smoothed over a region
$\cal V$ of comoving radius $R$ in a fixed direction $\hat{e}_{x}$ .
We will calculate the mgf of $X$ in order to obtain the moments (and
cumulants) and probability distribution $p(v_{x})$ of $X$ from it.
The moments of $X$, $\mu_{j} = < X^{j} > $ are given by
\begin{equation}
\mu_{j}=\left(\frac{d^{j}}{dt^{j}} \right)_{t=0} M_{X}(t)
\end{equation}
and the cumulants $c_{j}$ are defined by
\begin{equation}
c_{j}=\left(\frac{d^{j}}{dt^{j}} \right)_{t=0}\ln(M_{X}(t))
\end{equation}
The probability distribution $p(v_{x})$ can be expanded in an asymptotic series
called Edgeworth series (Scherrer \& Bertschinger 1991, Stuart \& Ord 1987) in
terms of the quantities
\begin{equation}
\lambda_{j} = c_{j}/c_{2}^{j/2}
\end{equation}
and Hermite polynomials $H_{n}(x)$ defined by
\begin{equation}
H_{n}(x) = (-1)^{n} e^{x^{2}/2} \frac{d^{n}}{dx^{n}} e^{-x^{2}/2}
\end{equation}
For distributions with vanishing odd moments as in our case, the expansion is
\begin{equation}
p(\delta) = \frac{e^{-\delta^{2}/2}}{\sqrt{2} \pi} [1+ \frac{\lambda_{4}}{24}
H_{4}(\delta) + \frac{\lambda_{6}}{720}H_{6}(\delta) +\frac{\lambda_{8} +
35 \lambda_{4}^{2}}{40320} H_{8}(\delta) + \cdots ]
\end{equation}
where $\delta =v_{x}/ \sigma$ , and $\sigma$ is the standard deviation of $X$.
Since it is an asymptotic expansion, the remainder is of the order of the last
term included (Erdelyi 1956).
The mgf has an important property which makes it useful for calculations.
For independent random variables $X$ and $Y$, the mgf of the sum is the
product of the individual moment generating functions
\begin{equation}
M_{X+Y}(t) = M_{X}(t) M_{Y}(t)
\end{equation}
The following calculations are carried out within an analytical model for the
string network
which has previously been used to obtain the temperature anisotropies in the
CMB and the magnitude of peculiar velocities from strings (see references given
in the introduction).
According to the scaling solution for cosmic strings, there is a fixed number
of long strings present per Hubble volume at any given time. After about one
expansion time of the universe (Hubble time), $t \rightarrow 2t$, these strings
will typically have self-intersected or intercommuted, so that the resulting
strings are uncorrelated with the ones at the previous Hubble time. We will
assume that during one expansion time $\nu$ long strings move across the Hubble
volume.
Each string is assumed to be straight over one horizon distance, and the effect
of all strings is taken to be the superposition of the effects of the
individual strings.
The fact that small-scale structure varies along strings is neglected, so that
we might somewhat underestimate the degree of non-Gaussianness of the velocity
distribution.
We also assume that the strings' positions, velocities and orientations at
each Hubble time are random and uncorrelated, although this is not strictly
true, since - to mention one reason - the string network has the form of a
self-avoiding random walk.
According to the two previous assumptions, the random variable for the total
peculiar velocity, $X$, is the sum of independent random variables for the
velocity perturbations from the individual strings.
By eq.(9) we can therefore reduce the calculation of the the mgf for $X$ to
that of the mgf
for the contribution of only one string, and take the products afterwards.
In fact
we know that there are $\nu$ strings per Hubble volume on average. The products
in eq.(9) simplify if we take a Poisson distribution for the number of strings
per Hubble volume instead of assuming the presence of exactly $\nu$ of them
(Scherrer \& Bertschinger 1993). We can picture this as having a
reservoir of $n$ strings, each with a probability $\nu/n$ of being present in
a particular Hubble volume, in the limit that $n \rightarrow \infty$. So if
$M_{Y_{i}}(t)$ denotes the mgf for a single component of peculiar velocities
of the region $\cal V$ due to one string present at time $t_{i}$ in the
region's Hubble volume, then
\begin{eqnarray}
M_{X_{i}}(t) & = & \lim_{n \rightarrow \infty} [ \frac{\nu_{i}}{n} \langle
\exp^{tY_{i}} \rangle + ( 1-\frac{\nu_{i}}{n} ) ]^{n} \nonumber \\
& = & \exp{[\nu_{i}(M_{Y_{i}}(t) - 1)]}
\end{eqnarray}
where $X_{i}$ is the random variable for the contribution of all strings at
Hubble time $t_{i}$ to the velocity perturbation of $\cal V$, and $\nu_{i}$
denotes the average number of strings having an effect on $\cal V$ at time
$t_{i}$, i.e. those strings which are within a distance of one Hubble radius
of $\cal V$ at time $t_{i}$ (see eq.(18)).
Since $X = \sum_{i=1}^{N} X_{i}$ ,
where $N$ is the number of expansion times since $t_{eq}$
\begin{equation}
N=\log_{2} \frac{t_{0}}{t_{eq}}
\end{equation}
and the $X_{i}$ are assumed to be independent,
\begin{eqnarray}
M_{X}(t) & = & \prod_{i=1}^{N} M_{X_{i}}(t) \nonumber \\
& = & \exp{[ \sum_{i=1}^{N} \nu_{i} ( M_{Y_{i}}(t) - 1)]}
\end{eqnarray}
We will now calculate the mgf for $Y_{i}$, the random variable for the
component
in the fixed direction ${\hat{e}}_{x}$ of the peculiar velocities of a region
$\cal{V}$ of comoving radius $R$ (the smoothing radius) due to one string
affecting the region at time $t_{i}$. We have to perform the ensemble average
over positions , orientations ${\hat{e}}_{s}$ and directions of velocity
${\hat{v}}_{s}$ of the string.
For a long straight string only transverse velocities are observable, and we
assumed ${\hat{e}}_{s}$ and ${\hat{v}}_{s}$ to be random unit vectors.
Therefore
the unit normal $\hat{e}={\hat{e}}_{s} \times {\hat{v}}_{s}$ of the plane swept
out by the string, i.e. the direction in which matter receives velocity
perturbations, is itself a random unit vector. Consequently, the projection
$s = \hat{e} \cdot {\hat{e}}_{x}$ is uniformly distributed over the interval
$[-1,1]$, and the magnitude of the velocity perturbation from one string has to
be multiplied by $s$ to get the component in direction ${\hat{e}}_{x}$.
During one expansion time a string sweeps out a plane towards which matter
within a distance of one Hubble radius receives velocity perturbations, which
have grown to $u_{i}$ (see eq.(2)) by today. The possible values $y_{i}$ for
the random variable $Y_{i}$, the projection of the peculiar velocity of
$\cal{V}$
in direction $\hat{e_{x}}$ due to one string at time $t_{i}$, are a function of
the perpendicular distance $r$ of the centre of $\cal{V}$ to this plane:
\[ y_{i}(r) = \;s u_{i} r/R_{i} \;\;\;\; {\rm{for}} \;\;\; 0 < r < R_{i}\]
\[ y_{i}(r) = \;s u_{i} \;\;\;\; {\rm{for}} \;\;\; R_{i} < r < H^{-1}_{i} - R_{i} \]
\begin{equation}
y_{i}(r) = \;s u_{i} \frac{H^{-1}_{i}-(r-R_{i})}{2 R_{i}} \;\; {\rm{for}} \;\;H^{-1}_{i}-R_{i}<
r<H^{-1}_{i}+R_{i}
\end{equation}
$R_{i}$ is the physical size of the comoving radius $R$ at time $t_{i}$.
Strings within a distance of $H^{-1}(t_{i}) \equiv H^{-1}_{i}$
of the region $\cal V$ can affect it. We distribute the centres $C$ of these
planes randomly within a sphere of radius $r_{max}^{i} = H^{-1}_{i}$ around the
centre of $\cal{V}$. So the probability $p(c)$ for $C$ to be a distance $c$
from the
centre of $\cal{V}$ is
\begin{equation}
p(c) = \frac{3c^{2}}{(r_{max}^{i})^{3}}
\end{equation}
Since the normal of this plane has random direction, $r$ can be smaller or
equal to $c$, with probability
\begin{equation}
p(r;c) \approx 2 \frac{r}{c^{2}}
\end{equation}
Integrating over all $c$ gives the probability for the plane to be a distance r
from the centre of $\cal{V}$ as
\begin{equation}
p(r)= \int_{r}^{r_{max}^{i}} dc \; p(r;c) \; p(c) =
\frac{6r}{(r_{max}^{i})^{3}}
(r_{max}^{i} -r)
\end{equation}
The ensemble average thus becomes an integral over $r$ and $s$
\begin{eqnarray}
M_{Y_{i}}(t) & = & \langle \exp^{t Y_{i}} \rangle \nonumber \\
& = & \int_{0}^{r_{max}^{i}} dr\;p(r) \int_{-1}^{1} ds \;
\frac{1}{2} \exp{(t y_{i}(r))}
\end{eqnarray}
These integrals can be done, and then eq.(12) can be used to obtain the mgf for
$X$ which includes the effect of all strings, with the number of strings
affecting $\cal V$ at time $t_{i}$ being given by
\begin{equation}
\nu_{i} = \nu(r_{max}^{i})^{3}/(H^{-1}_{i})^{3}
\end{equation}
There is one problem with the above. The formulas for $y_{i}(r)$ in equations
(13) are only true if the projection of $\cal{V}$ onto the plane swept out by
the string along its normal falls
completely into that plane, and is not (partly) outside of it. But the latter
can happen for large $c$ for some orientations of the plane, since one side of
the plane, in the direction of the string's motion, has a length $l{_i} =
H^{-1}_{i}$ or smaller, so that the distance of $C$ to the edge of the plane can be
smaller
than $H^{-1}_{i} /2$. For $l{_i} = H^{-1}_{i}$ one can show that less than half of the
strings miss $\cal{V}$ and give no perturbations to it, so that we can estimate
this effect by replacing $\nu$ by $\nu_{\rm{eff}} = 0.5 \nu$ .
If the strings are moving slowly, so that $l_{i}$ is even smaller, and not at
about the speed of
light, our model is not really applicable because the formula for the
imparted initial velocity perturbations would change.
Actually a string can affect $\cal{V}$ if $c \leq R_{i} + H^{-1}_{i}$ . But for $c$
larger than $H^{-1}_{i}$ we encounter the problem mentioned in the previous
paragraph, so that we overestimate the perturbations by using $r_{max}^{i} =
H^{-1}_{i} + R_{i}$.
Therefore we calculate the cumulants for both $r_{max}^{i} = H^{-1}_{i}$ and
$r_{max}^{i} = H^{-1}_{i} + R_{i}$ and take their average, and the model is more
accurate for smaller $R$. But since at scales below about $5h^{-1}$ Mpc
nonlinear effects become important, and we are only considering linear
perturbations, we must also keep above that scale.
\section{Probability Distribution and Comparison with Observations}
First we want to look at the shape of the probability distribution for $v_{x}$.
The nongaussianness
is largest on the smallest scales since smoothing makes things more
gaussian, and larger regions are affected by more strings. Therefore
we are going to compare $p(v_{x})$ from strings with the results from IRAS
at the smallest scale of $R = 6 h^{-1}$ Mpc considered in Kofman et al. (1994).
Also, the derivation of $M_{X}(t)$ is valid for scales of $R \leq l_{eq}$,
where $l_{eq} = 13 h^{-2}$ Mpc is the comoving size of the Hubble radius at the
time of equal matter and radiation.
The values $\Omega = 1$, $h = 1/2$ and $z_{eq} = 2.3 \cdot 10^{4} \Omega h^{2}$
are used.
Using a symbolic manipulation program (O'Dell 1991), the cumulants are
obtained from $M_{X}(t)$ according to eq.(5), giving the following values for
the $\lambda_{j}$ (eq.(6)) needed in the expansion of the probability
distribution (eq.(8)) for $R=6 h^{-1}$ Mpc :
\begin{eqnarray}
\lambda_{4} & = & 0.34/\nu \nonumber \\
\lambda_{6} & = & 0.20/\nu^{2} \nonumber \\
\lambda_{8} & = & 0.15/\nu^{3}
\end{eqnarray}
These values, as well as the standard deviations quoted below, are the averages
of the two cases $r_{max}^{i}=H^{-1}_{i}$ and $r_{max}^{i}=H^{-1}_{i} + R_{i}$ .
For two values of $\nu$, $p(v_{x}/\sigma)$ is plotted in Figure~1, up to
the term involving $H_{8}(\delta)$ in the expansion
(eq.(8)). For the higher value of $\nu = 10$ strings
per Hubble volume, the distribution is practically indistinguishable from
a gaussian one. For $\nu = 1$ there is a slight deviation. We want to
see if this deviation is significant by performing a $\chi^{2}$ -test with the
data given in Kofman et al. (1994), which has been grouped into bins of size
$v_{x}/\sigma = 0.25$. The data points fall practically on
a gaussian curve, so instead of taking the exact values from the data
we calculate the absolute frequencies $m_{j}$ in the $j$ bins expected if the
underlying
distribution were gaussian. The IRAS 1.9Jy survey maps out a sphere of radius
$80 h^{-1} \rm{Mpc}$ , so that there are $(80/6)^{3}$ independent smoothing
regions of radius $6h^{-1} \rm{Mpc}$. Let $n_{j}$ be the corresponding absolute
frequencies expected from the stringy distribution. Using 20 inner bins, we
find $\chi^{2} = 4.63$ for $\nu = 1$, where
\begin{equation}
\chi^{2} = \sum_{j=1}^{20} \frac{{(m_{j}-n_{j})}^2}{n_{j}}
\end{equation}
, which is much smaller than the $95 \%$ confidence upper limit of $38.58$ for
$19$ degrees of freedom, so that
the data is in agreement with the probability distribution from strings.
If we replace $\nu$ by $\nu_{\rm{eff}} = \nu / 2$ to take into account that
the side of the plane swept out by the string in the direction of its motion
is only half the diameter of the Hubble volume, then $\chi^{2} = 22.5$ for
$\nu = 1$.
Next we want to compare the magnitudes of velocities. The standard deviation
of the single velocity components is calculated to be
\begin{equation}
\sigma = 1.04 \nu^{1/2} \tilde{u} \;\;\; {\rm{for}} \;\;R=6 h^{-1} \rm{Mpc}
\end{equation}
\begin{equation}
\sigma = 0.99 \nu^{1/2} \tilde{u} \;\;\; {\rm{for}} \;\;\;R=12 h^{-1}
\rm{Mpc}
\end{equation}
where $\tilde{u} = 0.4 z_{eq}^{1/2} u$ , and $u$ is defined in eq.(1).
We compare these values with results from Peacock and Dodds (1994),
who used the power spectra of various observations to calculate the
3 dimensional rms velocities $v_{rms}$ of regions of radius $R$. For a
gaussian random variable with three independent gaussian variables as
components, the standard deviation of a single component is given by
$\sigma = v_{rms}/ \sqrt{3}$. Using this relation, the values given in Peacock
and Dodds (1994) are
\begin{equation}
\sigma = (381 \pm 156) \; {\rm{km/s}} \;\;\; {\rm{for}} \;\;R=6 h^{-1}
\rm{Mpc}
\end{equation}
\begin{equation}
\sigma = (337 \pm 138) \; {\rm{km/s}} \;\;\; {\rm{for}} \;\;\;R=12 h^{-1}
\rm{Mpc}
\end{equation}
where we have taken the fractional error of $1/\sqrt{6}$ quoted for
the actual measurement of $v_{rms}$ at a scale of $5 h^{-1}$ Mpc.
Comparison of $\sigma$ in the string model with the values obtained from
observations at these two scales, gives as an average value for $\alpha f$
\begin{equation}
\overline{\alpha f} = 3.2 \pm 0.9
\end{equation}
where $\alpha$ is the combination of parameters
\begin{equation}
\alpha= \sqrt{\nu} \frac{G\mu}{10^{-6}} \frac{\gamma_{s} v_{s}}{c}
\end{equation}
$\alpha$ can be constrained from the rms value of the temperature fluctuations
in the cosmic microwave background measured by COBE
to be (Perivolaropoulos 1993a)
\begin{equation}
\alpha = 1.0 \pm 0.2
\end{equation}
Using this value of $\alpha$ , we find
\begin{equation}
\bar{f} = 3.2 \pm 1.1
\end{equation}
This indicates that there must be some small scale structure on the strings
(see eq.(1)) in order to obtain the right magnitude of peculiar velocities and
consistency with CMB observations. This is also in agreement with recent
simulations (Bennett \& Bouchet 1988, Albrecht \& Turok 1989, Allen \& Shellard
1990), which show the presence of small-scale structure on cosmic strings. Our
analysis has been done in the string wake model, whereas strings with lots of
small scale structure accrete matter rather in the form of filaments. Therefore
the precise value of $f$ is not to be taken too seriously.
Numerical simulations of peculiar velocities from long strings without small
scale structure have been performed in a similar framework (Hara,
M\"{a}h\"{o}nen \& Miyoshi 1993), where
$\nu'$ strings are assumed to move across the horizon at every e-fold expansion
of the universe. The authors found that
\begin{equation}
\frac{G \mu}{10^{-6}} \frac{\gamma_{s}v_{s}}{c} \sqrt{\nu'} =( 4 \pm 1)
\end{equation}
yields good agreement with observations.
The number of strings at every two-fold expansion used in our analysis is
related to $\nu'$ by $\nu = \nu'\ln{2}$ . Therefore $\alpha f = 3.3 \pm 0.8 $
from these simulations,
which agrees quite well with our result of $\overline{\alpha f} = 3.2 \pm 0.9 $
{}.
\section{Discussion}
We have shown that the probability distribution of a single component of
peculiar velocities in the cosmic string wake model is very close to a normal
distribution on scales of several $h^{-1}$ Mpc, as suggested by a general
argument for seed models (Scherrer 1992), and in agreement with observations.
A comparison of the measured magnitude of peculiar velocities with that
expected from strings, using the normalization of string parameters from the
COBE quadrupole, suggested that strings have some small-scale structure.
Nongaussian features are more apparent in the velocity differences than in the
velocities themselves (Catelan \& Scherrer 1994), and it would be interesting
to calculate the
probability ditribution of these velocity differences within the cosmic string
model of structure formation.
\section{Acknowledgements}
I would like to thank Robert Brandenberger and Leandros Perivolaropoulos
for suggestions and helpful discussions. This work was supported in part by the
US Department of Energy under Grant DE-FG0291ER40688.
\newpage
{\bf References}
Albrecht A., Turok N., 1989, Phys. Rev. D 40 , 973.
Allen B., Shellard E.P.S, 1990, Phys. Rev. Lett. 64, 119.
Bennett D., Bouchet F., 1988, Phys. Rev. Lett. 60, 257.
Bennett D., Bouchet F., Stebbins A., 1992, ApJ (Lett.) 399, L5.
Brandenberger R., 1991, Phys. Scripta T36, 114.
Brandenberger R.,Perivolaropoulos L., Stebbins A., 1990, Int. J. Mod. Phys. A5,
1633.
Brandenberger R. et al., 1987, Phys. Rev. D36, 335.
Catelan P., Scherrer R., 1994, "Velocity Differences as a Probe of Non-Gaussian
Density \hspace*{.5in} Fields" (preprint SISSA Ref. 37/94/A).
Erdelyi A., 1956, "Asymptotic Expansions", (Dover).
Hara T., M\"{a}h\"{o}nen P., Miyoshi S., 1993, ApJ 415, 445.
Hara T., Miyoshi S., 1990, Prog. Theor. Phys. 81, 1187.
Kofman L. et al., 1994, ApJ 420, 44.
Moessner R., Perivolaropoulos L., Brandenberger R., 1994, ApJ 425, 365.
Nusser A. et al., 1991, ApJ 379, 6.
O'Dell J., 1991, "ALJABR", Fort Pond Research.
Peacock J., Dodds S., 1994, MNRAS 267, 1020 .
Perivolaropoulos L., 1993a, Phys. Lett. B 298, 305.
Perivolaropoulos L., 1993b, Phys. Rev. D 48, 1530.
Perivolaropoulos L., Brandenberger R., Stebbins A., 1990, Phys. Rev. D 41,
1764.
Perivolaropoulos L., Vachaspati T., 1994, ApJ 423, L77.
Sato H., 1986, Prog. Theor. Phys. 75, 1342.
Scherrer R., 1992, ApJ 390, 330.
Scherrer R., Bertschinger E., 1991, ApJ 381, 349.
Silk J., Vilenkin A., 1984, Phys. Rev. Lett. 53, 1700.
Stebbins A., 1986, ApJ (Lett.) 303, L21.
Stebbins A. et al., 1987, ApJ 322, 1.
Strauss M. et al., 1990, ApJ 361, 49.
Strauss M. et al., 1992, ApJ 385, 421.
Stuart A., Ord J., 1987, "Kendall's Advanced Theory of Statistics",
Vol.1 (London: \hspace*{.5in}Charles Griffin).
Traschen J., Turok N., Brandenberger R., 1986, Phys. Rev. D 34, 919.
Turok N.,1991, Phys. Scripta T36, 135.
Turok N., Brandenberger R., 1986, Phys. Rev. D 33, 2175.
Vachaspati T., 1986, Phys. Rev. Lett. 57, 1655.
Vachaspati T., 1992, Phys. Lett. B, 282, 305.
Vachaspati T., Vilenkin A., 1991, Phys. Rev. Lett. 67, 1057.
Veeraraghavan S., Stebbins A., 1992, ApJ Lett. 395, L55.
Vilenkin A., 1985, Phys. Rep. 121, 263.
Vollick D., 1992, Phys. Rev. D 45, 1884.
Yahil A. et al., 1991, ApJ 372, 380.
\newpage
\begin{center}
\bf Figure Captions
\end{center}
Figure 1: Stringy probability distribution $p(\delta= v_{x}/ \sigma)$ of a
single velocity component smoothed over regions of radius $R=6 h^{-1}$ Mpc for
for $\nu = 1$ (solid line) and
$\nu=10$ strings per Hubble volume (dotted line), compared with a normal
distribution (dashed line).
\end{document}
|
\section*{References} \begin{list}{}{
\setlength{\parsep}{0pt} \setlength{\itemsep}{0pt}
\setlength{\leftmargin}{0pt} } }{ \end{list} }
\title{Lognormal Properties of SGR 1806-20 and Implications for
Other SGR Sources}
\author{K. J. Hurley, B. McBreen, M. Delaney and A. Britton}
\date{astro-ph/9508074: presented at 29 ESLAB Symposium, April 1995}
\begin{document}
\maketitle
\begin{abstract}
The time interval between successive bursts from SGR 1806-20 and the
intensity of these bursts are both consistent with lognormal
distributions. Monte Carlo simulations of lognormal burst models
with a range of distribution parameters have been investigated. The
main conclusions are that while most sources like SGR 1806-20 should
be detected in a time interval of 25 years, sources with means about
100 times longer have a probability of about 5\% of being detected
in the same interval. A new breed of experiments that operate for
long periods are required to search for sources with mean recurrence
intervals much longer than SGR 1806-20.
\end{abstract}
\section{Introduction}
The lognormal properties of the soft repeater SGR1806-20 have been
previously reported by Hurley, K.J. \etal\ (1994). In particular, both
the time interval between repeater events and the luminosity function
of the source were fit with lognormal distributions (see Aitchison and
Brown, 1957, for a comprehensive introduction to lognormal
statistics). This analysis used the data-base of 111 events detected
by the International Cometary Explorer (ICE) mission, as reported by
Laros \etal\ (1987).
While the present number of events observed from the other two sources
(Norris \etal, 1991, Kouveliotou \etal, 1993) does not allow any
detailed analysis, the intervals between successive events of SGR
0526-66 (Golenetski\v{\i} \etal, 1987) is also suggestive of lognormal
behaviour. Continued observations by BATSE of these sources may reveal
lognormal properties for one or both of the remaining two repeaters if
either passes into a phase of activity similar to the behaviour of
SGR1806-20 during 1983.
The relationship between the number of active (i.e observable) sources
and the true number of SGRs in the galaxy is one which is the subject
of some debate (see discussions in Kouveliotou \etal, 1992,
Kouveliotou \etal, 1994 and Hurley, K. \etal, 1994). If the time
interval between SGR events proves to be lognormal then there may be
long quiescent periods where the source could be undetectable, leading
to an underestimate of the population.
\section{Simulations}
In order to investigate the behaviour of sources with much longer mean
recurrence times we generated Monte Carlo simulations with a variety
of distribution parameters. The Monte Carlo simulations were performed
using the random normal generator with Matlab 4.0 for Windows, which
is based on a random number generator algorithm given by Park and
Miller (1988) with the transformation to the standard normal variate
given by Forsythe, Malcolm and Moler (1977). The normal variates were
then transformed to lognormal variates using the relationship
$Y=e^{\sigma X+\mu}$ where $Y$ is lognormally distributed (with
parameters $\mu$\/ and \sig ) and $(\sigma \! X+\mu)$ is normally
distributed with mean $\mu$\/ and variance \sig.
\begin{figure}[thp]
\epsfxsize=\textwidth \epsfbox{mc_1806.eps}
\caption{Four separate Monte Carlo simulations of SGR1806-20 activity over a
hundred year period, generated from a lognormal distribution of
recurrence intervals with the same parameters $\mbox{$\mu_0$}=9.64$ and
$\mbox{$\sigma_0$}} \newcommand{\etal}{{\it et al.}=3.44$. The long gaps in activity of the source are the
contributions from the tail of this highly skewed distribution.}
\label{sgr-sim}
\end{figure}
The parameters of the lognormal density function which were fit to the
distribution of recurrence intervals for SGR1806-20 were $\mbox{$\mu_0$}=9.64$,
$\mbox{$\sigma_0$}} \newcommand{\etal}{{\it et al.}=3.44$ (Hurley, K.J. \etal, 1994). Initially we generated 100
year long simulations of SGR1806-20 (Fig.~\ref{sgr-sim}) using these
parameters, to check the algorithm. The samples produced were tested
for compatibility with a lognormal population using a $\chi^2$ test
(Sachs 1986) and were compatible at the 99\% confidence level,
indicating that the Monte Carlo simulator was functioning correctly.
Two further simulations were then performed to investigate how the
source behaviour varied as \mbox{$\mu$}} \newcommand{\sig}{\mbox{$\sigma$} \/ and \sig \/ varied. The results
(illustrated in Fig.~\ref{probplot}) are discussed below.
\section{Discussion}
\begin{figure}[thp]
\epsfxsize=\textwidth \epsfbox{probplot.eps}
\caption{Percentage of time with one or more events per 25 year interval: (a)
as a function of \mbox{$\mu$}} \newcommand{\sig}{\mbox{$\sigma$} ( with $\sigma = 3.0$ fixed) and (b) as a
function of \sig (with $\mu=9.5$ fixed). Errorbars are 1 standard
deviation values on the mean of 20 runs.}
\label{probplot}
\end{figure}
Presented in Figure~\ref{probplot} are the probabilities for source
activity in a 25 year period as calculated from the results of the two
simulations described above. Figure~\ref{probplot}(a) shows that as
the parameter \mbox{$\mu$}} \newcommand{\sig}{\mbox{$\sigma$} increases the chance of one or more event in 25
years falls from $\approx 80\%$ at $\mu=9.5 \approx \mu_0$ (geometric
mean of 0.15 days) to less than 5\% at $\mu=18$ (geometric mean of
$\sim \! 500$ days).
The chance of one or more events in 25 years for a source like
SGR1806-20 (that is with $\mu\approx\mu_0 = 9.64$ and
\sig$\approx\sigma_0 = 3.44$) as predicted by Figure~\ref{probplot}
indicates that the majority of this type of source should be observed
in $\sim$ 25 years. For $\mu\!\gg\!\mu_0$ experiments which operate
for a long time must be devised and maintain a continuous search over
the whole sky for longer periods than any spaceborne experiments
designed so far. Such experiments could reveal a larger population of
sources with significant gaps of inactivity.
The lognormal distribution arises in statistical processes whose
completion depend on a product of probabilities, arising from a
combination of independent events (Montroll and Shlesinger, 1982).
Lognormal statistics have previously been used in connection with
gamma-ray bursts by McBreen \etal\ (1994) and Brock \etal\ (1994). In
the context of this investigation the physical significance of this
statistical behaviour may lie in the connection between SGRs and
neutron stars. In their paper, Hurley, K.J. \etal\ also presented a
similar statistical analysis of the behaviour of micro\-glitches from
the Vela pulsar (Cordes, Downes and Krause-Polstroff, 1988). The time
separation and the intensity of these small ( \(\mid \! \Delta\nu/\nu
\! \mid \sim 10^{-9}\) ) frequency adjustments were both compatible
with lognormal distributions, and there was no correlation between
waiting time and intensity: just as observed with SGR1806-20 (Laros
\etal, 1987). This result, combined with the identification of X-ray
point sources (Murakami \etal, 1994, Rothschild \etal, 1994) embedded
in plerion-powered SNR (Kulkarni \etal, 1993) as counterparts to the
SGR sources, suggests structural adjustments in neutron stars may be
the cause of SGRs.
\section{Conclusion}
Previously it was shown that the time intervals between successive events
from SGR1806-20 and the associated luminosity function were both lognormally
distributed. Structural adjustments in neutron stars may be responsible for
this behaviour. The activity of sources with longer mean recurrence times
was investigated using Monte Carlo simulations. The results of the
simulations indicate that there could exist a significant population of SGRs
with means longer than SGR1806-20 that remain undetected. A new breed of
experiments with very long observation times will be required to search for
this type of source.
\begin{refs}
\item Aitchison, J. and Brown, J.A.C., 1957, The Lognormal
Distribution, Cambridge University Press: Cambridge.
\item Brock, M. \etal: 1994,
in Fishman, G.J.M, Brainerd, J.J, Hurley, K., ed(s), {\it A.I.P. Conf. Proc.
307,} 672.
\item Cordes, J.M., Downs, G.S. and Krause-Polstorff,
J., 1988, Astrophys. J., 330, 847.
\item Forsythe, G.E., Malcolm, M.A. and Moler C.B., 1977,
Computer Methods for Mathematical Computations, Prentice-Hall.
\item Golenetski\v{\i}, S.V. \etal, 1987, Sov. Astron. Lett.,
13(3), 166.
\item Hurley, K. \etal, 1994, Astrophys. J., 423, 709.
\item Hurley, K.J., McBreen, B., Rabbette, M. and Steel, S., 1994, Astron.
Astrophys., 288, L49.
\item Kouveliotou, C. \etal, 1992, Astrophys. J., 392, 179.
\item Kouveliotou, C. \etal, 1993, Nature, 362, 728.
\item Kouveliotou, C. \etal, 1994, Nature, 368, 125.
\item Kulkarni, S.R. \etal, 1994, Nature, 368, 129.
\item Laros, J.G. \etal, 1987, Astrophys. J., 320, L111.
\item McBreen, B., Hurley, K.J., Long, R. and Metcalfe,
L., 1994, MNRAS, 271, 662.
\item Montroll, E.W., Shlesinger, M.F., 1982, Proc Nat Acad Sci
USA, 79, 3380.
\item Murakami, T. \etal, 1994, Nature, 368, 127.
\item Norris, J.P., \etal, 1991, Astrophys. J., 366, 240.
\item Park, S.K. and Miller, K.W., 1988, Comm ACM, 32(10), 1192.
\item Rothschild, R.E., Kulkarni, S.R. and Lingenfelter, R.E., 1994, Nature
, 368, 432.
\item Sachs, L., 1986, Applied Statistics, Springer-Verlag: New York.,
\end{refs}
\end{document}
|
\section{introduction}
\label{sec:I}
The physics of the observed $T_{c}$ suppression in superconductors
that contain nonmagnetic disorder has been the subject of much debate in
recent years. Let us focus on homogeneously disordered thin superconducting
films, with the disorder parametrized by the normal-state sheet resistance
$R_{\Box}$, and on the BCS-like quasi transition that is well pronounced
in these films although the true superconducting transition is of
Kosterlitz-Thouless nature.\cite{films} In these systems, the BCS transition
temperature $T_c$, defined as the mid-point of the resistive transition,
is observed to decrease monotonically with increasing disorder.\cite{films,R}
A complete quantitative understanding of this effect within a
microscopic strong-coupling theory has proven difficult, although
the first perturbative calculations within a phenomenological BCS
model\cite{fukuyama} were rather promising.
Qualitatively, disorder-induced changes in the electron-phonon
coupling,\cite{schmid} in the Coulomb repulsion between the constituents
of the Cooper pairs,\cite{amr} and in the normal-state
density of states\cite{dbdos} have all been identified to be important.
The difficulty lies in the fact that some of these effects tend to
suppress $T_c$ while others tend to enhance it, and $T_c$ depends
exponentially on all of them so that subtle balancing effects result.
Also, the number of parameters that acquire a disorder dependence is
quite large, and fits of theoretical results to experimental data are
therefore not necessarily very conclusive. Indeed, theories that are
structurally quite different, and mutually inconsistent, have been
shown to fit the same $T_c$ data equally well.\cite{f,trkdb}
In this situation it is obvious that one should study the disorder
dependence of quantities other than the transition temperature in an
attempt to discriminate between various theories, and to obtain
independent information about the disorder dependence of the various
parameters that determine $T_c$. One possibility is to measure the
inelastic lifetime of quasiparticles.\cite{qp,pl} The experiment by
Pyun and Lemberger\cite{pl} on amorphous InO has been analyzed by the
present authors\cite{tpddb} in the framework of a strong-coupling theory
for disordered superconductors,\cite{db} and quantitative agreement
between theory and experiment has been achieved. Another possibility
is to study the influence of nonmagnetic disorder on the pairbreaking
induced by a small amount of magnetic impurities in addition to the
nonmagnetic ones. This has the advantage that the pair breaking parameter
is easier to measure than inelastic lifetimes, and that it can be
measured simultaneously with the $T_c$ suppression in a series of
films where the nonmagnetic disorder is varied {\it in situ} by
controlling the film thickness.
Such an experiment has recently been performed by Chervenak and
Valles,\cite{jj} who studied quench condensed ultrathin films of
Pb$_{0.9}$Bi$_{0.1}$ of varying degrees of
disorder ($150\ \Omega < R_{\Box} < 2.2\ {\rm k}\Omega$, leading to a
$T_{c}$ between $6$\ K and $2.35$\ K). Of each sample, one half was doped
with Gd, while the other half was left undoped.
Gd acts as a paramagnetic impurity and leads
to pair-breaking and a reduction of $T_{c}$. The transition
temperature as a function of Gd concentration was then studied as
a function of film thickness, which is correlated with $R_{\Box}$.
The remarkable result was that the pair-breaking parameter
$\alpha$ is only mildly dependent on disorder for films with
normal state sheet resistances $R_{\Box}$ ranging from
$0.15\ {\rm k}\Omega$ to $2.2\ {\rm k}\Omega$. The implication seems to be
that the effects of disorder that lead to lower transition
temperatures do not manifest themselves in the spin-flip pair breaking
rate. An attempt to understand this behavior by a phenomenological
modification of the Abrikosov-Gorkov result for $\alpha$,
using the renormalization of the density of states inherent in
Refs.\ \onlinecite{dbdos,db} and \onlinecite{trkdb}, failed.\cite{jj}
This poses the important question whether the success of these theories in
describing the $T_c$ suppression and the disorder and temperature
dependence of the inelastic lifetime was fortuitous, and whether they
are lacking some important physical ingredient that manifests itself
in the pairbreaking rate.
It is the purpose of the present paper to analyze these questions.
What we will find is that one needs to take the strong-coupling
corrections\cite{am} to the Abrikosov-Gorkov expression into
account before one generalizes to the disordered case in order to obtain
the correct structure of the theory. Once this is done, we find that
our previous theory\cite{db,tpddb} accounts very well for the
observed effect.
\section{formalism}
\label{sec:II}
Our starting point is our theory for the suppression of the
superconducting $T_c$,\cite{db}
and the enhancement of the inelastic scattering rate,\cite{tpddb}
by nonmagnetic disorder.\cite{TheoryChoice}
In this section we recall the most salient features of this theory.
First one uses an exact eigenstate formalism
to derive generalized Eliashberg equations for the normal Green
function $G(\epsilon,i\omega)$, and the anomalous Green function
$F(\epsilon,i\omega)$. Since the wavenumber is not a good quantum number
in the presence of static impurities, $G$ and $F$ are functions of
energy and frequency, rather than wavevector and frequency as in
Eliashberg theory. The Green functions are expressed, as usual, in terms
of an anomalous self-energy, $W(\epsilon,i\omega)$, and a normal
one.\cite{ssw}
The latter is split into a piece $i\omega Z(\epsilon,i\omega)$ that is
an odd function of frequency, and a piece $Y(\epsilon,i\omega)$ that
is even in $\omega$. In terms of these quantities, $G$ and $F$ read,
\begin{mathletters}
\label{eqs:1}
\begin{equation}
G(\epsilon,i\omega)=
{i\omega Z(\epsilon,i\omega)+\epsilon+Y(\epsilon,i\omega)\over{[i\omega
Z(\epsilon,i\omega)]^{2}-[\epsilon+Y(\epsilon,i\omega)]^{2}}}\quad,
\label{eq:1a}
\end{equation}
and
\begin{equation}
F(\epsilon,i\omega)={-W(\epsilon,i\omega)\over{[i\omega
Z(\epsilon,i\omega)]^2 - [\epsilon+Y(\epsilon,i\omega)]^{2}}}\quad.
\label{eq:1b}
\end{equation}
\end{mathletters}%
In clean systems, the normal self-energy piece $Y$ is a constant that
just shifts the chemical potential and can be omitted. In the presence
of disorder, however, $Y$ has been found to be of crucial
importance,\cite{db,blm,db2} and to reflect the physics of the Coulomb
anomaly in the density of states\cite{aa} in the context
of superconductivity.
The generalized Eliashberg equations then take the form of integral
equations in both energy and frequency for the three functions $Z$,
$W$, and $Y$. A solution of these equations has been obtained by
means of various approximations. In particular, the frequency dependence
of $Y$ was found to be weak and could be omitted, and its energy
dependence was approximated by the first term in a Taylor expansion
about a characteristic energy $\bar\omega$,
\begin{equation}
Y(\epsilon,i\omega) \simeq (\epsilon-\bar\omega) Y^{\prime}\quad,
\label{eq:2}
\end{equation}
with $Y^{\prime} = \partial Y/\partial\epsilon \vert_{\epsilon=\bar\omega}$,
and $\bar\omega$ an average phonon
frequency.\cite{vdg,db2} With some further approximations, the
energy integrations could then be performed, and the theory be cast
in the same form as Eliashberg theory. A two-square well approximation
then leads to a $T_c$ formula that has the structure of a generalized
McMillan or Allen-Dynes formula, viz.\cite{db,db2}
\begin{equation}
T_c = {\omega_{\log}\over 1.2}\exp\left[{-1.04(1+\tilde\lambda+Y^{\prime})
\over \tilde\lambda - \tilde\mu^{*}[1 + 0.62\tilde\lambda/(1+Y^{\prime})]}
\right]\quad.
\label{eq:3}
\end{equation}
Here $Y^{\prime}$ is the normal self-energy piece mentioned above,
and $\tilde\lambda$ and $\tilde\mu^{*}$ are disorder dependent
generalizations of the
electron-phonon coupling constant $\lambda$ and the Coulomb pseudotential
$\mu^{*}$, respectively, in Eliashberg theory. Explicit expressions for
all three of these quantities have been given in Ref.\ \onlinecite{db},
and will be evaluated for the case of thin films below.
In the presence of magnetic impurities, $T_c$ is reduced by pair
breaking.\cite{ag} Abrikosov-Gorkov theory has been modified to allow
for strong-coupling effects,\cite{am} with the only resulting change being
a factor of $1/Z$ in the pair breaking parameter. The result is
\begin{equation}
-\ln(T_{c}/T_{c0})=\psi(\alpha/2\pi T_{c}+1/2)-\psi(1/2)\quad,
\label{eq:4}
\end{equation}
with T$_{c0}$ the value of $T_c$ in the absence of the magnetic impurities,
$\psi$ the di-gamma function, and $\alpha$ the pair breaking
parameter $\alpha=(1/Z)(1/\tau_{s})$, with $1/\tau_{s}$ the spin-flip
scattering rate. In the case of clean superconductors, $Z=1+\lambda$.
In the presence of nonmagnetic disorder, and with the same approximations
that lead to the $T_c$ formula given by Eq.\ (\ref{eq:3}), it is
straightforward to repeat the calculation of Allen and Mitrovic\cite{am}
within the framework of the theory of Ref.\ \onlinecite{db}. The
result is again Eq.\ (\ref{eq:4}), but with $\alpha$ replaced by a
disorder dependent $\tilde\alpha$ which in turn is related to
disorder dependent parameters,
\begin{equation}
\tilde\alpha={1/\tilde\tau_s\over{1+\tilde\lambda}}\quad.
\label{eq:5}
\end{equation}
Here $\tilde\lambda$ is the same quantity as in Eq.\ (\ref{eq:3}), and
$1/\tilde\tau_s$ is the disorder dependent spin flip-scattering rate.
Throughout this paper, we choose units such that $\hbar = k_B = 1$.
In the next section, we derive an explicit form for $\tilde\alpha$ in a
disordered thin superconducting film.
\section{disorder dependence of $\alpha$}
\label{sec:III}
\subsection{Electron-phonon coupling $\tilde\lambda$}
\label{subsec:III.A}
The electron-phonon coupling strength $\tilde\lambda$ is defined as
an integral over the Eliashberg function $\alpha^2 F$,
\begin{equation}
\tilde\lambda=2\int {d\nu\over{\nu}}\ \alpha^{2}F(\nu)\quad.
\label{eq:6}
\end{equation}
$\alpha^2 F$, and hence $\tilde\lambda$, are disorder dependent due
to effects first discussed by Pippard\cite{pippard} in the context
of ultrasonic attenuation, and by Schmid\cite{schmid} for the electron-phonon
inelastic lifetime. The main physical point is that disorder decreases
the coupling of the electrons to longitudinal phonons due to collision drag,
but increases the coupling to transverse phonons due to the breakdown of
momentum conservation. For realistic parameter values the
latter effect is stronger than the former, leading to an overall increase
of $\tilde\lambda$ with disorder.\cite{dba2F} For Debye phonons in $3-d$
systems, $\tilde\lambda$ has been calculated in Ref.\ \onlinecite{db}.
Repeating that calculation in $d=2$ is straightforward, and very similar
to the corresponding calculation of the electron-phonon inelastic
lifetime.\cite{dbsds} The result is\cite{tpdthesis}
\begin{equation}
\tilde\lambda = 2\int_{0}^{\omega_{D}}{d\nu\over{\nu}}
{\nu^{2}l\over{\pi m}}
\sum_{b=L,T} {d_{b}\over{c_{b}^{3}}}\ f_{b}(\nu l/c_{b})\quad,
\label{eq:7}
\end{equation}
for a system with mean free path $l$. Here $c_{L,T}$ are the longitudinal and
transverse speeds of sound, respectively, $\omega_{D}$ is the Debye frequency,
the dimensionless constant
$d_{b}=k_{F}^{3}/16\pi \rho_{i}c_{b}$ with
$\rho_{i}$ the ion density and $k_F$ the Fermi wave number, and
the functions $f_{T,L}$ are given by,\cite{dbsds}
\begin{mathletters}
\label{eqs:8}
\begin{equation}
f_{T}(x)={8\over{x^{4}}}(1+x^{2}/2-\sqrt{1+x^{2}})\quad,
\label{eq:8a}
\end{equation}
\begin{equation}
f_{L}(x)=2\left({1\over{\sqrt{1+x^{2}}-1}} -{2\over{x^{2}}}\right)\quad.
\label{eq:8b}
\end{equation}
\end{mathletters}
Substituting Eqs.\ (\ref{eqs:8}) into Eq.\ (\ref{eq:7}) we obtain the
disorder dependence of $\tilde\lambda$,
\begin{mathletters}
\label{eqs:9}
\begin{eqnarray}
\tilde\lambda&=&{\lambda\hat R_{\Box}E_{F} c_{L}\over{\pi\omega_{D}v_{F}}}
\biggl[F_{L}\left({\pi\omega_{D}v_{F}\over{\hat R_{\Box}E_{F}c_{L}}}\right)
\nonumber\\
&+&2{c_{L}^{2}\over{c_{T}^{2}}} F_{T}\left({\pi\omega_{D}v_{F}\over
{\hat R_{\Box}E_{F}c_{T}}}\right)\biggr] \quad,
\label{eq:9a}
\end{eqnarray}
where we have defined two functions,
\begin{eqnarray}
&F_{L}(x)&=\sqrt{1+x^{2}}-1-\ln[(\sqrt{1+x^{2}}+1)/2]\quad, \nonumber \\
&F_{T}(x)&= {1-\sqrt{1+x^{2}}\over{2(1+\sqrt{1+x^{2}})}}
+\ln[(\sqrt{1+x^{2}}+1)/2]\quad,
\label{eq:9b}
\end{eqnarray}
\end{mathletters}%
with $\lambda = 4\omega_{D}d_{L}/\pi m c_{L}^{2}$ the electron-phonon
coupling in a clean $2-d$ system. The dimensionless resistance
$\hat R_{\Box} = R_{\Box} e^2/\hbar \approx R_{\Box}/4.1\ {\rm k}\Omega$
is a measure of the disorder in the material.
As in three dimensions, the size of the disorder renormalization of $\lambda$
depends on the ratio of the longitudinal to the transverse
speed of sound. This is a result of the abovementioned
competition between an increase in the coupling between electrons and
transverse phonons and a decrease of the
coupling to longitudinal phonons.
Since the transverse speed of sound is invariably smaller than the
longitudinal one, $\tilde\lambda$ increases with increasing disorder.
This effect tends to reduce the pair
breaking rate, Eq. (5). However, we also have to calculate the disorder
dependence of the spin-flip scattering rate in order to obtain the
disorder dependence of $\tilde\alpha$.
\subsection{Spin-flip scattering rate $1/\tau_{s}$}
\label{subsec:III.B}
The interaction between the electron spin and an impurity spin
${\vec S}({\vec r})$ at site ${\vec r}$
is described by a Hamiltonian,
\begin{equation}
H_{S}=\sum_{{\bf k},{\bf k^{\prime}},\mu,\nu} J_{{\bf k},{\bf k^{\prime}}}\
{\vec S}({\bf k}-{\bf k^{\prime}})\cdot
(c^{\dagger}_{{\bf k^{\prime}}\mu}\vec\sigma_{\mu\nu}c_{{\bf k}\nu})\quad.
\label{eq:10}
\end{equation}
Here $c^{\dagger}$ and $c$ are fermion operators,
$\vec\sigma = (\sigma_x, \sigma_y, \sigma_z)$ denotes the Pauli matrices,
and $J_{{\bf k},{\bf k^{\prime}}}$ denotes
the electron-magnetic impurity exchange
integral. We now calculate the electron self energy contribution,
$\Sigma$, due to this interaction in Born approximation. It is most
convenient to do this in
an exact eigenstate representation, in analogy to the calculation of
the Coulomb self energy in Ref.\ \onlinecite{aalr}. The calculation
is straightforward, and we obtain
\begin{equation}
\Sigma(\epsilon,i\omega)=\int {d\epsilon^{\prime}\over{N_{F}}}
G(\epsilon^{\prime},i\omega)\sum_{{\bf q}}V_{S}({\bf q})
F_{s}({\bf q},\epsilon-\epsilon^{\prime})\quad.
\label{eq:11}
\end{equation}
Here, $N_{F}$ is the free electron density of
states per spin at the Fermi level.
We only retain the $s-$wave component of the interaction so that
$V_{S}({\bf q})=n_{P}S(S+1) J^2$, where $n_{P}$ is the concentration of
paramagetic impurities, and J is a measure of the exchange interaction
strength.\cite{rkky} $G(\epsilon,i\omega)$
is the normal Green function in the superconductor and is given by
Eq.\ (\ref{eq:1a}). Finally, $F_s$ is the spin density analogue of the
density-density correlation function denoted by $F$ in
Ref.\ \onlinecite{aalr}. If we work to lowest order in the electron-impurity
spin interaction, and neglect Coulomb and finite temperature effects in
$\Sigma$, then $F$ and $F_s$ are identical.
We obtain the spin-flip scattering rate $1/\tau_s$ from the self energy
$\Sigma$ by analytically continuing to real frequencies,
$i\omega \rightarrow \omega + i0$, and going `on-shell', i.e. putting
$\epsilon = \omega$. For our purposes, we are interested in the
influence of spin-flip scattering on the superconducting $T_c$. The
physics that determines the latter is dominated by processes on a
frequency scale of $\bar\omega$, a typical phonon frequency. For the
same reason for which we take the parameter $Y^{\prime}$ in Eq.\ (\ref{eq:2})
at the frequency $\bar\omega$ we therefore define
\begin{equation}
1/\tau_s = -2{\rm Im}\Sigma(\bar\omega,i\omega\rightarrow\bar\omega + i0)
\quad.
\label{eq:12}
\end{equation}
In a clean system, the spin-density correlation function $F_s(q,\omega)$
is frequency independent. In that case we recover from Eq.\ (\ref{eq:12})
the well known result\cite{ag,am}
\begin{equation}
{1\over\tau_s}=n_{P}S(S+1) J^2 4N_{F}\quad.
\label{eq:13}
\end{equation}
In a disordered system, $F_s$ is diffusive,\cite{aalr} and in the
Green function $G$ we must keep the self energy piece $Y^{\prime}$ as
discussed above. We thus obtain
\begin{mathletters}
\label{eqs:14}
\begin{equation}
{1\over\tilde\tau_s}={2 n_{P}S(S+1)
J^2\over{N_{F}[1+Y^{\prime}]}}\sum_{\bf q}
F_s\left({\bf q},\bar\omega{\tilde\lambda \over 1 + Y^{\prime}}
\right)\quad,
\label{eq:14a}
\end{equation}
where
\begin{equation}
F_s({\bf q},\omega) = g(q) {Dq^2\over \omega^2 + (Dq^2)^2}\quad,
\label{eq:14b}
\end{equation}
\end{mathletters}%
with $g(q)$ the Lindhard function, which for simplicity we replace
by $N_F \Theta(2k_F - q)$. Here $D$ denotes the normal phase spin
density diffusion coefficient, which in
the noninteracting electron approximation coincides with the mass
or charge diffusion coefficient, so $D=\pi/m\hat R_{\Box}$.
Performing the wavenumber integral in Eq.\ (\ref{eq:14a}) we
finally obtain
\begin{mathletters}
\label{eqs:15}
\begin{equation}
{1\over\tilde\tau_s} = {1\over\tau_s}
\left\{1+{1\over1+Y^{\prime}}
{\hat R_{\Box}\over{8\pi}}
\ln\left[1 + \left({8\pi\over\hat R_{\Box}}{\epsilon_F\over
\bar\omega^{*}}\right)^{2}\right]\right\}\quad,
\label{eq:15a}
\end{equation}
with
\begin{equation}
\bar\omega^{*} = \bar\omega{\tilde\lambda\over 1 + Y^{\prime}}
\quad,
\label{eq:15b}
\end{equation}
\end{mathletters}%
and $1/\tau_s$ given by Eq.\ (\ref{eq:13}). $1/\tilde\tau_s$ depends on
disorder both explicitly, and implicitly through $Y^{\prime}$. Our
final task is therefore to calculate the dependence of $Y^{\prime}$
on $\hat R_{\Box}$.
\subsection{Normal Self Energy Piece $Y^{\prime}$}
\label{subsec:III.C}
In order to calculate $Y^{\prime}$ we again have to repeat the
calculations of Ref.\ \onlinecite{db} in $d=2$. Both the
electron-electron and the
electron-phonon contributions to the self energy contribute to the
self energy piece $Y$. Performing a Taylor
series expansion in energy around $\epsilon=\bar\omega$ of Eq. (2.12) of
Ref.\ \onlinecite{db}, we obtain
\begin{equation}
Y^{\prime}(\bar\omega)=\delta U_{C}^{Y}(\bar\omega)+4\int {d\nu\over{\nu}}
\delta\alpha^{2}F^{H}(\bar\omega,\nu)\quad.
\label{eq:16}
\end{equation}
$\delta U_{C}^{Y}(\bar\omega)$, which describes the Coulomb contribution
to $Y^{\prime}$, is taken from Ref.\ \onlinecite{db},
\begin{eqnarray}
& &\delta U_{C}^{Y}(\bar\omega)={1\over{\pi N_{F}}}\sum_{\bf q}
g({\bf q}){Dq^{2}\over{(Dq^{2})^{2}+\bar\omega^{2}}}\times
\label{eq:17} \\
& &\left\{
V_{C}({\bf q})-{2\over{g({\bf q})^{2}}}\sum_{\bf k,p}
\sum_{\bf k^{\prime},p^{\prime}}g_{\bf k,k^{\prime}}({\bf q})
g_{\bf p,p^{\prime}}({\bf q})V_{C}({\bf k-p})\right\},\nonumber
\end{eqnarray}
with the statically screened Coulomb potential,
\begin{equation}
V_{C}({\bf q})={1\over{2N_{F}}}{\kappa\over{\kappa+q}}\quad, \ \ \ \ \ \ \hfil
\kappa=4\pi e^{2}N_{F}\quad.
\label{eq:18}
\end{equation}
Using the prescription for performing momentum sums of this type as
described in Ref.\ \onlinecite{db}, the integrals can be done and yield
\begin{eqnarray}
\delta U_{C}^{Y}(\bar\omega)&=&{\mu \hat R_{\Box}\over{8\pi}}
\biggl[G\left({\hat R_{\Box}\bar\omega\over{8\pi E_{F}}}{4k_{F}^{2}\over{\kappa^{2}}},{\hat R_{\Box}\bar\omega\over{8\pi E_{F}}},{2k_{F}\over{\kappa}}\right)\nonumber \\&-&{2\over{\pi}} H\left({\hat R_{\Box}\bar\omega\over{8\pi E_{F}}},{2k_{F}\over{\kappa}}\right)\biggr]\quad,
\label{eq:19}
\end{eqnarray}
with the functions
\begin{eqnarray}
&G&(x,y,z)={z\over{1+x^{2}}}{1\over{\ln(1+z)}}
\nonumber\\
&\times&\biggl\{\ln\left[
{1+1/y^{2}\over{(1+z)^{4}}}\right]-\sqrt{{x\over{2}}}(1-x)\ln\left[
{1-\sqrt{2y}+y\over{1+\sqrt{2y}+y}}\right]\nonumber \\
&+&\sqrt{2x}(1+x)\tan^{-1}\left({\sqrt{2y}\over{y-1}}\right)
-2x\tan^{-1}(1/y)\biggr\}\quad; \nonumber \\
&H&(y,z)={z\over{\ln(1+z)}}\ln(1+1/y^{2}) {1\over{\sqrt{z^{2}-1}}}
\nonumber\\
&\times &\ln\left[{z+\sqrt{z^{2}-1}\over{z-\sqrt{z^{2}-1}}}\right]\quad.
\label{eq:20}
\end{eqnarray}
The Coulomb pseudopotential $\mu$ in $d=2$ is given by
\begin{equation}
\mu={\kappa\over{2\pi k_{F}}}\ln\left(1+{2k_{F}\over{\kappa}}\right)\quad.
\label{eq:21}
\end{equation}
As discussed in Ref.\ \onlinecite{db},
the phonon contribution to $Y^{\prime}$, which is given by the second
term on the right-hand side of Eq.\ (\ref{eq:16}), is related to
a stress-stress correlation function and can be calculated in a similar
manner as $\delta U_{C}^{Y}$. One obtains
\begin{eqnarray}
&4&\int {d\nu\over{\nu}} \delta\alpha^{2}F^{H}(\bar\omega,\nu)=\\
& &\lambda\hat R_{\Box}{c_{L}E_{F}\over{2\pi^{2}v_{F}\omega_{D}}}
\sin^{-1}\left[{\omega_{D}v_{F}\over{4 E_{F}c_{L}}}\right]
\ln\left[1+\left({8\pi E_{F}\over{\hat R_{\Box}\bar\omega}}\right)^{2}\right]\quad.
\nonumber
\end{eqnarray}
Finally, both contributions can be collected to give
\begin{eqnarray}
& &Y^{\prime}(\bar\omega)=\hat R_{\Box}\biggl\{{\mu\over{8\pi}}
[G-{2\over{\pi}} H] \nonumber \\
&+&\lambda{c_{L}E_{F}\over{2\pi^{2}v_{F}\omega_{D}}}
\sin^{-1}\left[{\omega_{D}v_{F}\over{4E_{F}c_{L}}}\right]
\ln\left[1+\left({8\pi E_{F}\over{\hat R_{\Box}\bar\omega}}\right)^{2}\right]\biggr\}\ ,
\label{eq:23}
\end{eqnarray}
where $G$ and $H$ have the same arguments as in Eq.\ (\ref{eq:19}).
The enhancement of $Y^{\prime}$ with increasing disorder is
due to the opening of a correlation gap in the (normal state)
density of states,
and contributes to the decrease of T$_{c}$.\cite{db,db2}
\section{Final Result and Discussion}
\label{sec:IV}
We are now in position to collect our results and thus
obtain the disorder dependence of the magnetic
pair breaking rate. Substituting Eqs.\ (\ref{eqs:9}), (\ref{eqs:15}), and
(\ref{eq:23}) into Eq.\ (\ref{eq:5}) yields our final result,
\begin{equation}
{\tilde\alpha\over\alpha}=
{1+\lambda\over{1+\tilde\lambda}}
\left\{1+{1\over{1+Y^{\prime}
}}{\hat R_{\Box}\over{8\pi}}\ln\left[1+\left({8\pi E_{F}\over{\hat R_{\Box}\bar\omega^{*}}}\right)^{2}\right]\right\}.
\end{equation}
The disorder renormalizations of the pair breaking rate appear both
in the numerator and denominator and therefore the rate may either
increase or decrease with increasing
disorder depending upon the material parameters
$\lambda,c_{L},c_{T},v_{F},E_{F},\omega_{D},$ and $\mu$.
We now address the experiment on Pb$_{0.9}$Bi$_{0.1}$ by Chervenak
and Valles.\cite{jj}.
To estimate the parameters entering into
Eq. (24), let us first consider the parameter values for bulk PbBi,
as far as available, given in Ref.\ \onlinecite{ad}.
Thereby we have $\bar\omega=56 K$,
and $\omega_{D}=108K$. The Bohm-Staver relation gives
\begin{equation}
{\omega_{D}\over{E_{F}}} {v_{F}\over{c_{L}}}=2 (2/Z)^{1/3}\quad.
\label{eq:25}
\end{equation}
For clean bulk Pb one has $E_{F}=1.1 \times 10^{5} K$, $c_{L}=$2050 m/s, $
c_{T}=710$ m/s, and $Z=4$. However, we do not expect the actual parameter
values to correspond to those for either bulk Pb or
bulk Pb$_{0.9}$Bi$_{0.1}$. First of all, the material in question is a thin
film, and moreover the substrate is expected to modify its properties,
in particular the acoustic ones. Evidence for this is provided by the fact
that the parameter values quoted above do not give
the correct value of the clean limit
$T_{c}$ as measured in Ref.\ \onlinecite{jj}.
It is therefore likely that the substrate on which the thin layer of
Pb is deposited strongly affects the phonon spectra of the film,
altering both $\lambda$ and the ratio $c_{L}/c_{T}$ compared to bulk Pb.
Accordingly, we choose a value for the bare $\lambda=1.12$ and
$\mu=0.1$ which (in the absence of disorder) reproduces the highest T$_{c}$ as
measured in Ref.\ \onlinecite{jj}. Lastly, since $c_{L}/c_{T}$ is not known
even for Pb$_{0.9}$Bi$_{0.1}$, we let $c_{L}/c_{L}$ be determined by a
fit to the data.
These parameters provide the curves shown in Figure 1. The solid curves
are the results for the disorder dependence of the normalized pair
breaking rate as given by Eq. (5). The points represent the data taken from
Ref. \ \onlinecite{jj} for two different runs. The decrease
at small $\hat R_{\Box}$ is due to the fact that for small disorder,
$\tilde\lambda$ and $Y^{\prime}$ grow more rapidly
than $1/\tilde\tau_{s}$. The normalized rate
goes through a shallow minimum at roughly $\hat R_{\Box}\sim 0.3$ at which
point the disorder renormalizations of $Y^{\prime},\tilde\lambda,$ and
$1/\tilde\tau_{s}$ are balanced and offset each other. With further
increasing disorder the enhancement of $1/\tilde\tau_s$ dominates,
and leads to a slowly increasing pair breaking rate.
We remark that the point at which the minimum occurs depends sensitively
on the ratio of the longitudinal and transverse speeds of sound.
To obtain the solid lines in Fig. 1a,
$c_{L}/c_{T}=1.9$ was used while 2.1 was used for Fig. 1b.
These values lie between the
value for bulk Pb (2.88) and the substrate (similar to
pyrex, 1.72) used in Ref. \ \onlinecite{jj} and thus does not seem
unreasonable. Larger values
of $c_{L}/c_{T}$ yield a more drastic reduction of the rate for small
disorder and the region of increasing $\alpha$
occurs at larger values of $R_{\Box}$. This sensitivity of the overall
shape of the curve to the material parameters may be reflected in the
slightly different results obtained for the two experimental runs in
Ref. \ \onlinecite{jj} as shown in Fig. 1. It would therefore be very
interesting to repeat
the experiments using substrates with different acoustic properties.
In summary, we have presented a theory for the paramagnetic pair breaking
rate in disordered superconducting films and have shown that the
disorder dependence of the rate depends delicately on the disorder
renormalizations of $Y^{\prime}$, $\tilde\lambda,$ and
$1/\tilde\tau_{s}$. As a result, the rate can
either increase or decrease with disorder, depending upon material parameters,
and in general it is not a monotonic function of disorder.
Our conclusion is that the disorder dependence of the rate as observed by
Chervenak and Valles in Pb$_{0.9}$Bi$_{0.1}$ films can be quantitatively
understood via an application of the
microscopic theories developed in Refs. \ \onlinecite{tpddb} and \
\onlinecite{db}.
\acknowledgements
We gratefully acknowledge helpful discussions with Jay Chervenak,
Jim Valles, and Martin Wybourne. Part of this work was performed at
the TSRC in Telluride, CO, and we thank the Center for its hospitality.
This work was supported by the NSF under grant numbers DMR-92-06023,
DMR-92-09879, and DMR-95-10185.
|
\section{Introduction}
\indent
Quantum chromodynamics (QCD) tells us that most, if not all,
of light hadron masses are
generated spontaneously by the breaking of chiral symmetry from
{$SU(N_f)\times SU(N_f)$} to diagonal $SU(N_f)$ where $N_f$ is the number of flavors,
equal to 2 without strangeness and 3 with. It is also widely believed that
as a hadronic system is heated to high temperature or compressed
to high density, the broken symmetry will get restored in a way paralleling
what happens in condensed matter physics.
A natural consequence of the restoration of the chiral symmetry must then
be that the spontaneously generated masses disappear as density
(and perhaps also temperature) is increased. The question we are
raising is how does this ``shedding of mass" occur?
This question is at the core of the fundamental theory of matter:
How is the mass generated, starting with the lightest object like neutrinos
to the heaviest detected particle like the top quark?
The aim of this talk is to describe how hadron properties get modified
in medium as the
system is heated or compressed. That is, immerse a hadron in medium and
compress the system or heat it. What does one expect to see happening?
To answer this question, let me start with the simplest nuclear system,
namely the deuterium. Let us look at what happens when a soft photon is
sent in to probe the system. Consider therefore the well-known inverse
process
\begin{eqnarray}
n + p\rightarrow d +\gamma\label{npcap}
\end{eqnarray}
at thermal neutron energy. This process was first explained in a
quantitative way by Riska and Brown \cite{riskabrown} in 1973. What
I will do here is to describe it more accurately in a modern QCD framework.
Since it is a very low-energy process, QCD can be
represented by an effective chiral Lagrangian field theory. This is because
at long wavelength limit, chiral perturbation theory (ChPT) is believed to be
exactly equivalent to QCD \cite{chptqcd}. This invites us to attempt to
describe
(\ref{npcap}) in terms of a chiral Lagrangian. There is one basic problem
in doing this and that has to do with the description of the deuteron
in QCD: we do not really know how to derive the deuteron starting {\it
directly}
from QCD. Since we are focusing on the chiral aspect of the problem, however,
the solution might be sought in a chiral Lagrangian approach to the
structure of the deuteron. If the number of colors $N_c$ is in some sense
big -- which gives rise to what is known as the ``large $N_c$ limit" --
then the effective Lagrangian is given by meson fields only as we know
from the skyrmion structure. There is some important progress in
obtaining a bound deuteron, recently through the work of Manton \cite{manton},
as a baryon number 2 skyrmion, but we are still far from
understanding it quantitatively. However there is an indirect approach to this
which is consistent with QCD and which has the potential to be quantitatively
accurate, namely that as we have known all along, the deuteron is made up of
a proton and a neutron bound by meson exchanges: In the framework of QCD,
the nucleon may be gotten from a large $N_c$ Lagrangian as a soliton
(skyrmion) but this is now known to be
equivalent, at least in the large $N_c$ limit,
to having the nucleon as a matter field
in the chiral Lagrangian. We are thus led to consider a chiral Lagrangian
that contains baryons (nucleon, $\Delta$ etc.), pseudo-Goldstone bosons $\pi^i$
(pions, kaons etc.), vectors $V_\mu=
\omega_\mu, \rho_\mu, \cdots$ etc. with suitable chiral invariant couplings.
There have been some attempts to compute the
deuteron from such a Lagrangian in chiral perturbation theory
\cite{vankolck} but at present the calculation can be done only at low orders
since higher order calculations would involve too many parameters to be
completely determined from available experiments. Luckily for our purpose,
we need not compute the deuteron from first principles as I shall
argue below.
The important point to note
is that certain aspects of the deuteron which have to do
with the chiral symmetry structure of hadrons can be probed by
the process (\ref{npcap}) {\it without knowing}
how to get the nucleus itself from
a chiral perturbation theory, as recently discussed by Park, Min and Rho
\cite{pmr}. Briefly the argument goes as follows:
For physics with energy scale much less than the chiral scale $\Lambda_\chi
\sim 4\pi f_\pi\sim 1$ GeV (where $f_\pi$ is the pion decay constant
$\sim 93$ MeV), the relevant Lagrangian is, schematically,
\begin{eqnarray}
{\cal L}=\sum_i {\cal L}_i [B, U, {\cal M}]
\end{eqnarray}
where $B_i$ are the baryon fields (both octet and decuplet), $U$ the unitary
pseudo-Goldstone fields $U=e^{i\pi/f_\pi}$ and ${\cal M}$ the quark mass matrix.
The degrees of freedom more massive than the chiral scale $\Lambda_\chi$ are
integrated out, appearing implicitly in the counter terms of the Lagrangian.
Chiral symmetry requires that there be only derivative couplings apart from
terms involving the
mass matrix and hence effectively the Lagrangian is an
expansion in $\partial/\Lambda_\chi$ and ${\cal M}/\Lambda_\chi^2$. Since the baryon
mass is $\sim \Lambda_\chi$, the baryons should be introduced as static
matter fields, so the derivative on the baryon field does not involve
time derivatives. ChPT is then a systematic perturbative expansion
in powers of $Q$, say, $Q^n$, where $Q$ is the energy-momentum scale
being probed,
with suitable counter terms to remove divergences and to take account of the
degrees of freedom that are
integrated out. Now in applying this theory to nuclear
systems, we need to separate the class of Feynman diagrams into two,
one ``irreducible" and the other ``reducible." It is in calculating the
irreducible diagrams that ChPT enters. The reducible diagrams -- that
cannot be treated by ChPT because of infrared singularity -- are incorporated
by solving a Lippman-Schwinger equation or Schr\"odinger equation with the
potential obtained with the irreducible graphs by ChPT. This is how
bound states are to be treated in ChPT. Now in calculating the process
(\ref{npcap}), we can write the EM current in two terms, one the
single-particle
current $j^{(1)}$ and the other the two-body current $j^{(2)}$
\begin{eqnarray}
j_\mu^{EM}=j_\mu^{(1)} + j_\mu^{(2)}.\label{current}
\end{eqnarray}
The former
is called ``impulse approximation current" and the latter ``exchange
current." For the system considered, we terminate with the two-body current.
Later we will see that in heavy nuclei there can enter many-body currents,
some of which become quite important.
Very accurate wave functions for the final deuteron and the initial
neutron-proton system obtained from some accurate phenomenological potential
such as the Argonne $v_{18}$ potential \cite{v18}
(or a potential calculated in a high order ChPT if it is feasible)
would correspond to a high order chiral expansion since
the Schr\"odinger equation sums a certain class of
chiral series to all orders and
presumably the phenomenological potential also subsumes all orders of chiral
perturbation. Now the idea is to compute the matrix element of the
current (\ref{current}) in chiral perturbation expansion in such a way
that is consistent with the calculation of the wave functions.
The calculation of the one-body current $j^{(1)}$ is without ambiguity.
That of the two-body exchange current is somewhat subtle, requiring
a careful sorting of irreducible and reducible contributions such that
the reducible ones are suitably accounted for in the one-body term
with the accurate wave functions. In this way, the Ward identities
associated with the conserved vector current are satisfied to a
given chiral order in the EM current. This is a numerically accurate
procedure.
\begin{figure}
\centerline{\epsfig{file=pmr3fig6.eps}}
\caption[np]{The np capture rate calculated in chiral perturbation theory.
The predicted capture rate (upper plot) $\sigma=(334\pm 3)$ mb agrees with
the experimental value $(334.2\pm 0.5)$ mb. The lower plot shows
the ratios of the matrix elements of the two-body terms over the one-body term.
``Tree" corresponds to the leading chiral order one-pion exchange term (with
the blob in Figs.\ref{graphs}(a,b) replaced by a bare vertex),
``$1\pi (\omega)$" and ``$1\pi (\Delta)$" correspond to the next-to-leading
order corrections involving the $\omega$ meson and $\Delta$ resonance,
respectively, to the one-pion exchange tree terms Figs.\ref{graphs}(a,b).
``$2\pi$"
is the genuine loop correction to the tree contribution. The hadrons appearing
in this calculation all have free-space properties. }\label{nprates}
\end{figure}
A recent calculation \cite{pmr} of the process (\ref{npcap}) is given in
Fig.\ref{nprates}. The calculation was done to order $Q^3$ which corresponds
to next-to-next-to leading order in chiral expansion, that is, to one-loop
order. Given the nucleon mass $m_N$, the pion decay constant
$f_\pi$, the pion mass $m_\pi$, the axial coupling constant $g_A$
and the vector meson mass $m_V$ all {\it determined in free space},
all the parameters that appear in the theory
are fixed in the theory except for
the hard core radius $r_c$ in the wave function reflecting on our inability
to handle in ChPT
very short-range physics. The remarkable agreement with
the experiment shows that the two-nucleon systems we are looking at
are made up of two nucleons with their properties as given in free space:
the chiral Lagrangian with the vacuum values of hadron parameters
describes nature remarkably well at low chiral orders.
\begin{figure}
\centerline{\epsfig{file=pmr3fig1.eps}}
\caption{Generic graphs contributing to exchange currents.
(a) and (b) are one-pion exchange and (c) represents multipion and/or
heavy-meson exchange currents. The large filled circles represent one-nucleon,
one-pion irreducible graphs, the solid line the nucleon, the dotted line the
pion and the wiggly line the current $j_\mu$.}\label{graphs}
\end{figure}
A very valuable information for heavier and denser nuclei
is lodged in some of the terms that are
{\it negligible} in the process (\ref{npcap}). As I showed elsewhere
\cite{mr91},
two-body currents involving four baryon fields (e.g. Fig.\ref{graphs}(c)
with the blob replaced by a point) like
\begin{eqnarray}
\bar{B}\Gamma_\mu B \bar{B}\tilde{\Gamma} B\label{fourbaryon}
\end{eqnarray}
with $\tilde{\Gamma}$ and $\Gamma_\mu$ representing some Lorentz scalar
and Lorentz vector quantities consistent with chiral symmetry, respectively,
are subdominant in the chiral counting and can be ignored. Now
terms like (\ref{fourbaryon}) show up in the chiral Lagrangian
as a result of integrating out heavy degrees of freedom with an energy scale
$E\roughly> \Lambda_\chi$. Consider for instance the scalar meson $\sigma$
that plays an important role in effective field theory of nuclei, a prototype
of which being the
Walecka model. In free space, there is no low-lying scalar that can appear
in the low-energy chiral Lagrangian. But there is a high-lying
scalar field that can be associated with the trace anomaly of QCD
\begin{eqnarray}
(T_\mu^\mu)_{QCD}=-\frac{\beta (g)}{2g} (G_{\mu\nu}^a)^2\sim \chi^4
\end{eqnarray}
where $\beta$ is the QCD beta function, $g$ the color gauge coupling constant,
$G_{\mu\nu}^a$ the gluon field tensor and $\chi$ the scalar glueball field.
Now the $\chi$ field is massive, with $m_\chi\sim 2$ GeV, so this
degree of freedom appears only in the counter terms. It will give rise to
a term like
\begin{eqnarray}
\kappa \bar{B}j_\mu^{(1)}B \bar{B}B\label{sigmacurrent}
\end{eqnarray}
with the coefficient $\kappa$ suppressed by the power $(Q/m_\chi)^2$.
In the process (\ref{npcap}), $Q$ is of order of 40 MeV, so $(Q/m_\chi)^2
< 10^{-3}$. Stated differently an effective two-body term like
(\ref{sigmacurrent}) will be screened by the short-range correlation implicit
in the wave functions.
The situation is quite different, however, in dense nuclear medium.
As discussed in \cite{brpr}, as density increases, the scalar
(or precisely the quarkish component of the scalar)
$\chi$ moves downwards in energy and at some high density, it joins
the triplet of nearly massless pions to make up the quartet of the $O(4)$
symmetry of chiral symmetry.
The merging presumably takes place at the chiral transition point
discussed below.
The point is that as discussed by Beane and van Kolck
\cite{beane}, in order to reconcile Weinberg's
``mended symmetry" \cite{mendedwein}
with effective chiral Lagrangians at some shorter length scale, the scalar
field must come down as a dilaton. Now if it comes down below
the chiral scale $\Lambda_\chi$, then we can no longer consider the scalar as a
counter-term contribution. It will strongly couple to low-mass multipion
excitations giving among others
what is usually taken as a scalar field $\sigma$ in
effective nuclear forces \cite{brpr}. All other hadrons (other than Goldstone
bosons) will also couple to this
scalar as well and will undergo a mass shift as density increases.
A chiral Lagrangian that accounts for this phenomenon has been shown to
lead to the Brown-Rho scaling \cite{br91}
\begin{eqnarray}
m_B^\star/m_B\approx m_V^\star/m_V\approx m_\sigma^\star/m_\sigma
\approx f_\pi^\star/f_\pi\approx \cdots\label{BR}
\end{eqnarray}
where $B$ stands for baryons, $V$ for vector mesons and $\sigma$ for the
dilatonic scalar. The star stands for density-dependent quantities.
Such a chiral Lagrangian effective in dense system will then contain
these effective
constants instead of the free-space values used in the process
(\ref{npcap}) while preserving the free-space chiral symmetry. As I will argue,
the consequence of this scaling can be significant in heavy nuclei and nuclear
matter.
One can already
see the effect of this scaling in finite nuclei. One clear case is
the axial charge transition in nuclei. It was shown in \cite{pmr,pmr93} that to
order $Q^3$ in chiral expansion the axial charge transition matrix element
in heavy nuclei
\begin{eqnarray}
A (0^\pm)\rightarrow B(0^\mp),\ \ \ \ \Delta T=1\label{axial}
\end{eqnarray}
is enhanced with respect to the impulse approximation
by the factor
\begin{eqnarray}
\epsilon_{\tiny MEC}= \frac{m_N}{m_N^\star} (1+ R)
\end{eqnarray}
where $R$ is the ratio of the exchange current matrix element Fig.\ref{graphs}
to the impulse approximation calculated
with matter-free-space constants. In (\ref{axial}), $R$ is
essentially given by Fig.\ref{graphs}(a) with the bare coupling and
is given by $R\approx 0.5$ with a small variation with density of the
system. Now at nuclear matter density, $m_N^\star/m_N\approx 0.75$, so
\begin{eqnarray}
\epsilon_{\tiny MEC}\approx 2.
\end{eqnarray}
In light nuclei, we expect $\epsilon_{\tiny MEC}\roughly> 1.5$.
These results are in agreement what was found experimentally \cite{warburton}.
One predicts a similar effect in magnetic moments of heavy nuclei but
here one has to include other effects of equal
importance present in the vector current case.
For instance, the ``back-flow" correction due to Galilean
invariance cancels almost completely the corrections coming from
the scaled nucleon mass.
We now turn to an important issue of making a bridge
between the chiral theory and Walecka mean field theory of nuclei
and nuclear matter which is found to be very successful. I wish
to show here, following \cite{br95-1},
that Walecka theory is equivalent to the chiral Lagrangian theory
at mean field {\it with} the BR scaling and that
this would allow a treatment of fluctuations into different
flavor directions (such as strangeness) in a way consistent with
the properties of normal nuclear matter. To do this we can focus on
the four-fermi interactions allowed in chiral Lagrangians that are
relevant in making contact with Walecka theory,
\begin{eqnarray}
{\cal L}_{4f}=\alpha \left(\bar{B}B\right)^2 +\beta \left(\bar{B} v^\mu B\right)^2
\label{4flag}
\end{eqnarray}
where $\alpha$ and $\beta$ are dimension -2 constants and we are using
the heavy-fermion formalism so that $v_\mu$ is the velocity four-vector
of the heavy baryon. Let us imagine that the first term of (\ref{4flag})
arises from integrating out the heavy chiral singlet scalar $\chi$ and
the second term from integrating out a heavy chiral singlet vector meson
$\omega$. We can include other degrees of freedom in a similar way but
we will not need them for symmetric nuclear matter that we shall consider.
In this case, we can identify the constants
\begin{eqnarray}
\alpha=\frac{g_\chi^2}{2m_\chi^2}, \ \ \ \
\beta=-\frac{g_\omega^2}{2m_\omega^2}.
\end{eqnarray}
We now ask what happens to this Lagrangian when it is immersed in dense and/or
hot matter. In mean field, we get the nucleon scalar potential $S_N$ and
vector potential $V_N$ as
\begin{eqnarray}
S_N &=& -\frac{{g_\sigma^\star}^2}{{m_\sigma^\star}^2} \rho_s,\label{SN}\\
V_N &=& \frac{9}{{8f_\pi^\star}^2} \rho\label{VN}
\end{eqnarray}
where $\rho_s$ is the scalar density and $\rho$ the vector density.
In obtaining (\ref{VN}), we have used $SU(3)_f$ relations together with
KSRF relation which is known to hold well and put the stars in (\ref{SN})
and (\ref{VN}) to indicate
that they are in-medium quantities. Now comparing with the phenomenology
with Walecka model, we find that the identification requires that
(\ref{BR}) holds with
\begin{eqnarray}
\frac{f_\pi^\star}{f_\pi}\approx 0.77.
\end{eqnarray}
Together with what we found in the case of the axial charge transition,
we come to the conclusion that the mean field Walecka theory is just the
mean field chiral Lagrangian theory {\it with BR scaling}.
It is a well-known defect of the mean field Walecka model that the
compressibility modulus $K_0$ is much too high in the model. Now
how does this defect get rectified? The answer must lie in higher loop
corrections going beyond the mean field as the scaling is known to fail
to give the nuclear matter saturation \cite{tjon}. This is also seen in
recent work of Furnstahl et al \cite{tang} who note that by giving
an anomalous dimension 2.7 to the scalar field $\sigma$ with the
Lagrangian suitably implemented with the trace anomaly, they can obtain
the low $K_0\approx 200$ MeV and the suppression of the many-body
terms $\sigma^n$, $n >2$. It is plausible that the anomalous dimension
is mocking up the
quantum loop effects that seem to be needed in the mean field approach given
in \cite{br95-1}. A remarkable observation is that at the anomalous
dimension of $d_a\approx 2.7$, two things happen simultaneously.
One is that the $K_0$ which is large at smaller anomalous dimensions
stabilizes at $\sim 200$ MeV for $d_a\approx 2.7$, stays at that value
for higher $d_a$'s and secondly, it is at this fine-tuned value of $d_a$
that {\it all} multi-body forces get suppressed. This clearly calls
for a simple explanation\cite{chaejun}.
The above result immediately suggests how to calculate kaon-nuclear
interactions in consistency with the nuclear matter properties as given
by Walecka theory. To see this, consider a part of the chiral Lagrangian
that figures importantly in the kaon-nuclear sector
\begin{eqnarray}
{\cal L}_{KN}=\frac{-6i}{8f^2}(\overline{B}\gamma_0 B)\overline{K}\partial_t K +
\frac{\Sigma_{KN}}{f^2}(\overline{B}B)\overline{K}K\equiv {\cal L}_\omega
+{\cal L}_\sigma\label{kaonL}
\end{eqnarray}
where $K^T=(K^+ K^0)$. The constant $f$ in (\ref{kaonL}) can be identified
in free space with the pion decay constant $f_\pi$. In medium, however,
it can be modified as we shall see shortly.
In chiral perturbation expansion, the first term corresponds to
${\cal O} (Q)$ and the second term to ${\cal O} (Q^2)$. There is one more
${\cal O} (Q^2)$ term proportional to $\partial_t^2$ which will be taken into
account in the numerical results quoted below
but they are not important except for quantitative details.
One can interpret the first term of (\ref{kaonL})
as arising from integrating out the $\omega$
meson as in the baryon sector. The resulting $K^- N$ vector potential in medium
can then be deduced in the same way as for $V_N$:
\begin{eqnarray}
V_{K^\pm}=\pm\frac{3}{8{f^\star_\pi}^2}\rho.
\end{eqnarray}
Thus in medium, we may set $f\approx f_\pi^\star$ and obtain
\begin{eqnarray}
V_{K^\pm}=\pm\frac 13 V_N.\label{omegascale}
\end{eqnarray}
This just says that the $\omega$ couples to a {\it matter field} kaon, hence
1/3 of the $\omega$ coupling to the nucleon. The reason for this matter-field
nature of the kaon is that all nonstrange hadrons become light in dense
medium, so the kaon becomes in some sense heavy. This dual character is known
from the hyperon structure which is well described by considering the kaon
to be heavy as in the Callan-Klebanov model.
As for the second term of (\ref{kaonL}), we use that the kaon behaves as
a massive matter field. We therefore expect that it be coupled to the chiral
scalar $\chi$ as
\begin{eqnarray}
{\cal L}_\sigma = \frac 13 2 m_K g^\star_\sigma \overline{K}K\chi
\end{eqnarray}
where the factor 1/3 accounts for one non-strange quark in the kaon as compared
with three in the nucleon.
When the $\chi$ field is integrated out as above, we will get,
analogously to the nucleon case,
\begin{eqnarray}
{\cal L}_\sigma = 2m_K \frac 13
\frac{{g^\star_\sigma}^2}{{m^\star_\sigma}^2}\overline{B}B \overline{K}K.
\end{eqnarray}
Comparing with the second term of (\ref{kaonL}), we find
\begin{eqnarray}
\frac{\Sigma_{KN}}{f^2}\approx 2\frac{m_K}{3}
\frac{{g^\star_\sigma}^2}{{m^\star_\sigma}^2}.\label{relation}
\end{eqnarray}
We can get the $\Sigma_{KN}$ from lattice calculations \cite{fukugita},
$\Sigma_{KN}\approx 3.2 m_\pi$. This gives $f\approx f_\pi^\star$.
Therefore we have
\begin{eqnarray}
S_{K^\pm}=\frac 13 S_N.
\end{eqnarray}
{\it To summarize: the kaon-nuclear potential gotten from a chiral Lagrangian
and the nucleon-nuclear potential given by Walecka mean field theory
are directly related through BR scaling.}
Given Walecka mean fields for nucleons, we can now calculate the corresponding
mean-field potential for $K^-$-nuclear interactions in symmetric nuclear
matter.
{}From the results obtained above, we have
\begin{eqnarray}
S_{K^-} +V_{K^-}\approx \frac 13 (S_N-V_N).
\end{eqnarray}
Phenomenology in Walecka mean-field theory gives
$(S_N-V_N)\roughly< -600\ {\mbox{MeV}}$ for $\rho=\rho_0$.
This leads to
the prediction that at nuclear matter density
\begin{eqnarray}
S_{K^-}+V_{K^-}\roughly< -200\ {\mbox{MeV}}.
\end{eqnarray}
This seems to be consistent with the result of the analysis in K-mesic atoms
made by Friedman, Gal and Batty
\cite{friedman} who find attraction at $\rho\approx 0.97\rho_0$ of
\begin{eqnarray}
S_{K^-}+V_{K^-}=-200\pm 20\ {\mbox{MeV}}.
\end{eqnarray}
An immediate consequence of this mean field description of the kaonic
sector is that kaons will condense in dense neutron star (or nuclear star)
matter at a density
\begin{eqnarray}
\rho_c\sim 2 \rho
\end{eqnarray}
as found by Lee et al \cite{LBMR} in ChPT to one-loop order.
\begin{figure}
\vskip -7cm
\centerline{\epsfig{file=ceres2.eps}}
\vskip -7cm
\caption[ceres]{The Li-Ko-Brown explanation of the dilepton data of the
CERES collaboration. The dotted line is the theoretical prediction {\it
without}
the scaling of the $\rho$ and $\omega$ mesons and the solid curve
{\it with} the BR scaling. The three lowest mass points are essentially
given by Dalitz pairs, so the relevant data points are the ones for
higher invariant masses. Note that the peak shown at $M\sim 800$ MeV
is predominantly given by
the $\omega$ decay outside of the medium.}\label{ceres}
\end{figure}
To conclude, I make a few remarks on the nature of the scaling
properties.
\begin{enumerate}
\item There are two points to the issue. One is that {\it some or all}
light hadrons may be undergoing a downward mass shift as
density or temperature is
increased. The second is that the scaling is like (\ref{BR}).
These are basically two different issues. To the extent that hadron
masses are generated spontaneously, it is inevitable that at least some
masses should drop. Indeed QCD sum rule calculations do predict the
drop for the vector-meson mass \cite{qcdsum}, the most recent value
being \cite{jin}
\begin{eqnarray}
m_\rho^\star/m_\rho=0.78\pm 0.08,\ \ \ \rho\approx \rho_0.
\end{eqnarray}
On the other hand, the ``universal scaling" (\ref{BR}), obtained at mean field,
may not be strictly valid. In fact, large $N_c$ arguments imply
$m_N^\star/m_N\approx \sqrt{g_A^\star/g_A} (f_\pi^\star/f_\pi)$. If we look
at the strict $N_c=\infty$ limit, $g_A^\star=g_A$ as we showed in \cite{br91},
the scaling (\ref{BR}) holds but we know that in nuclei, $g_A^\star\approx 1$
and thus in finite nuclei we expect that the nucleon scales somewhat faster
than the pion decay constant, at least up to $\rho\sim \rho_0$.
Whether the ``universal scaling" (\ref{BR}) is consistent
with nature remains to be seen.
\item An intriguing question is how far the scaling (\ref{BR}) can be
pushed in density and/or temperature. Can one use the mean field argument
all the way to the chiral phase transition?
This is a highly relevant question since there are arguments \cite{kocic}
that the second order chiral phase transition relevant to QCD with
2 flavors is of mean field type as in 3D Gross-Neveu model. This suggests that
the mean field chiral theory or Walecka theory could be used to discuss
the QCD chiral phase transition \cite{br95-2}. If correct, this theory
will be very useful for studying phase transitions in heavy ion collisions.
\item There are a large number of experimental projects to measure the
mass shift of hadrons in dense matter, particularly at GSI, CEBAF and
CERN using dilepton-pair production.
Recent data from the CERES collaboration \cite{CERES}
on $e^+ e^-$ pairs in S on Au collisions at 200 GeV/nucleon can be understood
in terms of the scaled mass of the vector mesons in medium, primarily due
to density effect. In a recent paper, Li, Ko and Brown \cite{likobrown}
have shown that the enhancement of the produced lepton pairs observed
in the range of invariant mass $300\ {\mbox{MeV}}\roughly< M \roughly< 550\
{\mbox{MeV}}$ can be explained simply by the BR scaling in the $\rho$
mass as the pairs are produced mainly through $\pi^+ \pi^-\rightarrow
\rho^\star
\rightarrow e^+ e^-$. The fit is given in Fig.\ref{ceres}.
The analysis is a complex one involving the assumption of an expanding
fireball in chemical equilibrium, but
the economy of the explanation and the quantitative success make it
quite compelling.
\end{enumerate}
\subsection*{Acknowledgments}
\indent
I am very grateful for continuing discussions with Gerry Brown with
whom most of the ideas described here have been developed. I
would also like to thank my young collaborators Chang-Hwan Lee, Kurt
Langfeld and Tae-Sun Park for their valuable help. This paper was
completed at the Center for Theoretical Physics (CTP) of Seoul National
University. I would like to thank Dong-Pil Min and the
members of the CTP for hospitality and support.
|
\section{Introduction}
In collisions between two heavy nuclei
at bombarding energies from a few hundred MeV
up to several GeV per nucleon,
hadronic matter at high density and temperature is formed.
In such collisions a large number
of energetic particles are produced
and may be used as probes
of the hot and dense phase of the reaction
\cite{Metag,Cassing,Mosel}.
Microscopic transport models,
such as BUU and QMD \cite{Cassing,Mosel,Wolf,Aichelin}
have been fairly successful
in describing particle production in heavy-ion reactions.
In these transport models,
the nucleons propagate in an effective one-body field
while subject to direct two-body collisions.
Sufficiently energetic nucleon-nucleon collisions may agitate
one or both of the colliding nucleons to a nucleon resonance,
especially $\Delta(1232)$, $N^*(1440)$, and $N^*(1535)$.
Such resonances propagate in their own mean field
and may collide with nucleons or other nucleon resonances as well.
Furthermore, the nucleon resonances may decay by meson emission
and these decay processes constitute the main mechanisms
for the production of energetic mesons \cite{Mosel}.
The transport descriptions normally employ
the vacuum properties of the resonances and mesons,
\ie\ the needed cross sections, decay widths, and dispersion relations
are taken according to their values in vacuum \cite{Wolf}.
However, in infinite nuclear matter,
a system of interacting $\pi$ mesons, nucleons, and $\Delta$ isobars
will couple to form spin-isospin modes.
Some of these modes are non-collective in their character,
dominated by a single baryon-hole excitation,
while other modes are collective
and correspond to meson-like states (quasimesons).
The non-collective spin-isospin modes correspond
to those already included in transport simulations
by promoting a nucleon from below to above the Fermi surface.
The collective spin-isospin modes can effectively
be regarded as particles of mesonic character (quasimesons),
which by means of a local density and temperature approximation
can be incorporated into the transport descriptions.
Some in-medium modifications have already
been employed in calculations of heavy-ion collisions,
both qualitatively \cite{Weise}
and by transport simulations \cite{Bertsch,Giessen,Texas}.
A more elaborate $\pi + N N^{-1} + \Delta N^{-1}$ model was
employed in ref.\ \cite{main} to derive several quantities useful for
implementation of in-medium properties in transport descriptions.
While it is straightforward
to apply the local density approximation
in the interior regions of the nuclear system,
there are conceptual problems of how to proceed
at the nuclear surface where the density approaches zero
and the quasimesons convert
to real physical particles.
The problem is that while in a real system
no hole states exist in vacuum,
a collective ($\Delta N^{-1}$-like) mode
can in a stationary infinite system
exist for an arbitrary small (but finite) density.
This is because the particles in the stationary system
have infinite time to explore the entire system
and form collective modes also at extremely low densities.
In this paper we will therefore discuss
how a proper conversion from quasimesons
to real particles at the surface
can be performed.
Earlier works \cite{Giessen,Texas} have treated this conversion
by various approximations
(see further the discussion in section \ref{sec_Qpions}).
In this work we will present a somewhat different approach,
based on ref.\ \cite{main},
and discuss how the approximations in the earlier works
can be improved upon.
The present paper constitutes a qualitative investigation
of treatment of pionic modes at a nuclear surface
in transport simulations.
To make the presentation simple and transparent
we will therefore restrict ourselves to some special cases
of the more complete investigation in ref.\ \cite{main}.
We will only consider the collective modes
in the spin-longitudinal channel (pion like),
since this is the dominant channel
at the energies we have in mind in this paper.
Collective modes
in the spin-transverse channel ($\rho$ meson like)
can be treated completely analogously.
Furthermore we will consider the zero-temperature case,
and the $\Delta$ width will not be included
in the calculation of the dispersion relations
(denoted as the reference case in ref.\ \cite{main}).
The justification for considering only $T=0$ in this paper
is that there is not a strong dependence
on the dispersion relations of the collective modes
for moderately large temperatures,
especially not at low densities at the surface.
Also, the temperature is not expected to be very high at the surface.
However we want to emphasize
that there is no principal difficulty
associated with incorporating $T>0$ in the treatment.
The motivation for omitting the $\Delta$ width,
in the calculation of the dispersion relations,
is somewhat more involved.
Including the $\Delta$ width self-consistently
in the calculation of the spin-isospin modes
encompasses decay processes like
\[
\tilde{\pi}_j \rightarrow \Delta N^{-1}
\rightarrow (N+\tilde{\pi}_k) N^{-1}\ .
\]
However, since such processes are already explicitly contained
in the transport simulation by processes like
\[
\tilde{\pi}_j + N \rightarrow \Delta
\rightarrow N+\tilde{\pi}_k \ ,
\]
it does not seem to be correct
to include the entire self-consistent $\Delta$ width
when calculating the collective modes
to be used in the transport description.
Instead it seems more correct
to use the results obtained with
the $\Delta$ width omitted,
both for the energies of the modes
and for the partial $\Delta$ widths
to be used in the decays $\Delta \rightarrow N +\tilde{\pi}_j$.
However one should note that
by omitting the $\Delta$ width in the dispersion relations
one also fails to take into account the fact
that the pionic modes
have a Breit-Wigner like energy distribution,
analogous with the $\Delta$.
The width and center of this distribution
are determined by the the self-consistent $\Delta$ width
and depends on the particular pionic mode and its momentum.
The center of the distribution approximately corresponds
to the energy found when the $\Delta$ width is omitted \cite{main}.
In section \ref{sec_Model} we will give
a brief presentation of the model.
The dispersion relations and amplitudes
of the spin-isospin modes obtained in infinite nuclear matter,
are presented and discussed in section \ref{sec_DispAmp}.
Section \ref{sec_Qpions} is devoted to a discussion
of the pionic modes at a nuclear surface
and the implications for transport descriptions,
while our results are summarized in section \ref{sec_Sum}.
In addition we present in appendix \ref{sec_Refl}
a discussion of reflection and transmission properties
at the nuclear surface,
and in appendix \ref{sec_RPAsolu}
some technical details for the RPA equations.
\section{The model} \label{sec_Model}
The model presented in this section
is treated and motivated in detail in ref.\ \cite{main}.
For convenience we here present a brief recapitulation
of the essential points.
Furthermore, the presentation in this section
only treats the spin-longitudinal channel
for the special case
when $T=0$ and the $\Delta$ width is omitted
in the calculation of the spin-isospin modes.
\subsection{Spin-isospin modes in an infinite system} \label{sec_Model-1}
We consider a system of interacting nucleons ($N$), delta isobars
($\Delta$) and pi mesons ($\pi$).
In order to investigate the in-medium properties of the interacting particles,
we employ a cubic box with side length $L$;
the calculated properties are not sensitive to the actual size,
so we need not take the limit $L\to\infty$ explicitly.
The in-medium properties are obtained by using the Green's function technique,
starting from non-interacting hadrons.
The non-interacting Hamiltonian can be written
\begin{equation}
H_0 = \sum_k e_k \hat{b}^{\dagger}_k \hat{b}^{ }_k +
\sum_l \hbar \omega_\pi(\mathbf{q_l})
\hat{\pi}^{\dagger}_l \hat{\pi}^{ }_l\ .
\label{eq_H0}
\end{equation}
Here the index $k=(\mathbf{p}_k; \, s_k, m_{s_k}; \, t_k, m_{t_k})$
represents the baryon momentum, spin, and isospin.
The spin and isospin quantum numbers, $s_k$ and $t_k$,
take the values $1\over2$ and $3\over2$ for $N$ and $\Delta$, respectively.
The energy of baryon $k$ moving in a (spatially constant) potential
is denoted $e_k$.
The baryon creation and annihilation operators,
$\hat{b}^{\dagger}_k$ and $\hat{b}^{ }_k$,
are normalized such that they satisfy the usual anti-commutation relation,
\begin{equation}
\{ \hat{b}^{\dagger}_k, \: \hat{b}^{\mbox{ }}_{k'} \} = \delta_{k,k'}\ .
\end{equation}
In the pion part of $H_0$,
the index $l$ represents the pion momentum and isospin,
$l = (\mathbf{p}_l,\lambda_l=0,\pm~1)$.
The meson energy is given by
$\hbar \omega_{\pi} = [m_{\pi}^2 + \mathbf{q}^2]^{1/2}$
and the creation and annihilation operators of the pion are
normalized such that they satisfy the usual commutation relation,
\begin{equation}
[ \hat{\pi}^{\dagger}_l, \: \hat{\pi}^{ }_{l'} ] = \delta_{l,l'}\ .
\end{equation}
Note that the $\Delta$ isobar described by $H_0$ has no decay width,
$\Gamma_\Delta=0$.
When the interactions are turned on,
the $\Delta$ width will emerge
and it will then automatically include also the free width.
\subsubsection{Basic interactions}
At the $N \pi N$ and $N \pi \Delta$ vertices we will use effective $p$-wave
interactions, $V_{N \pi N}$ and $V_{N \pi \Delta}$, which in the
momentum representation can be written as \cite{main,OTW}
\begin{eqnarray}
&~& V_{N \pi N} = i c \: \frac{(\hbar c)^{1\over2}}{L^3} \:
\left[ \frac{2m_N c^2}{m_N c^2 + \sqrt{s}} \right]^{1\over2}
\frac{f^\pi_{NN}}{m_\pi c^2}\ F_{\pi}(q) \;
(\mathbf{\sigma} \mathbf{\cdot} \mathbf{q}_{cm}) \:
\veC{\tau} \mathbf{\cdot} \Vec{\phi}_\pi(\mathbf{q})
\label{eq_Vnpn}
\\
&~& V_{N \pi \Delta} =
i c \: \frac{(\hbar c)^{1\over2}}{L^3} \:
\left[ \frac{2m_\Delta c^2}{m_\Delta c^2 + \sqrt{s}}\right]^{1\over2}
\frac{f^\pi_{N\Delta}}{m_\pi c^2}\ F_{\pi}(q) \;
(\mathbf{S^+} \mathbf{\cdot} \mathbf{q}_{cm}) \:
\veC{T}^+ \mathbf{\cdot} \Vec{\phi}_\pi(\mathbf{q})
+ {\rm h.c.} \phantom{123} \label{eq_Vnpd}
\end{eqnarray}
In these expressions, $\sqrt{s}$ is the center-of-mass energy
in the $N \pi$ system and $\mathbf{q}_{cm}$
is the pion momentum in the $N \pi$ center-of-mass system,
which in the non-relativistic limit is given by
\begin{equation}
\mathbf{q}_{cm} \approx \frac{m_N c^2}{m_N c^2 + \hbar \omega}\ \mathbf{q}
- \frac{\hbar \omega}
{m_N c^2 + \hbar \omega}\ \mathbf{p}_N\ ,
\label{eq_qcm}
\end{equation}
where $\hbar \omega$ and $\mathbf{q}$ is the pion energy and momentum, and
$\mathbf{p}_N$ is the nucleon momentum in an arbitrary frame.
The Pauli spin
and isospin matrices are denoted $\mathbf{\sigma}$ and $\vec{\tau}$, and
$\mathbf{S^+}$ and $\vec{T}^+$ are spin and isospin $1\over2$ to $3\over2$
transition operators normalized such that
$<{3\over2},{3\over2}|S^+_{+1}|{1\over2},{1\over2}>=1$.\footnote{For clarity,
we generally employ bold-face characters to denote quantities with
vector and tensor properties under ordinary spatial rotations,
while arrows are employed to indicate the transformation properties
under rotations in isospace.}
The momentum representation of the pion field is taken as
\begin{equation}
\phi^\pi_\lambda(\mathbf{q}) =
\frac{L^{3/2} \hbar c}{\sqrt{2 \hbar \omega_\pi(\mathbf{q})}}
\left[ \hat{\pi}^{ }_\lambda(\mathbf{q}) +
(-1)^\lambda \hat{\pi}_{-\lambda}^\dagger(-\mathbf{q}) \right]\ .
\end{equation}
The interactions contain a monopole form factor,
\begin{equation}
F_{\pi}(q) = \frac{\Lambda_\pi^2 - (m_\pi c^2)^2}{\Lambda_\pi^2 - (cq)^2}\ ,
\label{eq_Fpi}
\end{equation}
and the coupling constants are determined at
$(cq)^2 = (\hbar \omega)^2 - (c \, \mathbf{q})^2 = (m_\pi c^2 )^2$
and
$\sqrt{s} = m_N c^2$ or $\sqrt{s} = m_\Delta c^2$.
In addition we will also include effective short-range interactions at
nucleon-hole vertices, again written in momentum space,
\begin{equation}
V_{NN^{-1},NN^{-1}} = \: \left( \frac{\hbar c}{L} \right)^3 \:
g_{NN}' \left( \frac{f^\pi_{NN}}{m_\pi} \right)^2
|F_g(q)|^2\ (\mathbf{\sigma_1 \cdot \sigma_2})
(\veC{\tau_1} \cdot \Vec{\tau_2})\ ,
\label{eq_Vg}
\end{equation}
and the corresponding interactions obtained
when one (or two) of the nucleons is replaced by a $\Delta$.
The strength of the short-range interactions is determined by the
correlation parameters $g_{NN}'$, $g_{N\Delta}'$, and $g_{\Delta \Delta}'$.
\subsection{RPA approximation}
We want to calculate a spin-isospin mode Green's function
within the RPA approximation, symbolically
\begin{equation}
G^{\rm RPA}(\alpha,\beta;\omega) = G_{0}(\alpha,\beta;\omega) +
\sum_{\gamma,\kappa} G_{0}(\alpha,\gamma;\omega)\
{\cal V}(\gamma,\kappa;\omega)\
G^{\rm RPA}(\kappa,\beta;\omega)\ .
\label{eq_GRPA}
\end{equation}
The spin-isospin modes,
here represented by the Green's function $G^{\rm RPA}$,
will in this approximation
be obtained as an infinite iteration of
(non-interacting) pion, nucleon-hole, and $\Delta$-hole states,
represented by the diagonal Green's function $G_0$,
coupled with the interactions specified
in eqs.\ (\ref{eq_Vnpn}--\ref{eq_Vg})
which here are summarized by the symbolic interaction ${\cal V}$.
In nuclear collisions
at beam energies up to about one GeV per nucleon,
which is the domain of application that we have in mind,
only relatively few mesons and isobars are produced
and so the associated quantum-statistical effects may be ignored.
Accordingly, we assume $n_\Delta \ll 1$, and $n_\pi \ll 1$.
A set of RPA equations,
equivalent to eq.\ (\ref{eq_GRPA})
were derived in ref.\ \cite{main}.
{}From these equations eigenvectors and
eigenenergies are obtained
for the different spin-isospin modes.
The eigenvectors will yield the amplitudes
of the different components
($\pi$, $NN^{-1}$, $\Delta N^{-1}$)
forming the particular spin-isospin mode
with the given eigenenergy.
These RPA amplitudes contain important information
about the nature of the different spin-isospin modes.
The spin-isospin modes (or excited RPA states), $|\Psi_\nu>$,
are created by an operator
$Q^{\dagger}_\nu$,
\begin{equation}
Q^{\dagger}_\nu(\mathbf{q},\lambda) =
\sum_{jk} X^\nu_{jk}(\mathbf{q},\lambda)
\hat{b}^{\dagger}_j \hat{b}^{ }_k +
\sum_{k} Z^\nu_k(\mathbf{q},\lambda) \, \hat{\pi}_k^\dagger -
\sum_{k} W^\nu_k(\mathbf{q},\lambda) \, \hat{\pi}^{ }_k\ .
\label{eq_Qrpal}
\end{equation}
The quantity
$X^\nu_{jk}(\mathbf{q},\lambda)$
is here the amplitude of the baryon-hole
($N N^{-1}$ or $\Delta N^{-1}$)
component of the spin-isospin mode $|\Psi_\nu>$
at momentum $\mathbf{q}$ and isospin $\lambda$,
while
$Z^\nu_k(\mathbf{q},\lambda)$ and $W^\nu_k(\mathbf{q},\lambda)$,
in the same way, are the amplitudes of the pionic component.
The summation over baryon and meson states
in eq.\ (\ref{eq_Qrpal}) is restricted by taking
$X_{jk} \propto \delta_{\mathbf{p}_j, \mathbf{p}_k + \mathbf{q}}$,
$Z_k \propto \delta_{\mathbf{p}_k, \mathbf{q}} \delta_{\lambda_k, \lambda}$,
and
$W_k \propto \delta_{\mathbf{p}_k, -\mathbf{q}} \delta_{\lambda_k, -\lambda}$.
The RPA equations are obtained from the relation
\begin{equation}
<[\delta Q,[H,Q^{\dagger}]]> = \hbar \omega <[\delta Q,Q^{\dagger}]>\ ,
\label{eq_RPA-gen}
\end{equation}
with
$\delta Q=\hat{b}^{\dagger}_k\hat{b}^{ }_j$, $\hat{\pi}_r$, or
$\hat{\pi}_r^\dagger$,
and where the brackets $<\cdot>$ denote the expectation value
in the interacting ground state.
It can be shown that the set of RPA solutions
constitutes an orthonormal set.
For convenience the solutions
of the RPA equations of ref.\ \cite{main},
are recapitulated in appendix \ref{sec_RPAsolu}.
\subsubsection{The total $\Delta$ width}
The $\Delta$ self energy $\Sigma_\Delta$ is calculated according to the
diagrams in fig.\ \ref{fig_DseGraph}, by taking into account all the diagrams
corresponding to the $\Delta$ decaying into a spin-isospin mode and a nucleon,
which then again form a $\Delta$.
In the spin-longitudinal channel we obtain (ref.\ \cite{main})
\begin{equation}
\Gamma^l_\Delta(E_\Delta,\mathbf{p}_\Delta) = \mbox{Im } \frac{2}{3}
\left( \frac{\hbar c}{L} \right)^3
\sum_{\mathbf{q}} \,
\left[ \theta({\cal E}) - n(\mathbf{p}_\Delta-\mathbf{q}) \right]
\bar{M}(\Delta N,N \Delta)\ .
\label{eq_DSE2}
\end{equation}
where the energy available for the spin-isospin mode is given by
\begin{equation}
{\cal E} = E_\Delta - e_N(\mathbf{p}_\Delta-\mathbf{q})\ ,
\end{equation}
and
$\bar{M}(34,12)$ can be expressed as
\begin{equation}
\bar{M}(34,12)
=
\sum_{\omega_\nu > 0}
\left\{ \frac{ h(31;\nu) h(24;\nu) }
{ \hbar \omega - \hbar \omega_\nu + i \eta }
- \frac{ h(31;\nu) h(24;\nu) }
{ \hbar \omega + \hbar \omega_\nu - i \eta }
\right\}
+
\frac{ f^\pi_{31} f^\pi_{24} }{(m_\pi c^2)^2} F_g^2 g'_{34,12} \ .
\label{eq_Mrpa}
\end{equation}
The factor $h(jk,\nu)$ is obtained
from the interactions at the vertex
consisting of baryons $j$ and $k$,
and the spin-isospin mode $\nu$.
The interactions to be used depend
on the non-interacting states
that the mode consists of
and must therefore be multiplied
by the amplitude of the corresponding state,
\begin{equation}
h(jk,\nu) \vartheta^l(jk) = \frac{V_{j \pi k}}{\sqrt{2 \hbar \omega_\pi}}
\, [Z(\nu)+W(\nu)] +
\sum_{mn} V_{jk,mn} \, X_{mn}(\nu) \ ,
\label{eq_hMotiv}
\end{equation}
where $\vartheta^l(jk)$ is a short hand notation
for the spin-isospin matrix elements
in the spin-longitudinal channel,
$V_{j \pi k}$ is defined in (\ref{eq_Vnpn}) and (\ref{eq_Vnpd}),
$V_{jk,mn}$ is defined in (\ref{eq_Vg}),
and the amplitudes $X^l_{mn}$, $Z^l$, $W^l$
are defined in eq.\ (\ref{eq_Qrpal}).
The explicit expression for $h(jk,\nu)$
is somewhat lengthy
and has therefore also been relegated to appendix \ref{sec_RPAsolu}.
\subsubsection{Specific $\Delta$ channels}
The total $\Delta$ width
gives the transition probability per unit time
for the $\Delta$ resonance to
decay to any of its decay channels.
In a transport description one explicitly allows
the $\Delta$ resonance to decay
into specific final particles.
Consequently, one needs not only the total $\Delta$ width
(which is the sum of all decay channels)
but also the partial widths governing the decay
into specific RPA channels.
These decay channels consist
of a nucleon and one of the spin-isospin modes.
Since we have access to all the amplitudes
of a given spin-isospin mode
on the different unperturbed states,
it is possible to derive an expression
for the partial contribution to
$\Gamma_\Delta$ from the $\Delta$ decay to a specific mode $\nu$.
The right-hand side of fig.\ \ref{fig_DseGraph}
shows a diagrammatic representation of such a process.
The partial $\Delta$ width for a $\Delta$ decay to a nucleon
and a spin-longitudinal mode $\nu$ becomes \cite{main}
\begin{eqnarray}
\tilde{\Gamma}_\Delta^\nu(E_\Delta,\mathbf{p}_\Delta)
& = & \int \frac{d^3p_N}{(2\pi)^3} \,
\frac{d^3q}{(2\pi)^3} \;
| \frac{V_{\Delta \pi N}}{\sqrt{2 \hbar \omega_\pi}}
\cdot [Z(\nu)+W(\nu)] +
\sum_{mn} V_{\Delta N,mn} \cdot X_{mn}(\nu)|^2
\nonumber \\ & ~& \times
\bar{n}_N(\mathbf{p}_N) \,
(2\pi)^3 \delta( \mathbf{p}_\Delta - \mathbf{p}_N - \mathbf{q} )
2\pi \delta( E_\Delta - e_N - \hbar \omega ) \nonumber \\
& = & \frac{1}{3} \int \frac{d^3q}{(2\pi)^3} \:
| h(\Delta N,\omega_\nu) |^2 \,
\bar{n}_N(\mathbf{p}_\Delta - \mathbf{q}) \,
2\pi \delta( E_\Delta - e_N - \hbar \omega_\nu )
\label{eq_GnuReal}
\end{eqnarray}
where the factor, $\bar{n}_N = 1-n_N$,
takes into account the Pauli blocking of the nucleon.
Note that when this expression is to be used in transport models
the factor $\bar{n}_N$ should be omitted
since the Pauli blocking is treated explicitly
in the transport description.
The expression (\ref{eq_GnuReal}) is identical
to the contribution from one
of the $\nu$ terms in eq.\ (\ref{eq_DSE2}).
\section{ Dispersion relations and amplitudes }
\label{sec_DispAmp}
{}From eq.\ (\ref{eq_Xx}) in appendix \ref{sec_RPAsolu}
we calculate the energies
of the spin-isospin modes that are formed in the interacting system,
\ie\ their dispersion relations.
Fig.\ \ref{fig_DispT0r10} displays the dispersion relations
at normal nuclear density,
$\rho = \rho^0 = 0.153\ \fm^{-3}$.
In fig.\ \ref{fig_DispT0r10} a number of different modes
in the spin-longitudinal ($\pi$-like) channel are apparent.
Some of those are non-collective $NN^{-1}$ modes (solid curves),
which have their energies within the regions
\begin{eqnarray}
0 & \leq & \hbar \omega\
\leq\ \frac{q^2}{2 m_N^*} + \frac{q p_F}{m_N^*}\ ,
\qquad q < 2 p_F \ , \nonumber \\
\frac{q^2}{2 m_N^*} - \frac{q p_F}{m_N^*}
& \leq & \hbar \omega\
\leq\ \frac{q^2}{2 m_N^*} + \frac{q p_F}{m_N^*}\ ,
\qquad q > 2 p_F\ .
\label{eq_NhCont}
\end{eqnarray}
Since we are presenting our results for a box normalization
with a finite side length $L$,
we obtain a discrete number of non-collective $NN^{-1}$ modes.
The total number of spin-isospin modes
within the region (\ref{eq_NhCont})
depends on $L$ and tends towards a continuum
in the limit $L \rightarrow \infty$.
Similarly, a number of
non-collective $\Delta N^{-1}$ states emerge
in fig.\ \ref{fig_DispT0r10} which,
for a fixed $\mathbf{q}$, have their energies constrained to a band,
\begin{equation}
m_\Delta - m_N + \frac{q^2}{2 m_\Delta} - \frac{q p_F}{m_\Delta}\
\leq\ \hbar\omega\
\leq\ m_\Delta - m_N + \frac{q^2}{2 m_\Delta} +
\frac{q p_F}{m_\Delta}\ .
\label{eq_DhCont}
\end{equation}
The non-collective baryon-hole modes
correspond in a transport description
to propagation of uncoupled baryons ($N$ or $\Delta$).
This was discussed and studied in detail
in ref.\ \cite{main} and will therefore
not be further discussed in this paper.
In addition, two collective modes
appear in fig.\ \ref{fig_DispT0r10},
represented by dot-dashed curves.
The lower one
starts at $\hbar \omega = m_\pi c^2$ at $q = 0$
and continues into the $\Delta N^{-1}$ region
around $q \approx 360\ \MeV/c$.
This mode will in the following be referred to as $\tilde{\pi}_1$.
The upper collective mode
starts slightly above
$\hbar\omega \approx m_\Delta c^2 - m_N c^2$ at $q = 0$
and approaches
$\hbar\omega_\pi = [(m_\pi c^2)^2 + (cq)^2]^{1/2}$
at large $q$.
This mode is denoted $\tilde{\pi}_2$.
The incorporation of the two collective spin-isospin modes
into transport equations is more involved.
These modes can be regarded as separate particles of pionic character,
$\tilde{\pi}_1$ and $\tilde{\pi}_2$,
and treated in a manner analogous
to the standard treatment of the pion.
Since the pion is then fully included in the description,
it should no longer be treated explicitly.
The propagation of the two collective pionic modes
is governed by the effective Hamiltonians
\begin{eqnarray}
\tilde{H}_1(\mathbf{r},\mathbf{q}) & = &
\hbar \omega_{1}(\mathbf{q};\rho(\mathbf{r}))\ \equiv\
\hbar \tilde{\omega}_1\ ,
\nonumber \\
\tilde{H}_2(\mathbf{r},\mathbf{q}) & = &
\hbar \omega_{2}(\mathbf{q};\rho(\mathbf{r}))\ \equiv\
\hbar \tilde{\omega}_2\ ,
\label{eq_Hqpion}
\end{eqnarray}
where $\hbar \omega_{1}$ and $\hbar \omega_{2}$
are the energy-momentum relations
for the lower and upper collective modes
displayed in fig.\ \ref{fig_DispT0r10} for $\rho=\rho^0$.
Note that the spatial dependence of $\tilde{H}_1(\mathbf{r},\mathbf{q})$
is incorporated by representing $\rho(\mathbf{r})$ as a local quantity.
To facilitate center-of-mass transformations we parametrize
the dispersion relations of the pionic modes in the form
\begin{equation}
\hbar \tilde{\omega}(\mathbf{q};\rho)
\approx
\left\{ [c \, q - c \, q_0(\rho)]^2 + m_0(\rho)^2 c^4
\right\}^{\frac{1}{2}} + U_0(\rho)
\label{eq_DispParam}
\end{equation}
For convenience in the transport simulations,
we have chosen to use relatively simple expressions
for the parametrization,
rather than to try to optimize the fit.
In this way a quasipion moves like a relativistic particle
with the group velocity determined
by an effective momentum and energy
\begin{equation}
\frac{d \hbar \tilde{\omega}}{dq}
=
c \frac{c(q-q_0)}{\hbar \tilde{\omega} -U_0}
=
c \frac{c q^*}{\hbar \tilde{\omega}^*}\ .
\end{equation}
The density-dependent parameters $q_0$, $m_0$ and $U_0$
are presented in fig.\ \ref{fig_DispParam}.
Furthermore, in the collision term of the standard transport description
the process for the production and absorption of pions
$\Delta \leftrightarrow N + \pi$,
should be replaced by the two distinct processes
\begin{equation} \Delta
\leftrightarrow N + \tilde{\pi}_1 \quad \mbox{ and } \quad \Delta
\leftrightarrow N + \tilde{\pi}_2\ .
\label{eq_Ddecay}
\end{equation}
The $\Delta$ decay is governed
by the $\Delta$ decay width in the medium
to these two specific channels,
$\tilde{\Gamma}_\Delta^j$.
These partial widths,
should be employed in the same manner as the free width,
\ie\ they describe the probability for the $\Delta$ isobar
to decay into a nucleon and a pion.
The only difference is that several collective pionic modes
are available in the final state.
In fig.\ \ref{fig_GamParNPB} we present total and partial
$\Delta$ widths for a $\Delta$ with the momentum 300 MeV/$c$.
The reverse processes in (\ref{eq_Ddecay}) are characterized by
cross sections that were presented and discussed in ref.\ \cite{main}.
To obtain the partial $\Delta$ width from eq.\ (\ref{eq_GnuReal})
it is necessary to know the amplitudes
$Z$, $\sum X_{\Delta N^{-1}}$ and $\sum X_{N N^{-1}}$.
We therefore also present a parametrization of these quantities.
On the lower pionic mode the pionic component
dominates for small momenta, $q$,
and the $\Delta N^{-1}$ component dominates at larger momenta.
Therefore the sum of all individual $\Delta N^{-1}$ components
will for small $q$ increase with $q$.
However, when the lower pionic mode
enters the $\Delta N^{-1}$ region
the collectivity disappears gradually,
and the sum of all individual $\Delta N^{-1}$ components
starts to decrease with $q$.
Thus we employ the form
\begin{eqnarray}
\sum X^{\tilde{\pi}_1}_{\Delta N^{-1}}(\mathbf{q};\rho)
& \approx &
f_\Delta(q,\, \vec{C}[X^{\tilde{\pi}_1}_{\Delta N^{-1}}])\ ,
\label{eq_XDh-low-parm}
\\
\sum X^{\tilde{\pi}_1}_{N N^{-1}}(\mathbf{q};\rho)
& \approx &
f_N(q,\, \vec{C}[X^{\tilde{\pi}_1}_{ N N^{-1}}])\ ,
\label{eq_XNh-low-parm}
\end{eqnarray}
with
\begin{equation}
f_{\stackrel{\scriptstyle \Delta}{\scriptstyle N} }(q,\vec{C})
=
\frac{ \pm C_1^2 + C_2 \, cq }{ C_3^2 + (c q - C_4)^2 }\ .
\label{eq_f1(q,rho)-parm}
\end{equation}
The density-dependent coefficients $\vec{C}$
are presented in fig.\ \ref{fig_AmplParam}.
On the upper pionic mode we instead parametrize the amplitudes as
\begin{eqnarray}
\sum X^{\tilde{\pi}_2}_{\Delta N^{-1}}(\mathbf{q};\rho)
& \approx &
C_N'[X^{\tilde{\pi}_2}_{ \Delta N^{-1}}]
\sqrt{ f_2(q,\, \vec{C}'[X^{\tilde{\pi}_2}_{\Delta N^{-1}}]) }
\label{eq_XDh-upp-parm}
\\
\sum X^{\tilde{\pi}_2}_{ N N^{-1}}(\mathbf{q};\rho)
& \approx &
C_N'[X^{\tilde{\pi}_2}_{ N N^{-1}}]
\sqrt{ 1- f_2(q,\, \vec{C}'[X^{\tilde{\pi}_2}_{ N N^{-1}}]) }\ ,
\label{eq_XNh-upp-parm}
\end{eqnarray}
where
\begin{equation}
f_2(q,\vec{C}')
=
\left[ 1 + \exp \left( C_0' + C_1' \, cq + C_{-1}' / cq \right)
\right]^{-1}\ ,
\label{eq_f2(q,rho)-parm}
\end{equation}
with the density-dependent coefficients $\vec{C}'$,
displayed in fig.\ \ref{fig_AmpuParam}.
The amplitudes $Z$ of the pion component are obtained from the parametrization
of the squared amplitudes, eqs.\ (\ref{eq_Zl2-parm}) and (\ref{eq_Zu2-parm})
below, as $Z = \sqrt{Z^2}$.
The total $\Delta$ decay width,
has apart from the partial contributions
$\tilde{\Gamma}_\Delta^j$,
also the partial contributions
$\Gamma_\Delta^{N N^{-1}}$ and
$\Gamma_\Delta^{\Delta N^{-1}}$.
The partial width $\Gamma_\Delta^{N N^{-1}}$
gives the probability for the $\Delta$
to decay into a nucleon and a $N N^{-1}$ state.
In a transport description,
this implies that we initially have a $\Delta$
and after the decay process
we have two nucleons above the Fermi surface
and a hole left in the Fermi sea.
But this is the same process
as if the $\Delta$ would collide with a nucleon
below the Fermi surface to give two nucleons above the Fermi surface.
This process is normally already included
in the collision term in a standard transport description,
and the probability for such a collision is given by the
cross section for the process
$\Delta + N \rightarrow N + N$.
In a transport description it is therefore not correct
to both include a $\Delta$ decay according to
$\Gamma_\Delta^{N N^{-1}}$
and a collision term with
$\Delta + N \rightarrow N + N$.
Instead, the correct procedure should be to exclude
$\Gamma_\Delta^{N N^{-1}}$
and modify the cross section
$\sigma(\Delta + N \rightarrow N + N)$
to be the in-medium cross section.
Calculations of such in-medium cross sections was discussed
in ref.\ \cite{main}.
In the same way,
$\Gamma_\Delta^{\Delta N^{-1}}$
should be excluded in a transport description,
and $\sigma(\Delta + N \rightarrow \Delta + N)$
be the in-medium cross section.
Although the collective pionic modes
can thus be effectively treated as ordinary particles,
the fact that their wave functions contain components from
$\pi$, $NN^{-1}$ and $\Delta N^{-1}$ states
makes it difficult to picture them
in a physically simple manner.
Fortunately, their specific structure
is not important for the transport process,
as as long as these quasiparticles remain
well inside the nuclear medium.
First when such a quasiparticle penetrates a nuclear surface
and emerges as a free particle
is it physically meaningful to determine
what kind of real particle it is.
The gradual transformation of the collective quasiparticle
occurs automatically within the formalism,
because as the density is lowered,
$\rho \rightarrow 0$,
a pionic mode will acquire 100\%
of either the pion component or the $\Delta N^{-1}$ component,
depending on $\omega$ and $q$.
That is to say, it will turn into either a free pion
or an unperturbed $\Delta N^{-1}$ state.
The squared amplitudes have their values between zero and unity and we
therefore employ the parametrizations
\begin{eqnarray}
Z^{\tilde{\pi}_1}(\mathbf{q};\rho)^2
& \approx &
f_2(q,\, \vec{C}''[Z^{\tilde{\pi}_1}]),
\label{eq_Zl2-parm}
\\
Z^{\tilde{\pi}_2}(\mathbf{q};\rho)^2
& \approx &
1-f_2(q,\, \vec{C}''[Z^{\tilde{\pi}_2}])
\label{eq_Zu2-parm}
\\
\sum X^{\tilde{\pi}_1}_{\Delta N^{-1}}(\mathbf{q};\rho)^2
& \approx &
1-f_2(q,\, \vec{C}''[X^{\tilde{\pi}_1}_{\Delta N^{-1}}])
\label{eq_XDhl2-parm}
\\
\sum X^{\tilde{\pi}_2}_{\Delta N^{-1}}(\mathbf{q};\rho)^2
& \approx &
f_2(q,\, \vec{C}''[X^{\tilde{\pi}_2}_{\Delta N^{-1}}])\ ,
\label{eq_XDhu2-parm}
\end{eqnarray}
with $f_2(q,\vec{C}'')$ from eq.\ (\ref{eq_f2(q,rho)-parm}),
while the sum of the squared $N N^{-1}$ amplitudes
are obtained from the normalization,
\ie\ all squared amplitudes sum up to unity.
The density-dependent coefficients $\vec{C}''$,
are presented in fig.\ \ref{fig_Amp2Param}.
There remains the practical problem of how to represent
an unperturbed $\Delta N^{-1}$ state when $\rho \rightarrow0$.
However, as will be discussed in section \ref{sec_Qpions},
only a very small fraction of the pionic modes
(vanishing for a stationary density profile)
will emerge as unperturbed $\Delta N^{-1}$ states.
Note that this approach is different
from earlier works \cite{Giessen,Texas,KochPriv}
where the nature of the pionic mode
was determined already in the creation process,
\ie\ when the mode was created it was determined whether it represented
a free pion or an unperturbed $\Delta N^{-1}$ state.
This difference in approach will have crucial effects
for the collective modes that escape the system,
as will be seen in the next sections.
\section{ Quasipions at the nuclear surface }
\label{sec_Qpions}
In previous works the quasipions at a surface have been treated
in various approximations.
In ref.\ \cite{Giessen} effective dispersion relations were introduced,
corresponding to modes with either 100\% pionic
or $\Delta N^{-1}$ component.
The pionic mode was then propagated as a quasipion,
emerging as a free pion at the surface,
while the $\Delta N^{-1}$ mode was used
to derive a $\Delta$ potential for the uncoupled $\Delta$s.
In this way only pions and $\Delta$s escape the system,
but the drawback is that the effective dispersion relations
are quite distorted compared to the original ones,
and that the in-medium effects seem to be over-estimated
by allowing all uncoupled $\Delta$s propagate
in a collective potential.
In \cite{Texas} both collective modes were propagated,
and hence some modes escape the system
as unperturbed $\Delta N^{-1}$ states.
This was effectively taken care of
by converting these modes to free $\Delta$s
(neglecting baryon number conservation),
with the justification that the number of modes escaping the system
as unperturbed $\Delta N^{-1}$ states were found to be small.
We will in this paper report on alternative ways
to treat the modes at the surface
based on ref.\ \cite{main},
and how this treatment can improve
the descriptions in refs.\ \cite{Giessen,Texas}.
We first consider the simplified case
when the nuclear surface is stationary
\ie\ $\rho(\mathbf{r},t) = \rho(\mathbf{r},0)$ for all times $t$.
This means that there is no explicit time dependence
in the effective Hamiltonians in equation (\ref{eq_Hqpion}),
and thus the energies of the collective pionic modes are conserved.
For a pionic mode with energy
$\hbar \omega_\nu(\mathbf{q};\rho)$
propagating in a varying density
this means that the momentum
$\mathbf{q}$ will change as $\rho$ changes.
That is to say, the pionic mode effectively feels a potential.
In addition to momentum changes,
also the amplitudes
$X^\nu_{jk}(\mathbf{q};\rho)$,
$Z^\nu(\mathbf{q};\rho)$ and
$W^\nu(\mathbf{q};\rho)$
of the baryon-hole and pion components
will change as the density changes.
In the limit of vanishing density either
$X^\nu_{\Delta N^{-1}}$ or $Z^\nu$ will turn to unity, depending on
the mode $\nu$ and its energy and momentum.
In the latter case no problems arise,
the collective mode has simply been converted
to a free pion escaping the system.
However, in the former case
there is some inconsistency in the formalism
since there are no hole states in vacuum ($\rho=0$).
In a quantal description
of spin-isospin modes propagating at a surface
different scenarios could emerge.
There is some small probability that the mode
could be reflected at the surface.
Alternatively the mode could break up
in an uncoupled $\Delta$ escaping the system,
with the hole is trapped inside the nucleus.
In a transport description this could be handled
by allowing the test particle representing the mode $\nu$
to absorb a nearby nucleon,
converting it into a $\Delta$ isobar.
We consider in this section the quasipions
at a surface propagating from normal nuclear density to vacuum.
In this case the amplitude $W^\nu$ will be very small
and we will therefore omit it in the qualitative discussion
of this section.
In fig.\ \ref{fig_w12-rho}$a$ we present the energy
$\hbar \tilde{\omega}_1(\mathbf{q})$
of the lower collective mode $\tilde{\pi}_1$
for different densities in the range
$0.1 \rho^0 \leq \rho \leq \rho^0$.
The line closest to the free pion relation (dotted line)
represents the dispersion relation at the lowest density.
As the density is increased
the energy relation is (for each fixed $q$) lowered.
For small $q$ values the mode is completely dominated
by the pion component.
As $q$ is increased also the $\Delta N^{-1}$ components
starts to contribute substantially
(although to different extent depending on the density).
In the approximate range
$300\ \MeV/c \leq q \leq 400\ \MeV/c$
(depending on density)
the mode enters the $\Delta N^{-1}$ continuum
and changes character from collective mode to non-collective.
We will in this section discuss four different examples,
suitably chosen to illustrate the main features.
\subsection{ Lower pionic mode }
\label{sec_QpionsLow}
In our first example we consider the mode $\tilde{\pi}_1$
created at normal density
with energy $\hbar \tilde{\omega}_1 = 200\ \MeV$
propagating towards vacuum
without any interactions.
Initially this mode has the following characteristics:
\[
\begin{array}{cccc}
q \approx 220\ \MeV/c
&
|Z|^2 \approx 0.72
&
\sum_{\Delta N^{-1}} |X_{\Delta N^{-1}}|^2 \approx 0.25
&
\sum_{N N^{-1}} |X_{N N^{-1}}|^2 \approx 0.03\ .
\end{array}
\]
As the density decreases it will,
due to the energy conservation,
follow the path indicated
by the dashed line in fig.\ \ref{fig_w12-rho}$a$,
from $q \approx 220\ \MeV/c$ to $q \approx 150\ \MeV/c$.
In fig.\ \ref{fig_Amp-w}$a$ we see how the squared amplitudes
vary with the density for this particular energy
of the mode $\tilde{\pi}_1$.
As seen in fig.\ \ref{fig_Amp-w}$a$
only the pion component remains,
as the zero density limit is reached.
Thus in this particular case,
the mode will escape the system as a free pion.
In fig.\ \ref{fig_DhAmp-w}$a$
we show the squared amplitudes
of the individual $\Delta N^{-1}$ amplitudes
for the energy $\hbar \tilde{\omega}_1 = 200\ \MeV$
of the mode $\tilde{\pi}_1$.
We see that the mode is collective
at all densities.
In the discussion so far
we have not addressed the creation process.
How probable is it that we produce the mode $\tilde{\pi}_1$
at the particular energy $\hbar \tilde{\omega}_1 = 200\ \MeV$?
This information is given
by the partial $\Delta$ decay width
to the mode $\tilde{\pi}_1$.
In fig.\ \ref{fig_Gam_w-rho}$a$ we display
the partial $\Delta$ decay width,
$\tilde{\Gamma}_\Delta^1$,
for different densities,
as a function of the energy
of the emitted quasipion $\tilde{\pi}_1$.
Note that the width displayed
in fig.\ \ref{fig_Gam_w-rho}$a$
is for a $\Delta$ at rest,
and that the Pauli blocking of the emitted nucleon
has not been taken into account
(the Pauli blocking is treated explicitly in a transport description).
As seen in fig.\ \ref{fig_Gam_w-rho}$a$
the width is quite substantial
at the quasipion energy $\hbar \tilde{\omega}_1 = 200\ \MeV$,
about $80\ \MeV$ compared to the free width,
which is about $30\ \MeV$ at this energy.
Thus it is quite probable for a $\Delta$ to decay
to the mode $\tilde{\pi}_1$ around this quasipion energy.
Apart from penetrating the surface
there is also some probability
for the mode $\tilde{\pi}_1$
to be reflected at the surface.
This is difficult to exactly predict
since the reflection coefficient, ${\cal R}$,
will depend on the actual density profile.
However a first estimate can be obtained
by considering a one-dimensional scenario
with a density profile
that corresponds to a Wood-Saxon potential
with a surface thickness $a=0.65$ fm.
The height of the potential is then given
by the change in the momentum of the pionic mode,
and for the case of $\hbar \tilde{\omega}_1 = 200\ \MeV$
we obtain ${\cal R} \approx 0.017$,
see further the discussion in appendix \ref{sec_Refl}.
This number is very small,
which implies that the reflection
can practically be neglected for this particular case.
However, in other situations
where the momentum change is different or
the density profile is sharper,
the reflection coefficient may be larger.
We have therefore devoted appendix \ref{sec_Refl}
to discuss how effective local reflection coefficients
can be obtained and implemented in transport descriptions.
In our second example we also consider the mode $\tilde{\pi}_1$,
but now with energy $\hbar \tilde{\omega}_1 = 295\ \MeV$.
Initially this mode will now have the characteristics:
\[
\begin{array}{cccc}
q \approx 450\ \MeV/c
&
|Z|^2 \approx 0.0
&
\sum_{\Delta N^{-1}} |X_{\Delta N^{-1}}|^2 \approx 0.94
&
\sum_{N N^{-1}} |X_{N N^{-1}}|^2 \approx 0.06\ .
\end{array}
\]
As the density decreases it
follows the path indicated
by the dot-dashed line in fig.\ \ref{fig_w12-rho}$a$,
from $q \approx 450\ \MeV/c$ to $q \approx 250\ \MeV/c$.
In fig.\ \ref{fig_Amp-w}$b$ we see how the squared amplitudes
vary with the density for this particular energy
of the mode $\tilde{\pi}_1$.
As seen in fig.\ \ref{fig_Amp-w}$b$
also in this case only the pion component remains,
as the zero density limit is reached,
although $\tilde{\pi}_1$ initially was completely
dominated by the $\Delta N^{-1}$ component.
Thus also in this case
the mode will escape the system as a free pion.
In fig.\ \ref{fig_DhAmp-w}$b$
we show the squared amplitudes
of the individual $\Delta N^{-1}$ amplitudes
for the energy $\hbar \tilde{\omega}_1 = 295\ \MeV$
of the mode $\tilde{\pi}_1$.
Here the situation is different from the case at
$\hbar \tilde{\omega}_1 = 200\ \MeV$, since initially
the mode is dominated by a single $\Delta N^{-1}$ component,
\ie\ the mode is non-collective.
However as the density is lowered
the strength is spread over more $\Delta N^{-1}$ components
and at low densities the mode is completely collective.
{}From fig.\ \ref{fig_Gam_w-rho}$a$ we see that the partial width
$\tilde{\Gamma}_\Delta^1$ is very close to zero
at the quasipion energy
$\hbar \tilde{\omega}_1 = 295\ \MeV$ (dot-dashed curve)
at normal nuclear density.
Thus a $\Delta$ at normal nuclear density
will not decay to a quasipion with this energy.
The partial $\Delta$ width becomes very small
because the collective strength,
as well as the pion component,
is negligible in this case.
Although the mode at this energy
cannot be created at normal nuclear density,
it may still be created at lower densities,
as can be seen in fig.\ \ref{fig_Gam_w-rho}$a$.
Making the same assumptions as in the first example,
we find that the reflection coefficient,
becomes smaller than $10^{-4}$ for this case.
We thus conclude that the lower pionic mode
penetrating the surface will always emerge as a free pion.
For low energies this is natural
because as the density decreases
the dispersion relation of the pionic mode
approaches the free pion relation.
The cases of sufficiently large energy
to approach the unperturbed $\Delta N^{-1}$ branch
as the density is lowered
will never occur since no such modes will be created
from a decaying $\Delta$,
because the partial $\Delta$ width will be zero.
\subsection{ Upper pionic mode }
\label{sec_QpionsUpp}
Also on the upper collective mode,
$\tilde{\pi}_2$ a similar effect will occur.
However some properties are somewhat different
so we will therefore illustrate
also this case with two typical examples.
In our third example we thus consider the mode $\tilde{\pi}_2$,
with energy $\hbar \tilde{\omega}_2 = 320\ \MeV$.
Note that at this energy the mode can only exist
at densities up to about $0.5 \rho^0$.
Initially, at $\rho = 0.5 \rho^0$,
this mode will have the characteristics:
\[
\begin{array}{cccc}
q \approx 0\ \MeV/c
&
|Z|^2 \approx 0.0
&
\sum_{\Delta N^{-1}} |X_{\Delta N^{-1}}|^2 \approx 1.0
&
\sum_{N N^{-1}} |X_{N N^{-1}}|^2 \approx 0.0\ .
\end{array}
\]
As the density decreases we
follow the path indicated
by the dashed line in fig.\ \ref{fig_w12-rho}$b$,
from $q \approx 0\ \MeV/c$ to $q \approx 270\ \MeV/c$.
Note that on the upper collective mode the momentum increases as the density
decreases, corresponding to a negative potential step.
In fig.\ \ref{fig_Amp-w}$c$ we see how the squared amplitudes
vary with the density for this particular energy
of the mode $\tilde{\pi}_2$.
As seen in fig.\ \ref{fig_Amp-w}$c$,
contrary to previous examples,
only the $\Delta N^{-1}$ component remains,
as the zero density limit is reached.
In fig.\ \ref{fig_DhAmp-w}$c$
we show the squared amplitudes
of the individual $\Delta N^{-1}$ amplitudes
for the energy $\hbar \tilde{\omega}_2 = 320\ \MeV$
of the mode $\tilde{\pi}_2$.
Initially the mode is collective, but as the density is lowered the mode
becomes more and more non-collective.
{}From fig.\ \ref{fig_Gam_w-rho}$b$ we however see that the partial width
$\tilde{\Gamma}_\Delta^2$ actually is zero
at the quasipion energy
$\hbar \tilde{\omega}_2 = 320\ \MeV$ (dot-dashed curve)
at all densities.
Thus a $\Delta$
will not decay to a quasipion with this energy.
A mode at this energy with a very low momentum at half nuclear density,
which could occur in a time-dependent density,
could have a very large reflection coefficient,
approaching unity as the initial quasipion momentum approaches zero.
In our fourth and last example in this section
we consider the mode $\tilde{\pi}_2$
with energy $\hbar \tilde{\omega}_2 = 380\ \MeV$.
Initially this mode will now have the properties:
\[
\begin{array}{cccc}
q \approx 170\ \MeV/c
&
|Z|^2 \approx 0.12
&
\sum_{\Delta N^{-1}} |X_{\Delta N^{-1}}|^2 \approx 0.88
&
\sum_{N N^{-1}} |X_{N N^{-1}}|^2 \approx 0.0\ .
\end{array}
\]
As the density decreases we
follow the path indicated
by the dot-dashed line in fig.\ \ref{fig_w12-rho}$b$,
from $q \approx 170\ \MeV/c$ to $q \approx 350\ \MeV/c$.
In fig.\ \ref{fig_Amp-w}$d$ we see how the squared amplitudes
vary with the density for this particular energy
of the mode $\tilde{\pi}_2$.
As seen in fig.\ \ref{fig_Amp-w}$d$
in this case only the pion component remains,
as the zero density limit is reached,
although $\tilde{\pi}_2$ initially was
dominated by the $\Delta N^{-1}$ component.
Thus also in this case
the mode will escape the system as a free pion.
In fig.\ \ref{fig_DhAmp-w}$d$
we show the squared amplitudes
of the individual $\Delta N^{-1}$ amplitudes
for the energy $\hbar \tilde{\omega}_2 = 380\ \MeV$
of the mode $\tilde{\pi}_2$.
We see that the mode at all densities is completely collective.
{}From fig.\ \ref{fig_Gam_w-rho}$b$ we see that the partial width
$\tilde{\Gamma}_\Delta^2$ is quite substantial
at the quasipion energy
$\hbar \tilde{\omega}_2 = 380\ \MeV$ (dot-dashed curve)
also at normal nuclear density.
The reflection coefficient,
making the same assumptions as in the first example,
becomes for this case smaller than $10^{-3}$.
We thus conclude that also the upper pionic mode
penetrating the surface will always emerge as a free pion.
For high energies this is natural
because as the density decreases
the dispersion relation of the pionic mode
approaches the free pion relation.
The cases of low energy when
the unperturbed $\Delta N^{-1}$ branch is approached
as the density is lowered,
will never occur since no such modes will be created
from a decaying $\Delta$,
because the partial $\Delta$ width will be zero.
\subsection{ Refined scenarios }
\label{sec_QpionsOther}
If the surface changes with time
the energy of the pionic mode
need not to be conserved,
and there is some small possibility
for the pionic mode to end up
as an unperturbed $\Delta$-hole state.
The actual fraction of such modes is hard to estimate
without an explicit transport simulation,
but based on the scenario for the time-independent density profile,
it is reasonable to expect
that only a very small fraction of the pionic modes
will end up as unperturbed $\Delta$-hole states in vacuum.
In a quantum description
such a mode could be either reflected at the surface,
or the mode could break up into an uncoupled $\Delta$ and hole,
where the hole is trapped inside the nucleus,
and the $\Delta$ escape the system as a free $\Delta$.
Based on the results discussed
in the explicit examples of this section,
and the presentation in appendix \ref{sec_Refl}
we expect that the reflection at the surface
will be very small.
In a transport description the reflection
can be incorporated by a local transmission coefficient
as discussed in appendix \ref{sec_Refl}.
If the pionic mode is not reflected at the surface,
and its amplitude approaches 100\%
of the $\Delta N^{-1}$ component
as the density approaches zero,
the mode should thus break up
into an uncoupled $\Delta$ and hole.
In a transport simulation this could be practically handled
by allowing the pionic mode to absorb a nearby nucleon,
forming an uncoupled $\Delta$,
when the density falls below a specified value.
This prescription has some quantum mechanical justification
by the fact that at very low density
a wave packet representing the pionic mode,
will not be very well localized,
but instead have a large spatial spread.
\section{Summary}
\label{sec_Sum}
In-medium properties obtained in an infinite stationary system
consisting of interacting nucleons, nucleon resonances and mesons,
can be incorporated into transport descriptions
by a local density approximation.
While such a prescription is rather straightforward to implement
in the interior regions of the nuclear system,
conceptual problems exist at the nuclear surface.
When the nuclear density approaches zero,
collective mesonic modes formed in the medium
have to be converted to real particles in vacuum.
The problems arise since some collective modes
(e.g. $\Delta N^{-1}$-like)
may exists in the infinite
stationary system at arbitrary low (but non-vanishing) density,
but no corresponding real particle exists in vacuum.
This problem has been apparent in previous works \cite{Giessen,Texas}
where collective modes have been incorporated
into transport descriptions as quasimesons.
The character of the quasimesons (\ie\ realization in vacuum)
were in those works determined already at the time of creation.
Based on the formalism of ref.\ \cite{main},
we have in this paper employed a more elaborate
$\pi + N N^{-1} + \Delta N^{-1}$ model
(relative to the works \cite{Giessen,Texas})
to investigate a somewhat different treatment
of the collective pionic modes at a nuclear surface.
In this formalism we have obtained
not only density dependent dispersion relations
of the pionic modes,
but also density dependent amplitudes
of the components constituting the pionic mode.
These quantities are conveniently parametrized
with density dependent parameters,
in section \ref{sec_DispAmp}.
For the transport process it is not needed
to determine the character of the pionic modes
until they penetrate the surface and emerge as free particles.
This is automatically determined
within our formalism from the amplitudes at zero density.
We have further showed in section \ref{sec_Qpions}
that for a stationary density profile,
the conservation of the energy of the pionic mode
and the partial $\Delta$ decay width,
together leads to the fact
that only real pions are realized as free particles
when the pionic mode penetrates the surface.
Note that this finding is different from earlier works,
and it demonstrates the importance
of deriving dispersion relations and partial $\Delta$ widths
consistently within
a realistic\footnote{By ``realistic'' we here mean
that the $\Delta N^{-1}$ model
contains a continuum of
$\Delta N^{-1}$ and $N N^{-1}$ states,
as compared to the more simple
two-level $\Delta N^{-1}$ model
used for example in refs.\ \cite{Giessen,Texas}.
}
$\Delta N^{-1}$ model.
In a more refined scenario,
where the density changes with time,
deviations from this picture can be expected
and also the unperturbed $\Delta N^{-1}$ component
may be realized in the limit of vanishing density.
The actual fraction of such modes
is hard to estimate without an explicit transport simulation,
but based on the arguments in section \ref{sec_Qpions}
for the stationary surface,
we expect this fraction to be very small.
For the rare cases when the unperturbed $\Delta N^{-1}$ component
is realized in the limit of vanishing density,
the pionic mode must be converted to real particles.
In an extended description this could be made
by allowing the mode to break up into
an uncoupled $\Delta$ and a $N^{-1}$,
where the hole remains trapped in the nuclear system
and the $\Delta$ escapes the system.
Based on our formalism,
this seems to be the most probable scenario,
since the collective strength disappears on these modes
as the density approaches zero.
This could be implemented in a transport simulation
by letting the pionic mode absorb a nearby nucleon
to form an uncoupled $\Delta$,
as may be justified by the quantum-mechanical feature that
the wave packet representing the pionic mode is not well localized
at very low density.
Alternatively the pionic modes could in a quantal description
be reflected at the surface.
We have investigated
the reflection and transmission probabilities
for the collective modes
in a simplified one-dimensional scenario,
where the modes propagate perpendicular to the surface.
Exploring typical scenarios, we have found
that (with only very few exceptions)
the reflection of the pionic modes will be smaller than a few percent,
however, in appendix \ref{sec_Refl} we have
suggested how the reflection and transmission probabilities
could be incorporated into the transport descriptions
by using approximative local transmission coefficients. \\
\noindent
Stimulating discussions with Volker Koch are acknowledged.
This work was supported by the Swedish Natural Science Research Council,
and by the Director, Office of Energy Research, Office of High Energy
and Nuclear Physics, Nuclear Physics Division of the U.S. Department of
Energy under Contract No.\ DE-AC03-76SF00098.
|
\section{Introduction}
The problem of the gluon vacuum in QCD has a long history
\cite{savvidy,coleman,leutwyler,belavin,zakharov,shuryak}.
It is well known that at low energies perturbation theory does not apply
since the coupling constant becomes large and
the perturbative vacuum becomes unstable due to gluon self-interactions
\cite{savvidy}.
These self-interactions can lead to a 'reconstruction' of the vacuum and to
the appearance of a condensate \cite{nn,umt}. The presence of the condensate
is important for the physical properties of the low energy sector of QCD.
Of particular interest is the modification of the quasiparticle spectrum and
the occurence of massive collective excitations to be considered in this work.
In the literature two different kinds
of such condensates have been considered,
the {\it coherent} and the {\it squeezed} one.
In the coherent condensate gluon field excitations are
found by a {\it transitive} transformation, i.e.
shifting the fields to the solution of the classical
equations \cite{savvidy,coleman,leutwyler}
(e.g. instantons \cite{belavin,zakharov,shuryak}).
It is characterized by the condensation of single gluons and thus by
a nonvanishing vacuum expectation value of the gluon field $A$:
\begin{equation}
<A>\ \ \neq\ \ 0~.
\end{equation}
In the squeezed condensate on the other hand the gluon states are
constructed by a {\it multiplicative} Bogoliubov
transformation of the gluon fields \cite{celenza,biro1,biro2,mishra1,mishra2}.
The squeezed vacuum is characterized by the condensation of colourless,
scalar gluon pairs and could thus be realized through
\begin{equation}
<A>\ \ =\ \ 0\ ,\ \ <A^2>\ \ \neq 0\ .
\end{equation}
There has been a lot of activity to construct
a stable coherent vacuum in the gluon sector of QCD
\cite{savvidy,coleman,leutwyler}.
The problem in this case is that there are no stable
quasiclassical solutions to the
Yang-Mills equations in Minkowski space \cite{coleman,leutwyler}.
In recent years
the squeezed condensate (called here {\it Bogoliubov condensate})
has become a topic of great interest, see e.g.
\cite{celenza,biro1,biro2,mishra1,mishra2}.
Its investigation for non-Abelian fields faces the following problems:
\begin{enumerate}
\item
One has to find the adequate degrees of freedom to construct the gluon
condensation in a squeezed vacuum.
\item
Since the squeezed vacuum is most naturally described in the Hamiltonian
formalism one has to extract the gauge invariant oscillator - like
field variables from the non-Abelian QCD action. Recall that by using different
gauges for instance
Biro \cite{biro1,biro2} and Mishra \cite{mishra1,mishra2} get different
results for the condensate.
\item
The condensate leads to spontaneous gauge symmetry breaking with the
appearance of massive gluons. This has to
be accompanied by the corresponding
generation of the longitudinal components for the massive gluon fields.
Recent papers on ''squeezed'' gluon states
\cite{biro1,biro2,mishra1,mishra2} indeed obtain a constituent gluon mass.
However, since they fix the gauge, their gluon fields do not have a
longitudinal
component such that they are unable to provide the number of degrees of
freedom required for the description of a massive gluon vector field
\cite{slavnov}.
\end{enumerate}
With respect to the last topic, we mention that
mass generation for gauge fields via spontaneous gauge symmetry breaking
is a general problem familiar also from
other gauge field theories.
In the unified theory of electroweak interactions a
consistent description of massive vector fields can be given
due to the presence of a scalar Higgs field in the Lagrangian
which generates the longitudinal component of the massive gauge bosons
W and Z. The very interesting alternative, that the gauge bosons obtain
their mass by spontaneous gauge symmetry due to radiative corrections,
without introduction of an external Higgs field,
has been proposed by Coleman and Weinberg \cite{cw}.
This possibility is of great importance since for SU(N) gauge theories
the introduction of an external
Higgs field does not lead to a mass term for the gluons, as was shown by
Georgi and Glashow \cite{Georgi}.
The concept of spontaneous gauge symmetry breaking by radiative corrections
is therefore very attractive for the case of QCD and has been followed up
to now \cite{biro1,biro2,mishra1,mishra2}.
In the present work we study the possibility of gluon condensation
in a squeezed vacuum in view of the Bogoliubov model
\cite{nn} of the weakly nonideal Bose gas. In particular we investigate
the influence of a squeezed condensate on the gluon quasiparticle spectrum
in the low-energy region of QCD.
Note that the Bogoliubov theory was the first to explain the experimentally
observable spectrum of collective excitations of superfluid $^4$He.
This spectrum cannot be obtained by resummations of the conventional
perturbation theory series.
As Bogoliubov has shown, the collective
exitations are determined by the ''condensate'' of particles with zero
momentum and finite density. A first connection between Bogoliubov
condensation and the squeezed vacuum in field theory
(massless $\lambda \phi^4$)
was made by Castorina and Consoli \cite{CC}, see also \cite{PaBla}.
An application of the
concept of the Bogoliubov model to QCD,
however, has to our knowledge not been carried out by now.
In a first attempt to generalize the Bogoliubov model to QCD
we use the infrared
singularity of massless theories to squeeze the zero momentum mode
which leads to massive gluonic quasiparticles
in the nonzero momentum sector.
The free squeezing parameter is fixed from the $\eta-\eta'$ mass difference.
Our coresponding value for the gluon condensate is in
reasonable agreement with that obtained by Shifman, Vainshtein and Zakharov
\cite{zakharov} which supports our semi-phenomenological approach.
Concerning the above mentioned problems (2) and (3),
we try to solve the problem of
the appearance of a constituent gluon mass using
a gauge invariant scheme for the elimination of the unphysical components
of the gluon vector field \cite{dirac,perv1,perv2,fj,kp} which does not
require the gauge-fixing as initial supposition.
This scheme is based on
the construction of projection operators by formally solving the Gauss law
constraint. We show that these projectors are destroyed by
the interaction of gluons with the squeezed vacuum. As result
a constituent gluon mass appears together with the necessary
longitudinal components.
This is the central result of our paper and is quite
in analogy to spontaneous chiral symmetry breaking
in the quark sector \cite{ad,lopr,pk2s}. There, the appearance of
constituent quark masses due to the interaction of quarks with the squeezed
vacuum is accompanied by the destruction of the chiral projection
operator which leads to the necessary
increase of the number of spinor field components from two (Weyl spinors)
to four (Dirac spinors).
The paper is organized as follows:
In Section \ref{sec:field}, the Bogoliubov model
of a weakly nonideal Bose gas is generalized to field theory.
We give a field
theoretical description of the condensation phenomenon by the use of the
squeezed vacuum and discuss the conventional local
$\lambda \phi^4$ theory. The relation to the Bogoliubov model
for the weakly nonideal Bose gas is given in Appendix \ref{app:bogmod}.
In Section \ref{sec:qcd}
the homogeneous colourless
Bogoliubov condensate of gluons is introduced in QCD
where the unphysical degrees of freedom are eliminated by applying
projection operators instead of fixing a gauge.
We also discuss spontaneous gauge symmetry breaking and
the corresponding occurrence of a massive gluon quasiparticle spectrum.
In Section \ref{sec:app} the squeezing parameter is fixed
>from the $\eta' - \eta$ mass difference.
In Section \ref{sec:end} we present the conclusions.
\section{Bogoliubov condensation in quantum field theory}
\label{sec:field}
In order to introduce some notations and methods needed for our investigation
of the squeezed condensate in the rather complicated QCD,
we first consider massless $\lambda\phi^4$ theory with the Hamiltonian
\begin{equation}
H=\int d^3x[\pi(x)^2+\left(\partial_i\varphi(x)\right)^2+{\lambda\over 4!}
\varphi^4(x)]
\end{equation}
as the simplest example
of an interacting bosonic theory which is renormalizable.
The theory is quantized by turning the classical fields
$\varphi({\bf x},t),\pi({\bf x},t)$ to Schr\"odinger operators
$\varphi({\bf x}),\pi({\bf x})$
and imposing the canonical commutation relations
\begin{equation}
[\pi({\bf x}),\varphi({\bf x}')]=-i\delta({\bf x}-{\bf x}')~.
\end{equation}
In the momentum representation defined by
\begin{equation}
\label{FouTra}
\varphi_p={1\over\sqrt{V}}\int d^3x
\mbox{\large e}^{i{\bf p}{\bf x}}\varphi({\bf x}),\ \ \ \
\pi_p={1\over\sqrt{V}}\int d^3x
\mbox{\large e}^{i{\bf p}{\bf x}}\pi({\bf x})~,
\end{equation}
the Hamilton operator is
\begin{equation}
\label{hf1}
:{ H}[\varphi,\pi]:= \frac{1}{2} \sum_p \left[:\pi_p \pi_{-p}: + p^2
:\varphi_p \varphi_{-p}: \right]
+ \frac{\lambda}{4!V} \sum_{p_1 p_2 p_3 p_4} \delta_{p_1+p_2+p_3+p_4,0}
:\varphi_{p_1}\varphi_{p_2}\varphi_{p_3}\varphi_{p_4}: ,
\end{equation}
with the commutation relations
\begin{equation}
\label{crphipi}
[\pi_p,\varphi_{p'}]= - i\delta_{p,-p'} ~,~~
[\varphi_p,\varphi_{p'}]= [\pi_p,\pi_{p'}]= 0~.
\end{equation}
In (\ref{hf1}) we have introduced normal ordering with respect to the
creation and annihilation operators
$a_p, a_p^+$ defined according to
\begin{eqnarray}
\label{phipi}
\varphi_p =\sqrt{\frac{1}{2 \tilde{\omega}(p)}} (a_p + a^+_{-p})\ , \;\ \ \
\pi_p = i \sqrt{\frac{\tilde{\omega}(p)}{2 }} (-a_p + a^+_{-p}) ~,
\end{eqnarray}
with an arbitrary function $\tilde{\omega}(p)$.
The operators $a_p, a_p^+$ satisfy the commutation relations
\begin{equation}
\label{craa}
[a_p, a_{p'}^+]= \delta_{p, p'}~,~~[a_p, a_{p'}]= [a_p^+, a_{p'}^+]= 0 ~.
\end{equation}
The corresponding vacuum $|0> $ is defined by $a_p |0> = 0$, and the Fock
space is given as
\begin{eqnarray}
\label{fockstate}
\{\,|\,\Phi >\} = \,|\,0 >;\;\;\; a^+_p |0 > = |\,p >, ... ~.
\end{eqnarray}
We note that the special choice $\tilde{\omega}(p)=|p|$ diagonalizes the
free ($\lambda$ independent) part of the Hamiltonian. For this case,
however, $a_0$ and $a_0^+$ are not defined which corresponds to the well-known
infrared singularity of massless theories.
Generalizing the Bogoliubov model to field theory we should use the infrared
singularity of massless theories to squeeze the zero mode
by populating it macroscopically with massless particles.
Then, we diagonalize the nonzero mode single particle part of the
resulting squeezed Hamiltonian by changing from particles to quasiparticles
whose dispersion relation $\tilde{\omega}(p)$ is finally
determined selfconsistently. For the simple case of $\lambda\phi^4$
this has been carried out in detail \cite{PaBla}. Similar to the Bogoliubov
model this leads to renormalization of the bare parameters like the coupling
constant $\lambda$.
For the time being we leave $\tilde{\omega}(p)$ open and suppose
that the vacuum of the theory (\ref{hf1})
contains a large number of quasiparticles with zero momentum ($p=0$).
We construct this vacuum using the unitary squeezing operator
\begin{equation}
\label{ub}
U_B (\varphi_0,\pi_0) =
\exp \left (i\frac{f_0}{2} (\pi_0 \varphi_{0} + \varphi_0 \pi_{0})
\right )~,
\end{equation}
where $f_0$ is a very large parameter to be fixed later.
The operator $U_B$ transforms the Fock space of states to the Bogoliubov space
of states
\begin{eqnarray}
\label{bogstate}
|\, \Phi_B >\equiv U_B^{-1} \,|\, \Phi > ~.
\end{eqnarray}
In quantum optics these states are called
'squeezed states', see e.g. \cite{optics}.
Applying the unitary transformation (\ref{ub}), we can define
the new field operator $\varphi^B_0$ and its momentum $\pi^B_0$ by means of
\begin{eqnarray}
\label{bt}
\varphi^B_0 = U_B ^{-1}\varphi_0 U_B = {\rm e}^{-f_0} \varphi_0~,\nonumber\\
\\
\pi^B_0 = U_B ^{-1}\pi_0 U_B = {\rm e}^{f_0} \pi_0 \nonumber~,
\end{eqnarray}
which satisfy the same algebra of commutation relations
as the initial ones (\ref{crphipi}).
We shall now carry out the squeezing of the zero mode part of the Hamiltonian
by applying a Wick reordering procedure to the Bogoliubov vacuum $|0_B>$.
Note that under the squeezing transformation (\ref{bt}) the contraction
of a pair of field operators is left invariant,
\begin{equation}
C=
<0 \,|\,\varphi_{0}\varphi_{0} \,|\,0> =
<0_B\,|\,\varphi^B_{0} \varphi^B_{0}\,|\,0_B>~.
\end{equation}
The normal ordering
of the Bogoliubov fields $\varphi_B$ with respect to the Bogoliubov vacuum
(which is denoted as $::\varphi^B_{0} \varphi^B_{0} ::$)
has the same form as the normal ordering of the original
fields with respect to the Fock vacuum (\ref{fockstate})
\begin{eqnarray}
\label{no}
C = \varphi_{0} \varphi_{0} - :\varphi_{0} \varphi_{0}:
= \varphi^B_{0} \varphi^B_{0} - :: \varphi^B_{0} \varphi^B_{0} ::~~.
\end{eqnarray}
To reorder the Hamiltonian (\ref{hf1})
with respect to the new vacuum $|0_B >$ we use Eqs.(\ref{bt}) and (\ref{no}).
Reordering of the quadratic term gives
\begin{equation}
\label{bc}
:\varphi_{0} \varphi_{0}: =
::\varphi_{0}^B \varphi_{0}^B::~
\mbox{\large{e}}^{2f_{0}}~ + ~\tilde{C} ~,
\end{equation}
where
\begin{eqnarray}
\label{bfi}
\tilde{C}=
C
\left (\mbox{\large{e}}^{2f_{0}} - 1 \right )~.
\end{eqnarray}
Analogously we have
\begin{equation}
:\pi_{0} \pi_{0}: = ::\pi_{0}^B \pi_{0}^B::~
\mbox{\large{e}}^{-2f_{0}}~+~ C^{\pi}
\left (\mbox{\large{e}}^{-2f_{0}} - 1 \right ),
\end{equation}
with $ C^{\pi}=$
Reordering of the quartic term gives
\begin{eqnarray}
:\varphi_{0}\varphi_{0}\varphi_{0}\varphi_{0}: &=&
::\varphi_{0}^B \varphi_{0}^B \varphi_{0}^B \varphi_{0}^B::
\mbox{\large{e}}^{4f_{0} }
+:: \varphi_{0}^B \varphi_{0}^B::
\mbox{\large{e}}^{2f_{0}}
\tilde{C} +(5\,\, {\rm permutations})\nonumber\\
& & +\tilde{C}^2 + (2 \,\,{\rm permutations})~.
\end{eqnarray}
For applications in QCD we quote here also the general Wick reordering
result for any polynomial $F(\varphi_0)$
\begin{equation}
:F(\varphi_0):=\exp\{{1\over 2}\tilde{C}
{d^2\over d b^2}\}::F(\varphi^B_0\mbox{\large{e}}
^{f_0}+b)::\Big|_{b=0}~.
\end{equation}
As result of the reordering of $H=H_0+H'$,
where $H'$ is the Hamiltonian of nonzero momentum excitations ($p\neq 0$), we
obtain
\begin{eqnarray}
:H:&=& ::H_0(\phi_0):: + :H':~,\nonumber\\
\label{hfb}
:H'[\varphi,\pi]: &=&
E_0
+ :H^{(2)}[\varphi,\pi]:
+ :H^{(4)}[\varphi]:~,
\end{eqnarray}
where
\begin{eqnarray}
\label{h0}
E_0&=& 3{\lambda\over 4!V}\tilde{C}^2 + C^{\pi}\\
:{H}^{(2)}[\varphi,\pi]: &=& \frac{1}{2}\sum_{p\neq 0}
\left\{:\pi_{p} \pi_{-p}: +
\left[p^2 +
{\lambda\over 2V} \tilde{C} _{00}\right]
:\varphi_{p} \varphi_{-p}: \right\} ~,\\
:{ H}^{(4)}[\varphi]: &=&{\lambda\over 4!V}
\sum_{p_1,p_2,p_3,p_4\neq 0}\delta_{p_1+p_2+p_3+p_4,0}
:\varphi_{p_1}\varphi_{p_2}\varphi_{p_3}\varphi_{p_4} :~.
\end{eqnarray}
The zero momentum operator $::H_0(\phi_0)$ containing terms proportional to
$::\varphi_{0}^B \varphi_{0}^B::$ and
$::\varphi_{0}^B \varphi_{0}^B \varphi_{0}^B \varphi_{0}^B::$
describes excitations of the condensate and has not been written explicitely
here.
Note that for very large $f_0$ the second term in the expression (\ref{h0}) for
$E_0$ is much smaller than the first one (cf. (\ref{bfi})) and can be
neglected. Hence we find a condensate energy density
\begin{equation}
\label{Condene}
\epsilon_0\equiv{E_0\over V}=
{\lambda\over 8}\left({\tilde{C}\over V}\right)^2
\end{equation}
and a bosonic quasiparticle mass $m_B$,
\begin{equation}
\label{Bogmass}
m_B^2 \equiv \lambda {\tilde{C}\over 2V}~,
\end{equation}
appears.
Diagonality of the one-particle part $:H^{(2)}:$ of the reordered
Hamiltonian (\ref{hfb}) demands that we put the quasiparticle energy
$\tilde{\omega}(p)$ in Eq. (\ref{phipi}) to
\begin{equation}
\tilde{\omega}(p)=\sqrt{p^2+m_B^2}~.
\end{equation}
The effective Hamiltonian depends (through $\epsilon_0$ and $m_B$) on
the free parameter $\tilde{C}$.
Using Eq. (\ref{Bogmass}) we can eliminate the parameter $\tilde{C}$
>from the expression for the condensate energy density $\epsilon_0$ in Eq.
(\ref{Condene}) to obtain
\begin{equation}
\epsilon_0 =\frac{m_B^4 }{2 \lambda}~,
\end{equation}
which has the same nonanalytic dependence on the coupling constant as
that of the Higgs mechanism of mass generation.
Obviously, the Bogoliubov mechanism of spontaneous mass generation
presented above differs from the Higgs mechanism by the representations
of the vacuum and the interaction of the quasiparticles.
The Higgs mechanism corresponds to the coherent vacuum representation, while
the Bogoliubov one - to the squeezed vacuum, see Appendix \ref{app:bogmod}.
The introduction of the Bogoliubov condensate is related to the Wick
reordering procedure with respect to a new Fock space.
\section{Bogoliubov condensate in QCD}
\label{sec:qcd}
After the introductory generalization of Bogoliubov condensation for
superfluid $^4$He to
massless $\lambda\phi^4$ theory above it is attractive
to suppose that also the gluon vacuum of QCD can be considered as a
homogeneous colourless condensate of gluon pairs.
First steps in this direction where undertaken in Celenza and Shakin
\cite{celenza}.
The corresponding treatment in QCD is far more complicated than in
$\lambda\phi^4$ due to the
fact that QCD is a gauge theory with unphysical degrees of
freedom in the Lagrangian which have to be eliminated before quantization.
According to Dirac \cite{dirac}, only the spatial components of the gauge
fields are dynamical and have to be quantized. The time components obey
constraint equations (Gauss laws) and have to be eliminated.
As in the simpler case of massless $\lambda\phi^4$ theory the
squeezed condensate
is described by the procedure of Wick reordering with a free parameter
$\tilde{C}$ and leads both to a vacuum energy and to a mass term for the field.
In QCD, however, the presence of a condensate and the corresponding generation
of a mass term for the gauge field leads to spontaneous gauge symmetry
breaking, since the gauge invariance of the Hamiltonian is not shared by
the vacuum. The problem of
the appearance of a constituent gluon mass und the corresponding
resurrection of the longitudinal component of the massive quasigluons is
solved by using
a projection scheme for the elimination of the unphysical components
of the gluon vector field \cite{dirac,perv1,perv2,fj,kp} instead of
gauge-fixing.
The projectors are obtained by formally solving the Gauss law
constraint and appear in the kinetic energy term of a reduced gauge invariant
QCD Hamiltonian.
We show that these projectors are destroyed by
the interaction of gluons with the squeezed vacuum so that
the constituent gluon mass appears together with the necessary
longitudinal components.
The presence of a squeezed condensate leads to spontaneous gauge symmetry
breaking:
The gauge invariance of the QCD Hamiltonian is not shared by
the vacuum.
Finally we fix the free parameter $\tilde{C}$ of our squeezed vacuum
by estimating a value for the quasigluon mass from the $\eta'-\eta$
mass difference and comparing the corresponding condensate energy density
to the well known value obtained by Shifman et al. \cite{zakharov}.
We shall see that they are in good agreement.
\subsection{QCD Hamiltonian and Gauss law}
We start from the QCD Lagrangian
\begin{eqnarray}
\label{1}
{\cal L}(A) = -\frac{1}{4} F_{\mu\nu}^a F^{\mu\nu a}~,
\end{eqnarray}
where $F_{\mu\nu}^a$ is the field strength tensor
\begin{eqnarray}
F^a_{\mu\nu} = \partial_{\mu} A^a_{\nu} -
\partial_{\nu} A^a_{\mu} + g f^{abc} A^b_{\mu} A^c_{\nu}~.
\end{eqnarray}
In the following, we use the notation
\begin{eqnarray}
{A}_{\mu} = g \frac{A^a_{\mu} \lambda^a}{2i}~,
\end{eqnarray}
where $ g$ is the coupling constant.
Due to the gauge invariance
\begin{eqnarray}
\label{GI}
{A}^v_{\mu} = v({A}_{\mu} + \partial_{\mu})v^{-1}~,
\end{eqnarray}
\begin{eqnarray}
{\cal L} (A^v) = {\cal L} (A)~,
\end{eqnarray}
this classical Lagrangian contains only $3(N_c^2-1)$ degrees of
freedom instead of the $4(N_c^2-1)$ components of the ${A}$ field
($N_c$ is the number of colours).
For the construction of the Hamiltonian one usually introduces chromoelectric
and chromomagnetic fields
\begin{eqnarray}
E^a_i&=&\dot{A}^a_i - D^{ab}_i({\bf A}) A_0^b~,\\
B^a_i({\bf A}) &=& \frac{1}{2}~\epsilon_{ijk}
F_{jk}^a = \epsilon_{ijk}D^{ab}_j({\bf A})
A^b_k~,
\end{eqnarray}
with
$\dot{A}_i = {\partial_0} A_i $ and
the covariant derivative
\begin{eqnarray}
D^{ab}_i ({\bf A}) = \delta^{ab} \partial_i + g~{f^{acb}} A^c_i~.
\end{eqnarray}
The magnetic field satisfies the Bianchi identity
\begin{eqnarray}
\label{BI}
D^{ab}_i ({\bf A}) B^b_i({\bf A}) \equiv 0 ~,
\end{eqnarray}
which can be interpreted as a generalized transversality of the magnetic
field.
In order to construct the Hamiltonian, we have to find the canonical
momenta. We see that the Lagrangian (\ref{1}) does not contain time
derivatives of the zero components of the gluon fields. The corresponding
Euler-Lagrange equations are therefore constraints
(the Gauss laws):
\begin{eqnarray}
\label{Gaulaw}
D^2_{ab} ({\bf A}) A^b_0 = D^{ab}_i ({\bf A}) \dot{A}^b_i~,
\end{eqnarray}
where $D^2_{ab} = D^{ac}_i D^{cb}_i$.
In terms of the electric field the Gauss laws (\ref{Gaulaw}) read
\begin{equation}
\label{clGau}
G^a({\bf A},{\bf E})\equiv D^{ab}_i({\bf A})E^b_i=0~.
\end{equation}
The canonical momenta to the spatial fields $A_i^a$ are
the electric fields:
\begin{equation}
{\delta{\cal L}\over \delta \dot{A}^a_i}=E_i^a ~,~~i=1,2,3~.
\end{equation}
The Hamiltonian can now be written as
\begin{equation}
\label{clHam}
H({\bf A},{\bf E})=\int d^3x {1\over 2}\left[{E_i^a}^2+{B_i^a}^2\right]~.
\end{equation}
In the classical theory we thus have the Hamiltonian (\ref{clHam}) together
with the Gauss constraint (\ref{clGau}).
\subsection{Quantization}
In order to quantize the theory one could write the Hamiltonian
completely in terms of gauge invariant variables and their canonical
conjugate momenta and then impose the canonical commutation relations
only on these gauge invariant variables. Different to the case of QED,
this leads to inconsistencies in QCD, as shown in detail in Appendix B.
A more successful alternative way is Dirac quantization \cite{dirac},
where one imposes the
canonical commutation relations on the original $A_i$ and $E_i$:
\begin{equation}
\label{QCDCCR}
[E_i^a({\bf x}),A_j^b({\bf x}')]=
i\delta^{ab}\delta_{ij}\delta({\bf x}-{\bf x}')~.
\end{equation}
Both the Hamiltonian $H$ in (\ref{clHam}) and the Gauss function $G^a$ in
(\ref{clGau})
then become operators
satisfying
\begin{eqnarray}
\label{GIC}
[H,G^{a}({\bf x})] &=& 0~, \\ \quad
[ G^{a}({\bf x}),G^{b}({\bf x'})]
&=& if^{abc}G^c({\bf x})\delta({\bf x}-{\bf x}').
\end{eqnarray}
Since $G^a$ can be interpreted as the infinitesimal generator for gauge
transformations these two commutation relations express the gauge invariance
of the Hamiltonian (under small gauge transformations).
Note that the Hamilton operator obtained from the classical
Hamiltonian (\ref{clHam}) with the cartesian fields $E_i$ and $A_i$ has the
correct operator ordering \cite{ChrLee}.
As pointed out by Jackiw \cite{RJ},
the Gauss law $G^a=0$ cannot be taken as an operator equation
since it would lead to inconsistency with the Dirac commutation relations
(\ref{QCDCCR}).
Jackiw then suggested that the Gauss law should be implemented
by demanding that a physical state satisfies the Schr\"odinger equation
(with energy eigenvalue $\cal E$) and is annihilated by the Gauss law operator
\begin{eqnarray}
H({\bf A},{\bf E})|\Phi>&=&{\cal E} |\Phi>~,\\
\label{GauCon} G^a({\bf A},{\bf E})|\Phi>&=& 0~.
\end{eqnarray}
The second equation is the condition of gauge invariance of
the physical states.
The Gauss law constraint (\ref{GauCon}) can then at least in principle
be implemented by use of unitary transformations \cite{GJ} and
is still under lively discussion \cite{Lenz}.
The resulting kinetic term in the Hamiltonian is very complicated.
The requirement of gauge invariance of the physical states
expressed by (\ref{GauCon}), however, is too
restrictive. It does not allow for the possibility of spontaneous
breaking of the gauge symmetry in analogy to spontaneous symmetry
breaking in the Higgs model and spontaneous chiral symmetry breaking.
If the gauge symmetry is broken spontaneously, the gauge invariance
of the Hamiltonian is not shared by the vacuum.
We can arrive at a gauge invariant Hamiltonian
without demanding gauge invariance of the physical states, especially
of the vacuum
\begin{equation}
G^{a}({\bf E},{\bf A})|0_B>\neq 0~,
\end{equation}
by starting from a gauge invariant reduced
classical Hamiltonian. This is achieved by a projection method described in
the following.
Using the formal solution of the Gauss equations (\ref{Gaulaw})
\begin{equation}
\label{10}
A^a_0[{\bf A}] = \frac{1}{D^2_{ab}({\bf A})} D^{bc}_i ({\bf A})\dot{A}_i^c~,
\end{equation}
the electric field can be written as
\begin{equation}
\label{transvE}
E^a_i=\Pi_{ij}^{ab}({\bf A})\dot{A}_j^b~,
\end{equation}
with the projection operator
\begin{equation}
\label{pa}
\Pi_{ij}^{ab}({\bf A}) =
\delta_{ij}\delta^{ab}-D^{ac}_i({\bf A}){1\over D^2_{cd}({\bf A})}
D^{db}_j({\bf A})~.
\end{equation}
We assume that zero modes of the differential operator $D^2_{ab}(A)$
are absent. Consideration of zero modes is under current investigation
\cite{kp}.
In the case of $A=0$, this projection operator reduces to the transverse one
$\Pi_{ij}^{ab}({\bf A}=0) =\delta^{ab} \delta^T_{ij}\equiv\delta^{ab}
\left(\delta_{ij}-\partial_i\partial_j/\partial^2\right)$.
The gauge invariant reduced Lagrangian can be written as
\begin{equation}
\label{redLag}
{\cal L}^{Red}({\bf A})=
{1\over 2}\left[\left(\Pi_{ij}^{ab}({\bf A})\dot{A}_j^b\right)^2-
{B_i^a}^2({\bf A})\right]~.
\end{equation}
Note that we still have
\begin{equation}
{\delta{\cal L}^{Red}\over \delta \dot{A}^a_i}=E_i^a~, ~~i=1,2,3~.
\end{equation}
Due to the property
$\Pi^2=\Pi$ of the projection operator the gauge invariant reduced
Hamiltonian can be written in the form
\begin{eqnarray}
\label{redHam}
H^{Red}({\bf A},{\bf E})=
\int d^3x
{1\over 2}\left[E_i^{a}\Pi^{ab}_{ij}({\bf A})E_j^b+{B_i^{a}}^2({\bf A})
\right]~.
\end{eqnarray}
The non-Abelian projection operator has been inserted between the cartesian
electric fields $E_i$, which as variables of the Hamiltonian lost their
transversality property. The above form
(\ref{redHam}) is only one of many possible forms $E\Pi E$, $(\Pi E)^2$,
$(\Pi E)\Pi (\Pi E)$,... Whereas in QED they are equivalent due to
the property
$\Pi^2=\Pi$ and the possibility to perform partial integrations, in QCD they
are inequivalent due to the presence of the $A$ field in the covariant
derivatives and lead to different operator orderings of $E$ and $A$ after
quantization. The simplest choice $E\Pi E$ in (\ref{redHam})
will be correct at least for our investigation of a squeezed homogeneous
condensate, as discussed in the next paragraph.
Although the form (\ref{redHam}) is gauge invariant classically, we did not yet
succeed in showing explicitly, that the corresponding
Hamilton operator satisfies (\ref{GIC}).
Thus in our treatment the role of gauge fixing is played
by the projection operator (\ref{pa}), for deatils see \cite{perv1,perv2,kp}.
The nonabelian chromomagnetic field projects onto the generalized
transverse component of the $A$ field quite analogous to the form
$E_i^a \Pi_{ij}^{ab}(A) E_j^b$ for the chromoelectric field.
\subsection{Spectrum of quasigluon excitations}
\label{ssec:qge}
We shall consider the squeezed vacuum containing a colourless homogeneous
condensate of gluon pairs for the Hamiltonian
\begin{equation}
\label{QCDHAM}
:H({\bf A},{\bf E}): = \int {\rm d}^3 x ~ \frac{1}{2} \left[
:E_i^a \Pi_{ij}^{ab}({\bf A}) E_j^b: + :{B^a_i}^2({\bf A}):
\right]~.
\end{equation}
We have introduced normal ordering with respect to creation $a_p^+$ and
annihilation operators $a_p$ defined with respect to some open
$\tilde{\omega}(p)$ in close analogy to our definitions (\ref{FouTra}) and
(\ref{phipi}) introduced in Section II for the $\lambda\varphi^4$ model.
A four-gluon interaction term occurs in the $AAAA$ term of the magnetic part
${B^{a}_i}^2({\bf A})$ and in the kinetic term
$E_i^a \Pi_{ij}^{ab} E_j^b$.
In analogy to the $\lambda \varphi^4$ model we perform Wick reordering
to the new squeezed vacuum and consider a
homogeneous and colourless condensate ($f_{p\neq0}=0, f_0\neq 0$) with the
contraction
\begin{eqnarray}
\label{aa}
<0 \,|\,A_i^a(p_1)A_j^b(p_2) \,|\,0>&=& <0_B\,|\,(A^B)_i^a(p_1)
(A^B)_j^b(p_2)\,|\,0_B> =
\delta_{ij} \delta^{ab} \delta_{p_1,0}\delta_{p_2,0} C~~,
\end{eqnarray}
where $A({\bf p}),E({\bf p})$ are the Fourier transforms of
$A({\bf x}),E({\bf x})$ in analogy
to (\ref{bc}),
and the reordering formula for zero momentum gluon fields
\begin{equation}
\label{apa}
:A_k^a(p=0) A_l^b(p=0): =
::{(A^B)}^a_k(p=0){(A^B)}^b_l(p=0):: {\rm e}^{2f_0} + \tilde{C} \delta^{ab}
\delta_{k,l}~,
\end{equation}
is in analogy to Eq. (\ref{bc}), where
$\tilde{C}= C \left (\mbox{\large{e}}^{2f_{0}} -1 \right)$.
The corresponding reordering formula for the
quartic term is calculated in Appendix C.
The Wick reordering of the magnetic part of the
Hamiltonian (\ref{17}),
\begin{eqnarray}
\frac{1}{2}\int {\rm d}^3 x ~:{B_i^a}^2({\bf A}): &=&
\frac{1}{2}\int {\rm d}^3 x ~\left\{:(\partial_j A_k^a) \delta^T_{kl}
(\partial_jA_l^a):
+2 g f^{abc} :(\partial_j A_k^a) A_j^b A_k^c:
+\frac{1}{2} g^2 f^{abc} f^{ade} : A_j^bA_k^cA_j^dA_k^e : \right\}~,
\end{eqnarray}
leads to the result (see Appendix \ref{app:wick})
\begin{eqnarray}
\label{b2}
\frac{1}{2}\int {\rm d}^3 x ~: B_i^{a2}({\bf A}): &=& g^2\frac{3}{2}
\frac{N_c}{V} (N_c^2 - 1 ) {(\tilde{C})}^2
+\frac{1}{2} \sum_{p\neq 0} \left[(p^2 + 2 g^2 N_c \tilde{C}/V) \delta_{ij}
- p_i
p_j \right] : A_i^a(p) A_j^a(-p) :
\nonumber\\&&+\frac{1}{4 V} g^2 \sum_{p_1 \dots p_4\neq 0}f^{abc} f^{ade}
:A_j^b(p_1)A_k^c(p_2)A_j^d(p_3)A_k^e(p_4) :
\delta_{p_1+p_2+p_3+p_4,0} ~.
\end{eqnarray}
The first term corresponds to the conventional definition
of the gluon condensate \cite{zakharov}
\begin{eqnarray}
\label{26}
G^2 &=& \frac{ g^2}{\pi^2} < \frac{1}{4} F^a_{\mu\nu} F^{\mu\nu a} > =
\frac{g^2}{2 \pi^2} < B^2_i > \\ \nonumber
&=& \frac{N^2_c - 1}{2\pi^2 3N_c} (3N_c g^2 \tilde{C}/V)^2~.
\end{eqnarray}
The second term in Eq.(\ref{b2}) includes the mass of the quasigluons
which have both
transverse and longitudinal parts, as the operation of the
reordering destroys the projection properties of the non-Abelian
magnetic field. This fact can be understood as
spontaneous gauge symmetry breaking, and is the Bose-analogy
of the spontaneous chiral symmetry breaking for the constituent quarks
\cite{ad,lopr,pk2s}, which is also realized by the corresponding Bogoliubov
transformation of the ''squeezed'' type. It is well known that a
massive vector field requires for the description a number of degrees
of freedom which exceeds that provided after fixation of a gauge in QCD.
Note that the fixing of a gauge results in an elimination of the
longitudinal components of the vector field and is inconsistent with the
concept of a mass \cite{slavnov}. So, the method of projection onto gauge
invariant variables \cite{dirac} used here is more adequate to the
phenomenon of spontaneous gauge symmetry breaking
than the conventional gauge fixing method.
Similarly to the above magnetic energy the projector $\Pi_{ij}^{ab}$ in the
electric energy $E_i^a \Pi_{ij}^{ab}({\bf A}) E_j^b$ in (\ref{QCDHAM})
is destroyed by the Wick reordering procedure leading to the
kinetic energy contribution
\begin{eqnarray}
\label{ekin}
:E_i^a E_j^b : <\Pi_{ij}^{ab}({\bf A})> ~.
\end{eqnarray}
The expression $<\Pi_{ij}^{ab} ({\bf A})>$
can be determined
in the low energy limit $ p^2 \sim 0 $:
\begin{eqnarray}
\label{cpr}
<\Pi_{ij}^{ab}({\bf A})>
\big|_{p \sim 0}\simeq \delta_{ij}\delta^{ab} - <f^{agc} A^g_i
\frac{1}{\sum_k f^{cme} A^m_k f^{end} A^n_k}f^{dlb} A^l_j >
= \frac{2}{3} \delta_{ij}\delta^{ab} ~.
\end{eqnarray}
which can easily be checked by summation over colour and space indices.
The Eqs. (\ref{b2}), (\ref{ekin}) and (\ref{cpr}) allow us to find the
effective
Hamiltonian for the quasigluon excitations in the low-energy limit:
\begin{eqnarray}
\label{lowHam}
H_{\rm eff} ({\bf E},{\bf A}) = \frac{1}{2} (\frac{2}{3})
\int {\rm d}^3 x (E_i^2 + m_g^2 {A}_i^2 )~,
\end{eqnarray}
with the quasigluon mass
\begin{equation}
\label{mg2}
m_g^2 = 3 N_c g^2 \frac{\tilde{C}}{V}~.
\end{equation}
The corresponding effective low energy Lagrangian is
\begin{eqnarray}
\label{lquglu}
{\cal L}_{\rm eff} ({\bf A}) =
\frac{1}{2} (\frac{2}{3}) \int {\rm d}^3 x (\dot{A}_i^2
- - m_g^2 {A}_i^2 )~.
\end{eqnarray}
The quasigluon mass $m_g$ in the low energy limit is determined
by the vacuum expectation value using the relations (\ref{26})
and (\ref{mg2}):
\begin{eqnarray}
\label{mg}
m_g = \sqrt{\frac{3\pi G}{2}}~.
\end{eqnarray}
We have shown that the Bogoliubov condensation of gluon pairs leads to a
nonvanishing contraction $\tilde{C}$ of gluon fields
which results in the spontaneous
gauge symmetry breaking and the occurence of a gluon mass.
We have obtained a new gluon mass formula for the low energy limit of QCD.
In the following Section we examine consequences of the present approach to
the low energy sector of QCD for the $\eta'$ mass formula.
\section{Applications: Quasigluon mass and $\eta -\eta'$ mass difference }
\label{sec:app}
According to the Bogoliubov condensate approach, the contraction
$\tilde{C}$ is a phenomenological parameter of the "squeezed"
vacuum state and is
directly related to the macroscopic occupation of the zero momentum
quasigluon state and therefore to the gluon mass.
We suggest to use this relationship for the fixation of
the squared mass difference
\begin{eqnarray}
\label{eta}
m_{\eta'}^2 - m_{\eta}^2 = \Delta {m^2_{\eta'}} = 0.616\; {\rm GeV}^2.
\end{eqnarray}
According to conventional approaches to the determination of the $\eta'$ mass
\cite{rst},
we suppose that the mass difference (\ref{eta}) is determined by
the $\eta' \rightarrow \eta'$ transition
through the process of the anomalous decay of the $\eta'$ meson
into the gluon condensate $B_i^a$ and a collective gluon excitation
$E^a_i$.
The effective Lagrangian of such a process can be derived according to Ref.
\cite{rst}, see also \cite{volkov},
with the result
\begin{eqnarray}
\label{lag}
{{\cal L}_{\eta'}} = \frac{1}{4} F^a_{\mu\nu}\tilde{F}^{a\mu\nu}~\eta' c_\pi~,
\end{eqnarray}
where $c_\pi=\sqrt3 \alpha_s /({\pi F_{\pi}})$,
$\alpha_s = g^2/4\pi$ and $F_\pi = 93$ MeV.
For the derivation of an effective Hamiltonian for this process, we use
the sum of the Lagrangians (\ref{lag}) and (\ref{lquglu}),
\begin{eqnarray}
\label{ltot}
{\cal L}_{\rm eff}(A)={1 \over 2}(\dot{A}_i)^2- c_\pi \eta'\dot{A}_i B_i
+ \dots~.
\end{eqnarray}
From (\ref{ltot}) follows
\begin{eqnarray}
E_i=\dot{A}_i - c_\pi \eta'B_i~,
\end{eqnarray}
such that the effective Hamiltonian reads
\begin{eqnarray}
{\cal H}_{\rm eff}&=& E_i\dot{A}_i - {\cal L}_{\rm eff}(A)\noindent \\
&=& \frac{1}{2} (E_i + c_\pi \eta'B_i)^2 + \dots~.
\end{eqnarray}
The effective Hamiltonian of the $\eta' \rightarrow \eta'$ transition
has the form
\begin{eqnarray}
{{\cal H}}_{\eta'\rightarrow \eta'}=
{1 \over 2}(\eta')^2 (B_i^{a}(b))^2 c_\pi^2~,
\end{eqnarray}
where $b$ is the constant part of the $A$ field.
Thus, we have
\begin{eqnarray}
\label{eta2}
\Delta {m^2_{\eta'}} = (\frac{3\alpha_s^2}{\pi^2 F_{\pi}^2})~<
(B_i^a(b))^{2} >.
\end{eqnarray}
This equation together with Eqs.(\ref{26}) and (\ref{mg}) leads to a relation
between the quasigluon and $\eta'$ masses:
\begin{eqnarray}
\label{eta3}
\Delta {m^2_{\eta'}} =\frac{2}{3} \frac{\alpha_s}{\pi^3 F_{\pi}^2}~m_g^4~.
\end{eqnarray}
Choosing in the low-energy region $\alpha_s \simeq 1 $ we can estimate
>from this formula the value of the quasigluon mass and by formula (\ref{mg})
then the corresponding value of the gluon condensate as
\begin{eqnarray}
\label{mg1}
m_g = 0.71~ {\rm GeV}~~,
\ \ \ \ G^2 = 0.011 ~{\rm GeV}^4~,
\end{eqnarray}
which is in the agreement with earlier estimates \cite{biro1,zakharov}.
Note that the process of the decay of $\eta'$ into gluon fields
by means of the Hamiltonian (\ref{lag}) is forbidden, as
the vacuum expectation
value from the magnetic field $< 0_B|B_i^a(b)|0_B >$ is equal to zero.
\section{Conclusions}
\label{sec:end}
In the present paper we have considered the consequences of a squeezed
vacuum for the single particle excitation spectrum in the gluon sector of QCD
by applying the concept of the Bogoliubov theory of superfluidity to field
theory.
We have considered a squeezed homogeneous colourless condensate of zero
momentum gluon pairs.
The macroscopic occupation (squeezing) of the zero momentum mode has been
achieved through Wick reordering of the QCD Hamiltonian and is
characterized by a parameter $\tilde{C}$ which describes the magnitude of the
condensate.
The presence of the condensate leads to the occurrence of a gluon mass
and thus to spontaneous gauge symmetry breaking, i.e.
the gauge invariance of the Hamiltonian is not shared by the squeezed vacuum.
Instead of eliminating unphysical degrees of freedom by fixing a gauge
we use a projection operator method
resulting from the formal solution of Gauss law.
We show that the occurence of a condensate leads to a destruction of the
projection property
so that the generation of a mass is accompanied by the appearance
of the necessary longitudinal component for the gauge field.
We found that the quasigluon
spectrum depends on the parameter $\tilde{C}$
of the squeezed representation,
which yielded a relation between the quasigluon mass and the
gluon condensate.
We have fixed the quasigluon mass from the squared mass difference
$m_{\eta'}^2 - m_{\eta}^2 = 0.616 ~ {\rm GeV}^2$ of $\eta$ and
$\eta'$ and found that the corresponding value of the gluon condensate
$G^2 = 0.012 {\rm GeV}^4$ then agrees well with the
standard value $G^2 = 0.01 {\rm GeV}^4$ obtained by
Shifman, Vainshtein and Zakharov.
In this paper we have
populated the zero momentum state directly with quasigluons whose
dispersion relation was then determined selfconsistently by
demanding diagonality of the one-particle sector of the
Hamiltonian for nonzero momentum.
This should be considered as a first attempt to explain the concept
of the squeezed vacuum.
In a more rigorous treatment the zero momentum state should first be occupied
macroscopically with massless gluons using the freedom due to the
infrared singularity of massless theories, and the resulting
one-particle sector of the Hamiltonian should then be diagonalized
by transformation to quasiparticles.
This has been carried out for the much simpler $\lambda\phi^4$ theory
in a separate work \cite{PaBla} which shows how in a more rigorous treatment
the renormalization of both the mass and the bare coupling are included.
Important extensions of the present approach include the study of
small deviations from a homogeneous condensate, the inclusion of quark degrees
of freedom and the generalization to finite temperatures.
These issues are currently under investigation and will be reported in a
forthcoming paper.
\section*{Acknowledgement}
We are grateful to D. Ebert, A.V. Efremov, E.A. Kuraev, H. Leutwyler,
L.N. Lipatov, D.V. Shirkov, A.A. Slavnov,
O.I. Zav'ialov for fruitful discussions.
We thank S. Schmidt and D. Rischke for useful comments and
critical reading of the manuscript.
The work was supported in part by the RFFI, Grant No. 95-02-14411 and
the Federal Minister for Research and Technology (BMFT) within
Heisenberg-Landau Programme.
Two of us (V.N.P. and M.K.V.) acknowledge the financial support provided
by the Max-Planck-Gesellschaft and the hospitality of the MPG Arbeitsgruppe
''Theoretische Vielteilchenphysik'' at the Rostock University, where part
of this work has been done.
One of the authors (V.N.P.) acknowledges the hospitality of the
International Centre for Theoretical Physics in Trieste.
\begin{appendix}
\section{The weakly nonideal Bose gas model}
\label{app:bogmod}
The Bogoliubov theory of the weakly interacting Bose gas \cite{nn} is
described by the nonrelativistic Hamiltonian
\begin{eqnarray}
\label{w}
H = \sum_p \frac{p^2}{2m}\,a^+_p a_p +
\frac{U_0}{2V}\,\sum_{p_1 p_2 p_1^{'} p_2^{'}}\,
a^+_{p_1} a^+_{p_2} a_{p^{'}_2} a_{p^{'}_1}
\delta_{p_1+p_2, p_1^{'}+ p_2^{'}}~.
\end{eqnarray}
The operators $a^+_p, a_p$ are the creation and annihilation operators
of bosons in the state $p$ satisfy the commutation relations
\begin{eqnarray}
\label{cr}
[a_p,a^+_{p^{'}}]= \delta_{{pp}^{'}},
\end{eqnarray}
where $p$ stands for momentum and internal quantum numbers.
The coupling constant $U_0$ is defined by the scattering amplitude of
slow particles, $V$ is the volume of the system. We figure out the original
work of Bogoliubov \cite{nn},
for a more recent presentation of the theory of the
weakly interacting Bose gas see \cite{FW}, \cite{R}.
In the Bogoliubov derivation of the superfluid
spectrum one can distinguish three points:
\begin{enumerate}
\item A macroscopic occupation of the zero momentum
state ($p=0$) is assumed so that in the thermodynamic limit a finite density
of the condensate
\begin{eqnarray}
\label{nb}
n_B = {\mbox{\rm lim}}_{\rm th} \frac{N_0}{V} \neq 0
\end{eqnarray}
occurs, where $N_0$ denotes the number of particles in the condensate.
Therefore, the operators $a^+_0,a_0$ in the thermodynamic limit (\ref{nb})
are described as c-numbers
\begin{eqnarray}
\label{N}
a_0 \simeq a^+_0 \simeq \sqrt{N_0}.
\end{eqnarray}
This description, strictly speaking, should be completed by defining a
representation for the condensate which in the present work
is given below.
\item The next step is the expansion of the Hamiltonian (\ref{w})
around these c-numbers
\begin{eqnarray}
\label{exp}
\sum a^+_{p_1} a^+_{p_2} a_{p^{'}_2} a_{p^{'}_1} & = &
N^2_0 + N_0 \sum_{p \neq 0} (a^+_p a^+_{-p} + a_p a_{-p} + 4a^+_p a_p) +
O[a_{p\neq 0}^3] ~.
\end{eqnarray}
Taking into account the conservation of the total number of particles $N$ and
rewriting
$$
N_0=N - \sum_{p\neq 0}a^+_p a_p ~~~;
\left (\frac{N-N_0}{N} \ll 1
\right ),
$$
in Eq. (\ref{exp}) and neglecting terms of higher than second order
in the particle operators $a_{p\neq 0}, a^+_{p\neq 0} $,
the Hamiltonian (\ref{w}) transforms into
\begin{eqnarray}
\label{ha2}
H =\frac{N}{2} \nu + \sum_{p\neq 0} \left [a^+_p a_p \varepsilon _p +
\frac{\nu}{2}\;(a^+_p a^+_{-p} + a_p a_{-p}) \right ] + O[a_{p\neq 0}^3],
\end{eqnarray}
where
\begin{eqnarray}
\varepsilon_p =\frac{p^2}{2m} + \nu, ~~~ \nu=U_0\frac{N}{V}.
\end{eqnarray}
\item The last step is the diagonalization of (\ref{ha2}) using the Bogoliubov
transformation, i.e. the transition to the operators of quasiparticles $b^+_p $
and $b_p $ for $p \neq 0$
\begin{eqnarray}
\label{b1}
b_p &=& U^{-1} a_p U = {\rm cosh}(f_p) a_p + {\rm sinh}(f_p) a^+_{-p},
\nonumber\\
b^+_p&=& U^{-1} a^+_p U = {\rm cosh}(f_p) a^+_p + {\rm sinh}(f_p) a_{-p},
\\
\nonumber
\end{eqnarray}
where
\begin{eqnarray}
\label{b}
U =\exp \left \{\sum_p \frac{f_p}{2} (a^+_pa^+_{-p}\, -
\,a_p a_{-p}) \right \}~.
\end{eqnarray}
The $b_p$ satisfy the same commutation relations as the $a_p$.
The function $f _p $ is found from the requirement of the disappearance
of nondiagonal terms as
\begin{eqnarray}
\label{fp}
f_p = \frac{1}{2} {\rm arth} \left[\frac{\nu}{\varepsilon_p}\right]~,
\end{eqnarray}
so that the Hamiltonian (\ref{ha2}) gets the form
\begin{eqnarray}
H =\frac{N}{2} \nu -{1\over 2}\sum_{p\neq 0}
\left(\varepsilon_p-\omega_B(p)\right)
+ ~ \sum_{p\neq 0} b^+_p b_p~ \omega_B(p) +~ O[b_{p\neq 0}^3]~,
\end{eqnarray}
where $\omega_B$ is the spectrum of excitations in a superfluid liquid
\begin{eqnarray}
\label{bf}
\omega^2_B(p) = \varepsilon^2_p - \nu^2 =\left (\frac{p^2}{2m}\right )^2 +
\frac{p^2}{2m} \;\left ( 2U_0\; \frac{N}{V} \right ) \;\;,
\end{eqnarray}
which is determined by the condensate density
$n_B = N_0/V \cong N/V$ and by the coupling constant
$U_0$.
\end{enumerate}
In the low momentum region this expression describes the Landau sound
and the particle excitations with energy $\left (p^2/2m\right)$
disappear.
Note that the vacuum energy $E_0$ contains a divergent sum which can be
renormalized by expressing it in terms of the physical scattering length $a$
instead of the bare coupling $U_0$ (see \cite{FW}, p. 318).
In his paper \cite{nn}, Bogoliubov did not determine the
representation of the condensate state for which
Eq. (\ref{N}) is fulfilled.
Usually one assumes that of the coherent state
\begin{eqnarray}
\label{c}
|0_{\rm C} >~ = \exp\left\{\sum_p c_0(a^+_0 + a_0)\right\}~ |0>~,~
c_0 =\sqrt{N_0}~,
\end{eqnarray}
for which holds
\begin{eqnarray}
<0_{\rm C}|a_0|0_{\rm C}>~=~<0_{\rm C}|a^+_0|0_{\rm C}>~=\sqrt{N_0}~,
\end{eqnarray}
corresponding to Eq. (\ref{N}).
However, to get the Bogoliubov result it is enough to assume
the weaker condition
\begin{eqnarray}
\label{nn}
(a^+_0)^2 \simeq a^2_0 \simeq a^+_0\; a_0 \sim N_0~,
\end{eqnarray}
rather than (\ref{N}).
These relations are fulfilled for the representation of the condensate state
which is given by the same Bogoliubov transformation as for
$p \neq 0$ (\ref{b}):
\begin{eqnarray}
\label{s}
|\,0_{B} > = U_{B}^{-1}~~|~0 >~,
\end{eqnarray}
where
\begin{eqnarray}
\label{b0}
U_{B_0} =\exp \left \{ \frac{f_0}{2} (a^+_0a^+_{-0}\, - \,a_0 a_{-0}) \right
\}~.
\end{eqnarray}
The inverse of the unitary operator (\ref{b}) defines also
the transformation of the old into a new vacuum state for momenta $p \neq 0$.
In quantum optics the vacuum $ |\,0_B >$ is called 'squeezed vacuum', see e.g.
\cite{optics}.
For the ''squeezed'' vacuum representation
(\ref{s}) of the condensate we have the realization (\ref{nn})
\begin{eqnarray}
\label{a2}
<0_B\,|\,a^2_0\,|\,0_B>
& = & <0_B\,|\,(a^+_0)^2\,|\,0_B > ~= - {\rm cosh}{f_0} ~{\rm sinh}{f_0} \;\;,
\\
\nonumber
<0_B\,|\,a^+_0 a_0\,|\,0_B> & = & ({\rm sinh}{f_0})^2 = N_0~,
\label{a+a}
\end{eqnarray}
and at large $N_0$ (\ref{nn}) means that
\begin{eqnarray}
\label{n0}
- -{\rm cosh}{f_0}~ {\rm sinh}{f_0} &\simeq & ({\rm sinh}{f_0})^2
\simeq N_0 \rightarrow \infty\;\;, \nonumber\\
f_0 \sim -\frac{1}{2}\ln{4N_0}~.
\end{eqnarray}
The choice of the squeezed vacuum is more prefable from the point view
of a general consideration of all momenta, $p=0$ {\it and} $p \neq 0$.
Together, the Bogoliubov transformation now is given by the product $UU_{B}$.
\section{Gauge invariant variables}
\label{app:gaugi}
The unphysical components of the gluon fields are formally eliminated by the
transformation to gauge invariant variables \cite{dirac,perv1} which are
functionals constructed using the solution (\ref {10})
\begin{eqnarray}
\label{12}
{A}^I_i [{\bf A}] \equiv V({\bf A}) ({A}_i + \partial_i) V({\bf A})^{-1}~.
\end{eqnarray}
The matrix $V$ is defined from the equation
\begin{eqnarray}
\label{DefV}
V({A}_0[{\bf A}] + \partial_0) V^{-1} = 0 \Rightarrow V({\bf A}) =
T{\rm exp}(\int^t
{A}_0[{\bf A}] dt^{'})
\end{eqnarray}
(up to a stationary matrix as the time boundary condition).
The invariance of these functionals under arbitrary time dependent
gauge transformations $v({\bf x},t)$
\begin{equation}
\label{GI2}
{A}^I_i [{\bf A}^v] = V({\bf A}) v^{-1} v ({A_i} + \partial_i)
v\ v^{-1} V({\bf A}) = {A}^I_i [{\bf A}]~,
\end{equation}
follows from the transformation properties of $A_i$ in (\ref{GI})
and of $V({\bf A})$:
\begin{eqnarray}
V({\bf A}^v) = V({\bf A}) v^{-1}~.
\end{eqnarray}
which follows from (\ref{DefV}) and (\ref{GI}).
As consequence of this the variables (\ref{12}) represent only $2~(N_c^2-1) $
independent degrees of freedom. They contain hidden projection operators
onto generalized transverse components similar to the magnetic field which
satisfies the Bianchi identity (\ref{BI}).
The projection operator is contained (different to the QED case) not in the
$A_i^I$ themselves, but only in their time derivatives
$\dot{A}_i^I$ which satisfy the ``Bianchi type'' identities
\begin{equation}
\label{DAI}
D^{ab}_i({\bf A}^I) \dot{A}_i^{I^b} \equiv 0~.
\end{equation}
In the terms of the functionals (\ref{12}) the Lagrangian (\ref{1})
takes the form
\begin{equation}
\label{17}
{\cal L}^{Red} ({\bf A}^I) = \frac{1}{2}~ \left [\dot{A}^{I^a2}_i
- - B^{a2}_i({\bf A}^I) \right ]~.
\end{equation}
The canonical momenta to the spatial fields $A^{Ia}_i$
are
\begin{equation}
{E^I_i}^a \equiv\frac{\delta {\cal L}}{\delta \dot{A^I_i}^a}= \dot{A^I_i}^a~.
\end{equation}
The Hamiltonian becomes
\begin{equation}
H^{Red}({\bf A}^I,{\bf E}^I)=
\int d^3x {1\over 2}\left[E_i^{Ia2}+B_i^a({\bf A}^I)^2 \right]~.
\end{equation}
It follows from (\ref{DAI}) that the electric fields ${E^I}_i^a $ satisfy
the Gauss constraint
\begin{equation}
D^{ab}_i({\bf A}^I) {E^I}^b_i = 0~.
\end{equation}
Like the Bianci identity (\ref{BI}) for the magnetic field, this shows
the generalized transversality of the invariant
electric fields ${E^I}^a_i$.
In order to quantize the theory one could then like in QED
impose the following canonical commutation relations on the physical
variables $A^I$ and $E^I$:
\begin{equation}
[E_i^{Ia}({\bf x}),A_j^{Ib}({\bf x}')]=
i\delta^{ab}\delta_{ij}\delta({\bf x}-{\bf x}')~.
\end{equation}
In QCD, however, this leads to a contradiction when applying the
covariant derivative on it.
Instead one has to impose canonical commutation relations directly on
the three cartesian fields $A_i$ and $E_i$ and write both $E^I$ and $A^I$
as functionals of $E$ and $A$. Whereas the form of $A^I[{\bf A}]$ is known
by construction (\ref{12}), the functional form of $E^I[{\bf E},{\bf A}]$
in terms of
$E$ and $A$ has not been found yet, but is subject of intensive research
\cite{Kved}.
However, even if this problem is solved there remains still the question of
correct ordering of the operators $A$ and $E$.
\section{Wick reordering of gluon fields}
\label{app:wick}
In this appendix we perform the Wick reordering of the magnetic part of the
Hamiltonian (\ref{17}),
\begin{eqnarray}
\label{b2n}
\frac{1}{2}\int {\rm d}^3 x ~:{B_i^a}^2({\bf A}): &=&
\frac{1}{2}\int {\rm d}^3 x ~\left\{:(\partial_j A_k^a) \delta^T_{kl}
(\partial_jA_l^a):
+2 g f^{abc} :(\partial_j A_k^a) A_j^b A_k^c:
+\frac{1}{2} g^2 f^{abc} f^{ade} : A_j^bA_k^cA_j^dA_k^e : \right\}~,
\end{eqnarray}
which reads in momentum space
\begin{eqnarray}
\label{b2p}
\frac{1}{2}\int {\rm d}^3 x ~ : {B_i^a}^2({\bf A}) : &=&
\frac{1}{2}\bigg\{\sum_{p} : A_k^a(p)\left[p_kp_l - p^2
\delta_{kl}\right]A_l^a(-p) :\nonumber\\
& &+2 g^2 \frac{1}{\sqrt{V}} f^{abc} \sum_{p_1 p_2} i(p_1+p_2)_j :
A_k^a (p_1+p_2) A_j^b(-p_1) A_k^c(-p_2) :\nonumber\\
& & + \frac{1}{2} g^2 \frac{1}{V}f^{abc} f^{ade}
\sum_{p_1 \dots p_4} :A_j^b(p_1)A_k^c(p_2) A_j^d(p_3) A_k^e(p_4) :
\delta_{p_1+p_2+p_3+p_4,0}\bigg\}~.
\end{eqnarray}
The reodering of the first two terms on the r.h.s. of Eq. (\ref{b2p}) gives
no extra contraction contribution.
For the Wick reordering of the third term we write
\begin{eqnarray}
\label{a4}
f^{abc} f^{ade} : A_j^b(p_1)A_k^c(p_2)A_j^d(p_3)A_k^e(p_4) : &=&
f^{abc} f^{ade} A_j^b(p_1)A_k^c(p_2)A_j^d(p_3)A_k^e(p_4)\nonumber\\
&&-f^{abc} f^{ade} <A_j^b (p_1)A_j^d(p_3)> A_k^c(p_2) A_k^e(p_4)\nonumber\\
&&
- -f^{abc} f^{ade} A_j^b(p_1) A_j^d(p_3) <A_k^c(p_2) A_k^e(p_4)>\nonumber\\
&&-f^{abc} f^{ade} <A_j^b(p_1) A_k^c(p_2)> A_j^d(p_3) A_k^e(p_4)\nonumber\\
&&
- -f^{abc} f^{ade} A_j^b(p_1) A_k^e(p_4) <A_k^c(p_2) A_j^d(p_3)> \nonumber\\
&&+f^{abc} f^{ade} <A_j^b(p_1) A_j^d(p_3)> <A_k^c(p_2) A_k^e(p_4)>\nonumber\\
&&
+f^{abc} f^{ade} <A_j^b(p_1) A_k^c(p_2)> <A_j^d(p_3) A_k^e(p_4)>~.
\end{eqnarray}
Using formula (\ref{aa}) and the identities
\begin{eqnarray}
\sum_{abc} f^{abc} ~f^{abc} &=& N_c (N_c^2 -1)~,\nonumber\\
\sum_{ab} f^{abc} ~f^{abe} &=& N_c \delta_{ce}~,\nonumber
\end{eqnarray}
in Eq. (\ref{a4}) we thus have
\begin{eqnarray}
\label{25}
\sum_{p_1 \dots p_4}f^{abc} f^{ade} &:
& A_j^b(p_1)A_k^c(p_2)A_j^d(p_3)A_k^e(p_4) :
\delta_{p_1+p_2+p_3+p_4,0} = \nonumber\\
&& 6 N_c (N_c^2 - 1 ) C^2
- - 4 N_c C \sum_{p} A_i^a(p) A_i^a(-p) \nonumber\\
&& + \sum_{p_1 \dots p_4}f^{abc} f^{ade}
A_j^b(p_1)A_k^c(p_2)A_j^d(p_3)A_k^e(p_4)
\delta_{p_1+p_2+p_3+p_4,0}\nonumber\\
&=& 6 N_c (N_c^2 - 1 ) {(C^B)}^2
+4 N_c C^B \sum_{p} ::A_i^a(p) A_i^a(-p):: {\rm e}^{f_p +f_{-p}}
\nonumber\\&&+ \sum_{p_1 \dots p_4}f^{abc} f^{ade}
::A_j^b(p_1)A_k^c(p_2)A_j^d(p_3)A_k^e(p_4) ::
{\rm e}^{f_{p_1} +f_{p_2} +f_{p_3} +f_{p_4} }
\delta_{p_1+p_2+p_3+p_4,0} ~.
\end{eqnarray}
Putting all together,
\begin{eqnarray}
\label{b2r}
\frac{1}{2}\int {\rm d}^3 x ~: {B_i^a}^2({\bf A}): &=& g^2\frac{3}{2}
\frac{N_c}{V} (N_c^2 - 1 ) {(C^B)}^2
+\frac{1}{2} \sum_{p} \left[(p^2 + 2 g^2 N_c C^B/V) \delta_{ij} - p_i
p_j \right] :: A_i^a(p) A_j^a(-p) :: {\rm e}^{f_p +f_{-p}}
\nonumber\\&&+\frac{1}{4 V} g^2 \sum_{p_1 \dots p_4}f^{abc} f^{ade}
::A_j^b(p_1)A_k^c(p_2)A_j^d(p_3)A_k^e(p_4) ::
{\rm e}^{f_{p_1} +f_{p_2} +f_{p_3} +f_{p_4} }
\delta_{p_1+p_2+p_3+p_4,0} ~.
\end{eqnarray}
This proves the result of Eq. (\ref{b2}) used in the main text.
\end{appendix}
\newpage
|
\section{Introduction}
It is believed that in two-dimensions (2D) even small disorder localizes
all electron states.\cite{lee85} First hint for the absence of the
mobility threshold in 2D came from the calculation of the
weak localization correction, $\delta\sigma(\omega)$, to the conductivity
at finite frequency $\omega$. It was shown\cite{gor79} that in the
limit $k_Fl\gg 1$ one has
$\delta\sigma(\omega)/\sigma_0={1\over k_F l}\ln|\omega\tau|$,
where $k_F$ is the Fermi momentum, $l$ is the mean free path, and
$\tau$ is the elastic scattering time; $\sigma_0=(e^2/h)k_Fl$ is the
Drude conductivity. The logarithmic singularity in
$\delta\sigma(\omega)$ indicates that the zero-temperature ($T=0$)
conductivity is metallic only if $\omega\gg \omega_0$, where
\begin{eqnarray}\label{1}
\omega_0\sim {1\over\tau}e^{-\pi k_F l}
\end{eqnarray}
is the characteristic frequency which marks the crossover to the
exponentially localized regime.\cite{lee85} From the
finite size, $L$, correction to the zero-frequency conductivity, which
is of the order of ${1\over k_F l}\ln (L/l)$, one can estimate the
localization length $\xi$ as\cite{lee85}
\begin{eqnarray}\label{2}
\xi=l\exp\Biggl({\pi \over 2}k_F l\Biggr).
\end{eqnarray}
Therefore, starting from metal, one cannot describe the $T=0$ behavior
of the conductivity at frequencies $\omega\sim\omega_0$. The adequate
language for this region would be the language of the localized
states. Within this language the conductivity originates from the
transitions between the localized states induced by an external a.c.
field.
Let us briefly remind the corresponding derivation of
$\sigma(\omega)$, carried out by Mott\cite{mott70} in the strongly
localized regime. The Hamiltonian for strongly localized electrons reads
\begin{eqnarray}\label{3}
{\cal H}=\sum_i \epsilon_i c_i^{\dagger}c_i+
\sum_{ij} I_{ij} (c_i^{\dagger}c_j+c_j^{\dagger}c_i),
\end{eqnarray}
where $\epsilon_i$ is the energy of a localized
state centered at ${\bf r}={\bf r}_i$ and $I_{ij}$ is the overlap
integral which falls off exponentially with distance:
$I_{ij}=I_0\exp(-|{\bf r}_i-{\bf r}_j|/a)$, where $a$ is the
size of the wave-function of the localized electron. The general
expression for the conductivity is given by the Kubo formula
\begin{eqnarray}\label{4}
\sigma(\omega)={ie^2\omega\over \hbar A}
\sum_{\bf ij}|\langle {\bf i}|x|{\bf j} \rangle|^2
{n_{\bf i}-n_{\bf j} \over \omega + \omega_{\bf ij} + i0}
\end{eqnarray}
where $\langle {\bf i}|x|{\bf j} \rangle$ is the matrix element of
$x$ calculated from the {\em exact}
eigenstates $|{\bf i} \rangle$ and $|{\bf j} \rangle$ of the
Hamiltonian (\ref{3}) with energies $E_{\bf i}$ and $E_{\bf j}$,
$\hbar\omega_{\bf ij}=E_{\bf i}-E_{\bf j}$ and $n_{\bf i}$ is the
occupation number of the state $|{\bf i} \rangle$.
The dissipative conductivity, $\mbox{Re}\sigma(\omega)$, is determined by
the pairs of states with $\omega_{\bf ij}=\omega$.
At $\hbar\omega\ll I_0$ the spatial separation between bare
states $|i\rangle$ and $|j\rangle$ is much larger than $a$. Then in the
calculation of eigenstates of a resonant pair one should take
into account the overlap $I_{ij}$ within this pair {\em only} and neglect
the overlap with all the other localized states. This gives
\begin{eqnarray}\label{5}
|{\bf i}\rangle = {\epsilon_i -\epsilon_j\over\Gamma}|i\rangle +
{2I_{ij}\over\Gamma}|j\rangle,~~~~~~
|{\bf j}\rangle = {2I_{ij}\over\Gamma}|i\rangle +
{\epsilon_j-\epsilon_i \over\Gamma}|j\rangle.
\end{eqnarray}
The corresponging energies are
\begin{eqnarray}\label{6}
E_{\bf i,j}=(\epsilon_i+\epsilon_j)/2 \pm \Gamma/2,~~~~~
\Gamma=[(\epsilon_i-\epsilon_j)^2+4I_{ij}^2]^{1/2}=\hbar\omega_{\bf ij}.
\end{eqnarray}
Using (\ref{5}) the matrix element in (\ref{4}) takes the form
\begin{eqnarray}\label{7}
\langle {\bf i}|x|{\bf j}\rangle=(x_i-x_j)I_{ij}/\Gamma.
\end{eqnarray}
The contribution to $\mbox{Re}\sigma(\omega)$ from
pairs with shoulder ${\bf r}$ can be presented as
\begin{eqnarray}\label{8}
\mbox{Re}\sigma(\omega,{\bf r})=
{e^2\over\hbar}{\pi x^2 I^2(r)\over \hbar \omega}F(\hbar\omega,r),
\end{eqnarray}
where $F(\hbar\omega,r)$ is the density of pairs with shoulder
${\bf r}$ and
excitation energy $\hbar\omega$ [here $I(r)=I_0e^{-r/a}$]. The
density $F(\hbar\omega,r)$ is determined by the condition
that the pair is singly occupied (with energies on the opposite sides
from the Fermi level). Then we have
\begin{eqnarray}\label{9}
F(\hbar\omega,r)&=&g^2\int d\epsilon_1\int d\epsilon_2
\theta\Biggl({\hbar\omega\over 2}+{\epsilon_1+\epsilon_2\over 2}\Biggr)
\theta\Biggl({\hbar\omega\over 2}-{\epsilon_1+\epsilon_2\over 2}\Biggr)
\delta\Biggl[\sqrt{(\epsilon_1-\epsilon_2)^2+4I^2(r)}-\hbar\omega\Biggr]
\nonumber\\
&=&{2g^2(\hbar\omega)^2\over\sqrt{(\hbar\omega)^2-4I^2(r)}},
\end{eqnarray}
where $g$ is the density of states and $\theta(x)$ is the
step-function. Substituting (\ref{9}) into
(\ref{8}) and integrating over ${\bf r}$ we get
\begin{eqnarray}\label{10}
\mbox{Re}\sigma(\omega)=
{e^2\over\hbar}(2\pi^2 g^2 \hbar\omega)
\int_0^{\infty}{drr^3I^2(r)\over\sqrt{(\hbar\omega)^2-4I^2(r)}}.
\end{eqnarray}
For $\hbar\omega\ll I_0$ the main contribution to the integral comes
from $r\sim r_{\omega}=a\ln(2I_0/\hbar\omega)\gg a$ and one obtains
\begin{eqnarray}\label{11}
\mbox{Re}\sigma(\omega)=\sqrt{2}\pi^2
{e^2\over \hbar}(g^2a\hbar^2\omega^2 r_{\omega}^3).
\end{eqnarray}
Although we cannot show it explicitly, we argue below that the Mott
expression (\ref{11}) for $\mbox{Re}\sigma(\omega)$ is valid also
for the Anderson insulator with $k_Fl\gg 1$. When applying
(\ref{11}) to the Anderson insulator one should replace $a$ by $\xi$
from (\ref{2}) and substitute $g=m/2\pi\hbar^2$ ($m$ is the electron
mass). However, the question
remains: what is the magnitude of $I_0$ in this case? A plausible
estimate can be obtained in the spirit of the Thouless picture of
localization.\cite{tho77}
By definition, $I_0$ represents the splitting of energy
levels of two neighboring localized states (with centers at
distance $\sim \xi$). The estimate for $I_0$ emerges if one equates
this splitting to the mean energy spacing for localized
states centered within the area $\sim \xi^2$, so that $I_0\sim 1/g\xi^2$
(see also Ref.~\onlinecite{imr93}).
Note that in 1D case a similar argument leads to
$I_0\sim 1/g_1\xi$, $g_1$ being the 1D density of states.
Important is that in 1D the Kubo formula can be evaluated
exactly\cite{ber73} resulting in the 1D version of the Mott
formula, from which one can recover the above estimate for $I_0$.
Such a mapping was first established by Shklovskii and
Efros.\cite{shk81}
Using Eq.~(\ref{10}) one can formally evaluate
$\mbox{Re}\sigma(\omega)$ for $\hbar\omega > 2I_0$. In this case the
main contribution to the integral comes from $r\sim a$ and we get
\begin{eqnarray}\label{12}
\mbox{Re}\sigma(\omega)={3\pi^2\over 4}
{e^2\over \hbar}(I_0^2g^2a^4).
\end{eqnarray}
Certainly, the presentation of $I(r)$ in the form $I_0e^{-r/a}$ makes
sense only for $r\gg a$ This means that the numerical coefficient in
Eq.~(\ref{12}) is not reliable.
Clearly, at large frequencies $\omega\gg\omega_0$ the conductivity of
the Anderson insulator
should have the Drude form. The fact that Eq.~(\ref{12}), calculated
for strongly localized electrons, is also frequency independent allows
us to assume that the description of a.c. transport in the Anderson
insulator based on Hamiltonian (\ref{3}) is accurate within a
numerical coefficient. In other words, we assume that despite the
complex structure of the electron wave-functions in the Anderson
insulator the energy dependence of the matrix elements calculated
between these functions is still given by (\ref{7}).
Eq.~(\ref{12}) provides yet another way to estimate $I_0$ in the limit
$k_Fl\gg 1$. Namely, it matches $\sigma_0$ if we take
$I_0\sim (k_Fl)^{1/2}/g\xi^2$. We see that the dependence of $I_0$ on
$\xi$ in both estimates is the same; the extra factor $(k_Fl)^{1/2}$
presumably can be accounted for the $\ln(k_Fl)$ corrections to the
exponent of $\xi$.
There is also another argument in favor of the above estimate for
$I_0$. The frequency dependence of $\mbox{Re}\sigma(\omega)$ in the
Anderson insulator becomes strong for $\omega\ll\omega_0$, whereas in the
picture of strongly localized electrons the demarcation frequency is
$\omega\sim I_0/\hbar$. Equating $I_0$ to $\hbar\omega_0$ we get
$I_0=k_Fl/g\xi^2$, with another extra factor $k_Fl$.
The simplified description of the Anderson insulator based on the
Hamiltonian (\ref{3}) allows one to include into consideration
the Coulomb correlations (i.e., the correlations in the occupation
numbers of the localized states caused by electron-electron
interactions) using the ideas first spelled
out in Refs.~\onlinecite{efr81,shk81,efr85}. This is the main goal of the
present paper. We study the effect of Coulomb correlations on both
$\mbox{Re}\sigma(\omega)$ and $\mbox{Im}\sigma(\omega)$. The most
drastic conclusion we come to is that due to modification of
$\mbox{Im}\sigma(\omega)$ by the Coulomb correlations
{\em a system of localized
electrons can support surface plasmons within a certain frequency
range}. We also show that these plasmons cause an additional structure
in the behavior of $\mbox{Re}\sigma(\omega)$ at
$\omega > \omega_0 \sim I_0/\hbar$.
The paper is organized as follows. In Section \ref{sec2} we analyze
the polarizability of the localized system in the absence of Coulomb
correlations. In Section \ref{sec3} we introduce the Coulomb correlations
and find the dispersion law for the surface plasmons. In
Section \ref{sec3} we study the corrections to the dispersion law due
to the resonant scattering of plasmons by pairs of localized states.
In Section \ref{sec4} we calculate the plasmon contribution to
$\mbox{Re}\sigma(\omega)$. Section \ref{sec5} concludes the paper.
\section{Polarizability in the absence of Coulomb correlations}
\label{sec2}
In this section we demonstrate that without Coulomb correlations the
2D Anderson insulator cannot support surface plasmon.
In the framework of the linear response theory the dispersion law of a
plasmon, $\omega(q)$, is determined from the condition\cite{mahan}
\begin{eqnarray}\label{13}
1=v(q)\mbox{Re}{\bf P}(\omega,q),
\end{eqnarray}
where $v(q)$ is the Fourier component of the electron-electron
interaction $v(r)$ and ${\bf P}(\omega,q)$ is the polarization operator.
Within a standard approach ${\bf P}(\omega,q)$ is calculated for
non-interacting electrons described by Hamiltonian (\ref{3}):
\begin{eqnarray}\label{14}
P(\omega,q)={1\over A}
\sum_{\bf ij}|\langle {\bf i}|e^{i{\bf qr}}|{\bf j} \rangle|^2
{n_{\bf i}-n_{\bf j} \over \hbar\omega + E_{\bf i}-E_{\bf j} + i0}.
\end{eqnarray}
In the absence of disorder the eigenstates $|{\bf i} \rangle$ and
$|{\bf j} \rangle$ are the plain waves so that the matrix element in
(\ref{14}) reduces to the delta-function $\delta({\bf i-j-q})$. Then
evaluating $P(\omega,q)$ and substituting it into (\ref{13})
together with 2D Coulomb interaction $v(q)=2\pi e^2/\kappa q$
($\kappa$ is the dielectric constant) yields the surface
plasmon with the well-known dispersion law
\begin{eqnarray}\label{15}
\omega(q)=\Biggl[{2\pi n e^2 q\over m\kappa}\Biggr]^{1/2},
\end{eqnarray}
where $n$ is the 2D concentration of electrons.
The plasmon mode is undamped if $q < \omega/v_F$ where
$v_F=(4\pi n)^{1/2}\hbar/m$ is the Fermi velocity (for larger $q$
the Landau damping leads to a finite $\mbox{Im}P$). The latter
condition can be rewritten as $q<1/2a_B$ where
$a_B=\hbar^2\kappa/me^2$ is the Bohr radius.
In the case of a strong disorder the eigenstates ${\bf i}$ and ${\bf j}$
in (\ref{14}) are the localized states. The polarization operator
(\ref{14}) can be evaluated in a way similar to that employed in
the Introduction for calculation of a.c. conductivity. For small $q$
the matrix element in (\ref{14}) can be evaluated using (\ref{7}):
\begin{eqnarray}\label{16}
\langle {\bf i}|e^{i{\bf qr}}|{\bf j}\rangle=i{\bf qr}I(r)/\Gamma.
\end{eqnarray}
Then the contribution to $P(\omega,q)$ from the pairs
with the shoulder $r$ [density ${\cal P}(\omega,q,{\bf r})$ of
the polarization operator] can be presented as
\begin{eqnarray}\label{17}
\mbox{Re}{\cal P}(\omega,q,{\bf r})=({\bf qr})^2I^2(r)
\int_{2 I(r)}^{\infty}{d\Gamma\over \Gamma }
{F(\Gamma,r)\over (\hbar\omega)^2-\Gamma^2},
\end{eqnarray}
where $\Gamma$ and $F(\Gamma,r)$ is given by Eqs.~(\ref{6}) and
(\ref{9}), respectively (the integral is understood as principal
part). Substituting $F(\Gamma,r)$ into
(\ref{17}) and integrating over $\Gamma$ we obtain
\begin{eqnarray}\label{18}
\mbox{Re}{\cal P}(\omega,q,{\bf r})&=&2g^2({\bf qr})^2I^2(r)
\int_{2I(r)}^{\infty}d\Gamma
{\Gamma\over \sqrt{\Gamma^2-4I^2(r)}[(\hbar\omega)^2-\Gamma^2]}
\nonumber\\
&=&-{\pi g^2({\bf qr})^2I^2(r)\over \sqrt{4I^2(r)-(\hbar\omega)^2}},
{}~~~~~~~\mbox{for}~~\hbar\omega<2I(r),
\nonumber\\
&=&0,~~~~~~~~~~~~~~~~~~~~~~~~~~~~\mbox{for}~~\hbar\omega>2I(r).
\end{eqnarray}
We see that $\mbox{Re}{\cal P}(\omega,q,{\bf r})$ is either negative or
zero, so that Eq.~(\ref{13}) cannot be satisfied.
In fact, the result (\ref{18}) is almost obvious. Indeed, the
imaginary part of polarization operator density,
$\mbox{Im}{\cal P}(\omega,q,{\bf r})$,
at small $q$ differs from $\mbox{Re}\sigma(\omega,{\bf r})$ in
(\ref{8}) by a factor $e^2\omega/q^2$, so it follows from (\ref{8})
and (\ref{9}) that
\begin{eqnarray}\label{19}
\mbox{Im}{\cal P}(\omega,q,{\bf r})
&=&-{\pi g^2({\bf qr})^2I^2(r)\over \sqrt{(\hbar\omega)^2-4I^2(r)}},
{}~~~~~~~\mbox{for}~~\hbar\omega>2I(r),
\nonumber\\
&=&0,~~~~~~~~~~~~~~~~~~~~~~~~~~~~\mbox{for}~~\hbar\omega<2I(r).
\end{eqnarray}
Since $\mbox{Re}{\cal P}(\omega,q,{\bf r})$ and
$\mbox{Im}{\cal P}(\omega,q,{\bf r})$ are connected via the Kramers-Kronig
relation, the form (\ref{19}) immediately follows from (\ref{18}).
\section{Coulomb correlations and surface plasmon}
\label{sec3}
In the previous section the polarization operator was evaluated using
the pair density (\ref{9}) which was derived for non-interacting
electrons. As it was first pointed out by Efros\cite{efr81},
interactions modify strongly the density of singly-occupied pairs. The
underlying physics is the following. A pair can be singly-occupied
even if both energy states reside below the Fermi level. The right
condition for the pair to be singly-occupied is that the addition of a
second electron (which interacts with the first one) is energetically
unfavorable. Such a Coulomb correlations effectively enhance the
density of ``soft'' pairs (i.e., the pairs with small excitation
energy $\Gamma$).
Our goal is to apply the latter argument, which was presented for
strongly localized system, to the Anderson insulator with large
$\xi$. In order to do so we will adopt two assumptions:
$(i)$ The interactions do not change the localization radius $\xi$.
$(ii)$ The estimate for the overlap integral, $I_0\sim 1/g\xi^2$,
is unchanged in the presence of interactions.
In other words, we assume that switching on the interactions leads to
the Coulomb shifts of the eigenenergies but does not affect the wave
functions. Note that assumptions $(i)$ and $(ii)$ contradict
those made in Refs.~\onlinecite{ale94} and \onlinecite{pol93},
respectively.
As we will see below, the relevant pairs would be those with shoulder
$r\sim\xi$. In other words, the relevant transitions shift the
position of electron by $\sim\xi$. To establish the form of the density
of singly occupied pairs $F(\Gamma,r)$ with such a shoulder one can
argue as follows. An isolated region of a size $\xi$ can be viewed as a
small metallic granule. The transfer of an additional electron into
this granule leads to the charging energy $U=e^2/2C$, where $C$ is the
capacitance of a granule. In other words, the levels in the granule
get shifted\cite{shk82} by an amount $\sim U$. Consider now two
neighboring granules and assume that $U\gg I_0$. Due to aforementioned
charging effect the highest occupied levels in the two granules
typically differ by $\sim U$. Let for concreteness the highest
occupied level in the first granule be higher by $U$ than in the
second one. Then the sought singly occupied pair with frequency
$\Gamma$ can be composed from the top occupied states in the first
granule (these states should belong to the energy interval $\Gamma+U$
measured from the highest occupied level) and the empty states in the
second granule. Then the density of pairs $F(\Gamma,r)$
(with $r\sim \xi$) can be estimated as $g^2(\Gamma+U)$. With the
energy splitting taken into account it can be written as
\begin{eqnarray}\label{20}
F(\Gamma,r)=
{2g^2\Gamma(\Gamma +U)\over\sqrt{\Gamma^2-4I^2(r)}}.
\end{eqnarray}
Then at $U=0$ we return to (\ref{9}). Certainly, our consideration,
based on artificial arranging the localized states into the granules,
provides only the
order of magnitude estimate of $F(\Gamma,r)$. In particular, the
numerical coefficient in (\ref{20}) cannot be found from such a
consideration. Our choice of numerical coefficient in (\ref{20})
provides matching with a similar expression for strongly localized
regime.\cite{shk81,efr85,efr85b}
With the pair density (\ref{20}) we can now easily evaluate
$\mbox{Re}{\cal P}(\omega,q,{\bf r})$. Calculating the integral over
$\Gamma$ in (\ref{17}) yields
\begin{eqnarray}\label{21}
\mbox{Re}{\cal P}(\omega,q,r)
&=&-{\pi g^2({\bf qr})^2I^2(r)\over \sqrt{4I^2(r)-(\hbar\omega)^2}}
\Biggl\{1+{2\over\pi}{U\over\hbar\omega}
\arctan\Biggl[{\hbar\omega\over\sqrt{4I^2(r)-(\hbar\omega)^2}}\Biggr]\Biggr\},
{}~~~\mbox{for}~~\hbar\omega<2I(r),
\nonumber\\
&=&{U\over\hbar\omega}
{2 g^2({\bf qr})^2I^2(r)\over \sqrt{(\hbar\omega)^2-4I^2(r)}}
\ln\Biggl[{\sqrt{(\hbar\omega)^2-4I^2(r)}+\hbar\omega\over2I(r)}\Biggr],
{}~~~~~~\mbox{for}~~\hbar\omega>2I(r).
\end{eqnarray}
The expression (\ref{21}) for $\mbox{Re}{\cal P}(\omega,q,r)$ can be
also obtained, using the Kramers-Kronig relations, from
$\mbox{Im}{\cal P}(\omega,q,r)$ which has a simple form
\begin{eqnarray}\label{22}
\mbox{Im}{\cal P}(\omega,q,{\bf r})=
-{\pi ({\bf qr})^2I^2(r)\over (\hbar\omega)^2}F(\hbar\omega,r)
&=&-{\pi g^2({\bf qr})^2I^2(r)\over \sqrt{(\hbar\omega)^2-4I^2(r)}}
\Biggl(1+{U\over\hbar\omega}\Biggr),~~~\mbox{for}~~\hbar\omega>2I(r),
\nonumber\\
&=& 0,~~~\mbox{for}~~\hbar\omega<2I(r).
\end{eqnarray}
Note that as in (\ref{19}),
$\mbox{Im}{\cal P}(\omega,q,{\bf r})\neq 0$ only for
$\hbar\omega>2I(r)$.
We see that the enhancement in the density of pairs with small
$\Gamma$ leads to a {\em positive} sign of $\mbox{Re}{\cal P}(\omega,q,r)$
for $\hbar\omega > 2I(r)$. This is our main observation.
In order to obtain $\mbox{Re}P(\omega,q)$
one should integrate $\mbox{Re}{\cal P}(\omega,q,r)$ over ${\bf r}$.
For $\hbar\omega > 2I_0$ the main contribution to this integral comes from
$r\sim\xi$ [like in derivation of Eq.~(\ref{12})] and we get
\begin{eqnarray}\label{23}
\mbox{Re}P(\omega,q)={3\pi\over 4}
{q^2UI_0^2g^2\xi^4\over (\hbar\omega)^2}
\ln\Biggl({\hbar\omega\over I_0}\Biggr).
\end{eqnarray}
For generality we will assume that there is also a gate at a distance $d$
from the plane of 2D electrons. In this case the Fourier component of
electron-electron interaction has the form
\begin{eqnarray}\label{24}
v(q)={2\pi e^2\over \kappa q}\Biggl(1-e^{-2qd}\Biggr).
\end{eqnarray}
If the gate is close to the electron plane, that is $qd\ll 1$, we
can expand the exponent in (\ref{24}) and get $v(q)=4\pi e^2d/\kappa$.
If $d\ll \xi$ then the capacitance $C$ reduces to the capacitance of
two disks with area $S=\xi^2$ separated by a distance $d$, so that
$C=\kappa S/4\pi d$ and consequently
$U=4\pi e^2d/\kappa\xi^2$. With these $v(q)$ and $U$, after
substituting (\ref{23}) into the plasmon equation (\ref{13}), we obtain
\begin{eqnarray}\label{25}
1={3\pi\over 4}(q\xi)^2\Biggl({U\over\hbar\omega}\Biggr)^2
(I_0g\xi^2)^2\ln\Biggl({\hbar\omega\over I_0}\Biggr).
\end{eqnarray}
Certainly, the numerical coefficient in (\ref{25}) should not be taken
seriously. According to the assumtion $(ii)$, $I_0\sim g\xi^2$. Then
Eq.~(\ref{25}) yields the following dispersion law for the surface
plasmon
\begin{eqnarray}\label{26}
q(\omega)={1\over \xi}\Biggl({\hbar\omega\over U}\Biggr)
\ln^{-1/2}\Biggl({\hbar\omega\over I_0}\Biggr).
\end{eqnarray}
We see that the dispersion law is close to acoustic.
Let us establish the frequency range for the surface plasmon with
dispersion law (\ref{26}). The validity of expansion (\ref{16}),
$q\xi\ll 1$, implies that
$U\gg \hbar\omega\ln^{-1/2}(\hbar\omega/I_0)$. On the other hand,
$\hbar\omega>2I_0$. Then the frequency range for plasmon is
$2I_0\lesssim\hbar\omega\lesssim U$. The nesessary condition for this
range to be wide is $U\gg I_0$. The ratio $U/I_0=4\pi e^2d/\xi^2 I_0$
with $I_0\sim 1/g\xi^2$ can be presented as $8\pi^2 d/a_B$, where
$a_B=\hbar^2\kappa/me^2$ is the Bohr raduis. Thus, the condition $d\gg a_B$
insures that the plasmon equation (\ref{13}) has a solution within a
wide frequency range. If $d<a_B$, the screening of Coulomb interaction
by the gate is strong and the number of soft pairs is not sufficient to
change the sign of $\mbox{Re}P(\omega,q)$.
A similar condition can
be obtained from the analysis of $\mbox{Im}P(\omega,q)$, which
can be derived by integration of Eq.~(\ref{22}) over ${\bf r}$
\begin{eqnarray}\label{27}
\mbox{Im}P(\omega,q)={3\pi^2\over 8}
{q^2I_0^2g^2\xi^4\over (\hbar\omega)^2}(\hbar\omega+U)
\sim {q^2\over (\hbar\omega)^2}(\hbar\omega+U).
\end{eqnarray}
The origin of $\mbox{Im}P(\omega,q)$ is the interaction of a
plasmon with ``resonant'' pairs having excitation energy $\omega$. In
fact, $\mbox{Im}P(\omega,q)$ describes the resonant scattering of a
plasmon by a pair of localized states. One can introduce a mean free
path, $l$, associated with such a scattering and obtain
$ql\sim\ln^{1/2}(\hbar\omega/I_0)/(1+\hbar\omega/U)$.
Thus, the condition $2I_0\lesssim\hbar\omega\lesssim U$ reduces to the
condition $ql\gtrsim 1$.
In the absense of gate [$qd\gg 1$ in Eq.~(\ref{24})] one should take
$U=e^2/\kappa\xi$, so that $U/I_0\sim \xi/a_B\gg 1$. Then after a
simple algebra we obtain
\begin{eqnarray}\label{28}
q(\omega)={1\over \xi}\Biggl({\hbar\omega\over U}\Biggr)^2
\ln^{-1}\Biggl({\hbar\omega\over I_0}\Biggr).
\end{eqnarray}
It can be shown that the above analysis of the validity applies in
this case as well and leads to the same frequency range
$2I_0\lesssim\hbar\omega\lesssim U$. Within this range we again have
$q\xi\ll 1$.
Thus, one should use Eq.~(\ref{26}) for $qd<1$ and Eq.~(\ref{28}) for
$qd>1$. On the other hand, the magnitude of $U$ depends on the ratio
$d/\xi$. Note that for $d>\xi$ one still can have $qd<1$. In this case
the dispersion law is given by Eq.~(\ref{26}) with $U=e^2/\kappa\xi$.
\section{Renormalization of the plasmon dispersion law}
\label{sec4}
In the previous section, when calculating the polarization operator,
we took into account the Coulomb correlations within a pair, but
neglected the effect of polarization of surrounding pairs on a given
pair. On the other hand, by averaging of the polarization operator
(\ref{14}) over frequencies of pairs $\Gamma$ and their shoulders
${\bf r}$ we have effectively replaced the localized system by a
medium. The average polarization of this medium gave rise to a plasmon
mode. Within this procedure the ``feedback'' from surrounding pairs
reduces to the interaction of a given pair with plasmons. In the present
section we study the renormalization of the plasmon spectrum due to
this effect.
Generally, the plasmon excitation is defined as a pole in the
density-density correlation function, $\Pi(\omega, {\bf q},{\bf q}')$,
which is related to the polarization operator
${\bf P}(\omega, {\bf q},{\bf q}')$ by the Dyson equation
\begin{eqnarray}\label{29}
\Pi(\omega, {\bf q},{\bf q}')=
{\bf P}(\omega, {\bf q},{\bf q}')+\int {d{\bf q}_1\over (2\pi)^2}
{\bf P}(\omega, {\bf q},{\bf q}_1)v(q_1)\Pi(\omega, {\bf q}_1,{\bf q}').
\end{eqnarray}
Before averaging, both ${\bf P}(\omega, {\bf q},{\bf q}')$ and
$\Pi(\omega, {\bf q},{\bf q}')$ depend on two momentum variables
${\bf q}$ and ${\bf q}'$. The approximation we made above reduces to
replacing of ${\bf P}(\omega, {\bf q},{\bf q}')$ by its average
${\bf P}(\omega, q)$, so that the solution of (\ref{29}) takes
the form
\begin{eqnarray}\label{30}
\Pi(\omega, q)=
{{\bf P}(\omega, q)
\over 1-v(q){\bf P}(\omega, q)}.
\end{eqnarray}
Then the pole of $\Pi(\omega,q)$ is determined by the plasmon equation
(\ref{13}).
As a next step, we took for ${\bf P}(\omega, q)$ its expression
(\ref{14}) for non-interacting electrons, which represents a sum of
polarizations of pairs,
\begin{eqnarray}\label{31}
P(\omega,q)={1\over A}
\sum_{\bf ij}|\langle {\bf i}|e^{i{\bf qr}}|{\bf j} \rangle|^2
P_{\bf ij}(\omega),~~~~~
P_{\bf ij}(\omega)=
{n_{\bf i}-n_{\bf j} \over \hbar\omega + E_{\bf i}-E_{\bf j} + i0},
\end{eqnarray}
and performed the summation neglecting correlations between the pairs
but with the Coulomb correlations within a pair included.
The renormalized ${\bf P}_{\bf ij}(\omega)$ for a given pair
can be obtained from the following procedure.
The function $\Pi(\omega, q)$ has a diagrammatic presentation in
a form of a series of bubbles $({\bf ij})$, corresponding to
$P_{\bf ij}$, connected by the
Coulomb interaction lines (see Fig.~1a). First, we arrange into a single
block the sum over all combinations of bubbles which appear between two
bubbles $({\bf ij})$ (see Fig.~1b).
Then we replace this block by its average, so that the result
can be presented as two bubbles $({\bf ij})$ connected by
a plasmon propagator $\Pi(\omega, q)$ (see Fig.~1b) (there is
also an extra factor $v(q)$ in each vertex). The renormalized bubble
$({\bf ij})$ can be then obtained by summing up the series, consisting
from the bubbles $({\bf ij})$, connected by plasmon lines (see Fig.~1c).
The resulting expression for ${\bf P}_{\bf ij}(\omega)$ reads
\begin{eqnarray}\label{32}
{\bf P}_{\bf ij}(\omega)=
{P_{\bf ij}(\omega,)
\over 1-P_{\bf ij}(\omega)R_{\bf ij}(\omega)},
\end{eqnarray}
with
\begin{eqnarray}\label{33}
R_{\bf ij}(\omega)=\int {d{\bf q}\over (2\pi)^2}
|\langle {\bf i}|e^{i{\bf qr}}|{\bf j} \rangle|^2
v^2(q)\Pi(\omega, q).
\end{eqnarray}
Finally, replacing $P_{\bf ij}(\omega)$ in (\ref{31}) by
${\bf P}_{\bf ij}(\omega)$ we obtain
\begin{eqnarray}\label{34}
{\bf P}(\omega,q)={1\over A}
\sum_{\bf ij}|\langle {\bf i}|e^{i{\bf qr}}|{\bf j} \rangle|^2
{n_{\bf i}-n_{\bf j} \over \hbar\omega + E_{\bf i}-E_{\bf j} -
(n_{\bf i}-n_{\bf j})R_{\bf ij}(\omega)}.
\end{eqnarray}
The equations (\ref{30}), (\ref{33}), and (\ref{34}) form a closed
system which determines $\Pi (\omega, {\bf q})$ and, correspondingly, the
renormalized dispersion law of a plasmon in a self-consistent way. The
approximation made in order to get the closed system [replacement of a
block by a sought function $\Pi (\omega, q)$] is known as the
effective-medium approximation.
The analysis of the system (\ref{30},\ref{33},\ref{34}) reveals that
the renormalization of the plasmon dispersion law is weak. Namely, the
appearence of the term $R_{\bf ij}(\omega)$ in the denominator of
(\ref{34}) has the physical meaning that a pair $({\bf ij})$ acquires
a finite life-time $\tau_{\bf ij}$ due to the interaction with
plasmons. We will show that this life-time is long, i.e.,
$1/\tau_{\bf ij}\ll \omega_{\bf ij}$. It can be readily seen that the
difference between the renormalized polarization
$\mbox{Re}{\bf P}(\omega,q)$ and $\mbox{Re}P(\omega,q)$ originates
from resonant pairs with $(\omega -\omega_{\bf ij})\sim 1/\tau_{\bf ij}$
and $r\sim \xi$ (note that the main term, $\mbox{Re}P(\omega,q)$, is
determined by the entire interval $\omega_{\bf ij}\sim \omega$). If we
neglect the dependence of the matrix element in (\ref{34}) on
$\omega_{\bf ij}$ then the renormalization correction to
$\mbox{Re}P(\omega,q)$ would be identically zero. A finite correction
results from a slight asymmetry of the matrix element within the
narrow interval $(\omega -\omega_{\bf ij})\sim 1/\tau_{\bf ij}$. Then
the relative magnitude of the correction is of the order of
$1/\omega\tau_{\bf ij}$ with $\tau_{\bf ij}$ calculated for a pair
with $\omega_{\bf ij}=\omega$. Expecting $\tau_{\bf ij}$ to be long,
we can calculate it by substituting the nonrenormalized dispersion law
of a plasmon into Eq.~(\ref{33}). Performing the integration we obtain
\begin{eqnarray}\label{35}
R_{\bf ij}(\omega)={\hbar\over\tau_{\bf ij}}&\sim&\hbar\omega
\Biggl({\hbar\omega\over U}\Biggr)\Biggl({I_0\over U}\Biggr)^2
\ln^{-2}\Biggl({\hbar\omega\over I_0}\Biggr),~~~~~~~~~~\mbox{with gate,}
\nonumber\\
&\sim&\hbar\omega
\Biggl({\hbar\omega\over U}\Biggr)^3\Biggl({I_0\over U}\Biggr)^2
\ln^{-3}\Biggl({\hbar\omega\over I_0}\Biggr),~~~~~~~~\mbox{without gate.}
\end{eqnarray}
Since both ratios, $\hbar\omega/U$ and $I_0/U$ are small, the
correction to the dispersion law,
$\delta q(\omega)/q(\omega)\sim 1/\omega\tau_{\bf ij}$, is negligible.
In the next section we will see that the corresponding renormalization
of $\mbox{Re}\sigma (\omega)$ is much larger than the renormalization
of the dispersion law.
\section{Renormalization of the real part of the conductivity}
\label{sec5}
As we have seen in the previous section, the renormalization of the
polarization operator results in an appearance of
$i\hbar/\tau_{\bf ij}$ in the denominator of Eq.~(\ref{34}).
Correspondingly, the renormalized expression (\ref{4}) for
$\mbox{Re}\sigma(\omega)$ can now be rewritten as
\begin{eqnarray}\label{36}
\mbox{Re}\sigma(\omega)={e^2\omega\over \hbar A}\mbox{Im}
\sum_{\bf ij}\Biggl[{I_{ij}(x_i-x_j)\over \hbar\omega_{\bf ij}}\Biggr]^2
{1 \over \omega + \omega_{\bf ij} + i/\tau_{\bf ij}},
\end{eqnarray}
where the summation is performed over the singly occupied pairs
${\bf ij}$ and in this way the Coulomb correlations are taken into
account. In the limit $\tau_{\bf ij}\rightarrow\infty$ and for
$\omega >2I_0$ one should use the pair density $F(\Gamma,r)$, given by
Eq.~(\ref{20}), in order to perform the summation. The result is
determined by resonant pairs with
$\omega=\omega_{\bf ij}=\Gamma/\hbar$:
\begin{eqnarray}\label{37}
\mbox{Re}\sigma(\omega)={3\pi^2\over 4}
{e^2\over \hbar}(I_0g\xi^2)^2\Biggl(1+{U\over\hbar\omega}\Biggr)=
\sigma_0\Biggl(1+{U\over\hbar\omega}\Biggr).
\end{eqnarray}
We see that within the frequency interval $2I_0<\hbar\omega<U$ the
real part of the conductivity exceeds the Drude value due to Coulomb
correlations.
When calculating the correction to $\mbox{Re}\sigma(\omega)$ caused by
the finite value of $\tau_{\bf ij}$ it is important to realize that
$\hbar/\tau_{\bf ij}$ is maximal for soft pairs with small
$\omega_{\bf ij}$. This is because the matrix element
$\langle {\bf i}|e^{i{\bf qr}}|{\bf j} \rangle$ is proportional to
$1/\omega_{\bf ij}$. In the previous Section this was not important
since the correction to $\mbox{Re}P(\omega,q)$ came from the resonant
pairs only. Here, however, we have
$\mbox{Im}(\omega + \omega_{\bf ij} + i/\tau_{\bf ij})^{-1}
\propto 1/\tau_{\bf ij}\propto 1/\omega_{\bf ij}^2$, so that the soft
pairs give the main contribution to the correction
$\delta\mbox{Re}\sigma(\omega)$. Assuming $\omega_{\bf ij}\ll\omega$
we can present this correction in the form
\begin{eqnarray}\label{38}
\mbox{Re}\sigma(\omega)&=&{e^2\over \hbar\omega A}
\sum_{\bf ij}{[I_{ij}(x_i-x_j)]^2
\over (\hbar\omega_{\bf ij})^2\tau_{\bf ij}}
\nonumber\\
&=&{e^2\over \hbar\omega A}
\sum_{\bf ij}\Biggl[{I_{ij}(x_i-x_j)
\over \hbar\omega_{\bf ij}}\Biggr]^4
\int {d{\bf q}\over (2\pi)^2}
q^2v^2(q)\mbox{Im}\Pi(\omega, q),
\end{eqnarray}
where we have substituted $\hbar/\tau_{\bf ij}=R_{\bf ij}$ from
Eq.~(\ref{33}). The sum over pairs is again evaluated with the pair
density (\ref{20})
\begin{eqnarray}\label{39}
{1\over A}\sum_{\bf ij}\Biggl[{I_{ij}(x_i-x_j)
\over \hbar\omega_{\bf ij}}\Biggr]^4=
{1\over 4}\int d{\bf r}r^4I^4(r)
\int_{2I(r)}^{\infty}d\Gamma{F(\Gamma,r)\over\Gamma^4}.
\end{eqnarray}
The main contribution to the integral over $\Gamma$ comes from the
lower limit $\Gamma\sim I(r)$. Then the integral over $r$ is again
determined by $r\sim\xi$, so that the sum (\ref{39}) appears to be
$\sim I_0\xi^6g^2U$. As in the previous section, the value of the
integral in (\ref{39}) takes different values in the presence and in
the absence of a gate. Finally we obtain
\begin{eqnarray}\label{40}
{\delta\mbox{Re}\sigma(\omega)\over\sigma_0}
&\sim&
\Biggl({\hbar\omega\over U}\Biggr)^3\Biggl({\hbar\omega\over I_0}\Biggr)
\ln^{-2}\Biggl({\hbar\omega\over I_0}\Biggr),~~~~~~~~~~\mbox{with gate,}
\nonumber\\
&\sim&
\Biggl({\hbar\omega\over U}\Biggr)^5\Biggl({\hbar\omega\over I_0}\Biggr)
\ln^{-3}\Biggl({\hbar\omega\over I_0}\Biggr),~~~~~~~~\mbox{without gate.}
\end{eqnarray}
Comparing (\ref{35}) to (\ref{40}) we see that both corrections are of the
same order $\left(I_0/U\right)^2$ at $\hbar\omega\sim I_0$. However the
correction to ${\mbox Re}\sigma$ is much bigger at $\hbar\omega\gg I_0$.
This reveals a new mechanism of absorption of a.c. field:
by resonant excitation of plasmons.
More precisely, the field polarizes the soft pairs (with
$\omega_{\bf ij}\sim I_0/\hbar$) and the induced polarization
excites the plasmon waves. Therefore, the energy of a.c. field is effectively
absorbed by plasmons. The rapid increase of $\delta\sigma$ with $\omega$
is caused by the number (phase volume) of plasmons which absorb the field.
Note that with correction $\delta\sigma(\omega)$ the total
conductivity ${\mbox Re}\sigma(\omega)$
exhibits a rather complicated behavior. For $\hbar\omega<2I_0$ the
conductivity increases with $\omega$. Then it passes through a maximum
at $\omega\sim I_0/\hbar$ and falls off with $\omega$ according to
Eq.~(\ref{37}). However at $\hbar \omega\sim U(I_0/U)^{1/4}$ (with
gate) and $\hbar \omega\sim U(I_0/U)^{1/6}$ (without gate) we have
$\delta{\mbox Re}\sigma(\omega)\sim {\mbox Re}\sigma(\omega)$
and the conductivity starts rising again. On the other hand, the
expression for $\delta{\mbox Re}\sigma$ was derived assuming that it
is small. Therefore in the region
$\delta{\mbox Re}\sigma >{\mbox Re}\sigma(\omega)$ the renormalization
of $\sigma(\omega)$ by plasmons is strong. In this case one cannot
calculate $\tau_{\bf ij}$ using the bare polarization operator. The
full analysis of the system (\ref{30},\ref{33},\ref{34}) in this
frequency range is out of the scope of the present paper.
\section{Conclusion}
\label{sec6}
In the present paper we argue that the wave of electric field can
propagate along the surface of the 2D Anderson insulator. The field
originates from the density fluctuations of localized electrons. One
should distinguish this wave from the usual plasmon in an ideal 2D gas
with the dispersion law given by Eq. (\ref{15}): ({\em i}) the
derivation of (\ref{15}) implies that $\omega\gg 1/\tau$ while we
predict the existence of a plasmon at much lower frequencies
$\omega\gtrsim I_0/\hbar$. The minimal frequencies $1/\tau$ and
$I_0/\hbar$ differ by a factor $\exp(\pi k_Fl)$; ({\em ii}) the
plasmon (\ref{15}) results from the solution of Eq. (\ref{13}) with
polarization operator calculated for free electrons. Within this
approximation Eq. (\ref{13}) has no solutions for {\em localized}
electrons. The solution appears only if one takes into account the
Coulomb correlations in the occupation numbers of the localized states.
The obvious consequence of the existence of a plasmon excitation is
that localized electrons ``feel'' a fluctuating electric field
${\cal E}(\omega,{\bf q})$ with the spectral density
\begin{eqnarray}\label{41}
\langle |{\cal E}(\omega,{\bf q})|^2\rangle=
{\kappa^2\over 2\pi e^2}{q^2v^2(q)\mbox{Im}{\bf P}(\omega,q)\over
[1-v(q)\mbox{Re}{\bf P}(\omega,q)]^2+[v(q)\mbox{Im}{\bf P}(\omega,q)]^2},
\end{eqnarray}
where ${\bf P}(\omega,q)$ is the polarization operator (\ref{34}). The
spectral density (\ref{41}) has a peak at $\omega=\omega(q)$
corresponding to the dispersion law of a plasmon, which is different
with and without gate.
The basic assumption of our theory is that in the presence of
interactions the Anderson insulator is characterized by two energy
scales: $I_0$ - the spacing between energy levels in the area $\xi^2$
and the charging energy $U$ which is the Coulomb interaction of two
localized electrons separated by a distance $\sim\xi$. This energy
modifies the pair density $F(\omega,r)$. Such a modification results
in an enhancement\cite{shk81,efr81,efr85,efr85b} of the dissipative
conductivity
$\mbox{Re}\sigma(\omega)$. If we denote as $\rho(\omega)$ the matrix
element of $r$ calculated for a pair with frequency $\omega$ then
\begin{eqnarray}\label{42}
\mbox{Re}\sigma(\omega)\propto \omega\rho^2(\omega)F(\omega, r_{\omega}).
\end{eqnarray}
As it was discussed in the Introduction, it is plausible to assume
that for the Anderson insulator in the absence of interactions
$\rho(\omega)$ behaves as
$\rho(\omega)\sim r_{\omega}$ for $\omega\lesssim I_0/\hbar$ and
$\rho(\omega)\sim\xi I_0/\omega$ for $\omega\gtrsim I_0/\hbar$. Since
the pair density in the absence of interactions is proportional to
$\omega$ we reproduce Eqs.~(\ref{11}) and (\ref{12}):
\begin{mathletters}\label{43}
\begin{eqnarray}
\mbox{Re}\sigma(\omega)&\propto&\omega^2,
{}~~~~~~~~~~~~\mbox{for}~~~\omega\lesssim I_0/\hbar,\label{43a}\\
\mbox{Re}\sigma(\omega)&=& const,
{}~~~~~~~~~\mbox{for}~~~\omega\gtrsim I_0/\hbar,\label{43b}
\end{eqnarray}
\end{mathletters}
Where we neglected a logarithmic factor in (\ref{43a}).
We have assumed that the frequency dependence of $\rho(\omega)$
remains the same in the presence of interactions.\cite{note} At the
same time, $F(\omega,\xi)$ is changed drastically by interactions:
$F(\omega,\xi)=const$ for $\omega\lesssim U/\hbar$ and
$F(\omega,\xi)\propto\omega$ for $\omega\gtrsim U/\hbar$. As a result,
we get the following frequency dependence of $\mbox{Re}\sigma(\omega)$
in the presence of interactions [see also (\ref{37})]:
\begin{mathletters}\label{44}
\begin{eqnarray}
\mbox{Re}\sigma(\omega)&\propto& \omega,
{}~~~~~~~~~~~~~~\mbox{for}~~~\omega\lesssim I_0/\hbar,\label{44a}\\
\mbox{Re}\sigma(\omega)&\propto& 1/\omega,
{}~~~~~~~~~~~\mbox{for}~~~I_0/\hbar\lesssim\omega\lesssim U/\hbar,\label{44b}\\
\mbox{Re}\sigma(\omega)&=& \mbox{const},
{}~~~~~~~~~\mbox{for}~~~\omega\gtrsim U/\hbar,\label{44c}
\end{eqnarray}
\end{mathletters}
This simple analysis shows that $\mbox{Re}\sigma(\omega)$ exhibits
a maximum at $\omega\sim I_0/\hbar$ (see also the end of the previous
section). In fact, this maximum is
intimately related to the existence of a plasmon. Indeed, for
Eq. (\ref{13}) to have solutions we need a positive sign of
$\mbox{Re}{\bf P}(\omega,q)$, which is equivalent to a positive
$\mbox{Im}\sigma(\omega)$. At the same time, $\mbox{Im}\sigma(\omega)$
can be obtained from $\mbox{Re}\sigma(\omega)$ using the
Kramers-Kronig relation. In order to trace how the modification of
$\mbox{Re}\sigma(\omega)$ by Coulomb correlations leads to the change
of sign of $\mbox{Im}\sigma(\omega)$, we can interpolate the
frequency dependence of $\rho(\omega)$ as
$\rho(\omega)\propto [(\hbar\omega)^2+4I_0^2]^{-1/2}$ and the
frequency dependence of $F(\omega,\xi)$ as
$F(\omega,\xi)\propto (\hbar\omega + U)$. Then we have
\begin{eqnarray}\label{45}
\mbox{Re}G(\omega)={\hbar\omega(\hbar\omega+U)
\over (\hbar\omega)^2+4I_0^2},~~~~~~~~~~\omega>0,
\end{eqnarray}
where $G={\sigma\over e^2/\hbar}$ is the conductance. It can be easily
seen that Eq. (\ref{45}) reproduces correctly all the limiting
cases (\ref{44}) and (\ref{44}) both with and without charging effect.
The imaginary
part of the conductance calculated from (\ref{45}) has a simple form:
\begin{eqnarray}\label{46}
\mbox{Im}G(\omega)=
-{\hbar\omega[2I_0-(2U/\pi)\ln(\hbar\omega/ 2I_0)]
\over (\hbar\omega)^2+4I_0^2},~~~~~~~~~~\omega>0,
\end{eqnarray}
In Fig. 2 we have plotted
$\mbox{Re}G(\omega)$ and $\mbox{Im}G(\omega)$ for different ratios
$U/2I_0$. We see that in the absence of Coulomb correlations ($U=0$),
$\mbox{Im}G(\omega)$ is strictly negative (in our calculation in
Section \ref{sec2} it turns to zero for $\omega>I_0/\hbar$). However,
at finite $U$ the change of sign occurs at
$\omega=(2I_0/\hbar)\exp(\pi I_0/U)$. For $U\gg I_0$ this frequency is
just $2I_0/\hbar$.
As a final remark, let us outline the difference between our approach
and that of Ref.~\onlinecite{fle78}. In Ref.~\onlinecite{fle78} the
authors addressed electron-electron interactions in the strongly
localized regime when the localization radius is much smaller than the
interpair separation. In this case singly-occupied pairs can be
considered as point-like dipoles. The dipole moment ${\bf p}_k$
induced by an external field ${\cal E}_0e^{-i\omega t}$ can be written
as
\begin{eqnarray}\label{47}
p_k^{\mu}=\alpha_k^{\mu\nu}[{\cal E}_0^{\nu}+
\sum_{l\neq k}{\cal E}_k^{(l)\nu}],
\end{eqnarray}
where ${\cal E}_k^{(l)\nu}$ is a component $\nu$ of the electric field
caused by polarization of a dipole $l$ which acts on the dipole $k$
(summation over repeating indices $\nu$ is implied). The
polarizability $\alpha_k^{\mu\nu}$ of a dipole $k$ has the form
\begin{eqnarray}\label{48}
\alpha_k^{\mu\nu}={2e^2\over\hbar}{\rho_k^{\mu}\rho_k^{\nu}\omega_k
\over \omega^2-\omega_k^2}
={e^2\over\hbar}\rho_k^{\mu}\rho_k^{\nu}\sum_{\bf ij}P_{\bf ij},
\end{eqnarray}
where $\rho_k^{\mu}$ and $\omega_k$ are correspondingly the matrix
element and the frequency of the dipole $k$. Here $P_{\bf ij}$ is
given by Eq.~(\ref{31}) with the states ${\bf i}$ and ${\bf j}$ making
up the dipole $k$. To keep the discussion
simple we will assume that the polarizability is isotropic, i.e.,
$\alpha_k^{\mu\nu}=\alpha_k\delta_{\mu\nu}$. The field
${\cal E}_k^{(l)\nu}$ can, in turn, be expressed through ${\bf p}_k$
as
\begin{eqnarray}\label{49}
{\cal E}_k^{(l)\mu}=-{\partial\over\partial R^{\mu}}
{{\bf p}_l{\bf R}\over R^3}\Biggl|_{{\bf R}={\bf R}_k-{\bf R}_l},
\end{eqnarray}
where ${\bf R}_k$ is the position of the dipole $k$.
Upon substituting (\ref{49}) into (\ref{47}) we obtain an infinite
system of linear equations. It can be easily shown that iterating this
system leads to the renormalization of polarizabilities of dipoles
$\alpha_k$
\begin{mathletters}\label{50}
\begin{eqnarray}
\alpha_k={e^2\over\hbar}\rho_k^2
\sum_{\bf ij}{P_{\bf ij}\over 1-P_{\bf ij}\Sigma_k}
={2e^2\over\hbar}{\rho_k^2(\omega_k+\Sigma_k)
\over \omega^2-(\omega_k+\Sigma_k)^2},\\
\Sigma_k=-\sum_l \alpha_l
\Biggl({\partial\over\partial R^{\mu}}{R^{\nu}\over R^3}\Biggr)
\Biggl({\partial\over\partial R^{\nu}}{R^{\mu}\over R^3}\Biggr)
\Biggl|_{{\bf R}={\bf R}_k-{\bf R}_l}+\cdots.
\end{eqnarray}
\end{mathletters}
Here $\Sigma_k$ is the self-energy. Fleishman and Anderson\cite{fle78}
anylized this self-energy using the arguments similar to those put
forward by Anderson\cite{and58} when he demonstrated the existence of
localization transition for eigenfunctions of the Schroedinger
equation with disorder. They argued that $\mbox{Im}\Sigma$ takes a
finite value with non-zero probability. This means that in the absence
of an external field the system
\begin{eqnarray}\label{51}
{\bf p}_k+\alpha_k\nabla\sum_{l\neq k}
{{\bf p}_l{\bf R}\over R^3}\Biggl|_{{\bf R}={\bf R}_k-{\bf R}_l}=0,
\end{eqnarray}
has delocalized solutions.\cite{lev90} In other words, by analogy to the
Schoedinger equation in the tight-binding approximation, the
eigenstates $\{{\bf p}_k\}$ of the system ({\ref{51}) extend throughout
the entire volume.
Fleishman and Anderson\cite{fle78} considered a three-dimensional
system and neglected the Coulomb correlations. Note that if one
rewrites the system (\ref{51}) in the momentum representation and
averages it by factorizing the average of the product $\alpha{\bf p}$
(mean-field) then one arrives at the plasmon equation (\ref{13}). As it was
demonstrated in Section \ref{sec2}, this equation has no propagating
solutions in the absence of Coulomb correlations. Our central point is
that {\em with} the Coulomb correlations the delocalized solution
exists even at the mean-field level. This solution, that is surface
plasmon, is specific for the two-dimensional system and is
characterized by the dispersion law $\omega(q)$.
An interesting question that could be addressed
within the same approach is how the Coulomb correlations modify the
a.c. Hall conductivity $\sigma_{xy}(\omega)$ of the Anderson
insulator.\cite{imr93,zha92}
\acknowledgments
The authors are indebted to B.~I. Shklovskii for numerous discussions.
In fact, Sections \ref{sec4}-\ref{sec6} have emerged from his hot contesting
of our basic assumptions.
|
\section*{Appendix}
\setcounter{equation}{0}
\defA\arabic{equation}{A\arabic{equation}}
It is convenient to define the mass dependent variables $a:=2+\sqrt\xi$,
$b:=2-\sqrt\xi$ and $w:=\sqrt{(1-\sqrt\xi)/(1+\sqrt\xi)}$. The integrals
$t_1,\ldots,t_{12}$ appearing in Eqs.~(\ref{eqn13})--(\ref{eqn14}) are then
given by
\begin{eqnarray}
t_1&:=&\ln\left(\frac{2\xi\sqrt\xi}{b^2(1+\sqrt\xi)}\right),\quad
t_2\ :=\ \ln\left(\frac{2\sqrt\xi}{1+\sqrt\xi}\right)\quad
\Rightarrow\quad t_1-t_2\ =\ \ln\left(\frac\xi{b^2}\right)\\
t_3&:=&\ln\left(\frac{1+v}{1-v}\right)\\
t_4&:=&\mbox{Li}_2(w)-\mbox{Li}_2(-w)+\mbox{Li}_2(\frac abw)-\mbox{Li}_2(-\frac abw)\\
t_5&:=&\frac12\ln\left(\frac{\sqrt\xi(2+\sqrt\xi)}{4(1+\sqrt\xi)}\right)\ln\left(\frac{1+v}{1-v}\right)
+\mbox{Li}_2\left(\frac{2\sqrt\xi}{a(1+w)}\right)
-\mbox{Li}_2\left(\frac{2\sqrt\xi}{a(1-w)}\right)\,+\nonumber\\&&
+\mbox{Li}_2\left(\frac{1+w}2\right)-\mbox{Li}_2\left(\frac{1-w}2\right)
+\mbox{Li}_2\left(\frac{a(1+w)}4\right)-\mbox{Li}_2\left(\frac{a(1-w)}4\right)\\
t_6&:=&\ln^2(1+w)+\ln^2(1-w)+\ln\left(\frac{2+\sqrt\xi}8\right)\ln(1-w^2)
\,+\nonumber\\&&
+\mbox{Li}_2\left(\frac{2\sqrt\xi}{a(1+w)}\right)
+\mbox{Li}_2\left(\frac{2\sqrt\xi}{a(1-w)}\right)
-2\mbox{Li}_2\left(\frac{2\sqrt\xi}a\right)\,+\nonumber\\&&
+\mbox{Li}_2\left(\frac{1+w}2\right)+\mbox{Li}_2\left(\frac{1-w}2\right)
-2\mbox{Li}_2\left(\frac12\right)\,+\nonumber\\&&
+\mbox{Li}_2\left(\frac{a(1+w)}4\right)+\mbox{Li}_2\left(\frac{a(1-w)}4\right)
-2\mbox{Li}_2\left(\frac a4\right)\\
t_7&:=&2\ln\left(\frac{1-\xi}{2\xi}\right)\ln\left(\frac{1+v}{1-v}\right)
-\mbox{Li}_2\left(\frac{2v}{(1+v)^2}\right)
+\mbox{Li}_2\left(-\frac{2v}{(1-v)^2}\right)\,+\nonumber\\&&
-\frac12\mbox{Li}_2\left(-\left(\frac{1+v}{1-v}\right)^2\right)
+\frac12\mbox{Li}_2\left(-\left(\frac{1-v}{1+v}\right)^2\right)\,+\\&&
+\mbox{Li}_2\left(\frac{2w}{1+w}\right)-\mbox{Li}_2\left(-\frac{2w}{1-w}\right)
-2\mbox{Li}_2\left(\frac{w}{1+w}\right)+2\mbox{Li}_2\left(-\frac{w}{1-w}\right)
\,+\nonumber\\&&
+\mbox{Li}_2\left(\frac{2aw}{b+aw}\right)-\mbox{Li}_2\left(-\frac{2aw}{b-aw}\right)
-2\mbox{Li}_2\left(\frac{aw}{b+aw}\right)+2\mbox{Li}_2\left(-\frac{aw}{b-aw}\right)
\nonumber\\
t_8&:=&\ln\left(\frac\xi 4\right)\ln\left(\frac{1+v}{1-v}\right)
+\mbox{Li}_2\left(\frac{2v}{1+v}\right)-\mbox{Li}_2\left(-\frac{2v}{1-v}\right)-\pi^2\\
t_9&:=&2\ln\left(\frac{2(1-\xi)}{\sqrt\xi}\right)\ln\left(\frac{1+v}{1-v}\right)
+2\left(\mbox{Li}_2\left(\frac{1+v}2\right)-\mbox{Li}_2\left(\frac{1-v}2\right)\right)
\,+\nonumber\\&&
+3\left(\mbox{Li}_2\left(-\frac{2v}{1-v}\right)
-\mbox{Li}_2\left(\frac{2v}{1+v}\right)\right)\\
t_{10}&:=&\ln\left(\frac4\xi\right),\quad
t_{11}\ :=\ \ln\left(\frac{4(1-\sqrt\xi)^2}\xi\right),\quad
t_{12}\ :=\ \ln\left(\frac{4(1-\xi)}\xi\right)
\end{eqnarray}
\newpage
|
\section{Introduction}
Laser interferometric detectors of gravitational waves such as the LIGO
\cite{LIGO} and VIRGO \cite{VIRGO} are expected to be operational by the
turn of the century.
Gravitational waves from coalescing binary systems of black holes and neutron
stars are relatively `clean' waveforms in the sense that they are easier to
model and for this reason they are amongst the most important
candidate sources for interferometric detectors.
Binary systems are also valuable sources of astrophysical information as
one can probe the universe up to cosmological distances. For instance,
statistical analysis of several binary coalescences enables the estimation
of the Hubble constant to an accuracy better than 10\% \cite {SCH86,Mark}.
Events that produce high signal-to-noise ratio can be potentially used
to observe such non-linear effects, such as gravitational wave
tails, and to put general relativity into test in the strongly non-linear
regime \cite {BS95}.
Due to the weak coupling of gravitational radiation with matter the
signal waveform is in general
very weak and will not stand above the detector noise.
In addition to the on-going efforts to reduce the noise, and hence increase the
sensitivity of the detector, a considerable amount of research activity
has gone into the development of efficient and robust data analysis techniques
to extract signals buried in very noisy data.
For a recent review on gravitational waves from compact objects and their
detection see Thorne \cite{Th95a,Th95b}.
Various data analysis schemes have been suggested for the detection of the
`chirp' waveform from such systems \cite{MS95,Sm87,Th300}. Among them the
technique of Weiner filtering is the most promising \cite{Th300,HEL,SCH89}.
Briefly, this technique involves correlating the detector output
with a set of templates, each of which is tuned to
detect the signal with a particular set of parameters. This requires the signal
to be known to a high level of accuracy. The fully general relativistic
waveform from a coalescing binary system of stars is as yet unavailable.
In the absence of such an exact solution, there have been
efforts to find solutions perturbatively. Most of the work done in
this area aims at computing the waveform correct to a high degree of
accuracy so that theoretical templates computed based on them will
obtain a large signal-to-noise ratio (SNR) when correlated with the
detector output if the corresponding signal is present.
In general, the number of parameters increases as we incorporate the
higher order corrections. It is clear that the number of templates depends
upon the number of signal parameters.
As a consequence, the computing power for an on-line analysis will
be greater for a larger number of parameters. In view of this restriction
in computing power it is
necessary to choose the templates in an optimal manner.
This paper in part deals with this question. Investigations
till now have been restricted to either choosing a finite subset of the signal
space as templates \cite{SD91,DS94} or choosing templates from the
`Newtonian' or the `first post-Newtonian' family of waveforms
\cite{BD94,KKT94,A95,KKS95}. We generalize this problem by using the language
of differential geometry. We show that it is unnecessary to restrict
oneself to templates that are matched exactly to any particular signal.
Differential geometry has been used in statistics before (see \cite{Am}
and references therein) and the
standard approach is to treat a set of parameterised probability distributions
corresponding to a particular statistical model as a manifold. The
parameters of the distribution serve as coordinates on this manifold.
In statistical theory one frequently comes across the Fisher information matrix
whose inverse gives a lower bound for the errors in the estimation of the
parameters of a distribution.
The Fisher information matrix turns out to be a
very natural metric on the manifold of probability distributions
and this metric can be used profitably
in understanding the properties of a particular statistical model. Here in our
paper we treat the set of coalescing binary signals corresponding to
various parameters of the binary as a submanifold in the linear space of all
detector outputs. We show in this paper that both the above mentioned
manifolds are equivalent as far as their metrical properties are concerned.
The geometric approach turns out to be useful not only in
clarifying various aspects of signal analysis but also helps us to pose the
question of optimal detection in a more general setting.
Once a signal has been detected we can estimate the parameters of the binary.
We assume that the parameters of the signal are the same as those of
the template with which the maximum correlation is obtained. The errors
involved in such an estimation has been worked out by several
authors \cite{BS95,KKS95,F92,FC93,CF94,BS94,Kr,KLM93,PW95,JKKT},
for the case of `high' SNR and for the Newtonian and post-Newtonian
waveforms using a single and a network of detectors.
For the case of low SNRs one has to resort to numerical simulations.
We have started a project to carry out exhaustive
numerical simulations specifically designed to compute the errors in the
estimation of parameters and covariances among them at various post-Newtonian
orders, for circular and eccentric orbits, with and without spin effects and
for different optical configurations of the interferometer.
In this paper we report the results for the case of the initial
LIGO configuration taking
into account only the first post-Newtonian corrections and
assuming circular orbits. Going beyond this needs tremendous amount of
computing power which is just becoming available.
The rest of the paper is organized as follows.
In section \ref{waveforms} we describe the waveform from a coalescing binary
system at various post-Newtonian orders. We introduce, following \cite{Sat94},
a set of parameters called `chirp times'. These parameters are found to
be very convenient when we carry out Monte Carlo simulations.
It turns out that the covariance matrix is independent of these parameters
and hence it is sufficient to carry out the simulations only for a particular
set of parameters. In section \ref{sm} we develop a geometric interpretation
of signal analysis. We begin by introducing a metric on the manifold
from a scalar product, which comes naturally from the theory of matched
filtering and
then show that this metric is the same as the one used by Amari \cite{Am}.
Using the geometric approach we address the question of optimal filter
placement and show that for the purpose of detection it is optimal to
choose the templates outside the signal manifold.
The covariance matrix of errors and covariances is shown to be the inverse
of the metric on the manifold. In section \ref{sec_espar} we discuss
the results of our simulations and
compare the numerically obtained values and those suggested by
the covariance matrix. We find substantial discrepancies in the predictions
of the two methods. It is believed that the coalescing binary waveform
shuts off abruptly at the onset of the plunge orbit. This has a major effect on
the computations of the covariance matrix as well as on the Monte Carlo
simulations. We discuss the effects of higher post-Newtonian corrections
to the waveform. We also emphasisize the use of the instant of coalescence
as a parameter in order to determine the direction to the
source rather than the time of arrival \cite {BSD95}.
Finally in section \ref{sec_con} we summarise our results and indicate future
directions.
\section{Coalescing Binary Waveforms}
\label{waveforms}
For the purpose of constructing templates for on-line detection,
it is sufficient to work with the so called {\it restricted} post-Newtonian
gravitational waveform. In this approximation
the post-Newtonian corrections are incorporated only in the phase of the
waveform, while ignoring corresponding corrections to the amplitude
\cite {3mn}. Consequently, the restricted post-Newtonian waveforms
only contain the dominant frequency equal to twice the orbital frequency
of the binary computed up to the relevant order.
In the restricted post-Newtonian approximation the gravitational
waves from a binary system of stars, modeled as point masses orbiting
about each other in a circular orbit, induce a strain $h(t)$ at the
detector given by
\begin {equation}
h(t) = A (\pi f(t) )^{2/3} \cos \left [\varphi (t) \right ],
\label {wave}
\end {equation}
where $f(t)$ is the instantaneous gravitational wave frequency,
the constant $A$ involves the distance to the binary, its reduced
and total mass, and the antenna pattern of the detector \cite{Th300},
and the phase of the waveform $\varphi (t)$ contains several pieces
corresponding to different post-Newtonian contributions which
can be schematically written as
\begin {equation}
\varphi(t) = \varphi_0(t) + \varphi_1(t) + \varphi_{1.5}(t) + \ldots.
\label {phase}
\end {equation}
Here $\varphi_0(t)$ is the dominant Newtonian part of the phase
and $\varphi_n$ represents the $n$th order post-Newtonian
correction to it. In the quadrupole approximation we include
only the Newtonian part of the phase given by \cite {Th300}
\begin {equation}
\varphi (t) = \varphi_0 (t) =
{16 \pi f_a \tau_0 \over 5}
\left [ 1 - \left ({f\over f_a}\right )^{-5/3} \right] + \Phi,
\label {phaseN}
\end {equation}
where $f(t)$ is the instantaneous Newtonian gravitational
wave frequency given implicitly by
\begin {equation}
t - t_a = \tau_0
\left [ 1 - \left ( {f \over f_a} \right )^{-8/3} \right ],
\label {frequencyN}
\end {equation}
$\tau_0$ is a constant having dimensions of time given by
\begin {equation}
\tau_0 = {5 \over 256} {\cal M}^{-5/3} (\pi f_a)^{-8/3},
\label {NCT}
\end {equation}
and $f_a$ and $\Phi$ are the instantaneous gravitational
wave frequency and the phase of the signal, respectively, at $t=t_a.$
The time elapsed starting from an epoch when the
gravitational wave frequency is $f_a$ till the epoch when
it becomes infinite will be referred to as
the {\it chirp time} of the signal.
In the quadrupole approximation $\tau_0$ is the chirp time.
The Newtonian part of the phase
is characterised by three parameters:
(i) the {\it time of arrival} $t_a$
when the signal first becomes {\it visible} in the detector,
(ii) the {\it phase} $\Phi$ of the signal at the time of arrival
and (iii) the {\it chirp mass} ${\cal M} = (\mu^3 M^2)^{1/5},$
where $\mu$ and $M$ are the reduced and the total mass
of the binary, respectively.
At this level of approximation two coalescing binary signals of
the same chirp mass but
of different sets of individual masses would be degenerate and thus exhibit
exactly the same time evolution. This
degeneracy is removed when post-Newtonian corrections are included.
When post-Newtonian corrections are included
the parameter space of waveforms acquires an extra dimension.
In this paper we show that even when
post-Newtonian corrections up to relative order $c^{-4},$
where $c$ is the velocity of light, are included in the phase of
the waveform it is possible to make a judicious choice of the
parameters so that the parameter space essentially remains only
three dimensional as far as the detection problem is concerned.
It should, however, be noted that
the evolution of the waveform must be known to a reasonably high
degree of accuracy and that further off-line analysis would be
necessary to extract useful astrophysical information.
With the inclusion of corrections up to second post-Newtonian order the
phase of the waveform becomes \cite {BDI}
\begin {equation}
\varphi (t) = \varphi_0 (t) + \varphi_1 (t) + \varphi_{1.5} (t)
+ \varphi_2 (t)
\label {phasetotal}
\end {equation}
where $\varphi_0(t)$ is given by (\ref {phaseN}) and the various
post-Newtonian contributions are given by
\begin {equation}
\varphi_1 (t) =
4 \pi f_a \tau_1 \left [ 1 - \left ( {f\over f_a} \right )^{-1} \right ]
\label {phase1PN}
\end {equation}
\begin {equation}
\varphi_{1.5} (t) = - 5 \pi f_a \tau_{1.5}
\left [ 1 - \left ( {f\over f_a} \right )^{-2/3} \right ]
\label {phase1.5PN}
\end {equation}
and
\begin {equation}
\varphi_2 (t) = 8 \pi f_a \tau_2
\left [ 1 - \left ( {f\over f_a} \right )^{-1/3} \right ].
\label {phase2PN}
\end {equation}
Now $f(t)$ is the instantaneous gravitational wave frequency correct up to
second post-Newtonian order given implicitly by
\begin {equation}
t - t_a =
\tau_0 \left [ 1 - \left ( {f \over f_a} \right )^{-8/3} \right ] +
\tau_1 \left [ 1 - \left ( {f \over f_a} \right )^{-2} \right ] -\\
\tau_{1.5} \left [ 1 - \left ( {f \over f_a} \right )^{-5/3} \right ]
+ \tau_2 \left [ 1 - \left ( {f \over f_a} \right )^{-4/3} \right ].
\label {frequencyPN}
\end {equation}
In the above equations $\tau$'s are constants having dimensions of time
which depend only on the two masses of the stars and the lower frequency
cutoff of the detector $f_a$.
The total chirp time now consists of four pieces: the Newtonian contribution
$\tau_0$ is given by (\ref {NCT}) and the various post-Newtonian contributions
are
\begin {equation}
\tau_1 = {5 \over 192\mu (\pi f_a)^2} \left ({743\over 336} + {11\over 4} \eta
\right ),
\end {equation}
\begin {equation}
\tau_{1.5} = {1\over 8 \mu} \left ( m \over \pi^2 f_a^5 \right)^{1/3}
\end {equation}
and
\begin {equation}
\tau_2 = {5\over 128 \mu} \left ({m\over \pi^2 f_a^2}\right )^{2/3}
\left ({3058673 \over 1016064} + {5429\over 1008} \eta +
{617\over 144}\eta^2\right)
\end {equation}
where $\eta=\mu/M.$
The phase (\ref {phasetotal})
contains the reduced mass $\mu$ in addition to the chirp mass $\cal M.$
Taking ($\cal M$, $\eta$) to be the post-Newtonian mass parameters
the total mass and the reduced mass are given by
$M = {\cal M} \eta^{-3/5},$ $\mu = {\cal M} \eta^{2/5}.$
Note that in the total chirp time $\tau$ of the signal
the 1.5 post-Newtonian
contribution appears with a negative sign thus shortening the epoch of
coalescence.
With the inclusion of higher order post-Newtonian corrections a
chirp template is characterised by a set of
four parameters which we shall collectively denote by $\lambda^\mu,$
$\mu=1,\ldots, 4.$
At the first post-Newtonian approximation instead of working with the
parameters $\lambda^\mu =\{t_a,~\Phi,~{\cal M},~\eta\}$ we can equivalently
employ the set $\{t_a,~\Phi,~\tau_0,~\tau_1\}$ for the purpose of
constructing templates.
This, as we shall see later, has some advantageous. However,
at post-Newtonian orders beyond the first we do not have a unique
set of chirp times to work with.
The parameters $t_a$ and
$\Phi$ are {\it kinematical} that fix the origin of the
measurement of time and phase, respectively, while the Newtonian and the
post-Newtonian chirp times
are {\it dynamical} parameters in the sense that they dictate the
evolution of the phase and the amplitude of the signal.
It may be mentioned at this stage that in most of the literature
on this subject authors use the set of parameters $\{t_C,~\Phi_C,~{\cal M},~
\eta\}$ where $t_C$ is the instant of coalescence and $\Phi_C$ is
the phase of the signal at the instant of coalescence. In terms of
the chirp times we have introduced, $t_C$
is the sum of the total chirp time and the time of arrival
and $\Phi_C$ is a combination of the various chirp times and $\Phi:$
\begin {equation}
t_C = t_a + \tau_0 + \tau_1 - \tau_{1.5} + \tau_2; \quad
\end {equation}
\begin {equation}
\Phi_C = \Phi + {16\pi f_a \over 5} \tau_0 +
4 \pi f_a \tau_1 - 5\pi f_a \tau_{1.5} + 8\pi f_a \tau_2.
\end {equation}
In the stationary phase approximation
the Fourier transform of the restricted first-post-Newtonian
chirp waveform for positive frequencies is given by \cite
{Th300,SD91,FC93,CF94}
\begin {equation}
\tilde h (f) = {\cal N} f^{-7/6} \exp \left
[i\sum_{\mu=1}^6\psi_\mu(f)\lambda^\mu
- i {\pi \over 4} \right ]
\label {FT}
\end {equation}
where
$${\cal N} = A\pi^{2/3} \left ( {2\tau_0 \over 3}\right )^{1/2} f_a^{4/3} $$
is a normalisation constant, $\lambda^\mu,$ $\mu=1,\ldots, 6,$ represent
the various post-Newtonian parameters
\begin {equation}
\lambda^\mu = \left \{t_a, \Phi, \tau_0, \tau_1, \tau_{1.5}, \tau_2
\right \},
\end {equation}
and
\begin {eqnarray}
\label {eqs1}
\psi_1 & = & 2\pi f, \\
\psi_2 & = & -1, \\
\psi_3 & = & 2 \pi f -{ 16 \pi f_a \over 5}+ {6\pi f_a \over 5}
\left ( {f\over f_a} \right )^{-5/3},\\
\psi_4 & = & 2\pi f - 4\pi f_a + 2\pi f_a \left ({f\over f_a}\right)^{-1},\\
\psi_5 & = &-2\pi f + 5\pi f_a - 3\pi f_a \left ({f\over f_a}\right)^{-2/3},\\
\psi_6 & = & 2\pi f - 8\pi f_a + 6\pi f_a \left ({f\over f_a}\right)^{-1/3}.
\label {eqs2}
\end {eqnarray}
For $f<0$ the Fourier transform is computed using the identity $\tilde
h(-f) = \tilde h^*(f)$ obeyed by real functions $h(t).$ In addition to the
above mentioned
parameters we shall introduce an amplitude parameter $\cal A$ in section
\ref{sm}.
\section{A geometric approach to signal analysis}
\label{sm}
In this section we apply the techniques of differential geometry to the
problem of detecting weak signals embedded in noise. In section \ref{sms} we
introduce the concept of the signal manifold and elaborate on the relationship
of our approach with that of Amari \cite{Am}. Our discussion of the
vector space of all detector outputs is modeled after the discussion
given in \cite{SS94}.
In section \ref{detect} we deal with the problem of choosing a set of
filters for on-line analysis which would optimise the task of detection of
the signal. In section \ref{2pnsec} we deal with the dimensionality of the
chirp manifold when we incorporate higher order post-Newtonian corrections.
It is found, that due to covariances between the parameters, it is possible
to introduce an effective dimension which is less than the dimension of the
manifold.
This has very important implications for the detection problem.
\subsection{Signal manifold}
\label{sms}
The output of a gravitational wave detector such as the
LIGO, will comprise of data segments, each of
duration $T$ seconds, uniformly sampled with a sampling interval of
$\Delta$, giving the number of samples in a single data train to be
$N = T/\Delta$. Each data train can be considered as a $N$-tuple
$(x^0,x^1,\ldots,x^{N-1})$\, $x^i$ being the value of the output of the
detector at time $i\Delta$. The set of all such $N$ tuples constitutes a
$N$-dimensional vector space $\cal V$ where the addition of two vectors is
accomplished by the addition of corresponding time samples.
For later convenience we allow each sample to take complex values. A natural
basis for this vector space is the {\em time basis} ${\bf e}_m^i =
\delta^i_m$ where $m$ and $i$ are the vector and component indices
respectively. Another basis which we shall use extensively is the Fourier
basis which is related to the time basis by a unitary transformation
$\hat U$:
\begin{eqnarray}
{\bf\tilde e}_m &=& \hat U^{mn}{\bf e}_n = \frac{1}{\sqrt N}
\sum\limits_{n=0}^{N-1}
{\bf e}_n\exp
\left[\frac{2\pi i m n}{N}\right],\\
{\bf e}_m &=& \hat U^{\dagger mn}{\bf\tilde e}_n = \frac{1}{\sqrt N}
\sum\limits_{n=0}^{N-1}{\bf\tilde e}_n\exp
\left[-\frac{2\pi i m n}{N}\right].
\end{eqnarray}
All vectors in $\cal V$ are shown in boldface, and the Fourier basis vectors
and
components of vectors in the Fourier basis are highlighted with a `tilde'.
In the continuum case each data train can be expanded in a Fourier series
and will contain a finite number of terms in the expansion, as the
output will be band limited. The expansion is
carried out over the exponential functions $\exp\left(2\pi i m t/T\right)$
which are precisely the Fourier basis vectors defined above. Though the index
$m$ takes both positive and negative values corresponding to positive and
negative frequencies it is both, possible and convenient to
allow $m$ to take only positive values \cite{NUM}. Thus the
vector space $\cal V$ can be considered as being spanned by the $N$ Fourier
basis vectors implying immediately that the number of independent vectors in
the time basis to be also $N$. This is the content of the Nyquist theorem
which states that it is sufficient to sample the data at a
frequency which is twice as large as the bandwidth of a real valued signal,
where the bandwidth
refers to the range of positive frequencies over which the signal spectrum is
non zero.
This factor of two does not appear in the vector space picture as we allow
in general for complex values for the components in the time basis.
A gravitational wave signal from a coalescing binary system
can be characterised by a set of parameters
$\bbox{\lambda} = (\lambda^0,\lambda^1,\ldots,\lambda^{p-1})$ belonging to some
open set of the $p$-dimensional real space $R^p$.
The set of such signals
${\bf s}(t;\bbox{\lambda})$ constitutes a $p$-dimensional manifold $\cal S$
which is
embedded in the vector space $\cal V$. The parameters of the binary act as
coordinates on the manifold. The basic problem of signal analysis is thus to
determine whether the detector output vector $\bf x$ is
the sum of a signal vector and a noise vector, ${\bf x} = {\bf s} + {\bf n}$,
or just the noise vector, ${\bf x} = {\bf n}$, and furthermore to identify
which
particular signal vector, among all possible.
One would also like to estimate the errors in such a measurement.
In the absence of the signal the output will contain
only noise drawn from a stochastic process which can be described by a
probability distribution on the vector space $\cal V$.
The covariance matrix of the noise $C^{ij}$ is defined as,
\begin{equation}
C^{ij} = \overline{n^in^{*j}},
\end{equation}
where an * denotes complex conjugation and an overbar denotes an
average over an ensemble. If the noise is assumed to be stationary and ergodic
then
there exists a noise correlation function $K(t)$ such that $C_{ij} =
K(|i-j|\Delta)$. In the Fourier basis it can be shown that the components
of the noise vector are statistically independent \cite{HEL} and the
covariance matrix in the Fourier
basis will contain only diagonal terms whose values will be strictly positive:
$\tilde C_{ii} = \overline{\tilde n^i\tilde n^{*i}}$.
This implies that the covariance matrix
has strictly positive eigenvalues. The diagonal elements of this
matrix $\tilde C_{ii}$ constitute the discrete representation of the power
spectrum of the noise $S_n(f)$.
We now discuss how the concept of matched filtering can be used to induce a
metric
on the signal manifold. The technique of matched filtering involves correlating
the
detector output with a bank of filters each of which is tuned to detect the
gravitational wave from a binary system with a particular set of parameters.
The output of the filter, with an impulse response $\bf q$, is given in the
discrete case
as
\begin{equation}
\label{corr}
c_{(m)} = \frac{1}{\sqrt N}\sum\limits_{n=0}^{N-1}\tilde x^n \tilde q^{*n}
\exp\left[-2\pi imn/N\right].
\end{equation}
The SNR ($\rho$) at the output is defined to be the mean of $c_{(m)}$ divided
by the square root of its variance:
\begin{equation}
\label{rho1}
\rho \equiv \frac{\overline{c_{(m)}}}{\left[\overline{\left(c_{(m)} -
\overline{c_{(m)}}\right)^2}
\right]^{1/2}}.
\end{equation}
By maximising $\rho$ we can obtain the expression
for the matched filter $\bf q_{(m)}$ matched to a particular signal ${\bf
s}(t;\lambda^\mu)$ as
\begin{equation}
\label{matched}
\tilde q^n_{(m)}(\lambda^\mu) = \frac{\tilde s^n(\lambda^\mu)\exp\left[2\pi
imn/N\right]}{\tilde C_{nn}},
\end{equation}
where $\rho$ has been maximised at the $m^{th}$ data point at the output and
where $\mu = 1,2,\ldots,p$
where $p$ is the number of parameters of the signal. We now introduce a scalar
product in $\cal V$.
For any two vectors $\bf x$ and $\bf y$,
\begin{equation}
\label{scal}
\left\langle{\bf x},{\bf s}\right\rangle =
\sum\limits_{i,j=0}^{N-1}C^{-1}_{ij}x^iy^j =
\sum\limits_{n=0}^{N-1}\frac{\tilde x^n \tilde s^{*n}(\lambda^\mu)}{\tilde
C_{nn}}.
\end{equation}
In terms of this scalar product, the output of the matched filter $\bf q$,
matched to a signal
${\bf s}(\lambda^\mu)$, can be written as,
\begin{equation}
\label{eqsc}
c_{(m)}(\lambda^\mu) = \frac{1}{\sqrt N}\left\langle{\bf x},{\bf
s}\right\rangle.
\end{equation}
As $\tilde C_{ii}$ is strictly positive the scalar product defined
is positive definite.
The scalar product defined above on the vector space $\cal V$ can be used to
define a norm
on $\cal V$ which in turn can be used to induce a metric on the manifold. The
norm
of a vector $\bf x$ is defined as $\|{\bf x}\| = \left\langle{\bf x},{\bf x}
\right\rangle^{1/2}$.
The ratio for the matched filter can be
calculated to give $\rho = \left\langle{\bf s},{\bf s}\right\rangle^{1/2}$.
The norm of the noise vector will be a random variable $\left\langle{\bf
n},{\bf n}\right\rangle^{1/2}$
with a mean of $\sqrt{N}$ as can be seen by writing the expression for the norm
of the noise
vector and subsequently taking an ensemble average.
The distance between two points infinitesimally separated on $\cal S$ can be
expressed as
a quadratic form in
the differences in the values of the parameters at the two points:
\begin{eqnarray}
g_{\mu\nu}d\lambda^\mu d\lambda^\nu &\equiv& \|{\bf s}(\lambda^\mu +
d\lambda^\mu) -
{\bf s}(\lambda^\mu)\| = \|\frac{\partial{\bf
s}}{\partial\lambda^\mu}d\lambda^\mu\|\\
&=& \left\langle\frac{\partial{\bf s}}{\partial\lambda^\mu},
\frac{\partial{\bf s}}{\partial\lambda^\nu}\right\rangle d\lambda^\mu
d\lambda^\nu
\end{eqnarray}
The components of the metric in the coordinate basis are seen to be the scalar
products
of the coordinate basis vectors of the manifold.
Since the number of correlations we can perform on-line is finite, we cannot
have a
filter corresponding
to every signal. A single filter though matched to a particular signal will
also
detect signals in a small neighbourhood of that signal but with a loss
in the SNR. The metric on the manifold quantifies the drop in the
correlation
in a neighbourhood of the signal chosen.
Taking the output vector to be $\bf x$ and two signal vectors
$\bf s(\bbox{\lambda})$ and $\bf s(\bbox{\lambda}+d\bbox{\lambda})$
and using Schwarz's inequality we have,
\begin{eqnarray}
\left\langle{\bf x},{\bf s(\bbox{\lambda}+d\bbox{\lambda})}\right\rangle -
\left\langle{\bf x},{\bf s(\bbox{\lambda})}\right\rangle
= \left\langle{\bf x},{\bf s}(\bbox{\lambda}+d\bbox{\lambda}) - {\bf
s}(\bbox{\lambda})
\right\rangle
&\le& \|{\bf x}\|\|{\bf s}(\bbox{\lambda} + d\bbox{\lambda}) - {\bf
s}(\bbox{\lambda})\|\\
&=& \|{\bf x}\|g_{\mu\nu}d\lambda^\mu d\lambda^\nu.
\end{eqnarray}
As is apparent the drop in the correlation can be related to the metric
distance on the manifold between the two signal vectors.
We now discuss Amari's \cite{Am} work in the context of using differential
geometry in statistics and elaborate on the relationship with the approach
we have taken. The set of parametrised probability distributions corresponding
to a statistical model constitute a manifold. The parameterized probability
distributions in the context of signal analysis of gravitational waves from
coalescing binaries are the ones which specify the probability that the output
vector will lie in a certain region of the vector space $\cal V$ given that a
signal
${\bf s}(t;\bbox{\lambda})$ exists in the output which we denote as $p({\bf x}|
{\bf s}(t;\bbox{\lambda}))$
Since it is not our intention
to develop Amari's approach any further we will be brief and will make
all the mathematical assumptions such as infinite differentiability of
functions,
interchangeability of the differentiation and expectation value operators, etc.
The set of probability distributions $p({\bf x}|{\bf s}(\bbox{\lambda}))$
where $\bbox{\lambda} \in R^p$, constitutes a manifold $\cal P$ of dimension
$p$.
At every point
on this manifold we can construct a tangent space $T^0$ on which we can define
the
coordinate basis vectors as ${\bf\partial}_\mu =
\frac{\bf\partial}{{\bf\partial}\lambda^\mu}$.
Any vector $\bf A$ in this tangent space can be written as a linear combination
of these coordinate basis vectors.
We now define $p$ random variables $\sigma_\mu =
-\frac{\partial}{\partial\lambda^\mu}\log(p({\bf x}|{\bf s}(\bbox{\lambda})))$.
It can easily be shown that $\overline{\sigma_\mu} = 0$. We assume that these
$p$ random
variables are linearly independent. By taking all possible linear combinations
of these random
variables we can construct another linear space $T^1$. Each vector
$\bf B$ in $T^1$ can be written as ${\bf B} = B^\mu\sigma_\mu$. The two vector
spaces
$T^0$ and $T^1$
are isomorphic to each other, which can be shown explicitly by
making the correspondence $\sigma_\mu \leftrightarrow {\bf\partial}_\mu$. The
vector space
$T^1$ has a natural inner product defined on it which is the covariance
matrix of the $p$ random variables $\sigma_\mu$. This scalar product can be
carried
over to $T^0$ using the correspondence stated above. The metric on the
manifold can be defined by taking the scalar product of the coordinate basis
vectors
\begin{equation}
\label{met}
g_{\mu\nu} = \left\langle \partial_\mu,\partial_\nu\right\rangle =
\overline{\sigma_\mu\sigma_\nu}
\end{equation}
In statistical theory the above matrix $g_{\mu\nu}$ is called the Fisher
information
matrix. We will also denote the Fisher matrix, as is conventional, as
$\Gamma_{\mu\nu}$.
It is clearly seen that orthogonality between vectors in the tangent
space of the manifold is related to statistical independence of random
variables
in $T^1$.
If we take the case of Gaussian noise the metric defined above
is identical to the one obtained on the signal manifold by matched filtering.
Gaussian noise can be described by the distribution,
\begin{equation}
\label{ndis}
p({\bf n}) = \frac{\exp\left[
-\frac{1}{2}\sum\limits_{j,k=0}^{N-1}{C^{-1}_{jk} n^j n^{k*}}\right]}
{ \left[\ (2\pi)^N \det\left[C_{ij}\right]\ \right]^{1/2}} = \frac{\exp\left[
-\frac{1}{2}\sum\limits_{j,k=0}^{N-1}{\tilde C^{-1}_{jk} \tilde n^j
\tilde n^{k*}}\right]} { \left[\ (2\pi)^N \det\left[\tilde C_{ij}\right]\
\right]^{1/2}}
= \frac{\exp\left[
-\frac{1}{2}\sum\limits_{i=0}^{N-1}{\frac{\tilde n^i
\tilde n^{i*}}{\tilde C_{ii}}}\right]} { \left[\ (2\pi)^N \det\left[\tilde
C_{ij}\right]\ \right]^{1/2}},
\end{equation}
where in the last step we have used the diagonal property of the matrix $\tilde
C^{ij}$
which implies that $\tilde C^{-1}_{ii} = 1/\tilde C_{ii}$.
As the noise is additive $p({\bf x}|{\bf s}(\bbox{\lambda}))$ can be written
as $p({\bf x} - {\bf s}(\bbox{\lambda}))$. Assuming Gaussian noise we can
write the
expressions for the random variables $\sigma_i$ as,
\begin{equation}
\label{sig}
\sigma_\mu =
\left(\frac{1}{2}\right)\frac{\partial}{\partial\lambda^\mu}\left\langle{\bf
x}-{\bf s}(\bbox{\lambda}),
{\bf x} - {\bf s(\bbox{\lambda})}\right\rangle = \left\langle{\bf
n},\frac{\partial}{\partial\lambda^\mu}
{\bf s}(\bbox{\lambda})\right\rangle,
\end{equation}
where in the last step we have used ${\bf x} = {\bf s}(\bbox{\lambda})+{\bf
n}$. The covariance matrix
for the random variables $\sigma_\mu$ can be calculated to give
\begin{equation}
\label{sigco}
\overline{\sigma_\mu\sigma_\nu} =
\left\langle\frac{\partial}{\partial\lambda^\mu}{\bf s}(\bbox{\lambda}),
\frac{\partial}{\partial\lambda^\nu}{\bf s}(\bbox{\lambda})\right\rangle,
\end{equation}
which is the same metric as defined over the signal manifold. Thus, both the
manifolds
$\cal S$ and $\cal P$ are identical with respect to their metrical properties.
We will henceforth restrict our attention to the signal manifold $\cal S$.
For the purpose of our analysis we will choose a minimal set of parameters
characterizing the gravitational wave signal from a coalescing binary. We
consider
only the first post-Newtonian corrections. In section \ref{waveforms} we have
already
introduced the four parameters $\lambda^\mu = \{t_a,\Phi,\tau_0,\tau_1\}$. We
now introduce
an additional parameter for the amplitude and call it $\lambda^0 = \cal A$.
The signal can now be written as $\tilde s(f;\bbox{\lambda}) =
{\cal A} h(f;t_a,\Phi,\tau_0,\tau_1)$, where,
$\bbox{\lambda} \equiv \{{\cal A},t_a,\Phi,\tau_0,\tau_1\}$. Numerically
the value of the parameter $\cal A$ will be the same as that of the SNR
obtained for the matched filter. We can decompose the signal manifold into a
manifold containing normalised chirp waveforms and a one-dimensional manifold
corresponding
to the parameter $\cal A$. The normalised chirp manifold can therefore be
parameterized by
$\{t_a,\Phi,\tau_0,\tau_1\}$.
This parameterization is useful as the coordinate basis vector
$\frac{\partial}{\partial
{\cal A}}$ will be orthogonal to all the other basis vectors as will be seen
below.
In order to compute the metric, and equivalently the Fisher information matrix,
we use the continuum version of the scalar product as given in \cite{F92},
except that we
use the two sided power spectral density.
This has the advantage of showing clearly
the range of integration in the frequency space though we get the same result
using the discrete
version of the scalar product. Using the definition of the scalar product we
get
\begin {equation}
g_{\mu\nu} = \int_{f_a}^\infty {df\over S_n(f)}
{\partial \tilde s(f; \bbox{\lambda})\over \partial \lambda^\mu}
{\partial \tilde s^* (f; \bbox{\lambda}) \over \partial \lambda^\nu}
+ {\rm c.c.}
\label{gamma2}
\end {equation}
Recall that in the
stationary phase approximation the Fourier transform of
the coalescing binary waveform is given by $\tilde h(f) = {\cal N} f^{-7/6}
\exp \left [ i\sum_\mu\psi_\mu(f) \lambda^\mu \right]$ and $\tilde s(f) = {\cal
A}\tilde h(f)$,
where $\psi_\mu(f)$ are given
by equations (\ref {eqs1}-\ref{eqs2}), $\mu=1,\ldots, 4$ and
$\lambda^\mu = \{t_a,~\Phi,~\tau_0,~\tau_1\}.$
Note, in particular, that in the phase of the waveform
the parameters occur linearly thus enabling a very concise expression
for the components of $g_{\mu\nu}.$ The various partial
derivatives are given by
\begin {equation}
{\partial \tilde s (f;\bbox{\lambda}) \over \partial {\lambda^\mu}} =
i \psi_\mu (f) \tilde s(f;\bbox{\lambda}),
\label{derivatives}
\end {equation}
where we have introduced $\psi_0= -i/{\cal A}.$
On substituting the above expressions for the partial derivatives in
eq. (\ref {gamma2}) we get,
\begin {equation}
g_{\mu\nu} = \left < \psi_\mu h, \psi_\nu h \right >
= 2 \int_{f_a}^\infty {{\psi_\mu(f) \psi^*_\nu(f) \left |\tilde h(f)\right|^2}
\over S_n(f)} df
\label{gamma3}
\end {equation}
The above definition of the amplitude parameter ${\cal A},$
as in Culter and Flannagan \cite {CF94}, disjoins the amplitude of the waveform
from the rest of the parameters. Since $\psi_0$ is pure imaginary and
$\psi_\mu$'s are real, it is straightforward to see from eq. (\ref {gamma3})
that
\begin {equation}
g_{00} = 1 \mbox{ and } g_{0\mu} = 0; \ \ \mu=1,\ldots,4.
\end {equation}
The rest of the components $g_{\mu\nu}$ are seen to be independent of all the
parameters
except $\cal A$ {\em i.e.} $g_{\mu\nu} \propto {\cal A}^2$. As ${\cal A}$ is
unity for the
normalised manifold the metric on the normalised manifold is flat. This implies
not only that the manifold is intrinsically flat (in the stationary phase
approximation)
but also that the coordinate system used is Cartesian. If instead of the chirp
time
$\tau_0$ we use the parameter $\cal M$ then the metric coefficients will
involve that parameter
and the coordinate system will no longer remain Cartesian.
\subsection{Choice of filters }
\label{detect}
We now use the formalism developed to tackle the issue of optimal filter
placement. Till
now, it has been thought necessary to use a finite subset of the set of chirp
signals
as templates for detection. We show that this is unduly restrictive. We suggest
a procedure by which the
detection process can be made more `efficient' by moving the filters out of
the manifold.
It must be emphasised that the algorithm presented below is both, simplistic
and quite adhoc and
is not necessarily the best.
Moreover, we have implemented the algorithm only for the Newtonian case where
the computational
requirements are not very heavy. The signal manifold corresponding to higher
post-Newtonian
corrections will be a higher dimensional manifold and here the computational
requirements
will be substantial. The choice of optimal filters which span the manifold will
then be crucial.
Detection of the coalescing binary signal involves computing the scalar
product
of the output of the detector with the signal vectors. Subsequently one would
have to maximise
the correlations over the parameters and the number so obtained would serve as
the
statistic on the basis of which
we can decide whether a signal is present in the given data train.
Geometrically, this maximization
corresponds to minimizing the angle between the output vector and the vectors
corresponding
to the normalised signal manifold. Using the cosine formula,
\begin{equation}
\cos(\theta) =
\frac{\left<{\bf s}(\bbox{\lambda}),{\bf x}\right>}{\|{\bf x}\|\|{\bf
s}(\bbox{\lambda})\|}
= \frac{\|{\bf s}(\bbox{\lambda})\|^2 + \|{\bf x}\|^2 -
\|{\bf x} - {\bf s}(\bbox{\lambda})\|^2}
{2\|{\bf x}\|\|{\bf s}(\bbox{\lambda})\|},
\end{equation}
it is clear that as $\|{\bf s}(\bbox{\lambda})\|$ is unity for the vectors
belonging to the
normalised signal manifold,
maximising the scalar product is equivalent to minimising
$ \|{\bf x} - {\bf s}(\bbox{\lambda})\|$ which is the
distance between the tip of the output vector and the manifold.
Given the constraints of computational
power one would be able to evaluate only a finite number of these
scalar products, say $n_F$, in a certain amount of time depending on the data
train length
and the padding factor.
It is therefore necessary to be able to choose the $n_F$ filters in such a
manner that
the detection probability is maximal.
We will need efficient on-line data analysis for two reasons:
(i) To isolate those
data trains which have a high probability of containing a signal and (ii) to
determine the parameters
of the binary early on during the inspiral and to use them for dynamical
recycling techniques.
Due to the finiteness
of the filter spacing the signal parameters will in general not correspond
to any of the $n_F$ filters chosen and this will lead to a drop in the maximum
possible
correlation.
Till now attention has been focussed on identifying a set of optimal set of
filters which are
a discrete subset of the manifold.
If detection is the sole purpose then the differential geometric
picture suggests that confining the filter vectors
to the signal manifold is an unnecessary restriction and in fact non optimal.
Thus it
is worthwhile to explore making a choice of filters outside the manifold.
The filter vectors will thus belong to $\cal V$ but will not, in general,
correspond to any signal.
It is, of course, true that we are sacrificing on the maximum possible
correlation obtainable (when
the signal's parameters coincide with those of the filter). Thus the problem
essentially is
to select $n_F$ filter vectors which optimize the detection the efficiency of
which depends
upon the properties of the manifold.
In general a single filter vector would have to pick up signals over a region
of the manifold. The extent
of this region is determined by fixing a threshold on the correlation between
the filter and
any signal in the region. We will denote this threshold by $\eta$, where $\eta$
takes a value
which is close to, but less than unity. The typical value suggested for $\eta$
is $\sim 0.8$,
\cite{SD91}.
For a given filter $\bf q$ and a threshold
$\eta$ the region on the manifold corresponding to the filter will be denoted
as
${\cal S}_{\bf q}(\eta)$, where ${\cal S}_{\bf q}(\eta) \subset \cal S$.
Geometrically this region is the intersection of an open ball in $\cal V$ of
radius
$2^{1/2}(1-\eta)^{1/2}$ (using the distance defined by the scalar product),
with center $\bf q$, and the manifold.
The $n_F$ filters taken together would have to `span' the manifold which means
that the union of
the regions covered by each filter would be the manifold itself {\em i.e.},
$\bigcup_{\bf q} {\cal S}_{\bf q}(\eta) = \cal S$.
If the filter $\bf q$ lies on the manifold, then the correlation
function $c_{\bf q}(\bbox{\lambda}) = \left<{\bf s}(\bbox{\lambda}),
{\bf q}\right>$ will reach its maximum value of unity in
${\cal S}_{\bf q}(\eta)$ when ${\bf q} = {\bf s}(\bbox{\lambda})$
and will fall off in all directions. This means that the signals
in the region which are further away from the filter are less
likely to be picked up as compared to those in the immediate
neighbourhood of the filter $\bf q$.
We assume that a finite subset of the normalised signal manifold has been
chosen to act as filters
by some suitable algorithm \cite{SD91}, which taken together span the manifold.
The number of filters will be determined by the available computing power.
Consider one of these filters $\bf q$, the region corresponding to it for a
threshold of $\eta$,
${\cal S}_{\bf q}(\eta)$, and an arbitrary normalised vector ${\bf q}_o$
which belongs to $\cal V$
but not necessarily $\cal S$. By correlating the vector ${\bf q}_o$ with
vectors in
${\cal S}_{\bf q}(\eta)$ we obtain the correlation function $c_{{\bf
q}_o}(\bbox{\lambda}) =
\left<{\bf s}(\bbox{\lambda}),{\bf q}_o\right>$. We select that ${\bf q}_o$ to
serve as a more
optimal filter which maximises the average of this correlation function:
\begin{equation}
\left<{\bf s}(\bbox{\lambda}),{\bf q}_o\right>_{av} =
\frac{1}{\int\limits_{{\cal S}_{\bf q}(\eta)}
\sqrt{g}d^p\lambda}\int\limits_{{\cal S}_{\bf q}(\eta)}\left<{\bf
s}(\bbox{\lambda}),
{\bf q}_o\right>\sqrt{g}d^p\lambda= \left<
\int\limits_{{\cal S}_{\bf q}(\eta)}{\bf s}(\bbox{\lambda})d^p\lambda\ ,\ {\bf
q}_o\right>,
\end{equation}
where $g = det\left[g_{\mu\nu}\right]$. In the last step above
the integration and the scalar product operations in the
above equation have been interchanged. Moreover, for the normalised
chirp manifold the metric does not depend upon the parameters in the coordinate
system we have
chosen and therefore, $g_{\mu\nu}$ is a constant and the factor $\sqrt{g}$
cancels.
We now use Schwarz's inequality to maximise the average correlation to obtain,
\begin{equation}
\label{avcorr}
{\bf q}_o = {\cal N}\int\limits_{{\cal S}_{\bf q}(\eta)}{\bf
s}(\bbox{\lambda})d^p\lambda,
\end{equation}
where $\cal N$ is just a normalisation constant.
We implemented the above algorithm for filter placement for the case of
Newtonian
signals with certain modifications.
The normalised chirp waveform consists of three parameters $(\Phi,t_a,\tau_n)$.
If we keep $t_a$ and $\tau_n$ fixed then the tip of the signal vector traces
out a circle
as we vary $\Phi$. As any circle lies on a plane we can express a signal vector
as a
linear sum of two vectors where the two vectors differ only in the phase
parameter
and we take this phase difference to be $\pi/2$. Thus we need
only two mutually orthogonal
filters to span the phase parameter. The time of arrival parameter
$t_a$ is also a `convenient parameter' as by the use of fast Fourier transforms
the correlations
for arbitrary time of arrivals can be performed at one go. It is therefore not
profitable for us
to maximise the average correlation over the phase parameter $\Phi$ and the
time of arrival
$t_a$.
In view of the above restrictions we modified the the filter placement
algorithm.
We consider the correlation function for the case when the filter vector is on
the manifold.
We define the `line of curvature' to be the curve on the manifold along which
the
correlation function falls the least. Figure (\ref{fig_dets}) illustrates the
correlation function plotted
as a function of $\tau_0$ along the line of curvature.
It is seen from the contour diagram of the numerically computed
correlation function that the line of curvature lies nearly on the
submanifold $t_a + \tau_n = $ constant, of the normalised chirp manifold. We
take two curves
passing through the point $\bf q$ in the region ${\cal S}_{\bf q}(\eta)$:
\begin{enumerate}
\item $t_a + \tau_n = $ constant,\ \ $\Phi = 0$ and
\item $t_a + \tau_n = $ constant,\ \ $\Phi = \pi/2$.
\end{enumerate}
We obtain one filter for each of the two curves by evaluating eq. \ref{avcorr}
where the
domains of integration correspond to the segments of the curves defined.
Having determined the two filters we again plot the correlation function along
the line of curvature as a function of $\tau_0$ in figure (\ref{fig_dets}).
The region of the manifold selected corresponds to a range of $5.8$~secs to
$6.0$~secs
in the parameter $\tau_0$. It is interesting to note that we could have
translated these
values keeping the difference same without affecting our results.
It can be seen that the correlation has a minimum at the center.
In order to get a flatter correlation curve we select a linear
combination of the original filter and the one obtained by averaging with
suitable
weights attached to each filter. This performs reasonably well as shown by the
thick curve in
the figure. The importance of having a flatter correlation function lies in the
fact
all the signals in a region can be picked up with equal efficiency and the drop
in the maximum possible correlation can be compensated for by lowering the
threshold.
The average correlation obtained for the optimal filter is only marginally
better than
that obtained for the filter placed on the manifold.
In the discussion above, we had started with a fixed number of filters $n_F$
on the manifold and obtained another set of $n_F$ filters which performs
marginally
better than the former set.
Equivalently we can try to increase the span of each filter
retaining the same threshold but reducing the number of filters
required. In Figure \ref{fig_dets} we observe that the optimal filter chosen
spans
the entire region considered with a threshold greater than $0.9$, whereas the
filter on the manifold
spans about half the region at the same threshold. This indicates that by
moving the filters out of the
manifold in the above manner it may be possible to reduce the number of filters
by a factor of two
or so. One must however, bear in mind that the bank of filters obtained in this
way are not optimal. There is scope to improve the scheme further. Our analysis
is indicative of this feature.
\subsection {Effective dimensionality of the parameter space of a second
order post-Newtonian waveform}
\label{2pnsec}
It has already been shown that the first post-Newtonian waveform is essentially
one-dimensional \cite{Sat94}.
We argue in this subsection that even the second post-Newtonian waveform is
essentially
one-dimensional and a one-dimensional lattice suffices to filter the waveform.
A Newtonian waveform is characterised by a set of three parameters
consisting of the time of arrival, the phase of the signal at the
time of arrival and chirp mass (equivalently, Newtonian chirp time).
In this case, for the purpose of detection,
one essentially needs to employ a one-dimensional lattice
of filters corresponding to the chirp mass, the time of arrival
being taken care by the fast Fourier transform algorithm and the
phase being determined using a two-dimensional basis of orthogonal templates.
When post-Newtonian corrections are included in the phase of the
waveform the number of parameters increases apparently implying that
one needs to use a two-dimensional lattice of filters corresponding to,
say, the chirp and reduced masses (equivalently the Newtonian and
post-Newtonian
chirp times) which in turn means that the number of templates through which
the detector output needs to be filtered goes up by several orders of
magnitude. One of us (BSS) has recently shown that for the purpose
of detection it is sufficient to use a one-dimensional lattice of filters
even after first post-Newtonian corrections are included in the phase of
the waveform and the relevant parameter here is the sum of the Newtonian
and post-Newtonian chirp times. What happens when corrections beyond
the first post-Newtonian order are incorporated in the phase of the waveform?
The coalescing binary waveform is now available up to second post-Newtonian
order \cite {BDI,BDIWW}. Blanchet et al. argue that the phase correction
due to the second order post-Newtonian (2PN) term induces an accumulated
difference of 10 cycles in a total of 16000. Consequently, it is
important to incorporate the 2PN terms in the templates.
When the 2PN terms are included it is useful to consider that the full
waveform is parameterised by three additional parameters, corresponding to the
chirp times at the 1, 1.5 and 2PN order (cf. Sec \ref{waveforms}), as compared
to the Newtonian waveform. Of course, as far as the detection problem
is concerned there is only one additional parameter since the chirp times
are all functions of the two masses of the binary. However, for the purpose
of testing general relativity one can consider each of the chirp times
to be independent of the rest \cite{BS95,BS94}. Our problem now is
to find the dimensionality of the parameter space of a 2PN waveform.
To this end we consider the {\it ambiguity function} $C({\bbox
{\lambda}}^{'},{\bbox {\lambda}})$
which is nothing but
the correlation function of two normalised waveforms one of whose parameters
(${\bbox {\lambda}}$) are varied by holding the parameters of the other fixed
(${\bbox {\lambda}}^{'}$) :
\begin {equation}
C({\bbox {\lambda}}^{'}, {\bbox {\lambda}}) = \left < q({\bbox {\lambda}}^{'}),
q({\bbox {\lambda}}) \right >;
\ \ \left < q({\bbox {\lambda}}^{'}), q({\bbox {\lambda}}^{'}) \right > =
\left < q({\bbox {\lambda}}^{\phantom{'}}),({\bbox {\lambda}}) \right > = 1.
\end {equation}
It is useful to think of ${\bbox {\lambda}}^{'}$ to be the parameters of a
template and ${\bbox {\lambda}}$ to be that of a signal. With this
interpretation
the ambiguity function simply gives the span of a filter in the parameter
space.
The ambiguity function for the full waveform is a four-dimensional
surface since there are four independent parameters. To explore the
effective dimensionality of the parameter we consider the set of
parameters to
be $\{t_a, \Phi, m_1, m_2\},$ where $m_1$ and $m_2$ are the two masses
of the binary. We have shown the contours of the ambiguity function maximised
over
$t_a$ and $\Phi$ (since these two parameters do not explicitly
need a lattice of templates) in Fig. \ref{ambcont}.
The template at the centre of the plot corresponds to a binary
waveform with $m_1=m_2=1.4 M_\odot$ and the signal parameters are
varied over the entire astrophysically interesting range of masses:
$m_1, m_2 \in [1.4,10]M_\odot.$ From these figures we find that
the ambiguity function is almost a constant along a particular
line in the $m_1$--$m_2$ plane. This means that a template at the
centre of the grid spans a relatively large area of the parameter
space by obtaining a correlation very close to unity for all signals
whose masses lie on the curve along which the ambiguity function
roughly remains a constant. It turns out that the equation of this
curve is given by
\begin {equation}
\tau_{0} + \tau_{1} - \tau_{1.5} + \tau_{2} = \mbox{const.}
\end {equation}
Let us suppose we begin with a two-dimensional lattice of filters
corresponding to a certain grid (albeit, nonuniform) laid in the
$m_1$--$m_2$ plane. Several templates of this set will have their
total chirp time the same. Now with the aid of just one template
out of all those
that have the same chirp time we can effectively span the region that
is collectively spanned by all such filters.
More precisely, we will not have an appreciable loss in the SNR in
replacing all templates of a given total chirp time by one of them.
Consequently, the signal manifold can be spanned by a one-dimensional
lattice of templates.
\section {Estimation of parameters}
\label{sec_espar}
In this section we discuss the accuracy at which the various parameters
of a coalescing binary system of stars can be estimated. All our results
are for a single interferometer of the initial LIGO-type which has a
lower frequency cutoff at 40 Hz. At present it is beyond the computer
resources available to us to carry out a simulation for the advanced
LIGO. In the first
part of this section we briefly review the well known results obtained
for the variances and covariances in the estimation of parameters using
analytical methods. Analytical methods assume that the SNR is sufficiently
large (the so-called strong signal approximation) and implicitly use a
continuum of the parameter
space. In reality, however, these assumptions are not necessarily valid
and hence it is essential to substantiate the results obtained using
analytical means by performing numerical simulations.
In the second part of this section we present an exhaustive discussion
of the Monte Carlo simulations we have performed to compute the errors and
covariances of different parameters. As we shall discuss
below the computation of errors using the covariance matrix is erroneous even
at an SNR of 10-20. Our estimation of 1-$\sigma$ uncertainty in the
various parameters, at low SNRs, is substantially larger than those computed
using
the covariance matrix. However, for high values of the SNR
($>25$--$30$) Monte Carlo estimation agrees with the analytical results.
\subsection {Covariance matrix} \label {anacovar}
In recent years a number of authors have addressed issues related
to the variances expected in the parameter estimation
\cite{BS95,KKS95,F92,FC93,CF94,BS94,Kr,KLM93,PW95,JKKT},
In the standard method of computing the variances in the estimation of
parameters one makes the assumption that the SNR is
so large that
with the aid of such an approximation one can first construct
the Fisher information matrix $\Gamma_{\mu\nu}$
and then take its inverse to obtain the covariance matrix
$C_{\mu\nu}.$
In the strong signal approximation the Fisher information matrix
and the covariance matrix are given by
\begin {equation}
g_{\mu\nu} = \Gamma_{\mu\nu} = \left < {\partial {\bf s} \over \partial
\lambda^\mu},
{\partial {\bf s} \over \partial \lambda^\nu} \right >;
\ \ \ C_{\mu\nu} = \Gamma^{-1}{\mu\nu}.
\label{gamma1}
\end {equation}
As we have seen before the Fisher information and consequently the covariance
matrix is block diagonal and hence there is no cross-talk,
implying vanishing of the covariances between the amplitude and
the other parameters. Consequently, we need not construct, for the
purpose of Weiner filtering, templates corresponding to different amplitudes.
The judicious choice of parameters has also allowed a very elegant expression
for the Fisher information matrix. It is particularly interesting to note
that the information matrix does not depend on the values of the various
parameters, except for the amplitude parameter, and hence is a constant as far
as the parameter space of the normalised waveforms is considered.
For the purpose of numerical simulations
it is convenient to choose the set
$\lambda^\mu = \{{\cal A}, t_a, \Phi, \tau_0, \tau_1\}$
where $\cal A$ is the amplitude parameter, $t_a$ and $\Phi$ are the
time of arrival of the signal and its phase at the time of arrival,
and $\tau_0$ and $\tau_1$ are the Newtonian and the post-Newtonian
coalescence times.
For noise in realistic detectors,
such as LIGO, the elements of the Fisher information matrix cannot be
expressed in a closed form and, for the set of parameters employed,
it is not useful to explicitly write
down the covariance matrix in terms of the various integrals since the
errors and covariances do not have any dependence on the parameters.
We thus evaluate the information matrix numerically and then take
its inverse to obtain the covariance matrix. Instead of dealing with
the covariance matrix $C$ it more instructive to work with the matrix
of standard deviations and correlation coefficients
$D$ which is related to the former by
\begin {equation}
{D_{\mu\nu}} = \left \{ \begin{array} {ll}
\sqrt {C_{\mu\nu}}, & {\rm if} \ \mu=\nu\\
C_{\mu\nu}/(\sigma_\mu \sigma_\nu) & {\rm if}\ \mu \ne \nu,
\end {array}\right.
\label {varcorr}
\end {equation}
where $\sigma_\mu={D}_{\mu\mu}$ is the 1-$\sigma$ uncertainty in
the parameter $\lambda_\mu.$ The off-diagonal elements of $D$ take
on values in the range $[-1, 1]$ indicating
how two different parameters are correlated: For $\mu\ne\nu,$
${D}_{\mu\nu}=1$, indicates that the two are perfectly correlated,
${D}_{\mu\nu}=-1$ means that they are perfectly anticorrelated and
${D}_{\mu\nu}=0$ implies that they are uncorrelated.
Since the information matrix is block-diagonal,
the amplitude parameter is totally uncorrelated
with the rest and thus an error in the measurement of $\cal A$ will not reflect
itself as an error in the estimation of the other parameters and vice versa.
In contrast, as we shall see below,
Newtonian chirp time is strongly anticorrelated to
post-Newtonian chirp time, which implies that if in a given experiment
$\tau_0$ happens to be estimated larger than its true value then
it is more likely that $\tau_1$ will be estimated to be lower than
its actual value. Such correlations are useful as far as detection
is concerned since they tend to reduce the number of templates needed
in filtering a given signal.
On the other hand, strong correlations
increase the volume in the parameter space to which an event can
be associated at a given confidence level.
It seems to be in general true that a given set of parameters do not satisfy
the twin properties of
having small covariances and reducing the effective dimension of the manifold
for the purpose of filtering.
We elaborate on this point below.
Given a region in a parameter space, it is useful to know the
proper volume (as defined by the metric) of the manifold corresponding
to the said region. In choosing a discrete set of filters for the detection
problem one has
to decide upon the maximum allowable drop in the correlation due to the finite
spacing. Once
this is fixed, the number of filters can be determined from the total volume of
the manifold.
For the detection problem it is beneficial to have a small volume whereas if
the waveform is parameterized in such a way such that the manifold
corresponding to it covers
a large volume, then one can determine the parameters to a greater accuracy. As
a simple
example let us consider a two-dimensional toy model: $\bbox{\lambda} =
\{\lambda_1,\lambda_2\}$.
We compare different signal manifolds each corresponding to a different
parameterizations of
the waveform.
We assume that the covariance matrix and its inverse, the Fisher information
matrix, to be:
\begin{equation}
\label{volumes}
C_{\mu\nu} = \left( \begin{tabular}{cc}
$\sigma_{11}$&$\sigma_{12}$\\
$\sigma_{12}$&$\sigma_{22}$
\end{tabular}
\right)
\mbox{\ \ \ and\ \ \ }
\Gamma_{\mu\nu} =
\left( \begin{tabular}{cc}
$\gamma_{11}$&$\gamma_{12}$\\
$\gamma_{12}$&$\gamma_{22}$
\end{tabular}
\right) = \frac{1}{\left(\sigma_{11}\sigma_{22}
-\sigma_{12}^2\right)}
\left(
\begin{tabular}{cc}
$\sigma_{22}$&$-\sigma_{12}$\\
$-\sigma_{12}$&$\sigma_{11}$
\end{tabular}
\right).
\end{equation}
The volume of the manifold corresponding to a region $\cal K$ of the parameter
space is given as,
\begin{equation}
\label{vol2}
V_{\cal K} = \int_{\cal K}\gamma_{11}\gamma_{22}
\left[1 -
\frac{\gamma_{12}^2}{\gamma_{11}\gamma_{22}}\right]d\lambda_1d\lambda_2
= \int_{\cal K}\gamma_{11}\gamma_{22}\left[1 -
\epsilon\right]d\lambda_1d\lambda_2,
\end{equation}
where $\epsilon=\gamma_{12}^2/(\gamma_{11} \gamma_{22})$ is the correlation
coefficient.
It can be clearly seen that if the correlation coefficient is small,
keeping the variances in the parameters, fixed then the volume of the manifold
is maximal.
Since the parameters $\tau_0$ and $\tau_1$ are highly anticorrelated the proper
volume corresponding
to the region reduces to zero showing that the effective dimensionality of the
manifold is less.
Though, in principle, the variances and covariances are independent
of the chirp time, in reality there arises an indirect dependence
since one terminates a template at a frequency $f=1/(6^{3/2}\pi M)$
(where $M$ is the total mass of the binary) corresponding to the plunge
radius at $a=6M.$ Therefore, larger mass binaries are tracked
over a smaller bandwidth so much so that there is less frequency
band to distinguish between two chirps of large, but different, total mass.
Consequently, at a given SNR the error in the estimation of chirp times is
larger for greater mass binaries. This is reflected by the fact that
the integrals in eq. (\ref {gamma2})
are somewhat sensitive to the value of the upper cutoff.
(This also explains why the errors in the estimation of the chirp
and reduced masses are larger for greater mass binaries \cite{FC93,CF94}.)
In the following we assume that the noise
power spectral density is that corresponding to the initial LIGO
for which a fit has been provided by Finn and
Chernoff \cite{FC93}.
For an SNR of 10 the matrix $D$ is given by
\begin {equation}
{D_{\mu\nu}} = \left ( \begin{array} {rrrrr}
1.0 & 0 & 0 & 0 \\
& 8.37 & 0.999 & -0.999 \\
& & 3.16 & -0.998 \\
& & & 8.4 \\
\end {array} \right ),
\end {equation}
for the Newtonian signal, and by
\begin {equation}
{D_{\mu\nu}} = \left ( \begin{array} {rrrrr}
1.0 & 0 & 0 & 0 & 0 \\
& 20.4 & 0.997 & -0.972 & 0.911 \\
& & 6.7 & -0.954 & 0.881 \\
& & & 45.1 & -0.982 \\
& & & & 25.98 \\
\end {array} \right ),
\end {equation}
for the first post-Newtonian corrected signal.
While computing varianaces and covariances the
integrals in equation (\ref {gamma3}) are evaluated by
chosing a finite upper limit of 1 kHz.
In the above matrices the entries are arranged in the order
$\{{\cal A}, t_a, \Phi, \tau_0\}$ in the Newtonian case,
$\{{\cal A}, t_a, \Phi, \tau_0, \tau_1\}$ in the post-Newtonian case,
off diagonal elements are dimensionless correlation coefficients and,
where appropriate, diagonal elements are in ms.
The values quoted in the case of the Newtonian waveform are consistent
with those obtained using a different set of parameters by
Finn and Chernoff \cite {FC93}.
In Fig.~\ref {fdependence} we have plotted $\sigma$'s, at an SNR of 10,
as a function of the upper frequency cutoff $f_c$ for Newtonian and
post-Newtonian chirp times and the instant of coalescence $t_C.$
We see that $\sigma$ is larger for higher mass binaries but this is because
we have fixed the SNR. However, if we consider binaries of different
total masses, all located at the
same distance, then a more massive binary produces a stronger SNR
so that in reality it may be possible to determine its parameters more
accurately than that of a lighter binary.
In Fig.~\ref{truedependence} we have plotted
$\sigma$'s for binaries all located at the same distance
as a function of total mass. We fix one of the masses at a value of
$1.4~M_\odot$
and vary the other from $1.4~M_\odot$ to $10~M_\odot$.
In computing the $\sigma$'s plotted in this
figure we have terminated the waveform at the plunge orbit and
normalised the SNR of a $10 M_\odot$-$1.4 M_\odot$ binary system
to 10. As a function of $M$ the
uncertainties in $\tau_0$ and $\tau_1$ initially falloff since the increase in
the SNR for
larger mass binaries more than compensates for the drop in the upper frequency
cutoff. However, for $M$ larger than a certain $M_0$ the increase in SNR is not
good enough
to compensate for the drop in $f_c$ so much so that the uncertainties in
$\tau_0$ and $\tau_1$
increase beyond $M_0$. The parameter $t_a$, however, falls off monotonically.
\subsection {Monte Carlo estimation of parameters}
In this Section we present the first in a series of efforts to compute
the covariance matrix of errors through numerical simulations for
a coalescing binary waveform
at various post-Newtonian orders. Analytical computation of the
covariance matrix, as in the previous section, gives us an
idea of the covariances and variances but, as we shall see in this
Section, at low SNR's it grossly underestimates the errors.
Quite apart from the fact that the assumptions made in deriving
the covariance matrix might be invalid at low SNR's, in a realistic detection
and data analysis, other problems, such as discreteness of the
lattice of templates, finite sampling of the data, etc., do occur.
It therefore seems necessary to check the analytical calculations
using numerical simulations to gain further insight into the
accuracy at which physical parameters can be measured.
This Section is divided into several parts: In the first part we
highlight different aspects of the simulation, in the second part we briefly
discuss the choice of templates for the simulation, in the third we elaborate
on the Monte Carlo method that we have adopted to carry out our simulations
and in the fourth we discuss problems that arise in a numerical simulation.
The results of our study are discussed in the next Section.
\subsubsection {Parameters of the simulation}
Let $s(t)$ be a signal of strength
$\cal A$ characterised by a set of parameters $\hat{\bbox{\lambda}}$
\begin {equation}
s(t; \hat{\bbox{\lambda}}) = {\cal A} h(t;\hat{{\bbox {\lambda}}}),\ \
\left<h,h\right >=1.
\end {equation}
In data analysis problems one considers a discrete version
$\{s^k | k=0,\ldots,N-1\}$ of the waveform $s(t)$ sampled at
uniform intervals in $t:$
\begin {equation}
s^k \equiv s(k\Delta); \ \ \ k=0,\ldots,N-1,
\end {equation}
where $\Delta$ denotes the constant interval between consecutive samples
and $N$ is the total number of samples. The sampled output $x^k$ of the
detector consists of the samples of the noise plus the signal:
\begin {equation}
x^k = n^k + s^k.
\end {equation}
The {\it sampling rate,} $f_s=\Delta^{-1}$
(also referred to as the {\it sampling frequency})
is the number of samples per unit time interval. In a data analysis
problem the sampling frequency is determined by the signal bandwidth.
If $B$ is the signal bandwidth, i.e., if the Fourier transform of the
signal is only nonzero over a certain interval $B,$ then it is
sufficient to sample at a rate $f_s=2B.$ In our case there is a lower
limit in the frequency response of the detector since the detector
noise gets very large
below a seismic cutoff at about 10--40 Hz. As mentioned in the last Section
there is also an upper limit in frequency up to which a chirp signal is
tracked since one does not accurately know the waveform beyond
the last stable orbit of the binary. This
corresponds to gravitational wave frequency $f_c=1/(6^{3/2}\pi M).$
For a neutron star-neutron star (ns-ns) binary $f_c \sim 1525$~Hz while
for a ns-black hole (of 10 $M_\odot$) (ns-bh) binary $f_c \sim 375$~Hz.
Due to constraints arising out of limited computational power, we terminate
waveforms at $750$ Hz even when $f_c$ is larger than 1000 Hz. Such a shutoff
is not expected to cause any spurious results since, even in the case of
least massive binaries of ns-ns, which we consider in this study,
more than 99 \% of the `energy' is extracted by the time the
signal reaches 750 Hz.
We have carried out simulations with two types of upper cutoff:
\begin {enumerate}
\item one in which all templates, irrespective of their total mass,
are shutoff beyond 750 Hz.,
\item a second in which the upper frequency cutoff is chosen to be $750$ Hz
or $f_c,$ whichever is lower.
\end {enumerate}
Consistent with these cutoffs the sampling rate is always taken to be $2$ kHz.
(We have carried out simulations with higher sampling
rates and found no particular advantage in doing so
nor did we find appreciable changes in our results.)
In all our simulations,
as in the previous section, we take the detector noise power spectral density
$S$ to be that corresponding to initial LIGO
\cite {FC93}. For the purpose of simulations we need to generate noise
corresponding to such a power spectrum. This is achieved by the following
three steps:
\begin {enumerate}
\item generate Gaussian white noise $n'^k$ with zero mean and unit variance,
$$\overline {n'^k}=0,\ \ \overline{n'^k n'^l} = \delta^{kl}$$
where an overbar denotes average over an ensemble,
\item compute its Fourier transform
$$\tilde n'^k\equiv \frac{1}{\sqrt{N}}\sum_{l=0}^{N-1} n'^l \exp (2\pi i k l /
N)$$
and
\item multiply the Fourier components by the square root of the
power spectral density,
$$\tilde n^k = \sqrt {S^k} \tilde n'^k.$$
\end {enumerate}
The resultant random process has the requisite power spectrum.
In the above, the second step can be eliminated since the Fourier transform
of a Gaussian random process is again a Gaussian, but with a different
variance. In other words we generate the noise directly in the Fourier domain.
The simulated detector output, in the presence of a signal
$s^k,$ in the Fourier domain is given by
\begin {equation}
\tilde x^k = \tilde n^k + \tilde s^k
\end {equation}
where $\tilde s^k$ is the discrete Fourier transform of the signal.
\subsubsection {Choice of templates}
To filter a Newtonian signal we employ the set of parameters
$\{t_a, \Phi, \tau_0\}$ and to filter a post-Newtonian signal we
employ the set $\{t_a, \Phi, \tau_0, \tau_1\}.$ Templates need not
explicitly be constructed for the time of arrival since computation
of the scalar product in the Fourier domain (and the availability of
fast Fourier transform (FFT) algorithms) takes care of the time of arrival
in essentially one computation ($N\log_2 N$ operations as opposed to
$N^2$ operations, where $N$ is the number of data points). Moreover,
there exists a two-dimensional basis for the phase parameter which
allows the computation of the best correlation with the aid of just
two filters. Consequently, the parameter space is
essentially one-dimensional in the case of Newtonian signals
and two-dimensional in the case of post-Newtonian signals. (However,
as shown in Section \ref {2pnsec}
it is to be noted that for the purpose of detection the effective
dimensionality of the parameter space, even with the inclusion
of second post-Newtonian corrections, is only one-dimensional.) We adopt the
method described in Sathyaprakash and Dhurandhar \cite {SD91}
to determine the templates needed for chirp times. As described
in \cite {SD91,Sat94} filters uniformly spaced in $\tau_0$ and $\tau_1$
covers the parameters space efficiently.
\subsubsection {Monte Carlo method} \label {mcmethod}
In order to compute variances and covariances numerically,
we employ the Monte Carlo method. The basic idea here is to mimic
detection and estimation on a computer by performing
a very large number of simulations so as to minimize the uncertainties
induced by noise fluctuations.
In our simulations we generate a number of detector outputs $\{x^k\}$ each
corresponding to a definite signal $s^k(\hat{{\bbox {\lambda}}})$ of a certain
strength,
but corresponding to different realisations
of the random process $\{n^k\}.$ Computation of the covariance matrix involves
filtering each of these detector outputs through an a priori chosen
set (or lattice) of templates $\{q(t; {{\rm _t}\bbox{\lambda}_k}) |
k=1,\ldots,n_f\},$
where $n_f$ denotes the number of templates. The templates of the lattice
each has a distinct set of values of the {\em test} parameters
${{\rm _t}\bbox{\lambda}_k}$ and together
they span a sufficiently large volume in the parameter space.
The simulated detector output is correlated with each member of
the lattice to obtain the corresponding filtered output $C ({{\rm
_t}\bbox{\lambda}}_k):$
\begin {equation}
C({{\rm _t}\bbox{\lambda}}_k) = \left < x, q ({{\rm _t}\bbox{\lambda}}_k)
\right >.
\end {equation}
For a given realisation of noise a particular template obtains the
largest correlation and its parameters are the {\it measured} values
${{\rm _m}{\bbox{\lambda}}}$ of the signal parameters. Thus, the measured
values of the
parameters are defined by
\begin {equation}
\max_k C({{\rm _t}\bbox{\lambda}}_k) = C({{\rm _m}{\bbox{\lambda}}}).
\end {equation}
The measured values, being specific to a particular realisation of noise,
are random variables. Their average provides an
estimation ${{\rm _e}{\bbox{\lambda}}}$ of the true parameter values and their
variance
is a measure of the error $\sigma_{\bbox {\lambda}}$ in the estimation:
\begin {equation}
{{\rm _e}{\bbox {\lambda}}} = \overline{{\rm _m}{\bbox {\lambda}}},\ \ \ \ \
\sigma_{\bbox {\lambda}}^2 = \overline{ \left ( { {{\rm _m}{\bbox {\lambda}}} }
-
\overline{ {{\rm _m}{\bbox {\lambda}}}} \right )^2}, \ \ \ \
D^{\mu\nu} = \overline{ {{\rm _m}\lambda}^\mu \ {{\rm _m}\lambda}^\nu
\over \sigma_\mu \sigma_\nu}, \ \ (\mu\ne \nu),
\label {mccovar1}
\end {equation}
where $D_{\mu\nu}$ are the correlation coefficients between
parameters $\lambda_\mu$ and $\lambda_\nu.$
In order to accurately determine $\sigma_{\bbox {\lambda}}$ a large number of
simulations
would be needed. If the measured values ${{\rm _m}{\bbox {\lambda}}}$ obey
Gaussian
statistics then after $N_{\rm t}$ trials the variance is determined
to a relative accuracy $1/\sqrt N_{\rm t}$ and estimated values can differ from
their
true values by $\sigma_{\bbox{\lambda}}/\sqrt N_{\rm t}.$ We have performed in
excess of 5000
trials, for each input signal, and thus our results are accurate to better
than 1 part in 70. Even more crucial than the number of simulations
is the number of templates used and their range in the parameter space.
We discuss these and other related issues next.
The actual templates chosen, say for the parameter $\tau_0,$
in a given `experiment' depend on the
true parameters of the signal, the number of noise realisations employed
and the expected value of the error. Let us suppose
we have a first guess of the error in $\tau_0$, say $\sigma_{\tau_0}$.
Then, we choose 51 uniformly spaced filters around $\hat {\tau}_0$ (where
$\hat{\tau_0}$ is the signal chirp time) such that:
\begin{equation}
{{\rm _t}\tau_0} \in \left [ \hat{\tau}_0 - 5 \sigma_{\tau_0},
\hat{\tau}_0 + 5 \sigma_{\tau_0} \right ].
\end{equation}
This implies that we are covering a 5$\sigma$ width in $\tau_0$ at
a resolution of
$\sigma_{\tau}/5.$ The probability, that a template between 4$\sigma$ and
5$\sigma$
from the true signal `clicks', being $\sim 6\times 10^{-5},$
we are on safe grounds since, in a given simulation, we consider no more
than 5000 trials.
(In comparison, the probability that a template between 3$\sigma$ and 4$\sigma$
clicks is $2.2\times 10^{-3}$ corresponding to an expected 13 events
in 5000 trials.) For a post-Newtonian signal, which in effect
needs to be spanned by a two-dimensional lattice of filters, the
above choice of templates implies a requirement of $2601 \times 2$
filters in all. Here a factor of 2 arises because for each filter
in the $\tau_0$--$\tau_1$ space we will need two templates corresponding
to the two independent values of the phase $\Phi$: $0$ and $\pi/2.$
In the case of a Newtonian signal, the lattice being one-dimensional,
one can afford a much higher resolution and range. Even with the aid of
a mere 201 templates we can probe at a $\sigma/10$ resolution
with a $10\sigma$ range.
We start off a simulation with the pretension that there is
no knowledge of what the $\sigma_{{\bbox {\lambda}}}$'s are. Thus, we choose
as our first trial a very large $\sigma_{\bbox {\lambda}}$ and lay the lattice
of templates. With this lattice we perform a test run of 400 trials and
examine the distribution of the measured values. If the distribution
is not, as expected, a Gaussian then we alter $\sigma_{\bbox {\lambda}}$: we
decrease it if the distribution is too narrow and increase it if the
distribution is too wide and does not show the expected falloff.
In particular, we
make sure that the templates at the boundary of the chosen range do
not click even once and the skewness of the distribution
is negligible. When for a certain $\sigma_{\bbox {\lambda}}$ a
rough Gaussian distribution is observed then we carry out a
simulation with a larger number of trials (typically 5000).
We subject the measured values in this larger simulation to
the very same tests described above. We only consider for further
analysis such simulations which `pass' the above tests
and determine the estimates, variances and covariances of the
parameters using the measured values, with the aid of equation
(\ref {mccovar1}).
\subsubsection {Numerical errors and remedies}
There are several sources of numerical errors that tend to
bias the results of a simulation unless proper care is exercised
to rectify them. In this Section we point out the most
important ones and show how they can be taken care of.
Due to memory restrictions, the present version of our codes work with
single precision except the FFT, which is implemented in double precision.
In future implementations we plan to carry out all computations in
double precision. This will possibly reduce some of the numerical noise
that occurs, especially at high SNRs, in the present simulations.
\begin {enumerate}
\item {\it Orthonormality of filters:} For the sake of simplicity
it is essential that the filters are
normalised in the sense that their scalar product is
equal to unity: $ \left < q, q \right > = 1. $
A waveform is normalized numerically using the discrete version of the scalar
product:
\begin {equation}
{\cal N} = {1 \over \sum_{k=0}^{N-1} S^{-1}_k \left | \tilde q^k \right |^2 }
\end {equation}
As mentioned earlier we use
a two-dimensional basis of filters for the phase parameter. Choosing
the two filters to be orthogonal to each other makes the maximisation
over phase easier. However, here care must be exercised.
Two filters $q(t; t_a, \tau_0, \tau_1, \Phi=0)$ and
$q(t; t_a, \tau_0, \tau_1, \Phi=\pi/2)$
are apparently orthogonal to each other. The numerically computed `angle'
between the two filters, chosen in this manner, often turns out to
be greater than $\sim 10^{-2}$ radians. Consequently, one obtains
erroneous correlations. In order to circumvent this problem we first
generate two filters that are roughly orthogonal to each other,
as above, and then use the Gram-Schmidt method
to orthogonalise the two vectors.
If an explicit numerical orthogonalisation such as this is {\em not}
implemented
then the measured values of the various parameters show spurious
oscillations in their distribution and the estimated values of the
parameters tend to get biased.
\item {\it Correlation function:}
The scalar product of two normalised templates
$q(t; {{\rm _t}{\bbox {\lambda}}})$ and $q(t; {{\rm _t}{\bbox {\lambda}}}')$
is given by
\begin {equation}
C({{\rm _t}\lambda}_\mu, {{\rm _t}\lambda}_\nu') =
\left < q(t; {{\rm _t}\lambda}_\mu), q(t; {{\rm _t}\lambda}_\nu') \right >,
\ \ \left < q(t; {{\rm _t}{\bbox {\lambda}}}), q(t; {{\rm _t}{\bbox
{\lambda}}}) \right > =
\ \ \left < q(t; {{\rm _t}{\bbox {\lambda}}}'), q(t; {{\rm _t}{\bbox
{\lambda}}}') \right > = 1,
\end {equation}
where we have indicated the dependence of the scalar product
on the various parameters by explicitly writing down the parameter subscripts.
Let us fix the parameters of one of the templates, say ${{\rm _t}{\bbox
{\lambda}}},$
and vary the parameters of the other template. Of particular interest is the
behaviour of $C$ maximised over all but one of the parameters,
say $\lambda_{\nu_0}$:
\def\ooo #1 #2{\vphantom{S}^{\raise-0.5pt\hbox{$#1$}}_{\raise0.5pt
\hbox{$#2$}}}
\begin {equation}
C_{\rm max}({ {\rm _t}\lambda}_\mu, {{\rm _t} \lambda}_{\nu_0}') =
\max_{\ooo {\scriptstyle \ {{\rm _t}\lambda}_\nu'}
{\scriptstyle\nu \ne \nu_0}}
C ({{\rm _t}\lambda}_\mu, {{\rm _t} \lambda}_\nu').
\end {equation}
$C_{\rm max}$ is expected to drop monotonically as
$|\lambda_{\nu_0} - \lambda_{\nu_0}'|$ increases. However, we have
observed departures from such a behaviour possibly arising out of
numerical noise. Such a behaviour causes
bias in the estimation of parameters, and consequently in the determination of
their covariances, especially at high SNRs. We have found no remedy to
this problem and some of our results at high SNRs may have biases
introduced by this effect. (Sampling the templates at a higher rate
did not help in curing this problem.)
\item {\it Grid effects:}
The parameters of a signal chosen for the purpose of
simulation and detection can in principle be anything and
in particular it need not correspond to any of the templates
of the lattice. However, in practice we find that whenever the signal
parameters do not correspond to a member of the lattice then the resultant
simulation has a bimodal distribution of the measured values. This is,
of course, expected since a signal not on the grid is picked up by
two nearest templates along each direction in the parameter space.
Sometimes we do find that the peaks corresponding to the bimodal distribution
does not belong to the nearest neighbour filters but slightly away.
This is related to the fact that the correlation function maximised over
the time of arrival and the phase of the signal falls off much too
slowly along the $\tau_0$--$\tau_1$ direction and small deviation
from a monotonic fall can cause biases. (Such biases would be
present in the case when a signal corresponds to one of the grid points
though the magnitude of the effect would be lower.) In order to
avoid this problem, and the consequent shifts in the estimation of
parameters and errors in the determination of variances and covariances,
we always choose the parameters of the signal to be that corresponding
to a template.
\item {\it Upper frequency cutoff and its effect on parameter estimation}
The Fisher information matrix computed using the stationary phase
approximation
in Section \ref{sm} does
not include the effect of truncating the waveform at $a = 6M$ --- the plunge
cutoff.
As mentioned before, we
have carried out simulations for both with, and without, incorporating the
upper cutoff.
As the covariance matrix incorporating the upper cutoff is not available we
have been able to compare the Monte Carlo results with the covariance matrix
only
for the latter case, where the cutoff is held fixed at $750$~Hz. If we
incorporate the
upper cutoff into the Monte Carlo simulations the errors in the parameters are
reduced drastically.
The effect of the upper cutoff is expected to be more important for the higher
mass binaries such
as the ones we have considered. The ambiguity function, in this case, no longer
remains independent
of the point on the manifold. In other words, the correlation between two
chirps depends not only on the difference between the parameters of the
signals, but also
on the absolute values of the parameters. The correlation surface also ceases
to be
symmetric {\em i.e.} the correlation between two chirps also depends on the
sign of
$\delta\lambda$, where, $\delta\lambda$ is the difference in the values of the
parameters. As the computational power
required for carrying out simulations for lower mass binaries is not available
to us
the simulations have been restricted to ns-bh star binaries, where the effect
of the upper cutoff is important.
\item {\it Boundary effects}
For the purpose of simulations a grid of filters has to be set up `around' the
signal.
The grid must be large enough so that the estimated parameters do not
overshoot the boundary
of the grid. This causes a problem as every value in the \{$\tau_0,\tau_1$\}
plane
does not lead to a meaningful value for the masses of the binary system. This
does not
however prevent us to construct a waveform with such a value for
\{$\tau_0,\tau_1$\} even
though the signal in general does not correspond to any `real' binary system.
This is
valid, and even necessary, if we are to compare the numerical results with the
covariance matrix.
\item {\it Incorporating the cutoff in the presence of boundary effects}
If we wish to incorporate the effects of the upper cutoff in simulations then
we run into a
serious problem, as we would have to know the total mass of the binary in order
to compute the
upper cutoff. For an arbitrary \{$\tau_0,\tau_1$\} we can end up with negative
and even complex
values of the total mass and hence the upper cutoff at $a = 6M$ is not
meaningful.
Thus, we cannot even construct a waveform for an arbitrary combination of
\{$\tau_0,\tau_1$\}. Therefore, in such cases,
we restrict ourselves to simulations where the grid lies entirely
within valid limits for \{$\tau_0,\tau_1$\}.
\end {enumerate}
\subsection {Results and Discussion}
Our primary objective is to
measure the variances and covariances following the method described in
Sec. \ref{mcmethod} and study their departure from that predicted by
analytical means (cf. Sec. \ref {anacovar}).
We have carried out simulations for several values of the masses of
the binary and in each case the signal strength (which is a measure
of the SNR) is varied in the range 10--40. However,
since the variances and covariances are independent of the absolute
values of the parameters, for the parameter set that we employ,
results are only quoted corresponding to a typical binary system.
(See Sec. \ref {anacovar} for a discussion of the covariance matrix.)
Similar results are obtained in other cases too. We use two sets of
parameters to
describe our results. Monte Carlo simulations allow us to directly measure
the amplitude, the time of arrival, the phase at the time of arrival
and the chirp time(s). This is the set $\{{\cal A}, t_a, \Phi, \tau_0,
\tau_1\}.$
As we shall see below, the instant of coalescence can be measured much more
accurately than the time of arrival.
As a consequence of this, the direction to the source can be determined
at a {\em much greater accuracy} by employing $t_C$ as a
parameter instead of $t_a.$
Thus, we also quote estimates and errors for the parameter $t_C.$
Since the error in the estimation of the phase is quite large, even at
high SNRs, we ignore it in our discussions.
We first deal with the
Newtonian signal and highlight different aspects of the simulation
and discuss the results at length. We then consider the
first post-Newtonian corrected signal.
\subsubsection {Newtonian signal}
In the case of Newtonian signals the parameter space is
effectively one-dimensional and, as mentioned earlier, in this case
the lattice of templates covers a $10$--$\sigma$ range of the parameters
at a resolution of $\sigma/10$ centred around the true parameters
of the signal.
In Fig. \ref {nmcfig} we have shown
the error $\sigma_{\bbox {\lambda}}$ in the estimation of parameters
$t_a,$ $\tau_0$ and $t_C,$
as a function of SNR, deduced using the covariance matrix as
solid lines and computed using Monte Carlo simulations as dotted lines.
The curve corresponding to the covariance matrix is obtained
using an upper frequency cutoff $f_c= 750$~Hz consistent with that
used in our simulations. The errorbars in the estimation of
$\sigma_{{\bbox {\lambda}}}$'s are obtained using 4 simulations, each
with 4000 trials. At low SNR's $\sigma_{{\bbox {\lambda}}}$'s have a larger
uncertainty, as expected, and for $\rho > 30$ this uncertainty is
negligible, and sometimes smaller than the thickness of the curves,
except in the case of $\sigma_{t_C}$ (see below, for a possible
explanation).
At low SNRs (10--15) there is a large
departure of the various $\sigma_{\bbox {\lambda}}$'s from that inferred using
the covariance matrix. At an SNR $\sim 17$ the two curves
merge (except in the case of $\sigma_{t_C}$)
indicating the validity of the covariance matrix results for
this and higher SNR's. Interestingly,
the agreement between Monte Carlo simulation results and those
obtained using the covariance matrix is reached roughly at the same SNR
irrespective of the parameter in question.
We note that
in spite of the fact that the time of arrival and the chirp time have
large errors, the instant coalescence can be estimated very
accurately---an order of magnitude better than either.
What is puzzling, however, is that, in the case of $t_C,$ the Monte Carlo
curve drops below
the covariance matrix curve above an SNR of 15 and the two curves do
not seem to converge to one another
even at very high SNR's. Coincident with the crossover of the two
curves, the error in the estimation of $\sigma_{t_C}$ increases,
contrary to what happens for the other parameters,
signalling that there is a large fluctuation in the estimation of
$\sigma_{t_C}.$ This behaviour, we guess, is an artifact of the low
value of the sampling rate. Of course, our sampling rate is sufficiently
high to respect the sampling theorem. However, since $t_C$
is determined to an accuracy an order of
magnitude better than either $t_a$ or $\tau_0,$ a much higher
resolution in template-spacing would be needed for
determining the error in the instant of coalescence
than that used for estimating the errors in the
time of arrival or the chirp time(s).
Testing this claim, unfortunately, is beyond the computer resources at
our disposal since we would need a sampling rate of about 10 kHz with
a filter-spacing $10^{-4}$~s. We hope to be able to resolve this issue
in course of time.
Nevertheless, the fact that the error in the estimation of $\sigma_{t_C}$
first decreases with SNR and increases only after the two curves
crossover, hints at the above possibility as a cause for this anomalous
behaviour. This effect is also observed in the case of a post-Newtonian
signal.
In Table \ref{nmctab}
we have given the actual signal parameters $\hat {{\bbox {\lambda}}},$
estimated
values of the parameters ${{\rm _e}{\bbox {\lambda}}}$ (cf. equation
(\ref{mccovar1}))
and the corresponding errors in their estimation $\sigma_{\bbox {\lambda}},$
for
several values of the SNR. Errors inferred from the covariance matrix
can be read off from Fig.~\ref{nmcfig}.
The estimated values are different from the true values, some of them being
overestimated
and some others being underestimated.
However, the deviations are often larger than what we expect.
In a simulation that uses $N_{\rm t}$ trials the estimated parameters
${{\rm _e}{\bbox {\lambda}}},$ assuming a Gaussian distribution for the
measured
parameters ${{\rm _m}{\bbox {\lambda}}},$ can be different from the true values
by
$\sigma_{\bbox {\lambda}}/\sqrt{N_{\rm t}}.$ (In contrast, the measured values
${{\rm _m}{\bbox {\lambda}}}$ can differ from their true values by
$\sigma_{\bbox {\lambda}}$
or more.) However, we often obtain a slightly larger deviation,
\begin {equation}
2 {\sigma_{\bbox {\lambda}}\over \sqrt {N_{\rm t}}} \la \left |{{\rm_e}{\bbox
{\lambda}}} - \hat{{\bbox
{\lambda}}}\right | \la 3 {\sigma_{\bbox {\lambda}}\over \sqrt {N_{\rm t}}},
\end {equation}
and we are unable to resolve this discrepancy. A more concrete test
for the simulations is the histogram $n({{\rm _m}{\bbox {\lambda}}})$ of the
measured parameters, namely the frequency at which a given parameter
clicks in a simulation. This is shown plotted in Fig.~\ref {nhist}
for an SNR of 10.
The skewness
of the measured value is less than $10^{-2}.$ These results lend further
support to the Monte Carlo simulations. There are visible asymmetries
in the distributions of $\tau_0$ and $t_a$ and the asymmetries in the
two cases are of opposite sense. This can, of course, be understood from the
fact that $t_a$ and $\tau_0$ have a negative correlation coefficient.
The histogram of $t_C,$ even at an SNR of 10, has very few non-zero bins.
This of course reflects the fact that it is determined very accurately.
We are unable to resolve the central peak in $n(t_C)$ since,
as mentioned earlier, the sampling rate and resolution in
$\tau_0$ are not good enough to do so.
\subsubsection {Post-Newtonian signal}
As opposed to the Newtonian case
here we have essentially a two-dimensional lattice of filters corresponding
to $\tau_0$ and $\tau_1.$ For the purpose estimating variances and
covariances we lay a mesh consisting of $2601 \times 2$ uniformly spaced
filters around the true parameters of the signal.
As pointed out in Sec. \ref {detect} not all filters in the mesh, unlike
in the Newtonian case, would correspond to the waveform from a realistic
binary but that does not preclude their use in the Monte Carlo simulations.
We shall see that the results of our simulations lend
further support to the claim that for the purpose of detection,
the parameter space need only be one-dimensional \cite {Sat94}.
The results obtained for the first post-Newtonian
signal are qualitatively similar to that of a Newtonian signal
and we refer the reader, where appropriate, to the Newtonian case for
a more complete discussion.
In Fig. \ref {pnmcfig} we have shown
the error in the estimation of parameters $\tau_0,$ $\tau_1,$
$t_C$ and $t_a,$ clockwise from top left, respectively,
as a function of SNR.
The solid and dotted curves are as in Fig. \ref{nmcfig}.
Here again the upper frequency cutoff is taken to be 750 kHz.
Just as in the case of a Newtonian signal here too the results obtained from
Monte Carlo simulation are much higher than those obtained by employing
the covariance matrix. At an SNR of 10, which is the expected value
for a majority of events that initial LIGO and VIRGO interferometers
will observe, the Monte Carlo values are more than thrice as much as their
corresponding covariance matrix values and at an SNR of 15 the errors
are roughly twice that expected from the covariance matrix.
In absolute terms, however, the errors are still quite small compared to
the actual parameter values: for a ns-ns binary, at an SNR of 10,
\begin {equation}
{\sigma_{\tau_0}\over \tau_0} \sim 2.4 \%, \ \
{\sigma_{\tau_1}\over \tau_1} \sim 9.4 \%.
\end {equation}
At an SNR of 10 the time of arrival can be measured to an accuracy of
72 ms in contrast to a value of 20 ms expected from the covariance matrix.
As is well known, with the inclusion of the post-Newtonian terms error
in the estimation of the time of arrival and Newtonian chirp time
increases by about a factor of 2 and 3, respectively \cite{FC93,CF94}.
As in the Newtonian case here again we see that the Monte Carlo curves
approach the corresponding covariance matrix curves at a high SNR
the only difference being that the agreement is reached at a
higher SNR $\sim 25.$ For SNRs larger than this the
two curves are in perfect agreement with each other. As mentioned
earlier, $\sigma_{t_C}$ shows an anomalous behaviour possibly
arising out of insufficient resolution in the time of arrival and
the chirp times.
In Table \ref {pntab} we have listed the true parameters $\hat{{\bbox
{\lambda}}},$
the estimated values ${{\rm _e}{\bbox {\lambda}}}$ and the Monte Carlo errors
$\sigma_{\bbox {\lambda}}$ for different SNRs.
As in the Newtonian case here too the estimated values show a larger departure,
than expected,
from the true values. Histograms of the various measured parameters
including $t_C$ are shown in Fig. \ref {pnhist} for a signal strength of 10.
The skewness is below its standard deviation
of $\sqrt {15/N_{\rm t}}$ \cite {NUM}
indicating the Gaussian nature of the various distributions. Even in the case
of a post-Newtonian signal $\sigma_{t_C}$ is so small that we only have three
non-zero bins in $n(t_C).$
We now turn attention to other, more general, issues arising out of the
simulations.
In Sec. \ref {2pnsec} we have argued, on the basis of the behaviour
of the noise-free correlation function, that the effective
dimensionality of the parameter space for the purpose of detection,
even in the case of a post-Newtonian signal, is only one-dimensional.
The results of our Monte Carlo simulation unambiguously show that this
is indeed true even in the presence of noise.
We investigated the two-dimensional histogram, that gives the number of
occurrances of
different templates in the lattice, in a particular simulation.
The templates that `click' are all aligned along the line
$\tau_0 + \tau_1=$~const. In a total of 5000 realisations corresponding
to this simulation there is only one instance when a filter outside
this region clicks giving a probability of less than $10^{-3}$ for
a template outside this region to give a maximum. Consequently, it is
only necessary to choose a single filter along this curve.
The distribution of the maximum correlation
$ C_{\rm max}(\hat {{\bbox{\lambda}}}, { {\rm _t}{\bbox {\lambda}}})$ obtained
from
different noise realisations needs a special mention since it
has an inherent bias.
In Fig. \ref {snrhist} we have shown the distribution of the maximum
correlation taken from one of our simulations corresponding to an
SNR of 10.
Notice a slight shift of the distribution towards a higher
value and this cannot be accommodated within the expected fluctuation
in the mean. The measured value
of the standard deviation $\sigma_{\cal A}$ is $0.95.$ Since the number
of simulations is 5000 we expect that the signal strength should differ
from the true value of 10 by no more than $\sigma_{\cal A}/\sqrt {5000}=0.014.$
However, the mean
value is 10.26 giving a deviation of $0.26$ which is about 20 times larger
than that expected.
This occurs at all SNRs and for both Newtonian and post-Newtonian signals.
This of course does not mean there is bug in the way
we are computing the maximum correlation.
In the process of maximisation, values greater
than the signal strength are favoured and consequently the mean of the
maximum correlation shows a shift towards a higher value.
This suggests that that the maximum of the correlation is a biased estimator of
the
signal strength.
We find, consistent with the covariance matrix calculation, that
the amplitude parameter is uncorrelated with the rest of the parameters;
crosscorrealtion coefficients $D_0\mu,$ $\mu\ne 0,$ (cf. eq. \ref {varcorr})
inferred from our Monte Carlo simulations are less than $\sim 10^{-6}.$
Finally, it is of interest to note how the phase parameter $\Phi$
is correlated with the time of arrival. A plot of ${{\rm _m}\Phi}$
versus ${{\rm _m}t}_a$ is shown in Fig. \ref{taphi}.
We find that the measured values of the time of arrival and the phase are
such that $2\pi f_{0} \ {{\rm _m}t}_a = {{\rm _m}\Phi},$ where $f_{0}$
has a value of approximately $51$~Hz. When the
time of arrival shifts by more than a cycle of the signal the phase jumps
by a factor $2 \pi$ leading to the points seen in the top-left
and bottom-right corner of the figure. This makes the estimation
of the phase and the error in its estimation pretty involved.
\subsubsection {Incorporating the effects of upper cutoff}
As mentioned before,
incorporating an upper cutoff at the onset of the plunge has a drastic effect
on
the estimation of parameters. The incorporation of the upper cutoff is
implemented
by stopping the waveform when the instantaneous frequency reaches the frequency
associated with the onset of the plunge or
$750$~Hz, whichever is lower. However,
due to computational constraints we have carried out the simulations only for
high
mass binaries and hence the upper cutoff plays an important role in all our
simulations.
It is to be noted that the discussion of the
ambiguity function in section \ref{2pnsec} is not valid when the effect of the
upper
cutoff is taken into account though for low mass binaries, such as ns-ns
binaries,
the results there will still
be valid. The further dependence of the signal waveform on the total mass of
the
system through the upper cutoff means that we can estimate the individual
masses
more exactly though the computational power is bound to increase.
In order to carry out the simulations for the present case we selected
a $10 M_\odot-1.2 M_\odot$ binary system as this enables us to choose the
filter grid well
within valid limits of $\tau_0$ and $\tau_1$. The simulations were carried out
for various values of
SNR starting from 10.
The histograms of the estimated parameters at an SNR of 10 is
shown in fig. \ref{fig_phistu}.
At this SNR the errors obtained are $\sigma_{\tau_0} = 39.3$ ms,
$\sigma_{\tau_1} = 22.4$ ms, $\sigma_{t_a} = 23.1$ ms and $\sigma_{t_C} = 0.6$
ms.
These can be compared with the values in table \ref{pntab} and we can see that
except for
the parameter $t_C$ the errors are substantially lesser when the upper cutoff
is incorporated
into
the waveform. It is necessary to recompute the covariance matrix, as emphasized
before,
including the effect of the upper cutoff in order to compare these numerically
obtained values
with the covariance matrix. In order to do this it is not enough to replace the
upper limit
in the integral in eq. (\ref{gamma3}) with the upper cutoff. The waveform now
depends on
the total mass of the system through the upper cutoff and this information has
to be
incorporated into the waveform.
We carried out the simulations for various SNRs for the same value of masses
quoted above.
In the absence of the estimates of the covariance matrix when the upper cutoff
is incorporated
we assume that at an SNR of 40 the Monte Carlo estimates are consistent with
those of the covariance matrix.
In Figure~\ref{fig_psigu} we illustrate the results of our simulations.
The points on the dotted line
are the values obtained through Monte Carlo estimates and the continuous line
is obtained by fitting a
$1/\rho$ dependence of the errors on the SNR {\em assuming } consistency at an
SNR of 40.
It is seen as in the
previous simulations that except for the parameter $t_C$ the other parameters
are fairly consistent with a
$1/\rho$ dependence of the errors on the SNR at an SNR greater than 25.
\section{Conclusions}
\label{sec_con}
In this paper we have introduced the use of differential geometry in studying
signal analysis
and have addressed issues pertaining to optimal detection strategies
of the chirp waveform. We have also carried out Monte Carlo simulations to
check how well the
covariance matrix estimates the errors in the parameters of the chirp waveform.
We summarize below our main results.
\begin{enumerate}
\item We have developed the concept of a signal manifold as a subset of a
finite dimensional
vector space of detector outputs. Using the correlation between two signal
vectors as a scalar
product we have induced a metric on the signal manifold.
With this geometric picture it is possible
to pose the question of optimal detection in a more general setting. We
suggest that
the set of template waveforms for the detection of the chirp signal need not
correspond to
any point on the chirp manifold. We propose an algorithm to choose templates
off the signal manifold and show that the drop in the correlation due
to the discreteness of the set of templates is reduced.
This algorithm, though certainly not the best, motivates the search of more
efficient templates. In addition,
the chirp manifold corresponding to the second post-Newtonian waveform
is shown to be effectively
one-dimensional. This has important implications for the computational
requirement for the on-line detection of the chirp signal.
The use of a convenient set of parameters of the chirp waveform for carrying
out numerical and analytical simulations is stressed.
These parameters are such that the metric components are independent of
the parameters which implies that the manifold is flat and the corresponding
`coordinate system' is Cartesian.
As the metric defined is nothing but the Fisher information matrix, the
covariance matrix, being the
inverse of the Fisher information matrix, is also independent of the
parameters.
\item Monte Carlo simulations have been carried out for the case of the initial
LIGO
to find out whether the actual errors in
the estimation of parameters is consistent with the values predicted by the
covariance matrix.
Simulations have been carried out for both the Newtonian as well as the
post-Newtonian
waveforms. We have restricted ourselves to the case of high mass binary
systems, such as bh-ns binaries, where the computational requirement is
not very heavy since the length of the data train, in such cases,
works out to be less than $8$~s. Nevertheless, as has
been shown in this paper, the covariance matrix is independent of the
parameters identified by us when waveforms are terminated at a constant
upper cutoff irrespective of their masses. Consequently, our results will hold
good
for binary systems of arbitrary masses. We point out the major
problems that arise while performing a numerical simulation
and, where appropriate, we suggest how they may be taken care of.
In particular, the effect of incorporating the upper cutoff in the
frequency of the gravitational wave at
the onset of the plunge, which essentially depends on the total mass
of the binary, is extremely important for high mass binaries.
Since the covariance matrix with the inclusion of such a mass-dependent
upper cutoff is not available we have carried out most of our simulations
using a constant upper cutoff. This enables us to directly compare the
results of our Monte Carlo simulations with those of the
analytically computed covariance matrix. Since
for binaries with total mass less than $5 M_\odot$ the plunge
induced upper cutoff is larger than that induced by the detector noise
these effects can be ignored for such binaries.
The numerical experiments indicate that the covariance matrix underestimates
the errors in the
determination of the parameters even at SNRs as high as 20. In the Newtonian
case the correlation
coefficient of the time of arrival $t_a$ and the Newtonian chirp time
$\tau_0$ is found to be very close to $-1,$ so much so
that even at a SNR of $7.5$, the instant of coalescence $t_C = t_a + \tau_0$
remains practically a constant.
The error in the estimation of $\tau_0$ for the post-Newtonian waveform is
about four times the
error obtained in the case of the Newtonian waveform at the same SNR. This is
expected as the
first post-Newtonian correction to the waveform introduces a new parameter
$\tau_1$
(called the first post-Newtonian chirp time) which is highly
(anti)correlated with $\tau_0$.
For the post-Newtonian waveform at an SNR of $10$ the error
in $\tau_0$ is about $3$ times that predicted by the covariance matrix.
This corresponds to a factor of $2$ in the chirp mass $\cal M.$
The distributions for the parameters
have been obtained and are seen to be unimodal distributions and
are slightly more sharper than a Gaussian.
When the plunge induced upper cutoff is incorporated into the
waveform the errors in the estimation of parameters
decrease by a factor of about $2.5$. The correlation coefficient
between $\tau_0$ and $\tau_1$ is also found to decrease, which is
consistent with our discussion in Section \ref{anacovar}.
The results obtained suggest that higher moments than that used in
obtained in computing the covariance matrix may be important
in the determination of the errors in parameter estimation.
In the geometric picture this amounts to taking into account curvature effects,
either
intrinsic or extrinsic.
\item We suggest that $t_C$ is a more suitable
parameter to estimate the direction to the source than the time of
arrival.
The latter is a kinematical parameter that fixes the time at which
the gravitational wave frequency reaches the lower cutoff of
the detector while the parameter $t_C$
has the physical significance of being the instant of coalescence.
At an SNR of 10 the error in $t_a$ is too large (20 ms)
to deduce the direction to the source accurately, whereas the error
in the parameter $t_C$ is less than $0.5$~ms.
This will further
go down substantially for the advanced LIGO. A detailed analysis of the
determination of the direction to the binary using
delays in $t_C$ between different detectors
is carried out in Bhawal and Dhurandhar \cite{BD95}
(also see \cite {BSD95}).
\end{enumerate}
We now suggest further work which needs to be done along the lines of this
paper.
A full understanding of the chirp signal manifold when higher post-Newtonian
corrections
are incorporated into the waveform is in order. This will help in the
development of more efficient algorithms
for the choice of templates in the detection problem and facilitate reduction
in computational time.
The Monte Carlo simulations which we have carried our
are for the case of a binary waveform correct up to first post-Newtonian order.
Moreover, only circular orbits are considered.
The effect of eccentricity is currently being investigated \cite{GBS95}.
Performing simulations when higher post-Newtonian corrections
are taken into account
calls for an immense amount of computational time. Fortunately,
matched filtering algorithm being amenable to parallelization
\cite {SD93}, one could aim
at using the massively parallel computers, that are now becoming available
the world over, in performing such simulations.
\acknowledgments
The authors would like to thank the members of the gravitational wave group at
IUCAA
especially S.D. Mohanty and B. Bhawal
for many helpful discussions. R.B. is being supported by the Senior Research
Fellowship
of CSIR, India.
|
\section*{\Large\bit 1) Introduction}
\par
Although the Kerr-geometry, when considered classically, represents
a vacuum solution of the Einstein-equations, there have been various
attempts \cite{Israel1,Israel2,Burinskii,Lopez}
to find a singular matter distribution as its source.
The natural candidate for locating such an energy-momentum
distribution is the singular region of the spacetime.\par
Using distributional techniques it is indeed possible to find
a tensor-distribution \cite{BaNa2}, which is supported in the
singular region and which from the distributional viewpoint
represents the right-hand side of the Einstein equations.
A central ingredient in this endeavour is the Kerr-Schild
decomposition of the metric which is the main reason for
the applicability of distributional techniques to the non linear
Einstein equations. Moreover, the flat (background) part of the
decomposition provides us with a natural notion of boosts as being
its associated isometries. However, as already noted in the
Schwarzschild-case \cite{AiSe}, boosting the metric itself does
not produce a sensible result. Therefore the strategy is to shift
from the metric to the energy-momentum tensor which has a
well-defined limit \cite{BaNa1}.
The resulting limit is a pp-wave \cite{JEK}.
Solving the Einstein-equations with this inhomogeneity produces
the so-called Aichelburg--Sexl (AS) geometry describing the
gravitational field of a massless point-particle.
The result is independent of the direction of the boost which is
due to the spherical symmetry of the original geometry. \par
This is, however, no longer the case for Kerr, which is only
axis-symmetric. In a first step \cite{BaNa3} the authors considered
therefore a boost along the preferred direction, namely the axis of
symmetry. The aim of the present work is to lift this restriction
and to investigate the form of the limit by boosting along an
arbitrary direction. Finally we will discuss the extremal case
where the boost direction becomes perpendicular to the axis of symmetry
in some detail.
\section*{\Large\bit 2) The general boost}
The main ingredient of our approach
is the Kerr-Schild decomposition of the Kerr geometry
\begin{equation}\label{KerrSchild}
g_{ab}=\eta_{ab} + f \>k_a k_b,
\end{equation}
where $\eta_{ab}$ denotes the flat background part, with respect
to which boosts
find their natural home. $k^a$ denotes a geodetic null vector field and $f$
a scalar function.
With respect to Kerr-Schild coordinates \cite{HaEl} $\eta_{ab}$ becomes
manifestly flat and $k^a$ and $f$ are given by ($\rho^2 =x^2 + y^2 $)
\begin{eqnarray*}
k^a=(1,k^i), && k^i = \frac{r\rho}{r^2+a^2} e^i_\rho -
\frac{a\rho}{r^2+a^2} e^i_\phi \\
f=\frac{2 m r}{\Sigma} , && \Sigma = \frac{r^4 + a^2 z^2}{r^2}
\end{eqnarray*}
where $r$ is subject to $r^4-r^2 (\rho^2 +z^2-a^2) -a^2 z^2 =0.$
The Ricci-tensor for geometries in the Kerr-Schild class takes
the form
\begin{equation}\label{Ricci}
R^a{}_b = \frac{1}{2} (\partial^a\partial_c(fk^c k_b) + \partial_b\partial_c(fk^ck^a) -
\partial^2(fk^a k_b) ),
\end{equation}
which in the Kerr case gives rise to the distributional
energy-momentum tensor \cite{BaNa2}
\begin{align}\label{KerrEM}
T^a{}_b &= \frac{m\delta(z)}{8\pi}\left\{ \frac{2}{a}\left(
\left[ \frac{a^2\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} ^3}\right ] -\frac{\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} } -
\delta(\rho- a) \right )(dt)^a(\partial_t)_b \right. \nonumber\\
&+((\partial_t)^a(e_\phi )_b - (e_\phi )^a (dt)_b)\left(
2\left [\frac{\rho\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} ^3}\right ]
-\frac{\pi}{a}\delta(\rho- a) \right ) \nonumber\\
&+\frac{2}{a}\left( - \left[\frac{\rho^2\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} ^3}\right ]
- \frac{\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} }
+ 2\delta(\rho- a) \right )(e_\phi )^a(e_\phi )_b \nonumber\\
&\left. -\frac{2}{a}\frac{\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} }(e_\rho )^a(e_\rho )_b\right \}
\end{align}
The square-bracket terms in (\ref{KerrEM}) represent distributional
extensions of the corresponding non-locally integrable functions
to the whole of test function space.
Their definition may be exemplified by
$$
\left(\left[ \frac{\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} ^3} \right] ,\varphi \right) :=
\int\limits_{\rho\leq a} d^2x\frac{1}{\sqrt{a^2-\rho^2} ^3 }(\varphi(x)
-\varphi(a e_\rho)),
$$
where $e_\rho$ denotes the radial unit-vector with respect to
polar coordinates. For a more detailed discussion on the origin of
these terms the reader is referred to \cite{BaNa2}.
Interpreting Kerr-Schild coordinates as being asymptotically at rest
we may rewrite $T^a{}_b$ with respect to an arbitrary Lorentz-frame.
The boost-plane is spanned by the timelike vector $P^a= m(\partial_t)^a$
and its orthogonal spacelike counterpart $Q^a= m e^a$. Without loss of
generality we may take $e^a$ to lie in the {\it x-z} plane
\begin{align}\label{newvars}
&m(e_z)^a = Q^a \cos\alpha - m\bar{e} ^a \sin\alpha
\nonumber\\
& m(e_x)^a = Q^a \sin\alpha + m\bar{e} ^a \cos\alpha
\end{align}
where $\alpha$ denotes the angle between the axis of symmetry
and the direction of the boost and $\bar{e} ^a$ the spacelike direction,
which spans together with $(e_y)^a$ the two-plane orthogonal to the
boost. With respect to (\ref{newvars}) the $\delta(z)$ factor in
(\ref{KerrEM}) fixes $(Q x)= m\bar{x} \tan\alpha,\,\bar{x} = ( \bar{e} x) $,
which in turn implies that
\begin{equation}\label{radID}
\rho^2 = x^2+y^2 =
\frac{1}{m^2}( (Qx) \sin\alpha + m\bar{x} \cos\alpha )^2 + y^2=
\frac{\bar{x} ^2}{\cos^2\alpha} + y^2 =: \bar{\rho}^2.
\end{equation}
So one ends up with the following expression for the
energy-momentum tensor
\begin{align}\label{covEM}
T^a{}_b = & \frac{\delta(Qx-m\bar{x} \> \tan\alpha )}{8\pi\cos\alpha}
\left\{ -\frac{2}{a}\left(
\left[ \frac{a^2\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} ^3}\right ] -\frac{\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} } -
\delta(\rhob- a) \right )P^a P_b \right. \nonumber\\
&+\left( 2\left [\frac{\bar{\rho}\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} ^3}\right ]
-\frac{\pi}{a}\delta(\rhob- a) \right ) \left \{ P^a \>m (e_{\phi})_b +
m (e_{\phi})^a\> P_b \right \}\nonumber\\
&+\frac{2}{a}\left( - \left[\frac{\bar{\rho}^2\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} ^3}\right ]
- \frac{\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} } + 2\delta(\rhob- a) \right ) m (e_{\phi})^a\>
m (e_{\phi})_b
\nonumber\\
&\left. -\frac{2}{a}\frac{\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} } m (e_{\rho})^a\> m (e_{\rho})_b
\right \}
\end{align}
where due to the $\delta$-factor in (\ref{covEM})
\begin{align}
m (e_{\phi})^a &= \frac{1}{\rho}( x\, m (e_y)^a\! - y\, m (e_x)^a )
= \frac{1}{\bar{\rho}} \left( -y\sin\alpha\>Q^a\!
- \frac{m}{\cos\alpha}
(y\cos^2\alpha\>(\bar{e} )^a\! - \bar{x} \>(e_y)^a ) \right) \nonumber\\
m (e_{\rho})^a &= \frac{1}{\rho}(x\, m (e_x)^a\! + y \, m (e_y)^a)
=\frac{1}{\bar{\rho}}\left( \bar{x} \tan\alpha\>Q^a\! +
m(\bar{x} \>(\bar{e} )^a\! +y\>(e_y)^a ) \right)
\end{align}
Although this expression may look rather unwieldy in comparison with
(\ref{KerrEM}) it allows a simple ultrarelativistic limit by letting
$m\to 0$ and replacing $P^a$ and $Q^a$ by their null limit $p^a$.
\begin{align}\label{urEM}
T^a{}_b =
& \frac{\delta(px)}{8\pi\cos\alpha}
\left\{ -\frac{2}{a}\left(
\left[ \frac{a^2\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} ^3}\right ] -\frac{\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} } -
\delta(\rhob- a) \right ) \right. \nonumber\\
&+\left( 2\left [\frac{\bar{\rho}\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} ^3}\right ]
-\frac{\pi}{a}\delta(\rhob- a) \right )\frac{2y\sin\alpha}{\bar{\rho}} \nonumber\\
&+\left( - \left[\frac{\bar{\rho}^2\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} ^3}\right ]
+ 2\delta(\rhob- a) \right )\frac{2y^2\sin^2\alpha}{a\bar{\rho}^2}
\nonumber\\
&\left. -\frac{\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} }\frac{2\sin^2\alpha}{a} \right \}p^a p_b
=: -\frac{1}{16\pi}g(\bar{x} ,y) \delta(px) \, p^a p_b
\end{align}
As expected the resulting energy-momentum tensor is that of a
pp-(shock)wave. For $\alpha\to 0$ only the first term in the curly bracket
survives and (\ref{urEM}) coincides with the result obtained in
\cite{BaNa3}.
\section*{\Large\bit 3) Perturbative evaluation}
In order to find the metric corresponding to the distributional
energy-momentum tensor (\ref{urEM}), one has to solve the
Einstein equations that in this setting take the form of the Poisson
equation
\begin{equation}\label{Poisson}
( \partial_{\bar x}^2 + \partial_{y}^2 ) f(\bar x,y) = g(\bar x,y)
\end{equation}
for the profile-function $f(x)\, \delta(px)$
of the pp-wave. It can be solved straight-forwardly in a
perturbative way except for the particular case of the orthogonal
boost $\alpha\to\pi /2$ which needs special care and will therefore
be dealt with in the next chapter.
Rescaling $\bar{x} $ by $\cos\alpha$ and denoting the new variable by $x$
(\ref{Poisson}) becomes
\begin{equation}\label{genPERT}
(\Delta + \tan^2 \!\! \alpha \; \partial_x^2\; )\sum\limits_{n=0}^\infty
f_n \sin^n\alpha =\frac{1}{\cos\alpha}(g_0 + \sin \alpha\: g_1 +
\sin^2\! \alpha\: g_2)
\end{equation}
where the $g_i$ may be read off from (\ref{urEM}).
Expanding $\cos\alpha$ and $\tan\alpha$ into power series with respect
to $\sin\alpha$ and grouping corresponding powers together yields
\begin{align}
&\Delta f_0 = g_0,\qquad \Delta f_1 = g_1,\nonumber\\
&\Delta f_{2n} + \sum\limits_{k=0}^{n-1} \partial_x^2 f_{2k} =
\frac{\poch{1/2}{n}}{n!} g_0+
\frac{\poch{1/2}{n-1}}{(n-1)!} g_2, \qquad n\geq 1,\nonumber\\
&\Delta f_{2n+1} + \sum\limits_{k=0}^{n-1} \partial_x^2 f_{2k+1}=
\frac{\poch{1/2}{n}}{n!}
g_1,\qquad n\geq 1.
\end{align}
Let us explicitly derive the first order perturbation $f_1$
which is determined by
\begin{align}\label{pert}
&\Delta f_1(x) = -8\frac{y}{\rho}\left(
\left[\frac{\rho\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} ^3}\right ]
-\frac{\pi}{2a}\delta(\rho- a) \right ).
\end{align}
In the region $0< \rho < a$ the classical analogue of (\ref{pert})
may be separated employing polar coordinates. Decomposing $f_1(x)$
into $\tilde{f}_1(\rho)\sin\phi$ we obtain the radial equation
\begin{align}
&\frac{1}{\rho}\partial_\rho(\rho\partial_\rho\tilde{f}_1 ) -
\frac{1}{\rho^2}\tilde{f}_1 =
-\frac{8\rho}{\sqrt{a^2-\rho^2} ^3}\nonumber\\
\intertext{which may be simplified by replacing
$\rho$ by $a e ^u$}\nonumber
&(\partial_u^2 -1)\tilde{f}_1 = -\frac{8e^{3u}}{\sqrt{1-e^{2u}}^3}.
\end{align}
This equation is easily solved by using the Green-function
$$\vartheta(u)\sinh u = \frac{1}{2}\vartheta(\rho-a)
\left( \frac{\rho}{a} -\frac{a}{\rho}\right )
$$
that gives rise to the particular solution
$$
f_1(x)=\frac{8}{\rho}\sqrt{a^2-\rho^2} \sin\phi.
$$
Taking into account the distributional identities
\begin{align}
\Delta\left( \frac{\vartheta(a -\rho) }{\rho}\sqrt{a^2-\rho^2} \sin\phi \right ) &=
-\left[\frac{\rho\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} ^3}\right ]\sin\phi +\nonumber\\
&\hspace*{2cm}\frac{\pi}{2a}\delta(\rho- a) \sin\phi +
2\pi a\partial_y\delta^{(2)}(x)\nonumber\\
\Delta\left( \frac{1}{\rho}\sin\phi\right)&=2\pi \partial_y \delta^{(2)}(x),
\end{align}
we find
\begin{equation}\label{perSOL}
f_1(x) = \vartheta(a -\rho) \left(\frac{8}{\rho}(\sqrt{a^2-\rho^2} -a) \right )
\sin\phi - 8\vartheta(\rho-a) \, \frac{a}{\rho}\sin\phi.
\end{equation}
The choice of the solution is dictated by the distributional
nature of the inhomogeneity and natural boundary conditions
for $\rho\to\infty$ which ensure that $f_1$ vanishes asymptotically,
which is equivalent to the fact that the whole solution tends
to the AS-geometry at large distances.
The last property may also be derived from the fact that $f_1$ tends to
zero in the limit $a\to 0$. Comparing (\ref{perSOL}) with the result of
the boost along the axis of symmetry \cite{BaNa3} shows that the
rotational contributions being inverse powers
die off only in the limit and not at some finite value. This behaviour is
easily understood
by interpreting (\ref{Poisson}) as a 2-dimensional electrostatic problem.
In the spherically symmetric case the corresponding charge distribution
produces only a monopole momentum and may therefore be replaced by
a pointlike distribution of total charge outside its support.
\section*{\Large\bit 4) Transversal Limit}
In order to calculate the limit $\alpha\to \pi/2$ of (\ref{urEM}) one has to
evaluate the expression on an arbitrary test function $\varphi$.
Only two different types of expressions occur in the calculation, namely
those which arise from the square-bracket terms and the concentrated
(delta) contributions respectively.
In the following we will exemplify the respective limits on typical
representatives from each class.
\begin{align}\label{example}
&\lim_{\alpha\to \pi/2}\frac{1}{\cos\alpha}
\left( \left[ \frac{\vartheta(a -\rhob) }{\sqrt{a^2-\rhob^2} ^3}
\right ],\varphi \right ) =
\lim_{\alpha\to \pi/2}
\left( \left[ \frac{\vartheta(a -\rho) }{\sqrt{a^2-\rho^2} ^3}
\right ],\tilde{\varphi} \right )= \nonumber\\
&\lim_{\alpha\to \pi/2}\int\limits_{\rho\leq a} d^2 x
\frac{1}{\sqrt{a^2-\rho^2} ^3}\left( \varphi( \cos\alpha x,y) -
\varphi( \cos\alpha a\cos\phi,a\sin\phi)\right)=\nonumber\\
&\int\limits_{\rho\leq a} d^2 x\frac{1}{\sqrt{a^2-\rho^2} ^3}\left( \varphi(0,y) -
\varphi(0,a\sin\phi)\right),
\end{align}
where $\tilde{\varphi}(x,y):=\varphi(\cos\alpha x,y)$.
Unfortunately the above result is not in a very useful form.
Further simplification of (\ref{example}) may be achieved by
integrating out
$x$ in the first term, and $\rho$ in the second.
However, in order to perform these integrations we have to restrict the
domain of integration to a disk of radius $\bar{a} <a$ and do the limit
$\bar{a} \to a$ in the end. More explicitely we find
\begin{align*}
&\int\limits_{\rho\leq\bar{a} } d^2 x\frac{1}{\sqrt{a^2-\rho^2} ^3}
\varphi(0,y)=
\int\limits_{-\bar{a} }^{\bar{a} } dy\varphi(0,y)
\int\limits_{-\sqrt{\bar{a} ^2-y^2}}^{\sqrt{\bar{a} ^2-y^2} } dx
\frac{1}{\sqrt{a^2-\rho^2} ^3}=\\
& \frac{2}{\sqrt{a^2-\bar{a} ^2}} \int\limits_{-\bar{a} }^{\bar{a} } dy \varphi(0,y)
\frac{\sqrt{\bar{a} ^2-y^2 } }{a^2-y^2}\\
\intertext{and}
&\int\limits_{\rho\leq\bar{a} } d^2 x\frac{1}{\sqrt{a^2-\rho^2} ^3}
\varphi(0,a\sin\phi)=
\int\limits_0^{\bar{a} } \frac{\rho d\rho}{\sqrt{a^2-\rho^2} ^3}
\int\limits_0^{2\pi} d\phi \varphi(0,a\sin\phi)=\\
&\left(\frac{1}{\sqrt{a^2-\bar{a} ^2}}-\frac{1}{a} \right )
2\int\limits_{-a}^a\frac{dy}{\sqrt{a^2-y^2} }\varphi(0,y).
\end{align*}
Using l'Hospital's rule (\ref{example}) becomes
\begin{align*}
&\frac{2}{a} \int\limits_{-a}^{a} \frac{dy}{\sqrt{a^2-y^2} }\varphi(0,y)-
2 \lim_{\bar{a} \to a}
\sqrt{a^2-{\bar{a} }^2}\int\limits_{-\bar{a} }^{\bar{a} }
\frac{dy}{\sqrt{\bar{a} ^2-y^2}(a^2-y^2)}\varphi(0,y)=\\
&\frac{2}{a}\int\limits_{-a}^{a} \frac{dy}{\sqrt{a^2-y^2} }\varphi(0,y)
- \frac{\pi}{a}(\varphi (0,a) + \varphi (0,-a))=\\
&\left(\frac{2}{a}\frac{\vartheta(a-|y|)}{\sqrt{a^2-y^2} }\delta(\bar{x} ),\varphi\right ) -
\frac{\pi}{a}\left(\delta(a-|y|) \delta(\bar{x} ) ,\varphi\right).
\end{align*}
The limit of the simplest concentrated contribution gives
\begin{align}\label{example1}
&\lim_{\alpha\to \pi /2}\frac{1}{\cos\alpha}\left( \delta(\rhob- a) ,\varphi\right)=
\lim_{\alpha\to \pi /2} \left(\delta(\rho- a) ,\tilde{\varphi} \right)=\nonumber\\
&\lim_{\alpha\to \pi /2}
a \int\limits_0^{2\pi} d\phi \varphi(\cos\alpha a\cos\phi,a\sin\phi)=
a\int\limits_0^{2\pi} d\phi \varphi(0,a\sin\phi)=\nonumber\\
&2a\int\limits_{-a}^a \frac{dy}{\sqrt{a^2-y^2} }\varphi(0,y) =
2a\left(\frac{\vartheta(a-|y|)}{\sqrt{a^2-y^2} } \delta(\bar{x} ) ,\varphi \right)
\end{align}
Dealing with the remaining terms of (\ref{urEM}) in the same way we
finally end up with the ``transversal'' energy-momentum tensor
\begin{align}\label{orthEM}
T^a{}_b= & \delta(px) \delta(\bar{x} )\left\{ \frac{1}{2}\delta(a-|y|)
-\frac{1}{2}\left(\frac{y}{a}\right )\delta(a-|y|) \right\}p^ap_b = \nonumber\\
& \delta(px) \delta(\bar{x} )\delta(y+a)p^ap_b .
\end{align}
It is interesting to note that all contributions localized on the
line segment $|y| \le a$ compensate each other and the energy-momentum
tensor turns out to be concentrated on a pointlike region only.
The corresponding profile function is obtained by solving the
Poisson equation with (\ref{orthEM}) as inhomogeneity, which
gives
\begin{equation}\label{transPRO}
f(x)= -8\log\left( \frac{\sqrt{\bar{x} ^2 + (y+a)^2} }{\rho_0}\right),
\end{equation}
where $\rho_0$ denotes the length-scale of the AS-geometry \cite{BaNa3}.
The limit $a\to 0$ in fact reproduces the
AS-profile function.
\newpage
\section*{\Large\bit 5) Conclusion}
In the present paper we showed how to calculate the ultrarelativistic
limit of the Kerr-geometry without putting any restriction on the
direction of the boost. Our method is based upon the energy-momentum tensor
of the original Kerr-geometry, which in contrast to
metric admits a well-defined limit, thus avoiding possible
ambiguities arising from the removal of the infinities
of the metric-limit.
The resulting energy-momentum tensor has the form of a pp-(shock)wave,
depending parametrically on the angle $\alpha$
between the boost direction and the axis of rotation.
This dependence turns the support of the energy-momentum tensor into
an elliptical region in the two-dimensional subspace of the $px=0$-plane.
Therefore only the limiting cases $\alpha=0$ and $\alpha=\pi/2$,
where the support becomes a circle and a line-segment respectively,
admit a solution in closed form.
Nevertheless the general case allows a perturbative treatment
if one suitably rescales the coordinates and expands the resulting
expression with respect to $\sin\alpha$. An explicit calculation
shows that the ''screening'' behaviour of the longitudinal ($\alpha=0$)
case, where the solution turned into that of AS outside the disk with radius
$a$,
gets modified by contributions such that the whole solution displays only
asymptotic AS-behaviour.
The perturbative expansion breaks down in the limiting case $\alpha\to \pi/2$
where the direction of the boost becomes perpendicular to the axis of symmetry.
Taking into account the distributional nature of the limit it is nevertheless
possible to calculate the profile function of the limiting case in closed form.
\par\noindent
It would be interesting to investigate the dependence of particle scattering
on the angle $\alpha$ in comparison to the $\alpha=0$ case, since the latter
displays exactly the AS-behaviour outside the disk with radius $a$.
Work in this direction is currently under progress.
\newpage
|
\section*{Introduction}
\addcontentsline{toc}{section}{\bf Introduction}
Finding topological objects to characterize or even
classify tilings could be seen as a
major motivation to study tilings. This article is less
ambitious but already has the taste of it in that it is largely
devoted to the study of such an object: the integer group of coinvariants
associated to the tiling and the range of a state on it. But it takes
its stimulation from two areas of application. One is the
$K$-theoretical gap labelling of Schr\"odinger operator s describing the motion of a particle
in a tiling (and potentially many more features of solids for which a model
by a tiling is appropriate), and the other concerns topological dynamical
systems on the Cantor set. Let us briefly put these two areas into
context.
1) Tilings furnish (discrete) models of solids which may not be
periodic as for instance
quasicrystals.
A typical question arrising is that after the nature of the
spectrum of a particle moving in the tiling (e.g.\ a phonon or an electron).
The corresponding Schr\"odinger operator\ lies in the algebra of observables but its
diagonalization is for non periodic tilings to difficult to carry out
at present. To obtain at least some
qualitative description of the spectrum Bellissard proposed the $K$-theoretical
gap labelling \cite{Be4}. This gap labelling
requires the computation of a topological
invariant of a $C^*$-algebra\ which may be taken to be the algebra of observables,
namely of the range of a tracial state on its $K_0$-group.
The algebra of observables coincides with the algebra associated to the
tiling and the tracial state comes from a trace on it. The $K_0$-group
contains the integer
group of coinvariants (at least for a large class of tilings)
and the range of the tracial state equals its range on the coinvariants.
This range is a countable subgroup of $\mbox{\rm \twlset R}$ which contains the possible
labels
of gaps in the spectrum of a Schr\"odinger operator, i.e.\ its elements are the gap labels
predicted by $K$-theory.
As it is determined by
a topological invariant of the $C^*$-algebra\ associated to the tiling
the $K$-theoretical gap labelling will be of use if the
nature of the tiling is incorporated in the operator in a strong enough way
so that
all values of the above group (lying between $0$ and $1$) correspond to gaps.
Although the quantification of the latter property is in general (in particular
in higher dimensions)
an unsolved problem we find it still worth persuing as there are no
analytical alternatives in higher dimensions so far.
2) A certain class of tilings, the decorations of $\mbox{\rm \twlset Z}^d$, yield
$d$-dimensional topological dynamical systems of the Cantor set.
Such a topological dynamical system is given by $d$ commuting homeomorphisms,
and a tiling which consists of decorated
$d$-dimensional unit cubes touching nicely at their faces yields an example,
the $d$ independent shifts yielding commuting homeomorphisms of the hull of the
tiling. In one dimension the study of the $K$-group of the associated
$C^*$-algebra\ (the one usually associated to the dynamical system coincides with the
one associated to the tiling)
turned out to be very fruitful. The ordered $K_0$-group with order unit,
which coincides with the integer group of coinvariants as an unordered group,
furnishes an invariant classifying the classes
of strong orbit equivalent dynamical systems \cite{HPS,GPS}.
In higher dimensions the interpretation of the $K$-group in terms of
the properties of dynamical systems is yet less clear but it was recently
realized that the (unordered) $K$-groups may be decomposed into cohomology
groups of the group $\mbox{\rm \twlset Z}^d$ \cite{FoHu}. In particular the group of
coinvariants furnishes part of the $K_0$-group thus motivating its computation
which is carried out below for tilings which allow for an invertible substitution.
In turn, most of the analysis done for decorations of $\mbox{\rm \twlset Z}^d$ remains valid
for a much larger class of tilings including the Penrose tilings, as will
be shown below using the concept of reduction.\bigskip
Although the above applications are formulated in the framework of
$K$-theory of $C^*$-algebra s we shall not emphasize the $C^*$-algebra ic aspects in this
article since the topological invariant discussed here may be defined purely on
the level of spaces, i.e.\ here of groupoids.
The article is organized as follows.
In the first section the groupoid associated to a tiling is described.
We proceed along lines sligthly different from \cite{Ke2}
by first introducing a multiplicative structure on the pattern classes
of the tiling. This multiplicative structure is almost a groupoid, and
using the notion of the radius of a pattern class one obtains
the groupoid associated to the tiling in a way which makes its
local nature transparent. In contrast to \cite{Ke2} this groupoid is only
principal if the tiling is not periodic whereby we gain that it yields
the right groupoid $C^*$-algebra\ in the periodic case, too.
We then define the integer group of coinvariants associated to the tiling.
Naturally, local manipulations, i.e.\ deformations of patterns, appear
as maps which
are close to being homomorphims of the introduced almost-groupoid s.
In fact, a substitution\ can be algebraically defined as a particular kind of such a
homomorphism. We discuss the concept of reduction and the notion of a
decoration of $\mbox{\rm \twlset Z}^d$ and show that many tilings occuring in the description
of quasicrystals reduce to decorations of $\mbox{\rm \twlset Z}^d$.
Section 2 is devoted to the $C^*$-algebra ic characterization and contains some
details about the $K$-theoretical gap labelling. We first recall the
definition of the algebra associated to the tiling. The main result
needed concerns tilings which reduce to decorations of $\mbox{\rm \twlset Z}^d$ and states
that their $C^*$-algebra\ is stably isomorphic to a crossed product with $\mbox{\rm \twlset Z}^d$.
In particular the
$K$-theory of such a tiling is that of a topological dynamical system.
The third section, which is kept independent of
sections 1.4 and 2, is restricted
to tilings which allow for an invertible substitution. Geometrically such a substitution\
may be obtained from an deflation but the analysis remains purely algebaic.
The existence of an invertible substitution\ reflects in the
possibility of obtaining an injective coding of the tiling in terms of the
paths over a graph. This is technically the
basis for the computation of the integer group of coinvariants and the range
of its state.
For one dimensional substitution s similar approaches have been successfully
carried out, both,
for the range of the state \cite{Q,BBG} as well as for the group itself
\cite{For,Hos}. In higher dimensions results regarding the
range of the state have been obtained under the restriction that
the substitution\ forces its border \cite{Ke2}.
The results obtained here are independent of the dimension of the tiling,
the only requirements being that the substitution\ is (locally) invertible
and, for the determination of the range of the state, primitive.
In fact, we obtain a family of codings and graphs for a tiling which
for substitution s which force their border all coincide and have
the substitution\ matrix as
connectivity matrix but are more complicated in the general case.
As applications we discuss the Thue-Morse substitution\ sequence and the
Penrose tilings.
\section{Tilings and groupoids}
Groupoids and almost-groupoid s play a central role in the description of tilings.
They carry a structure which generalizes a group structure in so far as
the multiplication of two elements $x$ and $y$ is only defined provided
$(x,y)$ lie in a specified set. We shall indicate this by $x \vdash y$.
As a general reference to groupoids which includes groupoid-$C^*$-algebra s we
use \cite{Ren}.
\subsection{Almost-groupoids}
\begin{df}
An almost-groupoid\ is a set $\Gamma$ together with a subset
$\Gamma^{\vdash}\subset\Gamma\times\Gamma$ of composable elements (we write
$x\vdash y$ for $(x,y)\in\Gamma^{\vdash}$)
and a product map $x\vdash y\mapsto xy$:
$\Gamma^{\vdash}\rightarrow\Gamma$ and an inversion
map $x\mapsto x^{-1}$: $\Gamma\rightarrow\Gamma$, such that the
following relations hold:
\begin{itemize}
\item[I1] $(x^{-1})^{-1} = x$,
\item[I2] $x\vdash y$ implies $y^{-1}\vdash x^{-1}$, in which case
$(xy)^{-1}=y^{-1}x^{-1}$,
\item[I3] $x\vdash x^{-1}$ and $x\vdash x^{-1}x$ and $xx^{-1}x=x$,
\item[A]
$(x\vdash y$ and $xy\vdash z)$ whenever $(x\vdash yz$ and $y\vdash z)$,
in which case $(xy)z = x(yz)$.
\end{itemize}
\end{df}
The elements of $\Gamma^0=\{x^{-1}x|x\in\Gamma\}$ are called units.
We will make frequent use of the maps $L,R:\Gamma\rightarrow\Gamma^0$
(often denoted $r,d$)
defined by
\begin{equation}
L(x) = xx^{-1} \quad\quad R(x)=x^{-1} x.
\end{equation}
An almost-groupoid\ is called principal
if the map $(L,R):\Gamma\to\Gamma^0\times\Gamma^0$ is injective.
It defines orbits in the space of units: $u$ and $v$ are called
orbit equivalent, $u\sim_{o} v$, whenever there is an $x$ such that
$L(x)=u$ and $R(x)=v$.
Note that $x\vdash y$ is equivalent to $x\vdash yy^{-1}y$ and $yy^{-1}\vdash y$
so that it is by $A$ equivalent to $x\vdash yy^{-1}$. In particular,
$R(x)=L(y)$ implies
$x\vdash y$.
\begin{df}
A groupoid is an almost-groupoid\ for which cancelation holds, i.e.
\begin{itemize}
\item[C]
$x\vdash y$ and $x\vdash z$ imply $y= z$.
\end{itemize}
\end{df}
This definition of a groupoid is equivalent
to the usual one as e.g.\ presented in \cite{Ren}.
It is not difficult to show that the cancelation axiom C indeed
implies that
\begin{itemize}
\item
$x\vdash y$ implies
$R(x)=L(y)$,
\item
$x\vdash y$ and $y\vdash z$ imply $x\vdash z$.
\end{itemize}
In particular the above implications do not have to hold for almost-groupoid s
and associativity is only guaranteed provided all
multiplications are defined.
The lack of cancelation may be put into an order.
\begin{df}
The order of an almost-groupoid\ is defined by
\begin{equation}
x\preceq y\quad\mbox{iff}\quad x\vdash y^{-1}\quad\mbox{and}\quad xy^{-1}=yy^{-1}.
\end{equation}
\end{df}
A topological almost-groupoid\ is an almost-groupoid\ which carries a topology
such that the product and the inversion map are continuous,
$\Gamma^\vdash$ carrying the relative topology.
A (locally compact) groupoid is called $r$-discrete if $\Gamma^0$ is open.
A homomorphism of almost-groupoid s is a map which maps composable elements resp.\ units
onto composable elements resp.\ units
and preserves multiplication and inversion.
In particular it preserves the order.
A homomorphism of topological almost-groupoid s is a continuous homomorphism
of almost-groupoid s.
\begin{df}
An inverse semi-group is an almost-groupoid\ for which $\Gamma^{\vdash}=\Gamma\times\Gamma$.
\end{df}
Any almost-groupoid\ can be made into an inverse semi-group by introducing an extra
element $0$ and extending the multiplication to any
$(x,y)\in\Gamma\times\Gamma$ by
\begin{equation}
x y = \left\{ \begin{array}{ll}
xy & \mbox{if $x\vdash y$} \\
0 & \mbox{else}
\end{array} \right.
\end{equation}
and $0 x = x 0 = 0$, $0 0 = 0$. $0$ is its own inverse and
greater than any other element.
Conversely we could see an almost-groupoid\ as an inverse semi-group with an element
$0$ satisfying $0 x = x 0 = 0$ for all $x$.
However, homomorphisms of almost-groupoid s may in general not be extended to
homomorphisms of inverse semi-groups and vice versa.
The difference which makes homomorphisms of inverse semi groups unsuitable
for the purposes of this article
is that for a homomorphism $\varphi$ of an almost-groupoid,
even if $\varphi(x)\vdash \varphi(y)$, it has only to satisfy
$\varphi(x)\varphi(y)=\varphi(xy)$ if $x\vdash y$.\bigskip
{\em Examples:}
1) The principal groupoid which will be associated to a tiling is
given by an equivalence relation, i.e.\
its elements are pairs of equivalent elements of some set $X$.
Multiplication is only defined for pairs $(x,y),(x',y')$ if $x'=y$ and then
given by $(x,y)(y,z)=(x,z)$, and inversion by
$(x,y)=(y,x)^{-1}$.
The topology of the groupoids in question need in
general not to coincide with the relative topology from $X\times X$.
2) Sometimes the equivalence may be expressed as orbit equivalence under
the (right) action of a group $S$ acting on $X$: $x\sim y$ whenever
$\exists s\in S:y=x\cdot s$.
This leads to the consideration of another kind of groupoid which is called
transformation group \cite{Ren}.
Its space is the Cartesian product of $X$ with $S$ (here always considered
to carry the product topology) and the groupoid
structure is defined by $(x,s)(x',t)=(x,st)$ provided $x'=x\cdot s$ and
$(x,s)^{-1}=(x\cdot s,s^{-1})$.
We write it as $X\times_\alpha S$, $\alpha$ indicating the action. It
may be viewed as the groupoid defined by orbit equivalence
only if $S$ acts freely on $X$.
3) Any groupoid $\Gamma$ defines an inverse semi-group ${\cal ISG}(\Gamma)$
consisting of so-called $\Gamma$-sets. A $\Gamma$-set is a subset of
$\Gamma$ for which the restrictions of $L$ and $R$ to it are both injective.
If $X,Y$ are such $\Gamma$-sets (which may be empty)
then $XY:=\{xy|x\in X,y\in Y,x\vdash y\}$
and $X^{-1}:=\{x^{-1}|x\in X\}$. This multiplication of sets is not only
for $\Gamma$-sets defined. We will make frequent use of it.
\subsection{Tilings}
Borrowing the terminology for $d=2$ from \cite{GrSh}
a $d$ dimensional tiling is a (countable) family of
closed subsets of $\mbox{\rm \twlset R}^d$, called the
tile s, which cover $\mbox{\rm \twlset R}^d$, overlap at most at their boundaries and may
carry an additional decoration, e.g.\ to break symmetries.
We restrict the possible shape of tile s through the requirements that
they are of finite size, i.e\ they fit inside an $r$-ball for finite $r$,
and they are the closure of their interior.
We shall consider the geometrical objects as
equivalence classes under translations in $\mbox{\rm \twlset R}^d$ (but not under rotations
or reflections).
A tiling or its class $T$ defines an almost-groupoid\ as follows.
Consider the set
of pattern class es (with resp.\ to translation) of $T$
where a pattern of a tiling is given by a finite subset of its tile s.
A pattern class\ $M$ together with two tiles chosen from it (pointed out)
will be called a doubly pointed pattern class. It will be denoted by $M_{xy}$ where $x$ is the first
and $y$ the second pointed tile. The set of all doubly pointed pattern class es of $T$ will be
denoted by $\mtxx{T}$ or simply by ${\cal M}_{\rm I\!I}$ if the tiling is clear from the
context.
${\cal M}_{\rm I\!I}$ carries an order structure: $M_{x_1x_2}\dot{\preceq} N_{y_1y_2}$ iff
the pattern class\ of $N_{y_1y_2}$ contains the pattern class\ of $M_{x_1x_2}$ at such a
position that tile s $x_1$ resp.\ $x_2$ become $y_1$ resp.\ $y_2$, c.f.\ figure.
\epsffile[0 0 430 110]{dFig6.ps}
Moreover ${\cal M}_{\rm I\!I}$ carries
the structure of an almost-groupoid\ whose order structure coincides with the one just
introduced:
Two elements $M_{x_1x_2}$, $N_{y_1y_2}$ are composable iff there is
an $L_{z_1z_2}$ and a third tile $z$ of $L$ such that
$M_{x_1x_2}\dot{\preceq} L_{z_1z}$
and $N_{y_1y_2}\dot{\preceq} L_{zz_2}$.
We then define the product
\begin{equation}
M_{x_1x_2} N_{y_1y_2} = \min\{L_{z_1z_2}|
\exists z\in\kt{L}
:M_{x_1x_2}\dot{\preceq} L_{z_1z}, N_{y_1y_2}\dot{\preceq} L_{zz_2}\}
\end{equation}
where the minimum is taken with respect to $\dot{\preceq}$ and $\kt{L}$ denotes
the tiles of the pattern class $L$. E.g.:
\epsffile[0 0 430 120]{dFig5.ps}
The inversion is given by
\begin{equation}
(M_{x_1x_2})^{-1} = M_{x_2x_1}.
\end{equation}
It is straightforward to see that $\dot{\preceq}$ equals the order relation of the
almost-groupoid. So we will leave the dot away.
We consider ${\cal M}_{\rm I\!I}$ as topological almost-groupoid\ carrying the discrete topology.
The space of units ${{\mTxx}^0}$ is given by the elements of the form $M_{xx}$
which we simply write $M_x$ and call pointed pattern class es. The maps $L$ and $R$ act as
\begin{equation}
L(M_{xy})=M_{xx}\quad\quad R(M_{xy})=M_{yy}.
\end{equation}
Let $\kt{T}$ be the set of tile s of $T$ where we identify tile s if they
are mapped onto each other by a translation which is a multiple of a
period of the tiling.
(Since $T$ is an equivalence class we are forced to do this.)
In particular, for tilings which are completely periodic, i.e.\
periodic in $d$ independent directions, $\kt{T}$ is a finite set.
We call a tiling together with a chosen tile a pointed tiling and denote it
by $T_x$.
Let $M_r(T_x)\in{{\mTxx}^0}$ be the smallest pointed pattern class\
which covers all balls of radius $r$ having their center inside tile\ $x$
and keep this tile chosen.
Define $r_c(T_x,T_{x'}):=\sup\{r|M_r(T_x)=M_r(T_{x'})\}$.
Then $d(T_x,T_{x'}):=e^{-r_c(T_x,T_{x'})}$ yields a metric on $\{T_x|x\in\kt{T}\}$
\cite{Ke2}.
\begin{df}
The hull $\Omega$ of $T$ is the completion of $\{T_x|x\in\kt{T}\}$ with respect to
the tiling metric $d(T_x,T_{x'})=e^{-r_c(T_x,T_{x'})}$.
\end{df}
Note that $\Omega$ is a finite set if and only if $T$ is completely periodic.
The elements of $\Omega$ may be interpreted as pointed tiling s since any of their
patterns is eventually determined by some $T_x$.
In other words $\omega\in\Omega$ may be thought of as being
approximated by an increasing chain of pointed pattern class es
(which we now write shorter $u$).
\begin{df}
A sequence of pointed pattern class es $\{u_\nu\}_\nu$ is called approximating
if
\begin{enumerate}
\item
any finite number of elements is composable, i.e.\
$\forall k\geq 1:u_1\dots u_k \vdash u_{k+1}$,
\item
$\mbox{\rm rad}(u_\nu)\to\infty$ where $\mbox{\rm rad}(u)$ is the largest $r$ such that all
balls of radius $r$ having their center inside the pointed tile are
covered by $u$.
\end{enumerate}
\end{df}
An approximating sequence approximates a unique pointed tiling\ which we write
as its limit, namely
$\omega=\lim u_\nu$ is the pointed tiling\
which contains all pointed patterns $u_\nu$ at its pointed tile.
It is clear that any pointed tiling\ has an approximation
(just choose a divergent sequence $\{r_\nu\}_\nu$ in $\mbox{\rm \twlset R}^+$
and take $M_{r_\nu}(\omega)$) so that
we may carry over the notion $\preceq$
by writing
$u\preceq \omega$ if $u\preceq u_\nu$ for some element of an approximation
of $\omega$.
The pattern class es of $\omega$ are pattern class es of $T$ as well, i.e.\
for all $\omega\in\Omega$, $\mtxx{\omega}\subset {\cal M}_{\rm I\!I}$. One says
that $\omega$ is locally homomorphic to $T_x$ (or $T$).
Consequently $\omega$ (or its class) is called locally isomorphic
to $\omega'$ if $\mtxx{\omega}=\mtxx{\omega'}$.
$T$ is called minimal if it is locally isomorphic to any $\omega\in\Omega$
which is equivalent to saying that the closure of the orbit of any $\omega$
equals $\Omega$.
Since the range of $d(T_x,\cdot)$
is discrete away from $0$ the hull is completely disconnected
(zero dimensional). This means that the topology is generated
by sets which are both open and closed. In fact,
to any unit $u\in{{\mTxx}^0}$
we may assign an open and closed subset
\begin{equation}
U_u:=\{\omega\in\Omega|u\preceq \omega\}
\end{equation}
of $\Omega$. The collection of all of these yields a basis for the topology.
Let ${\mTxx}_{2,\neq}$ denote the set of all doubly pointed pattern class es which are composed of two
neighbored tiles (i.e.\ which have boundaries with nonzero intersection)
and such that the first pointed tile is not equal to the second.
${\mTxx}_{2,\neq}$ generates ${\cal M}_{\rm I\!I}\backslash {\mTxx}_{1}$ where ${\mTxx}_{1}$ are the pointed pattern class es
consisting of one tile only.
The following compactness condition will be required furtheron.
\begin{itemize}
\item
${\mTxx}_{2,\neq}$ is a finite set.
\end{itemize}
Under this condition $\Omega$ is compact. In fact,
this condition is equivalent to the requirement that
for any $r>0$ there are only finitely many
pointed pattern class es which are covered by a ball of radius $r$,
and the latter has been used
to prove compactness of $\Omega$ in \cite{Ke2}.
To keep contact with \cite{Ke2} where proofs of the above statements may be
found, we visualize a (non periodic) pointed tiling $T_x$
as a particular representative of $T$ in $\mbox{\rm \twlset R}^d$. For this we
choose a reference point $0$ in $\mbox{\rm \twlset R}^d$ and introduce a puncture for each
pattern class\ of a tile (i.e.\ the puncture is a point of the tile)
so that $T_x$ is the representative for which tile\
$x$ lies with its puncture on $0$.\bigskip
The almost-groupoid\ ${\cal M}_{\rm I\!I}$ acts on $\Omega$ from the right (and analogously from the
left). Let
\begin{equation}
\Omega^\rhd=\{(\omega,c)|L(c)\preceq \omega\}\subset \Omega\times{\cal M}_{\rm I\!I}
\end{equation}
with relative topology. Define $\gamma:\Omega^\rhd\rightarrow \Omega:\,
(\omega,c)\mapsto\omega\cdot c$ through an approximating sequence:
Let $\lim u_\nu = \omega$ and set
\begin{equation}
(\lim u_\nu)\cdot c := \lim R(u_\nu c).
\end{equation}
More pictorially with $\omega=T'_x$ and $M_y\preceq T'_x$
we have $T'_x\cdot M_{yy'}:=T'_{x'}$
where $x'$ is the tile\ of $T'$ which coincides with $y'$ once $M_y$ has been
realized as the pattern\ at $x$ in $T'$.
Note that
\begin{equation} \label{24071}
L(c)\preceq \omega \mbox{ and } L(c')\preceq \omega\cdot c \quad\iff \quad
c\vdash c'\mbox{ and } L(cc')\preceq \omega
\end{equation}
and that
in this case
\begin{equation}\label{24072}
\gamma(\gamma(\omega,c),c')=\gamma(\omega,c c').
\end{equation}
Moreover
$\gamma(\cdot,M_{xy}):U_{M_x}\rightarrow U_{M_y}$ is continuous
so that we might call it a continuous right action of an almost-groupoid.
Let us define an equivalence relation on $\Omega^\rhd$ using approximations
for the elements of $\Omega$:
\begin{equation}
(\lim u_\nu,c)\sim (\lim u_\nu,c')\quad\mbox{whenever}\quad \exists n:
u_n c = u_n c'.
\end{equation}
It is straightforward to see that this definition is independent of the
approximation and that the relation is transitive. We denote
the equivalence classes by $[\omega,c]$. These classes may be visualized
as doubly pointed tiling s, namely $[T'_x,M_{yz}]$ can be represented by $T'_{xz}$, $z$ being
the tile\ of $T'_x$ corresponding to the $z$ in $M_{yz}$ once the latter has
been identified in $T'_x$ such that $y$ coincides with $x$.
\begin{lem}\label{l12041}
Let ${\cal R}$ be quotient of $\Omega^\rhd$ by the above equivalence relation
with quotient topology and consider the groupoid structure
defined by
\begin{equation}\label{06041}
[\omega,c][\omega', c']=[\omega, c c']
\quad\mbox{provided}\quad
\omega'=\omega\cdot c
\end{equation}
and
\begin{equation}
[\omega,c]^{-1}=[\omega\cdot c,c^{-1}].
\end{equation}
Then ${\cal R}$ is an $r$-disrete groupoid. Its space of units is homeomorphic
to $\Omega$.
\end{lem}
{\em Proof:}
First, note that (\ref{06041}) is well defined because of
(\ref{24071},\ref{24072}).
It is straightforward to check that ${\cal R}$ satisfies the axioms of
a groupoid.
The topology of ${\cal R}$ is generated by sets of the form
\begin{equation}
{\cal U}_c:=\left[U_{L(c)}\times\{c\}\right]=\{[\omega,c]|L(c)\preceq\omega\}.
\end{equation}
${\cal U}_c^{-1}={\cal U}_{c^{-1}}$ shows continuity of the inversion and
$m^{-1}({\cal U}_c)=\bigcup_{c_1\vdash c_2, c\preceq c_1c_2}{\cal U}_{c_1}\times{\cal U}_{c_2}$
that of multiplication $m:{\cal R}^\vdash\to{\cal R}$.
The space of units is $\{[\omega,u]|u=L(u)\preceq\omega\}$ which is
open and homeomorphic to $\Omega$ via $[\omega,u]\mapsto \omega$.\hfill$\Box$
\begin{df}
We call the groupoid ${\cal R}$ of Lemma~\ref{l12041} the groupoid associated to
$T$.
\end{df}
The orbit of ${\cal R}$ through $\omega$ is given by
$[\omega]_o=\{\omega\cdot c|L(c)\preceq \omega\}$ or more
pictorial $[T'_x]_o=\{T'_x|x\in\kt{T'}\}$.
Hence $[T'_x]_o$ is the equivalence class under translation $T'$:
$T_x \sim_o T'_y$ whenever $T=T'$.
If $T$ is not periodic then the quotient topology on the
space of orbits $[\Omega]_o$ is not Hausdorff. Such a
space is an example of a non commutative space in the sense of Connes
\cite{Cone}.
But the reader should be warned that what is called
the non commutative space of
Penrose tilings in \cite{Cone} does not agree with the space of orbits
of the hull of a Penrose tiling but is rather a quotient of it.
We shall comment on that in section~\ref{03081}.
Let $\tilde{\gamma}:{\cal R}\rightarrow\Omega\times\Omega$ be given by
$\tilde{\gamma}[\omega,c]=(\omega,\omega\cdot c)$.
\begin{df} The principal groupoid ${\cal S}$ associated to $T$
is the image of ${\cal R}$ under $\tilde{\gamma}$ with weak topology
and groupoid structure defined by the orbit equivalence under ${\cal R}$.
\end{df}
\begin{lem}
${\cal S}$ is isomorphic to ${\cal R}$ if and only if $T$ is non periodic.
\end{lem}
{\em Proof:} Clearly $\tilde{\gamma}$ is a surjective and open
homomorphism of topological groupoids.
But it is only injective if $\gamma$ is free in the sense that
$\omega\cdot c=\omega$ implies that $c$ is a unit. This is precisely the
case if $T$ is not invariant under a translation.\hfill$\Box$\bigskip
In particular ${\cal R}$ is principal if $T$ has no periodicity.
Note that the topology of ${\cal S}$ does not coincide with the relative topology
inherited from $\Omega\times\Omega$. Its topology is generated by
\begin{equation}
U_{M_{xy}}:=\tilde{\gamma}[{\cal U}_{M_{xy}}]=\{(T'_{x'},T'_{y'})|M_x\preceq T'_{x'},
T'_{y'}=T'_{x'}\cdot M_{xy}\}.
\end{equation}
In \cite{Ke2} the principal groupoid of a tiling was used instead of
${\cal R}$ but it turned out that from the point of view of physics
the groupoid-$C^*$-algebra\ of ${\cal S}$ is not appropriate for tilings
which have a periodicity in that it does not contain
all translation operators in that case.
\subsubsection{The integer group of coinvariants associated to a tiling}
An important ingredient for the characterization of a tiling is the
integer group of coinvariants of its associated groupoid.
This notion is borrowed from the case of transformation groups
$X\times_\alpha S$ where the action of $S$ on $X$ gives rise to an
action on the group $C(X,\mbox{\rm \twlset Z})$ of continuous integer-valued functions,
$\alpha(s)^*(f)(x)=f(x\cdot s )$. The coinvariants are then
the quotient of $C(X,\mbox{\rm \twlset Z})$ by the subgroup generated by
$\{f-\alpha(s)^*(f)|f\in C(X,\mbox{\rm \twlset Z}),s\in S\}$.\bigskip
The ample semi group ${\cal ASG}(\Gamma)$ of an $r$-discrete groupoid
$\Gamma$ is the sub-inverse-semi-group of
${\cal ISG}(\Gamma)$ given by its compact open $\Gamma$-sets.
${\cal ASG}(\Gamma)$ acts on ${\cal ASG}(\Gamma)^0$ by conjugation,
$U\cdot {\cal U} := {\cal U}^{-1} U{\cal U}$,
leading to a partially defined action on $\Gamma^0$ \cite{Ren}:
for $u\in L({\cal U})$,
\begin{equation}
\{u\}\cdot {\cal U}={\cal U}^{-1}\{u\}\,{\cal U}=\{c^{-1}uc\}
\end{equation}
where $c=L^{-1}(u)$. The integer group of coinvariants of
$\Gamma$
(or group of coinvariants with coefficients in $\mbox{\rm \twlset Z}$)
shall be the coinvariants of the $\mbox{\rm \twlset Z}$-module
$C_c(\Gamma^0,\mbox{\rm \twlset Z})$
of integer valued continuous functions over $\Gamma^0$ with
compact support
with respect to the above partially
defined action of ${\cal ASG}(\Gamma)$. More precisely define
$\eta:{\cal ASG}(\Gamma)\to C_c(\Gamma^0,\mbox{\rm \twlset Z})$ by
\begin{equation} \label{23061}
\eta({\cal U}) := \chi^{}_{L({\cal U})}-\chi^{}_{R({\cal U})}
\end{equation}
where $\chi^{}_{U}$ is the characteristic function on $U$ and
let $E_\Gamma$ be the subgroup of $C_c(\Gamma^0,\mbox{\rm \twlset Z})$ generated by $\mbox{\rm im}\,\eta$.
\begin{df}
The integer group of coinvariants of an $r$-discrete groupoid $\Gamma$ is
$$H(\Gamma)=C_c(\Gamma^0,\mbox{\rm \twlset Z})/E_\Gamma.$$
\end{df}
If ${\cal U}$, ${\cal U}'$ are disjoint $\Gamma$-sets and ${\cal U}\cup{\cal U}'$ is a $\Gamma$-set
as well then $\eta({\cal U}\cup {\cal U}') = \eta({\cal U}) +\eta({\cal U}')$.
For transformation groups ${\cal ASG}(X\times_\alpha S)$ contains
disjoint unions of sets of the form
$U\times\{s\}$ where $s\in S$ and $U\subset X$ is open and compact.
Since
$\eta(U\times\{s\}) = \chi^{}_U-\chi^{}_{U\cdot s}$
the above definition is indeed a generalization.
Let us consider groupoids associated to tilings.
Since the restrictions of $L$ and of $R$ to the sets ${\cal U}_c$ generating the
topology
are both injective ${\cal ASG}({\cal R})$ contains unions of these sets.
These unions may be chosen to be disjoined because the
hull is zero dimensional.
Now using $L({\cal U}_c)={\cal U}_{L(c)}$ one obtains
\begin{equation}
H({\cal R}) = \langle\{\chi^{}_{U_u}|u\in {{\mTxx}^0}\}\rangle/
\langle\{\chi^{}_{U_{L(c)}}-\chi^{}_{U_{R(c)}}|c\in {\cal M}_{\rm I\!I}\}\rangle.
\end{equation}
We call $H({\cal R})$ also the integer group of coinvariants associated to $T$.
\subsection{Reduction}
The concept of reduction is very important in the sequel.
In the context of dynamical systems it is sometimes called induction
but we prefer the notion established for groupoids \cite{Ram}.\bigskip
Let ${\cal N}$ be a sub-almost-groupoid\ of ${\cal M}_{\rm I\!I}$ and define
\begin{equation}
\Omega_{\cal N} := \bigcup_{c\in{\cal N}} U_{R(c)}
\end{equation}
as well as
\begin{equation}
\Omega^\rhd_{\cal N} := \Omega^\rhd\cap(\Omega_{\cal N}\times {\cal N})
\end{equation}
with relative topologies.
We shall often consider the case in which ${\cal N}$ is finitely generated,
i.e.\ ${\cal N}= \langle{\cal C}\rangle$ for
a finite subset ${\cal C}$ of ${\cal M}_{\rm I\!I}$, then
$\Omega_{\cal N} = \bigcup_{c\in{\cal C}} (U_{R(c)}\cup U_{L(c)})$.
In fact,
since $\Omega$ is compact and the $U_{u}$ form a basis of its topology
any open and closed subset of $\Omega$ arrises in this manner with a
finitely generated ${\cal N}$.
There are two groupoids which naturally arrise from such a situation, one
is the so-called reduction of ${\cal R}$ by $\Omega_{\cal N}$:
\begin{equation}
{\cal R}_{\Omega_{\cal N}}:=\{[\omega,c]\in{\cal R}|\,\omega,\omega\cdot c\in\Omega_{\cal N}\}
\end{equation}
(with relative topology), and the other is
\begin{equation}
{\cal R}_{\cal N}:=\left[\Omega_{\cal N}^\rhd\right]_{\cal N}
\end{equation}
where $[\omega,c]_{\cal N}=\{(\omega,c')\in [\omega,c]|c'\in{\cal N}\}$
(with quotient topology).
\begin{df}
We call ${\cal N}$
approximating if any $\omega\in\Omega_{\cal N}$ is approximated by a
sequence $\{u_\nu\}_\nu$ with elements in ${\cal N}$ and generating if
$[\omega,c]_{\cal N}\mapsto [\omega,c]:\,{\cal R}_{\cal N}\to{\cal R}_{\Omega_{\cal N}}$ is an
isomorphism.
We call ${\cal N}$ regular
if any orbit of ${\cal R}$ intersects $\Omega_{\cal N}$, i.e.\ if
$[{\cal R}]_o=[{\cal R}_{\cal N}]_o$.
\end{df}
Note that $[\omega,c]_{\cal N}\mapsto [\omega,c]:\,{\cal R}_{\cal N}\to{\cal R}_{\Omega_{\cal N}}$
provides an isomorphism if it is surjective.
Note also that for minimal $T$ any non-empty ${\cal C}\in{\cal M}_{\rm I\!I}$
generates a regular sub-almost-groupoid\ $\erz{\C}$.
None of the properties of the definition implies
any other as counterexamples show. However,
if ${\cal N}$ is generating then
$\tilde{{\cal N}}=\{c_1uc_2|c_i\in{\cal N},c_1\vdash c_2,R(c_1),L(c_2)\preceq u\in{{\mTxx}^0}\}$
is approximating and ${\cal R}_{\cal N}={\cal R}_{\tilde{{\cal N}}}$.
\begin{lem}
If ${\cal N}$ is approximating then the
topology of $\Omega_{\cal N}$ is generated by sets of the form
$U_u$ with $u\in{\cal N}^0$.
\end{lem}
{\em Proof:}
It is clear that the topology of $\Omega_{\cal N}$ is generated by sets of the form
$U_{uv}$ where $u\in {\cal N}^0$ and $u\vdash v\in{{\mTxx}^0}$. If ${\cal N}$ is approximating
then there exists for any $\omega\in U_{uv}$ a $u_\omega\in{\cal N}^0$
such that $uv\preceq u_\omega\preceq\omega$. Hence $U_{uv}=\bigcup_{uv\preceq\omega}
U_{u_\omega}$.
\hfill$\Box$
\begin{lem}\label{18071}
If ${\cal N}$ is regular then
for any $u\in{{\mTxx}^0}$ there is a finite subset $\{c_i\}_i\subset{\cal M}_{\rm I\!I}$
with $U_u=\dot{\bigcup}_i U_{L(c_i)}$ and $U_{R(c_i)}\subset\Omega_{\cal N}$.
\end{lem}
{\em Proof:}
By regularity we may find for $\omega\in U_u$ a $c\in{\cal M}_{\rm I\!I}$ such
that $\omega\cdot c\in\Omega_{\cal N}$.
Let
$r=\sup_{\omega\in U_u}
\inf\{r'|u\preceq M_{r'}(\omega),\exists c\in{\cal M}_{\rm I\!I}:\omega\cdot c\in\Omega_{\cal N},L(c)\preceq
M_{r'}(\omega)\}$.
Since $U_u$ is compact $r$ is finite.
Choose for any $\omega\in U_u$ a $c_\omega$ so
that $\omega\cdot c_\omega\in\Omega_{\cal N}$ and $L(c_\omega)=M_r(\omega)$.
In particular $U_{R(c_\omega)}\subset\Omega_{\cal N}$ and
$\omega\in U_{L(c_\omega)}\subset U_u$.
Hence $U_u=\bigcup_{u\preceq\omega}U_{L(c_\omega)}$. Since
$U_{M_r(\omega)}\cap U_{M_r(\omega')}$ is either $U_{M_r(\omega)}$ or empty
the union can be made disjoint and then, by compactness, finite. \hfill$\Box$\bigskip
The
embedding $i:{\cal R}_{\Omega_{\cal N}}\hookrightarrow {\cal R}$ furnishes a
surjective group homomorphism $i^*:C(\Omega,\mbox{\rm \twlset Z})\to C(\Omega_{\cal N},\mbox{\rm \twlset Z})$ which
satisfies $i^*\circ\eta=\eta\circ i^{-1}$. Therefore
$E_{{\cal R}_{\Omega_{\cal N}}}=E_{\cal R}\cap C(\Omega_{\cal N},\mbox{\rm \twlset Z})$ and $i^*$ induces a surjective
homomorphism $H({\cal R})\to H({\cal R}_{\Omega_{\cal N}})$.
\begin{lem} \label{23051}
If ${\cal N}$ is regular then
$H({\cal R})=H({\cal R}_{\Omega_{\cal N}})$.
\end{lem}
{\em Proof:}
We have to show that
that for $u\in{{\mTxx}^0}$ there is a finite $\{u_i\}_i\subset{{\mTxx}^0}$ with $U_{u_i}
\subset\Omega_{\cal N}$ and $[\chi^{}_u]_{E_{\cal R}}=\sum_i [\chi^{}_{u_i}]_{E_{{\cal R}}}$.
This follows immediately from the foregoing lemma taking $u_i=R(c_i)$. \hfill$\Box$\bigskip
{\em Remark:} Lemma~\ref{18071} implies even that the groupoids
${\cal R}$ and ${\cal R}_{\Omega_{\cal N}}$ are continuously similar: Let ${\cal C}$ be the
collection of $c_i$'s such that $\Omega=\dot{\bigcup}_i U_{L(c_i)}$ and
$U_{R(c_i)}\subset\Omega_{\cal N}$. Define $\Omega\to{\cal C}:\omega\mapsto c_\omega$,
$c_\omega$ being the unique element satisfying $L(c_\omega)\preceq \omega$, and
$\epsilon:{\cal R}\to{\cal R}_{\Omega_{\cal N}}$:
\begin{equation}
\epsilon[\omega,c]=[\omega\cdot c_\omega,c_\omega^{-1}cc_{\omega\cdot c}].
\end{equation}
Clearly $\epsilon$ is continuous and
$\epsilon\circ i = id_{{\cal R}_{\Omega_{\cal N}}}$
for the embedding $i:{\cal R}_{\Omega_{\cal N}}\to {\cal R}$.
But also
$i\circ\epsilon$ is similar to
the identity on ${\cal R}$, namely the map $\Theta:\Omega\to{\cal R}$:
$\Theta(\omega)=[\omega,c_\omega]$ satifies
$\Theta(L([\omega,c]))\epsilon([\omega,c])=[\omega,c]\Theta(R([\omega,c]))$.
This is a continuous version of the theorem of Ramsay \cite{Ram}
guaranteeing that the continuous cohomology of ${\cal R}_{\Omega_{\cal N}}$ is
isomorphic to that of ${\cal R}$ \cite{Ren}.
\subsection{Locally defined maps between tilings}\label{11071}
We want to formulate in the present framework
the concept of locally deriving one tiling from another,
a concept which has been introduced for tilings in \cite{BSJ}.
Roughly speaking a local derivation amounts to a deformation of pattern class es,
e.g.\ by adding or deleting or deforming boundaries or decorations of tiles.
Such a deformation will be given by a certain class of maps from
sub-almost-groupoid s of ${\cal M}_{\rm I\!I}$ to the doubly pointed pattern class es ${\cal M}_{\rm I\!I}'=\mtxx{T'}$ of the other
tiling class $T'$. The results will be partly needed in the third section.
\begin{lem}\label{27061}
Let ${\cal N}$ be an approximating sub-almost-groupoid\ of ${\cal M}_{\rm I\!I}$ and
$\hat{\varphi}:{\cal N}\to{\cal M}_{\rm I\!I}'$ be a map which preserves composability,
commutes with the inverse map
and satisfies
\begin{itemize}
\item[1]
$\hat{\varphi}(c)\hat{\varphi}(c') \preceq \hat{\varphi}(cc')$
\item[2]
(growth condition)
there is a $t>0$ such that $|\mbox{\rm rad}(\hat{\varphi}(u))-t\,\mbox{\rm rad}(u)|$ is
a bounded function on ${{\mTxx}^0}$.
\end{itemize}
Then
\begin{equation}
\varphi[\lim u_\nu,c]_{\cal N} := [\lim\hat{\varphi}(u_\nu),\hat{\varphi}(c)]
\end{equation}
defines a homomorphism $\varphi:{\cal R}_{\cal N}\rightarrow{\cal R}'$.
\end{lem}
{\em Proof:}
Let us first show that $\svarphi$ preserves the order. $c\preceq c'$ is by
definition $c\vdash {c'}^{-1}$ and $c'=cR(c')$. But then
$\svarphi(c)\preceq \svarphi(c)\svarphi(R(c'))\preceq \svarphi(c')$.
Now let $\{u_\nu\}_\nu$ be an approximating sequence. Then for any $1<k$,
$u_1\dots u_k\vdash u_{k+1}$. The first condition and preservation of
composability immediatly implies that
$\svarphi(u_1)\dots \svarphi(u_k)\vdash \svarphi(u_{k+1})$. By the growth condition
the radius of
$\{\svarphi(u_\nu)\}_\nu$ diverges so it
is an approximating sequence as well. Hence for $\omega=\lim u_\nu$,
$\varphi(\omega)=\lim \hat{\varphi}(u_\nu)$ is well defined and clearly continuous.
As well is
$\Omega_{\cal N}\times {\cal N}\to \Omega'\times{\cal M}_{\rm I\!I}'$: $(\omega,c)\mapsto
(\varphi(\omega),\hat{\varphi}(c))$.
Thus we are left to show that $\varphi$ induces a map on $[\Omega_{\cal N}^\rhd]_{\cal N}$
and preserves the groupoid structure.
(By definition of the topology its continuity is then guaranteed.)
We have $\svarphi(L(c))\succeq L(\svarphi(c))$ and hence
$\svarphi(R(u_\nu c))\succeq R(\svarphi(u_\nu)\svarphi(c))$. Therefore
does $L(c)\preceq\omega$ imply $\svarphi(L(c))\preceq\varphi(\omega)$ and
$\varphi(\omega\cdot c)=\lim\svarphi(R(u_\nu c))=\lim R(\svarphi(u_\nu)\svarphi(c))=
\varphi(\omega)\cdot\svarphi(c)$.
This shows that
$\varphi(\Omega_{\cal N}^\rhd)\subset{\Omega'}^\rhd$ and that $\varphi$ maps equivalent
elements onto equivalent ones, because
equivalence may be expressed as
$(\omega,c)\sim (\omega,c')$ whenever $\exists c'':c,c'\preceq c''$ and $L(c'')\preceq\omega$.
That $\varphi$ commutes with inversion on the level of groupoids follows from
the corresponding hypothesis on the level of the almost-groupoid s.
Finally, multiplication is preserved if
$(\varphi(\omega),\svarphi(c)\svarphi(c'))\sim (\varphi(\omega),\svarphi(cc'))$. In view of the
above characterization of the equivalence this follows from
$\hat{\varphi}(c)\hat{\varphi}(c')\preceq\hat{\varphi}(cc')$. \hfill$\Box$\bigskip
We say that $\varphi:{\cal R}_{\cal N}\rightarrow{\cal R}'$ is locally defined.
The corresponding homomorphism on the principal groupoids is simply
$(\varphi,\varphi):{\cal S}_{\cal N}:={\cal S}\cap(\Omega_{\cal N}\times\Omega_{\cal N})\to{\cal S}'$.
Details of the proof of the following theorem, which is
constructive, will be used in section 3.
\begin{thm}\label{25071}
Let $\varphi:{\cal R}\to{\cal R}'$ be a locally defined homomorphism.
If $\varphi$ has a left inverse then this left inverse is locally defined.
\end{thm}
{\em Proof:}
Let $a\in{\mTxx}_{1}$, ${\mTxx}_{1}$ denoting the set of pointed pattern class es of tiles.
Since $\varphi(U_a)$ is open and compact we may find a finite subset
$\Phi(a)$ of ${\cal M}_{\rm I\!I}'$ satisfying
\begin{equation} \label{26061}
\varphi(U_a)=\bigcup_{v\in \Phi(a)}U_{v}.
\end{equation}
Let $E^0(u)=\{e\in{\cal M}_{\rm I\!I}|R(e)=u\}$ and denote by $x(e)$ the first pointed
tile of $e$.
The expression $u=\prod_{e\in E^0(u)} e^{-1}x(e)e$ motivates to define
\begin{equation}
\Phi^e(u):=\{\hat{\varphi}(e^{-1})\}\Phi(x(e))\{\hat{\varphi}(e)\}
\end{equation}
where we have used set multiplication, and
\begin{equation}
\Phi(c):=\left(\prod_{e\in E^0(L(c))}\Phi^e(L(c))\right)\{\hat{\varphi}(c)\}.
\end{equation}
We claim that
\begin{equation}\label{27062}
\varphi({\cal U}_c) = \bigcup_{d\in\Phi(c)}{\cal U}_d.
\end{equation}
Indeed, using that $u\preceq\omega$ whenever $\forall e\in E^0(u):
x(e)\preceq\omega\cdot e^{-1}$ we get
\begin{eqnarray*}
\varphi({\cal U}_c) & = & \bigcap_{e\in E^0(L(c))}
\{[\varphi(\omega),\hat{\varphi}(c)]|\omega\cdot e^{-1}\in U_{x(e)}\} \\
& = &
\{[\omega',\hat{\varphi}(c)]|
\forall e\in E^0(L(c)): \omega'\in
\bigcup_{v\in\Phi(x(e))}U_{\hat{\varphi}(e^{-1})v\hat{\varphi}(e)}\} \\
&= &
\{[\omega',\hat{\varphi}(c)]|\omega'\in
\bigcup_{v\in\Phi(L(c))}U_v\}
\end{eqnarray*}
from which the claim follows. In particular, setting
${\cal N}=\Phi({\cal M}_{\rm I\!I})$, we have $\mbox{\rm im}\,\varphi={\cal R}'_{\cal N}$.
Furthermore let $\Phi(c)\vdash \Phi(c')$, which shall mean that
$\exists d\in\Phi(c)\,\exists d'\in\Phi(c'):d\vdash d'$. Then from
(\ref{27062}) we may conclude that $U_{R(c)}\cap U_{L(c)}\neq\emptyset$. Hence
\begin{equation}\label{27063}
\Phi(c)\vdash \Phi(c')\quad\mbox{implies}\quad c\vdash c'.
\end{equation}
Now set $W(d)=\{c\in{\cal M}_{\rm I\!I}|d\in\Phi(c)\}$ which by construction is finite.
If $c,c'\in W(d)$ then by (\ref{27062}) ${\cal U}_c\cap{\cal U}_{c'}\neq\emptyset$
and hence $c$ and $c'$ have a common greater element, e.g.\ $L(c)c'$.
Since $L(d)\in\Phi(L(c))\cap \Phi(L(c'))$ also $L(d)\in\Phi(L(c)L(c'))$
so that $d\in\Phi(L(c)c')$. It follows that
$\min\{c|\forall c'\in W(d):c'\preceq c\}=\max W(d)$. Define
$\hat{\psi}:{\cal N}\to{\cal M}_{\rm I\!I}$:
\begin{equation}
\hat{\psi}(d):=\max W(d).
\end{equation}
It commutes with the inverse.
Because of $d\in\Phi(\hat{\psi}(d))$ and (\ref{27063})
$d\vdash d'$ implies $\hat{\psi}(d)\vdash \hat{\psi}(d')$.
But it also implies $W(d)W(d')\subset W(dd')$
and hence $\hat{\psi}(d)\hat{\psi}(d')\preceq\hat{\psi}(dd')$.
Since $\Phi(A)$ is finite $\hat{\psi}$ and $\hat{\psi}\circ\hat{\varphi}$
satisfy the growth condition (with $t^{-1}$ resp.\ $1$ in place of
$t$) and ${\cal N}$ is approximating. Hence they extend to a homomorphisms
${\cal R}'_{\cal N}\to {\cal R}$ resp.\ ${\cal R}\to {\cal R}$.
Since $\psi\circ\varphi$ extends to the identity
$\psi$ extends to a left inverse.\hfill$\Box$\bigskip
We say that $\varphi$ is locally invertible (on its image).
\begin{lem}
If $T$ is non-periodic then the almost-groupoid\ ${\cal N}=\Phi({\cal M}_{\rm I\!I})$ occurring in the proof
of the above theorem is generating.
\end{lem}
{\em Proof:}
Consider
the commuting diagram
\begin{eqnarray*}
{\cal R} & \stackrel{\varphi}{\leftrightarrow} & {\cal R}'_{{\cal N}}\\
\tilde{\gamma}\updownarrow & & \downarrow \tilde{\gamma}' \\
{\cal S}& \stackrel{(\varphi,\varphi)}{\to} &
{\cal S}'_{{\cal N}}
\end{eqnarray*}
where $\tilde{\gamma}$ is an isomorphism since $T$ is non periodic.
Clearly $(\varphi,\varphi)$ is injective. Suppose that it is not
surjective.
Then there must be an $x\sim_o y$ such that
$\varphi^{-1}(x)\not\sim_o \varphi^{-1}(y)$. But this contradicts that
$\varphi^{-1}$ is a left inverse which preserves (like any
homomorphism of groupoids) orbits.
Therefore $\tilde{\gamma}'$ is bijective and hence ${\cal N}$ generating.\hfill$\Box$\bigskip
Counterexamples show that the requirement for $T$ to be non periodic is
neccessary.
\subsection{Decorations of $\mbox{\rm \twlset Z}^d$}
For many tilings the groupoid is a transformation group the group being
$\mbox{\rm \twlset Z}^d$. This is e.g.\ the case
if the tiling is built from unit cubes which nicely touch at the faces.
But since we allow for decorat ions these tilings do not have to be periodic.
They are called decorations of $\mbox{\rm \twlset Z}^d$.
In any case they allow for $d$ commuting homeomorphisms of the hull
which are given by the $d$ independent shifts such that the orbits of ${\cal R}$
become orbits under the action of $\mbox{\rm \twlset Z}^d$ and ${\cal R}$
a transformation group $\Omega\times_\alpha\mbox{\rm \twlset Z}^d$.
The structure of the $K$-groups of $C^*$-algebra s defined by such
transformation groups is fairly well understood and it turns out to be
useful as well in the case where ${\cal R}$ merely has
a reduction ${\cal R}_{\Omega_{\cal N}}$ which is isomorphic to $\Omega_{\cal N}\times_\alpha\mbox{\rm \twlset Z}^d$.
In view of Lemma~\ref{23051} this already implies for regular ${\cal N}$ that
the integer group of coinvariants coincide.
\begin{df}
We say that $T$ reduces to a decoration of $\mbox{\rm \twlset Z}^d$ if there is
a regular ${\cal N}$ such that ${\cal R}_{\Omega_{\cal N}}\cong\Omega_{\cal N}\times_\alpha\mbox{\rm \twlset Z}^d$.
If ${\cal N}={\cal M}_{\rm I\!I}$ then $T$ is called a decoration of $\mbox{\rm \twlset Z}^d$.
\end{df}
Tilings which reduce to
decorations of $\mbox{\rm \twlset Z}^d$ yield by definition $d$-dimensional topological
dynamical systems. The topological space is $\Omega_{\cal N}$ which is homeomorphic
to the Cantor set. These systems are
topologically transitive and the notion of minimality in the sense of
tilings and in the sense of dynamical systems coincide.\medskip
There is a large class of tilings
which reduce to decorations of $\mbox{\rm \twlset Z}^d$. These
tilings are composed of $d$ dimensional
parallel epipeds which may be decorated.
For their definition consider a set of $N$
vectors $\xi_1,\cdots,\xi_N$ which span $\mbox{\rm \twlset R}^d$.
A set $J\subset \{1,\cdots,N\}$ containing $n$ elements, $n\leq d$, such that
$\{\xi_i\}_{i\in J}$ are linear independent defines a subset
$\{\sum_{i\in J}c_i\xi_i|c_i\in [0,1]\}$ of $\mbox{\rm \twlset R}^d$ which is an
$n$ dimensional parallel epiped. We will call a translate of it
an $n$-facet of type $J$.
Their boundaries are $n'$-facets,
$n'<n$, of type $J'\subset J$.
Let $T$ be a tiling of $\mbox{\rm \twlset R}^d$ consisting
of possibly decorat ed \pe{d}s which are arranged
in such a way that
\begin{itemize}
\item[D1] the boundaries of tile s overlap, if at all, at common complete
$d'$-facets, $d'<d$.
\end{itemize}
Provided the
number of different decorat ions is finite, which we assume,
there is only a finite number of pattern class es consisting
of two tiles which touch each other so that the hull is compact.
If $d=N$ then $T$
itself a decoration of $\mbox{\rm \twlset Z}^d$ so here we are
interested in $d<N$.
Now fix a type $I_0$ of some tile\ appearing in $T$ and set
\begin{equation}\label{19061}
{\cal C}=\{a\in{\mTxx}_{1}|a\mbox{ is of type }I_0\}.
\end{equation}
where ${\mTxx}_{1}$ denotes the set of pointed pattern class es of tiles.
For any $i\in I_0$ choose a (common) normal for $d\!-\!1$-facets of type
${I_0\backslash\{i\}}$, i.e.\ a positive and a negative side,
and let ${\cal B}_i\subset{\mTxx}_{2,\neq}$ consist of those doubly pointed pattern class es which have as
common boundary a \pe{d-1}\ of type ${I_0\backslash\{i\}}$ and such that
their first resp.\ second pointed tile is on the positive resp.\ negative
side of the common boundary. Then $\Omega_{\erz{\C}}=\bigcap_i\Omega_{\erz{{\cal B}_i}}$.
Define $\breve{\alpha}_i:\Omega_{\erz{{\cal B}_i}}\to\Omega_{\erz{{\cal B}_i}}$ by
\begin{equation}
\breve{\alpha}_i(\omega) = \omega\cdot c
\end{equation}
where $c$ is the unique element of ${\cal B}_i$ for which $L(c)\preceq\omega$.
Clearly $\breve{\alpha}^{-1}_i(\omega) = \omega\cdot c^{-1}$
with the unique $c$ for which $R(c)\preceq\omega$.
Now we require in addition that
\begin{itemize}
\item[D2]
for all $i\in I_0$ and all $\omega\in \Omega_{\erz{\C}}$ there are natural numbers
$n,m>0$ such that $\breve{\alpha}_i^n(\omega)\in \Omega_{\erz{\C}}$ and
$\breve{\alpha}_i^{-m}(\omega)\in \Omega_{\erz{\C}}$.
\end{itemize}
Let $n^\pm_i(\omega)$ be the smallest $n>0$
for which $\breve{\alpha}_i^{\pm n}(\omega)\in \Omega_{\erz{\C}}$.
This defines maps $n^\pm_i:\Omega_{\erz{\C}}\to\mbox{\rm \twlset Z}^{>0}$ which
are continuous by the compactness of $\Omega_{\erz{\C}}$.
Consequently $\alpha:\Omega_{\erz{\C}}\to\Omega_{\erz{\C}}$
\begin{equation}
\alpha_i(\omega):=\breve{\alpha}_i^{n^+_{i}(\omega)}(\omega)
\end{equation}
are continuous maps (first return maps) having inverses
$\alpha_i^{-1}:=\breve{\alpha}_i^{n^-_{i}}$.
\begin{lem}
Let $T$ be a $d$ dimensional
tiling which is composed of (possibly decorat ed) parallel epipeds
and satisfies conditions {\rm D1,D2}.
Then the homeomorphisms $\alpha_i$ defined above commute pairwise.
Moreover, $\erz{\C}$ defined by (\ref{19061}) is regular.
\end{lem}
{\em Proof:}
Draw in any tile\ which has \pe{d-1}s of type
${I_0\backslash\{i\}}$ a line segment
joining the middle points of the boundary facets of
this type. (The middle point of an \pe{n}\ of type $J$ is at
$\sum_{i\in J}\frac{1}{2}\xi_i$.) This segment goes through the
middle point of the tile.
All the line segments in a tiling fit together to yield a set of infinite
line s in $\mbox{\rm \twlset R}^d$.
Such a line\ will be called $i$-line\ its type being $i$.
As an example the below figure
shows a patch of an octogonal tiling with its line segments of two types.
The four vectors $\{\xi_1,\dots,\xi_4\}$ are given to the right,
the line s are of type $1$ and $3$.
\epsffile[0 0 420 260]{gFig1.ps}
In general the line s have the properties:
\begin{itemize}
\item[1]
If $l\neq l'$
then $l\cap l'$ contains at
most one point (which is then the middle point of the tile through which
they both go),
and if $l$ and $l'$ are of the same type then $l\cap l'=\emptyset$.
\item[2]
Let $\pi_i$ be the orthogonal projection along the span of $\{\xi_j\}_{j\in
I_0\backslash\{i\}}$.
Then, for $d=2$ and any line\ which is of type $i$, $\pi_i(l)$
equals to the span of $\xi_i$.
\end{itemize}
Any type of line\ shall now be given the direction of the chosen
normal of $I_0\backslash\{i\}$.
The tile s
belonging to it can be ordered and the application of
$\breve{\alpha}_i$ can be geometrically interpreted as a shift
to the next tile along an $i$-line .
Condition D2 then reads that any line\ of type $i\in I_0$
contains in both directions
infinitely many tile s of type $I_0$.
Let us single out two elements of $I_0$ which we
denote by $1$ and $2$.
Let $l_0,m_0$ be the line s of type $1$ resp.\ $2$ which go through
$x$, the pointed tile of $T_x\in\Omega_{\erz{\C}}$. Let $m_1,l_1,m'$ be the
successor resp.\ of $m_0$ along $l_0$, of $l_0$ along $m_0$, and of
$m_0$ along $l_1$, c.f.\ the below figure where this
situation is indicated topologically.
\epsffile[0 0 420 180]{gFig2.ps}
Then $\alpha_1\circ\alpha_2(T_x)=\alpha_2\circ\alpha_1(T_x)$ if and only if
$m'=m_1$.
We first proof that this is the case for $d=2$.
By property 1 above
$m'$ can neither intersect $l_1$ at a second
point nor $m_0$ at all
nor can it intersect $l_0$ at a lower point than $m_1$ does.
The same reasoning applies to $m_1$ so that, since $\pi_2(m')$ is the span
of $\xi_2$,
$m'\cap m\neq\emptyset$.
Hence $m'=m_1$.
The case $d>2$ can be traced back to $d=2$ reducing $T_x$ to a
two dimensional tiling $\overline{T_x}$ by the following process.
Let $J_0=I_0\backslash\{1,2\}$.
We define the $J_0$-neighborhood of a tile $a$ which has
\pe{d-2}s of type $J_0$ to consist of this tile an those tiles
whose boundary intersects the boundary of $a$ at a \pe{d-2}\ of type $J_0$.
Let $\overline{T_x}$
be the smallest connected component consisting of tiles having
\pe{d-2}s of type $J_0$ in such a way that, first, it contains with a tile
its $J_0$-neighborhood, and second, it contains the pointed tile of $T_x$.
In particular $\overline{T_x}$ is a connected topological manifold
which contains complete $1$- and $2$-line s.
Let $\pi$ be the orthogonal projection along
$\langle\{\xi_i\}_{i\in J_0}\rangle$. Then $\pi(\overline{T_x})$ is a
two dimenional tiling to which we can apply the above arguement leading
to $\pi(m_1)=\pi(m')$.
To see that this implies $m_1=m'$ consider the homotopy
$F_t(\xi_i)=(1-t)\xi_i+t\pi(\xi_i)$ for $i\notin J_0$ and $F_t(\xi_i)=\xi_i$
for
$i\in J_0$. $F_1(\overline{T_x})$ may be understood as the
Cartesian product of $\pi(\overline{T_x})$ with a $d\!-\!2$-facet of type
$J_0$.
Certainly no new intersections between lines
in $F_t(\overline{T_x})$ occurr for $t>0$ so that
the decomposition shows that $\pi(m_1)=\pi(m')$ whenever $m_1=m'$.
By compactness of $\Omega_{\erz{\C}}$ the distance of two neighbored line s has an
upper and a lower bound. Hence for some $r$ any $u\in{{\mTxx}^0}$ of radius larger
than $r$ contains a
tile of type $I_0$. Therefore $\erz{\C}$ is regular.
\hfill$\Box$
\begin{thm}
Let $T$ be a $d$ dimensional
tiling which is composed of (possibly decorat ed) parallel epipeds
and satisfies conditions {\rm D1,D2}. Then it reduces to a $Z^d$ decoration.
\end{thm}
{\em Proof:}
The last lemma implies that the map
$\Omega_{\erz{\C}}\times_\alpha \mbox{\rm \twlset Z}^d\to{\cal R}_{\Omega_{\erz{\C}}}$: $(\omega,e_i)\mapsto
[\omega, c]$
where $c$ is uniquely determined by
$c=c_1\dots c_{n^+(\omega)}$ with $c_k\in{\cal B}_i$, $c_k\vdash c_{k+1}$,
and $L(c)\preceq\omega$, is
an isomorphism of groupoids.\hfill$\Box$\bigskip
All tilings which are obtained
by the so-called generalized
dual (or grid) method \cite{SoSt2} fall under the vicinity of the above theorem
provided they satisfy D2.
In fact, the lines of different type can be understood objects dual
to the hyperplanes of the grids so that we just inverted the construction of
\cite{SoSt2}.
In particular all tilings obtained by the
cut and projection method \cite{DuKa}
(which is equivalent to the
grid method using special grids only)
reduce to decorations as they all satisfy D2.
Among the substitution\ tilings which reduce to decorations of $\mbox{\rm \twlset Z}^2$
are the Ammann-Beenker-
the Socolar- and the Penrose tilings.
\section{$C^*$-algebra ic characterization of tilings}
We review the $C^*$-algebra ic constructions neccessary to formulate the $K$-theoretical
gap labelling for particles moving in a tiling.
As already mentionned the
$K$-theoretical gap labelling may be seen as an invariant of a $C^*$-algebra.
The natural candidate for this algebra is the algebra $A_T$ associated to
the tiling $T$ in which the particle is moving since it
contains all local operators involving translations and multiplications
with pattern dependent functions.
\begin{df}
The algebra ${\cal A}_T$ associated to a tiling $T$ is the reduced groupoid-$C^*$-algebra\
$C^*_{red}({\cal R})$ of the groupoid ${\cal R}$ associated to $T$.
\end{df}
Any $r$-dicrete groupoid defines a reduced groupoid-$C^*$-algebra\ \cite{Ren}.
Specified to the groupoid ${\cal R}$ it is defined as follows:
Let $C_c({\cal R})$ be the $*$-algebra of continuous
functions $f:{\cal R}\rightarrow \mbox{\rm \twlset C}$
with compact support and multiplication and involution given by
\begin{eqnarray} \label{25061}
f*g\,[\omega,c] & = &
\sum_{[\omega\cdot c,{c'}]\atop\omega,c\: \mbox{\tiny fixed}}
f[\omega,cc']\: g[\omega\cdot cc',{c'}^{-1}] , \\
f^*\,[\omega,c] & = & \overline{f[\omega\cdot c,c^{-1}]} \label{25062}.
\end{eqnarray}
$C_c({\cal R})$ is generated by the characteristic functions
$e_c$ onto ${\cal U}_c$ which satisfy the relations
\begin{eqnarray}
e_c*e_{c'} &=& \left\{
\begin{array}{ll}
e_{c c'} & \mbox{if $c\vdash c'$} \\
0 & \mbox{else}
\end{array} \right.\\
e_c^*&=&e_{c^{-1}}.
\end{eqnarray}
Hence in any representation $e_c$ is represented as a partial isometry.
$C^*_{red}({\cal R})$ is the closure of $C_c({\cal R})$
taken with respect to the reduced norm.
This norm is defined using the family of representations $\pi_{[\omega]_o}$,
one for each point in the non commutative\ space,
acting on the Hilbert space of square summable functions
$\Psi:[\omega]_o\rightarrow \mbox{\rm \twlset C}$ through
\begin{equation}
(\pi_{[\omega]_o}(e_{c})\Psi)(\omega)=\left\{
\begin{array}{ll}
\Psi(\omega\cdot c) & \mbox{if $L(c)\preceq\omega$} \\
0 & \mbox{else}
\end{array} \right..
\end{equation}
Now the reduced norm is given by
$\|f\|_{red}=\sup_{[\omega]_o\in[\Omega]_o}\|\pi_{[\omega]_o}(f)\|$
where $\|\pi_{[\omega]_o}(f)\|$ is the operator norm.
Since the topology of ${\cal R}$ has a countable basis
${\cal A}_T$ is separable.\bigskip
As an example consider a tiling which is a decoration of $\mbox{\rm \twlset Z}^d$.
That is there are $d$ commuting homeomorphisms $\alpha_i$ of the
hull such that ${\cal R}=\Omega\times_\alpha\mbox{\rm \twlset Z}^d$.
To any continuous
$f:\Omega\times_\alpha\mbox{\rm \twlset Z}^d\to \mbox{\rm \twlset C}$ with compact support one may assign
the function $\hat{f}:\mbox{\rm \twlset Z}^d\to C(\Omega)$ through
$\hat{f}(k)(\omega) = f(\omega,k)$. Carried over from (\ref{25061},\ref{25062})
multiplication and involution then become convolution resp.\ involution twisted
by
$\alpha$:
\begin{eqnarray}
\hat{f}*\hat{g}\:(k)& =& \sum_{m\in\mbox{\rm \twlset Z}^d} \hat{f}(m)\,
(\hat{g}(k\!-\!m)\circ\alpha(m)) \\
\hat{f}^*(k)& =& \overline{\hat{f}(-k)\circ\alpha(k)}.
\end{eqnarray}
The closure $C^*_{red}(\Omega\times_\alpha\mbox{\rm \twlset Z}^d)$ is
isomorphic to $C(\Omega)\times_\alpha\mbox{\rm \twlset Z}^d$,
the crossed product of $C(\Omega)$ with $\mbox{\rm \twlset Z}^d$ by the action
$\alpha(k)(\hat{f}(m))=\hat{f}(m)\circ\alpha(k)$,
$\alpha(k)=\alpha_1^{k_1}\circ\dots \circ \alpha_d^{k_d}$.
\bigskip
Remember that $T$ is minimal if the closure of the orbit of any $\omega$
equals $\Omega$.
The lattice of closed (twosided) ideals of ${\cal A}_T$
may be identified with the lattice of
open invariant subsets of $\Omega$, i.e.\ those open subsets
which contain next to an element $T$ all its equivalent elements \cite{Ren}.
The groupoid ${\cal R}$ is called minimal if $\Omega$ does not contain
any proper invariant open subset. But this is just excluded for minimal
tilings. So
if $T$ is minimal then ${\cal R}$ is minimal
which implies that ${\cal A}_T$ is simple \cite{Ren}.
\bigskip
Lemma~\ref{23051} has a $C^*$-algebra ic counterpart.
Two $C^*$-algebra s ${\cal A}$ and ${\cal B}$ are called stably isomorphic if
${\cal A}\otimes{\cal K}$ is isomorphic to ${\cal B}\otimes{\cal K}$, where ${\cal K}$ is the algebra of
compact operators (on an infinite dimensional separable Hilbert space).
A theorem of Brown \cite{Bro} states
that a full reduction ${\cal B}$ of ${\cal A}$ is stably isomorphic\ to ${\cal A}$.
In this context a reduction of ${\cal A}$ is determined by a projection $p\in{\cal A}$,
namely given by ${\cal A}_p:=\{x\in{\cal A}|px=xp=x\}$,
and ${\cal A}_p$ is full if the twosided ideal
generated by $p$ is dense in ${\cal A}$.
\begin{lem} \label{22061}
Let ${\cal N}\subset{\cal M}_{\rm I\!I}$ be regular. Then $C^*_{red}({\cal R}_{\Omega_{\cal N}})$
is stably isomorphic\ to ${\cal A}_T$.
\end{lem}
{\em Proof:}
Let $\chi^{}_{\Omega_{\cal N}}$ be the projection in ${\cal A}_T$ which is the characteristic
function onto $\Omega_{\cal N}$ and $f\in C_c({\cal R})$.
Then $f*\chi^{}_{\Omega_{\cal N}}=\chi^{}_{\Omega_{\cal N}}*f=f$ if and only if
$\mbox{supp}f\subset {\cal R}_{\Omega_{\cal N}}$.
Hence $C_c({\cal R}_{\Omega_{\cal N}})$ is a dense subalgebra of $({\cal A}_T)_{\chi^{}_{\Omega_{\cal N}}}$
and, since the closure of $C_c({\cal R}_{\Omega_{\cal N}})$ with respect to the
norm of ${\cal A}_T$ on the one hand and with respect to reduced norm defined
for ${\cal R}_{\Omega_{\cal N}}$ on the other coincide, we have
$({\cal A}_T)_{\chi^{}_{\Omega_{\cal N}}} = C^*_{red}({\cal R})_{\chi^{}_{\Omega_{\cal N}}}$.
To show that this reduction is full we show that
the two sided ideal generated by $\chi^{}_{\Omega_{\cal N}}$ contains $1$,
the unit of the algebra,
if ${\cal N}$ is regular. By
Lemma~\ref{18071} there is a finite set $\{c_i\}_i$ satisfying
$\Omega=\dot{\bigcup}_i U_{L(c_i)}$ and $U_{R(c_i)}\subset\Omega_{\cal N}$.
Then $1=\sum_i e_{c_i}e_{c_i^{-1}}=\sum_i e_{c_i}\chi^{}_{\Omega_{\cal N}}e_{c_i^{-1}}$
showing that 1 is contained in the ideal.\hfill$\Box$
\subsection{$K$-theoretical gap labelling}
A local selfadjoint operator $H$ describing the motion of a particle
in the tiling $T$ is an operator in
the representation $\pi_T$ of ${\cal A}_T$, namely
\begin{equation} \label{14061}
H\psi(T_x)=\sum_{x'\in X(T)} H_{x,x'}\psi(T_{x'}).
\end{equation}
Locality refers to the requirement that
the matrix element $H_{x,x'}$ depends only on a the doubly pointed pattern class\
of a certain size to which $x,x'$ belong, i.e.\ that $H=\pi_T(h)$ for some
$h\in C_c({\cal R})$.
In particular $h$ may be of the
form $h=-\Delta + \sum_i v_i e_{u_i}$ where $\{u_i\}_i$
is a collection of pointed pattern class es and $v_i\in\mbox{\rm \twlset R}$
and $\Delta$ is the discrete Laplacian. The latter takes
the form $\Delta=\sum_{A\in{\mTxx}_{2,\neq}} b_A e_A -\sum_{a\in{\mTxx}_{1}} b_a e_a$ where
$b_a\in\mbox{\rm \twlset R}$ and $b_A^*=b_{A^{-1}}\in\mbox{\rm \twlset C}$.
The values of the integrated density of states $N_H(E)$ of $H$ at energies
$E$ lying in a gap of its spectrum serve as labels for the gaps;
they are insensitive to certain perturbations of the operator.
In Bellissard's $K$-theoretic formulation of the
gap labelling \cite{Be4,Be1} these values
are recognized as elements of
$\mbox{\rm tr}_*K_0({\cal A})$ where ${\cal A}$ is a $C^*$-algebra\
represented on the above Hilbert
space, $H$ an element of its representation, $\mbox{\rm tr}$
a trace on ${\cal A}$, and $\mbox{\rm tr}_*$ the induced state on its $K_0$-group $K_0({\cal A})$:
The $K_0$-group of a unital
$C^*$-algebra\ ${\cal A}$ is obtained via Grothendieck's construction
from the monoid of projection classes of ${\cal A}\otimes{\cal K}$, i.e.\
the equivalence classes of projections of ${\cal A}\otimes{\cal K}$
under $p\sim q$ whenever $\exists u\in{\cal A}\otimes{\cal K}:
p = uu^*,q = u^*u$ may be added orthogonally, $[p]+[q]=[p\oplus q]$,
yielding a monoid $V({\cal A})$, and this monoid may be completed to an abelian group
whose elements are classes of pairs under
$([p],[q])\sim ([p'],[q'])$ whenever
$\exists [r]\in V(A):[p]+[q']+[r] = [q]+[p']+[r]$, see \cite{Bla,Mur} for
details.
Any trace of ${\cal A}$ extends to a trace on ${\cal A}\otimes{\cal K}$ and
defines a linear map $\mbox{\rm tr}_*:K_0({\cal A})\to\mbox{\rm \twlset R}$:
$[p]\mapsto \mbox{\rm tr}(p)$. The gap labelling in the above formulation requires the
equality
\begin{equation} \label{28071}
N_H(E) = \mbox{\rm tr}(\chi_{h\leq E})
\end{equation}
$h$ being the element which is represented by $H$ and $\chi_{h\leq E}$
the spectral projection
of $h$ to energies smaller or equal to $E$. This equality involves
validity of Shubin's formula \cite{Be1} by which the trace is
equated to the operator trace per volume in the corresponding representation.
Taking ${\cal A}$ to be the algebra associated to the tiling and considering
the above representation the $K$-theoretical gap labelling then reads:
if $E$ lies in a gap
\begin{equation}
N_H(E)\in
\mbox{\rm tr}_*K_0({\cal A}_T)\cap[0,1]
\end{equation}
provided (\ref{28071}) holds. In other words $\mbox{\rm tr}_*K_0({\cal A}_T)\cap[0,1]$
is the set of gap labels predicted by $K$-theory.
One motivation for using ${\cal A}_T$ is that
it is expected not to yield to many values on the r.h.s.\ so that
if the couplings $v_i$ are strong and diverse enough all elements of
$\mbox{\rm tr}_*K_0({\cal A}_T)$ actually occur as labels for
open gaps of $H$. In that case,
and if $\mbox{\rm tr}$ is faithful on ${\cal A}_T$,
the density of the values of the integrated density of states
on gaps in $[0,1]$
expresses the fact that the continuous part of the
spectrum is a Cantor set.
In fact, conclusions on the nature of the spectrum may partly be drawn
without the need to connect the gap labelling with the values of
the integrated density of states, any faithful trace may be used.
For instance if $\mbox{\rm tr}$ is faithful and the set of gap labels
${\cal L}_{\mbox{\rm tr}} (h)=\{\mbox{\rm tr}(\chi_{h\leq E})|E\notin\sigma(h)\}$ dense in $[0,1]$
the spectrum $\sigma(h)$ cannot contain a proper closed interval, for,
if $[a,b]\in\sigma(h)$, then by faithfulness
$\mbox{\rm tr}(\chi_{h\leq b})>\mbox{\rm tr}(\chi_{h\leq a})$ -- here $\mbox{\rm tr}$ has to be extended
to measurable functions over $\sigma(h)$ -- so that
$[0,1]\backslash{\cal L}_{\mbox{\rm tr}} (h)$
would contain the open interval $(\mbox{\rm tr}(\chi_{h\leq a}),\mbox{\rm tr}(\chi_{h\leq b}))$.
However, up to now there is no $K$-theoretic formulation of a condition for
a Schr\"odinger operator\ $h$
under which ${\cal L}_{\mbox{\rm tr}} (h)$ coincides with $\mbox{\rm tr}_*K_0({\cal A}_T)$.
For a $C^*$-algebraic formulation of such a condition for the discrete
magnetic Laplacian on $\mbox{\rm \twlset Z}^2$ see \cite{Sh2}.
Another consequence of the realization of $H$ as an element of a
representation of $A_T$ is that for minimal $T$
its spectrum has no discrete part, since the spectral projection onto
the eigenspace of a discrete eigenvalue with finite multiplicity
would have to be represented by a compact operator.
In fact, for minimal $T$ the algebra ${\cal A}_T$ is antilimial \cite{Ped},
and hence does not contain compact operators, namely
it is neither limial itself -- its primitive
spectrum contains one single point whereas two different representations
$\pi_{[\omega]_o}$ and
$\pi_{[\omega']_o}$ are not unitarily equivalent
-- nor can it
contain a limial ideal.
\subsubsection*{Traces on ${\cal A}_T$}
Any normalized trace $\mbox{\rm tr}$ on ${\cal A}_T$ restricted to a linear functional
$\mu$ on $C(\Omega)$ defines a normalized
measure on $\Omega$ also denoted by $\mu$ through
$\mu(f)=\mbox{\rm tr}(f)=\int f d\mu$. This measure is invariant under the groupoid
in the sense that $\mu(U_{L(c)})=\mu(U_{R(c)})$.
A direct consequence is that the values of $\mu$ on integer valued
continuous functions over $\Omega$ lie in the $\mbox{\rm \twlset Z}$-module generated by
the traces of projections of the algebra, and therefore
\begin{equation} \label{11051}
\mu(C(\Omega,\mbox{\rm \twlset Z}))
\subset\mbox{\rm tr}_*K_0({\cal A}_T).
\end{equation}
The existence of a faithful trace implies that
the notion of positive elements of $H({\cal R})={\cal C}(\Omega,\mbox{\rm \twlset Z})/E_{\cal R}$ as being those
which have a representative which is a positive function turns $H({\cal R})$ into
an ordered group. That is the positive elements $H^+({\cal R})$ satisfy
$H^+({\cal R})+H^+({\cal R})\subset H^+({\cal R})$, $H^+({\cal R})-H^+({\cal R})=H({\cal R})$ and
$H^+({\cal R})\cap -H^+({\cal R})=\{0\}$. In fact,
the existence of a faithful trace rules out that
a nonzero element can be both positive and negative. It being zero on
$E_{\cal R}$ it moreover defines a state (if normalized) on $H({\cal R})$.
A similar statement holds true for $K_0({\cal A}_T)$: The existence of a normalized
faithful trace guarantees that ${\cal A}_T$ is stably finite and therefore
the usual notion of positive elements of $K_0({\cal A}_T)$ as those
which have a representative which is a projection in ${\cal A}_T\otimes{\cal K}$
turns $K_0({\cal A}_T)$ into an ordered group (and the trace
induces a state on that group) \cite{Bla}.
Conversely any ${\cal R}$-invariant normalized
measure $\mu$ on the hull $\Omega$ defines a normalized
trace through
\begin{equation} \label{24061}
\mbox{\rm tr}(f):=\int_\Omega\mbox{P}(f)\,d\mu
\end{equation}
where $\mbox{P}:\,C^*_{red}({\cal R})\rightarrow C(\Omega)$ is the
restriction map. $\mbox{P}$ is the unique conditional expectation on
$C(\Omega)$ and is faithful \cite{Ren}.
Moreover if $T$ is minimal every non trivial
invariant measure has to have closed support $\Omega$ so that $\mbox{\rm tr}$ defined
by (\ref{24061}) is faithful.\bigskip
One of the goals of this article is the determination of $\mbox{\rm tr}_*K_0({\cal A}_T)$.
The question under which circumstances a given trace satisfies
Shubin's formula and (\ref{28071}) will not be addressed here, but see
\cite{Be1,BBG,Ke2} for investigations in this directions.
\subsection{$K_0$-groups for tilings}
The $K_0$-group of a
$C^*$-algebra\ depends together with its order structure only on its stable
isomorphism class.
The same holds true for the $K_1$-group which may be understood as
the $K_0$-group of the suspension of ${\cal A}$.
For this reason we may apply the known results on crossed products
to obtain the structure of the $K$-theory for tilings which reduce to
decorations. We are not able to present any results on $K$-groups
in case the tiling algebras are not stably isomorphic to crossed
products.\bigskip
A lot is known about the $K$-groups of crossed products of the form
$C(\Omega)\times_\alpha\mbox{\rm \twlset Z}^d$ in particular for zero dimensional $\Omega$.
Recently a relation of these $K$-groups with
the group cohomology $H(\mbox{\rm \twlset Z}^d,C(\Omega,\mbox{\rm \twlset Z}))$
of $\mbox{\rm \twlset Z}^d$ with coefficients in $C(\Omega,\mbox{\rm \twlset Z})$
was discovered \cite{FoHu}. We will not discuss group cohomology here
but what is important for us is
that the cohomology group of highest nonvanishing degree (which is $d$)
coincides with the group of
coinvariants of $\Omega\times_\alpha\mbox{\rm \twlset Z}^d$ and that the $K_0$-group
decomposes into
\begin{equation} \label{08075}
K_0(C(\Omega)\times_\alpha\mbox{\rm \twlset Z}^d)\cong H(\Omega\times_\alpha\mbox{\rm \twlset Z}^d)\oplus H'
\end{equation}
where $H'$ is trivial for $d=1$ and equal to $\mbox{\rm \twlset Z}$ in $d=2$.
(In general $H'$ is a direct sum of cohomology groups of degrees
$d-2n$, $0<n\leq \frac{d}{2}$ and $K_1(C(\Omega)\times_\alpha\mbox{\rm \twlset Z}^d)$
is a direct sum of cohomology groups of degrees
$d-1-2n$, $0\leq n<\frac{d}{2}$ \cite{FoHu}.)
Up to $d=3$ this result was
obtained before \cite{Els2} in an even more explicit form
in which in particular becomes clear
that
the image of the state on $K_0(C(\Omega)\times_\alpha\mbox{\rm \twlset Z}^d)$ induced by a
trace on
$C(\Omega)\times_\alpha\mbox{\rm \twlset Z}^d$ satisfies\footnote{
In \cite{Els2} ergodic measures have been used, but ergodicity not essential
for the proof of (\ref{06061}).}
\begin{equation} \label{06061}
\mbox{\rm tr}_*K_0(C(\Omega)\times_\alpha\mbox{\rm \twlset Z}^d)=\mbox{\rm tr}_*(H(\Omega\times_\alpha\mbox{\rm \twlset Z}^d)) =
\mu(C(\Omega,\mbox{\rm \twlset Z})).
\end{equation}
(\ref{06061}) holds for arbitrary $d$ under the restriction that
$\Omega\times_\alpha\mbox{\rm \twlset Z}^d$ splits into a Cartesian product
$\Omega\times_\alpha\mbox{\rm \twlset Z}^d=(\Omega_1\times_{\alpha_1}\mbox{\rm \twlset Z})\times
\dots\times(\Omega_d\times_{\alpha_d}\mbox{\rm \twlset Z})$
\cite{Ke2}.
(\ref{08075}) immediately carries over to tilings which reduce
to decorations of $\mbox{\rm \twlset Z}^d$ due to Lemma~\ref{22061}
\begin{equation}
K({\cal A}_T)\cong K(C(\Omega_{\cal N})\times_\alpha\mbox{\rm \twlset Z}^d).
\end{equation}
This also concerns the order structure which is however not known for $d>1$.\medskip
As for (\ref{06061})
recall that in case the tiling reduces to a decoration
${\cal A}_{\cal N}=C(\Omega_{\cal N})\times_\alpha\mbox{\rm \twlset Z}^d$ is a subalgebra of ${\cal A}_T$, i.e.\ there is
an embedding $\imath:{\cal A}_{\cal N}\rightarrow{\cal A}_T$.
This embedding induces an isomorphism
$\imath_*$ from $K_0({\cal A}_{\cal N})$ onto $K_0({\cal A}_{T})$.
In fact, for separable $C^*$-algebra s, ${\cal A}$ being stably isomorphic to ${\cal B}$
is equivalent to the existence
of a (strong) Morita equivalence ${\cal A}$-${\cal B}$-bimodule
which may be viewed as an element of $K\!K({\cal A},{\cal B})$ and is a special
case of a $K\!K$-equivalence \cite{Bla,Ska,Cone}.
Any $K\!K$-equivalence between ${\cal A}$ and
${\cal B}$ yields an isomophism from $K\!K(\mbox{\rm \twlset C},{\cal A})$ onto
$K\!K(\mbox{\rm \twlset C},{\cal B})$, namely
by multiplying it from the right, the multiplication being the Kasparov
product.
Translated into $K_0$-groups, $K\!K(\mbox{\rm \twlset C},{\cal A})$ being isomorphic to
$K_0({\cal A})$, the right multiplication of elements of $K\!K(\mbox{\rm \twlset C},{\cal A}_{\cal N})$
with the canonical Morita equivalence ${\cal A}_{\cal N}$-${\cal A}_T$-bimodule,
which as a linear space is
${\cal A}_{\cal N}{\cal A}_T$,
precisely becomes $\imath_*$.
Now $\mbox{\rm tr}\circ\imath$ is a trace
on ${\cal A}_{\cal N}$ which is normalized to
$\mbox{\rm tr}(\imath(1_{{\cal A}_{\cal N}}))=\mbox{\rm tr}(\chi^{}_{\Omega_{\cal N}})=\mu(\Omega_{\cal N})$.
Since the invariant measure
on $\Omega_{\cal N}$ corresponding to $\mbox{\rm tr}\circ\imath$ is $\mu|_{\Omega_{\cal N}}$ we get
\begin{equation}
\mbox{\rm tr}_*\circ \imath_*K_0({\cal A}_{\cal N})=\mu(C(\Omega_{\cal N},\mbox{\rm \twlset Z}))
\end{equation}
and therefore
\begin{equation}
\mbox{\rm tr}_*K_0({\cal A}_{T})=\mu(C(\Omega_{\cal N},\mbox{\rm \twlset Z}))\subset \mu(C(\Omega,\mbox{\rm \twlset Z})).
\end{equation}
Together with (\ref{11051}) this extends (\ref{06061}) to tilings
which reduce to decorations of $\mbox{\rm \twlset Z}^d$, $d\leq 3$.
$K_0({\cal A}_{\cal N})$ and $K_0({\cal A}_{T})$ differ only in their order units
(the images of the units of the algebras in $K_0$).
If one identifies them as above
the order unit of the former is the $K_0$-class of $\chi^{}_{\Omega_{\cal N}}$.
\section{Substitution tilings}
We have seen that the integer group of coinvariants furnishes part of the
$K$-theory
of the algebra of the tiling. In particular in one and two dimensions it
yields up to order the $K_0$-group. To determine this group we need further
structure which is provided by a locally invertible substitution.\bigskip
A substitution\ of a tiling may be thought of as a rule according to which the
tile s of the tiling
are to be replaced by pattern s which fit together to yield a new tiling.
An algebraic way to formulate this is by means of homomorphisms of
the almost-groupoid\ ${\cal M}_{\rm I\!I}$ incorporating thus their local nature.
For the definition recall that $E^0(R(c))=\{e\in{\cal M}_{\rm I\!I}|R(e)=R(c)\}$ and
that a homomorphism satisfies the growth condition with $t>0$ if
$|\mbox{\rm rad}(\hat{\rho}(u))-t\,\mbox{\rm rad}(u)|$ is a bounded function on ${{\mTxx}^0}$.
Let $x(c)$ be the first pointed tile of the doubly pointed pattern class\ $c$.
\begin{df}
A substitution\ of $T$ is a homomorphism of almost-groupoid s
\begin{equation}
\hat{\rho}:{\cal M}_{\rm I\!I}\to {\cal M}_{\rm I\!I}
\end{equation}
which satisfies the growth condition with $t>1$,
and for which the pattern class\ of
$\hat{\rho}(c)$ is composed without overlap (up to boundaries) of
the pattern class es of the $\hat{\rho}(x(e))$, $e\in E^0(R(c))$,
and the first resp.\ second pointed
tile of $\hat{\rho}(c)$ correspond to
the pointed tile of $\hat{\rho}(x(c))$ resp.\
$\hat{\rho}(x(c^{-1}))$.
\end{df}
Since $\hat{\rho}$ satisfies the growth condition it
defines a homomorphism of ${\cal R}$ into itself.
We shall call a tiling resp.\ its class
invariant under a substitution\ if $[\rho]_o(T)=T$. We call it a substitution\ tiling if it
allows for a substitution\ such that
$[\rho]_o(T)$ is locally isomorphic to $T$.
Like any homomorphism $\hat{\rho}$ is determined by its action on the generators
${\mTxx}_{2,\neq}$. But by the above definition this means that $\hat{\rho}$ may be
given, first, by its image on the pointed pattern class es of tiles, ${\mTxx}_{1}$,
and second, by the relative
position of these images.
A substitution\ can be iterated and it is
called primitive if for some $n$ and all $a\in {\mTxx}_{1}$ the pattern class\
$\rho^n(a)$ contains at least one tile of each tile class.\medskip
If a tiling allows for a substitution\ which has a right inverse, and this is
equivalent with locally invertible as we have seen, then the substitution\
can be used to compute the coinvariants associated to the tiling class.
We shall carry this out below but first give a brief discussion on
geometric realizations of substitution s.
\subsection{Deflation}
Geometrically we view a pointed tiling as a representative of $T$ in the
Euclidian space $\mbox{\rm \twlset R}^d$ with fixed origin $0$. To be precise we
choose a point for each tile class (its puncture). Then $T_x$ shall correspond
to the representative of $T$
for which the puncture of the pointed tile coincides with $0$ thereby
getting a bijective correspondance between pointed tilings and representatives
in $\mbox{\rm \twlset R}^d$ having the puncture of one of its tiles at $0$.
Substitutions appear then as deflations followed by rescaling.
Let $t^{-1}T$ denote $T$ rescaled by $t^{-1}$ and ${\cal M}(T)$ the (unpointed)
pattern class es of $T$.
A deflation
is given by:
\begin{itemize}\item[1]
a pattern class $\rho_t(a)\in{\cal M}(t^{-1}T)$ for each tile class $a$,
$t>0$.
\item[2]
a relative position between
$a$ and $\rho_t(a)$ for each tile class $a$.
\end{itemize}
Let ${\cal T}$
be a representative of $T$ in $\mbox{\rm \twlset R}^d$.
In the process of applying the deflation
any tile of class $a$ in ${\cal T}$ is to be replaced by a pattern of class
$\rho_t(a)$ (its replacement)
at the relative position given above.
To be more precise choose a tile in $\rho_t(a)$.
Its puncture shall now indicate the
position of a pattern of class $\rho_t(a)$ in $\mbox{\rm \twlset R}^d$. Then the
relative position i.e.\ the difference
between the position of the tile and its replacement shall be $x_a\in\mbox{\rm \twlset R}^d$.
Thus if a tile of ${\cal T}$ of class $a$ is at $x\in\mbox{\rm \twlset R}$ then
a pattern of class $\rho_t(a)$ is in $\rho_t({\cal T})$ at $x-x_a$.
The object $\rho_t({\cal T})$
is a composition of patterns of the classes $\rho_t(a)$, and it is required
that this yields
a tiling
which is locally isomorphic to $t^{-1}T$. In particular no
overlap and no gaps are allowed.
Such a deflation defines a substitution\ in the algebraic sense (as a homomorphism
of ${\cal M}_{\rm I\!I}$). By the above choice of a tile in $\rho_t(a)$
the latter may now be understood
as a pointed pattern class\ of $t^{-1}T$.
We let $\hat{\rho}(a)$ be $t\rho_t(a)$ which is $\rho_t(a)$ scaled by $t$,
the pattern class\ of $\hat{\rho}(M_{x_1x_2})$ be
$t\rho_t(M)$, and the first resp.\ second pointed tile be the one
corresponding to the
pointed tile of $\hat{\rho}(x(M_{x_1x_2}))$ resp.\ $\hat{\rho}(x(M_{x_2x_1}))$.\medskip
A deflation is locally invertible
if the determination of whether or not a pattern\ of $\rho_t({\cal T})$
is a replacement of a tile\ of ${\cal T}$ may be uniquely carried out by
inspection of the larger patterns around that pattern\ up to a given finite size.
In other words, for each $a$
there is a finite set of pointed pattern classes
$\Phi_t(a)\subset\mtx{t^{-1}T}$ with
$\forall v\in\Phi_t(a):\rho_t(a)\preceq v$
such that whenever one of them occurrs
in $\rho_t({\cal T})$ at $x\in\mbox{\rm \twlset R}$ then a tile of class $a$ occurrs in
${\cal T}$ at $x+x_a$.
This condition allows one to locally obtain a preimage of any
representative of any $t^{-1}T'$, $T'\in[\Omega]_o$.
But this furnishes also
a right inverse of $\rho$ on $\Omega$ as follows:
Let ${\cal T}$ be the representative corresponding to $\omega\in\mbox{\rm im}\,\rho$.
Then a preimage of $\omega$ under $\rho$ is given by
the pointed tiling corresponding to $\rho_t^{-1}(t^{-1}{\cal T})-x_a$ where
$t^{-1}{\cal T}$ is rescaled in such a way that the puncture remains on $0$ and
$a$ is
determined through the unique replacement $\rho_t(a)$ whose puncture is on $0$.
Clearly the right inverse of $\rho$ on $\Omega$ extends to ${\cal S}$.
And since non periodicity is forced by the existence of a locally
invertible deflation \cite{GrSh,Ke2} the corresponding substitution\ $\hat{\rho}$
is locally invertible in the sense of section~\ref{11071}.
\subsection{Path spaces over graphs and their dimension groups}
This section is meant to fix the notation thereby giving an overview
on the structures that will be needed.\bigskip
A graph $\Sigma$
is a set of vertices $\Sigma^{(0)}$ and a set of edges $\Sigma^{(1)}$ with
two maps $s,r:\Sigma^{(1)}\to \Sigma^{(0)}$, the source and the range map.
Its connectivity matrix\ is the $|\Sigma^{(0)}|\times|\Sigma^{(0)}|$ matrix with
coefficients
\begin{equation}
\sigma_{xy} := \mbox{number of edges which have source $y$ and range
$x$}.
\end{equation}
A path $\xi=\xi_1\dots\xi_n$ of length $n$ over $\Sigma$ is a sequence of
$n$ edges such that $r(\xi_k)=s(\xi_{k+1})$.
We denote its lentgh by $|\xi|$. One sets
$r(\xi)=r(\xi_{n})$ and $s(\xi)=s(\xi_{1})$. Two paths $\xi,\xi'$ with
$r(\xi)=s(\xi')$
may be concatenated to yield longer path
$\xi\circ\xi'=\xi_1\dots\xi_{|\xi|}\xi'_1\dots\xi'_{|\xi'|}$.
The set of all (half) infinite paths over the graph yields a space
which carries a compact
metric topology, it is called the path space ${\cal P}_{\Sigma}$ of the
graph.
Its topology is generated by sets
\begin{equation}
U_{\xi} = \{\xi\circ\gamma|s(\gamma)=r(\xi)\}.
\end{equation}
These sets are closed as well so that ${\cal P}_{\Sigma}$ is a zero dimensional.
Two (infinite) paths $\gamma$, $\gamma'$ are called confinal if for some $n$,
$\forall i\geq n: \gamma_i=\gamma'_i$. Cofinality is an equivalence relation
and the subspace of ${\cal P}_{\Sigma}\times{\cal P}_{\Sigma}$ of all cofinal pairs may be given
a topology which is generated by
\begin{equation}
{\cal U}_{\xi\xi'} = \{(\xi\circ\gamma,\xi'\circ\gamma)|s(\gamma)=r(\xi)\}
\end{equation}
where it is required that $|\xi|=|\xi'|$ and $r(\xi)=r(\xi')$.
With this topology the groupoid defined by the
equivalence relation becomes
an $r$-discrete principal groupoid ${\cal R}_\Sigma$.
Its groupoid-$C^*$-algebra\ ${\cal A}_{\Sigma}$ is finitely approximated, i.e.\ an $AF$-algebra.
The integer group of coinvariants of $H({\cal R}_\Sigma)$ of this groupoid is
$C({\cal P}_{\Sigma},\mbox{\rm \twlset Z})/E_\Sigma$ where $E_\Sigma$ is generated by elements of the form
$\eta({\cal U}_{\xi\xi'}) = \chi^{}_{U_\xi}-\chi^{}_{U_{\xi'}}$.
$C({\cal P}_{\Sigma},\mbox{\rm \twlset Z})/E_\Sigma$ carries an order structure given by the notion of
positive functions, e.g.\ $[f]_\Sigma>0$ whenever it has a representative
$f>0$. With that order structure and the constant function $1$ representing
the order unit $H({\cal R}_\Sigma)$
is also the dimension group or
scaled ordered $K_0$-group of the $AF$-algebra ${\cal A}_{\Sigma}$ \cite{Eff}.
We will not discuss the
$C^*$-algebra ic and $K$-theoretic details here
but only concentrate on the computation of the
dimension group which will be used lateron to determine the coinvariants of
${\cal R}$.
The main point is that $H({\cal R}_\Sigma)$ can be obtained by taking the algebraic
limit of the directed
system $(G_n,\sigma)$ where
$G_n=\mbox{\rm \twlset Z}^{|\Sigma^{(0)}|}$ and
$\sigma:G_n\to G_{n+1}$ is the homomorphism given by the connectivity matrix\
once the vertices have been identified with the standard base.
The algebraic limit of the above system
is a universal object which is a group $G$ together with
homomorphisms $j_n:G_n\to G$ such that $j_{n+1}\circ \sigma = j_n$ \cite{Lan}.
It is up to isomorphism determined by the property that, if there is any
other group $G'$ and $j'_n:G_n\to G'$ such that
$j'_{n+1}\circ \sigma = j'_n$ then there is a unique homomorphism
$j:G\to G'$ such that $j'_n=j\circ j_n$.
Moreover, $G$ inherits an order structure and the order unit from the
standard order structures on the $G_n$.
And $H({\cal R}_\Sigma)$ coincides with $G$ as ordered group with order unit.
There are several "standard" realizations for this limit \cite{Lan,Mur}.
We will use neither of them here but instead one which is less general
but more suitable for our means, c.f.\ \cite{Hos}.
In this realization the group
is isomorphic to the quotient
\begin{equation}\label{17061}
H({\cal R}_\Sigma) \cong \{x\in \mbox{\rm \twlset R}^k|\exists n\geq 0:\sigma^n(x)\in\mbox{\rm \twlset Z}^k\}/
\{x\in \mbox{\rm \twlset R}^k|\exists n\geq 0:\sigma^n(x)=0\}
\end{equation}
with $k=|\Sigma^{(0)}|$.
The maps $j_{n}:G_{n}=\mbox{\rm \twlset Z}^k\to H({\cal R}_\Sigma)$ are given by
$j_{n}=\pi\circ \sigma^{-n}$
where $\sigma^{-1}$ is taking the preimage. If $G'$ and $j'_n:G_n\to G'$
with $j'_{n+1}\circ \sigma = j'_n$ is any other realization then
$j(\pi(x))=j'_n(\sigma^n(x))$ for $\sigma^n(x)\in\mbox{\rm \twlset Z}^k$ yields the unique
homomorphism
$j:H({\cal R}_\Sigma)\to G'$ satisfying $j'_n=j\circ j_n$.
To express the order structure in this realization
we restrict for simplicity to the case where $\sigma$ is
primitive so that it has a Perron-Frobenius-eigen\-value\ $\tau$ with left-Perron-Frobenius-ei\-gen\-vec\-tor\ $\nu$.
If $\pi$ denotes the canonical projection of the above quotient,
the positive cone is
\begin{equation}
H^+({\cal R}_\Sigma)\cong\{\pi(x)\in H({\cal R}_\Sigma)|\sum_i\nu_i x_i >0\}\cup\{0\},
\end{equation}
and the order unit $\pi(w)\in H({\cal R}_\Sigma)$ with $w_i=1$.
Since $\sigma$
is primitive, there is a unique normalized measure $\mu$ on
${\cal P}_{\Sigma}$ which is invariant under the groupoid ${\cal R}_\Sigma$,
i.e.\ satisfies
$\mu(U_\xi)=\mu(U_{\xi'})$ in case $|\xi|=|\xi'|$ and $r(\xi)=r(\xi')$
\cite{Eff}.
If the left Perron-Frobenius-ei\-gen\-vec\-tor\ of
$\sigma$ normalized to $\sum_i\nu_i=1$ then
$\mu(U_\xi)=\tau^{-|\xi|}\nu_{r(\xi)}$
and its range on $C({\cal P}_{\Sigma},\mbox{\rm \twlset Z})$ is given by
\begin{equation}\label{27071}
\mu(C({\cal P}_{\Sigma},\mbox{\rm \twlset Z}))
= \{\tau^{-n}\sum_i\nu_i n_i|n\geq 0,n_i\in\mbox{\rm \twlset Z}\}.
\end{equation}
In particular $\mu$ is well defined on $H({\cal R}_\Sigma)$ and the order
may be expressed as
\begin{equation} \label{11073}
H^+({\cal R}_\Sigma) = \{x\in H({\cal R}_\Sigma)|\mu(x)>0\}\cup\{0\} .
\end{equation}
Elements which are neither positive nor negative are called infinitesimal.
\subsection{Path spaces determined by substitution s}
Any substitution, invertible or not,
defines an obvious graph which however does not even for all
locally invertible substitution s contain
enough information of the tiling. In analogy to the
one dimensional case we call it with \cite{For} the improper graph.
It has the substitution\ matrix as connectivity matrix.
But other graphs may also be attached to the substitution. These are related
to the improper graph and coincide with it in the border forcing case.
They allow for a coding of the tilings which yields a homeomorphism
between their path space and the hull.
This will be used to solve the
$K$-theoretical gap labelling for a certain class of substitution\ tilings.
\bigskip
Let for $a\in{\mTxx}_{1}$
\begin{equation}
E^n(a) = \{e\in{\cal M}_{\rm I\!I}|R(e)=\hat{\rho}^n(a)\}.
\end{equation}
The improper graph
$\Sigma$ is the graph which has vertices and edges resp.\
\begin{eqnarray}
\Sigma^{(0)}&=&{\mTxx}_{1} \\
\Sigma^{(1)}&=&\bigcup_{a\in{\mTxx}_{1}}\{(e,a)|e\in E^1(a)\}
\end{eqnarray}
and range resp.\ source maps $r,s:\Sigma^{(1)}\to\Sigma^{(0)}$ given by
\begin{eqnarray}
r(e,a)&=&a \\
s(e,a)&=&x(e).
\end{eqnarray}
Thus the edges of the graph which
have range $a$ are in bijective correspondence to the tiles in $\hat{\rho}(a)$.
$\Sigma$ has as connectivity matrix\
\begin{equation}
\sigma_{a_1a_2} = |\{e\in E^1(a_1)|x(e)=a_2\}|,
\end{equation}
i.e.\ $\sigma_{a_1a_2}$ equals the number of $a_2$'s in $\hat{\rho}(a_1)$.
This matrix (or sometimes its transpose) is also called the substitution\
matrix of $\hat{\rho}$.
A path of finite length like
$((e_1,a_1),\dots,(e_n,a_n))$ depends
due to the particular form of the range
map only on
$(e_1,\dots,e_n)$ and $a_n$. It shall be written shorter
$(e_1,\dots,e_n;a_n)$. \bigskip
The improper graph does not contain enough information of the
tiling to yield the right integer group of coinvariants.
To improve this
we have to incorporate the neighborhood of substitute s.
To a given doubly pointed pattern class\ $c$ let ${{\cal F}}(c)$ be the set of all possible
neighborhoods of $c$,
i.e.\ the set of pattern class es
occuring in $T$ which are composed of $c$
together with all the tiles
the boundaries of which have
a non empty intersection of with $c$.
The pointed tiles are those of $c$.
An integer $\upsilon\geq 0$ will parametrize a set of graphs
$\Lambda_\mb$. (This generality is needed lateron for applications.)
Define
\begin{equation} \label{30061}
{\cal B}^\upsilon_1:=\{(a,f)|a\in{\mTxx}_{1},
f\in{{\cal F}}(\hat{\rho}^\upsilon(a)),\rho^\upsilon(U_a)\cap U_f\neq\emptyset\}.
\end{equation}
The extra condition insures that only those neighborhoods of $\hat{\rho}^\upsilon(a)$
are taken into account which are neighborhoods of the $\upsilon$-fold substitute\ of $a$.
Let $\Lambda_\mb$ be the graph with vertices and edges
\begin{eqnarray}
\Lambda_\mb^{(0)}&=&{\cal B}^\upsilon_1\\
\Lambda_\mb^{(1)}&=&\{(e,a,f)|(a,f)\in{\cal B}^\upsilon_1,e\in E^1(a)\}
\end{eqnarray}
and range and source map
given by
\begin{eqnarray}
r(e,a,f)&=&(a,f)\\
s(e,a,f) &=& (x(e),f') \mbox{ with
$f'\preceq L(\hat{\rho}^\upsilon(e)\hat{\rho}(f))$}.
\end{eqnarray}
(The $f'\in{\cal F}(\hat{\rho}^\upsilon(a))$ for which
$f'\preceq L(\hat{\rho}^\upsilon(e)\hat{\rho}(f))$ is unique.)
The connectivity matrix\ $\lambda_\mb$ of $\Lambda_\mb$ has coefficients
\begin{equation}
{\lambda_\mb\,}_{(a_1,f_1)(a_2,f_2)} =
|\{e\in E^1(a_1)|x(e)=a_2,f_2\preceq L(\hat{\rho}^\upsilon(e)\hat{\rho}(f_1))\}|.
\end{equation}
Like for $\Sigma$ a path
$((e_1,a_1,f_1),\dots,(e_n,a_n,f_n))$ depends only on $(e_1,\dots,e_n)$
and $(a_n,f_n)$ and may be abbreviated as
$(e_1,\dots,e_n;a_n,f_n)$. Note that such a path satisfies
$R(e_k)=\hat{\rho}(x(e_{k+1}))$ and $R(e_n)=\hat{\rho}(a_n)$.
Therefore the map $\delta:{\cal P}_{\Lambda_\mb}^n(a,f)\to E^n(a)$ given by
\begin{equation} \label{16062}
\delta(e_1,\dots,e_n;a,f) := e_1\hat{\rho}(e_2)\dots\hat{\rho}^{n-1}(e_n)
\end{equation}
is a bijection for fixed $(a,f)$.
Furthermore
let
$\beta_\upsilon:{\cal B}^\upsilon_1\to {\mTxx}_{1}$:
\begin{equation}
\beta_\upsilon(a,f) = a.
\end{equation}
Extending $\beta_\upsilon$ first to edges via $\beta_\upsilon(e,a,f):=(e,a)$ and then
to paths on obtains a map
$\beta_\upsilon:{\cal P}_{\Lambda_\mb}^n\to{\cal P}_{\Sigma}^n$:
\begin{equation}
\beta_\upsilon(e_1,\dots,e_n;a_n,f_n)=(e_1,\dots,e_n;a_n)
\end{equation}
which is not only surjective but also
injective on ${\cal P}_{\Lambda_\mb}^n(a,f)$ for fixed $f$.
In particular $\beta_\upsilon$ extends to a continuous surjective map
from ${\cal P}_{\Lambda_\mb}$ onto ${\cal P}_{\Sigma}$.
\begin{lem} \label{18061}
Let $T$ be a substitution\ tiling with locally invertible substitution\ $\hat{\rho}$.
For each $\omega\in\Omega$ there is a unique
$(e,a)\in\Sigma^{(1)}$
such that $\omega\cdot e\in\mbox{\rm im}\,\rho$.
\end{lem}
{\em Proof:}
By definition of a substitution\ does for any $c\in{\cal M}_{\rm I\!I}$ hold
the alternative:
either $c=c^{-1}$ or $\forall e_i\in E^1(x_i(c)):\:e_1\rho(c) e_2^{-1}
\neq (e_1\rho(c) e_2^{-1})^{-1}$ where $x_1(c)=x(c)$ and $x_2(c)=x(c^{-1})$.
Recall from the proof of Theorem~\ref{25071}
that $\rho(\Omega)=\Omega_{\cal N}$ where
${\cal N}=\Phi({\cal M}_{\rm I\!I})$.
Since for all $d\in\Phi(c)$, $\hat{\rho}(c)$ and $d$ have a common greater element
the above alternative carries over
for any $d\in {\cal N}$ in the form:
either $d=d^{-1}$ or $\forall e_i\in E^1(x_i(d)):\:e_1 d e_2^{-1}
\neq (e_1 d e_2^{-1})^{-1}$.
Since $[\rho]_o(T)$ is locally isomorphic to $T$,
${\cal N}$ has to be regular and there is a $c\in{\cal M}_{\rm I\!I}$ such that
$\omega\cdot c\in \Omega_{{\cal N}}$.
And as ${\cal N}$ is approximating and generating ($T$ has to be non periodic)
$c$ can be written as $c=e d$ where $d\in{\cal N}$ and $e\in E^1(x(d))$.
We claim that $e$ is unique, for if not then
$L(c)=e d{d'}^{-1}e'^{-1}$ and since this is a unit
$d{d'}^{-1}$ must be equal to its inverse
by the above alternative. Hence $e=e'$.
Finally, if $\omega\cdot ed\in\mbox{\rm im}\,\rho$ then also
$\omega\cdot e\in\mbox{\rm im}\,\rho$.
The condition $a\preceq \rho^{-1}(\omega\cdot e)$ determines uniquely
the $a$ for which $e\in E^1(a)$.
\hfill$\Box$\bigskip
Let ${\kappa_\mb}:\Omega\to\Lambda_\mb^{(1)}$:
$\omega\mapsto(e,a,f)$ where $(e,a)$ is
determined by the above lemma and $f\preceq\rho^{\upsilon-1}(\omega\cdot e)$.
Furthermore, the first component of $(e,a)$
shall be used to define
an extension of $\rho^{-1}$ to all of $\Omega$ through
\begin{equation}
\psi(\omega):=\rho^{-1}(\omega\cdot e).
\end{equation}
\begin{lem}
$s({\kappa_\mb}(\psi^{n}(\omega)))=r({\kappa_\mb}(\psi^{n-1}(\omega))$.
\end{lem}
{\em Proof:}
Let ${\kappa_\mb}(\psi^{n}(\omega))=(e_2,a_2,f_2)$ and
${\kappa_\mb}(\psi^{n-1}(\omega))=(e_1,a_1,f_1)$. We have to show that,
first, $x(e_2)=a_1$, and second, that the $f'\in{\cal F}(\hat{\rho}^\upsilon(a_2))$
which satisfies
$f'\preceq L(\hat{\rho}^\upsilon(e_2)\hat{\rho}(f_2))$ equals to $f_1$.
The first follows from the
definitions: $a_1\preceq\rho^{-1}(\psi^{n-1}(\omega)\cdot e_1)=\psi^n(\omega)$
and $L(e_2)\preceq \psi^n(\omega)$.
As for the second, $f_2\preceq\rho^{\upsilon-1}(\psi^n(\omega)\cdot e_2)$ implies that
$\rho(L(\hat{\rho}^{\upsilon-1}(e_2)f_2))\preceq\rho^\upsilon(\psi^n(\omega))=
\rho^{\upsilon-1}(\psi^{n-1}(\omega)\cdot e_1)$. Hence
$f'\preceq\rho^{\upsilon-1}(\psi^{n-1}(\omega)\cdot e_1)$ which is the relation
determining $f_1$.\hfill$\Box$\bigskip
As a consequence we may define a coding which is a map
$Q_\mb:\Omega\to{\cal P}_{\Lambda_\mb}$ where the $n$th edge of $Q_\mb(\omega)$ is
\begin{equation}
Q_\mb(\omega)_n:= {\kappa_\mb}(\psi^{n-1}(\omega)).
\end{equation}
\begin{thm} \label{16063}
Under the conditions of Lemma~\ref{18061} is $Q_\mb:\Omega\to{\cal P}_{\Lambda_\mb}$ a homeomorphism.
\end{thm}
{\em Proof:}
Local invertibility implies that, for $n\geq\upsilon$
\begin{equation} \label{23062}
{\cal V}_\upsilon^n(a,f,e_1,e_2) := \{[\omega\cdot e_1^{-1},e_1\hat{\rho}^{n-\upsilon}(f) e_2^{-1}]|
\omega \in \rho^n(U_a)\cap\rho^{n-\upsilon}(U_f)\}
\end{equation}
is compact and open for all $(a,f)\in{\cal B}^\upsilon_1$ and
$e_i\in E^n(a)$. (In fact, these are ${\cal R}$-sets.)
Let $\xi$ be a path of length $n$ with $r(\xi)=(a,f)$.
$Q_\mb(\omega)\in U_\xi$ if and only if $L(\delta(\xi))\preceq\omega$, $a\in\psi^n(\omega)$,
and $f\preceq \rho^\upsilon(\psi^{n}(\omega))$.
Since $\omega\cdot \delta(\xi) \in
\mbox{\rm im}\,\rho^n$ and $\psi^n(\omega)=\rho^{-n}(\omega\cdot\delta(\xi))$ this means
\begin{equation} \label{18062}
Q_\mb^{-1}(U_\xi)={\cal V}_\upsilon^{|\xi|}(r(\xi),\delta(\xi),\delta(\xi)).
\end{equation}
Hence $Q_\mb$ is continuous. Since
$Q_\mb^{-1}(U_\xi)$ is not empty, any path over $\Lambda_\mb$ is an
accumulation point of $\mbox{\rm im}\, Q_\mb$. As the latter is closed $Q_\mb$ is surjective.
Finally, to prove injectivity,
let $r(Q_\mb(\omega)_k)=(a_k,f_k)$ and
$\tilde{e}_n:=\delta(Q_\mb(\omega)_1\dots Q_\mb(\omega)_n)$.
Then $\hat{\rho}^{n-\upsilon}(f)\preceq\omega\cdot\tilde{e}_n$ and
$$u_n(\omega):=L(\tilde{e}_{n}\hat{\rho}^{n-\upsilon}(f_n))\preceq \omega.$$
Since $L(\tilde{e}_{n}\hat{\rho}^{n}(a_n))\preceq
L(\tilde{e}_{n}\hat{\rho}^{n-\upsilon}(f_n))$
the radius of $u_n(\omega)$ diverges exponentionally with $n$.
Hence $u_n(\omega)$ is an approximation which approximates $\omega$ and is
uniquely determined by $Q_\mb(\omega)$.
Thus $Q_\mb$ is injective.
\hfill$\Box$\bigskip
Theorem~\ref{16063} implies that
$Q_\mb^*:C({\cal P}_{\Lambda_\mb},\mbox{\rm \twlset Z})\to C(\Omega,\mbox{\rm \twlset Z})$: $Q_\mb^*(f)=f\circ Q_\mb^{-1}$ is an
isomorphism of groups. We may extend $Q_\mb$
to a map
$Q_\mb:{\cal S}\to{\cal S}_{\Lambda_\mb}$ by restricting $Q_\mb\timesQ_\mb$ to ${\cal S}$ and
since $T$ is not periodic we may view it as a map $Q_\mb:{\cal R}\to{\cal R}_{\Lambda_\mb}$.
Then alike (\ref{18062})
\begin{equation}
Q_\mb^{-1}({\cal U}_{\xi\xi'}) = {\cal V}^{|\xi|}_\upsilon(r(\xi),e_\xi,e_{\xi'})
\label{08071}
\end{equation}
showing that this map is continuous and surjective, but it is not
injective.
Since
$\eta\circQ_\mb^{-1}=Q_\mb^*\circ\eta$ the image under $Q_\mb^*$
of the subgroup $E_{\Lambda_\mb}$
is contained in $E_{\cal R}$. Hence there is an induced surjective
homomorphism
$[Q_\mb^*]:H({\cal R}_{\Lambda_\mb})\to H({\cal R})$.
Its kernel is $E_{\cal R}/Q_\mb^*(E_{\Lambda_\mb})$.
In particular, since $\lambda_\mb$ is primitive, we get a corrolary
from Theorem~\ref{16063}:
\begin{cor}
Let $T$ be a substitution\ tiling with
primitive locally invertible substitution\ $\hat{\rho}$
which reduces to a decoration of $\mbox{\rm \twlset Z}^d$, $d\leq 3$, and $\mbox{\rm tr}$ be a trace on
$A_T$. Then
\begin{equation}
\mbox{\rm tr}_*(K_0({\cal A}_T))
= \{\tau^{-n}\sum_i\nu_i n_i|n>0,n_i\in\mbox{\rm \twlset Z}\}
\end{equation}
where $\tau$ is the Perron-Frobenius-eigen\-value\ and $\nu$ the left Perron-Frobenius-ei\-gen\-vec\-tor\ of
$\lambda_\mb$ normalized to $\sum_i\nu_i=1$.
\end{cor}
{\em Proof:}
If $\mu$ is the measure on $\Omega$ obtained by restricting the trace
then $\mu\circQ_\mb^{-1}$ is a measure on ${\cal P}_{\Lambda_\mb}$ which is invariant under
${\cal R}_{\Lambda_\mb}$. Using (\ref{06061}) one obtains
$\mbox{\rm tr}_*(K_0({\cal A}_T))
= \mu(Q_\mb^{-1}(C({\cal P}_{\Lambda_\mb},\mbox{\rm \twlset Z})))$ and with (\ref{27071}) the statement.\hfill$\Box$\bigskip
This solves the $K$-theoretical version of the gap labelling for
substitution\ tilings
with primitive locally invertible substitution.
\bigskip
{\em Remark:}
Let us only remark here that
one obtains not only the above embedding of groups by $Q_\mb^*$ but
in fact a unital embedding $i:{\cal A}_{\Lambda_\mb}\rightarrow{\cal A}_T$
of the $AF$-algebra
defined by $\Lambda_\mb$ into
the algebra associated to the tiling
similar to embedding of ${\cal A}_{\Sigma}$ in ${\cal A}_T$
described in \cite{Ke2}. In fact, a glance on (\ref{08071})
tells us that the characteristic functions on
${\cal V}^{|\xi|}_\upsilon(r(\xi),e_\xi,e_{\xi'})$
generate ${\cal A}_{\Lambda_\mb}$ as a subalgebra of ${\cal A}_T$.
The embedding induces an order homomorphism $i_*$ of $K$-groups.
Then $K_0({\cal A}_{\Lambda_\mb})=H({\cal R}_{\Lambda_\mb})$
as ordered group with order unit, and, under the hypothesis
(\ref{08075}),
$i_*:K_0({\cal A}_{\Lambda_\mb})\to H({\cal R})$ coincides with $[Q_\mb^*]$.
However, in case $i_*$ is not surjective, i.e.\ the dimension of the tiling
is bigger than one, our analysis will not determine the order structure on
$K_0({\cal A}_T)$ but only a subcone of the positive cone $K^+_0({\cal A}_T)$.
\subsection{The integer group of coinvariants for substitution\ tilings}
Recall that the integer group of coinvariants is given by
$H({\cal R})=H({\cal R}_{\Lambda_\mb})/\ker [Q_\mb^*]$.
To compute $\ker [Q_\mb^*]=E_{\cal R}/Q_\mb^*(E_{\Lambda_\mb})$ we need to control pattern class es
consisting of two tiles. In analogy to (\ref{30061},\ref{23062})
define\footnote{In principle the $\upsilon$ below could be chosen different from
the one above, but we shall not make use of this generality.}
for $n\geq\upsilon$
\begin{equation}
{\cal B}^\upsilon_{2\neq}:=
\{(A,F)|A\in{\mTxx}_{2,\neq},F\in{\cal F}(\hat{\rho}^\upsilon(A)),\rho^\upsilon(U_A)\cap U_F\neq\emptyset\}.
\end{equation}
\begin{equation} \label{28061}
{\cal V}_\upsilon^n(A,F,e_1,e_2) := \{[\omega\cdot e_1^{-1},e_1\hat{\rho}^{n-\upsilon}(F) e_2^{-1}]|
[\omega,\hat{\rho}^{n-\upsilon}(F)] \in \rho^n(U_A)\cap\rho^{n-\upsilon}(U_F)\}
\end{equation}
where $e_i\in E^{n}(x_i(A))$, $x_1(c)=x(c)$ and $x_2(c)=x(c^{-1})$.
If $e_i$ are units we write ${\cal V}_\upsilon^n(A,F)$ and resp.\
${\cal V}_\upsilon^n(a,f)$ for the above.
It should be clear that
for fixed $e_1,e_2$ the sets ${\cal V}_\upsilon^n(x,y,e_1,e_2)$ are for different
$(x,y)\in{\cal B}^\upsilon_1\cup{\cal B}^\upsilon_{2\neq}$ pairwise disjoint.
$[\,\cdot\,]_{\Lambda_\mb}$ denotes equivalence classes
with respect to $Q_\mb^*(E_{\Lambda_\mb})$. \bigskip
\begin{lem} \label{07075}
Let $T$ be a substitution\ tiling with locally invertible substitution\ and $\Lambda_\mb$ and $Q_\mb$
as above. Then
$E_{\cal R}/Q_\mb^*(E_{\Lambda_\mb})$ is generated by elements of the form
$\left[\eta({\cal V}_\upsilon^n(A,F))\right]_{\Lambda_\mb}$
for $(A,F)\in{\cal B}^\upsilon_{2\neq}$.
\end{lem}
{\em Proof:} Since ${\mTxx}_{2,\neq}$ generates ${\cal M}_{\rm I\!I}\backslash{\mTxx}_{1}$
a generating set for $E_{\cal R}$ is provided by the set of
elements of the form
$\eta(\chi^{}_{c})$ where $c=uA$, $u\in{{\mTxx}^0}$, $u\vdash A\in{\mTxx}_{2,\neq}$.
For such a $c$ let
\begin{eqnarray*}
I_1^{n}(c)&:=&\{(a,f,e_1,e_2)|(a,f)\in{\cal B}^\upsilon_1,e_i\in E^{n}(a)|
\exists c':c, e_1 \hat{\rho}^{n-\upsilon}(f) e_2^{-1}\preceq c'\}\\
I_{2\neq}^{n}(c)&:=&\{(A,F,e_1,e_2)|(A,F)\in{\cal B}^\upsilon_{2\neq},e_i\in
E^{n}(x_i(A))|
\exists c':c, e_1 \hat{\rho}^{n-\upsilon}(F) e_2^{-1}\preceq c'\}
\end{eqnarray*}
Consider $\omega\in U_{L(c)}$ and let $n\geq \upsilon$.
By Lemma~\ref{18061} there is a unique
$(a,f,e_1)$ such that $\omega\in{\cal V}_\upsilon^n(a,f,e_1,e_1)$.
It follows that $e_1^{-1}\vdash c$. Now suppose that
$\exists e_2\in E^n(a)\exists c':e_1e_2^{-1}, c\preceq c'$. Then this $e_2$
is uniquely determined and
$[\omega,c]\in {\cal V}_\upsilon^n(a,f,e_1,e_2)$.
In particular ${\cal V}_\upsilon^n(a,f,e_1,e_2)$ are for $(a,f,e_1,e_2)\in I_1^{n}(c)$
pairwise disjoint.
If the above assumption is not satisfied then, because of the
form $c=uA$, there must be a $(A,F)$ with $a=L(A)$ and $f\preceq L(F)$
such that
$\exists e_2\in E^n(x_2(A))\exists c':e_1\hat{\rho}^{n-\upsilon}(F)e_2^{-1}, c\preceq c'$.
Again $e_2$ is uniquely determined
and $[\omega,c]\in {\cal V}_\upsilon^n(A,F,e_1,e_2)$. This shows that
${\cal V}_\upsilon^n(A,F,e_1,e_2)$ are for $(A,F,e_1,e_2)\in I_{2\neq}^{n}(c)$
pairwise disjoint and hence
\begin{equation} \label{30064}
{\cal U}_c \subset
\dot{\bigcup_{I^n_1(c)}}{\cal V}_\upsilon^n(a,f,e_1,e_2)\:\dot{\cup}\:
\dot{\bigcup_{I_{2\neq}^n(c)}}{\cal V}_\upsilon^n(A,F,e_1,e_2).
\end{equation}
Since for all $(x,y)\in{\cal B}^\upsilon_1\cup{\cal B}^\upsilon_{2\neq}$ and $e_i\in E^{n}(x_i(x))$,
$\mbox{\rm rad}(L(e_1\hat{\rho}^{n-\upsilon}(y)))$
diverges
there is an
$n$ such that
for all $y$ and all $e_i$ either $c\preceq e_1\hat{\rho}^{n-\upsilon}(y)e_2^{-1}$
or $\not\exists c':c, e_1 \hat{\rho}^{n-\upsilon}(y) e_2^{-1}\preceq c'$.
In other words,
either ${\cal U}_c\cap{\cal V}^n_\upsilon(x,y,e_1,e_2)={\cal V}^n_\upsilon(x,y,e_1,e_2)$
or that intersection is empty. Thus the inclusion in (\ref{30064})
is an equality for large enough $n$.
By (\ref{08071}),
${\cal V}_\upsilon^n(a,f,e_1,e_2)\in Q_\mb^{-1}({\cal ASG}({\cal R}_{\Lambda_\mb}))$ so that
$[\eta({\cal V}_\upsilon^n(a,f,e_1,e_2))]_{\Lambda_\mb}=0$. Thus
the first part of the union (\ref{30064}) will not contribute.
As for the second, observe that
\begin{equation}\label{08072}
L({\cal V}_\upsilon^n(A,F,e_1,e_2))=\{\omega\cdot e_1^{-1}|
\omega\in\rho^n(U_{L(A)})\cap\rho^{n-\upsilon}(U_{L(F)})\}
\end{equation}
and
${\cal V}_\upsilon^n(a,f,e,e)=\{\omega\cdot e^{-1}|
\omega\in\rho^n(U_{a})\cap\rho^{n-\upsilon}(U_{f})\}$. As above, if $l$ is big enough
one gets the alternative
\begin{equation}\label{30063}
L({\cal V}_\upsilon^n(A,F,e_1,e_2))\cap {\cal V}^{n+l}_1(a,f,e,e)=\left\{
\begin{array}{l}
L({\cal V}_\upsilon^n(A,F,e_1,e_2)) \\
\emptyset \end{array}\right..
\end{equation}
It follows that (\ref{08072}) (which does not depend on
$e_2$)
can be written as a disjoint union
$$L({\cal V}_\upsilon^n(A,F,e_1,e_2))=\dot{\bigcup_{(a,f,e)\in J^l(A,F,e_1)}}
{\cal V}^{n+l}_1(a,f,e,e)$$
for an appropriate $J^l(A,F,e_1)$, and moreover,
$J^l(A,F,ce_1)$ with $R(c)=L(e_1)$ equals to
$\{(a,f,ce)|(a,f,e)\in J^l(A,F,e_1)\}$. Thus
$[L({\cal V}_\upsilon^n(A,F,e_1,e_2))]_{\Lambda_\mb}=[L({\cal V}_\upsilon^n(A,F))]_{\Lambda_\mb}$ and we
end up with
\begin{equation}
[\eta({\cal U}_c)]_{\Lambda_\mb} = \sum_{(A,F,e_1,e_2)\in
I_{2\neq}^n(c)}[\eta(V^n_\upsilon(A,F))]_{\Lambda_\mb}.
\end{equation}
Since $\eta(V^n_\upsilon(A,F))\in E_{\cal R}$
the lemma is proven.\hfill$\Box$
\begin{lem}
Let $(a,f)\in{\cal B}^\upsilon_1$, $(A,F)\in{\cal B}^\upsilon_{2\neq}$, $e\in E^{l}(a)$,
$n\geq \upsilon$,
and $l\geq \upsilon$ large enough for the alternative
(\ref{30063}) to hold. Then $L({\cal V}_\upsilon^n(A,F))
\cap {\cal V}_\upsilon^{n+l}(a,f,\hat{\rho}^n(e),\hat{\rho}^n(e))\neq\emptyset$ whenever
$L(\hat{\rho}^\upsilon(e^{-1})F)\preceq\hat{\rho}^l(f)$ and $L(e^{-1}A)\preceq\hat{\rho}^{l-\upsilon}(f)$.
\end{lem}
{\em Proof:}
$$ L({\cal V}_\upsilon^n(A,F))
\cap {\cal V}_\upsilon^{n+l}(a,f,\hat{\rho}^n(e),\hat{\rho}^n(e))\neq\emptyset
$$
is for $n\geq \upsilon$ equivalent to
$$
\{\omega\cdot \hat{\rho}^\upsilon(e) |\omega\in\rho^\upsilon(U_{L(A)})\cap U_{L(F)}\}
\cap \rho^{l+\upsilon}(U_a)\cap \rho^l(U_f)
\neq\emptyset.
$$
That the latter implies
$L(\hat{\rho}^\upsilon(e^{-1})F)\preceq\hat{\rho}^l(f)$ and $L(e^{-1}A)\preceq\hat{\rho}^{l-\upsilon}(f)$
is clear provided $l$ is large enough.
Now let $(a,f)\in{\cal B}^\upsilon_1$, $(A,F)\in{\cal B}^\upsilon_{2\neq}$, and
$\omega\in \rho^{l+\upsilon}(U_a)\cap \rho^l(U_f)$. Then
$L(\hat{\rho}^\upsilon(e^{-1})F)\preceq \hat{\rho}^l(f)$ implies
$L(F)\preceq \omega\cdot\hat{\rho}^\upsilon(e^{-1})\in \mbox{\rm im}\,\rho^\upsilon$, and
$L(e^{-1}A)\preceq\hat{\rho}^{l-\upsilon}(f)$ implies
$L(A)\preceq \rho^{-\upsilon}(\omega\cdot\hat{\rho}^\upsilon(e^{-1}))$. \hfill$\Box$\bigskip
Define the $|{\cal B}^\upsilon_1|\times |{\cal B}^\upsilon_{2\neq}|$ matrices
with coefficients $K^{(l,\mb)}_{(a,f)(A,F)}$ resp.\ $\KK^{(l,\mb)}_{(a,f)(A,F)}$
through
\begin{equation} \label{07021}
K^{(l,\mb)}_{(a,f)(A,F)}:= |\{e\in E^{l}(a)|
L(\hat{\rho}^\upsilon(e^{-1})F)\preceq\hat{\rho}^l(f), L(e^{-1}A)\preceq\hat{\rho}^{l-\upsilon}(f)
\}|.
\end{equation}
with $l\geq \upsilon$ and
\begin{equation}
\KK^{(l,\mb)}_{(a,f)(A,F)}:=K^{(l,\mb)}_{(a,f)(A,F)}-K^{(l,\mb)}_{(a,f)(A^{-1},F^{-1})}.
\end{equation}
\begin{thm}
Let $T$ be a substitution\ tiling with locally invertible substitution\ and
$\lambda_\mb$ and $\KK^{(l,\mb)}$ as above.
If $l$ is large enough so that the alternative (\ref{30063}) holds then,
with $k=|{\cal B}^\upsilon_1|$,
\begin{equation}
H({\cal R})\cong \{x\in \mbox{\rm \twlset R}^k|\exists n\geq 0:\lambda_\mb^n(x)\in\mbox{\rm \twlset Z}^k\}/
\{x\in \mbox{\rm \twlset R}^k|\exists n\geq 0:\lambda_\mb^n(x)\in\mbox{\rm im}\,\KK^{(l,\mb)}\}.
\end{equation}
\end{thm}
{\em Proof:}
By the last lemma
$$
[\chi^{}_{L({\cal V}_\upsilon^n(A,F))}]_{\Lambda_\mb} = \sum_{(a,f)\in{\cal B}^\upsilon_1}
[\chi^{}_{{\cal V}_\upsilon^{n+l}(a,f)}]_{\Lambda_\mb} K^{(l,\mb)}_{(a,f)(A,F)}
$$
provided alternative (\ref{30063}) holds, and hence
\begin{equation} \label{07071}
[\eta({{\cal V}_\upsilon^n(A,F)})]_{\Lambda_\mb} = \sum_{(a,f)\in{\cal B}^\upsilon_1}
[\chi^{}_{{\cal V}_\upsilon^{n+l}(a,f)}]_{\Lambda_\mb} \KK^{(l,\mb)}_{(a,f)(A,F)}.
\end{equation}
As we already saw,
$[\chi^{}_{{\cal V}_\upsilon^{n}(a,f)}]_{\Lambda_\mb}=[Q_\mb^*(\chi^{}_\xi)]_{\Lambda_\mb}$ with $|\xi|=n$
and $r(\xi)=(a,f)$ so that these elements generate
$H({\cal R}_{\Lambda_\mb})$.
Moreover
\begin{equation} \label{07022}
[\chi^{}_{{\cal V}_\upsilon^{n}(a,f)}]_{\Lambda_\mb} = \sum_{(a',f')}
[\chi^{}_{{\cal V}_\upsilon^{n+1}(a,f)}]_{\Lambda_\mb} {\lambda_\mb\,}_{(a',f')(a,f)}
\end{equation}
so that
\begin{equation} \label{30062}
{\cal L}^{(l+n,\upsilon)}=\lambda_\mb^n \KK^{(l,\mb)}.
\end{equation}
Recall (\ref{17061}) that
$$H({\cal R}_{\Lambda_\mb}) = \{x\in \mbox{\rm \twlset R}^k|\exists n\geq 0:\lambda_\mb^n(x)\in\mbox{\rm \twlset Z}^k\}/
\{x\in \mbox{\rm \twlset R}^k|\exists n\geq 0:\lambda_\mb^n(x)=0\}.$$
The image of $E_{\cal R}/Q_\mb^*(E_{\Lambda_\mb})$ in $H({\cal R}_{\Lambda_\mb})$
is by Lemma~\ref{07075} and (\ref{07071},\ref{30062})
generated by the images of
$j_n\circ\KK^{(l,\mb)}=\pi\circ\lambda_\mb^{-n}\circ\KK^{(l,\mb)}$.
Now
$\pi(x)\in\mbox{\rm im}\, j_n\circ\KK^{(l,\mb)}$ whenever
$\lambda_\mb^n(x)\in\mbox{\rm im}\,\KK^{(l,\mb)}$ for some
representative $x$ of $\pi(x)$. \hfill$\Box$\bigskip
Suppose that $\Omega$ carries an ${\cal R}$-invariant (normalized) measure and that the
substitution\ is primitive. Then this measure is unique (and the tiling minimal
\cite{Ke2}) and the order structure of
$H({\cal R})$ can be expressed as
$[x]>0$ for $[x]\in H({\cal R})$ if and only if $\sum_i\nu_ix_i>0$
where $\nu$ is the left-Perron-Frobenius-ei\-gen\-vec\-tor\ of $\lambda_\mb$. In particular the latter
inequality is independent of the chosen representative
(the elements of $\ker[Q_\mb^*]$ are infinitesimal).
\subsubsection{Simplifications}
The determination of $\KK^{(l,\mb)}$ can be quite cumbersome.
But simplifications occurr under certain circumstances. \medskip
1) In case that for all $(a,f)\in{\cal B}^\upsilon_1$, $a$ is uniquely determined by $f$
(we write it as $a(f)$) $L(\hat{\rho}^\upsilon(e^{-1})F)\preceq \hat{\rho}^l(f)$ implies
$L(e^{-1}A(F))\preceq\hat{\rho}^{l-\upsilon}(f)$.
Thus $K^{(l,\mb)}$ simplifies to
\begin{equation}
K^{(l,\mb)}_{fF} :=
K^{(l,\mb)}_{(a,f)(A,F)} =
|\{e\in E^{l}(a(f))|L(\hat{\rho}^\upsilon(e^{-1})F)\preceq \hat{\rho}^l(f)\}|.
\end{equation}
This is for instance the case if $\upsilon=0$.
2) In some sense the other extreme is that the $f$ is determined by $a$.
This is case if the substitution\ forces its border.
It has been partly analyzed in \cite{Ke2}. In the present notation it
is expressed as follows:
\begin{df}
A locally invertible substitution\ forces its border if there is a $\upsilon$
such that $\beta_\upsilon:{\cal B}^\upsilon_1\to{\mTxx}_{1}$ is a bijection.
\end{df}
Not only the graph simplifies enormously, in that it coincides with the
improper graph, but also we may take $l=\upsilon$.
Moreover $L(e^{-1}A)\preceq\hat{\rho}^{l-\upsilon}(f(a))$
implies $L(\hat{\rho}^\upsilon(e^{-1})F(A))\preceq\hat{\rho}^l(f(a))$ where we wrote
$f(a)$ for the $f$ determined by $a$. Hence for substitution s which force
their border
\begin{equation}
K^{(\upsilon,\upsilon)}_{aA}:= K^{(\upsilon,\upsilon)}_{(a,f)(A,F)} =
|\{e\in E^{\upsilon}(a)|L(e^{-1}A)\preceq f(a)\}|.
\end{equation}
Note that a substitution\ can force its border only for $\upsilon\geq 1$, because otherwise
any pointed tiling would be determined by its pointed tile and hence
periodic which contradicts local invertibility. \medskip
As for the determination of $H({\cal R})$ we have:\medskip
3) If $\lambda_\mb$ is invertible over $\mbox{\rm \twlset Z}$ then
$H({\cal R})=\mbox{\rm \twlset Z}^k/\langle\bigcup_n
\lambda_\mb^{-n}\mbox{\rm im}\,\KK^{(l,\mb)}\rangle$.
\subsection{Examples}
The above machinerie has been designed to tackle higher dimensional
tilings. We therefore will present the computation of the coinvariants
associated to the Penrose tilings.
But it is also applicable to one dimensional substitution s
in the sense of \cite{Q}.
To allow the reader a comparison with technics used elsewhere to obtain
the integer group of coinvariants, c.f.\ \cite{For,Hos}, we present the
Thue Morse substitution\ as an example.
\subsubsection{The Thue Morse substitution}
The Thue Morse substitution\ $\varrho$ is defined on the two letter alphabet $\{a,b\}$
by
\begin{eqnarray}
a &\mapsto& ab\\
b &\mapsto& ba
\end{eqnarray}
and extended to words as
$\varrho(a_1\dots a_k)= \varrho(a_1)\dots \varrho(a_k)$.
It may be viewed (like any other one dimensional substitution of the kind
in \cite{Q}) as a substitution\ in the algebraic sense of the tiling which is
a fixed point under $\varrho$:
Consider the sequence
over $\mbox{\rm \twlset Z}^{\geq 0}$ with values in $\{a,b\}$ given by
$\varrho^\infty(a)$ and
complete it to a sequence over $\mbox{\rm \twlset Z}$ by reflection
(i.e.\ by $\varrho^{2\infty}(a)$ to the left).
This represents
a one dimensional pointed tiling the pointed tile (letter) being
the one on $0$, i.e.\
the first one of (the right) $\varrho^\infty(a)$.
The (finite) words appearing in $\varrho^\infty(a)$
with two letters chosen are doubly pointed pattern class es.\footnote{
If we do not have a geometric interpretation of the tilings as
sequences of decorated intervals we have to restrict to pattern class es which
are connected.}
We indicate the first by
a grave and the second by an acute, and if both coincide by a check.
(An example of a multiplication is
$(\lp{a}\fp{b})(b\fp{a}b\lp{b})=(ba\fp{b}\lp{b}))$.
A substitution\ as a homomorphism of the almost-groupoid\ of doubly pointed pattern class es is then given by
$\hat{\rho}(w_1\fp{a}_iw_2\lp{a}_jw_3)=\varrho(w_1)\fp{a}_{i_1}a_{i_2}
\varrho(w_2)\lp{a}_{j_1}a_{j_2}\varrho(w_3)$ where we have used the writing
$\varrho(a_i)=a_{i_1}a_{i_2}$.
The substitution\ is locally invertible but does not force its border and
there is no advantage in using large $\upsilon$. We therefore take $\upsilon=0$.
The connectivity matrix\ of $\Sigma_0$ is
$${\lambda_0\,}_{f_1f_2} =\mbox{number of $e\in E^1(a(f_1))$ with
$f_2\preceq L(e\check{\rho}(f_1))$}. $$
We have
$${\cal B}^0_1=\{a\bp{a}b,b\bp{a}a,b\bp{a}b,b\bp{b}a,a\bp{b}b,a\bp{b}a\}$$
$$\hat{\rho}({\cal B}^0_1)=\{ab\bp{a}bba,ba\bp{a}bab,ba\bp{a}bba,
ba\bp{b}aab,ab\bp{b}aba,ab\bp{b}aab\}$$
$$E^1(a)=\{\bp{a}b,\lp{a}\fp{b}\}\quad E^1(b)=\{\bp{b}a,\lp{b}\fp{a}\}$$
Taking the elements of ${\cal B}^0_1$ and $\hat{\rho}({\cal B}^0_1)$ in the above order one
obtains
$$\lambda_0 =
\left(\begin{array}{cccccc}
0 & 0 &1 &0 &1 &0 \\
1 & 0 &0 &0 &0 &1 \\
1 & 0 &0 &0 &1 &0 \\
0 & 1 &0 &0 &0 &1 \\
0 & 0 &1 &1 &0 &0 \\
0 & 1 &0 &1 &0 &0
\end{array}\right).
$$
The Perron-Frobenius-eigen\-value\ is $2$ and the normalized left-Perron-Frobenius-ei\-gen\-vec\-tor\ is
$\nu=\frac{1}{6}(1,1,1,1,1,1)$. Hence
\begin{equation}
\mu(H({\cal R})) = \frac{1}{3}\mbox{\rm \twlset Z}[\frac{1}{2}].
\end{equation}
To obtain the full group we may split up
${\cal B}^0_{2\neq}={\cal B}^0_{2<}\cup ({\cal B}^0_{2<})^{-1}$ with
$${\cal B}^0_{2<}=\{a\fp{a}\lp{b}a,a\fp{a}\lp{b}b,b\fp{a}\lp{b}a,b\fp{a}\lp{b}b,
b\fp{a}\lp{a}b,b\fp{b}\lp{a}b,b\fp{b}\lp{a}a,a\fp{b}\lp{a}b,a\fp{b}\lp{a}a,
a\fp{b}\lp{b}a\}.$$
Thus (\ref{30063}) holds for $l=1$, i.e.\ we have to determine
$$K^{(1,0)}_{fF} =\mbox{number of $e\in E^1(a(f))$ with
$L(F)\preceq L(e\hat{\rho}(f))$}. $$
Let $ {\cal L}^{(1,0)}_{<}$ be the restriction of ${\cal L}^{(1,0)}$ to indices
$fF$ with $F\in{\cal B}^0_{2<}$. Then, again with respect to the above order,
$$ {\cal L}^{(1,0)}_{<}=
\left(\begin{array}{cccccccccc}
0 & 0 &0 &0 &0 &0 &0 &-1 &0 &1\\
0 & 0 &0 &0 &-1 &0 &0 &1 &0 &0\\
0 & 0 &0 &0 &-1 &0 &0 &0 &0 &1\\
0 & 0 &-1 &0 &1 &0 &0 &0 &0 &0\\
0 & 0 &1 &0 &0 &0 &0 &0 &0 &-1\\
0 & 0 &0 &0 &1 &0 &0 &0 &0 &-1
\end{array}\right).
$$
And, since ${\cal L}^{(1,0)}_{fF^{-1}} = - {\cal L}^{(1,0)}_{fF}$,
$\mbox{\rm im}\,{\cal L}^{(1,0)}=\mbox{\rm im}\,{\cal L}^{(1,0)}_{<}\cong\mbox{\rm \twlset Z}^3$.
$\lambda_0{\cal L}^{(1,0)}_{<}$ is up to a permutation of the columns
${\cal L}^{(1,0)}_{<}$, i.e.\ $\lambda_0$ preserves $\mbox{\rm im}\,{\cal L}^{(1,0)}$.
In fact, $\lambda_0$ is diagonalizable, it has eigenvalues $2,1,0,-1$,
$-1$ occurring with multiplicity $3$, and $\mbox{\rm im}\,{\cal L}^{(1,0)}$ is spanned by
the (right) eigenvector to eigenvalue $1$ together with a two dimensional
subspace of the eigenspace to eigenvalue $-1$.
A system of (right) eigenvectors is given by
$$
\left(\begin{array}{c} 1\\1\\1\\1\\1\\1 \end{array}\right)\quad
\left(\begin{array}{c} -1\\0\\-1\\1\\0\\1 \end{array}\right)\quad
\left(\begin{array}{c} -1\\-1\\-1\\1\\1\\1 \end{array}\right)\quad
\left(\begin{array}{c} 0\\-1\\-1\\0\\1\\1 \end{array}\right)\quad
\left(\begin{array}{c} -1\\1\\0\\-1\\1\\0 \end{array}\right)\quad
\left(\begin{array}{c} 1\\-1\\-1\\1\\0\\0 \end{array}\right).
$$
It follows that
\begin{equation}
H({\cal R})\cong\mbox{\rm \twlset Z}[\frac{1}{2}]\oplus\mbox{\rm \twlset Z}.
\end{equation}
The elements of the second summand are infinitesimal
since the pairing between the left Perron-Frobenius-ei\-gen\-vec\-tor\ $\nu$ and all vectors from the
(right) eigenspace to eigenvalue $-1$ is zero. Thus
the positive cone is $\mbox{\rm \twlset Z}[\frac{1}{2}]^+$.
\subsubsection{Penrose tilings} \label{03081}
There are several well known
variants of tilings which are called Penrose tilings \cite{Pen,GrSh}
and which are a priori to be distinguished as they
lead to non-isomorphic groupoids. But they may be transformed
into each other by purely local manipulations
which implies that one can find maps satisfying the conditions
of Lemma~\ref{27061} and leading to isomorphisms between reductions of the
corresponding groupoids. Since any such tiling is minimal
all reductions lead to the same ordered integer group of coinvariants
differing possibly in the order unit.
The version which is most suitable for our purposes is the one which has
triangles as tile s, cf.\ Figure~7.
The triangles are decorated (with a little circle) to break the mirror
symmetry. There are $40$ pattern class es of them.
The orientational symmetry of a tiling (or its class)
is the largest subgroup of $O(d)$,
acting on a pointed tiling being identified with a representative in
$\mbox{\rm \twlset R}^d$ in the obvious way, which leaves the hull invariant.
The orientational symmetry of a Penrose tiling by triangles
possesses $20$ elements. It is
generated by a rotation around $\frac{\pi}{5}$ together with a
mirror reflection at a boundary line of a triangle \cite{Ke2}.
We denote by ${\cal M}_{\rm I\!I}$ resp.\ ${\cal R}$ the almost-groupoid\ resp.\
groupoid associated to these tilings.
The well known deflation of these tilings
$\rho_t$ with $t=\frac{1+\sqrt{5}}{2}$ is displayed below.
Since it is covariant with respect to the orientational symmetry
it suffices to give it for one orientation only.
\epsffile[0 0 430 170]{dFig1.ps}
A replacement $\rho_t(a)$ taking exactly the same space as the tile $a$ the
relative position measured with respect to the centers of gravity is $0$.
Choosing pointed tiles for $\rho_t(a)$ we obtain a substitution\ $\hat{\rho}$.
Those tilings which have an exact five-fold
symmetry are invariant under $\hat{\rho}^4$.
The substitution\ is primitive, locally invertible,
and forces its border with $\upsilon=4$ \cite{Ke2}.
Hence $\Lambda_4=\Sigma$ and $\lambda_4=\sigma$, the substitution\ matrix of
$\hat{\rho}$.
Since $\sigma$ is invertible over the integers $H({\cal R}_\Sigma)\cong\mbox{\rm \twlset Z}^{40}$.
To simplify the computation of $E_{\cal R}/Q_4^*(E_\Sigma)$ we make
use of the symmetry properties of the tiling and in particular of the fact
that the boundaries of substitute s $\hat{\rho}^4(a)$ are local mirror axes so that
the pattern classes of those $A\in{\mTxx}_{2,\neq}$ which cross the boundaries of
$4$-fold substitute s are always mirror symmetric \cite{Ke2}, c.f.\ below where
the boundaries are indicated through fatter lines.
\epsffile[0 0 430 190]{dFig7.ps}
Let $\alpha$ denote the direction of a boundary line of a tile resp.\ a substitute.
There are
$10$ different ones and we order them anti-clockwise identifying them with
$\{0,\cdots,9\}$. Saying that a substitute\ has boundary $\alpha$ if it has a boundary
with that direction we define the $40\times 40$ matrices with entries
\begin{equation}
N^\alpha_{ab} := \mbox{number of $a$'s at boundary $\alpha$ of $\hat{\rho}^4(b)$} .
\end{equation}
In particular $N^\alpha_{ab}=0$ in case $a$ or $\rho(b)$ do not have
boundary $\alpha$. Let $\alpha(a)$ be the mirror
image of $a$ with respect to the mirror axis $\alpha$ and define
\begin{equation}
{\cal D}^\alpha_{ab} = N^\alpha_{ab}-N^\alpha_{a \alpha(b)}.
\end{equation}
Then
\begin{equation}
{\cal L}^{(4,4)}_{a A} = \left\{
\begin{array}{cl}
{\cal D}^\alpha_{ab} & \mbox{if } |A|=b \alpha(b)\mbox{ and } x(A)=b \\
-{\cal D}^\alpha_{ab} & \mbox{if } |A|=b \alpha(b)\mbox{ and }x(A)=\alpha(b)\\
0 & \mbox{else}
\end{array}
\right.
\end{equation}
$|A|=b\alpha(b)$ indicating that the pattern class of $A$
is composed of $b$ and $\alpha(b)$
in such a way that the common boundary is the symmetry axis.
Hence
$\mbox{\rm im}\, {\cal L}^{(4,4)}=\erz{\mbox{\rm im}\,{\cal D}^\alpha,\alpha=0,\dots,9}$.
${\cal D}^\alpha$ is related to ${\cal D}^{0}$ by symmetry, i.e.\
${\cal D}^\alpha=R^{-\alpha}{\cal D}^{0} R^\alpha$,
$R$ being the matrix which acts as a rotation around $\frac{\pi}{5}$.
To be very explicit let us use a basis
$\{e_{10 k+\alpha}\}_{0\leq k \leq 3, 0\leq\alpha\leq 9}$ of
$H({\cal R}_\Sigma)$ with
$e_{10 k+\alpha}=[\chi^{}_{U_{a_{10 k+\alpha}}}]_\Sigma$
where $a_{10 k+\alpha}$ corresponds to the pattern class\ of the
triangle in Figure~7.k
rotated around an angle of $\frac{\alpha\pi}{5}$.
\epsffile[0 0 430 100]{dFig2.ps}
In terms of the rotation matrix $\omega$, which has
entries $\omega_{\alpha\beta}=\delta_{\alpha-\beta,1\;mod\;10}$,
$R$ and the substitution\ matrix $\sigma$ are given by
$$
R = \left(
\begin{array}{cccc}
\omega & 0 & 0 & 0 \\
0 & \omega & 0 & 0 \\
0 & 0 & \omega & 0 \\
0 & 0 & 0 & \omega
\end{array}
\right), \quad
\sigma = \left(
\begin{array}{cccc}
\omega^4 & \omega^0 & 0 & \omega^6 \\
\omega^0 & \omega^6 & \omega^4 & 0 \\
\omega^3 & 0 & \omega^7 & 0 \\
0 & \omega^7 & 0 & \omega^3
\end{array}
\right).
$$
The matrix $N^{0}$ may be read of from Figure~8
(and completed by symmetry); it is given below (\ref{19011}).
Moreover ${\cal D}^0=N^0-N^0S$ where $S$ implements the reflection at
$\alpha=0$, explicitly, with
$s_{\alpha\beta}=\delta_{\alpha+\beta,5\;mod\;10}$ (counting rows and
columns form $0$ to $9$)
$$
S = \left(
\begin{array}{cccc}
0 & s & 0 & 0 \\
s & 0 & 0 & 0 \\
0 & 0 & 0 & s \\
0 & 0 & s & 0
\end{array}
\right).
$$
It turns out that $\mbox{\rm im}\,{\cal L}^{(4,4)}$ is generated by
the orbit under $R$ of (the transpose of) the four vectors
$$\begin{array}{c}
v_1=(0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,
\!0,\!0,\!0,\!0,\!0,\!
0,\!0,\!1,\!0,\!0,\!0,\!0,\!-1,\!0,\!0,\!0,\!0,
\!0,\!0,\!0,\!0)\\
v_2=(0,\!0,\!0,\!0,\!0,\!0,\!1,\!0,\!0,\!0,\!0,\!-1,\!0,\!0,\!0,
\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!
0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0)\\
w_1=(0,\!0,\!0,\!0,\!0,\!0,\!0,\!1,\!0,\!0, -1,\!0,\!0,\!0,\!0,
\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!
0,\!0,\!0,\!0,\!1,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0, -1)\\
w_2=(0,\!0,\!0,\!0,\!0,\!1,\!0,\!0,\!0,\!0,\!0,\!0, -1,\!0,
\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!
0,\!0,\!0,\!0,\!0, -1,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!0,\!1,\!0)
\end{array}
$$
but $8$ of the $40$ vectors thus obtained are linearly dependent of the others.
Moreover $\mbox{\rm im}\,{\cal L}^{(4,4)}$ is invariant
under $\sigma$. Dividing it out yields no torsion so that we obtain
\begin{equation}\label{31071}
H({\cal R})\cong\mbox{\rm \twlset Z}^8.
\end{equation}
To obtain the order structure we look for the smallest subgroup $I$ of
$\mbox{\rm \twlset Z}^{40}$
which is invariant under $\sigma$ and spans a real vector space containing
the right Perron-Frobenius-ei\-gen\-vec\-tor\ of $\sigma$. Then the positive cone
of $\mbox{\rm \twlset Z}^{40}$ is given by those $x\in I$ which pair with the left-Perron-Frobenius-ei\-gen\-vec\-tor\
$\nu$ of $\sigma$ to $\nu x >0$.
All other elements must pair with $\nu$ to zero and are thus infinitesimal.
Clearly $I$ is spanned by (the transpose of)
$$ x_1 = (\overbrace{1,\dots,1}^{20},\overbrace{0,\dots,0}^{20}),
\quad x_2 = (\overbrace{0,\dots,0}^{20},\overbrace{1,\dots,1}^{20}) $$
and $\nu$ is given by $\nu = \frac{2-t}{20}(tx_1+x_2)^T$ (in its normalized
form).
Thus $I\cap \mbox{\rm im}\,{\cal L}^{(4,4)}=\emptyset$ and
\begin{equation}\label{03083}
H({\cal R})\cong I \oplus \mbox{\rm \twlset Z}^6
\end{equation}
the elements of the second summand being infinitesimal.
The range of the unique state on $H({\cal R})$, which
of course coincides with that on $H({\cal R}_\Sigma)$, is
\begin{equation} \label{03082}
\mu(H({\cal R}))=\frac{1}{20}(\mbox{\rm \twlset Z} + t\mbox{\rm \twlset Z}).
\end{equation}
This concludes the computation of the integer group of coinvariants and
the range of its state for the Penrose tilings.
But it may be instructive to
look at result (\ref{31071}) from a different point of view.
Let $H_1$ be the sublattice of $\mbox{\rm \twlset Z}^{40}$
which is generated by $\{R^\alpha v_i\}_{i=1,2;\,\alpha=0,1\cdots,9}$.
$H({\cal R}_\Sigma)/H_1=\mbox{\rm \twlset Z}^{20}$ has a basis with natural geometrical
interpretation.
It is formed by the classes of
characteristic functions on the pattern class es of rhombi which
are always formed by either
two smaller triangles or two larger triangles.
This indicates that
one should look at the Penrose tilings by rhombi.
Such a tiling, which we denote by $T_{Rh}$, has only $20$ pattern class es
of tile s.
That the groupoid
associated to
the Penrose tilings by rhombi is isomorphic to a reduction of ${\cal R}$ may be
seen as follows:
Let ${\cal C}\subset{\mTxx}_{1}$ consist of the classes of triangles
that are obtained from those of Figure~7.1 and 7.3 by a rotation.
Let ${Rh}\subset{\cal M}_{\rm I\!I}$ be the almost-groupoid\ given by elements of the form
$aca'$ where $a,a'\in{\cal C}$ and $c\in{\cal M}_{\rm I\!I}$ is such that
all its triangles pair to rhombi. Clearly ${Rh}$ is approximating generating
and regular.
Now deleting the diagonal
which coincides with the base of the two touching triangles
yields an isomorphism
of almost-groupoid s ${Rh}\to\mtxx{T_{Rh}}$ which satifies the growth condition with
$t=1$. Hence
the groupoid associated to the rhombus version is isomorphic to ${\cal R}_{Rh}$.
The largest subgroup of $O(2)$ leaving $\Omega_{Rh}$ invariant
consists only of the rotations which are multiples of $\frac{\pi}{5}$
since the mirror image of a $\omega\in\Omega_{Rh}$ lies in $\Omega\backslash\Omega_{Rh}$.
However, to our knowledge there is no substitution\ for Penrose
tilings by rhombi which is covariant even under this
reduced orientational symmetry.
But there are non covariant ones,
namely $10$ of them, the deflation
corresponding to $\hat{\rho}_0$ being given in Figure~9
and $\hat{\rho}_\alpha$
being obtained from $\hat{\rho}_0$ just by rotation of the whole figure
around $\frac{\alpha\pi}{5}$, $\alpha=0,\cdots,9$.
(The relative positions are indicated by a cross.)
That all these
substitution s are primitive locally invertible and force their border\ carries over
from $\hat{\rho}$.
It turns out that
\begin{equation}
\ker\sigma_\alpha = \erz{\{[R^\alpha w_i]_{H_1}\}_{i=1,2}}
\end{equation}
where $\sigma_\alpha$ is the substitution\ matrix of $\hat{\rho}_\alpha$ and $[\,\cdot\,]_{H_1}$
denotes the classes in $H({\cal R}_\Sigma)/H_1$.
Since the restriction of $\sigma_\alpha$ to its image is an automorphism and
$\ker\sigma_0=\ker\sigma^2_0$ we have
\begin{equation}
H({\cal R}_{\Sigma_\alpha})
\cong H({\cal R}_\Sigma)/\erz{\{H_1,R^\alpha w_i,i=1,2\}}
\end{equation}
$\Sigma_\alpha$ denoting the graph having $\sigma_\alpha$ as connectivity
matrix.
Any of the substitution s $\hat{\rho}_\alpha$ leads to a homeomorphism $Q_\alpha$ between
$\Omega_{Rh}$ and the path space ${\cal P}_{\Sigma_\alpha}$ and to a
surjective homomorphism $[Q^*_\alpha]:H({\cal R}_{\Sigma_\alpha})\rightarrow H({\cal R})$.
Let
$\pi_\alpha:H({\cal R}_\Sigma)\to H({\cal R}_{\Sigma_\alpha})$ be the natural projection.
Then $[Q_4^*]=[Q_\alpha^*]\circ \pi_\alpha$ for all $\alpha$ and therefore
\begin{equation}
\ker [Q_4^*] \supset\erz{\{H_1,R^\alpha w_i,i=1,2;\,\alpha=1,\cdots,9\}}\cong\mbox{\rm \twlset Z}^{32}.
\end{equation}
This shows independently that
$H({\cal R})\subset\mbox{\rm \twlset Z}^{8}$ but not the opposite inclusion.
A computation of $\mbox{im}\,{\cal L}^{(4,4)}$ for e.g.\ $\rho_0$
would have been more complicated due to the lack of symmetry.
Connes associates to the Penrose tilings yet another graph, the folded
$A_4$ Coxeter graph, making use
of a coding of a tiling by $0,1$ sequences obeying the condition that
no consecutive $1$'s can appear \cite{Cone}. But this coding, which was found
by Robinson \cite{GrSh}, does not distinguish between a tiling and
its image under an element of the orientational symmetry. In fact, the
coding yields a homeomorphism between the hull modulo the orientational
symmetry and the path space of the folded $A_4$ graph, with the effect that
the groupoid arising is that given by cofinality. In other words,
the tilings are considered
as equivalence classes under translations {\em and} rotations and reflections.
One then obtains an $AF$-algebra as groupoid-$C^*$-algebra\
whose $K_1$-group is trivial and whose $K_0$-group coincides with the
integer group of coinvariants of the groupoid
and may be identified with the group $I$ in (\ref{03083})
as ordered group. Although the range of the tracial state on it
coincides
with (\ref{03082}) up to the order unit (the factor $\frac{1}{20}$
does not appear) it is a priori not clear that it predicts the right
gap labels since the $AF$-algebra does not contain the
discrete Laplacian.
\subsection*{Concluding Questions}
\addcontentsline{toc}{section}{\bf Concluding Questions}
We have computed the integer group of coinvariants and its order but
even in two dimensions and under the assumption that
$K_0({\cal A}_T)=H({\cal R})\oplus \mbox{\rm \twlset Z}$ it is not clear what the order structure
on the $K_0$-group is.\medskip
We did not mention groupoid cohomology but at least for
tilings which reduce to decorations the integer group of coinvariants
is a cohomology group of the groupoid ${\cal R}$ with coefficients in $\mbox{\rm \twlset Z}$.
The result of \cite{FoHu} on the connection between $K$-theory and
group cohomology
is easily seen to generalize to this situation since the stability of
$K$-theory under taking stably isomorphic algebras
is mirrored by
the stability of
(continuous) groupoid cohomology under taking
(continuously) similar groupoids. It is therefore tempting to
believe that $K$-theory of algebras associated to tilings is always related to
the groupoid cohomology with coefficients in $\mbox{\rm \twlset Z}$ of the associated
groupoid in a way like in \cite{FoHu}.
\newpage
\begin{equation} \label{19011}
N^{0} =
\left(\begin{array}{c}
3\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;1\;0\;2\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;1\;0\;0\;1\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 0\;1\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;2\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;1\;0
\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;1\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 1\;0\;0\;0\;1\;0\;1\;0\;0\;0\;1\;0\;0\;0\;0\;0\;1\;0\;0
\;0\;1\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;3\;0\;0\;0\;0\;0\;2\;0\;0\;0\;1
\;0\;0\;0\;1\;0\;0\;1\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\\
0\;0\;0\;0\;1\;0\;1\;0\;0\;0\;2\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;
0\;0\;0\;0\\ 0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;1\;0\;0\;0\;1\;0\;0\;0\;1\;0\;1\;0\;0\;0\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 1\;0\;0\;0\;2\;0\;1\;0\;0\;0\;3\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;1\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;
0\;0\;0\;1\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;1\;0\;0\;0\;1\;0\;1\;0\;0\;0\;1
\;0\;0\;0\;1\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
2\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0
\;1\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;1\;0\;0\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;1\;0\;0\;0\;1\;0\;0\;0\;2\;0\;0\;0\;0\;0\;3\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;1\;0\\ 1\;0\;0\;0\;1\;0\;0\;0\;0\;0\;1\;0\;0\;0\;1\;0\;1\;0\;0\;0
\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;
1\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 0\;0\;0\;0\;0\;2\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;0\;0\;1
\;0\;0\;1\;0\;0\;0
\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;1\;0\;0\;0\;0\\
1\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0
\;1\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 0\;1\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;2\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;1
\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;1\;0\\
1\;0\;0\;0\;1\;0\;0\;0\;0\;0\;1\;0\;0\;0\;1\;0
\;1\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;0\;0\;1\;0
\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;
0\;0\;0\;0\;0\;1\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;1\;0\;1\;0\;0\;0\;2\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0
\;0\;0
\;0\;0\;0\;0\;0\;0\;1\;0\;0\;
0\;0\;0\;0\\ 0\;0\;0\;0\;0\;1\;0\;0\;0\;1\;0\;1\;0\;0\;0\;1\;0\;0\;0\;1
\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;1\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;1\;0\;1\;0\;0\;0\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;1\;0\;0\;0\;1\;0\;0\;0\;1\;0\;1\;0\;0\;0\;1
\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
2\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;1\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0
\;1\;0\;0\;1\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;1\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\;0\;1
\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;1\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0\\ 1\;0\;0\;0\;1\;0\;1\;0\;0\;0\;1\;0\;0\;0\;0\;0\;1\;0\;0\;0
\;1\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;2\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0
\;0\;0\;0\;1\;0\;0\;1\;0\;
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;1\;0\;0\;0\;0\\
0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0
\;0\;0\;0\;0\;0\;0\;0\;0\;0\;0\;
0\;0\;0\;0
\end{array}\right)
\end{equation}
\newpage
\epsffile[0 0 465 615]{dFig3.ps}
Figure 8: $4$-fold substitutes of the triangles of Figure~7.0 and
7.1 in all directions.
\newpage
\epsffile[0 0 465 645]{dFig4.ps}
\newpage
\addcontentsline{toc}{section}{\bf References}
|
\section{Statement of the problem}
The Dirac-ADM canonical approach \cite{Dir58,ADM} to GR was the essential
improvement
on the way to quantization of the Einstein-Hilbert theory of gravity
developed by Wheeler, DeWitt and others \cite{Wheel,DeWitt,Mis,Ryan}.
This conventional scheme of the canonical quantization is based on the
(3+1) Dirac-ADM foliation
\cite{Dir58,ADM,kuchar,Zel} of the four dimensional manifold
$(x^\mu)$ along the some time-like vector (associated with
the rest frame of an observer):
$$
ds^2=N^2dt^2-{}^{(3)}g_{ik}\breve d x^i \breve d x^k \;\;;\;\;
(\breve d x^i=dx^i + N^idt)\,
$$
(that means the restriction of the group of general coordinate
transformations by the kinemetric ones
\cite{Zel}: $t\to t'(t)\;\;\;;\;\;\;x_i\to {x'}_i(t,x_1,x_2,x_3)$)
and on the ADM action ($W_{ADM}$) which differs from the initial
Einstein-Hilbert action \cite{Hil15}
\begin{equation} \label{action}
W_{GR}=\int d^4x\sqrt{-g}\left [-\frac{{}^{(4)}R(g)}{2\kappa^2} +
{\cal L}_{matter} \right ]\;\;;\;\;(\kappa^2=8\pi G),
\end{equation}
by the surface terms:
$$
W_{GR}=W_{ADM}+W_S+W_T,
$$
where
\begin{equation} \label{t}
W_T=-\int dtd^3x {\breve \partial}_0 \left [\frac{{\breve \partial}_0
\sqrt{{}^{(3)}g}}{N\kappa^2} \right ]\;\;;\;\;\left [{\breve \partial}_0f=
\dot f -\partial_k (N^kf)\right].
\end{equation}
\begin{equation} \label{s}
W_S=-\int dtd^3x \partial_k[\sqrt{{}^{(3)}g}({}^{(3)}g^{ik}
\partial_l N)]\frac{1}{\kappa^2}\;.
\end{equation}
For the derivation of local classical equations these surface terms are
not essential, however, they play an important role in the determination
of the global quantities of such a total energy \cite{Regge}.
It is obvious that the wave function also has global nature and
depends on these surface terms.
{\bf The statement of the problem consists in the canonical quantization of
the Einstein-Hilbert action~(\ref{action}) by taking into account the
surface terms (\ref{t}),~(\ref{s})}. We continue the attempts to solve this
problem in the papers \cite{Perv,Khved}.
\section{A new version of the Hamiltonian Formulation}
{\bf To solve this problem we should
I) consider the space-scale variable $a=[{}^{(3)}g]^{\frac{1}{6}}$
in the time surface term~(\ref{t})} as one of dynamical variables :
\begin{equation} \label{metric}
ds^2=a^2[N_c^2dt^2-\omega_{i\underline k}\omega_{j\underline k}{\breve d}x^i\breve d x^j],
\;\;\;;\;\;\;aN_c=N\;\;\;;\;\;\; {\rm det}\omega=1,
\end{equation}
(we use the triad form $\omega_{i\underline j}$ for the rest dynamic
variables~\cite{Smir})
{\bf II) apply the Ostrogradsky method} ~\cite{Ostr,gkp}
to the theory with second order derivative of the scale factor with respect
to the time coordinate. As the result, the Hilbert action (\ref{action})
in terms
of the canonical conjugate variables $ a,\pi_a, \Phi,\pi_\Phi$
(where $\Phi$ denotes the set of matter fields
including graviton $\omega$ and photon $A$) has the form
$$
W_{GR}=\int^T_0 dt \int_V d^3 x \left [-\pi_{(a)}\stackrel{\odot}{a}+
\frac{1}{2}{\breve\partial}_0(\pi_{(a)}a)+\sum_{\Phi =\omega ,A} \pi_{(\Phi)}
\stackrel{\odot}{\Phi}
-N_c{\cal H}_{EC}\right] + W_S,
$$
where
$$
\stackrel{\odot}{a}=\breve\partial{}_0a+\frac{2}{3}a\partial_kN^k,\;\;\;\;\;
\stackrel{\odot}{A}_k=\partial_0A_k-\partial_kA_0-N^lF_{lk}
$$
$$
\stackrel{\odot}{\omega}_{lk}=(\dot\omega {}_{l\underline s}\omega {}_{k\underline s}
+\omega {}_{l\underline s}\dot\omega {}_{k\underline s}
-\nabla_lN_k -\nabla_kN_l+
\frac{2}{3}\omega_{l\underline s}\omega_{k\underline s}\partial_jN^j),
$$
is the kinemetric invariant time derivative (we use here covariant derivative
in the metric $\omega_{i\underline k}\omega_{j\underline k}$,
including the Laplace operator
$\Delta f=\nabla_k\partial^kf$), and ${\cal H}_{EC}$ is the Einstein energy
density
\begin{equation} \label{eenergy
{\cal H}{}_{EC}=-\frac{\kappa^2}{12}\pi^2_{(a)}+ \pi^2_{(\omega)}\left(
\frac{2\kappa^2}{a^2}\right)+\left(\frac{a^2}{2\kappa^2}\right)\bar R+
{\cal H}_{(A)};\;\;\;\;
\end{equation}
with the photon energy
$$
{\cal H}{}_{(A)}= \frac{1}{2}\pi_{(A)}^k{\pi}_{k (A)}+\frac{1}{4}F_{ij}F^{ij}
\;\;
$$
and three dimensional curvature
$$
\bar R=a^2\;\;{}^{(3)}R(a^2\omega^2)={}^{(3)}R(\omega^2) +
8a^{-\frac{1}{2}}\Delta a^{\frac{1}{2}},
$$
{\bf III) perform the canonical transformation}
$(\pi_{(a)},a) \ =>\;(\Pi,\eta)$
which removes the time surface term $\frac{1}{2}\partial_0(\pi_{(a)}a)$
\begin{equation} \label{ct}
\pi_{(a)}\stackrel{\odot}{a}-\frac{1}{2}\breve\partial{}_0(\pi_{(a)}a)=
\Pi
(\dot\eta-N^k\partial_k\eta)\;.
\end{equation}
One can represent this transformation as
\begin{equation} \label{cantr}
\pi_{(a)}=2\sqrt{\frac{3\Pi}{\kappa^2\Gamma}} C(\eta);\;\;\;\;
a=\sqrt{\frac{\kappa^2\Gamma \Pi}{3}} S(\eta),
\end{equation}
where $ C(\eta), S(\eta) $ and $\Gamma $ are some particular
solution of the following equations
\begin{equation} \label{cs}
C(\eta)\frac{d}{d\eta} S(\eta)- S(\eta)\frac{d}{d\eta} C(\eta)=1;\;\;\;\;\;
\partial_0({\rm ln\Gamma})-N^k\partial_k({\rm
ln}\Gamma)+\frac{1}{3}\partial_kN^k=0\;.
\end{equation}
Finally, the Hilbert action reads in terms of the new variables as
\begin{equation} \label{ham}
W_{GR}=\int dtd^3x\left [\sum_{\Phi =\omega ,A,{\rm ln}\Gamma } \pi_{(\Phi )}
{\dot \Phi }-\Pi {\dot \eta }
-N_c{\cal H}{}_{EC}-N^k{\cal P}_k \right]+{\bar W} {}_S\;,
\end{equation}
with space surface term
$\bar W{}_S=W_S-2\int^T_0dtd^3x\partial_l(N^k\pi_{(h)}{}^l_k)$
and constraints
\begin{equation} \label{c1}
{\cal H}_{EC}=
-\frac{\Pi}{\Gamma}(C^2-S^2\frac{6}{\Gamma^2}\bar R)
+\pi^2_{(\omega)}\frac{6}{S^2\Pi\Gamma}+{\cal H}_{(A)}=0,
\end{equation}
\begin{equation} \label{c2}
{\cal P}_k=\pi_{(\Gamma)}\partial_k{\rm ln}\Gamma+\partial_k\pi_{(\Gamma)}+
\Pi \partial_k\eta +2\nabla_l\pi^l_{(\omega) k}+\pi_{(A)}^lF_{lk}=0.
\end{equation}
\section{Interpretation of new variables}
To treat the new variables $\Pi $ and $\eta $, we consider the flat-space
limit \cite{Khved}:
$\pi_{(\omega)}=\bar R=N^k=0$;\\$C(\eta)=1;\;S(\eta)=\eta,$
where the Hilbert action~(\ref{ham}) has the form
$$
W_{GR}=\int dt d^3x\left\{\pi^k_{(A)}(\dot A_k-\partial_kA_0)-
\Pi\dot \eta
-N_c({\cal H}_{(A)}-\frac{\Pi}{\Gamma})\right\}.
$$
After the reduction on the constraint shell
$ {\cal H}_{EC} = 0$
in the gauge $N_c=1$, we get the expression
for the conventional action of
electrodynamics
$$
W^{Red}=\int dt d^3x \left\{\pi^k_{(A)}(\dot A {}_k-\partial_kA_0)-
{\cal H}_{(A)}\right\},
$$
Note that because
\begin{equation}
\label{em}
\dot\eta=\frac{1}{\Gamma}.
\end{equation}
{\bf in this limit the variable $\eta$ can be treated as the time and
the quantity $\Pi /\Gamma $
in the reduced phase space is a "reduced energy" like the
the quantity $ \sqrt{p^2+m^2} $ is the spectral energy for a
relativistic particle.}
\section{The wave function of the Universe}
Suppose that the constraint ${\cal H}_{EC}=0$ has the set of
solutions
$\Pi={\cal H}^{red}_{(\alpha)}\;\;\;,\alpha=1...m$.
For each solution one can write down the corresponding reduced Hilbert
action
\begin{equation} \label{wred
W^{red}_{GR(\alpha)}=\int dt \int d^3x \left(\sum_{\Phi=\omega ,
A,{\rm ln}\Gamma} \pi_{(\Phi)}\dot\Phi
-{\cal H}^{red}_{(\alpha)}\dot\eta\right) +\bar W_S
\end{equation}
The quantization of the action leads to the Schr\"odinger type evolution
equation
\begin{equation} \label{wdw}
\frac{1}{i}\frac{\delta}{\delta\eta}\Psi_{\alpha}=\hat{\cal
H}{}^{red}_{(\alpha)}
\Psi_{\alpha} .
\end{equation}
Let us consider the small time limit $(\eta)\sim 0 $
which corresponds to the small Universe
$(a\sim S(\eta)\to0)$.
According to equation~(\ref{cs}), in this region $S(\eta)\sim\eta$,
$C(\eta)\sim 1 $, and the graviton term ${\pi }_{(\omega)}^2$
dominates in the energy density (\ref{c1}). The corresponding solutions of
the constraints ${\cal H}_{EC}=0$
\begin{equation} \label{pi}
\Pi={\cal H}^{red}_\pm = \pm\sqrt{6\pi_{(\omega)}^2}\frac{1}{\eta}.
\end{equation}
and the reduced action reads
\begin{equation} \label{mis}
W_{GR\pm} =\int^T_0dt\int_Vd^3x\left(\sum \pi_{(\omega)}\dot\omega
\mp\sqrt{6\pi_{(\omega)}^2}\partial_0{\rm ln}\eta\right).
\end{equation}
One can verify that in the supposition of
homogeneousity of the space
\begin{equation} \label{homo
ds^2=a^2[N^2_c(t)dt^2-A^2(r)dx^2];\;\;\;
\;\;A(r)=\left(1+\frac{kr^2}{4r^2_0}\right)^{-1}
\end{equation}
\begin{equation} \label{fried}
a(t)N_c(t)dt=dT_{Fried};\;\;\;\;k=0,\pm1,
\end{equation}
it follows from (\ref{mis}) the action for the
Misner Universe \cite{Mis} (in details see~\cite{Ryan}):
\begin{equation} \label{1mis}
W_{GR\pm} = V_{(3)} \left(\pi_{(\omega)}(\omega(T)-\omega(0))\mp
\sqrt{6\pi_{(\omega)}^2}{\rm ln}\frac{\eta(T)}{\eta(0)}\right),
\end{equation}
and the spectral decomposition for the wave function
\begin{equation} \label{psq}
\Psi = \int d\pi_{(\omega)} \left( A^+_{\pi_{(\omega)}}e^{iW_{GR+}} +
A^+_{\pi_{(\omega)}}e^{iW_{GR-}} \right),
\end{equation}
where the role of the time is played by the logarithm of $\eta$,
and $A^\pm$ are the operators of creation and annihilation of the Universe.
Note that in the case of the homogeneous space the
particular solutions of eq.~(\ref{cs}) read
\begin{equation} \label{cos}
C(\eta)=1,{\rm cos{\eta}},{\rm cosh{\eta}}; \quad
S(\eta)= \eta,{\rm sin{\eta}},{\rm sinh{\eta}},
\quad \Gamma = r_0
\end{equation}
respectively for \(k = 0, 1, -1\).
The evolution of the Universe filled in by dust and radiation has been
considered in \cite{Khved}.
In this case the wave function of the Universe has the
form $\Psi_\pm=\exp{\{\imath
W_\pm^{red}(a)}\}$ with the reduced action
$$
W^{red}_\pm=\pm V_{(3)} \int_{0}^{T}da \left[\underline{\pi}_{(a)}-
\frac{1}{2}\frac{d}{da}(\underline{\pi}_{(a)}a)\right],
$$
where $V_{(3)}=\int d^3xA^3(r)$ and
$
\underline{\pi}_{(a)}=2\sqrt{\frac{3}{\kappa^2}}\left[{\cal H}_{(M)}-
\frac{3ka^2}{r_0^2\kappa^2}\right]^{1/2}
$
is the solution of the constraint (\ref{c1}) ${\cal H}_{EC}=0$. The phase
of this function coincides (up to the energy factor) with the Friedmann
time (\ref{fried}) for the dust case (${\cal H}_M=a\epsilon_{dust}$)
\begin{equation} \label{d}
W^{red}_{dust}=\frac{V_{(3)}\epsilon_{dust}}{2}T_{Fried}(a);\;\;\;\;\;\;\;\;
T_{Fried}(a)=\int_{a(0)}^{a(T)}da\frac{6a}{\kappa^2 \underline{\pi}_{(a)}}
\end{equation}
and with the conformal time $N_cdt=\eta (t) r_0$ for the radiation
(${\cal H}_M= \epsilon_{rad} $)
\begin{equation} \label{r}
W^{red}_{rad}=V_{(3)}\epsilon_{rad}\eta (a)r_0;\;\;\;\;\;\;\;\;
\eta (a)r_0=\int_{a(0)}^{a(T)}da\frac{6}{\kappa^2 \underline{\pi}_{(a)}}.
\end{equation}
It is worth to note that the this clear correspondence
between the quantum and classical physics ( the "phase time" and the
"interval time") arises due to the
maintenance of the time surface term $\frac{1}{2}
\partial (\underline{\pi}_{(a)}a)$ in the Hilbert action.
{\bf Thus, the time surface term helps us to establish the
correspondence between the time as the phase of the ADM wave function of
the Universe and the classical proper time as an invariant interval.}
\section{Quantum scenario
of the Weinberg-Salam Universe.}
Let us consider the quantum scenario of the evolution of the Universe filled
in by the Weinberg-Salam fields and described by the action
\begin{equation} \label{ws}
W=\int d^3xdt\sqrt{-g}\left\{-\frac{{}^{(4)}R}{2\kappa^2}
+\frac{{}^{(4)}R \Phi^* \Phi }{6}+ \partial_{\mu}\Phi^* \partial^{\mu} \Phi
-\gamma(\bar \Psi_{L}\Phi)\Psi_R+\cdots\right\}
\end{equation}
where $\Phi=\left(\begin{array}{c}\Phi_1\\
\Phi_2
\end{array}\right)$
is the doublet of the complex scalar fields; $\Psi_L$ and $\Psi_R$ are the left
and right fermions. We keep only the term of the scalar-fermion interaction
generating masses of the fermions \\
$m(\bar \Psi_{L2}\Psi_R+\bar \Psi_R\Psi_{L2})$
to show the evolution of the mass parameters with respect to the scale
$a$.
If we extract the scale factor \( a \) not only from the metric
(\ref{metric})
$g_{\mu\nu}=a^2\tilde g_{\mu\nu}$, but also from all other
matter fields
$\Phi=a\varphi,\;\;\; \Psi=a^{-3/2}\Psi_c$, we can
guarantee the classical limit of the massive fermion fields as the
Friedmann dust of the Universe.
The action in terms of the physical fields $\varphi,\Psi_c,\tilde g
(\sqrt{-\tilde g}=N_c)$ has the form
\begin{eqnarray
W&=&\int d^4x\{N_c\left[-\frac{{}^{(4)}\tilde R}{2\kappa^2}\left(a^2-
\frac{\kappa^2}{3}\varphi^*\varphi\right)-\frac{3}{\kappa^2}\partial_\nu a\partial^\nu a
+\partial_\nu \varphi^*\partial^\nu\varphi-\gamma(\bar
\Psi_{Lc}\varphi^)\Psi_{Rc}+\cdots\right]\nonumber\\
&&+\partial_\nu\left(N_c\left[\frac{3a^2}{\kappa^2}-\varphi^*\varphi
\right]\frac{\partial^\nu a}{a}\right)\}\nonumber
\end{eqnarray}
Note that the configuration $a^2 =\frac{\kappa^2}{3}\varphi^*\varphi $ represents
the singular point: in the vicinity of this point the sign
before the four- dimensional curvature is changing.
So, let us consider only the field configuration such that
$[a^2-\frac{\kappa^2}{3}\varphi^*\varphi]=\rho^2 > 0$.
For this case one can introduce new variables
$$
a=\rho \cosh(\xi);\;\;
\varphi_i=\sqrt{\frac{3}{\kappa^2}}\rho \sinh(\xi) n_i;\;\;
n_1=\cos(\Theta)\exp{\{\imath \chi_1\}};\;\;n_2=\sin(\Theta)\exp{\{\imath \chi_2\}},
$$
where $\xi,\Theta,\chi_1,\chi_2$ are the angles of the scale-scalar mixing.
For the homogeneous space~(\ref{homo}) with $V_{(3)}=1$ we get the action
\begin{eqnarray}
W&=&\int dt\{-\frac{D}{2N_c}\left(\frac{6}{\kappa^2}\right)+
N_c\left[\frac{k}{2r_0^2}\rho^2\left(\frac{6}{\kappa^2}\right)
-\gamma\sqrt{\frac{3}{\kappa^2}}\rho\sinh(\xi)(\bar
\Psi_{Lci}n_i)\Psi_{Rc}+\cdots\right]
\nonumber\\
&&+\partial_0\left[
\frac{\rho^2}{2N_c}\left(\frac{\dot \rho}{\rho}+\dot \xi \tanh \xi\right)
\right]\frac{6}{\kappa^2}\},\nonumber
\end{eqnarray}
where $D=-\dot \rho^2+\rho^2 \dot \xi^2+(\rho\sinh\xi)^2\left(\dot \Theta^2+
\sin^2(\Theta)\dot \chi^2_1+\cos^2(\Theta)\dot \chi^2_2\right)$
is the $SO(4,1)$-invariant differential form.
By the Ostrogradsky method this action can be rewritten in terms of momenta
$$
W=\int dt\{\sum_{\alpha=\xi,\Theta,\chi_1,\chi_2}\pi_{(\alpha)}\dot \alpha-
\pi_{(\rho)}\dot \rho-
N_c{\cal H}_{EC}+\frac{1}{2}\partial_0\left(\pi_{(\rho)}\rho\right)+
\frac{1}{2}\partial_0\left(\pi_{(\xi)}\tanh(\xi)\right)\}
$$
$$
{\cal H}_{EC}=-\frac{1}{2}\frac{\kappa^2}{6}\pi^2_{(\rho)}-\frac{k\rho^2}{r^2_0}
\frac{6}{\kappa^2}
+\frac{{\cal K}^2}{2\rho^2}\frac{\kappa^2}{6}
+\gamma\sqrt{\frac{3}{\kappa^2}}\rho\sinh(\xi)(\bar \Psi_{Lci}n_i)\Psi_{Rc}+\cdots
$$
where ${\cal K}^2$ is the Kazimir operator of the $SO(4,1)$ group
$$
{\cal K}^2=P^2_{(\xi)}+\frac{1}{\sinh^2(\xi)}\left(P^2_{(\Theta)}+
\frac{P^2_{\chi_1}}{\sin^2(\Theta)}+\frac{P^2_{\chi_2}}{\cos^2(\Theta)}\right).
$$
We see that the variable $\rho$ plays the same role as the scale $a$ in the
theory without scalar fields, and it is the time like variable.
Note that our transition to new variables
is similar to the Bekenstein transformation~\cite{Beken}.
Repeating the canonical transformation (\ref{cantr})
$\pi_{(\rho)}\dot \rho-
\frac{1}{2}\partial_0(\pi_{(\rho)}\rho)=\Pi\dot \eta$, where $\Gamma=r_0$, we get
the expression for the action
\begin{eqnarray}
W&=&\int dt\{\sum_{\alpha}P_{(\alpha)}\dot \alpha-\Pi\dot \eta-
N_c\left[
-\frac{\Pi }{r_0}+\frac{{\cal K}^2}{4\Pi r_0S^2(\eta)}+
\gamma\sqrt{\frac{3}{\kappa^2}}\rho\sinh(\xi)(\bar \Psi_{Lci}n_i)\Psi_{Rc}
+\cdots\right]\nonumber\\
&&+\frac{1}{2}\partial_0\left(P_{(\xi)}\tanh (\xi)\right)\}\nonumber
\end{eqnarray}
This action describes the following ADM-scenario of the evolution of the
Universe. In the small time limit $(\eta\sim 0)$ the Kazimir operator term
dominates and the reduced system on the constraint ${\cal H}_{EC}=0$ has
the form
$$
W^{red}_{(\pm)}=\int_{T_s(0)}^{T_s(T)}dT_s\left[\sum_{\alpha}
P_\alpha\frac{d\alpha}{dT_s}\mp{\cal
K}(P_\alpha)+\frac{1}{2}\frac{d}{dT_S}(P_\xi\tanh(\xi))
\right];\;\;\;\;dT_s=\frac{d\eta}{2S(\eta)}.
$$
The wave function of this system can be decomposed over the eigenfunctions
of the Kazimir operator with the eigenvalues ${\cal K}_
\epsilon$
$$
\Psi(T_s=T_s(T)-T_s(0)|\alpha)=\sum\limits_{\epsilon}
\left[A_\epsilon^{(+)}e^{+\imath {\cal K}_\epsilon T_s}\Psi_\epsilon(\alpha_T|\alpha_0)
+A_\epsilon^{(-)}e^{-\imath {\cal K}_\epsilon T_s}\Psi_\epsilon^*(\alpha_T|\alpha_0)\right],
$$
$$
\Psi_\epsilon(\alpha_T|\alpha_0)=Y_{SO(4,1)}(\alpha_T)Y^*_{SO(4,1)}(\alpha_0)
\exp\left\{\imath P\xi\left(\tanh(\xi_T)-\tanh(\xi_0)\right)\right\},
$$
where $A^{(\pm)}$ are the operators of the creation and annihilation of the
Universe, $Y_{SO(4,1)}$ is a unitary irreducible representation of the
$SO(4,1)$
group. This wave function reproduces the physical picture of the Misner
anisotropic Universe (\ref{mis}) discussed above in section 4.
In the large time limit, the $SO(4,1)$ symmetry is broken, the Kazimir
operator term disappears in comparison with the mass term.
In this case, the masses of elementary particles in the Weinberg-Salam
model are determined by the fixed values of angles of the scale-scalar
field mixing
and the ADM-observer gets the Friedmann cosmological models of radiation
and dust, considered above.
\vspace{0.3cm}
\section{Conclusion}
We have shown that including the time surface term in the canonical
Hamiltonian formulation of GR helps us to extract the time-like variable
and its conjugated momentum from the extended phase space. By taking
into account the time surface term we represented here the new version
of the Dirac - ADM Hamiltonian formalism for general relativity in the
reduced phase space with the Schr\"odinger - like equation for a wave
function describing the quantum evolution of the Universe. This evolution
coincides with the Friedmann classical evolution of the dust filled Universe
and shows that in GR like in special relativity there are two distinguished
invariant time variables:
{\bf the "phase time" of the ADM-observer (who constructs the Hamiltonian and
measures the time as a phase of the wave function of the expanding Universe )
and the geometrical time of the Friedmann observer
(who measures the time as an invariant proper interval and observes
this expansion on the earth)}.
In special relativity, the corresponding times are connected by
the Lorentz transformation and they coincide only in the case
when the rest frame of the Einstein observer coincides with the rest frame
of a particle.
Now the main question is to find the corresponding transformation
from the rest frame of the ADM - observer to the Friedmann one.
\vspace{0.3cm}
{\bf Acknowledgment}
\vspace{0.3cm}
The authors thank Profs. A. Ashtekar,
B M. Barbashov, G.T. Horowits, V.G. Kadyshevsky, K. Kuchar,
D.A.Kirzhnitz, L.N. Lipatov, D.V.Volkov for useful discussions.
One of the authors (V.N.P.) acknowledges the hospitality of the
International Centre for Theoretical Physics in Trieste where this paper
was finished.
\vspace{0.5cm}
|
\section{Introduction}
The origin of ${\cal CP}$ violation remains one of the outstanding problems in
particle physics. In the attempt to understand this problem many experimental
and theoretical efforts have been launched \cite{review}. One of the systems
where it is possible to search for ${\cal CP}$ violation is the non-leptonic
decay of hyperons. Although this has been known for many years \cite{marshak},
it is only recently that it has become conceivable to carry out an experimental
program to look for ${\cal CP}$ violating signals in the decays of $\Xi$ and
$\Lambda$ hyperons \cite{cernpro,proposal}.
Of particular interest is the upcoming experiment E871 that expects to
reach a sensitivity of $10^{-4}$ for the sum of asymmetries $A(\Lambda^0_-)
+ A(\Xi^-_-)$ \cite{proposal}.
Unfortunately, the calculation of these asymmetries is
plagued by theoretical uncertainties in the estimate of the hadronic matrix
elements involved. Nevertheless, a conservative study of these asymmetries
within the minimal standard model indicated that $A(\Lambda^0_-)$ is likely to
occur at the level of a few times $10^{-5}$. In view of this, the potential
results of E871 are very exciting.
One of the questions we would like to answer is whether the phase in the CKM
matrix of the three generation minimal standard model is the sole source of
${\cal CP}$ violation. The experimental information that we have so far is:
\begin{itemize}
\item A
non-zero value of the parameter $\epsilon$ in kaon decays \cite{pdb}:
\begin{equation}
|\epsilon| = 2.26 \times 10^{-3}
\end{equation}
\item A measurement
of the parameter $\epsilon^\prime$ \cite{epove}:
\begin{equation}
{\epsilon^\prime \over \epsilon} = \cases{ (2.3 \pm 0.65)\times 10^{-3} &
NA31 \cr
(0.74\pm0.52\pm0.29)\times 10^{-3} & E731 \cr}
\label{eppexp}
\end{equation}
\end{itemize}
The first result indicates that there is ${\cal CP}$
violation in nature, but it does not
pinpoint its origin. The best one can say is that it is possible for the
minimal standard model to accommodate this number.
If the second number turns out to
be non-zero it would establish the existence of
direct $|\Delta S|=1$ ${\cal CP}$ violation, ruling out some superweak models.
The current experimental numbers are
consistent with the minimal standard model, although the theoretical
calculations are also plagued with uncertainty from the
evaluation of hadronic matrix elements.
The present situation is, therefore, that there is no need for ${\cal CP}$ violation
beyond the phase in the three generation CKM matrix\footnote{Except perhaps
in the origin of the baryon asymmetry of the universe. We will not discuss that
issue in this paper.}, but that other sources of ${\cal CP}$ violation have not
been ruled out.
The question we want to address in this paper is whether it
is possible for E871 to find a non-zero asymmetry given its expected
sensitivity and the current values of $\epsilon$ and
$\epsilon^\prime/\epsilon$. To this end, and in keeping with the
results of all the precise experiments conducted to date, we will assume
that the minimal three-generation standard model is a very good low energy
approximation to the electroweak interactions. We will, therefore, discuss
any possible new physics in terms of an effective Lagrangian consistent
with the symmetries of the standard model and will only look only at
operators of dimension six.
Our paper is organized as follows. In Section 2 we review the notation
for ${\cal CP}$ violating observables in hyperon decays as well as the
standard model estimate of $A(\Lambda^0_-)$. In section~3 we compute
the contributions of ${\cal CP}$ violating four quark operators to
$A(\Lambda^0_-)$ and the constraints that result from the measurements
of ${\cal CP}$ violation in $K\rightarrow \pi\pi$. In section~4 we repeat this
analysis for the two-quark operators of dimension six (so called penguin
operators). Finally, we present our conclusions.
\section{${\cal CP}$ Violation in $\Lambda^0 \rightarrow p \pi^-$}
In this section we review the basic features of ${\cal CP}$ violation in
the reaction $\Lambda^0 \rightarrow p \pi^-$, denoted by $(\Lambda^0_-)$.
In the $\Lambda^0$ rest frame, $\vec{\omega}_{i,f}$ will denote unit vectors in
the directions of the $\Lambda$ and $p$ polarizations, and
$\vec{q}$ will denote the proton momentum.
The isospin of the final state is
$I= 1/2 {\rm ~or~} 3/2$, and each of these two states can be reached
via a $\Delta I = 1/2 {\rm ~or~} 3/2$ weak transition respectively.
There are also two possibilities for the parity of the final state. They are
the $s$-wave, $l=0$, parity odd state (thus reached via a parity
violating amplitude); and the $p$-wave, $l=1$, parity even state reached
via a parity conserving amplitude.
A model independent analysis of the decay can be done by writing the
most general matrix element consistent with Lorentz
invariance: \cite{marshak}
\begin{equation}
{\cal M} = G_F m_\pi^2 \overline{u}_P(A-B\gamma_5)u_\Lambda .
\label{gmatel}
\end{equation}
It is customary to introduce the quantities:
\begin{eqnarray}
s&=&A\nonumber \\
p&=&\biggl({|\vec{q}|\over E_P + M_P}\biggr) B
\end{eqnarray}
to write the total decay:
\begin{equation}
\Gamma = {|\vec{q}|(E_P+M_P) \over 4 \pi M_\Lambda}G_F^2 m_\pi^4
\left(|s|^2+|p|^2\right).
\label{drate}
\end{equation}
The angular distribution is proportional to:
\begin{equation}
{d\Gamma \over d\Omega} \sim 1+
\gamma\vec{\omega}_i\cdot\vec{\omega}_f + (1-\gamma)\hat{q}\cdot\vec{\omega}_i
\hat{q}\cdot\vec{\omega}_f
+ \alpha \hat{q}\cdot(\vec{\omega}_i+\vec{\omega}_f)
+ \beta \hat{q}\cdot(\vec{\omega}_f\times\vec{\omega}_i),
\label{msqsim}
\end{equation}
where we have used the standard notation \cite{marshak}:
\begin{equation}
\alpha \equiv {2 {\rm Re} s^*p \over |s|^2 + |p|^2}, \;\;
\beta \equiv {2 {\rm Im} s^*p \over |s|^2 + |p|^2}, \;\;
\gamma \equiv {|s|^2 - |p|^2 \over |s|^2 + |p|^2} .
\label{albega}
\end{equation}
If the proton polarization is not observed, $\alpha$ is the parameter
that governs the angular distribution:
\begin{equation}
{d\Gamma \over d\Omega}={\Gamma \over 4 \pi}
\left(1+\alpha \hat{q}\cdot \vec{\omega}_i\right).
\label{alonly}
\end{equation}
Similarly, if the initial $\Lambda$ is unpolarized,
$\alpha$ determines the polarization of the proton:
\begin{equation}
\vec{\cal P}_p = \alpha_\Lambda \hat{q}
\label{otheral}
\end{equation}
E871 will not measure the correlations governed by the
parameter $\beta$ so we will not deal with it in this paper.
The ${\cal CP}$-odd observable $A(\Lambda^0_-)$ is constructed by
comparing the parameter $\alpha$ in the reaction
$\Lambda^0 \rightarrow p \pi^-$ with the corresponding parameter
$\overline{\alpha}$ in the reaction
$\overline{\Lambda}^0 \rightarrow \overline{p} \pi^+$.
One can show that ${\cal CP}$ symmetry predicts that:
\begin{equation}
\overline{\alpha} = - \alpha
\label{cppred}
\end{equation}
so that a ${\cal CP}$ odd observable is:
\cite{donpa}
\begin{equation}
A \equiv {\alpha \Gamma + \overline{\alpha}\overline{\Gamma} \over
\alpha \Gamma - \overline{\alpha}\overline{\Gamma}}
\approx {\alpha + \overline{\alpha} \over \alpha - \overline{\alpha}}
\label{cpobs}
\end{equation}
Other possible ${\cal CP}$ odd observables have been discussed in the literature:
a rate asymmetry that is significantly smaller than $A$ \cite{donpa}; and
an asymmetry based on the parameter $\beta$ that won't be accessible to E871.
For these reasons we concern ourselves with the observable $A(\Lambda^0_-)$.
\footnote{In fact E871 will be sensitive to the
sum $A(\Lambda^0_-) + A(\Xi^-_-)$. An analysis of
$A(\Xi^-_-)$ parallels the one we will
carry out, but doesn't really affect our conclusions given the inherent
uncertainties in the computation of matrix elements. It has also been argued
that $A(\Xi^-_-)$ is probably smaller than $A(\Lambda^0_-)$ due to much smaller
strong rescattering phases \cite{lusawi}.}
It is convenient to decompose the amplitudes according to isospin, and
to introduce the following notation for the phases:
\begin{eqnarray}
s(\Lambda^0_-)= - \sqrt{2/3}s_{1}e^{i(\delta^s_1+\phi^s_1)}+\sqrt{1/3}
s_{3}e^{i(\delta^s_3+\phi^s_3)}\nonumber \\
p(\Lambda^0_-)= - \sqrt{2/3}p_{1}e^{i(\delta^p_1+\phi^p_1)}+\sqrt{1/3}
p_{3}e^{i(\delta^p_3+\phi^p_3)}
\label{overnot}
\end{eqnarray}
where $\delta^I_J$ is the strong rescattering phase for the pion nucleon
system and $\phi^I_J$ is the ${\cal CP}$ violating phase.
In terms of these quantities one finds: \cite{donpa}
\begin{eqnarray}
A(\Lambda^0_-)&=& -\tan\left(\delta^p_1-\delta^s_1\right)
\sin\left(\phi^p_1-\phi^s_1\right)\biggl[ 1 \nonumber \\
&+&{1\over \sqrt{2}}{s_3 \over s_1}\biggl( {\cos(\delta^p_1-\delta^s_3)
\over \cos (\delta^p_1-\delta^s_1)}-{\sin(\delta^p_1-\delta^s_3)
\over \sin(\delta^p_1-\delta^s_1)}{\sin(\phi^p_1-\phi^s_3)
\over \sin(\phi^p_1-\phi^s_1)}\biggr) \nonumber \\
&+&{1\over \sqrt{2}}{p_3 \over p_1}\biggl( {\cos(\delta^p_3-\delta^s_1)
\over \cos (\delta^p_1-\delta^s_1)}-{\sin(\delta^p_3-\delta^s_1)
\over \sin(\delta^p_1-\delta^s_1)}{\sin(\phi^p_3-\phi^s_1)
\over \sin(\phi^p_1-\phi^s_1)}\biggr) \biggr]
\label{approxas}
\end{eqnarray}
Experimentally we know the values of:
\begin{itemize}
\item the strong rescattering phases \cite{roper}:
\begin{equation}
\delta_1 \approx 6.0^\circ,\;\;\delta_3 \approx -3.8^\circ,\;\;
\delta_{11}\approx -1.1^\circ,\;\;\delta_{31}\approx-0.7^\circ
\label{stphex}
\end{equation}
with all the errors on the order of $1^\circ$.
\item the $\Delta I=3/2$ amplitudes are much smaller than the
$\Delta I=1/2$ amplitudes \cite{pdbee}:
\begin{equation}
s_{3}/s_{1} = 0.027 \pm 0.008, \;\; p_{3}/p_{1}=0.03\pm 0.037
\label{expamp}
\end{equation}
\item the $s$ and $p$ amplitudes (assuming they are dominated by the
${\cal CP}$ conserving, $\Delta I =1/2$, transitions):
\begin{eqnarray}
s \approx -\sqrt{2 \over 3} s_1 &=& 1.47 \pm 0.01 \nonumber \\
p \approx -\sqrt{2 \over 3} p_1 &=& \biggl({|\vec{q}|\over E_P + M_P}\biggr)
(9.98 \pm 0.24)
\end{eqnarray}
\end{itemize}
Substituting the experimental numbers for the amplitudes and strong
rescattering phases one gets:
\begin{equation}
A(\Lambda^0_-)\approx 0.13 \sin(\phi^p_1-\phi^s_1)
+0.001 \sin(\phi^p_1-\phi^s_3) -0.0024 \sin(\phi^p_3-\phi^s_1)
\label{alambda}
\end{equation}
\subsection{Standard model calculation}
In the case of the minimal standard model, the ${\cal CP}$ violating phase resides
in the CKM matrix. For low energy transitions, this phase shows up as the
imaginary part of the Wilson coefficients in the effective weak
Hamiltonian. In the notation of Buras \cite{buras},
\begin{equation}
H_W^{SM} = {G_F \over \sqrt{2}}V^*_{ud}V_{us}\sum_i c_i(\mu)Q_i(\mu) +
{\rm ~h.~c.}
\label{effweak}
\end{equation}
$Q_i(\mu)$ are four quark operators, and $c_i(\mu)$ are the Wilson
coefficients that are usually written as:
\begin{eqnarray}
c_i(\mu) &=&z_i(\mu)+\tau y_i(\mu) \nonumber \\
\tau &=& - {V^*_{td}V_{ts} \over V^*_{ud}V_{us}}
\end{eqnarray}
with the ${\cal CP}$ violating phase being the phase of $\tau$.
Numerical values for these coefficients can be found, for
example, in Buchalla {\it et. al.} \cite{buras}.
The calculation of the weak phases would proceed
by evaluating the hadronic matrix
elements of the four-quark operators in Eq.~\ref{effweak} to obtain real
and imaginary parts for the amplitudes, schematically:
\begin{equation}
\matel{p \pi}{H_w^{eff}}{\Lambda^0}|^I_\ell = {\rm Re}M^I_\ell + i {\rm Im}
M^I_\ell ,
\label{schematic}
\end{equation}
and to the extent that the ${\cal CP}$ violating phases are small, they can be
approximated by
\begin{equation}
\phi^I_\ell \approx { {\rm Im}M^I_\ell \over {\rm Re}M^I_\ell}.
\label{smallph}
\end{equation}
At present, however, we do not know how to compute the matrix elements so
we cannot actually implement this calculation.
For a simple estimate, we can take the real part of the matrix
elements from experiment (assuming that the measured amplitudes are real,
that is, that ${\cal CP}$ violation is small), and compute the imaginary
parts in vacuum saturation. This approach provides a conservative
estimate for the weak phases because the model calculation of the
real part of the amplitudes is smaller than the experimental value.
Nevertheless, the numbers should be viewed with great caution.
The approximate weak phases estimated in vacuum saturation are: \cite{steger}
\begin{eqnarray}
\phi^1_s &\approx& -3 y_6 {\rm Im}\tau \nonumber \\
\phi^1_p &\approx& -0.3 y_6 {\rm Im}\tau \nonumber \\
\phi^3_s &\approx& \left[3.6(y_1+y_2)+2.7(y_7+3y_8)
{B_0^2 \over m_K^2}\right] {\rm Im}\tau \nonumber \\
\phi^3_p &\approx& \left[0.5(y_1+y_2)-0.4(y_7+3y_8)
{B_0^2 \over m_K^2}\right] {\rm Im}\tau
\label{approxph}
\end{eqnarray}
These provide numerical estimates using the values for the
Wilson coefficients\footnote{For
$\mu=1$~GeV, $\Lambda_{QCD}=200$~MeV}
of Buchalla {\it et. al.} \cite{buras},
$y_6 \approx -0.08$; and the value of $B_0$ given in the
appendix. For the quantity ${\rm Im}\tau$ (we use the Wolfenstein
parameterization of the CKM matrix) we take:
\begin{equation}
{\rm Im}\tau = A^2 \lambda^4 \eta \leq 0.001
\label{imtau}
\end{equation}
Putting all the numbers together, and using the upper limit
in Eq.~\ref{imtau} yields:\footnote{See Ref.~\cite{steger} for
additional discussions of this calculation.}
\begin{equation}
A(\Lambda^0_-) \approx 3 \times 10^{-5}
\label{vsnumres}
\end{equation}
Other models of ${\cal CP}$ violation contain additional short distance operators
with ${\cal CP}$ violating phases \cite{wein,moha,senj} and predict different values
for $A(\Lambda^0_-)$ \cite{donpa,chang}.
A summary of results can be found, for example, in Ref.~\cite{hedpf}.
\section{Four-quark Operators}
We now study, in a model independent manner, the
contributions to $A(\Lambda^0_-)$ that occur
due to physics beyond the minimal standard model. In this section
we look at the effect of all the four-quark operators and in the next
section we discuss the two-quark operators.
We assume that the physics that lies beyond the minimal
standard model is characterized by a scale $\Lambda \gg M_W$ and, therefore,
that its most important low energy effects can be parameterized by the lowest
dimension operators of the most general effective Lagrangian consistent
with the symmetries of the standard model. Such a Lagrangian has been written
down by Buchm\"{u}ller and Wyler \cite{buchmuller}. In the appendix we list
all the operators that occur at dimension six with $|\Delta S| =1$.
The calculation then proceeds as in the previous section,
but with the effective Hamiltonian
\begin{equation}
H_{eff} = H_W^{SM} + {g^2 \over \Lambda^2}
\biggl(\sum_i\lambda_i {\cal O}^{new}_i + {\rm ~h.~c.}\biggr)
\label{effh}
\end{equation}
To the usual, QCD corrected, standard weak Hamiltonian of the previous section
we add all the four-fermion operators with $|\Delta S|=1$ that come from
the new physics sector. We will sidestep the issue of
the possible origin of the effective ${\cal CP}$ violating operators.
We use the notation of Ref.~\cite{buchmuller} as
detailed in the Appendix. These
operators violate ${\cal CP}$ if the coupling $\lambda_i$ has an imaginary
part. The normalization has been chosen for convenience.
\subsection{$K_L \rightarrow \pi \pi$ and $\epsilon^\prime / \epsilon$}
The standard notation for the $K \rightarrow \pi \pi$ amplitudes is:
\begin{eqnarray}
A(K^0 \rightarrow \pi^+ \pi^-) &=& A_0 e^{i\delta_0}+{A_2 \over \sqrt{2}}
e^{i\delta_2} \nonumber \\
A(K^0 \rightarrow \pi^0 \pi^0)&=& A_0 e^{i\delta_0}- \sqrt{2}A_2
e^{i\delta_2}
\end{eqnarray}
where $\delta_{0,2}$ are the strong $\pi\pi$ scattering phases in the
$I=0,2$ channel. The amplitudes $A_0$ and $A_2$ are real unless
there is ${\cal CP}$ violation.
Experimentally it is known that the $\Delta I = 3/2$ amplitude $A_2$
is much smaller than the $\Delta I = 1/2$ amplitude $A_0$:
\begin{equation}
\omega \equiv {{\rm Re}A_2 \over {\rm Re}A_0} \approx {1 \over 22}
\end{equation}
The contribution of the dominant penguin operator (${\cal O}_6$ in the notation
of Ref.~\cite{buras}) to $\epsilon^\prime/\epsilon$ is given by:
\begin{equation}
\biggl({\epsilon^\prime \over \epsilon }\biggr)_6
= -{\omega \over \sqrt{2} |\epsilon|}
{{\rm Im}(A_0)_6 \over |A_0|}=
{\omega \over 2 |\epsilon|}{G_F \over |A_0|} y_6 \lambda {\rm Im}\tau
<\pi^+\pi^-|{\cal O}_6|K^0>
\end{equation}
The hadronic uncertainty enters the calculation through the matrix
element of the four-quark operator. We will use the estimate
of Ref.~\cite{buras}
for the matrix element of ${\cal O}_6$:
\begin{equation}
<\pi^+\pi^-|{\cal O}_6|K^0>\bigg|_{I=0} = -4 \sqrt{2}f_\pi
{m_K^2 - m_\pi^2 \over
\Lambda_\chi^2}\biggl({m_K^2 \over m_s + m_d}\biggr)^2 \approx
-0.26 {\rm GeV}^3
\end{equation}
Using the values $A=0.9$, $\lambda =0.22$ and $\eta =0.5$ one finds that:
\begin{equation}
\biggl({\epsilon^\prime \over \epsilon}\biggr)_6\approx 1.5 \times 10^{-3}
\end{equation}
The usual standard model analysis of $\epsilon^\prime/\epsilon$ consists of
computing this contribution of the ``penguin'' operator,
and of normalizing all
other contributions to it in terms of a parameter $\Omega$:
\begin{equation}
{\epsilon^\prime \over \epsilon}=\biggl({\epsilon^\prime \over \epsilon}
\biggr)_6 \biggl( 1 -\Omega_{SM}-\Omega_{NEW}\biggr)
\end{equation}
$\Omega_{SM}$ is given, for example, in Ref.~\cite{buras}, and we have
introduced an analogous term $\Omega_{NEW}$ for the contributions of
the new four-quark operators. Given the experimental result
in Eq.~\ref{eppexp},
we will place bounds on the new physics by requiring, conservatively, that
$\Omega_{NEW}\leq 1$. We find:
\begin{eqnarray}
\Omega_{NEW} &=& 8\biggl({M_W \over \Lambda}\biggr)^2\sum_i
\biggl({ {\rm Im}\lambda_i
\over A^2 \lambda^5 \eta}\biggr)\biggl[
{<\pi^+\pi^-|{\cal O}_i|K^0>_{I=0}\over
y_6 <\pi^+\pi^-|{\cal O}_6|K^0>_{I=0} } \nonumber \\
&& - {\sqrt{2}\over \omega}
{<\pi^+\pi^-|{\cal O}_i|K^0>_{I=2}\over
y_6 <\pi^+\pi^-|{\cal O}_6|K^0>_{I=0} } \biggr]
\label{epprime}
\end{eqnarray}
Because there is no way at present to compute the matrix elements of
four-quark operators reliably, we will simply use vacuum saturation.
The new contributions to $\epsilon^\prime$ can thus be computed with the
aid of the matrix elements listed in Table~\ref{t: matelkpp}.
We use, as before, $A^2\lambda^5\eta \approx 2 \times 10^{-4}$ and we
explicitly separate the contributions from the different isospin
components of each operator for later convenience. We thus write:
\begin{equation}
\Omega_{NEW} = 4 \times 10^4
\biggl({M_W \over \Lambda}\biggr)^2\sum_i {\rm Im}\lambda_i
\biggl( \omega_{0i} + \omega_{2i} \biggr)
\label{numcoe}
\end{equation}
where $\omega_{0,2i}$ refers to the $\Delta I=1/2,3/2$ component of
${\cal O}_i$. We present numerical results for $\omega_{0,2i}$ in
Table~\ref{t: omega}.
\begin{table}[tbh]
\centering
\caption[]{Numerical coefficients for Eq.~\ref{numcoe}}
\begin{tabular}{|c|c|c|} \hline
Operator & $\omega_0$ & $\omega_2$ \\ \hline
${\cal O}_{qq}^{(1,1)}$ & $-0.06$ & 0 \\
${\cal O}_{qq}^{(8,1)}$ & $-0.3$ & 0 \\
${\cal O}_{qq}^{(1,3)}$ & $-0.3$ & 0 \\
${\cal O}_{qq}^{(8,3)}$ & $0.32$ & 0 \\ \hline
${\cal O}_{dd}^{(1)}$ & $0.08$ & $2.5$ \\
${\cal O}_{ud}^{(1)}$ & $-0.02$ & $-2.5$ \\
${\cal O}_{dd}^{(8)}$ & $0.1$ & $3.3$ \\
${\cal O}_{ud}^{(8)}$ & $0.2$ & $-3.3$ \\ \hline
${\cal O}_{qu}^{(1)}$ & $2.4$ & $-36.8$ \\
${\cal O}_{qu}^{(8)}$ & $0.1$ & $3.3$ \\
${\cal O}_{qd}^{(1)}$ & $1.5$ & 0 \\
${\cal O}_{qsd}^{(1)}$ & $-3.9$ & $36.8$ \\
${\cal O}_{qsd}^{(8)}$ & $-0.1$ & $-3.3$ \\ \hline
${\cal O}_{qsq}^{(1)}$ & $1.8$ & $-31.2$ \\
${\cal O}_{qsq}^{(8)}$ & $-3.5$ &$33$ \\
${\cal O}_{qqs}^{(1)}$ & $-3.5$ & $31$ \\
${\cal O}_{qqs}^{(8)}$ & $2.1$ & $-33$ \\
${\cal O}_{qq}^{(1s)}$ & $1.8$ & 0 \\
${\cal O}_{qq}^{(8s)}$ & $1.3$ & 0
\\ \hline
\end{tabular}
\label{t: omega}
\end{table}
Requiring that $\Omega_{NEW} < 1$, we can constrain the size of the
${\cal CP}$ violating couplings Im~$\lambda_i/\Lambda^2$. By assuming
that there is no accidental cancellation between the contributions
of different operators to $\Omega_{NEW}$ we may constrain each
operator separately. The isospin decomposition is useful because
it is possible to construct combinations of operators with definite
isospin transformation properties. The constraints that apply to
operators that are purely $\Delta I =1/2$ are different from those
that apply to operators that are purely $\Delta I = 3/2$.
\subsection{$K^0-\overline{K}^0$ Mixing and $\epsilon$}
In general, $\epsilon^\prime$ provides tighter constraints on new
${\cal CP}$ violating interactions that does $\epsilon$. Nevertheless,
it is necessary to consider constraints from $\epsilon$ because
the ones that arise from $\epsilon^\prime$ do not apply to parity
conserving operators that do not contribute to the decay $K^0 \rightarrow
\pi \pi$. In the operator basis that we are using, all the operators have
parity conserving and violating components. However, it is possible
to construct parity conserving combinations of operators just as it
is possible to construct combinations of operators with definite
isospin.
All of the $|\Delta S|=1$ four-quark
operators that we consider
contribute to $\epsilon$ when combined with a second
$|\Delta S|=1$ vertex from the usual weak Hamiltonian.
In terms of the $K^0-\overline{K}^0$ mixing matrix, each operator gives a
contribution to $\epsilon$ of the form:
\begin{equation}
|\epsilon|_i \approx {1 \over \sqrt{2}}{|{\rm Im}M_{12}|_i \over \Delta m_k}
\end{equation}
We estimate the long distance contributions to Im$M_{12}$
due to intermediate pion and eta poles \cite{dohowe}.
Using the matrix elements of Table~\ref{t: matelep} we find that there
is a cancellation between the contributions of the pion
and octet-eta poles at leading order in $SU(3)$ breaking. This
situation is unfortunate because it makes the estimates very unreliable.
For our purposes we will use the model of Ref.~\cite{dhlin}
to deal with this problem.
The contribution of each operator to $\epsilon$ is given by:
\begin{eqnarray}
|\epsilon|_{{\cal O}_i} & = & \sqrt{2} g^2 {g_8 \over M_W^2}\biggl(
{m_K \over \Delta m_K}\biggr) |{\rm Im} \lambda_i|
\biggl({M_W \over \Lambda}\biggr)^2{m_K^2 \over m_K^2-m_\pi^2}
|\xi_i|\nonumber \\
&\approx& 9.3 \ |{\rm Im} \lambda_i|
\biggl({M_W \over \Lambda}\biggr)^2|\xi_i|
\end{eqnarray}
where $g_8$ is defined in Eq.~\ref{smkpipole}, and
$\xi_i$ is given according to the model of Ref.~\cite{dhlin} by:
\begin{eqnarray}
\xi_i &=& {1\over \sqrt{2}f_\pi^2 m_K^2}
{f_K \over f_\pi} \biggl\{<\pi^0|{\cal O}_i|\overline{K}^0>
+ {<\eta_8|{\cal O}_i|\overline{K}^0> \over \sqrt{3} }
\nonumber \\
&&\cdot \biggl\{\biggl({m_K^2 - m_\pi^2 \over m_K^2 - m_\eta^2}\biggr)
\biggl[(1+\xi)\cos\theta+2\sqrt{2}\rho\sin\theta\biggr]
\biggl[{f_{\eta_8}\over f_\pi}\cos\theta-\sqrt{2}
{f_{\eta_0} \over f_\pi}\sin\theta\biggr] \nonumber \\
&&+ \biggl({m_K^2 - m_\pi^2 \over m_K^2 - m_\eta^{\prime 2}}\biggr)
\biggl[(1+\xi)\sin\theta-2\sqrt{2}\rho\cos\theta\biggr]
\biggl[{f_{\eta_8}\over f_\pi}\sin\theta+\sqrt{2}
{f_{\eta_0} \over f_\pi}\cos\theta\biggr] \biggr\}\biggr\}
\label{newep}
\end{eqnarray}
We choose the parameters that Ref.~\cite{dhlin} considers more
physical: $\theta = -20^\circ$, $\xi = 0.17$, $f_{\eta 8} =
1.25 f_\pi$, $f_{\eta 0}=1.04 f_\pi$.
Once more we present separate results for the $\Delta I = 1/2, 3/2$
components of each operator in Table~\ref{t: xi}.
We emphasize again that we present
our results in this form because it is possible to construct
combinations of operators that
have definite isospin transformation properties. For the
$\Delta I = 1/2$ component, there is sensitivity to the parameters
in the model of Ref.~\cite{dhlin}. We illustrate this by presenting
results for $\rho=0.8$, $\rho = 1.2$, and for just the pion pole.
For the $\Delta I= 3/2$ component there is only a pion pole.
\begin{table}[tbh]
\centering
\caption[]{Factors $\xi_i$ for Eq.~\ref{newep}.}
\begin{tabular}{|c|c|c|c|c|} \hline
Operator & $\xi_{i,1/2} (\rho =0.8)$ & $\xi_{i,1/2} (\rho =1.2)$
& $\xi_{i,1/2}$($\pi$-only) & $\xi_{i,3/2}$ \\ \hline
${\cal O}_{qq}^{(1,1)}$ & $ -0.24 $ & $0.14$ & $-0.04$ & $0$ \\
${\cal O}_{qq}^{(8,1)}$ & $-0.41 $ & $-0.12$ & $-0.22$ & $0$ \\
${\cal O}_{qq}^{(1,3)}$ & $-0.33 $ & $-0.17$ & $-0.21$ & $0$ \\
${\cal O}_{qq}^{(8,3)}$ & $-0.16 $ & $0.71$ & $0.22$ & $0$ \\ \hline
${\cal O}_{dd}^{(1)}$ & $-0.18 $ & $0.04$ & $-0.06$ & $-0.11$ \\
${\cal O}_{ud}^{(1)}$ & $-0.06 $ & $0.1$ & $0.01$ & $0.11$ \\
${\cal O}_{dd}^{(8)}$ & $-0.23 $ & $0.06$ & $-0.07$ & $-0.15$ \\
${\cal O}_{ud}^{(8)}$ & $-0.18 $ & $-0.18$ & $-0.15$ & $0.15$ \\ \hline
${\cal O}_{qu}^{(1)}$ & $-1.8 $ & $-1.7$ & $-1.5$ & $1.5$ \\
${\cal O}_{qu}^{(8)}$ & $-0.05 $ & $0.24$ & $0.07$ & $0.15$ \\
${\cal O}_{qd}^{(1)}$ & $-4.2 $ & $-1.2$ & $-2.2$ & $0$ \\
${\cal O}_{qsd}^{(1)}$ & $-2.4 $ & $0.57$ & $-0.75$ & $-1.5$ \\
${\cal O}_{qsd}^{(8)}$ & $-0.23 $ & $0.06$ & $-0.07$ & $-0.15$ \\ \hline
${\cal O}_{qsq}^{(1)}$ & $-1.7 $ & $-2.2$ & $-1.6$ & $1.2$ \\
${\cal O}_{qsq}^{(8)}$ & $0.46 $ & $-2.1$ & $-0.65$ & $-1.3$ \\
${\cal O}_{qqs}^{(1)}$ & $0.22 $ & $-2.6$ & $-1.0$ & $-1.2$ \\
${\cal O}_{qqs}^{(8)}$ & $-1.6 $ & $-1.6$ & $-1.3$ & $1.3$ \\
${\cal O}_{qq}^{(1s)}$ & $-1.5 $ & $-4.8$ & $-2.6$ & $0$ \\
${\cal O}_{qq}^{(8s)}$ & $-1.1$ & $-3.7$ & $-2.0$ & $0$ \\ \hline
\end{tabular}
\label{t: xi}
\end{table}
\subsection{$\Lambda \rightarrow p \pi^-$ and $A(\Lambda^0_-)$}
The starting point of the calculation is Eq.~\ref{alambda}. We study the
effect of the new physics one operator at a time and always assume that
the ${\cal CP}$ violating amplitudes are small, so that the
experimental value of the amplitudes is approximately equal to
the ${\cal CP}$ conserving amplitude. All the ${\cal CP}$ violating phases are then small
and we can write:
\begin{eqnarray}
A(\Lambda^0_-) &\approx & 3 \times 10^{-5} \nonumber \\
&+& \sum_i \biggl( 0.13 (\phi^p_1-\phi^s_1)
+0.001 (\phi^p_1-\phi^s_3) -0.0024 (\phi^p_3-\phi^s_1)\biggr)_i
\label{anew}
\end{eqnarray}
where the sum runs over all the operators in Eq.~\ref{effh}.
We carry out the calculation in the same manner as the standard model
analysis of the previous section \cite{steger}.
That is, we compute the imaginary
part of the amplitudes by taking matrix elements of each new four-quark
operator
in vacuum saturation. Further, we will not compute perturbative QCD
corrections to the effective Hamiltonian of the new physics sector.
We will also assume that the new physics does not
significantly alter the ${\cal CP}$ conserving amplitudes, but we will
comment on this later on. As discussed in
Ref.~\cite{steger}, this vacuum saturation calculation is not reliable at all,
nevertheless, we will use it for lack of anything better.
Calculating the imaginary part of the amplitudes
taking the real part from experiment as in
the previous section, we find that each operator ${\cal O}_i$ induces
the following phases:
\begin{eqnarray}
\biggl(\phi^p_1\biggr)_i &=&
-8 {G_F \over \sqrt{2}}
\biggl({M_W \over \Lambda}\biggr)^2 {\rm Im}\lambda_i
{<p \pi^-|{\cal O}_i|\Lambda>^P_1 \over 9.98 G_F m_\pi^2 }
\nonumber \\
\biggl(\phi^s_1\biggr)_i &=&
8 {G_F \over \sqrt{2}}
\biggl({M_W \over \Lambda}\biggr)^2 {\rm Im}\lambda_i
{<p \pi^-|{\cal O}_i|\Lambda>^S_1 \over 1.47G_F m_\pi^2 }
\nonumber \\
\biggl(\phi^p_3\biggr)_i &=&
8 {G_F \over \sqrt{2}}
\biggl({M_W \over \Lambda}\biggr)^2 {\rm Im}\lambda_i
{<p \pi^-|{\cal O}_i|\Lambda>^P_3 \over 0.21 G_F m_\pi^2 }
\nonumber \\
\biggl(\phi^s_3\biggr)_i &=&
-8 {G_F \over \sqrt{2}}
\biggl({M_W \over \Lambda}\biggr)^2 {\rm Im}\lambda_i
{<p \pi^-|{\cal O}_i|\Lambda>^S_3 \over 0.03 G_F m_\pi^2}
\nonumber \\
\label{hypphase}
\end{eqnarray}
The matrix elements are estimated in vacuum saturation and listed
in Table~\ref{t: matel} in the appendix. Numerically we find:
\begin{equation}
A(\Lambda^0_-) \approx 3 \times 10^{-5}
+ \biggl({M_W \over \Lambda}\biggr)^2 \sum_i {\rm Im}\lambda_i a_i
\label{newal}
\end{equation}
where the coefficients $a_i$ are listed in
Table~\ref{t: hyp}. We present two different values:
in the first column we include only the $\Delta I = 1/2$ component of each
operator, whereas in the second column we include both isospin
components. We can see from Table~\ref{t: hyp} that the ${\cal CP}$
violating asymmetry $A(\Lambda^0_-)$ is dominated by the interference
of the $s$ and $p$ waves in the $\Delta I =1/2$ amplitude, as can be
anticipated from Eq.~\ref{alambda}.
We also see from Table~\ref{t: hyp} that $a_i$ is of order one in some cases.
Eq.~\ref{newal} then tells us that a measurement of $A(\Lambda^0_-)$ at the
$10^{-4}$ level is sensitive, in principle, to new ${\cal CP}$ violating interactions
generated at a scale $\Lambda \leq 8$~TeV and is thus potentially interesting.
\begin{table}[tbh]
\centering
\caption[]{Factors $a_i$ for $A(\Lambda^0_-)_{NEW}$ in Eq.~\ref{newal}.}
\begin{tabular}{|c|c|c|} \hline
Operator &
$(a_i)_{1/2}$ & $a_i$ \\ \hline
${\cal O}_{qq}^{(1,1)}$ & $0.03$ & $0.03$ \\
${\cal O}_{qq}^{(8,1)}$ & $0.15$ & $0.14$ \\
${\cal O}_{qq}^{(1,3)}$ & $0.14$ & $0.14$ \\
${\cal O}_{qq}^{(8,3)}$ & $-0.15$ & $-0.15$ \\ \hline
${\cal O}_{dd}^{(1)}$ & $-0.05$ & $-0.06$ \\
${\cal O}_{ud}^{(1)}$ & $0.01$ & $0.02$ \\
${\cal O}_{dd}^{(8)}$ & $-0.06$ & $-0.08$ \\
${\cal O}_{ud}^{(8)}$ & $-0.1$ & $-0.1$ \\ \hline
${\cal O}_{qu}^{(1)}$ & $-1.3$ & $-1.1$ \\
${\cal O}_{qu}^{(8)}$ & $-0.05$ & $-0.08$ \\
${\cal O}_{qd}^{(1)}$ & $1.5$ & $1.5$ \\
${\cal O}_{qsd}^{(1)}$ & $-0.6$ & $-0.8$ \\
${\cal O}_{qsd}^{(8)}$ & $0.05$ & $0.08$ \\ \hline
${\cal O}_{qsq}^{(1)}$ & $-1.4$ & $-1.2$ \\
${\cal O}_{qsq}^{(8)}$ & $0.6$ & $0.7$ \\
${\cal O}_{qqs}^{(1)}$ & $-0.9$ & $-1.0$ \\
${\cal O}_{qqs}^{(8)}$ & $-1.1$ & $-1.0$ \\
${\cal O}_{qq}^{(1s)}$ & $1.8$ & $1.8$ \\
${\cal O}_{qq}^{(8s)}$ & $1.4$ & $1.3$
\\ \hline
\end{tabular}
\label{t: hyp}
\end{table}
We can use the constraints from ${\cal CP}$ violation in kaon decays to
place bounds on the magnitude of $A(\Lambda^0_-)$ that each of the
four quark operators can induce. In general, the bounds coming from
direct ${\cal CP}$ violation in $\epsilon^\prime$ are stronger than those
coming from $\epsilon$. However, it is necessary to consider both
because it is possible to construct parity conserving combinations of
operators that do not contribute to $K\rightarrow \pi \pi$ amplitudes and, thus,
evade the bounds from $\epsilon^\prime$. Similarly, $\epsilon^\prime$
places stronger constraints on $\Delta I =3/2$ operators than on $\Delta I
=1/2$ operators due to the enhancement factor of $1/\omega$ in
Eq.~\ref{epprime}. To take into account these distinctions, we list
in Table~\ref{t: result} the bounds on each of the weak phases separately.
The blank entries indicate that there is no bound because the particular
operator does not contribute to that amplitude.
\begin{table}[tbh]
\centering
\caption[]{Bounds on the phases that enter $A(\Lambda^0_-)$.}
\begin{tabular}{|c|c|c|c|c|} \hline
Operator & $\phi^p_1 \times 10^{5}$ & $\phi^s_1\times 10^{5}$
& $\phi^p_3\times 10^{5}$ & $\phi^s_3\times 10^{5}$\\ \hline
${\cal O}_{qq}^{(1,1)}$ & $2.9$ & $-10$ & $--$ & $--$ \\
${\cal O}_{qq}^{(8,1)}$ & $9.0$ & $-10$ & $--$ & $--$ \\
${\cal O}_{qq}^{(1,3)}$ & $10$ & $-10$ & $--$ & $--$ \\
${\cal O}_{qq}^{(8,3)}$ & $-24$ & $-10$ & $--$ & $--$ \\ \hline
${\cal O}_{dd}^{(1)}$ & $5.3$ & $-10$ & $400$ & $-16$ \\
${\cal O}_{ud}^{(1)}$ & $-3.7$ & $-10$ & $400$ & $-16$ \\
${\cal O}_{dd}^{(8)}$ & $5.3$ & $-10$ & $400$ & $-16$ \\
${\cal O}_{ud}^{(8)}$ & $14$ & $-10$ & $400$ & $-16$ \\ \hline
${\cal O}_{qu}^{(1)}$ & $14$ & $-9.3$ & $400$ & $-15$ \\
${\cal O}_{qu}^{(8)}$ & $-24$ & $-10$ & $400$ & $-16$ \\
${\cal O}_{qd}^{(1)}$ & $9$ & $22$ & $--$ & $--$ \\
${\cal O}_{qsd}^{(1)}$ & $5.4$ & $2.8$ & $-400$ & $-15$ \\
${\cal O}_{qsd}^{(8)}$ & $5.3$ & $-10$ & $400$ & $-16$ \\ \hline
${\cal O}_{qsq}^{(1)}$ & $16$ & $-14$ & $400$ & $-15$ \\
${\cal O}_{qsq}^{(8)}$ & $24$ & $-2.8$ & $-400$ & $15$ \\
${\cal O}_{qqs}^{(1)}$ & $-74$ & $4.2$ & $400$ & $-15$ \\
${\cal O}_{qqs}^{(8)}$ & $14$ & $-9.3$ & $400$ & $-15$ \\
${\cal O}_{qq}^{(1s)}$ & $29$ & $22$ & $--$ & $--$ \\
${\cal O}_{qq}^{(8s)}$ & $29$ & $22$ & $--$ & $--$ \\ \hline
\end{tabular}
\label{t: result}
\end{table}
The bounds on the $p$-wave phases arise from the contributions of the
operator to $\epsilon$, and are weaker than the bounds on the $s$-wave
phases that arise from the contributions to $\epsilon^\prime$.
For the operator basis that we have been using, the bounds on the
different components are not independent. This, however, is not an
important point because there is nothing special about this operator
basis. We prefer to illustrate separately the bounds on each parity
and isospin amplitude because it
is possible to construct operators with definite parity and isospin.
\section{Two-quark Operators of Dimension six}
In addition to the four-quark operators of dimension six considered in the
previous section, there are also two-quark operators of dimension six that
can contribute to the processes under consideration \cite{buchmuller}.
These operators are the $SU(3)\times SU(2) \times U(1)$ invariant versions
of ``penguin'' operators that naively appear to be dimension five
\cite{rujula}. There are two types of operators that contribute to
${\cal CP}$ violation in $|\Delta S|=1$ processes. The first one, in the
notation of Ref.~\cite{buchmuller} is:
\begin{equation}
{\cal O}_{dG} = (\overline{q}\sigma_{\mu\nu}\lambda^a d)\phi G^a_{\mu\nu}
\end{equation}
The operator of interest to us is obtained when the scalar
doublet, $\phi$, takes
its vacuum expectation value. This leads to the effective Lagrangian
(with the same overall normalization that we used before and
$v \approx 246$~GeV):
\begin{eqnarray}
{\cal L}_p&=&{g^2 \over \Lambda^2}{v \over \sqrt{2}}\biggl[
\lambda_{ds}\overline{d}\sigma_{\mu\nu}\lambda^a\biggl({1+\gamma_5 \over 2}\biggr)
sG^a_{\mu\nu} + \lambda^\star_{sd}\overline{d}\sigma_{\mu\nu}\lambda^a\biggl(
{1-\gamma_5\over 2}\biggr)sG^a_{\mu\nu}\biggr]+{\rm h.~c.}
\nonumber \\
&\equiv & {g^2 \over \Lambda^2}{v \over \sqrt{2}}
\overline{d}\sigma_{\mu\nu}t^a(f_{pc}+\gamma_5 f_{pv})sG^a_{\mu\nu} +{\rm h.~c.}
\label{penguinop}
\end{eqnarray}
There are also analogous operators where the Gluon field strength tensor
is replaced by field strength tensors for electroweak gauge bosons. The
matrix elements of these operators are suppressed by a power of
$\alpha = 1/137$ with respect to the gluon operator and we will,
therefore, neglect them.
\subsection{Constraint on the parity conserving coupling}
The parity conserving coupling $f_{pc}$ is constrained by the contribution
of Eq.~\ref{penguinop} to the parameter $\epsilon$. Unlike the four-quark
operators of the previous section, we cannot use vacuum saturation to compute
the matrix elements of this operator. However, this is the same
operator that arises in the Weinberg model of ${\cal CP}$ violation, and the
analysis has been carried out by Donoghue and Holstein \cite{dohowe}
using MIT bag model matrix elements. We can simply take over their
results to find:
\begin{equation}
|\epsilon|_p \approx 1.5 \times 10^{5} |\xi| \biggl({M_W\over
\Lambda}\biggr)^2|{\rm Im}f_{pc}|
\label{epconp}
\end{equation}
This contribution to $\epsilon$ is due to long distance effects as those
discussed in the previous section. In complete analogy we have introduced
the parameter $\xi$ which takes the values $\xi = 0.12$ for $\rho =0.8$
and $\xi = -0.48$ for $\rho =1.2$. We find that the sensitivity of the
result to the $SU(3)$ breaking parameters of the pole model is larger
in this case than it was for the four-quark operators.
\subsection{Constraint on the parity violating coupling}
The constraint on the parity violating coupling $f_{pv}$ comes from an
analysis of $\epsilon^\prime$. Just as we did for $f_{pc}$, we simply
take over the results of Ref.~\cite{dohowe} with a suitable identification
of the coupling. We find:
\begin{equation}
\biggl|{\epsilon^\prime \over \epsilon}\biggr|_p \approx
2.2 \times 10^5 \biggl({M_W \over \Lambda}\biggr)^2
|{\rm Im}f_{pv}|
\label{eppconp}
\end{equation}
\subsection{Contribution to $A(\Lambda^0_-)$}
Once again we use the fact that up to coupling constants, this operator
is the same one appearing in the Weinberg model of ${\cal CP}$ violation. Its
matrix elements using the MIT bag model can thus be taken from
Ref.~\cite{donpa}. We find:
\begin{eqnarray}
\phi^s_1 &\approx & 7 \times 10^4 \biggl({M_W \over \Lambda} \biggr)^2
{\rm Im}f_{pv} \nonumber \\
\phi^p_1 & \approx & -8 \times 10^4 \biggl({M_W \over \Lambda} \biggr)^2
{\rm Im}f_{pc}
\end{eqnarray}
{}From these it follows that:
\begin{equation}
A(\Lambda^0_-)_p\approx -10^4 \biggl({M_W \over \Lambda} \biggr)^2
\biggl({\rm Im}f_{pc}+0.9{\rm Im}f_{pv}\biggr)
\end{equation}
In the Weinberg model this operator appears with $f_{pv}=-f_{pc}$ and
there is a large cancellation between the two phases leading to a
smaller value for $A(\Lambda^0_-)$ than would have been obtained from
each phase individually \cite{donpa}.
In our general operator analysis, the bounds
from Eq.~\ref{epconp} and Eq.~\ref{eppconp} can be combined to obtain
(with $\xi =-0.5$):
\begin{eqnarray}
\biggl({M_W \over \Lambda} \biggr)^2 {\rm Im}f_{pc} &<& 2.7 \times 10^{-8}
\nonumber \\
\biggl({M_W \over \Lambda} \biggr)^2 {\rm Im}f_{pv} &<& 6.8 \times 10^{-9}
\end{eqnarray}
or in terms of the hyperon decay observable:
\begin{equation}
A(\Lambda^0_-) \leq \cases{ 3 \times 10^{-4} & Parity conserving operator \cr
6 \times 10^{-5} & Parity violating operator \cr}
\label{respen}
\end{equation}
Before ending this section we should comment on one class of two-quark
operators that we have not discussed. In the notation of Ref.~\cite{buchmuller}
it is:
\begin{equation}
{\cal O}_{qG} = i(\overline{q}\lambda^a \gamma_\mu D_\nu d)\phi G^a_{\mu\nu}
\label{other}
\end{equation}
and related operators with field strength tensors for electroweak gauge
bosons instead of the gluon. These latter ones will have matrix elements
suppressed by $\alpha$ compared to Eq.~\ref{other}. We have not found a
simple way to estimate the matrix elements of this operators and for this
reason we do not discuss them in detail. We do not expect the behavior of this
type of operator to be significantly different from the others that we have
discussed.
\section{Summary and Conclusions}
The minimal standard model of electroweak interactions
is in extraordinary agreement with all experiments conducted so far,
and there is no evidence for any new particles below $100$~GeV or so.
In view of this, it is reasonable to assume that any new physics beyond
the minimal standard model is associated with a scale $\Lambda \geq M_W$,
and it is, therefore, possible to represent the low energy effects of any
such new physics with an effective Lagrangian that respects the symmetries
of the standard model.
In this paper we have studied all the $|\Delta S| =1$, ${\cal CP}$ violating,
operators that occur at dimension six. We have investigated the constraints
that exist on the couplings of these operators from the measurements of
$\epsilon$ and $\epsilon^\prime$, and estimated what their largest
contribution to ${\cal CP}$ violation in $A(\Lambda^0_-)$ could be.
The operators that we have discussed also contribute
to CP conserving and flavor changing amplitudes. We might thus worry,
that the constraints on the real part of the couplings are such, that
it is not natural for the imaginary (${\cal CP}$ violating) part of the
couplings to attain the upper bounds allowed by the values of $\epsilon$
and $\epsilon^\prime$. We briefly address this issue in this section.
Consider the contributions to $K^0-\overline{K}^0$ mixing. If we fix
Im~$\lambda_i$ to its maximum allowed value, we find that the
constraint $2{\rm Re}M_{12,i} \leq \Delta m_K$ is also satisfied if:
\begin{equation}
{\rm Re}\lambda_i \leq {{\rm Im}\lambda_i \over 2 \sqrt{2}\epsilon}
\end{equation}
Therefore, the ${\cal CP}$ conserving constraint is also satisfied if
both real and imaginary parts of the couplings are of the same size
or if the imaginary part is smaller than the real part by a factor
of $\epsilon$.
The strongest constraints on flavor changing operators
in the ${\cal CP}$ conserving case are known to come from $K^0-\overline{K}^0$ mixing.
If we set the couplings to be of order one, we obtain
a lower bound on the scale of new physics $\Lambda$ requiring
that $2{\rm Re}M_{12,i} \leq \Delta m_K$. It is easy to check that
with couplings and scales satisfying this bound, the new operators do
not make any significant contributions to the real part of the amplitudes
in $K \rightarrow \pi \pi$ or $\Lambda \rightarrow p \pi$.
Therefore, we conclude that fixing the imaginary part of the couplings to
their maximum allowed value is not in conflict with ${\cal CP}$
conserving constraints.
In the minimal standard model we have estimated previously \cite{steger}
that $A(\Lambda^0_-)$ is of the order of a few times $10^{-5}$. For the
new physics considered in this paper we find that most of the operators
would naturally induce contributions to $A(\Lambda^0_-)$ at the $10^{-5}$
level, making them indistinguishable from the minimal standard model
(as long as precise calculations of the matrix elements are not available),
and inaccessible to the search to be conducted by E871. However, we have
also found that for certain operators, ${\cal O}_{qqs}^{(1)}$ and
${\cal O}_{dG}$, $A(\Lambda^0_-)$ could be as large as a few times
$10^{-4}$.
Given our crude estimate of the hadronic matrix elements involved,
all our numerical results should be viewed with caution. Nevertheless,
our results suggest that the search for ${\cal CP}$ violation in $A(\Lambda^0_-)$
at the $10^{-4}$ level of sensitivity that is expected for E871 is
potentially very interesting. Our results also suggest that this
measurement is complementary to the measurement of $\epsilon^\prime/
\epsilon$, in that it probes potential sources of ${\cal CP}$ violation
at a level that has not been probed by the kaon experiments. This
is particularly true for parity conserving interactions that do not
contribute to $\epsilon^\prime$ and are only constrained by $\epsilon$.
We conclude that it is possible for E871 to observe a ${\cal CP}$ violating
signal at the $10^{-4}$ level. Our study indicates that if such a signal
is observed, it would probably be evidence for physics beyond the minimal
standard model. However, a reliable determination of hadronic matrix
elements is necessary to reach any definite conclusion.
\section*{Acknowledgements}
The work of G.V. was supported in part by a DOE OJI award under contract
number DE-FG02-92ER40730. The work of X.G.He was supported in part by
DOE contract number DE-FG06-85ER40224.
G.V. thanks the Institute of Theoretical Science
at the University of Oregon for their hospitality while part of this work
was performed. We are grateful to J. F. Donoghue and N. Deshpande for
helpful discussions.
\clearpage
|
\section{#1} \setcounter{equation}{0}}
\newtheorem{theo}{Theorem}[section]
\begin{document}
\newtheorem{lem}[theo]{Lemma}
\newtheorem{prop}[theo]{Proposition}
\newtheorem{coro}[theo]{Corollary}
\title{Controlled Geometry via Smoothing \footnote{ 1991 {\em Mathematics
Subject Classification}. Primary 53C20.}}
\author{Peter Petersen\thanks{Partially supported by NSF and NYI grants} \\
\and Guofang Wei\thanks {Partially supported by NSF Grant \# DMS9409166.
}\\
\and Rugang Ye\thanks{Partially supported by NSF Grant \# DMS9401106}
}
\date{}
\maketitle
\begin{abstract}
We prove that Riemannian metrics with a uniform weak norm can be smoothed to
having arbitrarily high regularity. This generalizes all previous smoothing
results. As a consequence we obtain a generalization of Gromov's almost flat
manifold theorem. A uniform Betti number estimate is also obtained.
\end{abstract}
\newcommand{\mbox{Tr}}{\mbox{Tr}}
\newcommand{\mbox{inj}}{\mbox{inj}}
\newcommand{\mbox{vol}}{\mbox{vol}}
\newcommand{\mbox{diam}}{\mbox{diam}}
\newcommand{\mbox{Ric}}{\mbox{Ric}}
\newcommand{\mbox{conj}}{\mbox{conj}}
\newcommand{\mbox{Hess}}{\mbox{Hess}}
\newcommand{\mbox{div}}{\mbox{div}}
\sect{Introduction}
An ultimate goal in geometry is to achieve a classification scheme, using
natural geometric quantities to characterize the topological type
or diffeomorphism type of Riemannian manifolds. While this grand scheme seems
to be an impossible dream, its basic philosophy has been
a driving force in many important developments in Riemannian geometry. The
sphere theorems and various topological finiteness theorems
are typical examples. These results are concerned with control of global
topology of manifolds, and a crucial point therein is to control,
uniformly, the local topology.
Control of local topology often follows from control of local
geometry. Here, by local geometry, we mean the local behavior of the metric
tensor. On the other hand, control of local geometry is frequently also
the essential ingredient for control of global geometry, such as in
Cheeger-Gromov's compactness theorem and its various extensions,
which
can be named geometric finiteness theorems. Notice that some rudimental
topological finiteness results are direct corollaries of
geometric
finiteness theorems. But the significance of the latter goes beyond this. In
any case, control of local geometry is obviously a key
topic. An interesting and important aspect of this topic is various degrees of
control of local geometry needed or available in different
situations. In \cite{p}, the first author introduced a sequence of norms which
provide a certain quantitative measure for local geometric
control. These norms can be defined either in terms of the
$C^{k,\alpha}$-norms
or the $L^{k,p}$-norms for functions, and they are defined on a given
scale. For example, the $C^{k,\alpha}$-norm of a Riemannian manifold on scale
$r$ is bounded, if it is covered by coordinate charts of size comparable to $r$
such that the metric tensor expressed in the coordinates
is uniformly bounded in $C^{k, \alpha}$-norm, and that the coordinate
transition functions are uniformly bounded
in $C^{k+1, \alpha}$-norm. Note that the local topology is uniformly trivial if
one of these norms on some scale is bounded. To admit richer topological and
geometric structures under norm bounds, we shall introduce a weak version
of these norms. The essential new feature is that we allow coordinate maps
to have double points. In spirit, this is similar to replacing an injectivity
radius bound by a conjugate radius bound. (Of course, e.g. a weak (harmonic)
$C^{0, \alpha}$ bound is so weak that it is far from implying a conjugate
radius bound.)
Indeed, our basic theme is to try to find the minimal degree of control of
local geometry under which interesting geometric and
topological
consequences can be
drawn. Traditional geometric conditions such as curvature bounds imply various
degrees of local geometric control, so we can think from their perspectives.
Historically, sectional curvature bounds were the first to be systematically
studied. They can roughly be compared with weak $C^2$-norm bounds, at least
the latter imply
the former. Since understanding of sectional curvature bounds has been
reached on a
good level, it is natural to try to find the minimal degree of local geometric
control under which a metric can be approximated by metrics with sectional
curvature bounds or weak $C^2$-norm bounds (or better bounds). In this paper,
we present a result towards this goal,
along with some applications. The local geometric control we need is as weak
as a bound on the weak harmonic $C^{0, \alpha}$-norm, or a bound on the weak
$L^{1, p}$-norm. These do appear to be the sought-after minimal
degree of local geometric control in our set-up.
To formulate the result precisely, we introduce the following classes of
Riemannian manifolds. (The definition of the weak norms are given in \S 2.)
\newline
\noindent {\bf Definition}. Given $n \geq 2$, $0 < \alpha < 1$, $p > n$ and
function $Q: (0, \infty) \rightarrow [0, \infty)$ which is nondecreasing in $r$ and
satisfies $\lim_{r \rightarrow 0} Q(r) = 0$, we define
\[
{\cal M}(n, \alpha, Q) = \left\{ (M, g) \left| \begin{array}{l} (M,g) \mbox{ is
a complete Riemannian manifold}, \dim M = n, \\
\mbox{the weak harmonic}\ C^{0, \alpha}\ \mbox{norm}\
\|(M,g)\|^{W,h}_{C^{0,\alpha}, r} \leq Q(r)\\ \mbox{ for all positive }\ r \leq
1
\end{array} \right.
\right\},
\]
and
\[
{\cal M}(n, p, Q) = \left\{ (M, g) \left| \begin{array}{l} (M,g) \mbox{ is a
complete Riemannian manifold}, \dim M = n, \\
\mbox{the weak}\ L^{1,p}\ \mbox{norm}\ \|(M,g)\|^{W}_{L^{1,p}, r} \leq Q(r)\\
\mbox{ for all positive}\ r\leq 1
\end{array} \right.
\right\}.
\]
\noindent {\em Remark} 1. By Anderson-Cheeger's work \cite{ac} manifolds with a lower bound for
Ricci curvature and a positive lower bound for conjugate radius belong to these
two classes.
\noindent {\em Remark} 2. It is unknown whether (weak) $L^{1,p}$ bounds imply (weak) harmonic
$C^{0, \alpha}$ bounds. (By the Sobole
v embedding, they do imply $C^{0,\alpha}$ bounds, but not yet harmonic $C^{0,
\alpha}$ bounds. ) In other words,
it is unknown whether controlled $C^{0,\alpha}$ harmonic coordinates exist
under the assumption of a (weak) $L^{
1, p}$ bound. This question seems rather subtle. We plan to return to it in
the future. At the moment, we consider the said two bounds as independent
conditions. Note that one can further ask whether (weak) $C^{0, \alpha}$
bounds
imply (weak) harmonic $C^{0, \alpha}$ bounds. There does not seem to be any
evidence to support an answer in the affirmative.
\\
\begin{theo} \label{main}
For every manifold $(M,g)$ in ${\cal M}(n, \alpha, Q)$ and positive numbers
$\epsilon $, there are metrics $g_\epsilon$ on $M$ such that
\begin{eqnarray}
e^{-\epsilon}g \leq & g_\epsilon &\leq e^\epsilon g \label{t1},\\
\| (M,g_\epsilon)\|^{W}_{C^{0,\alpha}, r} & \leq & 2Q(r) \label{t2}, \\
\| (M,g_\epsilon)\|^W_{C^{k,\alpha}, r} & \leq & \widetilde{Q},
\end{eqnarray}
\noindent where $k$ is an arbitary positive integer, $0 < r \leq 1$ and
$\widetilde{Q} = \widetilde{Q} (n, k , \epsilon, \alpha, Q(r))$
denotes a positive number depending only on $n, k, \epsilon, \alpha$ and
$Q(r)$.
\end{theo}
\begin{theo} \label{main'}
For every closed manifold $(M,g)$ in ${\cal M}(n, p, Q)$ and positive numbers
$\epsilon $, there are metrics $g_\epsilon$ on $M$ such that
\begin{eqnarray}
e^{-\epsilon}g \leq & g_\epsilon &\leq e^\epsilon g \label{t1'},\\
\| (M,g_\epsilon)\|^{W}_{L^{1,p}, r} & \leq & 2Q(r) \label{t2'}, \\
\| (M,g_\epsilon)\|^W_{L^{k,p}, r} & \leq & \widetilde{Q},
\end{eqnarray}
\noindent where $k$ is an arbitary positive integer, $0 < r \leq 1$ and
$\widetilde{Q} = \widetilde{Q} (n, k , \epsilon, p, Q(r))$ denotes
a positive number depending only on $n, k, \epsilon, p$ and $Q(r)$.
\end{theo}
\noindent {\em Remark} Theorem 1.2 actually also holds for complete, non-compact manifolds. This
will be shown in a paper by the third author. \\
Thus a metric with some regularity (given by the weak norm) can be deformed or
smoothed to a nearby one with arbitrarily high regularity. In particular, {\it
manifolds with a lower bound on Ricci curvature and a positive lower bound on
conjugate radius can be smoothed.} Previous smoothing results have been
concerned with metrics with various curvature bounds, and involved two
independent techniques: the embedding method and the Ricci flow. The embedding
technique in smoothing as
used by Cheeger-Gromov \cite{cg} consists of embedding (or immersing) a given
manifold into a Euclidean space and then perturbing it
suitably by a smoothing operation, which is based on the classical convolution
process. The smoothing result in \cite{cg} is that metrics on closed manifolds
with lower and upper bounds on sectional curvatures and a positive lower bound
on injectivity radius can be smoothed to metrics with bounds on all derivatives
of the Riemann curvature tensor. Later, by embedding into a Hilbert space
instead of a finite dimensional space, Abresch \cite{a} was able to remove the
condition on injectivity radius and extend to complete manifolds. More recently
Shen \cite{sh} showed that manifolds with a lower bound on sectionl curvatures
and a positive lower bound on injectivity radius can be smoothed to having
two-sided sectional curvature bounds. The technique of Ricci flow is based on
the fundamental work of Hamilton \cite{ha}. Using this technique, Bemelmans-Min
Oo-Ruh \cite{bmr} obtained the same result as in \cite{cg} without injectivity
radius lower bound, and Shi \cite{s} obtained the same result as in \cite{a}.
Later work considers metrics with other kinds !
of curvature bound. For example,
in \cite{y1,y2}
Yang dealt with integral bounds on sectional curvatures. In \cite{dwy}, Ricci
curvature bounds were treated.
By virtue of the available constructions of controlled harmonic coordinates
under various curvature bounds, all these smoothing results are consequences
of Theorem~\ref{main} or Theorem~\ref{main'}.
As typical applications we present the following two results.
\begin{theo}[Betti number estimate] For the class of manifolds $M^n$ in \\
${\cal M}(n, \alpha, Q)$ or in ${\cal M}(n, p, Q)$, and satisfying $\mbox{diam}_M
\leq D$,
we have the estimate for the Betti numbers
\begin{equation}
\sum_i b^i(M^n) \leq C(n,D,\alpha,Q)\ \mbox{or}\ C(n,D,p,Q),\label{be}
\end{equation}
and the estimate for the number of isomorphism classes of rational homotopy
groups
\begin{equation}
\pi_q(M) \otimes Q \leq C(n,q,D,\alpha,Q)\ \mbox{or}\ C(n,q,D,p,Q)\
\mbox{for}\ q \geq 2. \label{hge}
\end{equation}
\end{theo}
(\ref{be}) follows from Theorem~\ref{main}, \ref{main'} and Gromov's uniform
betti number estimate regarding sectional curvature \cite{g2}. This estimate
can also be proved directly using Toponogov type comparison estimate introduced
in \cite{w}, see \cite{pw} for details. In \cite{w} the same estimate
(\ref{be}) is given for the class of manifolds satisfying $\mbox{Ric}_M \geq
-(n-1)H$, conj $\geq r_0$ and $\mbox{diam}_M \leq D$. (\ref{hge}) follows from
Theorem~\ref{main}, \ref{main'} and the results in \cite{r}.
\begin{theo} \label{aflat}
There exists an $\epsilon =\epsilon(n,\alpha,Q)$ or $\epsilon(n,p,Q)> 0$ such
that if a manifold $M^n$ belongs to
${\cal M}(n, \alpha, Q)$ or ${\cal M}(n, p, Q)$ and diam $\leq \epsilon$, then
$M$ is diffeomorphic to an infranilmanifold.
\end{theo}
This generalizes Gromov's almost flat manifold theorem \cite{g1} as well as its
generalization in \cite{dwy}. (The proof is simple: combine Theorem 1.1 with
\cite{g1}.)
The proof of Theorem~\ref{main} and Theorem 1.2 uses the embedding method in
\cite{a}. Roughly speaking, we embed a given Riemannian manifold into
the Hilbert space of $L^2$-functions on it, and then use the embedding to pull
back the $L^2$-metric of the Hilbert space.
The crucial point is of course to find a suitable embedding, such that the
pull-back metric will enjoy nice properties. In \cite{a}, the embedding is
defined in terms of distance functions. In our situation, these functions are
not appropriate, and we employ instead solutions of a canonical geometric
partial differential equation. Now if e.g. the harmonic $C^{0,\alpha}$-norm of
the manifold is bounded, then a uniform pointwise bound on sectional curvatures
will hold for the
pull-back metric, and hence we can apply the smoothing results for metrics
with sectional curvature bounds as given e.g. in \cite{a} or \cite{s}.
If we only assume that the weak harmonic $C^{0, \alpha}$-norm of the manifold
is bounded, i.e. it is in the class ${\cal M}(n, \alpha, Q)$,
the global embedding is generally not under control. To remedy the situation,
we follow the idea in \cite{a} of employing instead local embeddings. In
\cite{a}, Abresch uses the exponential map to lift local patches of the
manifold and
his local embeddings are exactly embeddings of these lifted patches. In our
situation, the exponential map is not suitable. Our substitute for it is the
coordinate maps.
Thus we use them to lift local patches, and construct embeddings of the lifted
patches via the same geometric partial differential equation as mentioned
before.
To make sure that the pull-back metrics induced by these local embeddings
descend to the local patches
and that the resulting metrics patch together to define a metric globally, it
is crucial to require the embeddings to be equivariant under isometries. Since
our embeddings are defined in terms of solutions of a canonical geometric PDE,
they
naturally share this equivariance property.
Basically, the above scheme also works for manifolds in the class ${\cal M}(n,
p, Q)$, but some modifications are necessary.
As before, the said pull-back metrics descend to
yield a new metric on the underlying manifold. But these metrics satisfy here
an integral bound on sectional curvatures rather than a poinwise bound. This is
a new situation. To handle it, we apply the Ricci flow and follow the arguments
in \cite{dwy}. A pointwise bound on Ricci curvature is used in several places
in\cite{dwy}. Since no such bound is available in our current situation, the
arguments in \cite{dwy} need to be improved and modified. The result we thus
arrive at not only completes the smoothing scheme for the class ${\cal M}(n, p,
Q)$, but also provides some new understanding of short time existence of the
Ricci flow.
\sect{Norm, Weak Norm and Smoothing}
Fix an integer $k \geq 0$ a number $0 \leq \alpha \leq 1$. The {\it
$C^{k,\alpha}$-norm } of an $n$-dimensional Riemannian manifold $(M,g)$ on
scale $r$, $\|(M,g)\|_{C^{k,\alpha}, r}$, is defined to be the infimum of
positive numbers $Q$ such that there exist embeddings: $$\varphi_s: B(0,r)
\subset R^n \rightarrow U_s \subset M$$
($B(0,r)$ denotes the closed ball of radius $r$ centered at the origin) with
images $U_s$, $s \in \cal S$ (an index set), with the following properties:\\
1) $e^{-Q} \delta_{ij} \leq g_{s,ij} \leq e^Q \delta_{ij}$, \\
2) Every metric ball $B(p, \frac{r}{10}e^{-Q}),\ p\in M$ lies in some set
$U_s$,\\
3) $r^{|j|+ \alpha} \|\partial^j g_{s,ij}\|_{C^\alpha} \leq Q$ for all
multi-indices $j$ with $0 \leq |j| \leq k$. \\
Here $g_{s,ij}$ denote the coefficients of $g_s=\varphi_s^*g$ on $B(0,r)$, and
$\delta_{ij}$ are the Kronecker symbols.
Note that this definition is slightly different from the corresponding one in
\cite{p}, where in addition the (rescaled) $C^{k+1, \alpha}$-norm of the
transition functions are required to be under control. For convenience, we can
call the $C^{k, \alpha}$-norm (of Riemannian manifolds) as defined in \cite{p}
the {\it strong $C^{k, \alpha}$-norm}. (Note however that the "strong"
harmonic $C^{k, \alpha}$-norm is equivalent to the harmonic $C^{k,
\alpha}$-norm.)
We define the harmonic $C^{k,\alpha}$-norm on scale $r$, $\|(M,g)\|^h_{C^{k,
\alpha}, r}$, by requiring additionally the following\\
4) $\varphi_s^{-1}: U_s \rightarrow R^n$ is harmonic, \\
which is equivalent to saying that \\
4$'$) id $: B(0,r) \rightarrow B(0,r)$ is harmonic with respect to $g_{s}$ on the
domain and the Euclidean metric on the target, which is in turn equivalent to
saying
that $$\sum_i \partial_i (g_s^{ij} \sqrt{ \det g_{s,ij}})= 0$$ for all $j$.
If $k \geq 1$ and $p >n$ (when $k=1$) or $p > \frac{n}{2}$ (when $k \geq 2$),
then we define the $L^{k,p}$-norm on the scale of $r$, $\|(M,g)\|_{L^{k,p},
r}$, by retaining 1) and 2), and replacing 3) by\\
3) $r^{|j|-\frac{n}{p}} \|\partial^j g_{s,ij}\|_{L^p} \leq Q$ for all $1 \leq
|j| \leq k$.
The harmonic $L^{k,p}$-norm is defined similarly. For any choice of these
norms, it is clear that the local topology is trivial on some uniform scale
for any
class of manifolds with uniformly bounded norm. (Note that the injectivity
radius may not be uniformly positive though.) To allow nontrivial local
topology, we introduce the weak norms $\|\ \ \|^W_{C^{k,\alpha},r}$ and $\|\ \
\|^W_{L^{k,p},r}$, which are defined in identical ways except that each
$\varphi_s: B(0,r) \rightarrow U_s$ is assumed to be a {\it local} diffeomorphism
instead of diffeomorphism. The corresponding weak harmonic norms $\|\ \
\|^{W,h}_{C^{k,\alpha},r}$ and $\| \ \ \|^{W,h}_{L^{k,p},r}$ are defined
in a similar way, with 4) being replaced by 4$'$).
Note that (weak) harmonic norms dominate (weak) norms on the same scale. We
also have $\|\ \ \|^W_{\ ,r} \leq \|\ \ \|_{\ ,r}$ and $\|\ \ \|^{W,h}_{\ ,r}
\leq \|\ \ \|^h_{\, r}$. All norms are continuous and non-decreasing in $r$.
If $(M,g)$ is sufficiently smooth, these norms converge to zero as $r \rightarrow 0$.
Furthermore, (weak) $C^{k,\alpha}\ (L^{k,p})$ norms vary continuously in the
$C^{k,\alpha}\ (L^{k,p})$ topology of Riemannian manifolds. See \cite{p} for
the relevant details.
We point out that $R^n$ is the only space with norm $=0$ on all scales. And
flat manifolds are the only spaces with weak norm $=0$ on all scales.
Conventional geometric conditions such as curvature bounds imply norm bounds.
Such implications are mostly contained in constructions of controlled harmonic
coordinates and are a crucial ingredient for various compactness theorems. To
have a clear perspective, we collect these results in the following
proposition.
\begin{prop}
There is a $Q(H,i_0,r,p)$ with $\lim_{r \rightarrow 0} Q(H,i_0,r,p) =0$ such that for
manifolds with \\
a) $|K| \leq H, \ \mbox{inj} \geq i_0$, then $\|(M,g)\|^h_{L^{2,p},r} \leq
Q(H,i_0,r,p)$; \\
b) $|K| \leq H$, then $\|(M,g)\|^{W,h}_{L^{2,p},r} \leq Q(H,r,p)$; \\
c) $|\mbox{Ric} | \leq (n-1)H,\ \mbox{inj} \geq i_0$, then $\|(M,g)\|^h_{L^{2,p},r} \leq
Q(H,i_0,r,p)$; \\
d) $\mbox{Ric} \geq -(n-1)H,\ \mbox{inj} \geq i_0$, then $\|(M,g)\|^h_{L^{1,p},r} \leq
Q(H,i_0,r,p)$; \\
e) $\mbox{Ric} \geq -(n-1)H$, conj $ \geq i_0$, then $\|(M,g)\|^{W,h}_{L^{1,p},r}
\leq Q(H,i_0,r,p)$.
\end{prop}
These results follow from works of Jost-Karcher \cite{jk}, Anderson \cite{an}
and Anderson-Cheeger \cite{ac}.
We now turn to the smoothing question. As explained in the introduction, our
strategy is to first achieve sectional curvature bounds by embedding into the
Hilbert space of $L^2$-functions. This is done in the next two sections. The
higher regularity smoothing then easily follows from known smoothing results.
Consider $(M,g) \in {\cal M}(n, \alpha, Q)$. We have a collection of local
diffeomorphisms
\[
\varphi_s:\ B(0,r) \rightarrow U_s \subset M
\]
satisfying 1), 2), 3) and 4$'$).
In the next section we will construct
a canonical embedding
\[
F_s: (B(0,r),g_s) \rightarrow L^2 (B(0,r), g_s),
\]
where $g_s = \varphi_s^*g$. We use $F_s$ to pullback the $L^2$ metric of $L^2
(B(0,r), g_s)$ to produce a new metric $\tilde{g}_s$ on $B(0,r)$. This
construction works for general metrics on $B(0,r)$, and has the following
equivariance property, which will be proved in \S 4. Namely, if $g_1,\ g_2$ are
two metrics on $B(0,r)$ such that there is an isometric embedding
\[
\psi: (B(0,r), g_1) \rightarrow (B(0,r),g_2)
\]
and if $\tilde{g}_1,\ \tilde{g}_2$ are obtained via the above construction,
then
\[
\psi: (B(0,r), \tilde{g}_1) \rightarrow (B(0,r), \tilde{g}_2)
\]
is also an isometric embedding. Granted this (see Proposition~\ref{4.4}) we
have
\begin{prop}
There exists a smooth metric $\bar{g}$ on $M$ such that the pullback of
$\bar{g}$ by $\varphi_s$ is exactly $\tilde{g}_s$.
\end{prop}
\noindent {\em Proof.} Let $r_1= \frac{r}{10} e^{-Q}$. Then for every $p \in M$, $B(p,r_1) \subset
U_s$ for some $s$. It follows that there exists a $\tilde{p} \in B(0,r)$ such
that $B_{g_s}(\tilde{p},r_1) \subset B(0,r)$ and $\varphi_s (\tilde{p}) = p$.
We now define the metric $\bar{g}$ as follows. If $X,Y \in T_pM$, then
\[
\bar{g}(X,Y) = \tilde{g}_s \left(((\varphi_s)_*|_{\tilde{p}})^{-1}(X),
((\varphi_s)_*|_{\tilde{p}})^{-1}(Y)\right).
\]
To show that this metric is well-defined, let $\tilde{p}'$ be another such
point, i.e. for some $s'$, $B_{g_{s'}}(\tilde{p}',r_1) \subset B(0,r)$ and
$\varphi_{s'} (\tilde{p}') =p$. Let $r_4 = \frac{r}{20} e^{-4Q}$ and $r_3 =
\frac{r}{20} e^{-3Q}$. Denote $g_0$ the Euclidean metric. Then we can show that
\begin{lem} \label{hom}
There is an isometric embedding
\[
\psi: (B_{g_0}(\tilde{p},r_4),g_s) \rightarrow (B_{g_{s'}} (\tilde{p}',r_3),g_{s'}).
\]
\end{lem}
\noindent {\em Proof.} First $\psi$ can be defined as follows.
Since $g_s$ is $e^Q$-quasi-isometric to $g_0$,
\begin{equation} \label{r34}
B_{g_0}(\tilde{p}, r_4) \subset B_{g_s} (\tilde{p}, r_3).
\end{equation}
For any point $q \in B(\tilde{p},r_4)$, connect $q$ to the center point
$\tilde{p}$ with a curve $\tilde{\gamma}$ in $B_{g_0}(\tilde{p},r_4)$ such that
the length of $\tilde{\gamma}$ $l_{g_s}(\tilde{\gamma}) < r_3$. Since
\[
\varphi_s:\ (B_{g_s} (\tilde{p}, r_3), g_s) \rightarrow B(p,r_3)
\]
is a local isometry and $\varphi_s (\tilde{p}) = p$. From (\ref{r34})
$\varphi_s$ maps the curve $\tilde{\gamma}$ to a curve $\gamma$ in $B(p,r_3)$
starting with $p$ and $l(\gamma) < r_3$. Again since $\varphi_{s'}$ is a local
isometry and $\varphi_{s'} (\tilde{p}') =p$. The curve $\gamma$ then can be
lifted via $\varphi_{s'}$ to a curve in $B_{g_{s'}}(\tilde{p}',r_3)$ starting
with $\tilde{p}'$. The other end point of this curve is defined to be the image
of $q$. (Note that, in general, lifting can not be done for incompelete space.
Here the map is a local isometry and the curve starts from the center, and we
have control on the length of the curve and the size of the metric ball, so it
will not hit the boundary during lifting.) Now we will show that $\psi$ is
well-defined, i.e. the image is independent of the choices of the curve
$\tilde{\gamma}$. If $\tilde{\gamma}_1$ is another curve in
$B_{g_0}(\tilde{p},r_4)$ connecting $q$ to the center point $\tilde{p}$ with
$l_{g_s!
}(\tilde{\gamma}_1) < r_3$, we ca
$g_s$ is $e^Q$-quasi-isometric to $g_0$. Then $\varphi_s$ maps
$\tilde{H}(s,t)$ to a homotopy $H(s,t)$ in $B(p,2r_3)$ with $l(H(s,\cdot)) <
2r_3$ for each $s$. Therefore $H(s,t)$ can be lifted via $\varphi_{s'}$ to a
homotopy in $B_{g_{s'}}(\tilde{p}',2r_3)$ starting with $\tilde{p}'$. By the
(localized) homotopy lifting lemma the other end points are all the same.
Therefore $\psi$ is well-defined.
Next we show that $\psi$ is one-to-one. Let $r_2 = \frac{r}{20} e^{-2Q}$.
Then
$$B_{g_{s'}} (\tilde{p}',r_3) \subset B_{g_0}(\tilde{p}',r_2) \subset
B_{g_{s'}} (\tilde{p}',\frac{1}{2}r_1).$$ Since $B_{g_0}(\tilde{p}',r_2)$ is
an Euclidean ball
one can construct ``inverse" $\phi$ similarly as above:
\[
\phi: \ (B_{g_{s'}} (\tilde{p}',r_3),g_{s'}) \rightarrow
(B_{g_s}(\tilde{p},\frac{1}{2}r_1),g_s).
\]
Thus $\psi$ is one-to-one. That $\psi$ is an isometric embedding follows from
the construction.
\hspace*{\fill}\rule{3mm}{3mm}\quad
Now using the equivariance, we have
\[
\psi^* \tilde{g}_{s'} = \tilde{g}_s.
\]
Therefore
\begin{eqnarray*}
\lefteqn{\tilde{g}_{s'}\left(((\varphi_{s'})_*|_{\tilde{p}'})^{-1}(X),
((\varphi_{s'})_*|_{\tilde{p}'})^{-1}(Y)\right)} \\
& = & \tilde{g}_{s'}\left((\psi)_* \circ ((\varphi_s)_*|_{\tilde{p}})^{-1}(X),
(\psi)_* \circ ((\varphi_s)_*|_{\tilde{p}})^{-1}(Y)\right) \\
& = & \psi^* \tilde{g}_{s'} \left(((\varphi_s)_*|_{\tilde{p}})^{-1}(X),
((\varphi_s)_*|_{\tilde{p}})^{-1}(Y)\right) \\
& = & \tilde{g}_s \left(((\varphi_s)_*|_{\tilde{p}})^{-1}(X),
((\varphi_s)_*|_{\tilde{p}})^{-1}(Y)\right).
\end{eqnarray*}
To show that $\varphi_s^* \bar{g} = \tilde{g}_s$, consider
\[
\varphi_s : \ B(0, \frac{9}{10}r) \rightarrow U_s.
\]
In particular, for any $\tilde{p} \in B(0, \frac{9}{10}r)$,
$B_{g_s}(\tilde{p},r_1) \subset B(0,r)$, and therefore $\tilde{p}$ can be used
to define the metric $\bar{g}$ at $\varphi_s (\tilde{p})$. It follows from the
definition that
\begin{equation} \label{mm}
\varphi_s^* \bar{g} = \tilde{g}_s.
\end{equation}
Finally, note that the smoothness of the metric $\bar{g}$ is an immediate
consequence of (\ref{mm}).
\hspace*{\fill}\rule{3mm}{3mm}\quad
\sect{Embedding I}
We continue with the above manifold $(M, g)$. Let $\Omega = B(0, r) \subset
R^n$ with a pull back metric $\varphi_s^*g$. For convenience, this metric will
be denoted by $g$. It is easy to see that $\|(\Omega, g)\|^h_{C^{0, \alpha},
r} \leq Q(r)$. We are going to construct an equivariant embedding of $(\Omega,
g)$ into $L^2(\Omega, g)$ by associating to every point $p \in \Omega$ a
geometric function $f_p \in L^2(\Omega) \equiv L^2(\Omega, g)$, which depends
nicely on $p$. A natural choice seems to be the distance function measured from
$p$. Indeed, it is used by Abresch in \cite{a}. However, under our rather
weak assumptions on the metric it is impossible to have uniform control of the
second order derivative of the distance function, which is needed to ensure
that the pull-back metric induced by the embedding satisfies a sectional
curvature bound. In fact, one can not even expect differentiability of the
distance function in balls of uniform size.
Our substitute for the distance function is solutions of a
canonical geometric partial differential equation.
Those solution functions have the crucial equivariance (
like the distance functions) and enjoy better regularity.
Many choices of ``canonical" PDE solutions are possible, e.g. in \cite{a}
Green's function is suggested. But Green's function is inconvenient because of
its singularity. We shall employ a very
simple and nicely-behaved PDE.
Denote
\[
\Omega_1 = \Omega \setminus \cup_{q \in \partial \Omega} \overline{B_g(q,
i_0)},
\]
where $i_0 =\frac{r}{10}$. ($B^g(q, \cdot)$ denotes the closed geodesic ball of
center $q$ and radius $\cdot$ measured in $g$.)
Then for $s \in \Omega_1$, let $h_s \in L^{1,2}_0 (\Omega)$ be the unique weak
solution of the following Dirichlet boundary value problem:
\begin{equation} \label{h}
\left\{ \begin{array}{rcll} \Delta h_s & = & -1 & \mbox{in} \ B_g(s,i_0) \\
h_s & \equiv & 0 & \mbox{on} \ \partial B_g(s,i_0).
\end{array} \right.
\end{equation}
Here the Laplace operator is defined with respect to the metric $g$. The
function $h_s$ will be extended to be zero outside the geodesic ball.
Since the harmonic $C^{0,\alpha}$-norm of $(\Omega, g)$ is uniformly bounded,
it is easy to see that a uniform Poincare inequality holds on the balls
$B_g(s,i_0)$ with dependence on $i_0$. A simple integration argument then
yields a uniform estimate of the Sobolev norm of $h_s$. Uniform interior $C^{2,
\alpha}$ estimates then follow readily, because in harmonic coordinates the
Laplace operator takes the form $\Delta = g^{ij} \partial_i \partial_j$. We
also have a uniform $L^{\infty}$ estimate up to boundary, but it seems
impossible to obtain better estimate up to boundary because the control of the
geometry of the boundary is very weak. At a first glance this appears to
threaten to destroy the embedding scheme. Fortunately we have a way to get
around it. On the other hand, we can not obtain control of the dependence of
$h_s$ on the center
$s$. To remedy this, we shall take a suitable average of $h_s$ over $s$. The
resulting new family of functions will depend nicely on the center.
Now let us state a few basic properties of the functions $h_s$ in the following
proposition, which will be proved at the end of this section. Here, as before,
we work under the assumption $\|(\Omega, g)\|^h_{C^{0, \alpha}, r} \leq Q(r)$.
\begin{prop} \label{bh}
Let $\bar{h}_s(p)$ be the solution of equation (\ref{h}) with respect to the
canonical Euclidean metric $g_0$ on the Euclidean ball $B_{g_0}(s,i_0)$. Then
for any $\epsilon > 0$ and fixed $0<R<1$, there is an $r_0 = r_0
(\epsilon,R,Q) > 0$ such that if $i_0 \leq r_0$,
\begin{eqnarray}
|h_s(p) - \bar{h}_s(p)| & < & \epsilon i_0^2, \label{c1} \\
| \frac{\partial}{\partial p} h_s(p) - \frac{\partial}{\partial p}
\bar{h}_s(p)| & < &\epsilon i_0 \label{c2}
\end{eqnarray}
for all $s$ and all $p$ with $d_{g_0}(s,p) \leq Ri_0$. It will follow from the
proof that $B_{g_0}(s,Ri_0) \subset B_g(s, i_0)$ so that these estimate make
sense. Also
\begin{equation} \label{ub}
| \frac{\partial^2}{\partial p^2} h_s(p) |\leq C(n, Q,R), \ \ |\frac{1}{i_0}
\frac{\partial}{\partial p} h_s(p) |\leq C(n, Q,R).
\end{equation}
\end{prop}
Note that
\begin{equation}
\bar{h}_s(p) = \frac{1}{2n} (i_0^2 - d_{g_0}^2(s,p)).
\end{equation}
Therefore $\frac{2n}{i_0^2} \bar{h}_s(p) \leq \frac{1}{5}$ when $d_{g_0}(s,p)
\geq \sqrt{\frac{4}{5}}i_0$. Choosing $R = \frac{10}{11}$ in
Proposition~\ref{bh}, we have $\frac{2n}{i_0^2} h_s(p) < \frac{1}{4}$ when $
\sqrt{\frac{4}{5}}i_0 \leq d_{g_0}(s,p) \leq \frac{10}{11}i_0$ and $i_0$ is
sufficiently small.
Let $\beta = \beta_n \in C^\infty_0 ([0,\infty))$ be the cut off function.
\[
\beta_n(t) = \left\{ \begin{array}{ll} 0 & \mbox{if} \ 0 \leq t \leq
\frac{1}{4} \\
B_n & \mbox{if} \ t \geq \frac{1}{2}
\end{array} \right.,
\]
where
$B_n$ is a constant which will be determined later.
Then $\beta \left( \frac{2n}{i_0^2}h_s(p)\right) = 0$ near the sphere
$d_{g}(s,p) = \frac{9}{10}i_0$ for all $i_0$ small. (Note that $d_{g}$
converges to $d_{g_0}$ when $i_0 \rightarrow 0$.) We define a new function which is
$\beta \left( \frac{2n}{i_0^2}h_s(p)\right)$ restricted to the ball $B(s,
\frac{9}{10}i_0)$
and identically zero outside. For simplicity we still denote this new function
by $\beta \left( \frac{2n}{i_0^2}h_s(p)\right)$.
As mentioned before, we have no control of the dependence of $h_s$ on the
center $s$. The said average function is given as follows
\begin{equation}
f_p(q) = \int_\Omega \beta \left( \frac{2n}{i_0^2}h_s(p)\right) \beta \left(
\frac{2n}{i_0^2}h_s (q) \right)\, ds.
\end{equation}
Note that $f_p(q)$ is symmetric in $p$ and $q$ and is $C^{2,\alpha}$ uniformly
bounded in both variables.
Now we define the embedding
\begin{eqnarray*}
F: \ \Omega_1 & \rightarrow & L^2(\Omega, g)\\
p & \rightarrow & i_0^{-\frac{3}{2}n+1} f_p(q)
\end{eqnarray*}
Note that
\begin{eqnarray}
d_{v_p}F : & q \longmapsto & 2ni_0^{-\frac{3}{2}n}\int_\Omega \beta'\left(
\frac{2n}{i_0^2}h_s(p)\right) \langle \frac{1}{i_0}\nabla_{v_p} h_s(p), v_p \rangle
\beta \left( \frac{2n}{i_0^2}h_s (q) \right)\, ds, \label{dff}\\
\nabla^2_{v_p,w_p}F: & q \longmapsto & 4n^2i_0^{-\frac{3}{2}n-1}\int_\Omega
\beta''\left( \frac{2n}{i_0^2}h_s(p)\right) \langle \frac{1}{i_0}\nabla h_s(p),
v_p \rangle \langle \frac{1}{i_0}\nabla h_s(p), w_p \rangle \beta \left(
\frac{2n}{i_0^2}h_s (q) \right)\, ds \nonumber \\
& & + 2ni_0^{-\frac{3}{2}n-1}\int_\Omega \beta'\left( \frac{n}{i_0^2}\right)
\nabla^2_{v_p,w_p} h_s(p) (v_p,w_p) \beta \left( \frac{n}{i_0^2}h_s (q)
\right)\, ds. \label{dff2}
\end{eqnarray}
We first show that when $\Omega$ is an Euclidean domain, we can normalize
$\beta$ so that $F$ is an isometric imbedding. In this case the imbedding
function
\[
\bar{f}_p(q) = \int_\Omega \beta \left(1-\frac{d_{g_0}^2(p,s)}{i^2_0}\right)
\beta \left(1-\frac{d_{g_0}^2(q,s)}{i^2_0}\right)\, ds.\]
By the symmetry of the integration domain, $B(p, \frac{\sqrt{3}i_0}{2}) \cap
B(q, \frac{\sqrt{3}i_0}{2})$, and the integrand, $\bar{f}_p(q)$ depends only on
$d_{g_0}(p,q)$ (and $\beta$). We can write $\bar{f}_p(q) = i^n_0 \tilde{f} (n,
\frac{1}{i_0}d_{g_0}(p,q))$,
$ d_{v_p}\bar{F}(q) = i_0^{-n/2}\tilde{f}' (n, \frac{1}{i_0}d_{g_0}(p,q)) \langle
\nabla d_{g_0}(p,q), v_p \rangle$. Then
\begin{eqnarray*}
\| d_{v_p}\bar{F}\|^2_{L^2(\Omega)} & = & i_0^{-n}\int_{B(p, 2i_0)}
\tilde{f}'^2 (n, \frac{1}{i_0}d_{g_0}(p,q)) \langle \nabla d_{g_0}(p,q), v_p \rangle^2
dq \\
& = & i_0^{-n}\int_0^{2i_0} r^{n-1} \int_{S^{n-1}} \tilde{f}'^2 (n,
\frac{1}{i_0}r) \langle \xi, v \rangle^2 d\xi dr \\
& = & i_0^{-n}\frac{\mbox{vol} (S^{n-1})|v|^2}{n}\int_0^{2i_0} r^{n-1}\tilde{f}'^2
(n, \frac{1}{i_0}r) dr \\
& =& \frac{\mbox{vol} (S^{n-1})|v|^2}{n}\int_0^{2}r^{n-1}\tilde{f}'^2 (n,r) dr.
\end{eqnarray*}
Choose $B_n$ in the defintion of $\beta$ so that $\frac{\mbox{vol}
(S^{n-1})}{n}\int_0^{2}r^{n-1}\tilde{f}'^2 (n,r) dr = 1$. Then we will have
achieved the following.
\begin{lem}
$\bar{F}$ is an isometric embedding.
\end{lem}
With the above choice of $\beta$ we will show that $F$ is an almost isometric
embedding when $\Omega$ is not necessarily an Euclidean domain and the second
derivative of $F$ is also uniformly bounded. More precisely we have
\begin{prop} \label{e-df}
For any given $\epsilon_0 >0$, there exists an $r_0>0$ such that
\begin{equation} \label{df1}
(1+ \epsilon_0)^{-2} |v|^2 \leq \|d_{v_p}F\|^2_{L^2(\Omega)} \leq (1+
\epsilon_0)^{2} |v|^2,
\end{equation}
for all $v\in T_p\Omega_1$ and $0 < i_0 \leq r_0$. And
\begin{equation} \label{df2}
\|\nabla^2_{v_p,w_p}F\|^2 _{L^2(\Omega)} \leq C(n,\alpha,Q) i_0^{-2}|v|^2 \cdot
|w|^2.
\end{equation}
\end{prop}
\noindent {\em Proof.} By definition
\[
\|d_{v_p}F\|^2_{L^2(\Omega)} = \int_{B(p,2i_0)} |d_{v_p}F(q)|^2 dq.
\]
Now the volume element of the metric $g$ is comparable with Euclidean one.
Namely
\begin{equation} \label{vv}
e^{-Q(i_0)} \mbox{vol}_{R^n} \leq \mbox{vol}_g \leq e^{Q(i_0)} \mbox{vol}_{R^n}.
\end{equation}
Therefore it suffices to prove that $|d_{v_p}F(q)|$ is close to
$|d_{v_p}\bar{F}(q)|$ when $i_0$ is small, which follows from (\ref{dff}) and
Proposition~\ref{bh}.
(\ref{df2}) also follows from (\ref{dff2}), (\ref{vv}) and
Proposition~\ref{bh}.
\hspace*{\fill}\rule{3mm}{3mm}\quad \\
{\em Proof of} Proposition~\ref{bh}. First we introduce some new functions.
Let $\tilde{h}_{s, i_0}(p)$ be the solutions of (\ref{h}) on
$B_{i_0^{-2}g}(s,1)$ with respect to the scaled metrics $i_0^{-2}g$, and
$\bar{\tilde{h}}_s(p)$ the solution of (\ref{h}) with respect to the Euclidean
metric on the Euclidean ball $B(s,1)$. Then
\begin{equation} \label{hh}
\tilde{h}_{s,i_0}(p)= i_0^{-2} h_s(p), \ \ \bar{\tilde{h}}_s(p) = i_0^{-2}
\bar{h}_s(p).
\end{equation}
(Here the variables $s,p$ are in domain with different metrics for different
functions.) Since
\begin{equation} \label{normb}
\|(B_{i_0^{-2}g}(s,1), i_0^{-2}g)\|_{C^{0,\alpha},1} = \|(B_g(s,i_0),
g)\|_{C^{0,\alpha},i_0} \leq Q(i_0) \leq Q(1), \ \mbox{for}\ i_0 \leq 1,
\end{equation}
for the same reasons as mentioned before for $h_s$, we have the following
estimates
\begin{equation} \label{el}
\| \tilde{h}_{s,i_0}\|_{L_0^{1,2}(B_g(s,1))} \leq C(n,Q(1))
\end{equation}
and
for any fixed $0<R<1$,
\begin{equation} \label{ell}
\| \tilde{h}_{s,i_0} \|_{C^{2,\alpha}(B_g(s,R))} \leq C(n, Q(1),R).
\end{equation}
On the other hand, the hypothesis $\|(B_{i_0^{-2}g}(s,1),
i_0^{-2}g)\|_{C^{0,\alpha},1} \leq Q(i_0)$ implies that
\[
e^{-Q(i_0)} d_{i_0^{-2}g_0}(p,s) \leq d_{i_0^{-2}g}(p,s) \leq e^{Q(i_0)}
d_{i_0^{-2}g_0}(p,s).
\]
Therefore
\[
B_{i_0^{-2}g_0} (s, e^{-Q(i_0)}R) \subset B_{i_0^{-2}g} (s, R) \subset
B_{i_0^{-2}g_0}(s, e^{Q(i_0)}R).
\]
{}From these estimates and the uniqueness of the weak solution $\bar{\tilde
h}_s$ it is easy to deduce the following: for each sequence of centers $s_k$
converging to some center $s_0$ and each sequence $i_0(k)$ converging to zero,
the corresponding rescaled solutions $\tilde h_{s_k, i_0(k)}$ converge weakly
to $\bar{\tilde h}_{s_0}$. Moreover, by the Arzela-Ascoli theorem, they also
converge uniformly in $C^1$ on proper compact subsets of $B_{g_0}(s_0, 1)$.
This convergence fact along with the smooth dependence of $\bar{\tilde h }_s$
on $s$ then imply that the $\tilde h_{s, i_0}$ converge uniformly with respect
to $s$ in $C^1$ on proper compact subsets of $B_{g_0}(s, 1)$ as $i_0$ goes to
zero. Consequently, for a fixed $R \in (0,1)$, given any $\epsilon >0$,
there is an $r_0 >0$ such that for all $s$ and $p$ with $d_{g_0}(p,s) < R$, if
$i_0 \leq r_0$, then
\begin{equation}
|\tilde{h}_s(p) -\bar{\tilde{h}}_s(p)| < \epsilon. \label{pc1}
\end{equation}
Similarly,
\begin{equation}
| \frac{\partial}{\partial p} \tilde{h}_s(p) -\frac{\partial}{\partial p}
\bar{\tilde{h}}_s(p)| < \epsilon. \label{pc2}
\end{equation}
Hence for all $s$ and all $p$ with $d_{g_0}(s,p) \leq Ri_0$,
\begin{eqnarray*}
|h_s(p) - \bar{h}_s(p)| & < & \epsilon i_0^2, \\
| \frac{\partial}{\partial p} h_s(p) - \frac{\partial}{\partial p}
\bar{h}_s(p)| & < &\epsilon i_0.
\end{eqnarray*}
(\ref{ub}) just follows from (\ref{ell}) and (\ref{hh}).
\hspace*{\fill}\rule{3mm}{3mm}\quad
\sect{Embedding II}
In this section we study the geometry of $F(\Omega_1)$ as a submanifold in
$L^2(\Omega)$. We will prove, among other things, two important properties of
$F(\Omega_1)$. That is, the induced metric of $F(\Omega_1)$ has uniformly
bounded sectional curvature and the embedding $F$ is equivariant.
The geometry of $F(\Omega_1)$ is completely determined by the second
fundamental form of its embedding into $L^2(\Omega)$, which in turn can be
described by the family of orthogonal projections.
$P(y):\ L^2(\Omega) \rightarrow T_yF(\Omega_1) \subset L^2(\Omega), y \in F(\Omega^1)$.
We have
\begin{lem} \label{4.1}
The sectional curvature of $F(\Omega_1)$ is given by the following formula:
\begin{equation} \label{curv}
R(z_1, z_2) z_3 = [d_{z_1} P, d_{z_2}P] z_3, \ \ z_1,z_2,z_3 \in T_z
F(\Omega_1).
\end{equation}
\end{lem}
\noindent {\em Proof.} Since $P^2 =P$, one has
\begin{equation} \label{p2}
(d_{z_1}P)P + P (d_{z_1}P) = d_{z_1}P.
\end{equation}
Let $\nabla$ be the connection on $F(\Omega_1)$ and $d_{z_1}$ the directional
derivative on the $L^2$ space. Then
\begin{eqnarray*}
\nabla_{z_1} z_2 & = & P(d_{z_1}z_2) \\
& = & d_{z_1}z_2 - (1- P)(d_{z_1}(Pz_2)) \\
& = & d_{z_1}z_2 - (1-P) \left[ (d_{z_1}P)z_2 + P (d_{z_1}z_2) \right] \\
& = & d_{z_1}z_2 - (1-P) (d_{z_1}P) (Pz_2) \\
& = & d_{z_1}z_2 - (d_{z_1}P)z_2.
\end{eqnarray*}
Here we have used (\ref{p2}) in the last equation.
Therefore
\begin{equation} \label{p1}
\nabla_{z_1} = d_{z_1} - (d_{z_1}P).
\end{equation}
Now formula \ref{curv} follows from (\ref{p1}) and the definition of the
curvature tensor.
\hspace*{\fill}\rule{3mm}{3mm}\quad
\begin{prop} \label{4.2}
Let $\alpha_0 = (1+\epsilon_0)^2 C(n,\alpha,Q)$. Here $\epsilon_0, r_0, C$ are
the same constants as in Proposition~\ref{e-df}. Then for all $0 < i_0 < r_0$,
\begin{equation} \label{ep}
\| d_{\dot{y}}P\|_{op} \leq \alpha_0 i_0^{-1} \| \dot{y} \|, \end{equation}
\end{prop}
\noindent {\em Proof.} Since $\left( 1-P(F(p)) \right) d_{w_p} F = 0$,
\[
d_{d_{v_p}F} P \cdot d_{w_p} F = \left( 1-P(F(p)) \right) \nabla^2_{v_p,w_p}F.
\]
By (\ref{df1}) and (\ref{df2}), $\| d_{\dot{y}}P\|_{op} \leq \alpha_0 i_0^{-1}
\| \dot{y} \|$.
\hspace*{\fill}\rule{3mm}{3mm}\quad \\
Therefore the metric $\tilde{g} = F^* g_{L^2}$, the metric on $\Omega_1$
obtained by pulling back the $L^2$ metric, has bounded sectional curvatures.
To prove the equivariance, we first note:
\begin{lem} \label{uh}
Let $h_s(p)$ be the function defined in (\ref{h}), and let $\psi: \Omega \rightarrow
\Omega'$ be an isometric embedding. Then
\begin{equation}
h_{\psi (s)} (\psi (p)) = h_s(p).
\end{equation}
\end{lem}
\noindent {\em Proof.} Since equation (\ref{h}) is invariant under isometry, this follows from the
uniqueness of solutions to (\ref{h}).
\hspace*{\fill}\rule{3mm}{3mm}\quad \\
Let $(\Omega,g)$ be as before and $F: \Omega_1 \rightarrow L^2(\Omega)$ the embedding
defined in \S 3. With the above lemma, we can now prove
\begin{prop} \label{4.4}
If $\psi : (\Omega,g) \rightarrow (\Omega',g')$ is an isometric embedding, then $$\psi
: (\Omega, \tilde{g}) \rightarrow (\Omega', \tilde{g}')$$ is also an isometric
embedding.
\end{prop}
\noindent {\em Proof.} First, we assume $\psi$ is actually an isometry. Then
\[
F \circ \psi (p) = f_{\psi (p)},
\]
where the function
\begin{eqnarray*}
f_{\psi (p)} (q) & = & \int_\Omega \beta \left( \frac{2n}{i_0^2}h_s(\psi
(p))\right) \beta \left( \frac{2n}{i_0^2}h_s (q) \right)\, ds.
\\
& = & \int_\Omega \beta \left( \frac{2n}{i_0^2}h_{\psi^{-1}(s)}(p)\right)
\beta \left( \frac{2n}{i_0^2}h_{\psi^{-1}(s)} (\psi^{-1}(q)) \right)\, ds.
\end{eqnarray*}
Here we have used Lemma~\ref{uh}. Since $\psi$ is an isometry, a change of
coordinates yields
\[
f_{\psi (p)}(q) = f_p (\psi^{-1} (q)).
\]
It follows then that
\[
F \circ \psi = (\psi^{-1})^* \circ F,
\]
where we have denoted by $(\psi^{-1})^*$ the map on $L^2(\Omega)$ induced by
$\psi^{-1}$. Therefore
\[
\psi^* F^*g_{L^2} = (F \circ \psi)^* g_{L^2} = F^* ((\psi^{-1})^*)^* g_{L^2} =
F^*g_{L^2}.
\]
This proves the equivariance when $\psi$ is an isometry. Since $\Omega,
\Omega'$ are both domains of $R^n$, the general statement follows by applying
the above to $\psi:\ \Omega \rightarrow \psi (\Omega)$.
\hspace*{\fill}\rule{3mm}{3mm}\quad \\
{\it Proof of} Theorem 1.1. This theorem is a consequence of Proposition 2.2,
Lemma 4.1, Proposition 4.2 and Proposition 3.3.
\hspace*{\fill}\rule{3mm}{3mm}\quad
\sect{Proof of Theorem 1.2}
We consider $(M,g) \in {\cal M}(n,p,Q)$ and $\Omega = B(0, r) \subset R^n$
with the pull-back metric $\varphi_s^*g$, where $\varphi_s$ is a coordinate
map. For convenience, we shall again denote the pull-back metric by $g$. We
have the inequality $\|(\Omega, g)\|_{L^{1,p}, r} \leq Q(r)$.
We employ the same embedding of $(\Omega, g)$ as before. Thus we use the same
functions $h_s \in L_0^{1,2}(\Omega)$, as given by (3.1). But the estimates for
$h_s$ are different now. By the $L^p$ elliptic theory we have uniform $L^{2,p}$
estimates for $h_s$ in the interior. A result similar to Proposition 3.1 then
holds, namely we have the estimates (3.2), (3.3) and the second one in (3.4),
while the first one in (3.4) is replaced by an estimate on the $L^p$-norm of
the second order derivative. Now it is clear that our smoothing process
produces a metric $\bar g$ on $M$ such that
the lifted metrics $\varphi_s^* \bar g$ on $B(0, r/2)$ have uniformly bounded
Sobolev constant and uniformly $L^p$-bounded sectional curvatures.
Next we apply the Ricci flow to deform $\bar g$. For this purpose, we assume
that $M$ is closed. We appeal to the arguments in \cite{dwy}. There, manifolds
with a pointwise bound on Ricci curvature and a conjugate radius bound are
treated.
These conditions are used to show that controlled harmonic coordinates exist
on lifted local patches, where the lifting is given by the exponential map.
In these coordinates, the Ricci curvature bound then implies an $L^p$-bound on
sectional curvatures. In our situation, we do have an $L^p$-bound
on sectional curvatures. But the Ricci curvature bound is also used in several
other places in \cite{dwy}. Since this bound is not available here, we need to
modify the arguments in \cite{dwy}.
In the key Proposition 3.1 (uniform short time
existence of the Ricci flow with a priori control) in \cite{dwy}, we drop the
estimate (3.4) on Ricci curvature. (Note that the conclusion of the proposition
without (3.4) still suffices for our purpose.)
We claim that the proposition then holds in our new situation. For
convenience, we shall call this proposition the "Key Proposition". In
\cite{dwy}, the proof of Key Proposition is based on four lemmata: Lemmata 3.2,
3.3, 3.4 and 3.5.
Now let's take a look at these lemmata in our new situation. Lemma 3.2
(pointwise estimate for Riemann curvature tensor) holds without change. The
proof of Lemma 3.3 ($L^p$-estimate for Riemann curvature tensor) depends on a
{\it covering estimate}, i.e. an estimate for certain covering number and
multiplicity regarding geodesic balls in the lifted patches. In \cite{dwy}, the
estimate comes from the Bishop-Gromov covering argument, which depends on a
pointwise lower bound for Ricci curvature.
Now we do not have such a bound. But we still have a covering estimate,
which follows from the properties of our coordinates and the basic control
over the pull-back metrics as given by (the modified versions of) Propositions
3.1 and
3.2 (in the present paper).
Another ingredient in the proof of Lemma 3.3 is
an isometry correspondance between geodesic balls on different lifted patches.
Now it is given by Lemma~\ref{hom}. (The proof of this correspondence given in
\cite{dwy} does not work here.) Thus Lemma 3.3 also holds. Since we have
dropped the estimate (3.4) about Ricci curvature in Key Proposition, Lemma 3.4
is no longer needed. Finally, note that the proof for Lemma 3.5 (estimate of
the Sobolev constant) in \cite{dwy} goes by computing the change rate of the
Sobolev constant along the flow. In \cite{dwy}, this rate is controlled by
a uniform bound on Ricci curvature, which is not valid here.
However, Lemma 3.2 contains an estimate for Ricci curvature at positive time
$t$, namely it is
dominated by a constant times $t^{-1/2}$. Since the function $t^{-1/2}$ is
integrable at $0$, it is clear that the change rate of the Sobolev constant
is still under control without a uniform Ricci curvature bound, and hence Lemma
3.5 carries over. (An alternative way of handling the Sobolev constant is to
apply Yang's estimate for it in \cite{y2}, which uses only an $L^p$-bound on
Ricci curvature and a positive lower bound on (local) volume. But that is more
involved. )
We leave to the reader to formulate precisely the independent result implied
by the above proof about short time existence of the Ricci flow.
|
\section{Introduction}
Studies of the dynamics of a scalar field together with other forms of
matter usually assume that this behaves as a perfect fluid. Few authors have
taken into account dissipative effects \cite{Sin88} \cite{Sin90} \cite{Roy92}
\cite{Roy93}, even though processes like particle creation might have been
important in the early universe.
In this paper we study the evolution of a universe filled with
a massless minimally coupled scalar field and a viscous fluid. We find exact
solutions of the Einstein equations in a Robertson-Walker metric and we
analyse their asymptotic stability by means of the Lyapunov method
\cite{Kra}.
\section{The model}
We investigate the evolution of a universe filed with a scalar field
and a viscous fluid. The scalar field $\phi $ is free and minimally
coupled to gravity, so that it obeys the equation $\Box \phi =0$.
In the case of the homogeneous isotropic Robertson-Walker
metric
\begin{equation}\label{1}
ds^2=dt^2-a^2(t)\left [{\frac{dr^2}{1-kr^2}}+r^2(d\theta ^2+\sin
{}^2\theta d\phi ^2)\right]
\end{equation}
\noindent
the scalar field equation becomes
\begin{equation}
\label{100}
\ddot \phi +3H\dot \phi =0
\end{equation}
Besides, only the bulk viscosity needs to be considered. Thus we replace in
the Einstein equations the equilibrium pressure $p$ by an effective pressure
\cite{Wei71}
\begin{equation}\label{2}
3H^2=\frac {1}{2}\dot\phi^2+\rho -3{\frac k{a^2}}+\Lambda
\end{equation}
\begin{equation} \label{3}
2\dot H+ 3H^{2} = -\frac{1}{2}\dot\phi^2 - p -\sigma-\frac{k}{a^2}+\Lambda
\end{equation}
\noindent where $H=\dot a/a$, $^{\cdot }=d/dt$, $\rho $ is the energy
density, $\sigma $ is the viscous pressure, $\Lambda$ is the cosmological
constant and we use units $c=8\pi G=1$.
As equation of state we take
\begin{equation}
\label{4}
p = ( \gamma -1 ) \rho
\end{equation}
\noindent with a constant adiabatic index $0\le\gamma \le 2$, and we assume
that $\sigma $ has the constitutive equation
\begin{equation}\label{45}
\sigma = - 3 \zeta H
\end{equation}
\noindent where $\zeta \ge 0$ is the bulk viscosity coefficient.
Thus, we must solve equation (\ref{100}) together with the Einstein equations
(\ref{2})(\ref{3}). Equation (\ref{100}) has the first integral
\begin{equation} \label{5}
\dot\phi=\frac{C}{a^3}\qquad
\end{equation}
\noindent
where $C$ is an arbitrary integration constant, and we are interested in the
case that $C\neq0$.
Then, adding (\ref{2}) and (\ref{3}), and eliminating $\rho$, we arrive at
\begin{equation} \label{6}
2\dot H+3\gamma H^2-3\zeta H=-\frac{2-\gamma}{2}\frac{C^2}{a^6}+
\left(2-3\gamma\right)\frac{k}{a^2}+\gamma\Lambda
\end{equation}
To find exact solutions of this equation we make the change of variable
$a=v^\nu$, and we consider two cases: $\zeta$ constant and $\zeta \sim
H$.
We assume now that $\zeta$ is a constant. This is a good approximation when
the thermodynamical variables do not change too much. This has been studied in
several papers \cite{Tre} \cite{Hel73} \cite{Hel74} \cite{Sus}
\cite{Roy83}. Choosing $\nu=2/(3\gamma)$, the equation (\ref{6}) becomes
\begin{equation} \label{7}
\ddot v+\frac{3}{2} \zeta\dot v+M v^m+N v^n-\frac{\gamma\Lambda}{2\nu}v=0
\end{equation}
\noindent
where $M=(3/8)\gamma\left(2-\gamma\right)C^2$, $m=1-4/\gamma$,
$N=(3/4)\gamma(3\gamma-2)k$, $n=1-4/(3\gamma)$.
It is linear for $\gamma=2$ and $k=0$, so that we can find its general
solution. In this case, the evolution of the scale factor is the same as in a
model without the scalar field and a bulk viscosity coefficient $\zeta/2$
\cite{Tre} \cite{Hel73}. Thus, we just quote the solutions without further
analysis.
\bigskip
\noindent
Let us consider first that $\Lambda=0$. We find
\bigskip
\begin{equation} \label{8}
a(t)=\left[A \exp\left(\frac{3}{2}\zeta t\right)+B\right]^{1/3}
\end{equation}
\begin{equation} \label{9}
\Delta\phi(t)=\frac{2C}{3\zeta B}\ln\left[\frac{\exp\left(\frac{3}{2}\zeta
t\right)} {A\exp\left(\frac{3}{2}\zeta t\right)+B}\right]
\end{equation}
\begin{equation} \label{10}
\rho(t)=\frac{\frac{3}{2}\zeta^2 A^2\exp(3\zeta t)-C^2}
{2\left[A\exp\left(\frac{3}{2}\zeta t\right)+B\right]^2}
\end{equation}
\noindent
where $A$ and $B$ are arbitrary integration constants. The requirement
$\rho\ge0$ implies that there is a minimum time for the validity of this
solution. Thus it describes a geodesically incomplete manifold, which in this
sense is singular \cite{Ger}.
Clearly, in its asymptotically de Sitter regime it is physically
justified the approximation of constant $\zeta$. However, for shorter times,
this assumption may not be justified.
The asymptotically de Sitter stage occurs for any other value of $\gamma$ and
$k$. In effect, assuming $a\sim\exp(H_0 t)$ for $t\to\infty$, with $H_0$ a
constant, we find that $H_0=\zeta/\gamma$. To verify the asymptotic stability
of this solution we use $a$ as the independent variable and we take the
Liapunov function $L=(H-H_0)^2$. Then, we find that
\begin{equation} \label{11}
L'=-3\gamma\frac{L}{a}+O\left(\frac{1}{a^3}\right)
\end{equation}
\noindent
is negative for large times.
Let us consider now that $\Lambda\neq0$.
There is a critical value for the cosmological constant
$\Lambda_0=-3\zeta^2/16$. Thus, we must distinguish several cases:
\noindent
Two kinds of solutions appear for $\Lambda>\Lambda_0$. Singular evolutions
\begin{equation} \label{12}
a(t)=a_0 e^{3\Delta t/4}\left[\sinh\left(\omega \Delta t
\right)\right]^{1/3}
\end{equation}
\noindent
or nonsingular evolutions
\begin{equation} \label{13}
a(t)=a_0 e^{3\Delta t/4}\left[\cosh\left(\omega \Delta t
\right)\right]^{1/3}
\end{equation}
\noindent
where $\omega=\left[3\left(\Lambda-\Lambda_0\right)\right]^{1/2}$, $\Delta
t=t-t_0$ and $a_0$, $t_0$ are integration constants. For the
scalar field we find
\begin{equation} \label{14}
\Delta
\phi(t)=-\frac{C}{B\mu}e^{(\omega-3\zeta/4)\Delta t}
{}_2F_1\left(1,\frac{1}{2}-\frac{3\zeta}{8\omega},\frac{3}{2}
-\frac{3\zeta}{8\omega},-\frac{A}{B}e^{2\omega \Delta t}\right)
\end{equation}
While $\rho\to (3/8)\zeta^2+(3/2)\zeta\omega>0$ for $t\to\infty$, in the
nonsingular solutions $\rho$ becomes negative before a minimum time. On the
other hand, for singular solutions, we find
\begin{equation} \label{15}
\rho\sim\left(\frac{1}{3}-\frac{C^2}{2a_0^3\omega^2}\right)\frac{1}{\Delta t^2}
\qquad \Delta t\to 0
\end{equation}
\noindent
so that the energy density may remain positive along all the evolution.
\noindent
In the case $\Lambda=\Lambda_0$, the evolution is always singular
\begin{equation} \label{16}
a(t)=a_0 (\Delta t)^{1/3} e^{3\Delta t/4}
\end{equation}
\noindent
In the case $\Lambda\le\Lambda_0$, the scale factor recollapses to a second
singularity
\begin{equation} \label{17}
a(t)=a_0 e^{3\Delta t/4}\left[\sin\left(|\omega| \Delta t\right)\right]^{1/3}
\end{equation}
\noindent
All singular evolutions have particle horizons because $a\sim \Delta t^{1/3}$
as $\Delta t\to 0$.
For other values of $\gamma$ and $k$, there is also a critical value of the
cosmological constant $\Lambda_0=-3\zeta^2/(4\gamma^2)$, so that for
$\Lambda>\Lambda_0$ an asymptotically de Sitter evolution occurs for two
values of $H$
\begin{equation} \label{18}
H_\pm=\frac{1}{2\gamma}\left[\zeta\pm\left(\zeta^2+\frac{4}{3}\gamma^2\Lambda
\right)^{1/2}\right]
\end{equation}
\noindent
We verify that the behavior $a\sim\exp(H_+t)$ is asymptotically stable by
means of the Liapunov function $L_+=(H-H_+)^2$, which satisfies $L_+'<0$ for
large times when $H>0$.
Nonlinear viscous effects has been shown to arise as a phenomenological
description of the effect of particle creation in the early universe
\cite{Ver}, and cosmological models with this kind of fluids has been
considered in \cite{Nov} and \cite{Rom}.
Here, we assume that $\zeta=\alpha H$, with a constant $\alpha$ such that
$\gamma>\alpha>0$, and we consider only expanding evolutions.
Following the same procedure as before, we arrive at the equation
\begin{equation} \label{20}
\ddot v+M v^m+N v^n-\frac{\gamma\Lambda}{2\nu}v=0
\end{equation}
\noindent
where now $\nu=2/[3(\gamma-\alpha)]$, $M=(3/8)(\gamma-\alpha)(2-\gamma)$,
$m=1-4/(\gamma-\alpha)$, $N=(3/4)(\gamma-\alpha)(3\gamma-2)k$ and
$n=1-4/[3(\gamma-\alpha)]$. As this case behaves as a conservative mechanical
system, we may obtain (at least formally) its general solution:
\begin{equation} \label{21}
\Delta t=\int\frac{dv}{\sqrt{2\left[E-V(v)\right]}}
\end{equation}
\noindent
where $E$ is an integration constant and
\begin{equation} \label{22}
V(v)=\frac{M}{m+1}v^{m+1}+\frac{N}{n+1}v^{n+1}-\frac{\gamma\Lambda}{4\nu}v^2
\qquad \gamma-\alpha\neq\frac{2}{3}
\end{equation}
\begin{equation} \label{23}
V(v)=\frac{M}{m+1}v^{m+1}+N\ln v-\frac{\gamma\Lambda}{4\nu}v^2
\qquad \gamma-\alpha=\frac{2}{3}
\end{equation}
A qualitative analysis of (\ref{21}) shows that expanding evolutions may begin
either at a singularity or a bounce. Both bounded and unbounded singular
solutions occur, but only for $\Lambda<0$ are there bouncing solutions that
reach a maximum.
We quote two cases for which the equation (\ref{20}) becomes linear, so that we
may obtain its general solution in closed form (for simplicity we take
$\Lambda=0$ ). The first one is $\gamma=2$ and $k=0$
\begin{equation} \label{24}
a(t)=\left(A\Delta t\right)^\nu
\end{equation}
\begin{equation} \label{25}
\Delta\phi(t)=\frac{C\Delta t^{1-3\nu}}{A^{3\nu}(1-3\nu)}
\end{equation}
\begin{equation} \label{26}
\rho(t)=\frac{3\nu^2}{\Delta t^2}-\frac{C^2}{2(A\Delta t)^{6\nu}}
\end{equation}
\noindent
The other solution is for $\gamma=2$, $\alpha=2/3$ and $k\neq0$. We find
\begin{equation} \label{27}
a(t)=\left|A\Delta t-2k\Delta t^2\right|^{1/2}
\end{equation}
\begin{equation} \label{28}
\Delta\phi(t)=-\frac{2C}{A^2}\frac{A-4k\Delta t}
{\left|A\Delta t-2k\Delta t^2\right|^{1/2}}
\end{equation}
\begin{equation} \label{29}
\rho(t)=\frac{3}{4}\frac{(A-4k\Delta t)^2}
{\left|A\Delta t-2k\Delta t^2\right|^2}-\frac{1}{2}\frac{C^2}
{\left|A\Delta t-2k\Delta t^2\right|^3}+\frac{3k}
{\left|A\Delta t-2k\Delta t^2\right|}
\end{equation}
We find also solutions of the form $a=a_0 \Delta t^\sigma$ for special values
of $\alpha$: a. $\gamma=2$, $\sigma=1$, $\alpha=(4/3)(1+k/a_0^2)$; b. $k=0$,
$\sigma=1/3$, $\alpha=\gamma+3(2-\alpha)C^2/(2a_0^2)-2$.
\section{Conclusions}
We have found exact solutions of the Einstein equations with a free scalar
field and a viscous fluid source, in a homogeneous isotropic metric. We have
considered cosmological models with two forms of the bulk viscosity pressure:
linear, with a constant coefficient, and cuadratic in the Hubble variable. For
the first case the picture is similar to the case without scalar field and we
show that the de Sitter evolution is asymptotically stable. Also, we find that
the physical requirement of positivity of the energy density of the fluid
makes some of these solutions geodesically incomplete.
For the case of a nonlinear viscous pressure we reduce the equations of this
model to that of a conservative mechanical system. Thus, we are able to give
its general solution in implicit form, and besides we show several cases for
which an explicit solution arise. Both, singular and bouncing solutions arise,
and we find cases in which the scale factor reach a maximum or grow without
bound. In the latter case, when $\Lambda=0$, the scale factor has a power
law behavior for large times.
Dissipative effects like particle production may have been important in the
early universe, for instance in the reheating period at the end of the
inflationary era. It is generally assumed that this stage of accelerated
evolution was driven by a self-interacting scalar field. Thus, we consider of
interest to investigate further the models of this paper, taking into account
also an interaction potential.
\newpage
|
\subsection*{To appear in Physical Review B.}
\subsection*{cond-mat/9508140}
\hspace*{0.5cm}
\section{ Introduction }
In our recent paper \cite{anyon} we have pointed out the possibility
that vortices in superconducting films might be anyons. There is a
classic paper by Haldane and Wu \cite{hawu}, which demonstrates that
vortices in superfluid helium layers are not anyons because there
is no well defined dilute limit.
We have reanalysed this question from the point of view of the
phenomenological time-dependent Ginzburg-Landau models. It was shown
that there are situations when one can go around the argument in \cite{hawu}.
A classic example are vortices in the Ginzburg-Landau models of the
fractional quantum Hall effect \cite{wl,glhe}. Chern-Simons interaction makes
them well localised objects and the dilute limit can be defined.
The main effect of the CS gauge field is to remove the divergent
gradient energies present in the global model. These divegencies are
also removed in the Ginzburg-Landau model for superconductors.
We have considered a gauged nonlinear Schrodinger equation as a minimal
version of time-dependent Ginzburg-Landau model. To model the structure
of the vortex core we assumed it was filled with normal fluid
so as to make the whole structure locally charge-neutral. The strenght of the
statistical interaction was proportional to the net deficit of the superfluid
replaced inside the core by the normal phase.
It is argued \cite{baps} that the time evolution of the condensate
is much faster then that of the normal fluid. Thus in the static
case the core is filled with the normal fluid but when the vortex
moves fast enough the normal fluid can not adjust itself and does
not follow the vortex motion. It is a crude approximation \cite{stoof}
and it does not invalidate our results. The statistical interaction
shows up in the adiabatic approximation which is quite an
oposite limit. In this limit we can assume the nonuniformity of the
normal fluid does follow the vortex motion. The effective Lagrangian
for vortex motion can be arranged term by term according to the powers
of vortex velocity. The statistical interaction together with the term
responsible for the Magnus force \cite{at} are the first terms in this
effective Lagrangian. The distortion of the normal fluid distribution from
the static one, which is at least of the first order in velocities, can
contribute but only to the term quadratic in velocities which is
the next to leading term. Thus at least for slow motion as compared
to characteristic velocities, when expansion in powers of velocities is
justified, the Magnus force and statistical interaction are not altered.
One can rewrite the local BCS model in
terms of the gap-function field \cite{abts,stoof,aatz} but the effective
theory appears to be nonlocal - it contains derivatives of arbitrary order.
In this way one goes from description in terms of electronic
degrees of freedom to description in terms of Cooper pairs.
The latter is well suited for the bulk of the superconductor.
However in the core of the vortex we can expect some decoherence,
which phenomenologically might be decribed by normal fluid.
Whenever degrees of freedom other than those of Cooper pairs
come into play the description in terms of Cooper pairs may
happen to be irrelevant. One way of dealing with the problem
is to truncate the effective theory on the lowest order in derivatives
and introduce more or less explicitly something like a normal
component. This approach is limited by the poor knowledge
about the nature of the normal fluid. Another approach is to take
as many orders in derivatives in the effective theory as possible to describe
also the normal fluid in terms of the gap function. The problem is that
in practice one would have to cut the expansion at certain order
and there is no warranty that such a cut theory would be self-consistent.
To go around these problems we will derive the statistical
interaction in the framework of the Bogolubov-de Gennes formalism
for pure samples at zero temperature. We will work in the quasi
two-dimensional regime of long parallel vortices. In the case of
superconducting layers thicker than $100\AA$ the penetration lenght
$\Lambda$ is still very close to the penetration lenght $\lambda$ in the
bulk superconductor. In this regime we expect modifications to be rather
quantitative in nature then qualitative.
An important first step was done by Gaitan \cite{gaitan} in his derivation of
the Berry phase responsible for the Magnus force. In this paper we are going
to reanalize his derivation and then to extend the method to the case of two
well separated vortices. In the microscopic theory the language
of the normal and superfluid is not very fruitful. The distinction of the
states below the Fermi surface into localised bound states and scattering
states appears to be more natural. The distinction does not
influence the value of the Magnus force but it is crucial
for the statistical interaction. The scattering states are common to
all vortices while bound states can be identified with particular ones.
\section{ Preliminaries on vortex solution in BCS theory }
The problem of the vortex solution in the BCS theory can be
conveniently posed within the Bogolubov-de Gennes formalism \cite{dg}. It
has not been completely solved although some qualitative fictures of the
solution are known \cite{bardeen}. We will restrict here to listing the basic
ingredients of the formalism.
The Bogolubov equation, defined in the Nambu spinor space \cite{nambu},
is
\begin{equation}\label{2.10}
(E_{n}-\hat{H}_{BOG})
\left(\begin{array}{c}
u_{n} \\
v_{n}
\end{array}\right)=0 \;\;,
\end{equation}
where $\hat{H}_{BOG}$ is the Bogolubov hamiltonian
\begin{equation}\label{2.20}
\hat{H}_{BOG}=
\left[\begin{array}{cc}
-\frac{1}{2}(\mbox{\boldmath $\nabla$}-ie\mbox{\boldmath $A$})^{2}-E_{F} &
\Delta(\mbox{\boldmath $x$}) \\
\Delta^{\star}(\mbox{\boldmath $x$}) &
\frac{1}{2}(\mbox{\boldmath $\nabla$}+ie\mbox{\boldmath $A$})^{2}+E_{F}
\end{array} \right] \;\;.
\end{equation}
$E_{F}$ is the Fermi energy.
$\Delta(\mbox{\boldmath $x$})=\Delta_{0}(r)\exp(-i\theta)$ denotes
the gap function of vortex solution. $\Delta_{0}(r)$ interpolates
between $0$ at the origin and a constant, which we will call
$\sqrt{\rho_{0}}$, at infinity. There are both positive and negative energy
solutions. If $(u_{n},v_{n})$ is an eigenstate with energy $E_{n}>0$,
then $(-v^{\star}_{n},u^{\star}_{n})$ is a solution with energy
$-E_{n}<0$. The equations (\ref{2.10}) have to be supplemented
by a self-consistency condition
\begin{equation}\label{2.25}
\Delta(\mbox{\boldmath $x$})=
g\sum_{n}u_{n}(\mbox{\boldmath $x$})v_{n}^{\star}(\mbox{\boldmath $x$}) \;\;,
\end{equation}
where $g$ is the BCS coupling constant, together with Maxwell equations
determining the vector potential.
The field operator for Nambu quasiparticles can be expanded in terms
of the solutions of Eq.(\ref{2.10}).
\begin{equation}\label{2.30}
\Psi( \mbox{\boldmath $x$} )=\left(\begin{array}{c}
\psi_{\uparrow}(\mbox{\boldmath $x$}) \\
\psi_{\downarrow}^{\dagger}(\mbox{\boldmath $x$})
\end{array}\right)=
\sum_{n}
\left[\begin{array}{c}
\gamma_{n\uparrow}
\left(\begin{array}{c} u_{n}(\mbox{\boldmath $x$}) \\
v_{n}(\mbox{\boldmath $x$}) \end{array}\right) +
\gamma_{n\downarrow}^{\dagger}
\left(\begin{array}{c} -v^{\star}_{n}(\mbox{\boldmath $x$}) \\
u^{\star}_{n}(\mbox{\boldmath $x$})
\end{array}\right)
\end{array}\right] \;\;.
\end{equation}
The negative energy states are occupied in the BCS ground state
\begin{equation}\label{2.40}
\mid BCS > = \prod_{n} \gamma_{n\downarrow} \mid 0 > \;\;.
\end{equation}
The eigenstates satisfy the following orthogonality
\begin{eqnarray}\label{2.50}
&&\int d^{3}x\;
[u_{n}(\mbox{\boldmath $x$})u^{\star}_{m}(\mbox{\boldmath $x$})
+v_{n}(\mbox{\boldmath $x$})v^{\star}_{m}(\mbox{\boldmath $x$})]=
\delta_{nm}
\;\;,\nonumber\\ &&\int d^{3}x\;
[u_{n}(\mbox{\boldmath $x$})v_{m}(\mbox{\boldmath $x$})
-v_{n}(\mbox{\boldmath $x$})u_{m}(\mbox{\boldmath $x$})]=0
\end{eqnarray}
and completeness relations
\begin{eqnarray}\label{2.60}
&&\sum_{n}[u_{n}(\mbox{\boldmath $x$})
u_{n}^{\star}(\mbox{\boldmath $x$}^{\prime})
+v_{n}^{\star}(\mbox{\boldmath $x$})
v_{n}(\mbox{\boldmath $x$}^{\prime})]
=\delta(\mbox{\boldmath $x$}-\mbox{\boldmath $x$}^{\prime})
\;\;,\nonumber\\ &&\sum_{n}[u_{n}(\mbox{\boldmath $x$})
v_{n}^{\star}(\mbox{\boldmath $x$}^{\prime})
-v_{n}^{\star}(\mbox{\boldmath $x$})
v_{n}(\mbox{\boldmath $x$}^{\prime})]=0 \;\;.
\end{eqnarray}
With the help of these relations the creation and annihilation operators
can be expressed as
\begin{eqnarray}\label{2.70}
&&\gamma_{n\downarrow}=\int d^{3}x\;
[-\psi^{\dagger}_{\uparrow}(\mbox{\boldmath $x$})
v_{n}^{\star}(\mbox{\boldmath $x$})
+\psi_{\downarrow}(\mbox{\boldmath $x$})
u^{\star}_{n}(\mbox{\boldmath $x$})] \;\;,
\nonumber\\
&&\gamma_{n\uparrow}^{\dagger}=\int d^{3}x\;
[\psi^{\dagger}_{\uparrow}(\mbox{\boldmath $x$})
u_{n}^{\star}(\mbox{\boldmath $x$})
+\psi_{\downarrow}(\mbox{\boldmath $x$})
v_{n}(\mbox{\boldmath $x$})] \;\;.
\end{eqnarray}
Adiabatic vortex motion \cite{gaitan} gives rise to a Berry phase
in the solutions of Eq.(\ref{2.10}),
$(u_{n},v_{n})\rightarrow \exp[i\phi_{n}](u_{n},v_{n})$.
These Berry phases sum up to the total Berry phase picked up
by the ground state $\mid BCS >\rightarrow \exp[i\Gamma] \mid BCS >$
which with the help of Eqs.(\ref{2.70},\ref{2.40}) can be established
to be
\begin{equation}\label{2.80}
\Gamma=-\sum_{n}\phi_{n} \;\;.
\end{equation}
To persue some of the questions we need more detailed knowledge
about the eigenstates. The axially symmetric ansatz takes the form
\begin{equation}\label{2.90}
\chi_{n}(\mbox{\boldmath $x$})=\left ( \begin{array}{c}
u_{n}(\mbox{\boldmath $x$}) \\
v_{n}(\mbox{\boldmath $x$})
\end{array} \right )=
e^{ik_{z}z} e^{i(\mu-\frac{1}{2}\sigma_{z})\theta} f_{n}(r) \;\;,
\end{equation}
where $\hbar k_{z}$ is the $z$-component of the momentum. $\mu$ must
be a half-integer for the expression to be single-valued. The functions
$f_{n}(r)$ have been investigated in \cite{bardeen} with the help of WKB
approximation. We will quote some more detailed results in the following.
\section{ Origin of the Magnus force }
Now we are going to rederive the Berry phase responsible for
the Magnus force following the argument of Gaitan \cite{gaitan}. In
comparison with \cite{gaitan} we clarify some points and remove some
unnecessary assumptions.
The general form of the Berry phase in two dimensions is
\begin{equation}\label{3.10}
\phi_{n}=i\int dt\; \int d^{2}x\;
\chi_{n}^{\dagger} (\frac{d}{dt}+i\frac{e}{\hbar}A_{0}) \chi_{n} \;\;.
\end{equation}
The time derivative is understood as a total derivative with respect
to slow degrees of freedom. For an adiabatic motion of a
single vortex the derivative has to be replaced by
$\dot{\mbox{\boldmath $r$}}_{0}\mbox{\boldmath $\nabla$}
_{\mbox{\boldmath $r_{0}$}}$, where $\mbox{\boldmath $ r_{0} $}$
is a position of the vortex singularity. The scalar potential
vanishes for the vortex solution so we will skip the second term in
what follows. With the axially symmetric ansatz (\ref{2.90}) the phase
becomes
\begin{equation}\label{3.20}
\phi_{n}=\int dt\;\int d^{2}x\; \dot{ \mbox{\boldmath $r$} }_{0} [
(\chi^{\dagger}_{n}(-\mu+\frac{1}{2}\sigma_{z})\chi_{n})
\mbox{\boldmath $ \nabla $}_{\mbox{\boldmath $r_{0}$}} \theta+
f^{\dagger}_{n}(r) \mbox{\boldmath $ \nabla $}_{\mbox{\boldmath $r_{0}$}}
f_{n}(r)]
\end{equation}
The second term vanishes by symmetry arguments. The contribution of the first
term to the total Berry phase is
\begin{equation}\label{3.30}
\Gamma=-\sum_{n}\phi_{n}=-\int dt\;\int d^{2}x\;
( \dot{ \mbox{\boldmath $r$} }_{0} \mbox{\boldmath $ \nabla $}
_{ \mbox{\boldmath $ r_{0} $} }\theta) S(r)
\end{equation}
where
\begin{equation}\label{3.40}
S(r)=\sum_{n}[\mid u_{n}(r)\mid^{2}(-\mu+\frac{1}{2})+
\mid v_{n}(r)\mid^{2}(-\mu-\frac{1}{2})] \;\;.
\end{equation}
$\hbar S$ is minus the z-component of the canonical angular momentum density.
It is not a gauge-invariant integral of motion. $S$ is the expectation
value density of the operator $-i\hbar\frac{\partial}{\partial\theta}$
instead of
the gauge-invariant
$-i\hbar\frac{\partial}{\partial\theta}+\sigma_{z}A_{\theta}$.
$S(0)=0$ because either $u_{n}$ ($v_{n}$)
or the factor $(-\mu\stackrel{+}{-}\frac{1}{2})$ vanishes at the origin.
To find out its asymptotic behavior at infinity we would need a much more
detailed knowledge about the solutions. We go around this problem by
resorting to the effective theory which is
equivalent to the microscopic formalism. The general term linear in
the covariant time derivative reads
\begin{equation}\label{3.50}
\int dt\;\int d^{2}x\;
i[\Delta^{\star}(\hbar\partial_{t}+2ieA_{0})\Delta-c.c.]
[G(\Delta^{\star}\Delta)+spatial\;derivative\;terms]\;\;.
\end{equation}
By "$spatial\;derivative\;terms$" we mean terms which are of at least
first order in the covariant spatial derivatives. $G$ is a function
of $\Delta^{\star}\Delta$ only which tends to $\rho_{s}/\rho_{0}$ as the gap
function $\Delta$ approaches its asymptotic equilibrium value. $\rho_{s}$
is the equilibrium Cooper pairs' density.
Let us consider the adiabatic rotation of the vortex solution
$\Delta=\Delta_{0}(r)e^{-i\theta}$ around its axis,
$\theta\rightarrow\theta-\omega t$. The action picks up a term
(to lowest order in $\omega$)
\begin{equation}\label{3.60}
\omega \int dt\;\int d^{2}x\; \{-\hbar\Delta^{\star}\Delta
[G(\Delta^{\star}\Delta)+spatial\;derivative\;terms]\} \;\;.
\end{equation}
The spatial integral is just the total angular momentum.
For large $r$, where $A_{0}$ tends to zero, the density of this angular
momentum is, by gauge invariance (\ref{3.50}), equal to $\hbar$ times
minus the bulk Cooper pairs' density
$\hbar S\approx-\hbar\rho_{s}=\hbar\lim_{r\rightarrow\infty}
\frac{\delta W}{\delta (2eA_{0})}$, where $W$ is the effective action
and $\rho_{s}>0$. We have to stress that we make use of the effective
theory only very far from the vortex core where it should be equivalent
to the microscopic treatement. In particular in the distant asymptotic
region there is no contribution from unpaired bound states which
can not be described in terms of Cooper pairs.
We have all we need to calculate the Berry phase. Let us expand
the integrand in Eq.(\ref{3.30}) around the vortex position
$\mbox{\boldmath $r_{0}$}=(X,Y)$ close to the origin, $(X,Y)=(0,0)$,
\begin{eqnarray}\label{3.70}
&&S=S(r)-S^{\prime}(r)[X\cos\theta+Y\sin\theta]
+O(r_{0}^{2}) \;\;,\nonumber\\
&&\dot{ \mbox{\boldmath $r$} }_{0}\mbox{\boldmath $\nabla$}
_{\mbox{\boldmath $r_{0}$}}\theta=
\dot{X}(\frac{\sin\theta}{r}+\frac{X\sin 2\theta-Y\cos 2\theta}{r^{2}})+
\dot{Y}(-\frac{\cos\theta}{r}-\frac{X\cos 2\theta+Y\sin 2\theta}{r^{2}})+
O(r_{0}^{2}) \;\;.
\end{eqnarray}
A straightforward integration yields
\begin{equation}\label{3.80}
\Gamma=-\pi [S(\infty)-S(0)] \int dt\;\varepsilon_{kl}X^{k}\dot{X}^{l}
+O(r_{0}^{2}) \;\;.
\end{equation}
The expression $O(r_{0}^{2})$ does vanish. We are considering single vortex
in absence of any driven current. Such a system is translationally
invariant and isotropic. The first term on the R.H.S. of Eq.(\ref{3.80})
is already the most general term linear in velocity which is, up to a total
time derivative, translationally invariant and isotropic. Thus we do not need
to consider finite $\mbox{\boldmath $r_{0}$}$ to obtain a generally valid
expression. The phase (\ref{3.80}) is remarkably simple to evaluate. For
a vortex with winding number $-1$ it reads
\begin{equation}\label{3.90}
\Gamma=\pi\rho_{s}\int dt\; \varepsilon_{kl}X^{k}\dot{X}^{l} \;\;.
\end{equation}
From our derivation of the Magnus force it is clear that
the Wess-Zumino term (in the gauge $A_{0}=0$)
\begin{equation}\label{3.100}
\int dt\;\int d^{2}x [\rho_{s}\partial_{t}\theta] \;\;,
\end{equation}
with $\rho_{s}=const$, does not make much sense as it stands.
$\rho_{s}$ is the same at the origin as at infinity so the Magnus force
vanishes. The formula (\ref{3.100})
is to be understood with an implicit assumption that a small area around
the phase singularity is excluded from the spatial integration. In
other words $\rho_{s}$ must be put equal to $0$ in this area.
It is not difficult to realise, by performing radial integration first
and then integration over the angle around the singularity, that the
way of regularisation does not matter. In particular it does not need
to be rotationally symmetric. The only factors that determine the Magnus
force are the two limit values of $\rho_{s}$. Thus vortices in a
condensate will always feel the Magnus force. It is not the case
for say Jackiw-Pi solitons \cite{jp}, where $\rho_{s}$ is zero both at
the origin and at infinity.
\section{ Mutual statistical interaction of vortices }
Let us consider two vortices: "1" at the origin and "2"
very far apart at $\mbox{\boldmath $R$}(t)$. It is important to realise
that the eigenstates of the Bogolubov hamiltonian can be divided
into common scattering states, which we will still denote by just
$u_{n},v_{n}$, and bound states which can be identified with a given vortex
$u^{(1,2)}_{n},v^{(1,2)}_{n}$. Vortices are very distant so there is no
overlap between their localised bound states.
The bound states of the stationary vortex "1" feel what
is going on around them through the pair potential
$\Delta(t,\mbox{\boldmath $x$})$ inside and around the core.
Vortices are well localised so a fairly good
approximation to a two-vortex gap function is the product ansatz
\begin{equation}\label{4.10}
\sqrt{\rho_{0}}\Delta(t,\mbox{\boldmath $x$})
=\Delta_{v}( \mbox{\boldmath $x$} )
\Delta_{v}[ \mbox{\boldmath $x-R$}(t) ] \;\;,
\end{equation}
where $\Delta_{v}(\mbox{\boldmath $x$})=\Delta_{0}(r)\exp(-i\theta)$ denotes
the gap function of a single vortex centered at the origin. Close to
$\mbox{\boldmath $x$}=0$ this expression can be further simplified
\begin{equation}\label{4.20}
\Delta(t,\mbox{\boldmath $x$})=\Delta_{v}(\mbox{\boldmath $x$})
e^{ -i\theta[\mbox{\boldmath $x-R$}(t)] } \;\;,
\end{equation}
Thus the bound states of the static vortex have to be modified as
\begin{equation}\label{4.30}
\chi_{n}^{(1)}[\mbox{\boldmath $x,R$}(t)] =
e^{-i\frac{\sigma_{z}}{2}\theta[\mbox{\boldmath $x-R$}(t)]}
\chi_{n}(\mbox{\boldmath $x$}) \;\;.
\end{equation}
Their contribution to the Berry phase is
\begin{eqnarray}\label{4.40}
-\int dt\;\int d^{2}x\; \{ \dot{ \mbox{\boldmath $R$} }
\mbox{\boldmath $\nabla$}
_{\mbox{\boldmath $R$}}\theta[\mbox{\boldmath $x-R(t)$}] \}
\sum_{bound\;st.}(\frac{1}{2}\mid u_{n}\mid^{2}-
\frac{1}{2}\mid v_{n}\mid^{2})\approx \nonumber\\
\{\frac{1}{2}\int d^{2}x\;
\sum_{bound\;st.}(\mid u_{n}\mid^{2}-\mid v_{n}\mid^{2})\}
\int dt\; \dot{ \mbox{\boldmath $R$} }\mbox{\boldmath $\nabla$}
_{\mbox{\boldmath $R$}}\theta[\mbox{\boldmath $R(t)$}]\;\;,
\end{eqnarray}
where the approximate equality is valid for small
$\frac{r_{c}}{R}$, where $r_{c}$ is a radius of the core.
The equality $\mid u_{n}\mid^{2}=\mid v_{n}\mid^{2}$ holds for the bound
states, at least up to the WKB approximation \cite{bardeen},
so their contribution to the Berry phase vanishes.
Now as the vortex "2" moves its bound states follow its trajectory
$\chi^{(2)}[\mbox{\boldmath $r-R$}(t)]$, similarly as in the single vortex
case
considered in the previous section. In addition, as an effect due to the
vortex "1" (\ref{4.10}), their components perform the relative phase rotation
\begin{equation}\label{4.50}
\chi_{n}^{(2)}[\mbox{\boldmath $x,R$}(t)] \approx
e^{-i\frac{\sigma_{z}}{2}\theta[\mbox{\boldmath $R$}(t)]}
\chi_{n}[\mbox{\boldmath $x-R(t)$}] \;\;.
\end{equation}
The contribution from this relative phase rotation is once again zero.
The bound states contribute but only to the Magnus term (\ref{3.90}) just
as in the single vortex case.
The bound states do not give rise to any new effects so let us consider
scattering states common to both vortices. In the vortex core region
the asymptotes of the scattering states must be close to those
of the scattering states for a single vortex but the phase has to be
replaced by the asymptote of the phase in the product ansatz (\ref{4.20}).
Close to the origin
\begin{equation}\label{4.60}
\chi_{n}[\mbox{\boldmath $x,R$}(t)]\approx e^{ik_{z}z}
e^{i(\mu-\frac{\sigma_{z}}{2})
\{\theta(\mbox{\boldmath $x$})+\theta[\mbox{\boldmath $x-R$}(t)]\}}
f_{n}(r) \;\;.
\end{equation}
The contribution from around the stationary vortex is
\begin{equation}\label{4.70}
-\int dt\;\int d^{2}x\;
\dot{\mbox{\boldmath $R$}}\mbox{\boldmath $\nabla$}_{\mbox{\boldmath $R$}}
\theta[\mbox{\boldmath $x-R$}(t)] \bar{S}(r) \;\;,
\end{equation}
where $\bar{S}$ is a part of the canonical angular momentum due to
the scattering states
\begin{equation}\label{4.80}
\bar{S}(r)=\sum_{scatt.\;st.}[\mid u_{n}(r)\mid^{2}(-\mu+\frac{1}{2})+
\mid v_{n}(r)\mid^{2}(-\mu-\frac{1}{2})] \;\;.
\end{equation}
Far from the core $\bar{S}\approx S\approx -\rho_{s}$.
If $\bar{S}$ were equal to $-\rho_{s}$ also in the core, the contribution to
the Berry phase from (\ref{4.70}) would be just the same as to the
Magnus term. Thus we are interested only in the effects due
to deviations of $\bar{S}$ from its asymptotic value $-\rho_{s}$.
Inside the core $\mid u_{n}\mid^{2}$ and $\mid v_{n}\mid^{2}$
are changed by a factor which is $>1$ for the states with $\mu$ negative
and $<1$ for $\mu$ positive \cite{bardeen}. The net deviation
$\delta\bar{S}(r)=\bar{S}(r)+\rho_{s}$ is positive.
At the very origin $\delta\bar{S}(0)=\rho_{s}$. The total change in the Berry
phase is twice that in (\ref{4.70}), as there are two vortices, and
amounts to
\begin{equation}\label{4.90}
\delta\Gamma=[2\int d^{2}x\;\delta\bar{S}(r)]
\int dt\; \dot{\mbox{\boldmath $R$}}\mbox{\boldmath $\nabla$}
_{\mbox{\boldmath $R$}}\theta(\mbox{\boldmath $R$})\;\;.
\end{equation}
Thus the total Berry phase for a dilute vortex system is
\begin{equation}\label{4.100}
\Gamma=\int dt\;
[ -\pi\rho_{s}\sum_{p} n_{p}\varepsilon^{kl}X^{k}_{(p)}\dot{X}^{l}_{(p)}+
\alpha\sum_{p<q}n_{p}n_{q}\frac{d}{dt}\Theta_{(p,q)}] \;\;,
\end{equation}
where the indices $p,q$ run over vortices, $n$'s are their winding numbers,
$\Theta_{(p,q)}$ is the angle between the $p$-th and $q$-th vortex
and the numerical factor $\alpha$ can be read from Eq.(\ref{4.90})
\begin{equation}\label{4.110}
\alpha=2 \int d^{2}x\; \delta\bar{S}(r) \;>\; 0 \;\;.
\end{equation}
$\alpha$ is roughly the number of electrons inside the core
and as such it can range from $\sim 1$ for high $T_{c}$ superconductors
to $\sim 10^{5}$ for some conventional type $II$ superconductors.
\section{ Vortex statistics within variational wave-function approach }
Once we have derived statistical interaction in the microscopic
setting it may be worthwhile to reanalise some earlier approaches
to similar problems. In the paper by Ao and Thouless \cite{at} the Magnus
force was derived with the help of the variational many-electron
vortex wave-function
\begin{equation}\label{at.10}
\psi_{v}[z]=\exp[\frac{i}{2}\sum_{k}\theta(z_{k}-z_{0})]\psi_{0}[z] \;\;,
\end{equation}
where $z_{0}$ is a complex vortex position, $z_{k}$'s are positions
of electrons and $\psi_{0}[z]$ is an antisymmetric variational function.
The phase factors in the wave-function are determined by the demand
of correct electronic quantumstatistics and by topological properties.
There are variational profile functions in $\psi_{0}$ which can not be
established without dynamical considerations. One can consider an adiabatic
vortex motion along some trajectory and calculate the Berry phase
picked up by the wave-function. This Berry phase coinsides with
Eq.(\ref{3.90}). For a closed path the Berry phase is proportional
to the number of electrons enclosed by the trajectory.
This setting is convenient to analise what is the dependence of
the Magnus force on impurities \cite{at}. An impurity can be viewed as an
attractive potential which traps some of electrons in localised
bound states. The trapped electrons disappear from the ansatz (\ref{at.10}).
The Berry phase is still proportional to the area enclosed by the trajectory
but this time the area should not be multiplied by the total density
of electrons but rather by the total density minus the density of electrons
trapped by impurities. Impurities lower the value of the Magnus force.
Now let us consider the effect of an exchange of two vortices.
More precisely, let us fix the position of one vortex and consider
another distant vortex moving around it. One could
argue there is no special effect because the net charge of any vortex
must be zero. Provided the trajectory is large enough, there is no
change in the number of enclosed electrons due to the enclosed vortex.
The last sentence is certainly true but the example with impurities
tought us that it is not the total number of electrons that really
matters but rather the number of electrons in the coherent state
described by the wave-function (\ref{at.10}). We know from the discussion
in the previous sections that inside vortex core the scattering
or continuum states are replaced by bound states.
Thus vortex can be viewed as a kind of impurity, which traps some
of electrons into localised bound states with energies within the energy gap
band. The localised electrons are removed from the wave-function
(\ref{at.10}). There is an additional Berry phase proportional
to the number of vortices enclosed by the trajectory. Each enclosed
vortex contributes a term proportional to the number of electrons
trapped inside its core.
We can consider a path for a chosen vortex in a more or less uniform
distribution of vortices. If we neglect possible intervortex correlation
effects, the background vortices could be regarded as uniform distribution
of impurities lowering the density of electrons in the coherent state
(\ref{at.10}). In this mean-field approximation the Magnus force
acting on a choosen vortex is lowered by the presence of another vortices.
This approximation is nothing else but the delocalisation procedure
so often applied to anyonic systems. The Magnus force can be interpreted
as Lorenz force due to interaction of effectively charged vortices with some
uniform effective magnetic field. The statistical interaction can be seen as
Aharonov-Bohm effect due to the fluxes attached to vortices. In the
mean-field approximation the fluxes, which are opposite to the external
flux, are delocalised and they lower the net uniform flux. In the same
way the real impurities can be interpreted as localised fluxes, opposite
to the external field, randomly distributed over the plane. If the M-F
approximation appears to work for real impurities, it will also work
for vortices.
\section{ Hall angle and vortex density }
The fact that the value of the Magnus force can be lowered with
increasing density of vortices can, in principle, be observable
in Hall experiments \cite{hl}. The vortex density should increase
and the M-F Magnus force should decrease with increasing real magnetic field.
This should manifest itself in the changes of the measured Hall
angle. Vortex equation of motion takes the form \cite{ah}
\begin{equation}\label{at.20}
m_{eff}\ddot{\mbox{\boldmath $r$}}=
\frac{\rho_{s}hd}{2}
( \dot{\mbox{\boldmath $r$}} - \mbox{\boldmath $v_{s}$} )\times
\hat{\mbox{\boldmath $z$}}
-\eta d \dot{\mbox{\boldmath $r$}} + \mbox{\boldmath $F_{pin}$}
+ \mbox{\boldmath $f$} \;\;,
\end{equation}
where $m_{eff}$ is a small effective vortex mass, $\eta$ is a vortex
viscosity, $\mbox{\boldmath $F_{pin}$}$ is a pinning force,
$\mbox{\boldmath $f$}$ is a fluctuating force, $d$ is a sample thickness
and $\mbox{\boldmath $v_{s}$}$ is a driven uniform superfluid velocity.
When we neglect pinning and average over fluctuations the stationary state
motion will be determined by the equation
\begin{equation}\label{at.30}
( \dot{\mbox{\boldmath $r$}} - \mbox{\boldmath $v_{s}$} )\times
\hat{\mbox{\boldmath $z$}}
=\tan(\theta_{H})\dot{\mbox{\boldmath $r$}} \;\;,
\end{equation}
where $\tan(\theta_{H})=\frac{2\eta}{\rho_{s}h}$. The solution is
\begin{equation}\label{at.40}
\dot{\mbox{\boldmath $r$}}=
\frac{ \mbox{\boldmath $v_{s}$}
+ (\hat{\mbox{\boldmath $z$}}\times\mbox{\boldmath $v_{s}$})
\tan(\theta_{H}) }{1+\tan^{2}(\theta_{H}) } \;\;.
\end{equation}
$\theta_{H}$ is the angle between the superfluid velocity
$\mbox{\boldmath $v_{s}$}$ and the stationary vortex velocity.
The angle is the larger the weaker is the Magnus force.
If the effective Magnus force is lowered with increased vortex density
the angle should also grow with external magnetic field, which
drives the rise in vortex density. The changes of the Magnus force
due to changes in vortex density should be the more rapid the larger is
the number of electrons trapped in the vortex core.
For this reason we would recommend experiments on mildly type $II$
conventional superconductors with large vortex cores
(large correlation length $\xi$). Rather strong viscosity
should be prefered for
the angle to be more sensitive to the strengh of the Magnus force.
The sample should be pure of pinning centers to avoid obscure pinning
effects.
\section{ Vortices' fractional quantum Hall effect }
To summarise our knowledge about the dynamics of planar vortices,
let us write down an effective Lagrangian for diluted vortices with
topological charge $-1$
\begin{equation}\label{5.10}
L_{eff}=\sum_{p} [\frac{1}{2}m_{eff}\dot{\mbox{\boldmath $X$}}_{p}
\dot{\mbox{\boldmath $X$}}_{p}
+\hbar\pi\rho_{s} \mbox{\boldmath $X_{p}$}
\times \dot{ \mbox{\boldmath $X$} }_{p} ]
+\sum_{p<q}[\hbar\alpha\frac{d}{dt}\Theta_{(p,q)}
-V_{eff}(\mid\mbox{\boldmath $X_{p}-X_{q}$}\mid)] \;\;.
\end{equation}
The indices $p,q$ run over vortices. $m_{eff}$ is the effective
vortex mass. It is usually estimated to be around
$10^{8}m_{e}/m=10^{-2}m_{e}/\AA$.
The second term in Eq.(\ref{5.10}) decribes interaction of vortices
with effective magnetic field. We stress that this field has nothing to do
with the real
magnetic field $B_{ext}$, which in this case is just a device
to drive the changes of vortex density. If we assumed the density
of electrons to be $\sim 10^{30} m^{-3}=1\AA^{-3}$ the effective magnetic
field defined by $\frac{eB_{eff}}{2}=\pi\hbar\rho_{s}$ would turn out to
be $\sim 10^{6} T/\AA$. When compared with the effective vortex mass
per $1\AA$ the magnetic field turns out to be incredibly strong.
Its effect on a vortex should be the same as that of the $10^{8}T$ magnetic
field on an electron. Vortices can be expected to be confined to the
lowest Landau level (LLL).
Now let us consider a single vortex at $z=X_{1}+iX_{2}$. What is the
magnetic lenght $l$ which determines the size of the LLL wavepacket
\begin{equation}\label{5.20}
\psi_{0}(z,\bar{z})=
\frac{1}{\sqrt{2\pi l^{2}}}e^{-\frac{z\bar{z}}{4l^{2}} }\;\;?
\end{equation}
The magnetic lenght is determined by the strength of the exactly known
Magnus force $l^{2}=\frac{1}{2\pi\rho_{s}}$. The area over which
the center of such a vortex fluctuates can be estimated to be
$\frac{2}{\rho_{s}}$, which is of the same order as the area per
one electron. This effect is significant for extremely type $II$
superconductors where the core is very thin.
Now we are prepared to address the question of the fractional Hall effect.
Vortices are anyons with a statistical parameter $\alpha$. A nontrivial
statistics would be sufficient to prevent them from overlapping if
the third term in (\ref{5.10}) were not regularised at short distances
and replaced by mutual charge-flux interaction \cite{kld}. Fortunately we
also have an effective short range mutual repulsion $V_{eff}$ which at low
temperatures may be sufficient to keep vortices at a distance.
On the other hand the potential is very weak as compared to the Landau
energy so the mixing with higher LL's should be in any case negligible.
Following Laughlin \cite{laughlin}, the trial wave-function for a many
vortex state can be written
\begin{equation}\label{5.30}
\psi_{m}[z]=\prod_{p<q}(z_{p}-z_{q})^{\alpha+2m}
\prod_{r} exp(-\mid z_{r} \mid^{2}/4l^{2}) \;\;,
\end{equation}
where $z_{p}$'s are positions of vortices. $m$ is a nonnegative integer
so that the exponent $(\alpha+2m)$ provides a correct quantumstatistics.
The density $n$ and the filling factor $\nu$ in such a state are
\begin{equation}\label{5.40}
\nu_{m}=2\pi l^{2} n_{m}=\frac{1}{\alpha+2m} \;\;.
\end{equation}
For $m=0$ the density is just $n_{0}=1/2\pi\alpha l^{2}=\rho_{s}/\alpha$.
$\alpha$ is roughly equal to the number of electrons
inside the core so in the $m=0$ state vortex cores
would have to overlap slightly. Certainly this density is close
to that of the Abrikosov lattice \cite{abr}. For larger $m$ the density
is smaller and finally we should get outside of the crystalline
regime. Then as the density of vortices is driven to change by the changes
of the real magnetic field $B_{ext}$ we should observe some plateaux
at the densities $n_{m}$ and maybe also at some other quantised
filling factors. For conventional superconductors, which are mildly type $II$,
$\alpha$ is large, so the difference between $n_{m}$ and $n_{m+1}$
is too small for the plateaux to be distinguishable
experimentally. For extremaly type $II$ superconductors or high $T_{c}$
superconductors we can expect $\alpha$ to be even $\sim 1$.
In such a case the first plateaux can be expected already
at $\sim 1/3$ and $\sim 1/5$ of the Abrikosov lattice density.
The main observation of this section is that if the Magnus force
is translated into the language of interaction with some effective uniform
magnetic field,
the field appears to be even $\sim 10^{6}T/\AA$. It has nothing
to do with the real magnetic field which is at best $\sim 10 T$.
The FQHE is a result of this huge effective
magnetic field and of repulsive intervortex potential but its existence
does not depend on the vortex statistics as the FQHE is possible even for
bosons \cite{jain}. However the pattern of the FQHE plateaux can help to
identify the quantumstatistics.
An experimental setup to detect the FQHE would consist of a planar sample
of some pure superconducting material in external uniform perpendicular
magnetic field. One would have to measure the total
flux $\Phi$, which penetrates through the sample. This flux gives the
actual number of vortices in the sample because each vortex carries one
flux quantum.
Just after continuously turning on weak magnetic field
the field lines would be pushed out of the sample. Only above
some threshold value of the flux the energy of the field could be
minimised by creating a vortex line. The story would repeat until
some stable FQHE plateau were reached. The plateau is a manifestation of
a very stable many-vortex state so an addition of one vortex to this state
would cost more energy then to a state far from the plateau. When the stable
state is approached from low densities the addition of one more vortex
should be much easier than usually. Quite opposite, when we lower
the external magnetic field it is easier to remove one flux quantum
from the sample just above the stable state and more difficult to remove
it just below the state, see the figure. The moduli of the change in the
driving flux $\Delta\Phi$ (or some equivalent quantity) necessary to change
the flux through the sample by one quantum will develop a characteristic
pattern around the stable density $n_{0}$. Two measurements, one with
adiabatically increasing and one with adiabatically
decreasing external magnetic field,
would give two curves with opposite polarisation. Taking their difference
will amplify the effects due to the plateau and remove the not neccesarily
constant bias.
It should be stressed here that our results are not in contradiction
with predictions of a bosonic Hall effect in Josephson junction arrays
\cite{jja}. The arrays are strictly planar devices. The penetration
lenght $\Lambda$ is likely to be greater then the sample size.
What is more, it seems to be possible to excite a vortex without any bound
states in the core, which indeed might be a boson. Our results are
exact in the limit of long straight-linear parallel vortices. In practice
it is sufficient that the sample is thick enough for the penetration
lenght $\Lambda$ to be close to that in the bulk. The sample thickness
would have to be a bit more then $d=100\AA$. For such a $d$ there is
still no space for vortex entanglement. In the absence of entanglement
vortices can be uniquely projected on a plane. In nonzero temperature
one should expect transversal modes to be excited. These excitations
do not affect the Berry phase picked up during vortices exchange as it
depends on the overall topological properties of vortices. Vortices
excited to different transversal states although in principle distinguishable
interact statistically. It is an example of mutual statistical interaction
\cite{mut} introduced first to describe interlayer phase correlations
in the double-layer Hall effect. The plateaux pattern for our anyonic
vortices is given by (\ref{5.40}). In distinction to bosonic Hall
effect the constant $\alpha$ is nonzero.
\section{ Summary }
We have given two new arguments why vortices in superconducting
films should be anyons. In addition we have discussed two
different experiments where this theoretical prediction could
be verified.
\paragraph{Aknowledgement.}
I would like to thank Prof. B.Halperin for stimulating comments on
Ref.\cite{anyon}. This research was supported in part by the KBN grant
No. 2 P03B 085 08 and in part by Foundation for Polish Science fellowship.
|
\section{Introduction}
Our best candidates, to date, of fundamental descriptions of nature,
now all seem to be connected by the duality conjectures
\cite{duff,schwarz,hull,witten}. The emerging picture is one of an
underlying, even more fundamental, theory in which the other theories
emerge as one approaches various limits in the moduli space.
Much effort has gone into understanding the $N=4$ \cite{schwarz,hull},
$N=2$ \cite{sei,kach,klemm,louis} and more recently the $N=1$
\cite{seib,vaf,har} cases. In particular the $K3$ manifold plays a central role
\cite{hull,witten,ferr,strom,klemm,louis,vaf,paul,town,sen} in many
scenarios.
Compactification of $d=11$ supergravity on a
compact manifold of $G_{2}$ holonomy
gives an effective four dimensional $N=1$ theory. Joyce has given many
examples of such manifolds \cite{J1,J2}. It is natural to conjecture
that these manifolds play a crucial role in the $N=1$
truncations of the conjectured dualities
between the Type IIA superstring on $K3$, the heterotic string on $T^{4}$
and $d=11$ supergravity on $K3{\times}S^{1}$.
In \cite{har} a manifold of $G_{2}$ holonomy was constructed by considering
freely acting involutions of a $K3{\times}T^{3}$ orbifold and resolving the
singularities. These involutions translate directly to actions on
the heterotic side giving an example of $N=1$ duality.
However, freely acting involutions of this type are very limited, and
the case when the involutions act with fixed points on the seven
coordinates on the supergravity side is less well understood. In fact,
all of the manifolds of \cite{J1,J2} are constructed by resolving
fixed point singularities, and if these manifolds are to play
a role in $N=1$ duality then this situation needs to be understood.
This work aims to shed some light on this problem. We work at a
generic point in moduli space throughout the following.
We focus on the simplest example given in \cite{J1},
which we denote by $J$, constructed with
three non-freely acting involutions of the seven torus, the singularities
of which are then resolved. In section two a brief outline of Joyce's
construction is given. Then in section three we present the heterotic dual
compactification. In particular, we are forced
to consider a new kind of 'overlapping orbifold' on
the heterotic side of the duality map. We end with some conclusions.
\section{Joyce's Construction}
In \cite{J1,J2} Joyce gave explicit constructions of 7-manifolds of
$G_2$ holonomy. When eleven-dimensional supergravity is compactified
on these manifolds, the resulting theory is four dimensional N=1
supergravity with $b_2$ vector multiplets and $b_3$ chiral multiplets,
where $b_2$ and $b_3$ are the non-trivial Betti numbers of the
7-manifold. The examples of \cite{J1} are all constructed by
orbifolding the seven-torus by various discrete isometries and then
resolving the singularities by replacing them with non-compact
Eguchi-Hanson geometries, a process that is now
more than familiar to
string theorists.
The simplest example given in \cite{J1},
which we denote by $J$ is constructed as follows:
Define the seven-torus coordinates as $(x_1,......,x_7)$.
Three $Z_2$ isometries of $T^7$ are defined by:
\begin{equation}
\alpha(x_1,....x_7) = (-x_1,-x_2,-x_3,-x_4,x_5,x_6,x_7)
\end{equation}
\begin{equation}
\beta(x_1,....x_7) = (-x_1,1/2-x_2,x_3,x_4,-x_5,-x_6,x_7)
\end{equation}
\begin{equation}
\gamma(x_1,....x_7) = (1/2-x_1,x_2,1/2-x_3,x_4,-x_5,x_6,-x_7)
\end{equation}
Obviously each of these $Z_{2}$'s has 16 fixed singular $T^{3}$'s
and each one defines an orbifold limit of a particular $K3{\times}
T^{3}$.
There are thus 48 fixed singular $T^{3}$'s of the surface. However
one must ask how the other two isometries act on the singular set
of each $Z_{2}$. In particular,
the $Z_{2}{\times}Z_{2}$'s generated by
$(\beta,\gamma)$, $(\alpha,\gamma)
$ and
$(\alpha,\beta)$ act freely on the singular sets of $\alpha$,
$\beta$ and $\gamma$ respectively. This implies that each isometry
has only four singularites of the original sixteen, giving a total
of 12 singular $ T^{3}$'s. Resolving each of these is crucial for the
consistency of the supergravity theory. This is done in the usual
way for $K3$ singularities - by inserting Eguchi-Hanson geometries, \cite{eh}.
The Betti numbers of the singular $T^{7}$ are $b_{2} = 0$ and $b_{3} = 7$.
The resolution of each singularity adds 1 to $b_{2}$ and 3 to $b_{3}$.
The Betti numbers of this Joyce manifold are thus $b_{2} = 12$
and $b_{3} = 43$.
\section{The Heterotic Dual and Generalised \newline Orbifold}
Eleven-dimensional Supergravity on $K3{\times}T^{3}$ is conjectured
to be dual to the heterotic string on $T^{6}$. Orbifolding the
$K3{\times}T^{3}$ by symmetries such as the Enriques involution
on the $K3$ part translates to a particular orbifold action
on the heterotic side. This has been illustrated successfully in
several examples \cite{ferr,sen,vaf,har}.
In particular, in \cite{har}
an example was considered in which $T^{7}$ was modded out
by a particular $Z_{2}{\times}Z_{2}{\times}Z_{2}$, one of
which defined a particular orbifold limit of $K3{\times}T^{3}$.
This is what has been called $\alpha$ in the construction of $J$.
The remaining $Z_{2}$'s acted freely, and we note that they are
essentially 'freely acting versions' of what we denoted by $\beta$
and $\gamma$ in constructing $J$. This is so because the example of \cite{har}
contained
half shifts of $S^{1}$'s defined by $x_{i} \rightarrow x_{i} + 1/2$.
In that example there is only one $K3$ present in constructing the
7-manifold and it is straightforward to map the freely acting
$Z_{2}{\times}Z_{2}$ to the heterotic side.
So what is the dual of the supergravity theory on $ J$?
$J$ was constructed by taking $T^{7}$, orbifolding by
three $Z_{2}$'s ({\sl each of which defines a particular $K3\times T^{3}$ })
and resolving all the singularities. In a sense, we have three
overlapping $K3$'s and when the singularities are suitably resolved
we give non-trivial holonomy to the whole $T^{7}$, promoting it to
$G_{2}$. This implies that on the heterotic side, the dual theory
should be 'an overlapping' of three $Z_{2}{\times}Z_{2}$ orbifolds,
because each of the three $K3$'s on the supergravity side should
be treated on an equal footing.
Denote by ${\alpha}\prime$, ${\beta}\prime$ and ${\gamma}\prime$ the
action of the $Z_{2}{\times}Z_{2}$
generated by $(\beta,\gamma)$, $(\alpha,\gamma)$ and $(\alpha,
\beta)$ respectively on the heterotic side. First we note that,
if treated separately, the $Z_{2}{\times}Z_{2}$ orbifolds given by
${\alpha}\prime$, ${\beta}\prime$ and ${\gamma}\prime$ each produce
the same massless spectra as the model condidered in \cite{har} \footnote
{Appropriate $S^{1}$ shifts are included in the
duality map, so that the adiabatic argument of \cite{vaf} applies
to each orbifold},
namely four vector multiplets and 19 chiral multiplets.
To find the massless spectrum of our 'overlapping orbifold', it is only
necessary to note that we have three copies of the same entity
on both sides of the duality map.{\sl On the supergravity side
resolution of singularities of {\it each} $K3$ gave a specific
massless spectrum. On the heterotic side, each of the three
orbifolds involved in the 'overlapping' each give essentially the same
compactification}. This suggests that
we sum the massless spectra separately,
which leads naturally to the definition
of the 'overlapping orbifold' as one in which
each orbifold involved in the process should
be treated separately, and the spectra summed.
However, we should only
count the dilaton and six moduli once. This gives a
spectrum with precisely 12 vector multiplets and 43 chiral
multiplets.
\section{Conclusions.}
It is further interesting to note that all the examples of Joyce
constructed from $Z_{2}$'s
have the Betti numbers of the singular $T^{7}$ as $b_{2} = 0$ and
$b_{3} = 7$. This should then always
correspond to the seven chiral multiplets
containing the dilaton and moduli. In our example, the number of singularities
of each isometry then corresponded on the heterotic side to
the number of surviving $ N=4$ vector multiplets, each of which
give rise to one $ N=1$ vector multiplet and three $ N=1$ chiral
multiplets on {\it both} sides.
Of the more general cases considered in \cite{J2}, it turns out
that for certain subsectors of the singularities there exists more
than one topologically distinct ways of resolving.
The different resolutions add different numbers to $b_{2}$ and $b_{3}$.
This then should correspond on the heterotic side to subsets of $N=4$
vector multiplets surviving the orbifold projection and possibly to
extra massless states from the twisted sectors.
It is thus natural to conjecture that
there is a heterotic dual for each of the manifolds given in \cite
{J2} and this deserves further investigation \cite{ba}.
\newline
{\bf Acknowledgements}.
The author is extremely indebted to Chris Hull,
Wafic Sabra, and Tomas Ortin for discussions and in particular
to Ashoke Sen for many
illuminating conversations.
The author would also like to thank PPARC, by whom this work is supported.
|
\section{Introduction}
In recent years strongly interacting bosons attracted a lot of
attention. Experiments on thin granular films show a zero temperature
superconductor to insulator (SI) phase transition as a function of the film
thickness~\cite{gold} or magnetic field~\cite{hpsi}, and experiments on
Josephson junction arrays (JJA) exhibit the same SI transition as a function
of the ratio between Josephson coupling and charging energy~\cite{jja}. These
findings are interpreted in terms of a bosonic model~\cite{efe,doni,fwgf} in
which the competition between the hopping of Cooper pairs and the Coulomb
interaction between them is responsible for the SI transition.
In relation to the effect of disorder on the SI transition a Bose-glass phase
is discussed in Refs.~\cite{fwgf,f-db,zimrev}.
An interesting extension of the minimal model with on-site interaction only is
to account for finite-range interactions, which are present in Josephson
junction arrays and $^4$He-films on substrates in two dimensions, and bulk
$^4$He in three dimensions. Artificially fabricated JJA's are ideal model
systems to study
the superconductor to Mott-insulator transition as a function of the coupling
constants or magnetic field~\cite{jja}. The finite range of the interaction in
these systems leads to commensurability and frustration effects.
Furthermore, coexistence phases may appear that are called supersolids. In
these phases a charge-density wave (CDW) that is stabilized by the interaction
may coexist with superfluidity. In other words the system has diagonal
long-range order (LRO) and off-diagonal LRO at the same time.
The supersolid phase was studied in the early seventies after Andreev and
Lifshitz~\cite{andlif} suggested that vacancies in a quantum crystal such as
$^4$He might Bose-Einstein condense at low temperatures without destroying the
crystal structure, thereby establishing a superfluid solid~\cite{l-ss}, or
supersolid. Normally bosons at zero temperature are either superfluid (with
off-diagonal LRO) or solid (with diagonal LRO). However, for a finite
range of the interactions between the bosons a coexistence phase was
predicted within mean-field approximations~\cite{mt-ss,lf-ss,bfs,rs-ss}.
Experiments have been performed on $^{4}$He, but no positive
identification of this coexistence phase has yet been made. There are,
however, hints towards such a phase \cite{m-ss,lg-ss}.
Recently several other kinds of coexistence phases were studied.
The possibility of a spontaneous vortex anti-vortex lattice in
superfluid films was explored in Ref.~\cite{vavl} and a coexistence
phase of superfluidity and hexatic orientational order was proposed
in Ref.~\cite{hex}. Orientational order in incompressible quantum Hall fluids
is discussed in Ref.~\cite{leon}. Collinear supersolids were studied in
Refs.~\cite{gergy,rs-mc}. Finally, we mention the relation between 2D bosons
and 3D flux-lines in type II superconductors (high $T_{c}$ materials)
in a magnetic field~\cite{nelson,fl-d,feigel}. Also in these
systems different kinds of LRO may coexist and the equivalent of
the supersolid is discussed in Refs.~\cite{blatt,frey}.
Related is the question whether or not vortices may form a disentangled
liquid, which would imply a normal ground state for bosons with
long-range Coulomb interaction~\cite{feigel}.
In this paper both the SI transition and the supersolid phase will be
studied in some detail. In the next section several related models for
interacting bosons will be discussed, of which the Bose-Hubbard model
is the most general. In the limit of infinite on-site interaction the
Bose-Hubbard model reduces to the spin $\frac{1}{2}$ XXZ model.
For a large number of bosons per lattice site the Quantum-Phase model
is applicable. The phase diagram of the latter two models are first
determined within mean-field theory in Section~\ref{sec-mf}.
A variational Ansatz for the ground state
wave function that treats all three models on equal footing~\cite{baltin}
is presented in Section~\ref{sec-vm}. In Section \ref{sec-nm} more
accurate results for the phase diagram will be obtained by means of
Quantum Monte Carlo. Part of the data has already been
published as Ref.~\cite{ow-ss}. Finally the paper closes with a
discussion of the various results in Section~\ref{sec-d}.
\section{Three model Hamiltonians}
\label{sec-tmh}
A convenient starting point for the description of interacting lattice
bosons is the Bose-Hubbard Hamiltonian
\begin{equation}
H_{BH}=\frac{1}{2}\sum_{ij}n_{i}U_{ij}n_{j} -\mu\sum_{i}n_{i}
-\frac{t}{2}\sum_{\langle ij\rangle}(b^{\dagger}_{i}b_{j}+
b^{\dagger}_{j}b_{i})\;,
\label{bh}
\end{equation}
where $b^{\dagger}, b$ are the creation and annihilation operators
for bosons that satisfy the commutation relation
$[b_i,b_j^{\dagger}]=\delta_{ij}$ and $n_i=b^{\dagger}_ib_i$ is the number
operator. The first term describes
the density-density interaction between the bosons. We will take it
to be short range, $U_{0}$ for on-site, $U_{1}$ for nearest neighbors,
and $U_{2}$ for second-nearest neighbors. The second term describes
the coupling of the density to a chemical potential, and the third
describes the hopping from site to site with hopping integral $t$. The
first two terms will cause the bosons to be localized if the interaction
dominates over the hopping. In the opposite case where the hopping
dominates the bosons will form a superfluid.
In the limit of large on-site interaction $U_{0}$, only the states with
0 and 1 boson per lattice site survive, and by identifying $b^{\dagger}
= S^{+} = S^{x}+iS^{y}$ and $n = S^{z}+\frac{1}{2}$ the mapping to the
spin $\frac{1}{2}$ XXZ Hamiltonian is obtained. It reads
\begin{equation}
H_{XXZ}=\frac{1}{2}\sum_{i,j}S^{z}_{i}U_{ij}S^{z}_{j}
-h\sum_{i}S^{z}_{i}
-t\sum_{\langle ij\rangle}[S^{x}_{i}S^{x}_{j}+S^{y}_{i}S^{y}_{j}]\;,
\label{xxz}
\end{equation}
where $h=\mu-\frac{1}{2}\sum_i U_{0i}$.
This Hamiltonian has been
investigated extensively on the mean-field level for nearest-neighbor
as well as next nearest-neighbor interactions.
In the limit of a large number of bosons per lattice site one may
parameterize $b^{\dagger}=\sqrt{\rho}e^{i\varphi}$ and take $\rho$
equal to the boson density. Thus, only the phase remains as a dynamic
variable and the Quantum-Phase Hamiltonian reads
\begin{equation}
H_{QP}=\frac{1}{2}\sum_{i,j}n_{i}U_{ij}n_{j} -\mu\sum_{i}n_{i}
-J\sum_{\langle ij\rangle}\cos(\varphi_{i}-\varphi_{j})\;,
\label{qp}
\end{equation}
where $J=\rho t$. The phase and number are canonical conjugate variables
and satisfy $[\varphi_i,n_j]=i\delta_{ij}$.
This is exactly the Hamiltonian for Josephson junction arrays with
an applied gate voltage in order to tune the chemical potential
\cite{bfkos}. It is convenient to introduce a parameter $n_0=
\mu/\sum_{i}U_{ij}$, which allows us to rewrite the Coulomb interaction
and chemical potential term in Eqs.~(\ref{bh}) and (\ref{qp}) as
\begin{equation}
H_{\mbox{\footnotesize int}}=
\frac{1}{2}\sum_{i,j}(n_{i}-n_{0})U_{ij}(n_{j}-n_{0}) \;.
\label{int}
\end{equation}
We expect that the properties of the Quantum-Phase model will be periodic in
the variable $n_{0}$ with period 1.
\section{Mean-field}
\label{sec-mf}
A first insight into the properties of these models may be gained
by considering their mean-field solutions. The mean-field approximation
for the spin model in Eq.~(\ref{xxz}) is straightforward and was carried
out by various authors \cite{mt-ss,lf-ss,bfs}. The question of a
supersolid phase in this model was first addressed in Ref.~\cite{lf-ss}
and investigated further in Ref.~\cite{bfs}.
We review the spin mean-field theory in Subsect.~\ref{subsec-smf}.
A mean-field approximation for the Quantum-Phase model which is able
to identify both diagonal LRO and off-diagonal LRO appeared in
Ref.~\cite{rs-ss}. It is reviewed in Subsect.~\ref{subsec-qpm}.
Mean-field approaches to the Bose-Hubbard Model have been discussed
in Refs.~\cite{fwgf,skpr}, but to our present knowledge there exists no
mean-field approximation for the Bose-Hubbard model which allows for the
identification of a supersolid phase.
In Sect.~\ref{sec-vm} we present a variational Ansatz for the many-body wave
function that allows the three models to be treated on an equal footing. This
is important, as it allows for a comparison between the models on the same
level of approximation. Since we use a factorizable Ansatz for the
wave function, it also has mean-field character.
The validity of the mean-field approximation in two dimensions is often
limited, since the lower critical dimension for models with continuous
symmetries is two, which is the dimension we are particularly interested
in. This is a serious problem at finite temperature where the models do not
even exhibit superfluid long-range order as formulated in the theorem of
Mermin and Wagner, and Hohenberg~\cite{rmwh}.
Useful results may, however, be gained from mean-field approximations at zero
temperature where quantum dynamics formally adds an extra dimension, the
imaginary time axis. The systems then do have long-range order and mean-field
results may be qualitatively correct. This will be shown in the present
article by scaling arguments and confirmed by the Monte Carlo simulations. In
the remainder of this section we present the different mean-field type
approaches.
\subsection{Spin Mean-Field}
\label{subsec-smf}
The spin mean-field theory can be formulated by linearizing the spin-spin
interaction. This yields the standard self-consistency equations for the
magnetization
\begin{equation}
\langle{\bf S}_{i}\rangle=-\frac{1}{2}
\frac{{\bf H}_{i}}{\mid{\bf H}_{i}\mid}
\tanh\left(\frac{\beta}{2}\mid{\bf H}_{i}\mid\right)\;,
\end{equation}
where the subscript $i$ refers to the sublattice and $\beta$ is the inverse
temperature (we use units in which $k_{B}=\hbar=c=1$). In order to identify
spatially varying solutions one needs to introduce several sublattices,
at least three for interactions including nearest and next-nearest
neighbors. The symmetry in the $xy$-plane reduces the number of
independent variables and most of the phase boundaries can be
determined analytically. If we allow for three different sublattices
($A$, $B$ and $C$), the effective field is given by
\begin{equation}
{\bf H}_{A}=(-4t\langle S^{x}_{B}\rangle,
-4t\langle S^{y}_{B}\rangle,2U_{1}\langle S^{z}_{B}\rangle
+2U_{2}\langle S^{z}_{C}\rangle-h) \;,
\end{equation}
and corresponding equations for ${\bf H}_{B}$ and ${\bf H}_{C}$.
The connection to hard-core bosons is made by remembering that $S^z$
corresponds to the particle number, and $S^{x/y}$ to the boson creation
operators. A staggered magnetization in $z$-direction implies a solid (
charge-density wave) ordering of the bosons.
Finite magnetization in $xy$-direction
translates into off-diagonal LRO for the boson model. The coexistence phase of
both types of LRO is the supersolid phase for the bosons. The result is shown
in Fig.~\ref{spinfig} for $U_{2}=0.1\;U_{1}$. Two different types of solid
ordering occur for different chemical potential, i.e., $(\pi,\pi)$ ordering
around half-filling, and additional $(\pi,0)$ and
$(0,\pi)$ ordering around quarter filling.
\subsection{Quantum-Phase Model}
\label{subsec-qpm}
Now we turn to the mean-field treatment of soft-core bosons following
Refs.~\cite{rs-ss,khw-d}. After decoupling the hopping and the interactions
between the different sites in Eq.~(\ref{qp}), one arrives at a local problem
\begin{equation}
H_{mf}^i\;|\psi_i\rangle = E_i\;|\psi_i\rangle \; ,
\end{equation}
with the $2\pi$-periodic solutions for the wave function in the phase
representation
\begin{equation}
\psi_i(\varphi_i) = \exp(i\eta_i x_i) \; f_i(x_i) \; ,
\label{eq:sol}
\end{equation}
where $\varphi_i=2\,x_i$, $\eta_i= 2\,n_{0} \;-\; \frac{2}{U_0}
\sum_{j \neq i} U_{ij}(\langle n_{j}\rangle-n_{0})$.
After these substitutions the following Mathieu equation for $f$ is obtained
\begin{equation}
\left\{ \frac{\partial^2}{\partial x_i^2} \;
+\;e_i\;+\;2\,r_i\cos(2\,x_i) \right\} f_i(x_i) \;=\;0
\end{equation}
\begin{equation}
\mbox{with}\ \ \ r_i\;=\;\frac{4J}{U_0}
\sum_{\beta}\langle\cos(\varphi_{i+\beta})\rangle \; .
\end{equation}
The sum runs over nearest neighbors $\beta$ of $i$. The parameter
$e_i$ is proportional to the energy and should be minimized. The ground
state is defined via the self-consistency equation
\begin{eqnarray}\nonumber
\langle n_i\rangle\ \ \ \ &=&\ \ \ \langle\psi_i^0\,|\;n_i\;
|\psi_i^0\rangle\\
\langle\cos(\varphi_i)\rangle&=&\langle\psi_i^0\,|\,
\cos(\varphi_i)\,|\psi_i^0\rangle \;.
\label{sc}
\end{eqnarray}
In addition the free energy has to be minimized in order to have an
unambiguous solution of Eq.~(\ref{sc}). The numerical solution of this set of
equations is straightforward. In order to allow for spatially varying solutions
we again define sublattices. Our numerical solutions confirm the results of
Ref.~\cite{rs-ss}, however, some details differ.
For on-site interaction we obtain Mott insulating lobes centered around
integer values $N$ of $n_0$, see Fig.~\ref{mottfig}. In these incompressible
lobes the density is pinned to $N$. At half integer values of $n_0$ the
critical value of the hopping vanishes, due to the degeneracy in energy of
states with $N$ and $N+1$ particles per site.
Including nearest-neighbor interaction the calculation has to be performed on
two sublattices. We obtain two different kinds of insulating lobes, one with
integer filling and homogeneous density, another with half-integer filling and
a checkerboard charge-density wave, or solid order, centered around half
integer values of $n_0$. In addition we find a region of supersolid phase
around the half lobe where the checkerboard charge-density wave and
superfluidity coexist, see Fig.~\ref{cbssfig}. Again, the superfluid phase
occupies the parts on the right of the phase diagram and also between the
insulating lobes down to zero hopping, in slight contrast to the result of
Ref.~\cite{rs-ss}.
The fraction of supersolid phase in the phase diagram depends on the ratio
$U_{1}/U_{0}$, see Fig.~\ref{numbfig}. In the hard-core limit $U_0 \rightarrow
\infty$ the supersolid phase vanishes. An expansion of the ground state wave
function $|\psi_i^0\rangle=\sum_{n}c^{i}_{n}|n\rangle$ into particle number
eigenstates $|n\rangle$ yields more information about the nature of the
supersolid phase. The inset of Fig.~\ref{numbfig} shows $|c_{2}|^{2}$ at
$n_{0}$=0.5 as a function of $J/U_{0}$. The figure indicates that double
occupancy is important for the supersolid phase. Thus, we conclude that the
supersolid phase is favored by fluctuations in the particle number.
The inclusion of further interactions leads to more structure in the phase
diagram. For next-nearest neighbor interaction we introduce three
sublattices. As a result, The phase diagram shows a lobe with quarter filling
and a phase where this quarter structure and superconductivity coexist,
see Fig.~\ref{nnnfig}.
\subsection{Hard-core vs. Soft-core}
\label{subsec-hcsc}
The previous subsections showed two extreme cases of the Bose-Hubbard model,
the hard-core limit and the limit of large particle numbers. Although the
phase diagrams look similar, these two limits behave qualitatively different
as far as the supersolid phase is concerned.
In the hard-core limit the existence of the supersolid is related to a finite
next-nearest neighbor interaction. The supersolid does not exist in the model
including nearest-neighbor interaction only~\cite{bfs}. Furthermore, there is
no supersolid phase at exactly half filling. This leads to the following
interpretation of the supersolid phase. At densities corresponding to half
filling the particles form an incompressible solid. Away from half filling
superfluidity is enabled by defects in the solid structure that Bose-Einstein
condense. This interpretation does, however, not explain why next-nearest
neighbor interaction is necessary for the formation of the supersolid.
In the limit of many particles per site the Quantum-Phase model applies. In
this case the supersolid phase already exists for nearest-neighbor interaction
only and at half filling. This is related to excitations which are forbidden
in the hard-core limit. We observe that a large nearest-neighbor interaction
$U_{1}$ or small on-site interaction $U_{0}$ favors the supersolid, in the
hard-core limit the supersolid is suppressed. In the soft-core limit the
supersolid phase is also present at half-filling at the tip of the
checkerboard lobe. Thus, it seems that the system itself generates the defects
(particle-hole excitations) that Bose condense, thereby turning the solid into
the supersolid.
\section{Variational Method}
\label{sec-vm}
Here we present a method which allows us to treat the three models within one
scheme~\cite{baltin}. The idea is to guess a variational wave function for the
problem \cite{gutz}. We use a variational wave function that is inspired by
earlier successful Gutzwiller-like treatments of the spin model~\cite{huse}
and the Bose-Hubbard model~\cite{roko,kraut}. The occupation number
representation is well suited for all three models. Explicitly we resolve the
identity as
\begin{equation}
1=\sum_{\{n_i\}}|\{n_i\}\rangle\langle\{n_i\}|
\end{equation}
and express the trace as
\begin{equation}
\mbox{Tr ...}=\sum_{\{n_i\}}\langle\{n_i\}| ... |\{n_i\}\rangle,
\end{equation}
where the $n_i$ are in $[0,1]$ for the spin $\frac{1}{2}$ XXZ model,
in $[0..\infty]$ for the Bose-Hubbard (BH) model, and in
$[-\infty..\infty]$ for the
Quantum-Phase (QP) model. Besides the difference in the allowed values of
$n_i$, the matrix elements differ for the three models, e.g.,
\begin{eqnarray}\nonumber
\mbox{BH model:}\;\;\;\;\;\; & & \langle\{n_i'\}|b_k|\{n_i\}\rangle
= \langle\{n_{i\neq k}'\}|\{n_{i\neq k}\rangle
\langle n_k'|n_k-1\rangle \sqrt{n_k} \\
\mbox{QP model:}\;\;\;\;\;\; & &
\langle\{n_i'\}|\exp{i\varphi_k}|\{n_i\}\rangle
= \langle\{n_{i\neq k}'\}|\{n_{i\neq k}\rangle
\langle n_k'|n_k-1\rangle
\end{eqnarray}
As our variational wave function we take the product of single-site wave
functions. This amounts to neglecting certain correlations and is therefore
similar to a mean-field approximation. We use the following Ansatz
$$
|\psi\rangle=\prod_{i=1}^{N}\frac{1}{\sqrt{Z_{i}}}
\sum_{\{n_{i}\}}\exp\left\{-k_{i}
(n_{i}-m_{i})^{2}/2\right\}|n_{i}\rangle
$$
\begin{equation}
\mbox{where }\ \ \ \ Z_{i}=
\sum_{\{n_{i}\}}\exp\left\{-k_{i}(n_{i}-m_{i})^{2}\right\} \;,
\label{ansatz}
\end{equation}
and $k_{i}$ and $m_{i}$ are real variational parameters. The single-site wave
function is a superposition of boson number eigenstates weighted with a
Gaussian. The average is controlled by $m_i$, the width by $k_i$, and spatial
variations (different sublattices $i$) are allowed.
By minimizing the energy expectation value $E=\langle\psi|H|\psi\rangle$ with
respect to $k_{i}$ and $m_{i}$ we obtain the ground state wave function within
our approximation. In the following, expectation values are understood
to be calculated with the ground state wave function. As the discrete
sums can in general not be performed analytically, we treat the problem
numerically. For on-site and nearest-neighbor interactions two sublattices
arranged in a checkerboard configuration
are introduced. We define the superconducting order parameter as
$\langle\exp(i\varphi)\rangle$ for the QP model, $\langle b\rangle$ for the BH
model, and $\langle S^{-} \rangle$ for the XXZ model. The structure factor
\begin{eqnarray}\label{eqstr}
S(\vec{q}) & = & \bigg(\frac{1}{N}\bigg)^{2}\sum_{i,j}
\exp[i\vec{q}\cdot(\vec{r}_{i}-\vec{r}_{j})]\langle n_{i}n_{j}\rangle
\end{eqnarray}
yields information about the solid order. A non-vanishing structure factor at
finite $\vec{q}$ signals diagonal LRO. By decomposing the lattice into two
sublattices we gain information about $S(\pi,\pi)$. A finite $S(\pi,\pi)$
corresponds to a checkerboard arrangement of the particles. For next-nearest
neighbor interactions we introduce four sublattices which yields additional
information about $S(\pi,0)$ and $S(0,\pi)$.
Figure~\ref{qpvarfig} shows the phase diagram for the QP model with
on-site and nearest-neighbor interaction obtained with our
variational Ansatz. Both the superfluid-insulator and the crystalline order
transition can be identified and agree with previous mean-field calculations
\cite{rs-ss,ow-ss}. The phase boundaries are periodic in $n_{0}$. We
find that superfluidity sets in simultaneously on both sublattices $A$
and $B$. Within the supersolid phase the value of the superconducting
order parameter on $A$ differs from that on $B$, but both are nonzero.
Beyond the crystalline order phase boundary the order parameter
does not exhibit any difference on the sublattices.
For the XXZ model the variational method confirms that there is no
supersolid for on-site and nearest-neighbor interaction only.
In the presence of next-nearest neighbor interaction we find
supersolid phases in perfect agreement with \cite{bfs} and Fig.~\ref{spinfig}.
The phase diagram for the BH Hamiltonian is shown in Fig.~\ref{bhvarfig}.
The size of the lobes decreases with increasing $n_0$. At point $\alpha$
the supersolid vanishes. This might be due to the lower bound for the
occupation numbers $n\geq 0$. For small $n_0$ charge fluctuations are
suppressed. Hence, we conclude that charge fluctuations are necessary for the
supersolid. For large $n_{0}$ the phase boundaries of the QP model approach
those of the BH model, if the hopping is rescaled according to
Section~\ref{sec-tmh}, i.e., if $t$ and $J/n_{0}$ are identified.
\section{Quantum Monte Carlo}
\label{sec-nm}
Several methods are available for determining the phase diagram for
the three model Hamiltonians by numerical means. Exact diagonalization
is possible for the XXZ model and small systems, i.e., less than about 26
lattice sites. Furthermore it is possible to perform Monte Carlo
simulations for all the three models on much larger systems.
The Bose-Hubbard model was studied in one dimension in
Refs.~\cite{bsz,nfsb}, and one data point for the phase boundary was
obtained in two dimensions in Ref.~\cite{kt-mc}.
The Bose-Hubbard model was more recently studied in the context of
a supersolid phase in Ref.~\cite{gergy}.
The two-dimensional Quantum-Phase model was studied in Ref.~\cite{swgy}
in relation to the universal conductivity at the SI transition and in
Refs.~\cite{rs-mc,ow-ss} in relation to supersolid phases.
In this section we present more results on Monte Carlo simulations
of the two-dimensional QP model in addition to those in Ref.~\cite{ow-ss}.
We discuss the on-site problem and compare to the analytic results of
$t/U$ expansions \cite{fm-bh,fm-ex}. The inclusion of nearest-neighbor
interactions allows us to study the supersolid phase.
\subsection{Duality Transformation}
\label{subsec-dt}
In order to study the Quantum-Phase model by means of Monte Carlo
we map the 2D quantum model onto a (2+1)D classical model of discrete
divergence-free currents. The essential feature of this mapping were
presented in
Ref.~\cite{fl-d} and the derivation makes use of duality
transformations that were developed in Ref.~\cite{jkkn}. We start
from the basic expression for the partition function $Z$ for the
Quantum-Phase model of Eq.~(\ref{qp})
\begin{equation}
Z=\mbox{Tr}\;\exp(-\beta H_{QP})\;,
\label{az}
\end{equation}
where $\beta$ is the inverse temperature. We go over to a Euclidean
path-integral formulation by introducing time-slices, i.e., dividing
$\beta$ in $N_{\tau}$ intervals of size $\epsilon$, such that
$N_{\tau}\epsilon=\beta$. Inserting complete sets of states at each
time slice we arrive at
$$
Z=\sum_{\{n_{i,\tau}=0,\pm1,\pm2,\cdots\}}
\int{\cal D}\varphi_{i,\tau}\exp{\Big\{
-\frac{\epsilon}{2}}\sum_{ij,\tau}
n_{i,\tau}U_{ij}n_{j,\tau}+ \epsilon\mu\sum_{i,\tau}n_{i,\tau}
$$
\begin{equation}
+i\sum_{i,\tau}n_{i,\tau}\dot{\varphi}_{i,\tau}+
\epsilon J\sum_{\langle ij\rangle,\tau}
\cos(\varphi_{i,\tau}-\varphi_{j,\tau}-A_{ij}){\Big\}}\;.
\label{laz1}
\end{equation}
We included a coupling to a vector potential $A_{ij}=(2\pi/\Phi_{0})
\int^{j}_{i}{\bf A}\cdot d{\bf l}$ for later convenience.
In order to proceed, we make one approximation, the Villain
approximation~\cite{vill} for the cosine term. This amounts to
expanding the cosine around all its minima
\begin{equation}
\exp\{\epsilon J\cos(\varphi_{i\tau}-\varphi_{j\tau}-A_{ij})\}
\approx\!\!\!\sum_{\{m_{ij,\tau}\}}
\exp\{-\frac{\epsilon J F(\epsilon J)} {2}
(\varphi_{i\tau}-\varphi_{j\tau}-2\pi m_{ij,\tau}-A_{ij} )^{2}\}\;,
\label{eq:av}
\end{equation}
where $m$ is a directed discrete field that lives on the bonds between
lattice sites. The function $F$ is determined by demanding that the
first two Fourier coefficients of the expressions on the left and
right hand side of Eq. (\ref{eq:av}) are equal~\cite{vill} and is given by
\begin{equation}
F(x)=\frac{1}{2x\log\{I_{0}(x)/I_{1}(x)\}}\;.
\label{vill}
\end{equation}
For large arguments $F$ approaches 1, as is clear from a direct
expansion of the cosine potential. After a subsequent Poisson
resummation, i.e., writing
$$
\sum_{\{m_{ij,\tau}\}}f[m_{ij,\tau}]=
\sum_{\{{\cal J}_{ij,\tau}\}}\int {\cal D}m f[m_{ij,\tau}]
\exp(2\pi i\sum_{ij,\tau}m_{ij,\tau}{\cal J}_{ij,\tau})\;,
$$
an integration over the fields $m_{ij,\tau}$ and the phases
$\varphi_{i,\tau}$ yields a representation in terms of divergence-free
discrete current loops
$$
Z=\left[\epsilon JF(\epsilon J)\right]^{-1}
\sum_{\{{\cal J}^{\nu}_{i,\tau}\}}
\delta(\nabla_{\nu}{\cal J}^{\nu})\exp\{-S[{\cal J}]\}\;,
$$
$$
S[{\cal J}]=\frac{\epsilon}{2}\sum_{ij,\tau}
{\cal J}^{\tau}_{i,\tau} U_{ij} {\cal J}^{\tau}_{j,\tau}
-\epsilon\mu\sum_{i,\tau}{\cal J}^{\tau}_{i,\tau}
+\frac{1}{2\epsilon J F(\epsilon J)}\sum_{i,\tau,a=x,y}
\left({\cal J}^{a}_{i,\tau}\right)^{2}
-i\sum_{i,\tau}\vec{A}_{i}\cdot\vec{{\cal J}}_{i,\tau}
$$
\begin{equation}
=\sqrt{\frac{2}{K}}\left\{\sum_{ij\tau}({\cal J}^{\tau}_{i,\tau}-n_{0})
\left(\delta_{ij}+\frac{U_{1}}{U_{0}}\delta_{\langle ij\rangle}\right)
({\cal J}^{\tau}_{j,\tau}-n_{0})+\!\!\sum_{i\tau,a=x,y}\!\!
\left({\cal J}^{a}_{i,\tau}\right)^{2}\right\}
-i\sum_{i,\tau}\vec{A_i}\cdot\vec{{\cal J}}_{i,\tau}\;,
\label{al}
\end{equation}
where $\nu=x,y,\tau$. The effective coupling constant is given by
$\sqrt{2/K}$ with $K=8F(\epsilon J)J/U_{0}$, where $\epsilon$ is
defined implicitely by $\epsilon=1/\sqrt{U_0F(\epsilon J) J}$.
The time components ${\cal J}^{\tau}_{i,\tau}$ of the current correspond
to the boson numbers $n_{i,\tau}$
along their world lines. We keep track of the energy dependence of
the determinant that arises from integrating out the $m_{ij,\tau}$.
This is relevant for determining the energy density and specific heat
by means of Monte Carlo.
For zero magnetic field ($\vec{A}=0$) the action in the representation
of Eq.~(\ref{al}) is real, and therefore suitable for Monte Carlo
methods. Using the standard Metropolis algorithm we generate
configurations of currents in a system of size $L\times L\times
L_{\tau}$.
The condition that the currents ${\cal J}^{\nu}$ be divergence-free is taken
into account by making Monte Carlo steps that preserve the property
$\nabla_{\nu}{\cal J}^{\nu}=0$. Thus, we create or annihilate small current
loops at every lattice site in all three directions, as well as
periodic current loops that go through the whole system which is
taken to have periodic boundary conditions in all three directions.
The generation of configurations may be done in a canonical as well as
in a grand-canonical way. Here we work in the grand-canonical ensemble
in order to make contact to the mean-field phase diagrams.
In the last line of Eq.~(\ref{al}) we used
$\epsilon=1/\sqrt{U_0F(\epsilon J) J}$.
With this choice the couplings are isotropic for on-site interaction,
leading to an efficient numerical algorithm. The choice of the time lattice
spacing $\epsilon$ needs some justification. Introducing time slices
in the partition function Eq.~(\ref{laz1}) is exact in the limit
$\epsilon \rightarrow 0$. This would produce extremely anisotropic
couplings in the current-loop model and the numerical simulation
would become impossible. Thus, a more reasonable choice for the time
lattice spacing is necessary. The plasma frequency
$\omega_{p}=\sqrt{JU_0}$ is a natural frequency for spin waves.
A cutoff beyond this frequency by introducing a time spacing
$\epsilon \approx 1/\omega_{p}$ should not do any harm. We tested
the dependence on the choice of $\epsilon$ for the quantities of
interest and observed that they are not sensitive to variations of
$\epsilon$ by one order of magnitude, provided that the temperature
is kept constant, i.e., $\beta=1/T=const.=N_{\tau}\epsilon$.
Additional justification for the choice of a finite time spacing
$\epsilon$ in the present paper is given by the fact that we are
only interested in the critical regime, where a high-frequency cutoff
should be irrelevant.
\subsection{Finite-Size Scaling}
The mapping to the current-loop model Eq.~(\ref{al}) is well suited
for simulating the behavior of the Quantum-Phase model with short
range interactions. In practice we are able to study system sizes
up to $ 12 \times 12 \times 12$, or $10 \times 10 \times 25$,
respectively. The factor that limits the largest systems to have $L$= 12 is
the necessity to make the periodic current loops in the spatial
directions. Only they can change the superfluid stiffness, which is a
topological quantity that cannot be changed by making local current
loops. As phase transitions take place only in the thermodynamic
limit, we need an additional ingredient that relates the data obtained
on finite-size systems to infinite system size properties as the
critical coupling and exponents. This is provided by finite-size scaling.
Let us first consider the onset of superfluidity in the system.
Therefore we study the behavior of the superfluid stiffness $\rho_0$,
a quantity which measures the response to a twist in boundary
conditions for the phase. A finite value of $\rho_0$ in the thermodynamic limit
reflects long-range phase coherence, i.e., off-diagonal LRO or
superfluidity. A twist in the boundary conditions for the phases by an
angle $\Theta$ increases the free energy density by
\begin{equation}
\Delta f=\frac{1}{2}\rho_0\left(\frac{\Theta}{L}\right)^2
\label{df}
\end{equation}
The twist may be realized by applying a magnetic field. In general
the frequency and wave-vector dependent stiffness $\rho(k,\omega_{\mu})$
can be defined as
\begin{equation}
\rho(k,\omega_{\mu})=\left(\frac{1}{2e}\right)^{2}
\sum_{i,\tau}\frac{-\delta^{2}\ln Z}
{\delta A^{x}_{i,\tau}\delta A^{x}_{0}} \Bigg|_{\vec{A}=0}
e^{-i\omega_{\nu}\tau-ikr_{i}}\;.
\end{equation}
The zero-frequency and zero wave-vector component of $\rho(k,\omega_{\mu})$
defines the superfluid stiffness. In terms of the currents it reads
\begin{equation}
\rho_0=\rho_{(k=0,\omega_{\mu}=0)}=\frac{1}{L^d L_{\tau}}
\left\langle \left(\sum_{i,\tau}{\cal J}^x_{i,\tau}\right)^2\right
\rangle=\frac{1}{L^{d-2} L_{\tau}}\langle w^2_x \rangle\,.
\end{equation}
The winding numbers $w_x=\frac{1}{L} \sum_{i,\tau} {\cal J}^x_{i,\tau}$
can be measured easily; they are only modified by periodic current
loops, local loops do not change their value. The symmetry in
$x/y$-direction further simplifies the numerics.
Direct measurement in the critical regime is ruled out by the
divergence of the correlation length $\xi$ at the continuous
transition. The powerful method of finite-size scaling, however,
takes advantage of this fact and allows us to study the critical
behavior.
Near a continuous phase transition the diverging coherence length
is cut off by the finite system size $L$. As a result all quantities
will depend on the ratio $\xi/L$. A finite-size scaling relation
for the superfluid stiffness $\rho_0$ is readily derived~\cite{cha}.
Assuming hyperscaling, the critical part of the free energy density
behaves as the inverse correlation volume $\xi^d\xi_{\tau}$. Here, $\xi$
is the correlation length in real space which may be different from
the correlation length $\xi_{\tau}$ in imaginary time, depending on
the quantum dynamics of the model. The dynamical critical exponent
$z$ can be introduced through $\xi_{\tau}\propto\xi^z$. Combining
Eq.~(\ref{df}), the hyperscaling relation and the fact that $\xi$
is cut off by $L$ ($\xi_{\tau}$ by $L_{\tau}$) we arrive at the
scaling behavior
\begin{equation}
\rho_{0}={L^{2-d-z}}\tilde{\rho}
(b L^{1/\nu}\delta, L_{\tau}/L^{z})\,.
\label{rhosc}
\end{equation}
Here, $b$ is a non-universal scale factor and $\tilde{\rho}$ is a
universal scaling function with a smooth dependence on $L_{\tau}/L^{z}$
and $\xi/L=\delta^{-\nu}/L$. We assumed a power law critical behavior
$\xi \propto \delta^{-\nu}$ with dimensionless distance to the
transition $\delta=(K-K^{\ast})/K^{\ast}$ and critical coupling
$K^{\ast}$. At the critical point ($\delta=0$),
$L^{d+z-2}\rho_{0}$ is a function of $L_{\tau}/L^{z}$ only. Thus, plots
of $L^{d+z-2}\rho_{0}$ vs.\ $K$ will intersect at the transition if
$L_{\tau}/L^{z}$ is kept constant, with no further fitting involved.
Furthermore, the data for $L^{d+z-2}\rho_{0}$ plotted as a function
of $L^{1/\nu}\delta$ for different system sizes should collapse onto
one single curve. This allows the exponent $\nu$ to be determined.
As we are interested also in diagonal LRO and a possible supersolid
phase, we measure the structure factor as well. The structure factor
is the $(\pi,\pi)$-component of the equal time density-density
correlation, see Eq.~(\ref{eqstr}), and may be derived as the second
variation of the free energy with respect to a staggered chemical
potential. Expressed in terms of the currents ${\cal J}^{\tau}$ it reads
\begin{equation}
S_{\pi}=\frac{1}{L^{4}L_{\tau}}\Big\langle\sum_{ij,\tau}(-1)^{i+j}
{\cal J}^{\tau}_{i,\tau}{\cal J}^{\tau}_{j,\tau}\Big\rangle\;.
\end{equation}
The scaling relation for the structure factor is
\begin{equation}
S_{\pi}=L^{-2\beta/\nu}
\tilde{S}(b'L^{1/\nu}\delta,L_{\tau}/L^{z})\;.
\end{equation}
This scaling relation arises, since the structure factor $S_{\pi}$
is related to the square of the staggered magnetization order parameter
$M=\sum_{i}(-1)^{i}{\cal J}^{\tau}_{i}$ for which the exponent $\beta$ is the
order parameter exponent. Near the phase transition,
$M\sim \delta^{\beta}$ and $\xi\sim \delta^{-\nu}$, from which
the quoted form follows. From a three-parameter fit to the scaling
relation for different system sizes, the exponents $\nu$ and $\beta$
as well as the critical coupling constant $K^{\ast}$ are determined.
Again, plots of $L^{2\beta/\nu}S_{\pi}$ will cross at the critical
point if the ratio $L_{\tau}/L^{z}$ is kept constant.
In both scaling relations a knowledge of $z$ is assumed. This is
well-known for the SI transition and the scaling of $\rho_0$, i.e.,
in general $z$=2, except for the tips of the lobes where $z$=1 as
dictated by particle hole-symmetry. This argument is based on a
coarse-graining treatment \cite{doni} that allows for the explicit
construction of a Ginzburg-Landau-Wilson free energy functional for
the superconducting order parameter \cite{fwgf,owfs,jk-cg,kz-cg},
from which $z$ can be read off.
For the transition related to the structure factor the situation
is less clear. We were, for instance, not yet successful in explicitly
constructing the Ginzburg-Landau-Wilson free energy functional for
the CDW transition. From symmetry arguments one expects $z$=1 at the
mean-field level~\cite{bafr}. Also the spin-wave analysis of
Ref.~\cite{gergy} predicts $z$=1. However, it is unclear whether or not
$z$ for the CDW will be renormalized by the coupling to the gapless
sound mode in the superfluid part. As a working hypothesis we take
$z$=1 and comment later on the consistency of this assumption.
\subsection{Numerical Results}
\label{subsec-nm}
We report about the SI transitions for on-site interaction and
nearest-neighbor interaction in the complete range of the chemical potential.
A phase diagram is mapped out and the scaling predictions are verified.
The existence of a supersolid phase in a finite region of the phase
diagram is firmly established.
The case of {\it on-site interaction} has been extensively studied in the
past years. A phase diagram consisting of lobe-like structures in the
$t-\mu$ or $J-n_{0}$ plane
is found. Mott-insulating lobes are centered around integer values of
$n_0$. This has been observed in mean-field approximations, $t/U$ expansions
\cite{fm-bh,fm-ex}, and in Monte Carlo simulations of the one-dimensional Bose
Hubbard model \cite{bsz}. The shape of the lobes in two dimensions is yet
unknown. Therefore we first map out this phase diagram which is shown in
Fig.~\ref{mottfig}.
We performed the simulations for fixed $n_0$ and tuned through the transition
by varying the effective coupling $K$. In the phase diagram in Fig.\
\ref{mottfig} this
corresponds to moving on horizontal lines through the phase transition.
At integer values of $n_0$ the system is particle-hole symmetric and we have
$z=1$. According to the scaling form of Eq.~(\ref{rhosc}) we use $L=L_{\tau}$
which keeps the second argument of the scaling function constant. Four
different system sizes with $L=6,8,10,12$ where studied. For the largest
systems we typically use $10^6$ sweeps through the system for equilibration
and measurement. In each sweep we try all possibilities of local and periodic
current loops. Without any fitting the curves of $\rho_0 L^{d+z-2}=\rho_0 L$
vs.\ $K$ should cross in the critical point $K^*$. This serves as a good test
of the scaling prediction. The critical exponent $\nu$ can be extracted by
fitting the data of different system sizes near the transition to one scaling
curve. This is shown in Fig.~\ref{dat1fig} where we plot the data as a
function of $\delta L^{1/\nu}$. For $n_0=0$ we obtain the critical coupling
$K^*=0.886\pm0.003$, and using a $\chi^2$ fit $\nu=0.69\pm0.06$. The exponent
$\nu$ agrees with the expected 3D-XY behavior, where $\nu\approx 2/3$.
Away from the symmetry points at the tips of the lobes the dynamical critical
exponent is $z=2$. In order to keep $L_{\tau}/L^z$ constant we simulate
systems
with $L_{\tau}=L^2/4$, i.e., $L\times L \times L_{\tau}=6 \times 6 \times 9$,
$8 \times 8 \times 16$ and $ 10\times 10\times 25$.
Now the curves of $\rho_0 L^2$ should cross in the critical point. We obtained
data for $n_0=$0.05, 0.1, ..., 0.45 as shown in Table \ref{table-ons}. From a
fit to the scaling form we find $0.42<\nu<0.52$ with errors of $\approx 20
\%$. This is in agreement with the expected mean-field transition where
$\nu=1/2$. A more accurate determination of the critical exponents turns out
to require an enormous numerical effort with our method and is beyond the
scope of our present work.
In Fig.~\ref{mottfig}, the Monte Carlo data are shown
together with mean-field phase boundaries and the results of the
$t/U$-expansion of Refs.~\cite{fm-bh,fm-ex} in the limit of large $n_{0}$ (or
$\mu$). A lobe shape is
observed, these lobes are sharper than predicted by mean-field, but smoother
than expected from the $t/U$-expansion. Approaching $n_0=0.5$ we observe that
the critical value of $J/U_0$ decreases and a simple extrapolation yields
$J^{\ast}=0$ for $n_0=0.5$. A numerical answer to the question whether
$J^{\ast}$ is zero at $n_0=0.5$ is not possible, since the acceptance rates
are very small for small $J/U_0$.
The inclusion of {\it nearest-neighbor interactions} yields a richer phase
diagram. The scaling for the superfluid stiffness is not modified by the
inclusion of short-range interactions, i.e., the universality class is
preserved. We expect $z=1$ in the case of particle-hole symmetry, $z=2$
otherwise. For nearest-neighbor interaction the lines with integer and
half-integer values of $n_0$ exhibit particle-hole symmetry. For the
transition related to charge-density wave order, we expect $z=1$ throughout
the phase diagram.
Guided by the
mean-field phase diagram we expect the existence of two different kinds of
lobes, with homogeneous particle densities centered around integer values of
$n_0$, and with a checkerboard charge-density wave centered around half integer
values of $n_0$. The supersolid phase is expected in the vicinity of the
checkerboard lobe.
For $n_{0}$= 0 and 0.5 we simulated $L\times L\times L$ systems, where
$L$= 6, 8, 10, 12, as suggested by particle-hole symmetry and $z=1$.
For $n_{0}$=0.1, 0.2 and 0.4 we obtain $z=2$ for the superfluid transition.
In order to keep the ratio $L_{\tau}/L^{z}$ constant, we simulated
$L\times L\times L^{2}/4$ systems, where $L$= 6, 8, 10. For $n_{0}$=0.4 near
the transition for the diagonal LRO we performed the calculations with both
$z=1$ and $z=2$. The results are summarized in
Fig.~\ref{dat2fig} and Fig.~\ref{fig:zdisc} and Table \ref{table-nn}.
In the following we discuss our data in detail.
First we present our data for $n_{0}$=0.5. Figure \ref{dat2fig} shows that
there are two separate transitions for diagonal and off-diagonal LRO with a
coexistence region in between where {\it both} the superfluid density {\it
and} the structure factor scale to a finite value in the thermodynamic
limit. This demonstrates the coexistence of diagonal LRO and off diagonal LRO
for soft-core bosons with nearest-neighbor interaction in 2
dimensions. Table~\ref{table-nn} shows that the exponent $\nu$ is different
for the two transitions. For the transition related to superfluidity (point
$\beta$ in Fig.~\ref{cbssfig}) we find $\nu=0.65\pm0.08$ which is consistent
with the 3D XY universality class. For the transition related to crystalline
order (point $\gamma$) the universality class is not known. We find
$\nu\approx$0.55 and $\beta\approx$0.21.
Also at $n_{0}$=0.4 we find two separate transitions that are the boundaries
for the supersolid phase in between, see Table \ref{table-nn}. As compared to
$n_{0}$=0.5 both transitions are shifted to smaller values of the coupling
constant $K$. This is consistent with the mean-field phase diagram. Again the
two transitions have different critical exponents. The transition related to
superfluidity (the line separating phases ``Sol'' and ``SSol'' in
Fig.~\ref{cbssfig})
has $\nu=0.44\pm0.08$ which is consistent with a mean-field transition in
$d+z=4$ effective dimensions. For the transition related to crystalline order
(between phases ``SSol'' and ``SF'') we compare the data for $z=2$ and $z=1$ in
Fig.~\ref{fig:zdisc}.
A fit to the expected scaling behavior is equally possible in
both cases. The statistical errors of the data for $z=2$ are larger,
since the simulation for $z=2$ requires larger $L_{\tau}$ (up to 25) which
decreases the acceptance ratio for periodic loops in this direction.
The values for the critical coupling and the critical exponents for
both $z=1$ and $z=2$ are shown in Table \ref{table-nn}, they coincide
within the error bars. With the values of Table~\ref{table-nn} the scaling
relation $2\beta=\nu(d+z-2+\eta)$ leads to $0.8\pm0.2=z+\eta$, with
correlation function exponent $\eta$ (small and usually positive). This rules
out $z=2$ and one is led to conclude that indeed $z\sim 1$ for the crystalline
transition, independent of $n_{0}$.
The small value of $\beta$ rules out mean-field behavior for the CDW
transition. Insight into the nature of this transition is gained by the
following consideration. In the neighborhood of this transition, i.e., far in
the superfluid phase, the x,y-components of the currents ${\cal J}$ fluctuate
strongly and may be integrated out as Gaussian fluctuations. In other words:
one expects the staggered magnetization order parameter to couple to the
gapless 4$^{th}$ sound mode of the superfluid.
This yields strong long-range interactions in the time direction for the
$\tau$-components of the currents ${\cal J}$. It is likely that these
long-range interactions are a relevant perturbation and suppress the
exponent $\beta$.
Finally, we discuss the data for $n_{0}$ = 0, 0.1, and 0.2. In these cases
there is only one phase transition, as the Mott-insulating lobes
(phase ``MI'' in
Fig.~\ref{cbssfig}) do not have any non-trivial crystalline order. Our
data are consistent with a transition in the 3D XY universality class
for $n_{0}$=0 and a mean-field phase transition for $n_{0}$ = 0.1 and 0.2.
\section{Discussion}
\label{sec-d}
We studied the $T=0$ phase diagram of two-dimensional interacting
bosons by various techniques. The combination of the various aspects
of these treatments allow us to draw the following conclusions.
For on-site interaction we have seen the lobe structure of the phase boundary
in the $t-\mu$ or $J-n_{0}$ plane. This structure is exactly periodic in the
chemical potential $n_{0}$ in the Quantum-Phase model of Eq.~(\ref{qp}). In
mean
field or variational treatments of the different models, the lobes are
parabolic near their tips. The Monte Carlo results, however, show that the
lobes are sharper than predicted by mean-field approximations. This deviation
near the tips of the lobes is expected by scaling arguments: due to the
additional particle-hole symmetry at the tips of the lobes, the effective
dimension is reduced to $d_{eff}=d+z=3$ and enhanced fluctuations destroy part
of the ordered (superfluid) phase. This is consistent with the value of the
critical exponent $\nu$ obtained from the simulations ($\nu\approx 2/3$) that
agrees well with a 3D XY critical point. Similar sharpening of the lobes is
seen in a higher-order $t/U$ expansion for the BH model~\cite{fm-bh,fm-ex}.
A central point of this paper is the consideration of finite-range
interactions. They lead to a richer phase diagram, due to commensurability
effects. In particular, charge-density wave or solid order
appears in parts of the phase diagram.
In our definition the solid phase is characterized by the breaking
of the translation symmetry of the underlying lattice, i.e., if a
density wave with a wave vector smaller than the lattice is present.
Upon inclusion of nearest-neighbor interactions new half-filled lobes
around half-integer $n_{0}$ appear in the phase diagram, in which the
checkerboard ($\pi,\pi$) density wave is stable. For
next-nearest
neighbor interactions additional density waves exist in which 1 or 3 out of
4 sites are occupied, around $n_{0}$=1/4 or 3/4. These pure density
waves are incompressible Mott-insulators.
Surprisingly, each Mott-insulating charge-density wave seems to have a
corresponding compressible supersolid phase in which the particular
CDW order coexists with off diagonal LRO.
For soft-core bosons the supersolid phases appear already for nearest-neighbor
interaction. Hard-core bosons, on the other hand, supersolidify only if at
least next-nearest neighbor interaction is present. Another difference is the
presence of a supersolid at the tip of the half lobe: hard-core bosons only
form a supersolid away from half-integer filling, whereas soft-core bosons do
so also at half-integer filling.
The difference between Bose-Hubbard and Quantum-Phase models is another
issue. Our results indicate that, on the mean-field level, the two models are,
apart from a trivial rescaling of the hopping matrix element, almost identical
for large $\mu$ or $n_{0}$. For small $\mu$ corresponding to a filling of less
than one boson per site the mean-field solution for the two models differ
essentially and are not described by a simple rescaling of the
parameters. Within our mean-field approximation we find that the supersolid
phase is suppressed around the solid lobe with exactly half filling
(alternating empty and singly-occupied sites). This may be one of the reasons
why in a recent numerical simulation of the BH model \cite{gergy} a supersolid
phase was not found for exactly half filling.
Our combined mean-field and Monte Carlo analysis has shown that the
density-wave order transition is not mean-field like. Both the exponents
($\beta\approx 0.21$ and $\nu\approx 0.55$) and the location of the
phase boundary deviate considerably from the mean-field results
($\beta= 1/2$ and $\nu= 1/2$), see also Fig.~\ref{cbssfig}.
This may be related to the coupling of the CDW order parameter to the
gapless $4^{th}$ sound mode in the superfluid and remains a subject for
further study.
A related issue is the value of the dynamical critical exponent $z$
for this transition. The value $z$=1 used in the simulations gives good
scaling. However, other values of $z$ turn out to give reasonable
scaling as well and yield identical values for $\beta$, $\nu$ and
$K^{\ast}$.
It seems that the CDW transition is rather indifferent
to the value of $z$ used to scale the systems studied numerically.
From the scaling relation
$2\beta/\nu=d+z-2+\eta$ we expect the combination $z$+$\eta\approx$
0.8$\pm$0.2 which would indicate either a small $\eta$ and renormalized
$z\approx$ 0.8 or $z\approx 1$ and $\eta$ negative. Further simulations
and renormalization group studies are required for solving this
problem.
In summary we performed an analysis of interacting bosons in 2D
within zero temperature lattice models, using and comparing both
mean-field theory and exact Quantum Monte Carlo simulations. We obtained
the full phase diagram for on-site interaction as well as for
nearest-neighbor interactions, in which case our simulations establish the
existence of a supersolid phase in which a CDW and superfluidity coexist.
{\bf Acknowledgments:} We thank G.~T. Zimanyi for motivation and
discussions, and J.~K. Freericks and H. Monien for discussions and their data
as plotted in Fig.~\ref{mottfig}. The work was partially supported by
'Sonderforschungsbereich 195' of the DFG and by the Swiss Nationalfonds (AvO).
\begin{table}
$$
\begin{array}{||c|c|c|c|c||c|c|c|c|c||}
\hline\hline
n_{0} & z & K^{\ast} & J^{\ast}/U_0 & \nu &
n_{0} & z & K^{\ast} & J^{\ast}/U_0 & \nu \\ \hline\hline
0.00 & 1 & 0.886\pm0.003 & 0.152 & 0.69\pm0.06 &
0.25 & 2 & 0.64\pm0.01 & 0.098 & 0.5\pm0.1 \\ \hline
0.05 & 2 & 0.85\pm0.01 & 0.144 & 0.5\pm0.1 &
0.30 & 2 & 0.56\pm0.01 & 0.081 & `` \\ \hline
0.10 & 2 & 0.82\pm0.02 & 0.137 & `` &
0.35 & 2 & 0.47\pm0.01 & 0.062 & `` \\ \hline
0.15 & 2 & 0.77\pm0.02 & 0.126 & `` &
0.40 & 2 & 0.37\pm0.01 & 0.042 & `` \\ \hline
0.20 & 2 & 0.71\pm0.01 & 0.113 & `` &
0.45 & 2 & 0.26\pm0.01 & 0.023 & `` \\ \hline
\end{array}
$$
\caption{Critical couplings and exponents for on-site interactions}
\label{table-ons}
\end{table}
\begin{table}
$$
\begin{array}{||c||c|c|c|c||c|c|c|c|c||}
\hline\hline
\multicolumn{10}{||c||}{\mbox{the transition for}}
\\ \hline
\multicolumn{1}{||c||}{}&\multicolumn{4}{|c||}{\mbox{off-diagonal LRO}}&
\multicolumn{5}{|c||}{\mbox{diagonal LRO}}
\\ \hline
n_{0} & z & K^{\ast} & J^{\ast}/U_0 & \nu &
z & K^{\ast} & J^{\ast}/U_0 & \nu & \beta \\ \hline\hline
0.5 & 1 & 0.775\pm0.005 & 0.127 & 0.65\pm0.08 &
1 & \; 0.837\pm0.005& 0.141 & \;0.55\pm0.05 &
0.21\pm 0.04 \\ \hline
0.4 & 2 & 0.645\pm0.008 & 0.099 & 0.44\pm0.08 &
2 & 0.749\pm0.006 & 0.122 & 0.5\pm0.1 & \; 0.25
\pm 0.10 \\ \hline
0.4 & & & & &
1 & 0.747\pm0.007 & 0.121 & 0.55\pm0.06 & \; 0.21
\pm 0.08 \\ \hline \hline
0.2 & 2 & 0.446\pm0.005 & 0.057 & 0.5\pm0.1 &
\multicolumn{5}{|c||}{\mbox{for comparison:}} \\ \hline
0.1 & 2 & 0.707\pm0.007 & 0.112 &\; 0.49\pm0.11 &
\multicolumn{3}{|c|}{\;\;\;\; \mbox{mean-field:}} & 1/2 & 1/2 \\ \hline
0.0 & 1 & \; 0.843\pm0.005& 0.142 & 0.61\pm0.08 &
\multicolumn{3}{|c|}{\;\;\;\;\;\;\;\;\;\; \mbox{3D XY: }} & 2/3 & 1/3 \\
\hline\hline
\end{array}
$$
\caption{Critical couplings and exponents for the different
transitions with nearest-neighbor interactions $U_1/U_0=0.2$}
\label{table-nn}
\end{table}
|
\section{Introduction}
\indent
The minimal supersymmetric Standard Model (MSSM) with the particle
content of the Standard Model (SM), including their supersymmetric (SUSY)
partners, conserves the discrete quantum number $R$, known as $R$ parity.
Under this symmetry, the SM particles are even whereas their superpartners
are odd. The quantum number of $R$ parity may conveniently be expressed
as~\cite{Rsusy,SW,HS,EGJRV}.
\begin{equation} \label{m1}
R=(-1)^{3B+L+2S} \, ,
\end{equation}
where $B$, $L$, and $S$ are the baryon number, the lepton number, and the
spin of a particle, respectively. Evidently, nonconservation of $R$ parity
results, in general, in baryon ($\not\!\! B$) and lepton ($\not\!\! L$)
number violating terms. However, the usually considered models retain only
the lepton-number-violating terms in order to preserve the stability of the
proton. In the presence of both, $\not\!\! L$ and $\not\!\! B$ terms, the
proton decay might render the model phenomenologically not
viable.
The explicit breaking of $R$ parity in a SUSY model means also that we
introduce additional coupling constants into the Lagrangian, but not additional
fields, {\em i.e.}\ the field content of the model is the same as in the MSSM.
This property of the model has some interesting consequences. One important
implication is that the explicit lepton-number violation allows, in
principle, for a mixing of the left-handed neutrinos with the gauginos
and higgsinos. As a consequence, such a mixing, if present,
will generate neutrino masses. This mechanism to render neutrinos massive
is quite distinctive from the standard options, such as the introduction of
right-handed neutrinos or the extension of the Higgs sector by adding an
exotic Higgs triplet. However, to make the afore-mentioned mixing possible,
one should ensure that the sneutrinos acquire vacuum expectation values
(vev's).
By a careful examination of the scalar potential of the $R$-parity broken
SUSY model, one can show that the appearance of non-vanishing vev's
for the scalar neutrinos is not always a direct consequence of the
theory. In fact, if a term of the form
$\varepsilon_i \varepsilon_{ab}\hat{L}^a_i \hat{H}^b_1$ is neglected
in the superpotential of the model, where $\hat{L}_i$ are the chiral lepton
fields and $\hat{H}_1$ is the super-Higgs fields with hypercharge $Y=1$, then
insistence of non-zero vev's of the sneutrino fields can lead, in some cases,
to a fine-tuning relation among the original parameters of the scalar
potential. For instance, this would happen if we assumed the scalar mass
terms $m_{ij}L^{\dagger}_iL_j +$ H.c.\ to be diagonal in $i$, $j$.
On the other hand,
the retention of the $\varepsilon_i$ terms leads unavoidably to non-zero vev's
of the sneutrinos ($w_i$). It is also easy to see that in the presence of
soft-SUSY breaking parameters, the $\varepsilon_i$ terms cannot be rotated
away by a unitary transformation. This is consistent with the observation
that such a term will be generated radiatively, even if one chooses
$\varepsilon_i=0$ at the Planck mass scale~\cite{HS}.
Therefore, one could argue that models which break the $R$ parity explicitly,
in which the $\varepsilon_i$ terms are neglected, may be considered to be not
general and hence incomplete. This fact is expected to influence the analytic
expressions of the mass and mixing matrices, especially when new parameters
enter the theory. Indeed, we will show how the $\varepsilon_i$ terms
change the phenomenological predictions of the model under consideration.
Models with an explicit mixing term between $\hat{L}_i$ and $\hat{H}_1$ have
been considered in the past~\cite{HS}. More recently, such models have been
re-examined by paying special attention to CP violation in the scalar potential
as well as to neutrino masses \cite{JN,JNCP}.
In Refs.~\cite{HS,JN,JNCP}, it was assumed that all
individual lepton numbers, $L_i$, are broken. It has been known for some time
that lepton-number-violating interaction can erase the existing baryon
asymmetry in the universe (BAU) if $\not\!\! L$ interactions are
in thermal equilibrium with the $B+L$-violating sphalerons.
The most stringent constraints on $w_i$ and $\varepsilon_i$ arise from such
considerations. However, it has been realized~\cite{DR}
that one can evade the erasure
of the BAU if we demand that one individual lepton number is conserved.
No constraints on lepton-number-violating couplings will then follow. The
most conservative constraints on lepton-number-violating couplings can then
be inferred from an analysis of the absence of possible
new-physics phenomena in various experiments.
In $R$-parity broken models, there are in general two sources that can
produce lepton-number- and lepton-flavour-violating interactions. Those that
are
induced by the $\varepsilon_i$ and $w_i$ parameters and give rise, {\em
e.g.}, to tree level off-diagonal $Z$-boson decays, and those that are
proportional to the so-called $\lambda$ and $\lambda'$ couplings in
the superpotential. The literature pertinent to constraints on the
$\lambda$ and $\lambda'$ couplings is immanent~\cite{hinch,BGH,GBDC,HKK,roh}.
Here, we will study effects of lepton-number and/or lepton-flavour
violation on the ordinary known matter, which can be induced by $W$-
and/or $Z$-boson interactions in an $R$-parity broken model. Therefore,
one may consider our study complementary to investigations of limits
on the $\lambda$ and $\lambda'$ couplings mentioned above. In an
$\not\!\! R$ SUSY model, direct limits on $w_i$ are mostly obtained
from upper bounds on light neutrino masses, {\em e.g.}\
using the laboratory bound $m_{\nu_{\tau}} < 31$~MeV~\cite{PDG94}.
However, we find that better limits may be deduced
by the nonobservation of non-SM processes, such as the decay of a muon
into three electrons. Moreover, cosmological and astrophysical
implications of a massive light neutrino for our model will be
discussed in Section~5.7.
The paper is organized as follows. In Section 2, we will give a detailed
discussion of the $R$-parity broken SUSY model where the emphasis
will be put on the scalar potential of the model.
Further analytic results on this topic are relegated in Appendix A.
Section 3 treats the mixing between the neutralinos and
neutrinos as well as that between the charginos and charged leptons.
In Section 4, we derive the
$W$- and $Z$-boson interaction Lagrangians in the seesaw approximation.
In Section 5,
we analyze a number of low-energy processes that can be induced by the
non-SM couplings present in $\not\!\! R$ SUSY models. We then discuss
new constraints that may be derived by laboratory experiments together
with constraints coming from cosmology and astrophysics. In addition,
we investigate the possibility that our $R$-parity broken SUSY model
can explain the KARMEN anomaly. Our numerical results are presented in
Section 6. We draw our conclusions in Section 7.
\setcounter{equation}{0}
\section{The $R$-parity violating SUSY model}
\indent
In this section, we set up our definition and notation and outline
in some more detail the scalar potential of the MSSM with explicit
$R$-parity breaking terms. Our main concern will be to argue that
the retention of a special term in the superpotential, which is usually
not taken into account, leads {\it naturally}, {\em i.e.}\ without
any fine tuning problems, to non-zero vev's of the sneutrinos.
As a consequence, neutrinos acquire masses through mixing with
gauginos and higgsinos and the standard $W$/$Z$-boson interactions with
fermions get modified, {\em e.g.}\ one gets tree-level off-diagonal $Z$
decays.
The novelty in our approach is the afore-mentioned term in the superpotential
whose coupling strength $\varepsilon_i$ enters the $W$/$Z$ interaction
Lagrangians.
\subsection{Superpotential}
\indent
We write the full superpotential $W$ as consisting of an $R$-parity
conserving part ($W_0$) and $R$-parity violating term
($W_{\not \! R}$), {\em i.e.}\
\begin{equation} \label{m2}
W\ =\ W_0+W_{\not \! R}\, .
\end{equation}
Let then $\hat{L}_i$ ($\hat{E}_i^C$) and
$\hat{Q}_i$ ($\hat{U}_i^C $,$ \hat{D}_i^C$) denote the lepton and
quark doublets superfields (lepton and quarks $SU(2)$ singlets) with
generation index $i$, respectively and let $\hat{H}_{1,\; 2}$ be
the super-Higgs fields. With the usual $U(1)_Y$ quantum
number assignment, $Y(\hat{L}_i)=-1$,
$Y(\hat{E}_i^C)=2$, $Y(\hat{Q}_i)=1/3$, $Y(\hat{D}_i^C)=2/3$,
$Y(\hat{U}_i^C)=-4/3$, $Y(\hat{H}_1)=-1$, $Y(\hat{H}_2)=1$,
the standard form for $W_0$ is
\begin{equation} \label{m3}
W_0\ =\ \varepsilon_{ab}\left [h_{ij}\hat{L}_i^a \hat{H}_1^b
\hat{E}_j^C + h'_{ij}\hat{Q}_i^a \hat{H}_1^b \hat{D}_j^C + h''_{ij}
\hat{Q}_i^a \hat{H}_2^b \hat{U}_j^C + \mu \hat{H}_1^a \hat{H}_2^b \right]\, ,
\end{equation}
where $a$, $b$ are $SU(2)$ group indices.
The explicit breaking of $R$ parity can be introduced through
$W_{\not \! R}$, which in its most general form is given by~\cite{SW,HS}
\begin{equation} \label{m4}
W_{\not \! R}\ =\ \varepsilon_{ab}\left( \lambda_{ijk}\hat{L}_i^a
\hat{L}_j^b
\hat{E}_k^C + \lambda'_{ijk}\hat{L}_i^a \hat{Q}_j^b \hat{D}_k^C
+ \varepsilon_i \hat{L}_i^a \hat{H}_2^b \right) + \lambda''_{ijk}
\hat{U}_i^C \hat{D}_j^C \hat{D}_k^C \, .
\end{equation}
Unlike the MSSM, the $R$-parity broken SUSY model allows for explicitly
broken lepton ($\not \! \! L$) and baryon ($\not \! \! B$) number
interactions. To be more precise, the terms in Eq.~(\ref{m3}) proportional to
$\lambda_{ijk}=-\lambda_{jik}$, $\lambda'_{ijk}$ and $\varepsilon_i$ violate
lepton number, whereas the baryon number is explicitly broken by the
$\lambda''_{ijk}$-term ($\lambda''_{ijk}=-\lambda''_{ikj}$).
The presence of both $\not\!\! B$- and $\not\!\! L$-type of terms in
the Lagrangian leads to unsuppressed proton decay.
Therefore, at the most, we can retain either the
$L$-violating or the $B$-violating terms in (\ref{m3}).
To account for this fact, hereafter we will set $\lambda''_{ijk}=0$.
Let us now comment on the term
$\varepsilon_{ab} \varepsilon_i \hat{L}_i^a \hat{H}_2^b$ which is usually
not taken into account in $W_{\not \! R}$ by using the argument of
field redefinitions to rotate away such bilinears. It has been discussed
in some detail in~\cite{JN,JNCP}
that this argument may not be valid once
we add to the Lagrangian soft-SUSY breaking terms.
Indeed, we are unable to absorb the $\varepsilon_i$ terms by using an
orthogonal
transformation of the $\hat{H}_1$ and $\hat{L}_i$ fields. The omission of
such terms is therefore not justified. We note here that exactly these terms
are in principle responsible for a non-zero vev's of the sneutrinos.
The $\varepsilon_i$ terms force the
vev's of the sneutrino fields to assume non-zero values. We will
return to this point while discussing the Higgs potential.
The $\lambda$-terms in $W_{\not \! R}$ lead to the
interaction Lagrangian
\begin{equation} \label{Llambda}
{\cal L}_{\lambda}\ =\ \lambda_{ijk}\biggl[\tilde{\nu}_{iL}
\bar{e}_{jR}e_{jL}+
\tilde{e}_{jL}\bar{e}_{kR}\nu_{iL}+\tilde{e}^*_{kR}
\bar{\nu}_{iL}^Ce_{jL}
-(i \leftrightarrow j)\biggr]\ +\ \mbox{H.c.}\, ,
\end{equation}
where $e_i$ denote charged leptons enumerated by generation index
$i$ and tilded symbols denote as usual the superpartners. Correspondingly,
the $\lambda'$ interaction Lagrangian reads
\begin{eqnarray} \label{Llambda'}
{\cal L}_{\lambda'}&=&\lambda'_{ijk}\biggl( \tilde{\nu}_{iL}
\bar{d}_{kR}d_{iL}
+\tilde{d}_{jL}
\bar{d}_{kR}\nu_{iL}+\tilde{d}^*_{kR}\bar{\nu}_{iL}^Cd_{jL}
- \tilde{e}_{iL}\bar{d}_{kR}u_{jL}
- \tilde{u}_{jL}\bar{d}_{kR}e_{iL}\nonumber\\
&& -\tilde{d}^*_{kR}\bar{e}_{iL}^Cu_{jL}
\biggr)\ +\ \mbox{H.c.}\, ,
\end{eqnarray}
where $d_i$ ($u_i$) are down-type (up-type) quarks.
Since the main concern of the present paper will be to constrain the
lepton-number-violating couplings,
it is worth discussing briefly constraints on
the $L$-violating couplings which come from baryogenesis. It has been known
for some time that non-perturbative anomalous $B$-violating interactions
of the electroweak theory can
wash out the baryon asymmetry generated initially at the GUT scale. This gives
rise to severe constraints on lepton-number-violating couplings, such as
the trilinear couplings $\lambda_{ijk}$ and $\lambda'_{ijk}$.
In this context, it has been noticed that to evade such limits is sufficient
to have one individual lepton number, $L_i$, conserved~\cite{DR}.
The latter will be assumed throughout this work. In particular, we have
\begin{eqnarray} \label{m7}
&&\lambda_{ijk}=0,\quad \mbox{for} \quad i\neq j \neq k\, , \nonumber \\
&&\lambda_{iki} \neq 0,\quad \mbox{for}\quad \not\! \! L_k \,\,\, L_i\, ,
\nonumber \\
&&\lambda'_{ijk} \neq 0,\quad \mbox{for }\quad \not\! \! L_i\, ,
\end{eqnarray}
where~$\not\! \! L_i$ ($L_i$) indicates which lepton number is violated
(conserved). Choosing $L_i$ as the conserved lepton number, we then have
four independent $\lambda$ couplings. For instance, conserving $L_e$, we
are left with non-zero $\lambda_{121}$, $\lambda_{131}$, $\lambda_{232}$, and
$\lambda_{323}$. This conservation of a separate lepton number,
{\em i.e.}\ $\Delta L_e=0$,
has profound consequences for exotic lepton-number-violating processes, which
can, in principle, occur at the tree level in an $R$-parity broken SUSY model.
If one $L_i$ is strictly conserved, then no tree-level $\lambda$-dependent
interaction can contribute, {\em e.g.}, to the process $\mu \to eee$.
The latter would proceed via a sneutrino mediated diagram only if {\em all}
individual lepton numbers were broken. If $\Delta L_i=0$ for only one
lepton number $L_i$, the bound on the $\lambda$ couplings derived
via $\mu \to eee$ in \cite{hinch} does not further apply. The reason is
that the process $\mu \to eee$ is still possible at the tree level,
however through a diagram containing off-diagonal $Z$ couplings
(see discussion in Section~5.1).
In the subsequent section, we will also address the question whether
having $\Delta L_i=0$ for one lepton number $L_i$ at the level of Lagrangian,
this particular symmetry gets broken spontaneously through a vev of the
sneutrino field, $w_i$, with the very same flavour index $i$. Indeed, we
will argue that this is not the case, when the couplings $\varepsilon_i$
are taken into account.
Laboratory constraints on $\lambda$ and $\lambda'$ couplings have been put
in~\cite{BGH} and~\cite{GBDC,HKK}. The detectability of possible direct
$R$-parity-violating signals through the $\lambda$ and $\lambda'$
couplings at the CERN Large Electron Positron (LEP) collider, planned
to operate at 200-GeV centre of mass energies (LEP-2), has been discussed
in~\cite{roh}.
We note here that there are low-energy processes to which both type of
$\not\!\! L$ interactions ---those induced by the trilinear $\not\!\! R$
$\lambda$- and $\lambda'$-dependent couplings and those emanating from
non-zero $\varepsilon_i$ and $w_i$ parameters--- will contribute. We will
see that a combined analysis is not necessary, {\em i.e.}\
deriving limits first on $\varepsilon_i$ and $w_i$ from, say,
$\mu \to eee$, and then applying the so obtained results to put constraints
on the $\lambda$ and $\lambda'$ couplings.
\subsection{The scalar potential}
\indent
With the superpotential given, it is straightforward to construct the scalar
potential of an $R$-parity broken SUSY model~\cite{Rsusy}.
Using the convention that
symbols without a hat, say $A$, are the spin-zero content of a chiral
superfield
$\hat{A}$, we first write down the relevant soft-SUSY breaking terms
\begin{eqnarray} \label{m8}
V_{soft}&=&m_1^2 H_1^{\dagger}H_1 + m_2^2 H_2^{\dagger} H_2
+(m_{L_{ij}}^2L_i^{\dagger}L_j\ +\ \mbox{H.c.})\nonumber \\
&&-(m_{12}^2\varepsilon_{ab}H_1^a H_2^b\ +\ \mbox{H.c.})
+ (\kappa'_i \varepsilon_{ab}
H_2^a L_i^b\ +\ \mbox{H.c.}) \nonumber \\
&&+ (\mu^2_{+ij}E_i^* E_j\ +\ \mbox{H.c.})+[\mu'_{ij}(H_1^{\dagger}L_i)E_j\ +\
\mbox{H.c.}] \nonumber \\
&&+ (\kappa'_{ijk}\varepsilon_{ab}L^a_j L^b_j E_k\ +\ \mbox{H.c.})\, .
\end{eqnarray}
In Eq.~(\ref{m8}), $E_i$ are the positively charged scalar singlet fields.
The full scalar potential (without squarks) can be written as the sum of five
terms,
\begin{equation} \label{m9}
V_{Scalar}\ =\ V^{2H} + V^L + V^{\not L} + V_+^L + V_+^{\not L}\, ,
\end{equation}
where $V^{2H}$ is the usual Higgs potential of the MSSM, $V^L$ and $V
^{\not L}$
contain the slepton doublets (both $L$-conserving and $L$-violating parts, the
latter connected with $R$-parity breaking couplings), and finally
$V_+^L$ and $V_+^{\not L}$
refers to the part of the potential which contains
the charged scalars $E_i$. The standard MSSM potential, $V^{2H}$, reads
\begin{eqnarray} \label{m10}
V^{2H}&=&\mu_1^2\phi_1^{\dagger}\phi_1 +
\mu_2^2 \phi_2^{\dagger} \phi_2
+{1 \over 2}\lambda_1(\phi_1^{\dagger}\phi_1)^2 +
{1 \over 2}\lambda_1(\phi_1^{\dagger}
\phi_2)^2
\nonumber \\
&&+\lambda_3 (\phi_1^{\dagger}\phi_1)(\phi_2^{\dagger}\phi_2)
-(\lambda_3 + \lambda_1)(\phi_1^{\dagger}\phi_2)(\phi_2^{\dagger}\phi_1)
\nonumber \\
&&+ \lambda_6 (\phi_1^{\dagger}\phi_2)
+ \lambda_6^*(\phi_2^{\dagger}\phi_1)\, ,
\end{eqnarray}
where we have used the notation
$\varphi_i \equiv L_i$, $\phi_2
\equiv H_2$, and $\phi_1 \equiv -i\tau_2 H_1^*$ [$\tau_2$ being
the Pauli matrix, ($i\tau_2)_{ab}=\varepsilon_{ab}$]. The new parameters in
Eq.~(\ref{m10}), which depend on the couplings entering the superpotential,
are the soft-SUSY breaking parameters given in (\ref{m8}), as well as the
$SU(2)_L$ coupling constant $g$ and the corresponding $U(1)_Y$ coupling
constant $g'$. These parameters are related as follows:
\begin{eqnarray} \label{m11}
&&\mu_1^2=m_1^2+\vert \mu \vert^2,\; \; \; \mu_2^2=m_2^2 +\vert
\mu \vert^2 + \varepsilon_i \varepsilon_i^*\, , \nonumber \\
&&\lambda_1={1 \over 4}(g^2 + g'^2), \; \; \; \lambda_3={1 \over
2}g^2- \lambda_1,\; \; \; \lambda_6=-m_{12}^2\, .
\end{eqnarray}
The lepton-number-conserving scalar potential containing the slepton doublet
fields $\varphi_i$ reads
\begin{eqnarray} \label{m12}
V^L &=& (\mu^2_{L_{ij}}\varphi_i^{\dagger}\varphi_j\ +\ \mbox{H.c.})
+ {1 \over 2}\lambda_1 (\sum_i \varphi_i^{\dagger}\varphi_i)^2 +
\lambda_1(\phi_1^{\dagger}\phi_1)(\varphi_i^{\dagger}\varphi_i)
\nonumber \\
&&- \lambda_1(\phi_2^{\dagger}\phi_2)(\varphi_i^{\dagger}\varphi_i)
+ (\lambda_3 + \lambda_1)(\phi_2^{\dagger}\varphi_i)(\varphi_i^{\dagger}
\phi_2) \nonumber \\
&&+ \Big[\kappa_{jk} -(\lambda_3 + \lambda_1)\delta_{jk}\Big](\phi_1^{\dagger}
\varphi_k)(\varphi_j^{\dagger}\phi_1) + \kappa_{nmij}(\varphi_i^T \tau_2
\varphi_j)(\varphi_n^T \tau_2 \varphi_m)^{\dagger}\, .
\end{eqnarray}
Here, we have used the definitions
\begin{eqnarray} \label{m13}
&&\mu_{L_{ij}}^2=m^2_{L_{ij}}+\varepsilon^*_i \varepsilon_j, \,\,\,
\kappa_{jk}=\kappa^*_{kj}\equiv h^*_{ji}h_{ki}, \nonumber \\
&&\kappa_{nmij}=-\kappa_{mnij}=-\kappa_{nmji}=\kappa^*_{ijnm} \equiv
\lambda^*_{nmk}\lambda_{ijk}\, .
\end{eqnarray}
$V^L$ given in Eq.~(\ref{m12}) may be contrasted with the corresponding
potential of the MSSM. In the MSSM, the term
proportional $\kappa_{nmij}$ is absent and
$\mu^2_{L_{ij}}$ should be replaced by the diagonal
mass parameters $\mu^2_{L_i}=m^2_{L_i}+\vert \varepsilon_i \vert^2$.
Finally, the lepton-number-violating part of the potential is given by
\begin{eqnarray} \label{m14}
V^{\not L}&=&[i\kappa_i (\phi_1^T \tau_2 \varphi_i)\ +\ \mbox{H.c.}]
+ [i\kappa'_i (\phi_2^T \tau_2 \varphi_i)\ +\ \mbox{H.c.}] \nonumber \\
&&- i\kappa_{nmj}(\phi_1^{\dagger}\varphi_j)(\varphi_n^T \tau_2
\varphi_m)^{\dagger}\, ,
\end{eqnarray}
with
\begin{equation} \label{m15}
\kappa_i \equiv \mu^* \varepsilon_i, \,\,\, \kappa_{nmj} \equiv
\lambda^*_{nmk}h_{jk}\, ,
\end{equation}
and $\kappa'_j$ is a soft-SUSY breaking parameter from Eq.~(\ref{m8}). The
parts
$V^{2H}$, $V^L$ and $V^{\not L}$
are sufficient to determine the minimization conditions of
the potential. The explicit form of the remaining
contributions to $V_{Scalar}$, $V_+^L$ and $V_+^{\not L}$, is given in
Appendix A. At the minimum, the fields take the values
\begin{equation} \label{m16}
\langle \phi_1 \rangle =\left(\begin{array}{c}
0 \\
v_1 \end{array}\right), \,\,\,
\langle \phi_2 \rangle =\left(\begin{array}{c}
0 \\
v_2 \end{array}\right), \,\,\,
\langle \varphi_i \rangle =\left(\begin{array}{c}
w_i \\
0 \end{array}\right),
\end{equation}
and the minimization conditions are found to be
\begin{eqnarray} \label{m17}
\mu_1^2 v^*_1 +\lambda_1 v_1^*(\vert v_1 \vert^2 -\vert v_2 \vert^2 +
\sum_k \vert w_k \vert^2)+\lambda_6^* v^*_2 -\sum_k \kappa_k w_k &=& 0\, ,
\nonumber \\
\mu_2^2 v^*_2 -\lambda_1 v_2^*(\vert v_1 \vert^2 -\vert v_2 \vert^2 +
\sum_k \vert w_k \vert^2)+\lambda_6 v^*_1 -\sum_k \kappa'_k w_k &=& 0\, ,
\nonumber \\
\sum_j \mu^2_{L_{ij}}w^*_j +\lambda_1 w_i^*(\vert v_1 \vert^2 -\vert v_2
\vert^2 +\sum_k \vert w_k \vert^2)- \kappa_i v_1-\kappa'_i v_2 &=& 0\, ,
\end{eqnarray}
where the last equation is valid for every generation index $i$. It is clear
from Eq.~(\ref{m17}) that $w_i \neq 0$ unless $\kappa_i=\kappa'_i=0$.
To further elucidate this point,
let us first consider the case of $\kappa_i$ and
$\kappa'_i$ not identical to zero ({\em i.e.}\ the case of
the $R$-parity broken SUSY model with $\varepsilon_i \neq 0$),
and assume $w_i =0$ for some generation index $i$.
{}From Eq.~(\ref{m17}), we then get a `fine-tuning' relation among the
original parameters of the potential (no summation convention)
\begin{equation} \label{m18}
(\mu^2_1 + \mu^2_2)\kappa_i^{*}\kappa'^*_i=\lambda^*_6 \kappa_i^{*2}+
\lambda_6 \kappa'^{*2}_i\, .
\end{equation}
Analogous to our $\not\!\!R$ model is the limiting case of the MSSM,
without explicit $R$-parity breaking, if one sets $\kappa_i=\kappa'_i=0$
and replaces $\mu^2_{L_{ij}}$ by the diagonal coupling $\mu^2_{L_i}$.
Again, the attempt to maintain $w_i \neq 0$ yields
the `fine-tuning' relation (see also \cite{comel}).
\begin{equation} \label{m19}
(\mu_1^2 - \mu^2_{L_i})(\mu^2_2 + \mu^2_{L_i})=\vert \lambda_6 \vert^2\, .
\end{equation}
This demonstrates nicely how, unlike the MSSM, the $R$-parity-breaking
couplings $\varepsilon_i$ from Eq.~(\ref{m3}) naturally lead to non-zero
vev's for the scalar neutrinos. Had we neglected the bilinear term
$\varepsilon_{ab} \varepsilon_i \hat{L}_i^a \hat{H}_2^b$ in Eq.~(\ref{m3}), as
done
usually, then there would have been no compelling reason to acquire non-zero
vev's for the scalar neutrino fields even in the $R$-parity broken case.
Moreover, if we conserve an individual lepton number, say $L_i$, in
the Lagrangian/superpotential of the $R$-parity broken SUSY model ---among
other couplings, this implies that $\varepsilon_i=0$---, this symmetry will
not break spontaneously; so, we are free to choose the corresponding
$\varphi_i$ to have a vanishing vev.
This follows again from Eqs.\ (\ref{m18}) and (\ref{m19}).
The spontaneous breaking of CP violation in the case of $V_{Scalar}$,
Eq.~(\ref{m9}), has been discussed in~\cite{JNCP}. We supplement these
considerations on CP properties by giving below all conditions necessary to
restore CP conservation in $V^{2H}+V^L+V^{\not L}$.
Denoting $\eta_{{}_1}$, $\eta_{{}_2}$
and $\eta_{{}_{L_i}}$ the CP phases of the fields
$\phi_1$, $\phi_2$, and $\varphi_i$, respectively, we find from the
requirement of CP conservation that (no summation convention below)
\begin{eqnarray} \label{m20}
&&\lambda_6^* \eta_{{}_1} \eta_{{}_2}^*=\lambda_6\, , \nonumber \\
&&\kappa_{jk}=\kappa^*_{jk}\eta_{{}_{L_i}}\eta^*_{{}_{L_k}}\, , \nonumber \\
&&\kappa_{nmjk}=\kappa^*_{nmjk}\eta_{{}_{L_n}}
\eta_{{}_{L_m}}\eta^*_{{}_{L_i}}\eta^*_{{}_{L_j}}\, ,
\nonumber \\
&&\kappa^*_i \eta^*_{{}_1} \eta^*_{{}_{L_i}}=\kappa_i\, , \nonumber \\
&&\kappa'^*_j \eta^*_{{}_2} \eta^*_{{}_{L_j}}=\kappa'_j\, , \nonumber \\
&&\kappa^*_{nmj}\eta_{{}_1} \eta^*_{{}_{L_j}}
\eta_{{}_{L_n}}\eta_{{}_{L_m}}=\kappa_{nmj}\, .
\end{eqnarray}
{}From Eq.\ (\ref{m20}), one can derive over twenty conditions for
CP conservation which in contrast
to (\ref{m20}) do not involve the CP phases $\eta$'s. However, it is obvious
that not all such conditions are independent. For instance, a set of
independent conditions that follows from (\ref{m20}) is given by
\begin{eqnarray} \label{m21}
&&\Im m(\lambda_6 \kappa_i \kappa'^*_j \kappa_{ij})=0\, , \nonumber \\
&&\Im m(\kappa_i \kappa^*_j \kappa_{ij})=0\, , \nonumber \\
&&\Im m(\kappa'_i \kappa'^*_j \kappa_{ij})=0\, , \nonumber \\
&&\Im m(\kappa^*_{ni}\kappa^*_{mj}\kappa_{nmij})=0\, , \nonumber \\
&&\Im m(\kappa^*_{n}\kappa_{mj}\kappa^*_{nmj})=0\, ,
\end{eqnarray}
where again no summation convention has been used.
After spontaneous symmetry breaking, the first equation in the
set (\ref{m21}) for $i=j$, {\em i.e.}\ $\Im m(\lambda_6 \kappa_i
\kappa^{'*}_i)=0$, translates into the following three equivalent
conditions:
\begin{eqnarray} \label{m22}
&&\Im m(\kappa_i w_i v_1)=0\, , \nonumber \\
&&\Im m(\kappa'_i w_i v_2)=0\, , \nonumber \\
&&\Im m(\lambda_6 v_1^* v_2)=0\, .
\end{eqnarray}
For the case at hand, it is, however, possible to have
complex vev's such that CP gets broken spontaneously.
For further details on this issue, the reader is referred to~\cite{JNCP}.
Some remarks on the vev's of sneutrinos, $w_i$, are in order.
One should notice that through the kinetic term
\begin{displaymath}
\sum_i \left(D_{\mu}{\cal S}_i
\right)^{\dagger}\left(D^{\mu}{\cal S}_i\right)\, ,
\end{displaymath}
where
$D_{\mu}$ is the covariant derivative and ${\cal S}_i$ are the scalar fields in
the theory, the sneutrino vev's contribute to the gauge bosons masses. In this
way, the SM vev is obtained by
\begin{equation} \label{mex2}
v \equiv \sqrt{v_1^2 + v_2^2 + \sum_i w_i^2}\ =\ \frac{2M_W}{g}\, .
\end{equation}
As a consequence, $w_i$ and the angle $\beta$ defined by
\begin{equation} \label{mex3}
\tan \beta = { v_1 \over v_2}\, ,
\end{equation}
for real $v_i$, may be regarded as free parameters of the theory,
while $v_i$ are not free any longer, but determined by
\begin{eqnarray} \label{mex4}
&&v_1 = \sin \beta \sqrt{v^2 -\sum_i w^2_i}\, , \nonumber \\
&&v_2 = \cos \beta \sqrt{v^2 -\sum_i w^2_i}\, .
\end{eqnarray}
Evidently, the vev's of the scalar neutrinos, $w_i$, cannot have arbitrarily
large values, but they are bounded from above,
as can be readily seen from Eqs.~(\ref{mex2}) and (\ref{mex4}).
\setcounter{equation}{0}
\section{Mass matrices}
\indent
In this section, we will present the mass matrices of the neutralino/neutrino
as well as those
of the chargino/charged lepton states. Since the lepton number is
explicitly broken, the neutrinos will mix with the neutralinos to give the
neutrinos mass. A natural seesaw mechanism emerges, in which
$\mu$, $v_1$, $v_2$, and the gaugino mass parameters $M$ and $M'$
act as the heavy scales, and the lepton-number-breaking couplings
$\varepsilon_i$
together with the sneutrino vev's $w_i$ constitute the light Dirac components
of the seesaw matrix.
It will turn out that only one neutrino becomes
massive through this mechanism at the tree level.
These considerations are relevant for putting limits on the
lepton-number-breaking parameters. In particular, one can already infer
constraints on those parameters from the $\tau$-neutrino mass.
Furthermore, the neutralino--neutrino or chargino--charged lepton mixing
will enter the interaction Lagrangians of $W$ and $Z$, giving rise
to non-SM processes, through which the new parameters can also be constrained.
\subsection{Neutralino--neutrino mixing}
\indent
In general, there are two mechanisms that can give rise to neutrino masses
in the Born approximation. For example, one possibility is to give masses
to the left-handed neutrinos through the vev of an exotic Higgs field which
transform under $SU(2)_L$ as a triplet. The other mechanism requires,
in general, the mixing of the left-handed neutrinos with other neutral
fields of the theory. The latter are usually taken to be the
right-handed neutrinos, introducing hereby additional fields in the theory. In
our minimal $R$-parity broken SUSY model, in which right-handed neutrinos
are absent, the r\^ole of the new neutral fields required for the
afore-mentioned mixing will be assumed by the gauginos and higgsinos.
In two component notation, let $\Psi'$ denote the column vector of neutrinos
and
neutralinos
\begin{equation} \label{m23}
\Psi^{'T}_0=(\psi^1_{L_1},\,\, \psi^1_{L_2},\,\, \psi^1_{L_3},\,\,
-i\lambda',\,\, -i\lambda_3,\,\, \psi^1_{H_1},\,\, \psi^2_{H_2})\, ,
\end{equation}
where $\psi^1_{L_i}$ are the neutrino fields ---the upper index indicates
the component of the doublet---, $-i\lambda'$ and $-i\lambda_3$ are the
unmixed photino and gaugino states, respectively, and the last two entries
refer to the two higgsino fields. In the Weyl basis, the Lagrangian
describing the neutralino/neutrino masses is then given by
\begin{equation} \label{m24}
{\cal L}^{\chi^0}_{mass}=-{1 \over 2}\Psi_0^{'T}{\cal M}_0 \Psi'_0\ +\
\mbox{H.c.}\, ,
\end{equation}
where the mass matrix has the general seesaw-type structure
\begin{equation} \label{m25}
{\cal M}_0=\left(\begin{array}{cc}
0 & m \\
m^T & M_4 \end{array}\right) .
\end{equation}
Here, the sub-matrix $m$ is the following $3 \times 4$ dimensional matrix:
\begin{equation} \label{m26}
m=\left(\begin{array}{cccc}
-{1 \over 2}g' w_e & {1 \over 2}g w_e & 0 & -\varepsilon_e \\
-{1 \over 2}g' w_\mu & {1 \over 2}g w_\mu & 0 & -\varepsilon_\mu \\
-{1 \over 2}g' w_\tau & {1 \over 2}g w_\tau & 0 & -\varepsilon_\tau
\end{array}\right) .
\end{equation}
In Eq.~(\ref{m25}), $M_4$ is the usual $4 \times 4$ dimensional
neutralino mass matrix of the MSSM, which has the form
\begin{equation} \label{m27}
M_4=\left(\begin{array}{cccc}
cM & 0 & -{1 \over 2}g'v_1 & {1 \over 2}g'v_2 \\
0 & M & {1 \over 2}gv_1 & -{1 \over 2}gv_2 \\
-{1 \over 2}g'v_1 & {1 \over 2}gv_1 & 0 & - \mu \\
{1 \over 2}g'v_2 & -{1 \over 2}gv_2 & - \mu & 0 \end{array}\right) ,
\end{equation}
where $M$ is the common gaugino mass parameter and
$\displaystyle{c={5 g'^2 \over 3g^2 }\simeq 0.5}$.
The seesaw hierarchy is now evident, when constraints on neutralino masses
and upper limits on lepton-number-violating couplings will be considered
in Section~5. We will then find that $(M_4)_{ij} \gg m_{kl}$ in agreement
with experimental constraints on neutralino and neutrino masses.
We can utilize this posterior fact to calculate the diagonalization
of ${\cal M}_0$ in an approximate way in terms of the small matrix-valued
quantity defined as
\begin{equation}\label{xi}
\xi=mM_4^{-1}\, .
\end{equation}
Parenthetically, we wish to draw the reader's attention to one
exact result in connection with the diagonalization of ${\cal M}_0$.
Because of the different hypercharge assignments of the two higgsinos
and the absence of light-neutrino masses at the tree level,
the first three lines together with the last line in ${\cal M}_0$ are not
linearly independent. As an immediate consequence of the latter, two neutrino
masses are exactly zero in the Born approximation \cite{JN}.
Let us now define the mass eigenstates $\Psi_0$ by the rotation
\begin{eqnarray} \label{m28}
\Psi_{0i}=\Xi_{ij} \Psi'_{0j}\, , \nonumber \\
\Xi^*{\cal M}_0 \Xi^{\dagger} =\widehat{{\cal M}}_0\, ,
\end{eqnarray}
where $\widehat{{\cal M}}_0$ is the diagonal matrix with neutrino/neutralino
masses as elements. To leading order in $\xi$ expansion,
the approximate form of $\Xi^*$ is readily estimated to be
\begin{equation} \label{m30}
\Xi^*=\left(\begin{array}{cc}
V_\nu^T & 0 \\
0 & N^* \end{array}\right)
\left(\begin{array}{cc}
1 -{1 \over 2}\xi \xi^{\dagger} & -\xi \\
\xi^{\dagger} & 1 -{1 \over 2}\xi^\dagger \xi
\end{array}\right),
\end{equation}
where the second matrix block-diagonalizes ${\cal M}_0$ to the form
$\mbox{diag}(m_{eff}, M_4)$ with
\begin{equation} \label{m31}
m_{eff}=-m\; M_4^{-1}\; m^T = {cg^2+ g'^2 \over D}\,
\left(\begin{array}{ccc}
\Lambda_e^2 & \Lambda_e \Lambda_\mu
& \Lambda_e \Lambda_\tau \\
\Lambda_e \Lambda_\mu & \Lambda_\mu^2
& \Lambda_\mu \Lambda_\tau \\
\Lambda_e \Lambda_\tau & \Lambda_\mu \Lambda_\tau & \Lambda_\tau^2
\end{array}\right).
\end{equation}
The quantities $\Lambda_i$ and $D$ newly introduced are defined as follows:
\begin{equation} \label{31}
\vec{\Lambda}\equiv \mu \vec{w} - v_1 \vec{\varepsilon}\, ,
\end{equation}
and
\begin{equation} \label{32}
D \equiv 4{det M_4 \over M}=2\mu \left[-2cM\mu + v_1v_2
\left(cg^2 + g'^2 \right)\right] .
\end{equation}
The sub-matrices $N$ and $V_\nu$ in Eq.~(\ref{m30}) diagonalize $M_4$ and
$m_{eff}$ in the following way:
\begin{equation} \label{33}
N^* M_4 N^{\dagger}=\mbox{diag}(m_{\tilde{\chi}^0_i})\, ,
\end{equation}
where $m_{\tilde{\chi}^0_i}$ are the heavy neutralino masses only.
For the diagonalization of $M_4$, we have retained the notation and
convention of Ref.~\cite{HK}.
For the neutrino case, we obtain
\begin{equation}
\label{m34}
V^T_{\nu}\; m_{eff}\; V_{\nu} = \mbox{diag}(0, \; 0, \; m_{\nu}) \, ,
\end{equation}
where the only non-zero neutrino mass is given by
\begin{equation} \label{m35}
m_{\nu}=tr(m_{eff})={cg^2+ g'^2 \over D}\, \vert \vec{\Lambda} \vert^2\, .
\end{equation}
Furthermore, an analytic calculation of the rotation matrix $V_{\nu}$
gives \cite{JN}
\begin{equation} \label{m36}
V_{\nu}=\left(\begin{array}{ccc}
\cos \theta_{13} & 0 & -\sin \theta_{13} \\
\sin \theta_{23}\sin \theta_{13} & \cos \theta_{23} & \sin \theta_{23}
\cos \theta_{13} \\
\sin \theta_{13} & \sin \theta_{23} & \cos \theta_{13}\cos \theta_{23}
\end{array}\right) ,
\end{equation}
where the mixing angles are expressed through the vector
$\vec{\Lambda}$ as follows:
\begin{equation} \label{m37}
\tan \theta_{13} = -{\Lambda_e \over
\sqrt{\Lambda_\mu^2 + \Lambda_\tau^2}}, \;\;\;\;\;
\tan \theta_{23} = {\Lambda_\mu \over \Lambda_\tau} \, .
\end{equation}
In~\cite{JN}, the baryogenesis constraint on all lepton-number-violating
couplings were applied, which led to a solution to the solar neutrino puzzle
through vacuum oscillations.
In that case, the neutrino mass $m_\nu$ came out rather naturally of order
$10^{-5}$ eV, while the mixing angle $\theta_{13}$ was predicted to be
large, {\em i.e.}\ $\tan\theta_{13} \simeq -1/\sqrt{2}$.
Since we can evade the constraints from BAU by conserving one individual lepton
number, our scenario regarding the light neutrino mass, $m_\nu$,
is quite different.
In Section 6, we will discuss some numerical examples of the neutrino mass
as well as the resulting constraints on $w_i$, $\varepsilon_i$ together with
the constraints emerging from exotic processes.
Here we note in passing
that appreciable values for $w_i$ and $\varepsilon_i$ in the GeV range
result in a tau-neutrino mass of ${\cal O}($MeV) which is still allowed
by laboratory constraints. Let us now demonstrate explicitly, by an
example, how the $\varepsilon_i$ terms can change
some of the phenomenological implications. We choose the following set
of parameters: $M=\mu=2M_W$, $\tan\beta =1$, $\varepsilon_{\tau}=w_{\tau}=0$,
$w_e=w_{\mu}\equiv w=1$~GeV, $\varepsilon_{e}=\varepsilon_{\mu} \equiv
\varepsilon$. If we now put $\varepsilon=0$, then $m_{\nu} \simeq 3.5$~MeV.
On the other hand, the same soft-SUSY parameters and vev's,
but having now $\varepsilon=4w$ instead, give $m_{\nu} \simeq 38.5$~MeV,
which already exceeds the laboratory limit on the tau-neutrino mass.
\subsection{Chargino--charged lepton mixing}
\indent
Similar to the case of neutralino--neutrino mixing, the explicit violation
of the lepton number allows also for chargino--charged lepton mixing.
In two component notation, the mass term takes the form
\begin{equation} \label{m38}
{\cal L}^{\chi^+}_{mass}= -\zeta^{'T} {\cal M}_+\, \omega'\ +\ \mbox{H.c.}\, ,
\end{equation}
where in the vector $\zeta'$, we gather the lower components of a charged
Dirac spinor in the Weyl representation, {\em i.e.}\
\begin{equation} \label{m39}
\zeta^{'T}=(\psi^2_{L_1},\,\, \psi^2_{L_2},\,\, \psi^2_{L_3},\,\,
-i\lambda_-,\,\, \psi^2_{H_1})\, ,
\end{equation}
whereas $\omega'$ contains the upper components
\begin{equation} \label{m40}
\omega^{'T}=(\psi_{R_1},\,\, \psi_{R_2},\,\, \psi_{R_3},\,\,
-i\lambda_+,\,\, \psi^1_{H_2})\, .
\end{equation}
In order to establish contact between the notation of the MSSM in~\cite{HK}
or that of our minimal $\not\!\! R$ model and the usual SM notation, we
note that the charged leptons are represented
by their charged conjugate fields, {\em i.e.}
\begin{displaymath}
l_i^C\ =\ \left( \begin{array}{c}
\psi_{R_i} \\ \bar{\psi}_{L_i}^2 \end{array} \right).
\end{displaymath}
In this basis, the chargino/charged-lepton mass matrix ${\cal M}_+$
appearing in Eq.~(\ref{m38}) may be written down as
\begin{equation} \label{m41}
{\cal M}_+=\left(\begin{array}{cc}
M_l & E \\
E' & S
\end{array}\right) ,
\end{equation}
where $S$ is the usual MSSM chargino mass matrix given by
\begin{equation} \label{m42}
S=\left(\begin{array}{cc}
M & \frac{1}{\sqrt{2}} g v_2 \\
\frac{1}{\sqrt{2}} g v_1 & \mu
\end{array}\right) .
\end{equation}
The sub-matrices $E$ and $E'$ which give rise to chargino--charged lepton
mixing
are defined as follows:
\begin{equation} \label{m43}
\displaystyle{E=\left(\begin{array}{cc}
{1\over \sqrt{2}} g w_e & \varepsilon_e \\
{1\over \sqrt{2}} g w_\mu & \varepsilon_{\mu} \\
{1\over \sqrt{2}} g w_\tau & \varepsilon_{\tau}
\end{array}\right)} ,
\end{equation}
and
\begin{equation} \label{m44}
E'=\left(\begin{array}{ccc}
0 & 0 & 0 \\
\Upsilon_e & \Upsilon_{\mu} & \Upsilon_{\tau}
\end{array}\right) ,
\end{equation}
where $\displaystyle{\Upsilon_l \sim {m_l \over v_1} w_l}$ and
$m_l$ are the lepton masses. For our numerical purposes,
we will assume that $M_l$ is a diagonal matrix whose elements can be
identified, to a high accuracy, with the physical lepton masses $m_i$.
In addition, we can neglect the elements of $E'$ as
compared to the other entries in Eq.~(\ref{m41}). Therefore, we will be
working in the approximation $E'=0$.
Let us now express the mass eigenstates $\zeta$ and $\omega$ in terms
of the states $\zeta'$ and $\omega'$ via the unitary transformations
\begin{equation} \label{m45}
\zeta_i=\Sigma_{ij}\zeta'_j, \,\,\, \omega=\Omega_{ij}\omega'_j\, .
\end{equation}
The bi-diagonalization leads then to the diagonal matrix $\widehat{{\cal M}}_+$
whose elements are the chargino and lepton masses
\begin{equation} \label{m46}
\Sigma^* {\cal M}_+ \Omega^{\dagger}=\widehat{{\cal M}}_+\, .
\end{equation}
Proceeding now as in the case of the neutralino--neutrino mixing, we
carry out an approximate diagonalization for ${\cal M}_+$.
In this way, the expansion parameters are found to be
\begin{eqnarray} \label{m47}
&&\xi_{{}_{L}}^*=ES^{-1} \, ,\nonumber \\
&&\xi_{{}_{R}}^*=M_l^{\dagger}ES^{-1}(S^{-1})^T
=M_l^{\dagger}\xi_{{}_{L}}^*(S^{-1})^T\, .
\end{eqnarray}
Note that $\xi_{{}_{R}} \sim \xi_{{}_{L}} m_l/M$.
To leading order in $\xi_{{}_L}$ and $\xi_{{}_R}$, the rotation matrices
are written down as
\begin{equation} \label{m48}
\Sigma^*=\left(\begin{array}{cc}
V_L & 0 \\
0 & U^* \end{array}\right)\left(\begin{array}{cc}
1 -{1 \over 2} \xi_{{}_{L}}^*\xi_{{}_{L}}^T & -\xi_{{}_{L}}^* \\
\xi_{{}_{L}}^T & 1 -{1 \over 2} \xi_{{}_{L}}^T\xi_{{}_{L}}^*
\end{array}\right) ,
\end{equation}
and
\begin{equation} \label{m49}
\Omega^{\dagger}=
\left(\begin{array}{cc}
1 -{1 \over 2} \xi_{{}_{R}}^*\xi_{{}_{R}}^T & \xi_{{}_{R}}^* \\
-\xi_{{}_{R}}^T & 1 -{1 \over 2} \xi_{{}_{R}}^T\xi_{{}_{R}}^*
\end{array}\right)
\left(\begin{array}{cc}
V_R^{\dagger} & 0 \\
0 & V^{\dagger} \end{array}\right).
\end{equation}
Adopting the convention of~\cite{HK} for the matrices, which
also appear in the MSSM, we have
\begin{eqnarray} \label{m50}
&& U^*SV^{\dagger}=\widehat{S}\, , \nonumber \\
&& V_L M_l V_R^{\dagger} =\widehat{M}_l\, ,
\end{eqnarray}
where, as before, the hatted matrices are diagonal.
\setcounter{equation}{0}
\section{The $W$- and $Z$-boson interaction Lagrangians}
\indent
In this section, we will derive the interaction Lagrangians of $Z$ and $W$
bosons with neutralinos/neutrinos and charginos/charged leptons. We will
first present general expressions and, subsequently, use the analytic
results of the approximate diagonalization of the mass matrices, given in
the preceeding section, to calculate the mixing matrices in the first
order approximation. Here and in the following,
because of the mixing, we collectively call $\Psi_0$ all the neutralinos,
with neutrinos being the light neutralinos, and $\zeta$, $\omega$
all the charginos, where the charged leptons are the light charginos.
\subsection{General expressions}
\indent
Starting from two component notation and defining for convenience the matrix
\begin{equation} \label{m51}
T^Z=\mbox{diag}(1,\, 1,\, 1,\, 0,\, 0,\, 1,\, -1)\, ,
\end{equation}
the interaction Lagrangian of $Z$ with neutralinos reads
\begin{equation} \label{m52}
{\cal L}^{Z\chi^0\chi^0}_{int}\ =\ -{g \over 2\cos\theta_w}
Z^{\mu}\bar{\Psi}'_{0i} T^Z_{ij}\bar{\sigma}_{\mu}\Psi'_{0j} \, .
\end{equation}
After replacing the weak eigenstates $\Psi'_{0i}$ by four component
Majorana mass eigenstates $\chi^0_i$ in Eq.~(\ref{m52}), we obtain
\begin{equation} \label{Znunu}
{\cal L}^{Z\chi^0\chi^0}_{int}\ =\ - \frac{g}{4\cos\theta_w}\, Z^\mu\,
\bar{\chi}^0_i \gamma_\mu\Big( i \Im m \tilde{C}_{ij}\
-\ \gamma_5 \Re e \tilde{C}_{ij}\Big) \chi^0_j\, ,
\end{equation}
where
\begin{equation} \label{m54}
\tilde{C}_{ij}\ =\ (\tilde{C}^{\dagger})_{ij}\ =\ (\Xi T^Z \Xi^{\dagger})_{ij}
\, .
\end{equation}
It is easy to check that Eq.~(\ref{Znunu}) reproduces the
$Z$-neutralino-neutralino interaction of the MSSM, when the leptonic
$\not\!\! R$ admixture is neglected.
To calculate the $Z$-chargino-chargino coupling, we again define two
auxiliary matrices
\begin{eqnarray} \label{m55}
T^Z_R &=& \mbox{diag}(0,\,0,\, 0,\,2,\, 1)\, , \nonumber \\
T^Z_L &=& \mbox{diag}(1,\,1,\, 1,\,2,\, 1)\, ,
\end{eqnarray}
such that ${\cal L}^{Z\bar{\chi}^-\chi^-}_{int}$ takes the form
\begin{eqnarray} \label{m56}
{\cal L}^{Z\bar{\chi}^-\chi^-}_{int}
&=&{g \over 2\cos \theta_w}Z^{\mu}\Big[
\bar{\zeta}'_i(T^Z_L)_{ij}\bar{\sigma}_{\mu}\zeta'_j
- \bar{\omega}'_i(T^Z_R)_{ij}\bar{\sigma}_{\mu}\omega'_j
\nonumber \\
&& +2 \sin^2 \theta_w ( \bar{\omega}'_i
\bar{\sigma}_{\mu}\omega'_i
-\bar{\zeta}'_i\bar{\sigma}_{\mu}\zeta'_i )\Big] .
\end{eqnarray}
Denoting by $\chi^{-}_i$ the physical charginos in the
four-component Dirac notation pertaining to the definition of Eq.~(\ref{m39}),
the above Lagrangian can be written down as
\begin{equation}\label{Zll}
{\cal L}^{Z\bar{\chi}^-\chi^-}_{int}\ =\ \frac{g}{2\cos\theta_w}\, Z^\mu\,
\bar{\chi}^-_i \gamma_\mu \Big( \tilde{A}^L_{ij} \mbox{P}_L\
+\ \tilde{A}^R_{ij} \mbox{P}_R\Big)\chi^-_j\, ,\\
\end{equation}
where P$_L$(P$_R)=[1-(+)\gamma_5]/2$ and
\begin{eqnarray} \label{m58}
\tilde{A}^L_{ij} &=& (\Sigma T^Z_L \Sigma^{\dagger})_{ij} - 2\delta_{ij}
\sin^2 \theta_w\, , \nonumber \\
\tilde{A}^R_{ij}&=&(\Omega^* T^Z_R \Omega^T)_{ij} - 2\delta_{ij}
\sin^2 \theta_w \, .
\end{eqnarray}
As done above for the neutral-current interactions, for the charged-current
case we first introduce two auxiliary matrices given by
\begin{equation} \label{m59}
T^L=\left(\begin{array}{ccccccc}
1 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 1 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 1 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & \sqrt{2} & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 1 & 0 \end{array}\right), \, \, \,
T^R=\left(\begin{array}{ccccc}
0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & -\sqrt{2} & 0 \\
0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 1 \end{array}\right) .
\end{equation}
In the Weyl weak eigenbasis, we can then write
\begin{equation} \label{m60}
{\cal L}^{W\chi^-\chi^0}_{int}\ =\ -{ g \over \sqrt{2}} \Big(
\bar{\zeta}'_i T^L_{ij}\bar{\sigma}^{\mu}\Psi'_{0j} +
\bar{\Psi'}_{0i} T^R_{ij}\bar{\sigma}^{\mu}\omega'_j\Big) W^-_{\mu}\
+\ \mbox{H.c.}\, ,
\end{equation}
and
\begin{equation}\label{Wlnu}
{\cal L}^{W\chi^-\chi^0}_{int} \ =\ -\frac{g}{\sqrt{2}}\, W^{-\mu}\,
\bar{\chi}^-_i \gamma_\mu \Big( \tilde{B}^L_{ij} \mbox{P}_L\
+\ \tilde{B}^R_{ij} \mbox{P}_R\Big) \chi^0_j\ +\ \mbox{H.c.}\, ,\\
\end{equation}
in four-component mass eigenbasis notation. The mixing matrices are
\begin{eqnarray} \label{m62}
\tilde{B}^L_{ij} &=& (\Sigma T^L \Xi^\dagger)_{ij}\, , \nonumber \\
\tilde{B}^R_{ij} &=& -[\Omega^* (T^R)^T \Xi^T ]_{ij}\, .
\end{eqnarray}
In the next section, we will give analytic approximate expressions for
all mixing matrices defined above.
\subsection{Mixing matrices}
\indent
In Section~4.1, we have derived the analytic expressions of
the interaction Lagrangians of the $W$ and $Z$ bosons with chargino and
neutralino states in the $R$-parity-violating SUSY model.
However, the mixings $\tilde{A}^L$, $\tilde{A}^R$,
$\tilde{B}^L$, $\tilde{B}^R$, and $\tilde{C}$ that govern
these interactions are high dimensional matrices, involving
a large number of parameters. Therefore, it is more convenient
to find approximative forms for the mixing matrices that will
enable us to appreciate the strength of the $Z$- and $W$-boson
couplings in the model under consideration.
To facilitate our presentation, we first introduce the following
auxiliary matrices:
\begin{eqnarray}
d &=& \mbox{diag}(2,1)\, ,\qquad t_z\ =\ \mbox{diag}(0,0,1,-1)\, ,\nonumber\\
t_{{}_L}& =& \left( \begin{array}{cccc}
0 & \sqrt{2} & 0 & 0 \\
0 & 0 & 1 & 0 \end{array} \right),\nonumber\\
t_{{}_R}& =& \left( \begin{array}{cccc}
0 & \sqrt{2} & 0 & 0 \\
0 & 0 & 0 & -1 \end{array} \right).
\end{eqnarray}
Substituting the unitary mixing matrices of Eqs.\ (\ref{m30}), (\ref{m48}),
and (\ref{m49}) into Eqs.~(\ref{m54}), (\ref{m58}), and (\ref{m62}),
and neglecting terms of ${\cal O}(\xi_{{}_L}^3,\xi^3 )$ and higher,
we obtain
\begin{eqnarray}
\tilde{A}^L &=& -2\sin^2\theta_w {\mbox{\bf 1}}\ +\ \left(
\begin{array}{cc}
1-\tilde{\xi}_{{}_L}(1-d)\tilde{\xi}_{{}_L}^\dagger &
\tilde{\xi}_{{}_L}(1-d)U^\dagger\\
U(1-d)\tilde{\xi}^\dagger_{{}_L} & UdU^\dagger \end{array}\right), \label{AL}\\
\tilde{A}^R &=& -2\sin^2\theta_w {\mbox{\bf 1}}\ +\ \left(
\begin{array}{cc}
0 & -\tilde{\xi}_{{}_R}dV^T\\
-V^*d\tilde{\xi}_{{}_R}^\dagger & V^*d V^T \end{array}\right), \label{AR}\\
\tilde{B}^L &=& \left( \begin{array}{cc}
V^l-\frac{1}{2}\tilde{\xi}_{{}_L}\tilde{\xi}^\dagger_{{}_L}V^l-
\frac{1}{2}V^l\tilde{\xi}^*\tilde{\xi}^T +
\tilde{\xi}_{{}_L} t_{{}_L} \tilde{\xi}^T
& (V^l\tilde{\xi}^* - \tilde{\xi}_{{}_L} t_{{}_L}) N^\dagger \\
U(\tilde{\xi}^\dagger_{{}_L}V^l-t_{{}_L}\tilde{\xi}^T) & Ut_{{}_L}N^\dagger
\end{array} \right),\label{BL}\\
\tilde{B}^R &=& \left( \begin{array}{cc}
0 & - \tilde{\xi}_{{}_R} t_{{}_R} N^T \\
-V^* t_{{}_R} \tilde{\xi}^\dagger & V^* t_{{}_R} N^T \end{array}
\right),\label{BR}\\
\tilde{C} &=& \left( \begin{array}{cc}
1- \tilde{\xi}^* (1-t_z)\tilde{\xi}^T & \tilde{\xi}^* (1-t_z) N^\dagger \\
N(1-t_z)\tilde{\xi}^T & N t_z N^\dagger \end{array} \right).\label{C}
\end{eqnarray}
Here, we have defined $\tilde{\xi}_{{}_L}
=V_L^*\xi_{{}_L}$,
$\tilde{\xi}_{{}_R}=\widehat{M}_l \tilde{\xi}_{{}_L} (S^{-1})^\dagger$,
$\tilde{\xi} = V_\nu^T \xi$, and $V^l=V_L^* V_\nu$.
Furthermore, the unitary matrices $V_L$, $V_R$, $V_\nu$,
$U$, $V$, $N$, together with the mixing matrices $\xi_{{}_L}$,
$\xi_{{}_R}$, and $\xi$ are defined in Section~3. In the derivation
of Eqs.~(\ref{AR}) and (\ref{BR}), we have also used the fact that
$\xi_{{}_R} ={\cal O}( \xi_{{}_L} m_l/M)$.
{}From Eqs.~(\ref{AL})--(\ref{C}), it is now easy to see how the
$R$-parity-violating couplings to ordinary leptons deviate from the
SM vertices. To leading order in $\xi_{{}_L}$ and $\xi$, we find that the
interactions of the $W$ and $Z$ bosons with left-handed charged leptons
and neutrinos are modified, whereas the corresponding couplings to
right-handed charged leptons remain unaffected, having
the SM form.
\setcounter{equation}{0}
\section{Laboratory and cosmological constraints}
\indent
Our aim is to constrain the parameter space of this
$\not \!\! R$ scenario, by taking laboratory and
cosmological constraints into account. For this purpose,
we will pay special attention to limits derived from
low-energy processes and LEP data, such as charged lepton decays
of the form $l^-\to l'^-l_1^-l_1^+$, flavour-changing
$Z$-boson decays $Z\to l_il_j$, the invisible width
of the $Z$ boson, charged-current universality in muon
and tau decays, lepton universality at the $Z$ peak,
and charged-current universality in pion decays.
In this vein, we will report some phenomenological implications
of our minimal model that may be relevant to explain the intriguing
anomaly found by the KARMEN collaboration~\cite{KARMEN}. In the last section,
we will discuss the viability of our model when cosmological constraints are
considered, such as the requirement of not washing out the primordial BAU and
the absence of large disruptive reheating effects caused by an unstable $\tau$
neutrino with $m_{\nu_\tau}={\cal O}(10)$~MeV.
\subsection{{\boldmath $l^-\to l'^-l_1^-l^+_1$}}
\indent
As has been found in Lagrangian~(\ref{Zll}), the model
predicts flavour-changing neutral current (FCNC) $Zll'$ couplings
at the tree level. These new $\not\!\! R$ interactions induce
$\tau$ and $\mu$ decays into three lighter charged leptons. In
this way, we obtain
\begin{equation}
B(l^-\to l'^-l_1^-l^+_1)\ =\
\frac{\alpha_w^2 m^4_l}{1536\pi M_W^4}\, \frac{m_l}{\Gamma_l}
\left( \left|\tilde{A}^L_{ll'}\right|^2
+\left|\tilde{A}^R_{ll'}\right|^2\right)\left(\left|\tilde{A}^L_{l_1l_1}
\right|^2 +\left|\tilde{A}^R_{l_1l_1}\right|^2\right), \label{BRlll}
\end{equation}
where $\alpha_w=g^2/4\pi$ and $\Gamma_l$ is the total width of the decaying
charged lepton $l$.
The experimental upper limit on the branching ratio of $\mu^- \to e^-e^-e^+$
is given by~\cite{PDG94}
\begin{equation}
B(\mu^- \to e^-e^-e^+) \ \leq\ 1.0\cdot 10^{-12}\, ,\label{Bexpmueee}
\end{equation}
at $90\%$ confidence level (CL).
Recently, CLEO collaboration~\cite{CLEO} has considerably lowered
experimental upper bounds on branching ratios of
neutrinoless $\tau$-lepton decays. They have found
\begin{eqnarray}
B(\tau^-\to e^- e^+ e^- ) &\leq& 3.3\cdot 10^{-6}\, ,\nonumber\\
B(\tau^-\to \mu^- e^+ e^- ) &\leq& 3.4\cdot 10^{-6}\, ,\nonumber\\
B(\tau^-\to e^- \mu^+ \mu^- ) &\leq& 3.6\cdot 10^{-6}\, ,\nonumber\\
B(\tau^-\to \mu^- \mu^+ \mu^- )&\leq& 4.3\cdot 10^{-6}\, ,\label{Bexptaulll}
\end{eqnarray}
at $90\%$ CL. Theoretical predictions obtained for the observables
given in Eqs.~(\ref{Bexpmueee}) and~(\ref{Bexptaulll}) will be discussed
in Section 6.
\subsection{{\boldmath $Z\to l^-l'^+$} and {\boldmath $Z\to \nu\nu$}}
\indent
The presence of FCNC $Zll'$ couplings at the tree level will also
give rise to flavour-violating $Z$-boson decays at LEP. The theoretical
prediction of their branching ratios is determined by
\begin{equation}
B(Z\to l^-l'^+\ \mbox{or}\
l^+ l'^-)\ =\ \frac{\alpha_w}{12\cos^2\theta_w}\,
\frac{M_Z}{\Gamma_Z}\left( \left|\tilde{A}^L_{ll'}\right|^2
+\left|\tilde{A}^R_{ll'}\right|^2 \right),\label{BZlilj}
\end{equation}
where $\Gamma_Z=2.49$~GeV is the total width of the $Z$ boson
measured experimentally~\cite{PDG94}. Furthermore, an analysis
of this kind of decays at LEP yields
\begin{eqnarray}
B( Z\to e^-\mu^+\ \mbox{or}\ e^+\mu^-) &\leq & 6.0\cdot 10^{-6}\,
,\nonumber\\
B( Z\to e^-\tau^+\ \mbox{or}\ e^+\tau^-) &\leq & 1.3\cdot 10^{-5}\,
,\nonumber\\
B( Z\to \tau^-\mu^+\ \mbox{or}\ \tau^+\mu^-) &\leq & 1.9\cdot 10^{-5}\, ,
\label{BexpZlilj}
\end{eqnarray}
at $95\%$ CL. In addition, the Lagrangian (\ref{Znunu}) modifies the
invisible width of the $Z$ boson through the non-universal and
flavour-dependent $Z\nu_i\nu_j$ tree-level couplings. It is then easy
to obtain the branching ratio for the total invisible $Z$-boson
width, which is assumed to be caused mainly by $Z\to \nu_i\nu_j$
\begin{equation}
B(Z\to \nu\bar{\nu})\ =\ \frac{\alpha_w}{24\cos^2\theta_w}\,
\frac{M_Z}{\Gamma_Z}\sum\limits_{\nu_i,\nu_j}\, \left|\tilde{C}_{\nu_i
\nu_j}\right|^2. \label{BZnunu}
\end{equation}
On the other hand, an experimental analysis on the $Z$ pole
gives~\cite{PDG94}
\begin{equation}
1 - \frac{B(Z\to \nu\bar{\nu})}{B_{SM}(Z\to \mbox{invisible} )}\
\leq \ 1.31\cdot 10^{-2}\, ,\label{BexpZnunu}
\end{equation}
where $B_{SM} (Z\to \mbox{invisible} )$ is the SM prediction for the
invisible width of the $Z$ boson.
In Section 6, we will analyze the phenomenological impact of
the new-physics decay channels mentioned above on restricting
our model.
\subsection{ Universality violation at the {\em Z} peak}
\indent
Interesting limits on $R$-parity breaking, nonuniversal, diagonal $Zll$
couplings can be extracted from measurements of lepton universality
on the $Z$-boson pole. In order to impose constraints, we will adopt the
LEP observable based on leptonic $Z$-boson partial width
differences studied in~\cite{BKPS}
\begin{equation}
U_{br}^{(ll' )}\ =\ \frac{\Gamma(Z\to l^+l^-)
- \Gamma (Z\to l'^+l'^-)}{\Gamma(Z\to l^+l^-) + \Gamma (Z\to l'^+l'^-)}
\ =\ \frac{ |\tilde{A}^L_{ll}|^2\ -\ |\tilde{A}^L_{l'l'}|^2}{
|\tilde{A}^L_{ll}|^2\ +\ |\tilde{A}^L_{l'l'}|^2}\, ,\label{Ubr}
\end{equation}
where $l\ne l'$. A combined experimental analysis for
the observable $U_{br}$ gives~\cite{PDG94}
\begin{eqnarray}
|U_{br}^{(ll')}| \ \leq \ 5.0\cdot 10^{-3}\, ,\label{Ubrexp}
\end{eqnarray}
at 1$\sigma$ level, almost independent of the charged leptons $l$ and $l'$.
Another relevant observable involving leptonic asymmetries, which
has been analyzed in~\cite{BP}, is
\begin{equation}
\Delta{\cal A}_{ll'}\ =\ \frac{{\cal A}_l\ -\ {\cal A}_{l'}}{
{\cal A}_l\ +\ {\cal A}_{l'}}\ =\ \left(\frac{1}{{\cal A}_l^{(SM )}}\
-\ 1\right)\, U_{br}^{(ll')}\, ,\label{DAll'}
\end{equation}
where ${\cal A}_l^{(SM)}=0.14$ is the leptonic asymmetry predicted
theoretically in the SM.
In the last step of Eqs.~(\ref{Ubr}) and~(\ref{DAll'}), we have used the
fact that, to a good approximation, the tree-level coupling of the $Z$ boson
to right-handed charged leptons is universal in our minimal $R$-parity
violating SUSY model, {\em i.e.}\ $\tilde{A}^R_{ll}=\tilde{A}^R_{l'l'}$
as can be seen from Eq.~(\ref{AR}).
Considering the experimental upper bound on $U_{br}$ given in
Eq.~(\ref{Ubrexp}), Eq.~(\ref{DAll'}) furnishes the upper
limit
\begin{equation}
\Delta{\cal A}_{ll'} \ \le\ 3.0\cdot 10^{-2}\, , \label{DAUbr}
\end{equation}
which is slightly below the present experimental sensitivity at
LEP~\cite{LEP} [$\Delta{\cal A}^{LEP}_\tau/{\cal A}^{(SM)}_l=0.07$, at
1$\sigma$] and Stanford Linear Collider (SLC)~\cite{SLD}
[$\Delta{\cal A}^{SLC}_e/{\cal A}_l^{(SM)}=0.04$, at 1$\sigma$].
It is also interesting to notice that the apparent difference of
$\Delta{\cal A}_{\tau e}\simeq - 10\%$ between the measured
leptonic asymmetries ${\cal A}^{SLC}_e$ and ${\cal A}^{LEP}_\tau$
cannot be predicted in our $\not\!\!\! R$ model, without invalidating
the inequality~(\ref{DAUbr}) at the same time.
\subsection{Decays {\boldmath $\mu\to e\nu\nu$} and
{\boldmath $\tau\to e\nu\nu$}}
\indent
Useful constraints can be obtained from possible deviations
of charged-current universality in $\tau$-lepton decays.
In fact, measures of such deviations can be defined
and straightforwardly be calculated as follows:
\begin{eqnarray}
R_{\tau e} &=& \frac{\Gamma (\tau\to e \nu\bar{\nu})}{
\Gamma (\mu\to e \nu\bar{\nu})}\ =\ R_{\tau e}^{SM}\,
\frac{\sum\limits_{\nu_i} \Big[ |\tilde{B}^L_{\tau\nu_i}|^2 +
|\tilde{B}^R_{\tau \nu_i}|^2 \Big]}{\sum\limits_{\nu_j}
\Big[ |\tilde{B}^L_{\mu\nu_j}|^2 + |\tilde{B}^R_{\mu \nu_j}|^2 \Big]}\, ,
\label{Rtaue}\\
R_{\tau \mu} &=& \frac{\Gamma (\tau\to \mu \nu\bar{\nu})}{
\Gamma (\mu\to e \nu\bar{\nu})}\ =\ R_{\tau\mu}^{SM}\,
\frac{\sum\limits_{\nu_i} \Big[ |\tilde{B}^L_{\tau\nu_i}|^2 +
|\tilde{B}^R_{\tau \nu_i}|^2 \Big]}{\sum\limits_{\nu_j}
\Big[ |\tilde{B}^L_{e\nu_j}|^2 + |\tilde{B}^R_{e \nu_j}|^2 \Big]}\,
.\label{Rtaumu}
\end{eqnarray}
In Eqs.~(\ref{Rtaue}) and~(\ref{Rtaumu}), the SM contributions
to the observables, $R_{\tau e}^{SM}$ and $R_{\tau\mu}^{SM}$,
have been factored out. Of course, deviations from the SM values
can also be induced by the $\lambda$-dependent interactions in
Eq.~(\ref{Llambda}). These observables are used to constrain
the couplings $\lambda_{ijk}$ as a function of the mass of the scalar
right-handed leptons~\cite{BGH}. To avoid excessive complication,
we assume that all $\lambda_{ijk}=0$ and focus our study mainly on
the phenomenological consequences originating from the $\varepsilon_i$
terms in the superpotential. Furthermore, experimental limits
related to the ratios $R_{\tau e}$ and $R_{\tau\mu}$ may be presented
in the following way~\cite{DAB}:
\begin{eqnarray}
1\ -\ \frac{R_{\tau e}}{R^{SM}_{\tau e}} &=& 0.040 \pm 0.024\,
,\label{Rexptaue}\\
1\ -\ \frac{R_{\tau\mu}}{R^{SM}_{\tau\mu}} &=& 0.032 \pm 0.024\,
,\label{Rexptaumu}
\end{eqnarray}
at 1$\sigma$ level. Constraints obtained from Eqs.~(\ref{Rexptaue})
and~(\ref{Rexptaumu}) on the parameters of our $\not\!\! R$ model
will be discussed in Section 6.
\subsection{Charged-current universality in pion decays}
\indent
Complementary to the physical quantities $R_{\tau e}$ and $R_{\tau\mu}$
are the constraints derived
from the ratio $R_\pi=\Gamma (\pi\to e\nu )/\Gamma (\pi\to \mu\nu )$
in the $\pi^-$ decays. $R_\pi$ is an observable that measures
possible deviations from charged-current universality in the $e-\mu$ system.
It is not difficult to obtain
\begin{equation}
R_\pi\ \ =\ R_\pi^{SM}\,
\frac{\sum\limits_{\nu_i} \Big[ |\tilde{B}^L_{e\nu_i}|^2 +
|\tilde{B}^R_{e \nu_i}|^2 \Big]}{\sum\limits_{\nu_j}
\Big[ |\tilde{B}^L_{\mu\nu_j}|^2 + |\tilde{B}^R_{\mu \nu_j}|^2 \Big]}\,
.\label{Rpi}
\end{equation}
In addition, the 1$\sigma$ experimental bound related to $R_\pi$ is
given by~\cite{DAB}
\begin{equation}
\frac{R_\pi}{R_\pi^{SM}}\ -\ 1 \ =\ 0.003\pm 0.003\, .\label{Rpiexp}
\end{equation}
It is again worth mentioning that similar deviations of $e-\mu$
universality can arise from the presence of $\lambda'$-dependent
couplings through the interaction Lagrangian~(\ref{Llambda'}).
In our analysis, we will assume that all $\lambda'_{ijk}=0$.
This may also be reflected by the fact that the current experimental
lower bound on the half-lifetime of the $^{76}$Ge $0\nu\beta\beta$ decay
leads to the tight constraint~\cite{HKK}
\begin{equation}
\lambda'_{111}\ \le\ 3.9\cdot 10^{-4}\, \left( \frac{m_{\tilde{q}}}{100\
\mbox{GeV}}\right)^2 \left( \frac{m_{\tilde{g}}}{100\ \mbox{GeV}}
\right)^{1/2},
\end{equation}
where $\tilde{q}$ ($\tilde{g}$) is the scalar quark (gluino).
\subsection{KARMEN anomaly}
\indent
Recently, the KARMEN collaboration, which operates at RAL, has
reported an anomaly~\cite{KARMEN} in the time-dependence of
decay spectra coming from stopped pions. To account for the
KARMEN anomaly, one can make the plausible assumption that a new
massive weakly-interacting particle, say $x$, is produced in the
pion decays, {\em i.e.}\ $\pi^+\to \mu^+ x$~\cite{KARMEN,BPS}. The
mass of this hypothetical particle should be $m_x\simeq 33.9$ MeV,
since it should explain the apparent $\sim 2\sigma$ bump present in
the time distribution of decaying muon events, which should normally
fall off exponentially.
This experimental peak occurs with a time delay of 3.6~$\mu$sec
after all pulsed pions have promptly decayed.
A recent study~\cite{BPS} suggests that the $x$ particle
should have similar features with those of a neutrino, but it cannot be
the $\nu_\tau$, because $m_{\nu_\tau}< 31$ MeV at 95$\%$ CL~\cite{PDG94},
or another predominantly-isodoublet neutrino, without affecting limits coming
from the supernova 1987A. The authors
in~\cite{BPS} further advocate that a mainly-sterile neutrino scenario
could, in principle, be compatible with all constraints ---both terrestrial and
astrophysical---, since the production of $x$ particles both in supernova
and in the early universe could then be suppressed. Although
in our $\not\!\! R$ model the coupling mixing matrices describing the
charged- and neutral-current interactions differ crucially from
usual singlet-neutrino scenarios~\cite{ZPC}, the above discussion
is still valid and translates into the requirement that one neutralino
state, {\em e.g.}\
$\chi$, should be light, having a mass $m_\chi=m_x$. Assuming
that the KARMEN anomaly gets resolved by the decay
$\chi\to e^-e^+\nu$, we have for the Majorana fermion $\chi$~\cite{BPS}
\begin{eqnarray}
|\tilde{B}^L_{e\chi}|\, |\tilde{B}^L_{\mu\chi}| &\simeq & 0.6\cdot 10^{-6}\, ,
\nonumber\\
|\tilde{B}^L_{e\chi}| &\stackrel{\displaystyle <}{\sim} & 2.5\cdot 10^{-4}\, ,
\nonumber\\
|\tilde{B}^L_{\mu\chi}| &\stackrel{\displaystyle <}{\sim} & 4.5\cdot 10^{-2}\,
{}.
\label{KARMENmix}
\end{eqnarray}
The bounds presented in Eq.~(\ref{KARMENmix}) are obtained from a number
of phenomenological requirements, such as the absence of a correction
to the Michael $\rho$ parameter in $\mu\to e\nu\nu$, negligible
decay events in neutrino beams, no anomalous contributions to
$\pi\to e\chi$, limits from neutrinoless double-$\beta$
decays, {\em etc}.
Because of the large number of parameters existing in our model,
it appears not difficult to accommodate the upper limits and
relations given in Eq.~(\ref{KARMENmix}). However, the soft-SUSY
breaking parameters in our model have to satisfy the following
hierarchy scheme:
\begin{eqnarray}
M\, (=2M') &\stackrel{\displaystyle >}{\sim}& 500\ \mbox{GeV}\, ,\nonumber\\
\mu &\stackrel{\displaystyle <}{\sim}& 30\ \mbox{MeV}\, ,\nonumber\\
\vec{\varepsilon} & \sim & \frac{\mu}{v_1}\vec{w}\, ,\label{KARMENsusy}
\end{eqnarray}
which is mainly prescribed by the fact that $m_\chi=33.9$~MeV.
{}From Eq.~(\ref{KARMENsusy}), we find that only SUSY models with
a $\mu$ at the scale of 10 MeV have a chance to account
for the KARMEN anomaly.
Similar $R$-parity broken SUSY models were also discussed in
Ref.~\cite{EGJRV}. Adapting the results of~\cite{EGJRV},
one can estimate that for $\tan\beta=1$ and $w_\tau<60$~GeV,
$B(Z\to \chi \chi)\simeq w_\tau^4/(3v^4)<1.\ 10^{-3}$ in compliance
with the LEP bound on invisible $Z$-boson decays in
Eq.~(\ref{BexpZnunu}).
However, such light-$\mu$ scenarios may encounter the
known $\mu$ hierarchy problem, where $\mu\sim M_{Pl}$
as derived naively from supergravity. Even though one
could invoke the Guidice--Masiero mechanism~\cite{GM}
to obtain a value of $\mu$ at the electroweak scale,
the small value of $\mu={\cal O}(10)$ MeV would, however,
require an additional unnatural suppression of the
gravitational couplings in the K\"ahler potential.
\subsection{Cosmological and astrophysical constraints}
\indent
The minimal SUSY model with explicit $R$ nonconservation contains
lepton-number violating interactions that can wash out any primordial
BAU generated at the GUT scale via the $B+L$-violating sphaleron
interactions~\cite{GTH,NM,KRS},
which are in thermal equilibrium above the critical temperature of the
electroweak phase transition~\cite{ADD}. Sphalerons generally conserve
the individual quantum numbers $B/3-L_i$~\cite{CDEO,DR}. In particular, it
has been shown in~\cite{DR} that
if only one separate lepton number is preserved in thermal equilibrium
({\em e.g.}\ $L_i$) and finite masses for the charged
leptons are taken into account in the analysis of chemical potentials, this is
then sufficient to protect any primordial excess in $L_i$, which can be
converted later on, via sphalerons, into the observed BAU. For our purposes,
we will assume that only one separate lepton number is conserved each time
in the full Lagrangian, when low-energy experiments are
considered. For definiteness, in our numerical analysis we will consider
that either $w_e=\varepsilon_e=0$ or $w_\tau=\varepsilon_\tau=0$. Of
course, one can use a complementary restriction and put
$w_\mu=\varepsilon_\mu=0$,
which, however, will not alter our phenomenological constraints discussed
in Sections~5.1--5.5 in an essential way.
There is a great number of bounds coming from astrophysics,
such as those obtained from the dynamics of red giants and white dwarfs,
or the absence of a distorted spectrum of the $2.73^\circ$ K
blackbody radiation background~\cite{GGR}. However, we find
more worrying the severe limits derived from possible reheating effects
of a decaying massive neutral relic with $m_\nu\simeq 10-40$
MeV and especially those obtained from the primordial
nucleosynthesis~\cite{EGLNS,EMR}.
In particular, $\tau$-neutrino decays with a lifetime bigger than about
1~sec or so may increase the elemental ${}^4$He abundance by making it
incompatible with astrophysical observations. Imposing the
latter constrain, we find
\begin{equation}
|\tilde{B}^L_{e\nu_\tau}|^2\ \stackrel{\displaystyle >}{\sim}\
10^{-4}\, \left(\frac{30\, \mbox{MeV} }{m_{\nu_\tau}}\right)^5\, ,
\label{Bastr}
\end{equation}
which is only applicable for $m_{\nu_\tau}
\stackrel{\displaystyle >}{\sim} 10-50$~MeV~\cite{EGLNS}.
In fact, the bound of Eq.~(\ref{Bastr}) is not so restrictive,
since it simply constrains only the mixing-matrix element
$V^l_{e\nu_\tau}>10^{-2}$ in Eq.~(\ref{BL}),
which is not excluded from solar neutrino oscillation scenarios.
In our analysis of laboratory observables, we sum up over all invisible
light neutrinos, so the unitary matrix $V^l$ becomes practically
redundant. Moreover, the nonobservation of a $\gamma$ ray burst from
the Solar Maximum satellite after the supernova 1987A neutrinos
were detected may point towards the fact that the $\tau$ neutrino mainly
decays inside the supernova core. This leads again to $\nu_\tau$ lifetimes
compatible with the approximate inequality of Eq.~(\ref{Bastr}).
Even though the predicted supernova luminosity will increase
in such a case, an allowed window of scenarios that maximally violate $L_\mu$
and $L_\tau$ may be present in the $\sim 3$ MeV neutrino-sphere~\cite{BBHH}.
As has also been pointed out by the authors in Ref.~\cite{BBHH},
there may exist viable cosmological models in which $\nu_\tau$ is
stable with a mass of ${\cal O}(10)$~MeV. Such a solution requires
an alteration of the standard cosmological picture by, {\em e.g.},
reheating the universe even after inflation to only a few MeV and
invoking low-temperature baryogenesis as well~\cite{DH}. Then, the resulting
$\nu_\tau$ may not overclose the universe but it can even constitute
the cold dark matter.
For neutrinos with $m_\nu\stackrel{\displaystyle <}{\sim} 0.1$~MeV,
the cosmological bound regarding their lifetimes, $\tau_\nu$,
is different. In fact, $\tau_\nu$ should not be larger the age
of the universe, {\em i.e.}\ $\tau_\nu \stackrel{\displaystyle >}{\sim}
10^{23} \displaystyle{\big(\frac{m_\nu}{1\, \mbox{eV}}\big)}$~sec~\cite{EMR}.
Furthermore, it is worth mentioning that radiative decays of massive
neutrinos with $0.1\stackrel{\displaystyle <}{\sim} m_\nu
\stackrel{\displaystyle <}{\sim} 10$~keV have also found some applications
in cosmology and astrophysics~\cite{DWS}. In this context, most noticeable
is probably the Gunn-Peterson test~\cite{GP}, {\em i.e.}, the search of
primordial elements in the intergalactic medium. There seems to be a
deficiency of neutral hydrogen and helium in the intergalactic
medium~\cite{GP}.
A source of photo-ionization of these elements might be a radiatively decaying
neutrino. Here, we simply comment on the fact that the mass range of neutrino
required for such an explanation is different from what is suggested by
the solar neutrino puzzle and the atmospheric neutrino problem. To ionize
singly ionized helium, $m_{\nu}$ should be bigger than 109~eV, since the
ionization potential is 54.4~eV. A recent investigation of this issue may
be found in~\cite{SKS}.
\setcounter{equation}{0}
\section{Numerical results}
\indent
In this section, we will present numerical predictions as well as
constraints on the basic parameters of our $\not\!\! R$ model,
which have been discussed in Section 5. Although there is a large number
of parameters that could vary independently, it is important to remark that
there exists a strong correlation between new-physics observables and light
neutrino masses. This seems to be a generic feature of most of the $R$-parity
broken SUSY models considered in the literature~\cite{BBHH,NRV}.
However, a novel feature of our minimal $\not\!\! R$ scenario is that
the size of the scalar-neutrino vev's and the $\varepsilon_i$ terms can,
in principle, be unconstrained. In fact, if $\vec{\Lambda}\simeq 0$ in
Eq.~(\ref{31}), which is a form of alignment in the flavour space between
the vev's of the sneutrinos, $\vec{w}$, and the $\not\!\! R$ terms,
$\vec{\varepsilon}$, this condition alone is sufficient to evade upper limits
on the tau-neutrino mass for any value of the SUSY parameters $M$, $M'$,
$\mu$, and $\tan\beta$.
In order to understand how all new-physics interactions are proportional
to $\Lambda_i$ and hence depend on $m_{eff}$ [or $m_{\nu_\tau}$] in
Eq.~(\ref{m31}) [Eq.~(\ref{m35})],
we evaluate the mixing matrix $\xi$ defined in Eq.~(\ref{xi}). Thus, we have
\begin{eqnarray}
\xi_{i1} &=& \frac{g'M\mu}{2\, \mbox{det}M_4}\, \Lambda_i\, ,\nonumber\\
\xi_{i2} &=& -\, \frac{g cM\mu}{2\, \mbox{det}M_4}\, \Lambda_i\, ,\nonumber\\
\xi_{i3} &=& \frac{\varepsilon_i}{\mu}\, +\, \frac{(cg^2+g'^2)Mv_2}{4\,
\mbox{det}M_4}\, \Lambda_i\, , \nonumber\\
\xi_{i4} &=& -\, \frac{(cg^2+g'^2)Mv_1}{4\, \mbox{det}M_4}\, \Lambda_i
\, ,\label{xii}
\end{eqnarray}
for $i=1\, (e),\, 2\, (\mu)$, and $3\, (\tau)$. It is now easy to see that
only the elements $\xi_{i3}$ contain the dominant contributions characterized
by being {\em not} proportional to $\Lambda_i$. However, these contributions
vanish identically in the relevant expression
\begin{equation}
\delta_{\nu_i\nu_j}\, -\, \tilde{C}_{\nu_i\nu_j}\ =\
[\tilde{\xi}^* (1-t_z) \tilde{\xi}^T ]_{\nu_i\nu_j},
\end{equation}
given in Eq.~(\ref{C}), since the element of the
diagonal matrix $(1-t_z)_{33}=0$. Consequently, in the limit
of vanishing $\tau$-neutrino mass, the invisible $Z$-boson width
predicted in our $\not\!\!\! R$ model will coincide with that found
in the SM.
Similar strong $m_{\nu_\tau}$ dependence occurs in the non-SM
part of the couplings $Zl_il_j$ and $Wl_i\nu_j$ via the mixing matrix
$\xi_{{}_L}$, which is given by
\begin{eqnarray}
(\xi^*_{{}_L})_{i1}=
\frac{g\, \Lambda_i}{\sqrt{2} (M\mu-\frac{1}{2}g^2v_1v_2)}\, ,
\nonumber\\
(\xi^*_{{}_L})_{i2}= \frac{\varepsilon_i}{\mu}\, -\,
\frac{g^2v_2\, \Lambda_i}{2\mu (M\mu-\frac{1}{2}g^2v_1v_2)}\, .\label{xiiL}
\end{eqnarray}
One can readily see that the dominant terms in $\xi_{{}_L}$
are contained in the
elements $(\xi_{{}_L})_{i2}$. However, in the $Zl_il_j$ coupling, the
new-physics
contributions are determined by
\begin{equation}
[\xi_{{}_L} (1-d) \xi^\dagger_{{}_L}]_{ij}\ =\ (\xi_{{}_L})_{i1}
(\xi^*_{{}_L})_{j1}\, ,
\end{equation}
and the elements $(\xi_{{}_L})_{i2}$ always get killed by the diagonal matrix
$(1-d)$. Thus, leptonic FCNC $Z$-boson decays and associated
universality-breaking effects are proportional to $\Lambda_i$ and are absent
if $\nu_\tau$ is massless. Moreover, we find that the non-SM contributions
present in the coupling $Wl\nu$ in Eq.~(\ref{BL}) are proportional to
\begin{equation}
\Big( -\xi_{{}_L}\xi_{{}_L}^\dagger\, -\,
\xi^*\xi^T\, +\, 2\xi_{{}_L}t_{{}_L}\xi^T\Big)_{ij}\ =\
-\, (\xi_{{}_L})_{i2}[(\xi^*_{{}_L})_{j2}-\xi_{j3}]\, -\,
\xi_{j3}[\xi^*_{i3}\, -\, (\xi_{{}_L})_{i2}]\, .\label{newW}
\end{equation}
Substituting Eqs.~(\ref{xii}) and (\ref{xiiL}) into Eq.~(\ref{newW}),
it is easy to verify that new-physics effects in
charged-current interactions are also very strongly correlated
with the light neutrino mass $m_\nu$.
For reasons mentioned above, we will work in the seesaw approximation
by keeping the mass of $\nu_\tau$ finite. For our illustrations, we will
consider the following modest $\not\!\! R$ SUSY scenarios:
\begin{equation}
\begin{array}{r|r|rrrr}
\mbox{Scenario}& (\mbox{type of line})
& \tan\beta & M\, \mbox{[GeV]} & \mu\, \mbox{[GeV]} &\varepsilon_\mu\
(\mbox{or}\ \varepsilon_\tau )\, \mbox{[GeV]} \\
\hline
\mbox{I} &\mbox{(solid)} & 1 & 50 & 500 & 0 \\
\mbox{II} &\mbox{(dashed)} & 1 & 50 & -50 & -0.5 \\
\mbox{III} &\mbox{(dotted)} & 1 & 100 & 200& 1 \\
\mbox{IV} &\mbox{(dash--dotted)} & 4 & 200 & 400& 2 \\
\end{array} \label{A-scenarios}
\end{equation}
where $M'=M/2$ and the type of line used in our plots is also
indicated.
First, we will study possible limits on the $\not\!\! R$
models in Eq.~(\ref{A-scenarios}) that may be derived by the
nonobservation of a muon decay into three electrons. In Fig.~1(a), numerical
predictions for $B(\mu^-\to e^-e^-e^+)$ as a function of $m_{\nu_\tau}$
are displayed for $w_\mu/w_e=1$.
The horizontal dotted line indicates the present experimental
limit. Fig.~1(a) also shows the strong quadratic dependence of
$B(\mu^-\to e^-e^-e^+)$ on $m_{\nu_\tau}$. In particular,
if $w_\mu$ and $w_e$ are comparable in size ({\em e.g.}, $w_\mu/w_e=1$),
this constraint is more severe. Qualitatively, we find
that
\begin{equation}
\frac{w_ew_\mu}{w_e^2+w_\mu^2}\, \frac{m_\nu}{M}
\ \stackrel{\displaystyle <}{\sim}\ 10^{-6}\, .\label{bound}
\end{equation}
Of course, this limit gets relaxed for large vev ratios $w_\mu/w_e$.
The bound derived from $B(\mu^-\to e^-e^-e^+)$ is more sensitive to
the soft-SUSY gaugino mass $M$. To be more precise, our analysis
yields the following upper limits on $m_\nu$:
\begin{equation}
\begin{array}{r|r}
\mbox{Scenario}& m_\nu\, \mbox{[MeV]} \\
\hline
\mbox{I} & 0.20 \\
\mbox{II} & 0.57 \\
\mbox{III} & 0.43 \\
\mbox{IV} & 0.89 \\
\end{array} \label{A-bound}
\end{equation}
at 90$\%$ CL. We also remark that $\tau$-lepton number
is assumed to be conserved so as to protect a primordial excess in $L_\tau$
from being erased by processes that are in thermal equilibrium.
It is then obvious that for scenarios with $w_\mu/w_e=1$ and $M=200$ GeV,
$m_{\nu_\tau}<0.9$ MeV. From~(\ref{A-bound}), we see that scenario I
gives a stronger limit than the experimental one on the mass of $\nu_\mu$,
which is currently $m_{\nu_\mu}<0.27$~MeV at 90$\%$ CL~\cite{PDG94}.
As the non-SM couplings depend crucially
on the $\tau$-neutrino mass, the less than 1 MeV upper bound on a massive
neutrino gives little chance to see new-physics effects in other observables.
However, if $\Delta L_e=0$ in the model, {\em i.e.}\ $w_e=\varepsilon_e=0$,
inequality~(\ref{bound}) is trivially fulfilled and the so-derived neutrino
mass bound does not apply any longer.
In Fig.~1(b), numerical estimates reveal that non-SM contributions
to the invisible $Z$-boson width are one order of magnitude smaller
than the present experimental sensitivity. As a result, experimental
searches for physics beyond the SM, based solely on neutrino counting at
the $Z$ peak, are bound to be inadequate to unravel the nature of our minimal
$\not\!\! R$ model.
{}From Fig.~2(a), it can be seen that our minimal $\not\!\! R$
model may predict universality-breaking effects via the observable
$U_{br}$ in excess of $10^{-3}$. Such new-physics phenomena might
be seen at LEP, if all the experimental data accumulated in the year
1995 are analyzed.
Furthermore, in Fig.~2(b), we give theoretical predictions for the
observable $R_\pi/R_\pi^{SM}-1$ given in Eq.~(\ref{Rpiexp}). Possible
deviations from lepton universality in charged-current interactions
turn out to be one order of magnitude smaller than those that
can be accessed in experiment. Also, beyond the realm of detection
are found to be possible violations of charged-current universality
in the decays $\tau\to e \nu\nu$ and $\mu\to e \nu\nu$, which are
measured by virtue of the physical quantities $R_{\tau e}$ and
$R_{\tau\mu}$. Theoretically, similar is predicted to be the
situation for the size of the FCNC $Z$-boson mediated decays, such as
$\tau^-\to \mu^- e^- e^+$ and $Z\to ll'$. More explicitly,
it is estimated that
\begin{eqnarray}
B(\tau^-\to \mu^- e^- e^+) &\stackrel{\displaystyle <}{\sim}&
1.\, 10^{-9}\, ,\nonumber\\
B(Z\to l^-l'^+\ \mbox{or}\ l^+l'^-) &\stackrel{\displaystyle <}{\sim}&
1.\, 10^{-8}\, ,\nonumber\\
1\, -\, \frac{R_{\tau e}}{R_{\tau e}^{SM}} &\stackrel{\displaystyle <}{\sim}&
1.\, 10^{-4}\, ,\nonumber\\
1\, -\, \frac{R_{\tau\mu}}{R_{\tau\mu}^{SM}} &\stackrel{\displaystyle <}{\sim}&
1.\, 10^{-4}\, .
\end{eqnarray}
There may also be other places where $R$-parity violation could
manifest its presence. Of course, if neutralinos are lighter than the
$Z$ boson, one could search for distinctive signatures caused by decays
of the form $Z\to \nu_\tau \chi^0$ or $\tau^\pm \chi^\mp$, where
$\chi^0$ and $\chi^\pm$ decay subsequently into two $b$-quark jets
accompanied by a large amount of missing mass~\cite{BBHH}. However,
if the production threshold of heavy neutralinos and charginos is above
the LEP centre of mass energy, one then has to rely on studies of
possible indirect non-SM signals via sensitive observables
devoid of ambiguities coming from the evaluation of hadronic matrix
elements, as those discussed in Sections 5.1--5.4.
In the same logic, $R$-parity violating effects may also be probed
in the $\nu_\mu e$ scattering, even though experimental data
do not impose very stringent constraints as compared to those resulting
from $B(\mu\to e^-e^-e^+)$~\cite{BGH,BGKLM}.
Since our minimal $\not\!\! R$ model
only modifies the leptonic sector, one may derive useful constraints
from atomic parity violation measurements of the effective `weak charge',
$Q_W$, of a heavy nucleus. In the case of $^{133}_{55}$Cs, one
has~\cite{BGKLM}
\begin{equation}
Q_W^{exp}(^{133}_{55}\mbox{Cs})-Q_W^{SM}(^{133}_{55}\mbox{Cs})\ =\
73.5\cdot [\tilde{\xi}_L (1-d) \tilde{\xi}_L^\dagger]_{11}\ \le\ 3.74
\, ,\label{QW}
\end{equation}
at 1$\sigma$. The above bound turns out to be rather weak when
compared to that derived from $B(\mu\to eee)$.
Finally, for reasons that have already been mentioned in Section 5.3,
possible limits obtained directly from forward-backward-asymmetry
observables similar to $\Delta{\cal A}_{ll'}$ are estimated to be much
weaker than those determined
by the universality-breaking parameter $U_{br}^{(ll')}$ in Eq.~(\ref{Ubr}),
and are therefore not taken into consideration here.
\section{Conclusions}
\indent
The minimal $R$-parity broken SUSY model contains bilinear
lepton-number-violating terms ($\varepsilon_i$), which cannot in general
be eliminated
by a re-definition of the superfields provided soft-SUSY breaking parameters
are simultaneously present in the superpotential. The consideration of these
$\varepsilon_i$ mass terms, which involve the chiral multiplets of the
left-handed leptons and the Higgs field with $Y=+1$, give rise naturally
to non-vanishing vev's, $w_i$, of the
scalar neutrinos after the spontaneous breaking of the gauge symmetry.
In particular, if the vectors $\vec{w}$ and $\vec{\varepsilon}$,
spanned in the flavour space, satisfy a kind of alignment relation,
$\vec{\Lambda}=0$, forced, {\em e.g.}, by some horizontal symmetry, the
afore-mentioned $w_i$ and $\varepsilon_i$ parameters are not restricted
by limits on the $\tau$-neutrino mass. Furthermore, constraints from primordial
nucleosynthesis and the observed BAU have been considered. Specifically,
to evade BAU constraints has been sufficient to impose that at least
one separate leptonic number has to be conserved in our $\not\!\! R$ model,
{\em e.g.}\ $w_\tau=\varepsilon_\tau=0$ and $w_e=\varepsilon_e=0$.
Our main interest has been to investigate the phenomenological implications
of this novel $\not\!\!\! R$ model in the light of a number of terrestrial,
astrophysical, and cosmological constraints.
To be more concrete, we have considered a typical set
of $\not\!\! R$ models as is stated in (\ref{A-scenarios}) and
confronted it with results obtained from LEP, CLEO and other
experiments. We have found that the resulting non-SM contributions to
the couplings $Z\nu\nu$, $Zll'$, and $Wl\nu$ show a strong correlation
with the $\tau$-neutrino mass and vanish in the massless limit. This
direct correlation between the size of $R$-parity-violating phenomena
and the magnitude of the neutrino mass appears to be a generic feature
of most of the $R$-parity broken models considered in the
literature~\cite{BBHH,NRV}. In our analysis, the most severe
constraint comes from $B(\mu\to eee)$ for $\not\!\! R$ scenarios,
where $\Delta L_\tau=0$, and $L_e$ and $L_\mu$ are maximally violated.
In this way, we have been able to set an upper bound on $m_{\nu_\tau}$
by means of Eq.~(\ref{bound}). For instance, for $M=\mu=2M_W$ and
$w_e=w_\mu$, we find that $m_{\nu_\tau}\stackrel{\displaystyle <}{\sim}
1$ MeV. Especially, for scenario I in Eq.~(\ref{A-scenarios}),
we have $m_\nu < 0.2$~MeV as has been given in Eq.~(\ref{A-bound}),
which is even tighter than the current experimental bound on the mass of
the $\mu$ neutrino.
The remaining observables leave the main bulk of the parameter
space unconstrained. The most encouraging prediction is obtained
for the universality-violating observable $U_{br}$, with
$U_{br}\stackrel{\displaystyle <}{\sim} 2.\, 10^{-3}$. Such phenomena
might be seen at LEP, when the analysis of all the data of the year 1995
is completed.
For our purposes, we need not study the combined effect of the trilinear
$R$-parity-violating couplings $\lambda$ and $\lambda'$, {\em i.e.}\
$\lambda_{ijk}=\lambda'_{ijk}=0$.
The reason is that the Yukawa couplings $\lambda_{ijk}$ and $\lambda'_{ijk}$
are not sufficient to explain possible new-physics phenomena that can be
shown up in certain low-energy processes and LEP observables,
such as $B(l^-\to l'^-l_1^-l^+_1)$, $B(Z\to ll')$, and $U_{br}$,
discussed in Sections~5.1--5.3. In this context, we
remark that the KARMEN anomaly can, in principle, be explained by
assuming the presence of a fourth light neutralino, even though an
unnaturally small value of $\mu={\cal O}(10)$~MeV may be required.
\vskip1cm
\noindent
{\bf Acknowledgements.} The authors gratefully acknowledge discussions
with Roger Phillips. M.N. would like to thank the theory group of
Rutherford Appleton Laboratory for hospitality extended to him during
a visit, when part of this work was done.
M.N. also gratefully acknowledges financial support by
the HCM program under EEC contract no. CHRY-CT 920026.
\newpage
\setcounter{section}{0}
\def\Alph{section}.\arabic{equation}{\Alph{section}.\arabic{equation}}
\begin{appendix}
\setcounter{equation}{0}
\section{Appendix}
\indent
For completeness, we present that part of the
scalar potential (\ref{m9}) that contains the charged singlet fields
$E_i$. This also consists of a lepton-number conserving contribution
($V_+^{L}$) and a lepton-number violating one ($V_+^{\not L}$).
The former reads
\begin{eqnarray} \label{a1}
V_+^{L}&=&[\mu^2_{+ij}(E_i^* E_j) + \mbox{H.c.}] +
[\mu_{ij}(\phi_2^{\dagger}\varphi_i)E_j + \mbox{H.c.}]
+[\mu'_{ij}(\phi_1^{\dagger}\varphi_i)E_j + \mbox{H.c.} ]\nonumber \\
&&+\lambda(\sum_k E_k^* E_k)^2 +(\tilde{\kappa}_{jk} - \lambda \delta_{jk})
(\phi_1^{\dagger}\phi_1)(E_j^* E_k) +
\lambda(\phi_2^{\dagger}\phi_2)(E_k^{*} E_k) \nonumber \\
&&+ (\mu_{ijnm} - \lambda \delta_{in}\delta_{jm})(\varphi_i^{\dagger}
\varphi_n)(E_j^{*} E_m) + 4\tilde{\kappa}_{nmij}(\varphi_n^{\dagger}
\varphi_i)(E_m^{*} E_j)\, ,
\end{eqnarray}
where we have defined
\begin{eqnarray} \label{a2}
&& \lambda = {1 \over 2} g'^2\, , \nonumber \\
&& \mu_{ij} = \mu^* h_{ij}\, , \nonumber \\
&& \tilde{\kappa}_{jk}= h_{ij}^* h_{ik}=\tilde{\kappa}_{kj}^*\, ,
\nonumber \\
&& \mu_{ijnm}= h_{ij}^* h_{nm}= \mu_{nmij}^* \, ,\nonumber \\
&& \tilde{\kappa}_{nmij}=\lambda_{knm}^* \lambda_{kij}=
\tilde{\kappa}_{ijnm}^*\, .
\end{eqnarray}
In Eq.~(\ref{a2}), $h_{ij}$ and $\lambda_{ijk}$ are couplings from the
superpotential (\ref{m3}) and (\ref{m4}), respectively.
$\mu_{+ij}^2$ and $\mu'_{ij}$ are soft-SUSY breaking parameters
from Eq.~(\ref{m8}).
For the lepton-number-violating contribution, we obtain
\begin{equation} \label{a3}
V_+^{\not L}= -2i\tilde{\kappa}_{ijk}(\phi_1^T \tau_2 \varphi_j)(E_i^*
E_k) + \kappa'_{ijk}(\varphi_i^T \tau_2 \varphi_j)E_k + \mbox{H.c.}\, ,
\end{equation}
where
\begin{equation} \label{a4}
\tilde{\kappa}_{ijk} = h_{ni}^* \lambda_{njk}\, ,
\end{equation}
and $\kappa'_{ijk}$ is a soft-SUSY breaking parameter contained in
Eq.~(\ref{m8}).
By analogy with Eq.~(\ref{m20}), from $V_+^L + V_+^{\not L}$, we can derive
conditions for not having CP violation in this part of the potential.
These conditions are listed below
\begin{eqnarray} \label{a5}
&& \mu^*_{ijnm}\eta^*_{{}_{L_n}} \eta_{{}_{L_i}} \eta^*_{{}_{+m}}
\eta_{{}_{+j}}= \mu_{ijnm}\, , \nonumber \\
&& \tilde{\kappa}^*_{jk} \eta^*_{{}_{+k}} \eta_{{}_{+j}}=\tilde{\kappa}_{jk}
\, ,\nonumber \\
&&\mu^*_{ij}\eta_{{}_{2}} \eta^*_{{}_{L_i}} \eta^*_{{}_{+j}}=\mu_{ij}
\, ,\nonumber \\
&&\mu'^{*}_{ij}\eta_{{}_{1}} \eta^*_{{}_{L_i}} \eta^*_{{}_{+j}}=\mu'_{ij}
\, ,\nonumber \\
&&\tilde{\kappa}^*_{nmij}\eta^*_{{}_{L_n}} \eta_{{}_{L_i}} \eta^*_{{}_{+m}}
\eta_{{}_{+j}}= \tilde{\kappa}_{nmij}\, , \nonumber \\
&&\tilde{\kappa}^*_{ijk}\eta^*_{{}_{1}} \eta^*_{{}_{L_j}} \eta_{{}_{+i}}
\eta^*_{{}_{+k}}= \tilde{\kappa}_{ijk}\, , \nonumber \\
&& \kappa'^{*}_{ijk}\eta^*_{{}_{L_i}} \eta^*_{{}_{L_j}} \eta^*_{{}_{+k}}
=-\kappa'_{ijk}\, ,
\end{eqnarray}
where summation convention is not implied.
In Eq.~(\ref{a5}), $\eta_{{}_{+j}}$ are the CP phases of the scalar fields
$E_j$, similar to the notation of Eq.~(\ref{m20}). In general, both sets,
(\ref{m20}) and (\ref{a5}), should not be viewed independently of one another.
For instance, using the equalities in Eqs.~(\ref{m20}) and (\ref{a5}),
one can derive
(no summation convention)
\begin{eqnarray} \label{a6}
&& \Im m(\lambda_6 \mu_{ij}\mu'^{*}_{nm}\tilde{\kappa}_{mj}\kappa_{in})=0
\, ,\nonumber \\
&& \Im m(\lambda_6 \mu^*_{ijnm} \mu_{nm}\mu'^{*}_{ij})=0
\, ,\nonumber \\
&& \Im m(\mu^*_{ijnm}\kappa_{ni}\tilde{\kappa}^*_{mj})=0\, ,
\end{eqnarray}
and many similar relations of this kind.
\end{appendix}
\newpage
|
\section{Introduction}
\label{introduction}
\cite{sandt:presupdrt} introduces \cite{lewis}' notion of {\em
accommodation\/} in Discourse Representation Theory \cite{KR:DtoL} as a
tool to account for gaps in the discourse. His theory of presupposition
projection takes presuppositions to behave like anaphora. Anaphoric
expressions normally are linked to antecedents that have previously been
established in the discourse. If example~\ref{kingoffrance} would
appear in a context where no {\sl king of France} is present -- hence no
antecedent is available -- then Van der Sandt's algorithm {\em
accommodates\/} the existence of a king of France.
\begin{examples}
\ex\exlab{kingoffrance}
When I give a party, the king of France always attends it.
\end{examples}
This is different from the situation where a definite description can be
{\em linked\/} to an antecedent that was previously introduced by an
indefinite description, as in~\ref{celebrity}. There is no need to
accommodate an antecedent, because there is already a suitable candidate
available.
\begin{examples}
\ex\exlab{celebrity}
When I invite a celebrity, the celebrity never comes.
\end{examples}
Example~\ref{barkeeper} however is slightly different. There is no
actual antecedent for the anaphoric expression {\sl the barkeeper}, but
because of {\sl a bar}, there isn't really a problem, apparently there
is some implicit antecedent. Van der Sandt's projection algorithm fails
to make this implicit link, and accommodates the existence of {\sl a
barkeeper} to the global context, in fact no theory on presupposition
that we know of can deal with these\footnote{
The closest comes probably Beaver's dynamic theory of
presupposition \cite{beaver}. }.
\begin{examples}
\ex\exlab{barkeeper}
When I go to a bar, the barkeeper always throws me out.
\end{examples}
Contrasting~\ref{barkeeper} with~\ref{playground} makes our point even
clearer; this sentence sounds truly infelicitous. The hearer tries to
somehow link this {\sl barkeeper} with familiar information, and fails.
\begin{examples}
\ex[?]\exlab{playground}
When I go to a playground, the barkeeper always throws me out.
\end{examples}
{\sl A bar} provides sufficient information to license {\sl the
barkeeper}, but in {\sl a playground} there is nothing that can establish
such a link. Making a link between the new discourse referent (i.e. {\sl
the barkeeper}) to the network of discourse referents that is already
established, is called {\em bridging\/} (\cite{bridging},
\cite{heim:diss}). Definite descriptions that can be bridged to existing
information do not need the accommodation of new referents;
example~\ref{kingoffrance} requires accommodation, but \ref{barkeeper}
can be solved with mere bridging. An adequate theory of presupposition
obviously needs a serious explanation of bridging to account for the
projection problem of presupposition.\\
To account for these phenomena, we borrow from \cite{pus:GL} and compare
{\em bridging\/} with {\em coercion\/}. Pustejovsky presents examples
like~\ref{beginabook}:
\begin{examples}
\ex\exlab{beginabook}
I would like to begin a new book tonight.
\end{examples}
Here, too, some information is missing: {\sl begin} implies some event,
but {\sl a new book} is an artifact. The fact that the speaker should be
interpreted as {\sl beginning to read the book}, or -- if he is a writer
-- {\sl to write} one, is motivated by what we know about {\sl book}.
Pustejovsky claims that such information should be considered lexical
knowledge of the noun, which is represented in a so-called {\em qualia
structure\/}. Based on this information, arguments of improper types can
be {\em coerced\/} to proper ones. We will see how a similar approach
can be followed to account for {\em bridging\/}.
In section~\ref{qscoercion} we will present Pustejovsky's ideas in more
detail, explaining concepts like {\em coercion\/} and {\em qualia
structure\/}. In Pustejovsky's work these ideas only get applied on the
sentence level. Section~\ref{bridging} will show how the ideas of Van
der Sandt and Pustejovsky fit together very nicely, even complementing
each other and we will show that {\em bridging\/} operates intra- as well
as inter-sentential. In section~\ref{examples} we will present some
examples of {\em linking\/}, {\em bridging\/} and {\em accommodation\/},
and in section~\ref{functionalcomposition} we will discuss the notion of
functional composition and coercion in this model.
\section{Qualia Structure and Coercion}
\label{qscoercion}
\subsection{Qualia Structure}
\label{qscoercionqs}
In \cite{pus:GL} and subsequent papers the notions of {\em coercion} and
{\em qualia structure} have been introduced. Qualia structure can be seen
as a set of lexical entailments. For instance, the word {\sl book}
entails at least the two events of {\sl reading} and {\sl writing} it,
besides the knowledge that it consists of several separate parts, like
the {\sl cover}, {\sl pages}, etc. Pustejovsky suggests four {\em qualia
roles} to represent such knowledge: {\sc formal}, {\sc constitutive},
{\sc telic} and {\sc agentive}. In \cite{pus:GL} these have been
defined as follows\footnote{
In more common AI-related terms we could rephrase them as:
{\sc formal} - {\sc isa}, {\sc constitutive} - {\sc part-of~/~hasa}, {\sc
telic} - {\sc purpose} and {\sc agentive} - {\sc
cause}.}:
\begin{itemize}
\item {\sc formal}: That which distinguishes the object within a larger
domain.
\item {\sc constitutive}: The relation between an object and its
constituents or proper parts.
\item {\sc telic}: Purpose and function of the object.
\item {\sc agentive}: Factors involved in the origin or "bringing about"
of an object.
\end{itemize}
\label{important}
The exact structure of this kind of lexical semantic knowledge seems to be
very intricate. Again for the same example, it is important for instance
to realize that a {\sl book} is at the same time a {\em physical object}
and an {\em information container}. The first description considers the
physical viewpoint, whereas the second defines the conceptual angle of
what constitutes our idea of a {\sl book}. What angle one takes ({\em
physical object} or {\em information container}) has immediate consequences
for the knowledge that is represented in the rest of the qualia structure.
The composing parts of the physical side of a book ({\sl pages}, {\sl cover},
etc.) are different from those of the conceptual side ({\sl title}, {\sl
sections}, {\sl paragraphs}, etc.). The same goes for the representation
of typical events a book is involved in. The physical 'quality' (qual) of
a book can be {\sl printed}, {\sl typeset} or even {\sl shelved}. The
information 'quality' can be said to undergo the events of {\sl reading},
{\sl writing} as mentioned before. It is however undeniable that the
two main qualities of book, along with all their entailments, are intimately
related to each other and should be represented accordingly in one
comprehensive (qualia) structure.
\subsection{Coercion}
The need for a rich lexical semantic knowledge representation like qualia
structure becomes clear in considering sentences like \ref{beginabook}
above, which is repeated here:
\begin{examples}
\ex\exlab{beginabook2}
I would like to begin a new book tonight.
\end{examples}
As mentioned before, the verb {\sl begin} expects an event here but has to
settle for an {\sl artifact} (book). We can now use the qualia structure of
this {\sl artifact} to infer some {\sl event} that is entailed by it and
which can stand in its place. This is an example of what Pustejovsky has
called {\em metonymic reconstruction} \cite{pus:GL} for cases where an
interpretation can be inferred from some partial meaning of the word in
question. In more general terms, anytime a word or phrase is not of the
desired type\footnote{
Possibly this use of the term {\em type} is not appropriate and
we should use {\em sort} instead. However, here we present the
terms as they have been defined by \cite{pus:GL}.}
(like {\sl artifact}, {\sl event}, etc.) we are allowed to {\em coerce} it
into one of its entailments that is of the appropriate type, where the
entailments are stored in its qualia structure. Another example of this is
the following sentence:
\begin{examples}
\ex\exlab{announcenewmodel}
BMW announced a new model.
\end{examples}
Here the verb {\sl announce} is looking for a subject of type {\em animate}
while only one of type {\em institute} is available. The qualia structure of
any {\em institute} however should represent the fact that they are made up
of people, which are {\em animate} entities. So, in this sentence we can
infer that some human at the BMW company did the actual announcement.\\
This summarizes Pustejovsky's program as described in \cite{pus:GL} and
subsequent papers. In this paper we extend coercion with the notion of
context, which seems not only a valid research topic but also
desperately needed because of the restricted explaining power of
coercion if context is not considered. Take for instance
sentence~\ref{beganabook} :
\begin{examples}
\ex\exlab{beganabook}
John began a book.
\end{examples}
Although above we assumed several times that one can infer {\sl read} and
{\sl write} events from the qualia structure of {\sl book} in order to make
this sentence semantically well formed, this can only be a default
approximation. We would need an actual context for this sentence to decide
what event exactly should be inferred. Imagine for instance a dinner,
organized by the {\sc literary and culinary society}, where all dishes are
shaped in the form of books$\ldots$\footnote{
Still another problem concerning the lack of context is
illustrated by the following examples where no argument
at all is available for coercion to take place, {\sl Monday} and
{\sl yesterday} are modifiers:
\begin{quote}
I propose Monday.\\
I began yesterday.
\end{quote}
Pustejovsky (personal communication) has termed this
loosely as {\em null coercion}, because although coercion
should take place it cannot be executed properly. Taking
context into account could be of help however to make the
sentences sound more natural, as the following examples show:
\begin{quote}
Let's make an appointment. I propose Monday.\\
Let's play darts. I begin.
\end{quote}
It seems that null coercion should coerce an anaphor
which is of the required type. In both examples this would
be event-type anaphors. }
This example is farfetched, but it may make the point more clear. We do
not assume that the qualia structure of {\sl book} should contain any
reference to this particular example. It is important however to realize
that any {\sl artifact} entails by default a number of events in which
it is engaged. In this particular context these events would be
overruled.
\section{Bridging in DRT with Qualia Structure}
\label{bridging}
This section shows how we deal with anaphora resolution in general,
and particularly bridging, in a version of DRT which uses extensively
{\em qualia\/} information. We define the language of
Discourse Representation Structures (DRSs) of our extended DRT,
show how resolution works, and finally give some detailed examples.
\subsection{A Sketch of the Architecture}
Basically, we extend Van der Sandt's theory of presupposition
with the notion of bridging anaphora. In short, Van der Sandt
views presupposition as anaphora with more descriptive content,
and uses one and the same mechanism for dealing with
both phenomena \cite{sandt:presupdrt}. Anaphoric information
can either be resolved to an antecedent that is available
from discourse, or if no antecedent is found, be accommodated.
We add a possibility of bridging to the resolution algorithm.
The basic architecture of the system is:
\begin{enumerate}
\item parse sentence: result is a sentence-DRS
\item merge sentence-DRS with main-DRS
\item perform anaphora resolution
\end{enumerate}
A {\em sentence-DRS} is a DRS with all anaphoric information
unresolved, and is the result of a bottom-up driven semantic
construction dependent on some syntactic structure.
A sentence-DRS can be viewed as a sort of under-specified
logical form with respect to anaphoric information.
Special types of DRSs ($\alpha$-DRSs) mark anaphoric
information. The {\em main-DRS} is the DRS of the context
interpreted so far. It is a {\em proper} DRS, i.e., a DRS with
no unresolved anaphoric information. Proper DRSs can be
interpreted as in standard DRT: they are {\em true} with respect to
a certain model if they can be {\em embedded\/} in that model
(\cite{KR:DtoL}. Before we explain how anaphora resolution works we
define DRSs and the merging operation.
\subsection{Discourse Representation Structures}
Let's introduce some terminology.
Discourse markers are variables ranging over objects in
the domain. Terms are either discourse markers or
DRSs. Furthermore,
we adopt a typed lambda-calculus for DRSs \cite{tilburg,muskens:cdrt}.
DRSs are defined as follows:
\begin{quote}
{\bf Definition 1. DRS}\\
If $U$ is a set of discourse markers,
$C$ is a set of DRS-Conditions, and t$_1$,...,t$_n$ terms, then
$<U,C>$ is a DRS,
$<U,C> \oplus <U^\prime,C^\prime>$ is a DRS,
$\lambda$ t$_1$,...,t$_n$. $<U,C>$ is a DRS. Nothing else is a DRS.
\end{quote}
\begin{quote}
{\bf Definition 2. DRS-Conditions}\\
If x$_1$,..,x$_n$ are discourse markers, P an n-place
condition, K and K$_1$ DRSs, then
P(x$_1$,...,x$_n$),
x$_1$ = x$_2$,
K $\to$ K$_1$,
$\lnot$ K,
K $\lor$ K$_1$,
$\alpha$:K, and
Q:K are DRS-Conditions. Nothing else is a DRS-Condition.
\end{quote}
The first five DRS-Conditions we already know from standard DRT
\cite{KR:DtoL} and need no further explanation. So called $\alpha$-DRSs
represent unresolved anaphoric information. DRSs that contain
$\alpha$-DRSs are therefore unresolved DRSs. Q-DRSs represent qualia
structure, with Q$_F$ for {\sl formal}, Q$_C$ for constitutive, Q$_A$
for agentive and Q$_T$ for telic. For notational purposes we use
$\cal{Q}$ to represents a {\em set\/} of qualia-DRSs\footnote{
As mentioned in section~\ref{qscoercionqs} on
page~\pageref{important}, the distinction that is made in the
formal role carries through in all other qualia roles. This
could be represented by embedding Q$_C$, Q$_A$ and Q$_T$ in
Q$_F$. This is beyond the scope of our paper.}. Now for merging:
\begin{quote}
{\bf Definition 3. Merging ($\oplus$)}.\\
$<U_1,C_1> \oplus <U_2,C_2> = <U_1 \cup U_2, C_1 \cup C_2>$
\end{quote}
The merge operation takes two DRSs and makes a union of the
sets of discourse markers and a union of the sets conditions.
Merging of DRSs is used both for constructing DRSs
(cf. \cite{tilburg}) and {\em coercive accommodation}.
The latter term brings us to the next definition.
Qualia-information, represented
in Q-DRSs is normally not
accessible and does not affect the truth-conditions of a DRS.
It is introduced in the lexicon and brought into discourse via
the DRS bottom-up construction algorithm. If necessary, for example
to play the role of antecedent, the
qualia structure is put forward to the surface by a process we
call {\em coercive accommodation}. It is defined as a function from
DRSs to sets of DRSs:
\begin{quote}
{\bf Definition 4. Coercive Accommodation (CA).}\\
{\sc ca}($<U,C>$) = $\{ <U,C> \oplus$ K $|$ Q:K $\in C\}$
\end{quote}
Note that CA is always local: it cannot accommodate qualia
information which is embedded. Note also that we have defined
CA only for DRS without lambda's: this will do for the
purposes of this paper.
Q-DRSs are also used for {\em type coercion}, which is
discussed later on in this paper.
In DRT the structure of DRSs restricts the choice of possible
antecedents of an anaphoric construction. For a discourse marker
to be the antecedent for an anaphor, it must be {\em accessible}
from the DRS which the anaphor is represented.
To define accessibility of DRSs and discourse markers
we first use the notion {\em subordination} between DRSs.
We adopt the notation C(K) meaning the set of conditions of
DRS (K), and U(K) meaning the set of discourse markers of K.
\begin{quote}
{\bf Definition 5. Subordination.}\\
If K$_1$, K$_2$, and K$_3$ are DRSs, then K$_2$ is subordinated
to K$_1$ (or K$_1$ subordinates K$_2$) if
K$_1$ $\oplus$ K$_2$,
K$_1$ $\to$ K$_2$ $\in$ C(K$_3$),
K$_2$ $\to$ K$_3$ $\in$ C(K$_1$),
$\lnot$ K$_2$ $\in$ C(K$_1$),
K$_2$ $\lor$ K$_3$ $\in$ C(K$_1$),
K$_3$ $\lor$ K$_2$ $\in$ C(K$_1$),
$\alpha$:K$_2$ $\in$ C(K$_1$),
Q:K$_2$ $\in$ C(K$_1$), and
K$_2$ is subordinated to K$_3$ and K$_3$ is subordinated to K$_1$.
\end{quote}
So, if x is a discourse marker and K$_1$ and K$_2$ are DRSs,
and x is in the domain of K$_1$ (x $\in$ U(K$_1$)),
then x is accessible from K$_2$ if K$_2$ is subordinated to K$_1$.
\subsection{Anaphora Resolution}
Left to explain is how anaphora resolution works.
We repeat for convenience that resolution can take place in three
different ways:
\begin{enumerate}
\item resolution to an accessible, suitable discourse marker ({\em linking})
\item resolution to coercively accommodated material of an
accessible DRS ({\em bridging})
\item accommodation of the anaphoric information to an
accessible DRS ({\em accommodation})
\end{enumerate}
We introduced accessibility already,
but haven't explained yet
the notion of `suitable' discourse marker, or
better: suitable DRSs. Suitability is an extra constraint
on the choice of antecedent. A DRS is suitable to another
DRS if there is a way you find a match between discourse
markers and conditions between both. More formally:
\begin{quote}
{\bf Definition 6. Suitability.}\\
A DRS K$_2$ is $m$-suitable to DRS K$_1$ if there is a mapping
$m$ such that scope($m$)=U(K$_2$) and for every x it is the
case that m(x) $\in$ U(K$_1$) and there is a DRS K$_3$ such that
C(K$_3$) $\subseteq$ C(K$_1$) if U(K$_3$) = $\{$ m(x) $|$ x $\in$
U(K$_1$)$\}$.
\end{quote}
We now introduce the heart of the system: anaphora resolution.
This algorithm works as follows. All anaphoric information in
the main-DRS (of course after merging it with the sentence-DRS of
the last processed sentence) is resolved. This information
is clearly marked because these are just our $\alpha$-DRSs. Resolution
either unifies this material with a suitable antecedent or
accommodates it, and as a result, $\alpha$-DRSs disappear.
After resolving all $\alpha$-DRSs, we are left with a proper-DRS,
a DRS which is fully specified with respect to anaphoric information.
This DRS is model-theoretically interpretable, as in standard DRT.
To describe the component,
we use K$_\alpha$ to indicate anaphoric DRSs, and K$_m$ for
the main-DRS. Definition 7 describes a function that takes a
certain main-DRS and a certain $\alpha$-DRS from it, and
returns a set of DRSs (since there could be more
than one possible antecedent or accommodation site)
with this $\alpha$-DRS resolved. The output of this function
can be fed back into the same function until all anaphoric
information is resolved (all $\alpha$-DRS have been
consumed).\footnote{The order of which resolution of anaphoric structure
takes place is important as well. We don't pay any attention to this,
but see \cite{sandt:presupdrt}.}
\begin{quote}
{\bf Definition 7. Anaphora Resolution.}\\
\begin{tabular}{ll}
{\sc ar}(K$_\alpha$,K$_m$)
& = $\{$ K' $|$ K' $\in$ {\sc link}(K$_\alpha$,K$_m$) $\}$ iff
$|${\sc link}(K$_\alpha$,K$_m$)$|$ $>$ 0\\
& = $\{$ K' $|$ K' $\in$ {\sc bridge}(K$_\alpha$,K$_m$) $\}$ iff
\parbox[t]{44mm}{$|${\sc link}(K$_\alpha$,K$_m$)$|$ $=$ 0 and\\
$|${\sc bridge}(K$_\alpha$,K$_m$)$|$ $>$ 0}\\
& = $\{$ K' $|$ K' $\in$ {\sc acc}(K$_\alpha$,K$_m$) $\}$ iff
\parbox[t]{44mm}{$|${\sc link}(K$_\alpha$,K$_m$)$|$ $=$ 0 and\\
$|${\sc bridge}(K$_\alpha$,K$_m$)$|$ $=$ 0}\\
\end{tabular}
\end{quote}
Note that this definition prefers linking to bridging, and
bridging to accommodation, which we assume is right.
{\sc link}, {\sc bridge}, and {\sc acc} are functions from
the main DRS to sets of DRSs. We use DRSsubstitution to
describe these operations ($[$ K$_1$ / K$_2$ $]$ K$_3$ means
that K$_1$ is substituted for K$_2$ in K$_3$).
\begin{quote}
{\bf Definition 8. Linking.}\\
{\sc link}(K$_\alpha$,K$_m$) = $\{$ \parbox[t]{100mm}{
$[$ K$_3$ / K$_2$ $]$ K$_m$ $|$
K$_\alpha$ is subordinated and m-suitable to K$_1$ \&
$\alpha$:K$_\alpha$ $\in$ C(K$_2$) \&
U(K$_3$)=U(K$_2$) $\cup$ U(K$_\alpha$) \&
C(K$_3$)=C(K$_2$)-$\alpha$:K$\alpha$ $\cup$ C(K$_\alpha$) $\cup$
$\{$ x=y $|$ m(x)=y $\}$
$\}$ }
\end{quote}
\begin{quote}
{\bf Definition 9. Bridging.}\\
{\sc bridge}(K$_\alpha$,K$_m$) = $\{$ \parbox[t]{100mm}{
K'$_m$~$|$~K$_\alpha$ is subordinated K$_4$ \&
K$_1$ $\in$ {\sc ca}(K$_4$) \& \linebreak
m-suitable to K$_1$ \& $\alpha$:K$_\alpha$ $\in$ C(K$_2$) \& \linebreak
U(K$_3$)=U(K$_2$) $\cup$ U(K$_\alpha$) \&
C(K$_3$)=C(K$_2$)-$\alpha$:K$\alpha$ $\cup$ C(K$_\alpha$) $\cup$
$\{$~x=y~$|$~m(x)=y~$\}$ \&
K'$_m$ = $[$ K$_3$ / K$_2$ $]$ K$_m$ \&
K'$_m$ = $[$~K$_1$~/~K$_4$~$]$~K$_m$
$\}$ }
\end{quote}
\begin{quote}
{\bf Definition 10. Accommodation.}\\
{\sc acc}(K$_\alpha$,K$_m$) = $\{$ \parbox[t]{100mm}{
K'$_m$ $|$
K$_\alpha$ is subordinated to K$_1$ \&
$\alpha$:K$_\alpha$ $\in$ C(K$_2$) \&
U(K$_3$)=U(K$_2$) \&
C(K$_3$)=C(K$_2$)-$\alpha$:K$_\alpha$ \&
K'$_m$=$[$~K$_1$~$\oplus$~K$_\alpha$~/~K$_1$ $]$ K$_m$ \&
K'$_m$=$[$~K$_3$ / K$_2$ $]$ K$_m$~$\}$ }
\end{quote}
Accommodation has its limits. First, it shouldn't introduce
free variables, and Van der Sandt introduces a number
of acceptability rules for accommodation. These are
briefly: resolution should not introduce contradictions and
require a contribution to discourse. For more discussion
on this issue the interested reader should consult
\cite{sandt:presupdrt}.
\subsection{Examples of Linking, Bridging and Accommodation}
\label{examples}
This section exemplifies the notions {\em linking}, {\em bridging}, and
{\em accommodation}, which we introduced in the previous section.
We will do this in view of the examples given in the introduction.
For each of these examples we give the DRS with all
anaphoric information unresolved, and the fully resolved derived
after anaphora resolution as well. For reasons of clarity, only
the relevant parts of the DRSs are deeply analyzed.
\subsubsection{Linking}
The first example involves simple linking between anaphor and
antecedent. Consider the unresolved DRS of \ref{celebrity}:
\begin{center}
\drs{}{\\
\drs{x}{celebrity(x)\\
I-invite(x)} $\to$
\drs{}{\\$\alpha$:\drs{y}{celebrity(y)}\\
never-comes(y)}\\}
\end{center}
The definite description introduces an $\alpha$-DRS for {\sl the
celebrity}, since this is presupposed information.
Trying to link this anaphoric information is successful,
since there is an accessible suitable discourse marker available.
The result is the resolved DRS:
\begin{center}
\drs{}{\\
\drs{x}{celebrity(x)\\
I-invite(x)} $\to$
\drs{y}{celebrity(y)\\
y=x\\
never-comes(y)}\\}
\end{center}
This DRS can be read as: {\sl If I invite a celebrity, he never comes}.
\subsubsection{Bridging}
Now for our bridging\footnote{
Bridging does not seem to be the preferred option in the case
of resolution of pronouns as the following examples show:
\begin{quote}
When I go to a bar, he always throws me out.\\
When BMW announced a new model, he looked very proud.
\end{quote}
In both sentences a reading for 'he' can be found by linking
to coercively accommodated material out of the Q-DRS from
respectively {\sl bar} and {\sl BMW}, i.e. {\sl a barkeeper} or
{\sl a spokesperson}. However they don't seem to be the
preferred readings as has been shown by
\cite{McGlashan,SanfordandGarrod} for similar examples.}
example.
The unresolved DRS of example~\ref{barkeeper} is (simplifying the Q-DRS
for convenience):
\begin{center}
\drs{}{\\
\drs{x}{bar(x)\\
Q:\drs{z}{barkeeper(z)\\of(z,x)}\\
I-go-to(x)} $\to$
\drs{}{\\
$\alpha$:\drs{y}{barkeeper(y)}\\
always-throws-me-out(y)}\\}
\end{center}
The presupposition trigger {\sl the barkeeper} introduces the
anaphoric information. Linking fails, the only available
discourse marker is not suitable since the condition of anaphoric
information does not match with it. Bridging is successful, though,
yielding the resolved DRS:
\begin{center}
\drs{}{\\
\drs{x z}{bar(x)\\
$\cal{Q}$\\
I-go-to(x)\\
barkeeper(z)\\
of(z,x)} $\to$
\drs{y}{barkeeper(y)\\
y=z\\
always-throws-me-out(y)}\\}
\end{center}
This DRS does not assume a particular barkeeper that throws
the speaker out, but a barkeeper that belongs to the bar the
speaker goes to -- the correct prediction.
\subsubsection{Accommodation}
Accommodation is our emergency case: {\em if everything fails, then
accommodate}. This happens in cases like \ref{kingoffrance}, which
unresolved DRS is:
\begin{center}
\drs{}{\\
\drs{x}{party(x)\\
I-give(x)} $\to$
\drs{}{\\
$\alpha$:\drs{y}{king-of-france(y)}\\ \\
$\alpha$:\drs{z}{}\\
always-attends(y,z)}\\}
\end{center}
The pronoun represented by the discourse marker {\sl z} can be linked to
{\sl x}. But we cannot link {\sl the king of France} to some accessible
discourse marker, nor is there a way to make bridging inference. The
only possibility left is to accommodate the king:
\begin{center}
\drs{y}{king-of-france(y)\\
\drs{x}{party(x)\\
I-give(x)} $\to$
\drs{z}{z=x\\
always-attends(y,z)}\\}
\end{center}
This DRS represents the reading: {\sl there is a king of France, and if
I give a party, he will attend it}. This is again the correct
prediction.
\section{Functional Composition and Coercion}
\label{functionalcomposition}
\subsection{Defining the notions}
Functional Composition, including type coercion,
is defined as follows, K$_1$ being the functor,
K$_2$ the argument, and $\vec{\sigma}$ a sequence of
terms such that K$_2$(t)($\vec{\sigma}$) is a proposition:
\begin{quote}
{\bf Definition 11. Functional Composition ($\odot$).}\\
\begin{tabular}{ll}
K$_1$ $\odot$ K$_2$ &
= $\lambda$ $\vec{\sigma}$.
K$_1$($\lambda$ v. (K$_2$(v)($\vec\sigma$)))
iff K$_1$ is of type $<\alpha$,t$>$ and v is of type $\alpha$; \\
& = K$_1$ $\odot$ K$_3$ (where K$_3$ $\in$ {\sc tc}(K$_2$)) otherwise.
\end{tabular}
\end{quote}
Clause one is like the functional composition rule (in
\cite{tilburg}). This rule has the nice property that
it doesn't need type-shifting of arguments. It always binds
the first argument position of the argument, and has functional
application as a special case ($\vec{\sigma}$ is empty then).
The second clause does the type coercion stuff (cf. \cite{pus:TC}):
\begin{quote}
{\bf Definition 12. Type Coercion (TC).}\\
{\sc tc}(K)=$\{$ K$^\prime$ $\odot$ K $|$ K$^\prime$ $\in$ {\sc qa}(K) $\}$
\end{quote}
\begin{quote}
{\bf Definition 13. Qualia Access (QA)}.\\
{\sc qa}(K)=$\{$ K$_Q$ $|$ Q:K$_Q$ $\in$ C(K)
or {\sc qa}(K') where K' is a sub-DRS of K $\}$
\end{quote}
Note that {\sc tc} also works for arbitrarily deep embedded DRSs by use
of the Qualia Access function. This is nice for quantified NPs like {\sl
every book}, where the qualia DRSs lexically introduced for {\sl book}
has been placed in the restrictor.
\subsection{Some lexical entries}
In this section we present some example lexical entries. In this paper we
will only assign (a simplified) qualia structure to nouns\footnote{
\cite{pus:book} assumes qualia structures for {\em all}
categories.},
see {\bf book}. Lexical entries can be abbreviated by their boldface
notation -- {\bf write} stands for the semantic part of the
lexical entry of {\sl write}. We use small e,x,y and z for variables
over type e (for entities, i.e. objects and events), capital P for
DRS of type $<$e,t$>$ (properties), and capital E for event-types
(normally $<e,t>$).
\begin{tabbing}
{\bf pres} \= : \= \kill
{\bf book}\>:\> $\lambda$z.\drs{z}{book(z)\\
\qdrs{F}{}{info\_cont(z)}\\ \\
\qdrs{C}{Z}
{sections(Z)\\
has(z,Z)}\\ \\
Q$_{\mbox{\tiny A}}$: {\bf write} \\
Q$_{\mbox{\tiny T}}$: {\bf read}}
\end{tabbing}
By introducing determiners ({\bf a, the, every}) we account for the
possibility to carry qualia structure through the derivation. Note the
difference between these three determiners. The article {\em the}
introduces an $\alpha$-DRS since it is a presupposition trigger.
\begin{tabbing}
{\bf pres} \= : \= \kill
{\bf a} \>:\>
$\lambda$ P$_1$ P$_2$. \drs{x}{} $\oplus$ P$_1$(x) $\oplus$ P$_2$(x)\\ \\
{\bf the} \>:\>
$\lambda$ P$_1$ P$_2$.
\drs{}{\mbox{}\\[-2ex] $\alpha$: \drs{x}{} $\oplus$ P$_1$(x)\mbox{}\\[-2ex]}
$\oplus$ P$_2$(x)\\ \\
{\bf every} \>:\>
$\lambda$ P$_1$ P$_2$.
\drs{}{\mbox{}\\[-2ex]\drs{x}{} $\oplus$ P$_1$(x) $\to$
P$_2$(x)\mbox{}\\[-2ex]}
\end{tabbing}
The proper name {\bf john} introduces an anaphoric DRS
which is merged with the representation of its predicate. Proper names
do {\em not\/} have qualia structure (see footnote earlier).
\begin{tabbing}
{\bf pres} \= : \= \kill
{\bf john}\>:\>
$\lambda$ P. \drs{}{\mbox{}\\[-2ex]$\alpha$:\drs{x}{john(x)}\mbox{}\\[-2ex]}
$\oplus$ P(x)
\end{tabbing}
The verbs
{\bf write} and {\bf read} introduce event-types. Lambda-operators bind the
variables that will fulfill the thematic roles {\em agent\/} and {\em theme}.
\begin{tabbing}
{\bf pres} \= : \= \kill
{\bf write} \>:\> $\lambda$ y x e. \drs{}{write(e)\\
agent(e,x)\\
theme(e,y)}\\ \\
{\bf read} \>:\> $\lambda$ y x e. \drs{}{read(e)\\
agent(e,x)\\
theme(e,y)}
\end{tabbing}
The aspectual verb
{\bf begin} expects something that expresses an event-type.
We here simply treat it as a modifier, and ignore its further
aspectual presuppositions.
Finally, tense {\bf pres}
applies to an event-type and binds off the event variable:the
result is a DRS of type t, i.e. a DRS with no lambda variables.
\begin{tabbing}
{\bf pres} \= : \= \kill
{\bf begin} \>: \> $\lambda$ E x e. \drs{}{begin(e)} $\oplus$ E(x)(e) \\ \\
{\bf pres} \> : \> $\lambda$ E. \drs{e}{now(e)} $\oplus$ E(e)
\end{tabbing}
\subsection{A sample derivation}
Let us now follow the derivation of `John begins a book'. Functional
composition of {\bf a} with {\bf book} yields a noun phrase that
contains the qualia structure of the noun and awaits a property to
merge with.
\begin{quote}
{\bf a} $\odot$ {\bf book} =
$\lambda$P.\drs{z}{book(z)\\
\qdrs{F}{}{info\_cont(z)}\\ \\
\qdrs{C}{Z}
{sections(Z)\\
has(z,Z)}\\ \\
Q$_{\mbox{\tiny A}}$: {\bf write} \\
Q$_{\mbox{\tiny T}}$: {\bf read}}
$\oplus$ P(z)
\end{quote}
Functional composition of {\bf begin} with {\bf a $\odot$ book} can only
work with a type coercion. The event that {\bf begin} requires cannot be
found directly, so no simple link can be made. From the qualia structure
of {\bf a $\odot$ book} we can for example coerce {\bf read}, and
this qualifies as the required event.
This coercion step is worked out later.
\begin{quote}
{\bf begin} $\odot$ ({\bf a} $\odot$ {\bf book}) =
$\lambda$ x. e.
\drs{y}
{begin(e)\\
read(e)\\
agent(e,x)\\
theme(e,y)\\
book(y)\\
$\cal{Q}$}
\end{quote}
The rest of the derivation follows straightforwardly. {\bf begin $\odot$
(a $\odot$ book)} functionally composed with {\bf john} results in a
lambda-DRS.
\begin{quote}
{\bf john} $\odot$ ({\bf begin} $\odot$ ({\bf a} $\odot$ {\bf book})) =
$\lambda$ e.
\drs{y}
{\mbox{}\\[-2ex]$\alpha$:\drs{x}{john(x)\mbox{}\\[-2ex]}\\
begin(e)\\
read(e)\\
agent(e,x)\\
theme(e,y)\\
book(y)\\
$\cal{Q}$
}
\end{quote}
Adding tense ({\bf pres}) to the lambda-DRS turns it into a proper
DRS -- all anaphoric information has been resolved and no antecedent for
the presupposed event needed to be accommodated.
\begin{quote}
{\bf pres} $\odot$ ({\bf john} $\odot$ ({\bf begin} $\odot$ ({\bf a} $\odot$
{\bf book}))) =
\drs{e y}{\mbox{}\\[-2ex]$\alpha$:\drs{x}{john(x)\mbox{}\\[-2ex]}\\
now(e)\\
begin(e)\\
read(e)\\
agent(e,x)\\
theme(e,y)\\
book(y)\\
$\cal{Q}$}
\end{quote}
\pagebreak
Naturally we could just as easily have taken {\bf write} instead of {\bf
read}, or for that matter, any of the other events that occur in the
qualia structure of {\bf book}. Since {\bf read} and {\bf write} are the
only events, the result of coercing {\bf a book} is as follows:\\
\begin{minipage}{\textwidth}
{\sc tc}($\lambda$P.\drs{z}{book(z)\\
\qdrs{F}{}{info\_cnt(z)}\\ \\
\qdrs{C}{Z}
{sections(Z)\\
has(z,Z)}\\ \\
Q$_{\mbox{\tiny A}}$: {\bf write} \\
Q$_{\mbox{\tiny T}}$: {\bf read}}
$\oplus$ P(z)) =
$\{$ $\lambda$x.e.\drs{z}
{book(z)\\
write(e)\\
agent(e,x)\\
theme(e,z)\\
$\cal{Q}$},
$\lambda$x.e.\drs{z}{book(z)\\
read(e)\\
agent(e,x)\\
theme(e,z)\\
$\cal{Q}$}$\}$
\end{minipage}
\section{Conclusions and Further Work}
\label{conclusions}
We have shown that Bridging and Coercion can be seen in very much the
same light, viz. as using implicit lexical information to accommodate
a missing antecedent. In doing so, we have extended Pustejovsky's ideas
on Coercion and placed it in a discourse perspective. On the other hand
we have extended Van der Sandt's algorithm with Bridging, and thus made
it more complete with respect to the linguistic data.
The work presented here is limited to definite descriptions; we have not
looked into other presupposition triggers. \cite{beaver} mentions the
following examples, where inferencing takes place.
\begin{examples}
\ex\exlab{beaver1}
Probably, if Jane takes a bath, Bill will be annoyed
that there is no more hot water.
\ex\exlab{beaver2}
If Spaceman Spiff lands on Planet X, he will be bothered
by the fact that his weight is higher than it would be
on Earth.
\end{examples}
In \ref{beaver1} the inference is made that taking a bath uses up a hot
water reservoir, in \ref{beaver2} that landing on a strange planet may
make changes to your weight. To fit with these examples in the framework
we presented in this paper remains for future research.
|
\section{Introduction}
The formation of cosmological structure in the universe,
inhomogeneities in the matter distribution like quasars at redshifts
up to $z\sim 5$, galaxies, clusters,
super clusters, voids and walls, is an outstanding basically
unsolved problem within the standard model of cosmology.
At first sight it seems obvious that small density enhancements can grow
sufficiently rapidly by gravitational instability. But global
expansion of the universe and
radiation pressure counteract gravity, so that, e.g., in the case of
a radiation dominated, expanding universe no density inhomogeneities
can grow faster than logarithmically. Even in a universe dominated by
pressure-less matter, cosmic dust, the
growth of density perturbations is strongly reduced by the expansion
of the universe.
On the other hand, we know that the universe was extremely homogeneous
and isotropic at early times. This follows from the isotropy of
the 3K Cosmic Microwave Background (CMB), which represents a relic of
the plasma of baryons, electrons and radiation at times before protons
and electrons combined to hydrogen. After a long series of upper bounds,
measurements with the COsmic Background Explorer satellite (COBE)
have finally established anisotropies in this radiation \cite{Sm} at
the level of
\[ \langle{\Delta T\over T}(\theta)\rangle \sim 10^{-5}
~~~\mbox{ on angular scales}~
7^o\le \theta\le 90^o ~.\]
On smaller angular scales the observational situation is
at present somewhat confusing and contradictory \cite{WJNPW}, but many
upper limits require
$ \Delta T/T < 4\times 10^{-5} $ on all scales $\theta <8^o$.
All observations together clearly rule out the simplest model of a
purely baryonic universe with density parameter $\Omega \sim 0.1$ and
adiabatic initial fluctuations (either the initial perturbations are too
large to satisfy CMB limits, or they are too small to develop into the
observed large scale structure).
The most conservative way out, where one just allows for non--adiabatic
initial perturbations (minimal isocurvature model), also faces severe
difficulties \cite{PS,GSS,PD,HBS}. In
other models one assumes that initial fluctuations are
created during an inflationary epoch, but that the matter content of
the universe is dominated by hot or cold dark matter or a mixture
of both. Dark matter particles do not interact with photons other than
gravitationally and thus induce perturbations in the CMB only via
gravitation. In these models, inflation generically leads to $\Omega =1$,
while the baryonic density parameter is
only $\Omega_Bh^2 \sim 0.02$, compatible with nucleosynthesis constraints.
With one
component of dark matter, these models do not seem to
agree with observations \cite{GSS,Os}, however, if a suitable mixture
of hot and cold dark matter is adopted, the results from numerical
simulations look quite promising \cite{Ho,XK,DSS}, although they might
have difficulties to account for the existence of clusters at a redshift
$z\sim 1$ \cite{Ba}.
In these dark matter models initial fluctuations are
generated during an inflationary phase. Since all worked out models
of inflation face difficulties (all of them have to invoke fine
tuning to obtain the correct amplitude of density inhomogeneities), we
consider it very important to investigate yet another possibility:
Density perturbations in the dark matter and baryons might have been
triggered by seeds.
Seeds are an inhomogeneously distributed form of energy which
makes up only a small fraction of the total energy density of the
universe.
Particularly natural seeds are topological defects. They can form
during symmetry breaking phase transitions in the early universe
\cite{Ki,CDTY}. Depending on the symmetry being gauged or global, the
corresponding defects are called local or global.
The fluctuation spectrum on large scales observed by COBE is
not very far from scale invariant \cite{Go}. This has been
considered a great success for inflationary models which generically
predict a scale
invariant fluctuation spectrum. However, as we shall see, also
models in which perturbations are seeded by global topological defects
yield scale invariant spectra of CMB fluctuations. To be specific, we
shall mainly
consider texture, $\pi_3$--defects which lead to event singularities
in four dimensional spacetime \cite{Tu,d95}. Global defects are viable
candidates for structure formation, since the scalar field energy
density, $\rho_S$, of global topological defects scales like
$\rho_S \propto 1/(at)^{2}$ (up to a logarithmic correction for global
strings) and thus always represents the same fraction of the total
energy density of the universe ($t$ is conformal time).
\begin{equation} \rho_S/\rho \sim 8\pi G\eta^2 \equiv 2\epsilon ~,
\end{equation}
where $\eta$ determines the symmetry breaking scale ({\sl see} Fig.~1). For
the background spacetime we assume a Friedmann--Lema\^{\i}tre universe
with $\Omega=1$ dominated by cold dark matter (CDM). We choose
conformal coordinates such that
\[ ds^2 = a^2(-dt^2 + \delta_{ij}dx^idx^j) ~. \]
Numerical analysis of CMB fluctuations from topological defects on large
scales has been performed in \cite{BR,PST}; a spherically symmetric
approximation is discussed in \cite{DHZ}. Results for intermediate scales
angular are presented in \cite{CFGT}. All these investigations (except
\cite{DHZ}) use linear cosmological perturbation theory in synchronous
gauge and (except \cite{PST}) take into account only scalar
perturbations.
Here we derive a fully gauge invariant and local system of
perturbation equations. The (non--local) split into
scalar, vector and tensor modes on hyper surfaces of constant time
is not performed. We solve the equations numerically in a cold dark
matter (CDM) universe with global texture. In this paper, we
detail the results outlined in a previous letter \cite{DZ}. Furthermore,
we present explicit
derivations of the equations, a description of our numerical
methods and we briefly discuss some tests of our codes.
Since there are no spurious gauge modes in our initial conditions,
there is no danger that these may grow in time and some of the
difficulties to choose correct initial conditions ({\sl see} e.g. \cite{PST})
are removed. However, as we shall discuss in Section~3, the results
do depend very sensitively on the choice of initial conditions.
Nevertheless, we should keep in mind, that we are investigating models
of structure formation which rely on the particle physics and cosmology
at temperatures of $T\sim T_{GUT} \sim 10^{16}GeV$. An energy scale
about which we have no experimental evidence whatsoever. The physical
model adopted for our calculations should thus always be considered as a
toy model, of which we hope it captures the features relevant for
structure formation of the 'realistic physics' at these energies.
Therefore, we suggest, not to take the results serious much beyond
about a factor of two or so. On the other hand, our
models show that the particle physics at GUT scale may have left its
traces in the distribution of matter and radiation in the present
universe, yielding the exciting possibility to learn about the physics
at the highest energies, smallest scales, by probing the largest
structures of the universe.
\vspace{0.2cm}
We calculate the CMB anisotropies on angular scales
which are larger than the angle subtended by the horizon scale at
decoupling of matter and radiation, $\theta> \theta_d$. For $\Omega =1$
and $z_d \approx 1000$
\begin{equation} \theta_d = 1/\sqrt{z_d+1} \approx 0.03 \approx 2^o
~. \end{equation}
It is therefore sufficient to study the generation and evolution
of microwave background fluctuations after recombination. During this
period, photons stream freely, influenced only by cosmic
gravitational redshift and by perturbations in the gravitational
field (if the medium is not re-ionized).
In Section~2 we derive a local and gauge invariant perturbation
equation to calculate the CMB fluctuations. In Section~3, we
put together the full system of equations which has to be solved
to investigate gravitationally induced CMB fluctuations and the dark
matter perturbation spectrum in a model with global topological defects.
We discuss the choice of initial conditions and the numerical treatment
of this system in Section~4. The next section is devoted to the
presentation and analysis of our numerical results. We end with
conclusions in Section~6.
\vspace{0.2cm}
{\bf Notation:} We denote conformal time by $t$. Greek indices run from
0 to 3, Latin indices run from 1 to 3. The metric signature is chosen
$(- + + +)$. We set $\hbar=c=k_{Boltzmann}=1$ throughout.
\section{A Local and Gauge Invariant Treatment of the Perturbed Liouville
Equation}
Collision-less particles are described by their one particle
distribution function which lives on the seven dimensional phase space
\[ {\cal P}_m = \{ (x,p)\in T\!{\cal M}| g(x)(p,p)=-m^2\} ~. \]
Here $\cal M$ denotes the spacetime manifold and $T{\cal M}$ its
tangent space. The fact that collision-less particles move on geodesics
translates to the Liouville equation for the one particle distribution
function, $f$. The Liouville equation reads \cite{Ste}
\begin{equation} X_g(f)=0 \label{L}~.\end{equation}
In a tetrad basis $(e_\mu)_{\mu=0}^3$ of $\cal M$, the vector field
$X_g$ on ${\cal P}_m$ is given
by ({\sl see} e.g. \cite{Ste})
\begin{equation}
X_g = (p^\mu e_\mu - \omega^i_{\:\mu}(p)p^\mu{\partial \over \partial p^i})
~, \label{liou} \end{equation}
where $\omega^\nu_{\:\mu}$ are the connection 1--forms of $({\cal M},g)$
in the basis $e^\mu$, and we have chosen the basis
\[(e_\mu)_{\mu=0}^3~~ \mbox{ and }~~~
({\partial\over \partial p^i})_{i=1}^3 ~~\mbox{ on }~~~~ T{\cal P}_m~,
~~~~~~ p=p^\mu e_\mu~. \]
We now apply this general framework to the case of a perturbed Friedmann
universe. The metric of a perturbed Friedmann universe with density
parameter $\Omega=1$ is given by $ds^2=g_{\mu\nu} dx^\mu dx^\nu$ with
\begin{equation}
g_{\mu\nu} = a^2(\eta_{\mu\nu} + h_{\mu\nu}) = a^2 \tilde{g}_{\mu\nu}
~, \end{equation}
where $(\eta_{\mu\nu}) = diag(-,+,+,+)$ is the flat Minkowski metric and
$(h_{\mu\nu})$ is a small perturbation, $|h_{\mu\nu}|\ll1$. We now use
the fact that the motion of photons is conformally invariant:
We show that for massless particles and conformally related metrics,
\[g_{\mu\nu}= a^2\tilde{g}_{\mu\nu}~,\]
\begin{equation}
(X_gf)(x,p)=0 ~~~\mbox{ is equivalent to }~~~
(X_{\tilde{g}}f)(x,ap)=0 ~.
\label{conform} \end{equation}
This is easily seen if we write $X_g$ in a coordinate basis:
\[ X_g = b^\mu\partial_\mu -\Gamma_{\alpha\beta}^ib^\alpha b^\beta{\partial\over \partial b^i}
~,\]
with
\[ \Gamma_{\alpha\beta}^i={1\over 2}g^{i\mu}(g_{\alpha\mu},_\beta +
g_{\beta\mu},_\alpha-g_{\alpha\beta},_\mu) ~. \]
The $b^\mu$ are the components of the momentum $p$ with respect to the
{\em coordinate} basis:
\[ p=p^\mu e_\mu = b^\mu\partial_\mu ~.\]
If $(e_\mu)$ is a tetrad with respect to $g$, then $\tilde{e}_\mu=ae_\mu$
is a tetrad basis for $\tilde{g}$. Therefore, the coordinates of
of $ap=ap^\mu\tilde{e}_\mu= a^2p^\mu e_\mu = a^2b^\mu\partial_\mu$ with
respect to $\partial_\mu$ on
$({\cal M},\tilde{g})$ are given by $a^2b^\mu$.
In the coordinate basis thus our statement Eq.~(\ref{conform}) follows, if we
can show that
\begin{equation}
(X_{\tilde{g}}f)(x^\mu,a^2b^i)=0
~~~\mbox{ iff }~~~ (X_gf)(x^\mu,b^i)=0
\label{conformco} \end{equation}
Setting $v=ap=v^\mu\tilde{e}_\mu = w^\mu\partial_\mu$, we have $v^\mu=ap^\mu$
and $w^\mu=a^2b^\mu$. Using $p^2=0$, we obtain the following relation
for the Christoffel symbols of $g$ and $\tilde{g}$:
\[ \Gamma^i_{\alpha\beta}b^\alpha b^\beta= \tilde{\Gamma}^i_{\alpha\beta}b^\alpha b^\beta
+{2a,_\alpha\over a}b^\alpha b^i ~.\]
For this step it is crucial that the particles are massless! For massive
particles the statement is of course not true.
Inserting this result into the Liouville equation we find
\begin{equation}
a^2X_gf = w^\mu(\partial_\mu f|_b -2{a,_\mu\over a} b^i{\partial f\over \partial b^i})
-\tilde{\Gamma}_{\alpha\beta}^iw^\alpha w^\beta {\partial f\over \partial w^i} ~,
\label{lstar} \end{equation}
where $\partial_\mu f|_b$ denotes the derivative of $f$ w.r.t. $x^\mu$ at
constant $(b^i)$. Using
\[\partial_\mu f|_b = \partial_\mu f|_w + 2{a,_\mu\over a}b^i{\partial f\over \partial b^i}
~,\]
we see, that the braces in Eq.~(\ref{lstar}) just correspond to
$\partial_\mu f|_w$. Therefore,
\[a^2X_g f(x,p)=w^\mu\partial_\mu f|_w-\tilde{\Gamma}_{\alpha\beta}^iw^\alpha w^\beta
{\partial f\over \partial w^i} = X_{\tilde{g}}f(x,ap) ~. \]
We have thus shown that the Liouville equation in a perturbed
Friedmann universe is equivalent to the Liouville equation in perturbed
Minkowski space,
\begin{equation} (X_{\tilde{g}}f)(x,v)=0 ~,
\label{LM} \end{equation}
with $v=v^\mu\tilde{e}_\mu = ap^\mu\tilde{e}_\mu$.\footnote{Note that
also Friedmann universes with non vanishing spatial curvature,
$K\neq 0$, are conformally flat and thus
this procedure can also be applied for $K\neq 0$. Of course, in this
case the conformal factor $a^2$ is no longer just the scale factor but
depends on position. A coordinate transformation which transforms the
metric of $K\neq 0$ Friedmann universes into a conformally flat form
can be found, e.g., in \cite{CDD}.}
We now want to derive a perturbation equation for Eq.~(\ref{LM}).
If $\bar{e}^\mu$ is a tetrad in Minkowski space,
$\tilde{e}_\mu = \bar{e}_\mu + {1\over 2}h_\mu^\nu\bar{e}_\nu$ is a
tetrad w.r.t the
perturbed geometry $\tilde{g}$. For
$(x,v^\mu\bar{e}_\mu )\in \bar{P}_0$, thus,
$(x,v^\mu\tilde{e}_\mu)\in \tilde{P}_0$. Here $\bar{P}_0$ denotes the
zero mass one particle phase space in Minkowski space and $\tilde{P}_0$
is the phase space with respect to $\tilde{g}$, perturbed Minkowski
space. We define the
perturbation of the distribution function $F$ by
\begin{equation}
f(x,v^\mu \tilde{e}_\mu) = \bar{f}(x,v^\mu \bar{e}_\mu) +
F(x,v^\mu\bar{e}_\mu)
~. \end{equation}
Liouville's equation for $f$ then leads to a perturbation equation
for $F$. We choose the natural tetrad
\[\tilde{e}_\mu=\partial_\mu -{1\over 2}h_\mu^\nu\partial_\nu\]
with the corresponding basis of 1--forms
\[\tilde{\theta}^\mu=dx^\mu +{1\over 2}h^\mu_\nu dx^\nu ~.\]
Inserting this into the first structure equation,
$d\tilde{\theta}^\mu= -\omega^\mu_{~~\nu}\wedge dx^\nu$, one finds
\[ \omega_{\mu\nu}=-{1\over 2}(h_{\mu\lambda},_\nu -
h_{\nu\lambda},_\mu)\theta^\lambda ~.\]
Using the background Liouville equation, namely that $\bar{f}$ is
only a function of $v=ap$, we obtain the perturbation equation
\[ (\partial_t +\gamma^i\partial_i)F = -{v\over 2}[(\dot{h}_{i0}-h_{00},_i)\gamma^i
+(\dot{h}_{ij}-h_{0j},_i)\gamma^i\gamma^j]{d\bar{f}\over dv} ~,\]
where we have set $v^i=v\gamma^i$, with $v^2=\sum_{i=1}^3(v^i)^2$.
Let us parameterize the perturbations of the metric by
\begin{equation} \left(h_{\mu\nu}\right) = \left(\begin{array}{ll}
-2A & B_i \\
B_i & 2H_L\delta_{ij}
+2H_{ij} \end{array}\right),
\label{scalar} \end{equation}
with $H_i^i=0$. Inserting this above we obtain
\begin{equation}
(\partial_t +\gamma^i\partial_i)F = -[\dot{H}_L +(A,_i +{1\over 2}\dot{B}_i)\gamma^i +
(\dot{H}_{ij}-{1\over 2}B_{i,j})\gamma^i\gamma^j]v{d\bar{f}\over dv} ~.
\label {LF} \end{equation}
From Eq.~(\ref{LF}) we see that the perturbation in the distribution
function in each spectral band is proportional to $v{d\bar{f}\over dv}$.
This shows ones more that gravity is achromatic. We thus
do not loose any information if we integrate this equation over
photon energies. We define
\[ m = {\pi\over \rho_ra^4}\int Fv^3dv ~.\]
$4m$ is the fractional perturbation of the brightness $\iota$,
\[ \iota = a^{-4} \int f v^3dv ~. \]
This is obtained using the relation
\begin{equation}
{4\pi}\int {d\bar{f}\over dv}v^4dv = -4\int \bar{f}v^3dvd\Omega
=-4\rho_ra^4 ~. \label{rel} \end{equation}
Setting $\iota = \bar{\iota}(T(\gamma,x))$, one finds that
$ \iota =(\pi/60) T^4(\bm{\gamma},x)$. Hence, $m$
corresponds to the fractional perturbation in the temperature,
\begin{equation} T(\gamma,x) = \bar{T}(1+m(\gamma,x)) ~.\label{T} \end{equation}
Another derivation of Eq.~(\ref{T}) is given in \cite{d94}.
Since the $v$ dependence of $F$ is of the form
$v{d\bar{f}\over dv}$, we have with Eq.~(\ref{rel})
\[ F(x^\mu,\gamma^i,v)=-m(x^\mu,\gamma^i)v{d\bar{f}\over dv} ~. \]
This shows that $m$ is indeed the quantity which is measured in a
CMB anisotropy experiment, where the spectral information is used
to verify that the spectrum of perturbations is the derivative of a
blackbody spectrum. Of course, in a real experiment located at a fixed
position in the Universe, the monopole and
dipole contributions to $m$ cannot be measured. They cannot be
distinguished from a background component and from a dipole due to our
peculiar motion w.r.t. the CMB radiation.
Multiplying Eq.~(\ref{LF}) with $v^3$ and integrating over $v$, we obtain
the equation of motion for $m$
\begin{equation}
\partial_tm+\gamma^i\partial_im= \dot{H}_L +(A,_i +{1\over 2}\dot{B}_i)\gamma^i +
(\dot{H}_{ij}-{1\over 2}B_i,_j)\gamma^i\gamma^j ~.
\label{Lm}
\end{equation}
It is well known that the equation of motion for photons only couples to
the Weyl part of the curvature (null geodesics are conformally invariant).
The r.h.s. of Eq.~(\ref{Lm}) is given by first derivatives of the metric only
which could at most represent integrals of the Weyl tensor. To obtain
a local, non integral equation, we thus rewrite Eq.~(\ref{Lm}) in terms of
$\triangle m$. It turns out, that the most suitable variable is however not $\triangle m$
but $\chi$, which is given by
\[ \chi = \triangle m - (\triangle H_L-{1\over 2}H,_{ij}^{ij}) -
{1\over 2}(\triangle B_i-3\partial^j\sigma_{ij})\gamma^i ~, \]
\[ \mbox{where }~~ \sigma_{ij}= -{1\over 2}(B_i,_j+B_j,_i) +
{1\over 3}\delta_{ij}B_l^{,l} +\dot{H}_{ij}. \]
Note that $\chi$ and $\triangle m$ only differ by the monopole contribution,
$\triangle H_L-(1/2)H^{ij},_{ij}$ and the dipole contribution,
$(1/2)(\triangle B_i -3\partial^j\sigma_{ij})\gamma^i$. The higher multipoles of
$\chi$ and $\triangle m$ agree.
An observer at fixed position and time cannot distinguish a monopole
contribution from an isotropic background
and a dipole contribution from a peculiar motion.
Only the higher multipoles, $l\ge 2$ contain information about
temperature anisotropies. For a fixed observer therefore, we can
identify $\triangle^{-1}\chi$ with $\delta T/T$.
In terms of metric perturbations, the electric and magnetic part of the
Weyl tensor are given by ({\sl see}, e.g. \cite{Ma,d94})
\begin{eqnarray}
E_{ij} &=& {1\over 2}[\triangle_{ij}(A-H_L) -\dot{\sigma}_{ij}
-\triangle H_{ij}-{2\over 3}H_{lm}^{,lm}\delta_{ij}
+ H_{il}^{,l},_j + H_{jl}^{,l},_i] \label{E} \\
B_{ij} &=& -{1\over 2}(\epsilon_{ilm}\sigma_{jm},_l + \epsilon_{jlm}\sigma_{im},_l ) ~,
\label{B} \end{eqnarray}
\[ \mbox { with }~~ \triangle_{ij} =\partial_i\partial_j -(1/3)\delta_{ij}\triangle ~.\]
Explicitly working out $(\partial_t+\gamma^i\partial_i)\chi$ using Eq.~(\ref{Lm}),
yields after some algebra the equation of motion for $\chi$:
\begin{equation}
(\partial_t +\gamma^i\partial_i)\chi = 3\gamma^i\partial^jE_{ij} +\gamma^k\gamma^j\epsilon_{kli}
\partial_lB_{ij} \equiv S_T(t,\bm{x},\bm{\gamma})~ ,
\label{Lchi} \end{equation}
where $\epsilon_{kli}$ is the totally antisymmetric tensor in three
dimensions with $\epsilon_{123}=1$.
The spatial indices in this equation are raised and lowered with
$\delta_{ij}$ and thus index positions are irrelevant. Double indices are
summed over irrespective of their positions.
In eqn. Eq.~(\ref{Lchi}) the contribution from the electric part of the
Weyl tensor does not contain tensor perturbations. On the other hand,
scalar perturbations do not induce a magnetic gravitational field. The
second contribution to the source term in Eq.~(\ref{Lchi}) thus represents a
combination of vector and tensor perturbations. If vector perturbations
are negligible, the two terms on the r.h.s of Eq.~(\ref{Lchi})
yield a split into scalar and tensor perturbations which is local.
Since the Weyl tensor of Friedmann Lema\^{\i}tre universes vanishes, the
r.h.s. of Eq.~(\ref{Lchi}) is manifestly gauge invariant (this is the so
called Stewart--Walker lemma \cite{SW}). Hence also the variable
$\chi$ is gauge invariant.
Another proof of the gauge
invariance of $\chi$, discussing the behavior of $F$ under infinitesimal
coordinate transformations is presented in \cite{d94}.
The general solution to Eq.~(\ref{Lchi}) is given by
\begin{equation}
\chi(t,\bm{x},\bm{\gamma}) =
\int_{t_i}^t S_T(t',\bm{x}+(t'-t)\bm{\gamma}, \bm{\gamma})dt'
~ + ~ \chi(t_i,\bm{x}+(t_i-t)\bm{\gamma}, \bm{\gamma}) ~,
\label{chi} \end{equation}
where $S_T$ is the source term on the r.h.s. of Eq.~(\ref{Lchi}).
Let us compare this result with the more familiar one, where one
calculates $\delta T/T$ by integrating photon geodesics (which is of
course equivalent to solving the Liouville equation). For simplicity,
we specialize to the case of pure scalar perturbations (the expressions
for vector and tensor perturbations given in \cite{d94} can be compared
with Eq.~(\ref{chi}) in the same manner.) For scalar perturbations,
integration of photon geodesics yields \cite{d94}
\begin{equation}
{\delta T\over T}(t_f,\bm{x}_f,\bm{n}) = -[{1\over 4}D_g^{(r)} +
V_i\cdot n^i + (\Psi-\Phi)]^f_i +
\int_i^f(\dot{\Psi}-\dot{\Phi})d\lambda ~. \label{geo}
\end{equation}
Here $\Psi$ and $\Phi$ denote the Bardeen potentials as defined, e.g.,
in \cite{KS,d94}.
On super horizon scales (which are the important scales for the
Sachs--Wolfe) contribution $V_i\cdot n^i$ can be neglected. Furthermore,
the contributions in the square bracket of Eq.~(\ref{geo})
from the final time $t=t_f$, only lead to
uninteresting monopole and dipole terms. We now use that
the electric contribution to the Weyl
tensor for purely scalar perturbations is given by \cite{d94})
\[ E_{ij} = {1\over 2}(\partial_i\partial_j -{1\over 3}\triangle)(\Psi-\Phi)
\equiv {1\over 2}\triangle_{ij}(\Psi-\Phi) ~.\]
Therefore $\partial_i(\Psi-\Phi)=3\partial^jE_{ij}$. Using furthermore
\[ -(\Psi-\Phi)\left|^f_i \right. = -\int_i^f[\dot{\Psi}-\dot{\Phi}+
(\Psi-\Phi),_in^i]d\lambda ~, \]
Eq.~(\ref{geo}) leads to
\begin{equation}
{\delta T\over T}(t,\bm{x},\bm{n}) =
{1\over 4}D_g^{(r)}(t_i,\bm{x}_i) -
3\int_i^f\triangle^{-1}\partial^jE_{ij}n^i dt ~ \label{geo2}
\end{equation}
If we take into account that the direction $\bm{n}$ in Eq.~(\ref{geo}),
the direction of an {\sl incoming } photon corresponds to $-\bm{\gamma}$
in Eq.~(\ref{chi}), we find that Eq.~(\ref{geo}) coincides with
Eq.~(\ref{chi}) for scalar perturbations, and that
\begin{equation} \chi(t_i,\bm{x}_i,\bm{\gamma})= {1\over 4}\triangle D_g^{(r)}(t_i,\bm{x}_i)
={1\over 4}\triangle D_g^{(r)}(t_i,\bm{x} -\bm{\gamma}(t-t_i)) ~.
\end{equation}
We now want to investigate this initial value and decompose
Eq.~(\ref{geo2}) into terms due to CDM and terms coming from the
source, the scalar field. We assume that dark matter and radiation
perturbations are adiabatic on {\em superhorizon scales},
\[D_g^{(r)}=(4/3)D_g^{(c)}~.\]
Since
radiation and CDM probably have been a single fluid at early times
(e.g. at the time of the phase transition), this assumption is
reasonable.
It is however, inconsistent to set $D_g^{(r)} =4/3D_g^{(c)}$ on
subhorizon scales. Due to the different equations of state for the
two components, adiabaticity cannot be maintained on sub-horizon
scales \cite{KS}.
We can then derive from the equations given in \cite{d94}
\[{1\over 4}D_g^{(r)} ={5\over 3}\Phi_C +
{2\over 3}\dot{\Phi}_C/(\dot{a}/a) +\Phi^S ~.\]
Here the Bardeen potentials are split into parts due to
cold dark matter ($_C$) and the scalar field ($_S$) respectively.
For cold dark matter $\Psi_C=-\Phi_C$. Using this, we can bring
Eq.~(\ref{chi}) into the form
\begin{eqnarray} {\delta T\over T}(t_f,\bm{x}_f,\bm{n}) &=&
{1\over 3}\Psi_C(t_i,\bm{x}_i) -
{2\over 3}\dot{\Psi}_C/(\dot{a}/a)(t_i,\bm{x}_i)
+2\int_i^f\dot{\Psi}_C dt \nonumber \\
&& +\Phi^S(t_i,\bm{x}_i) -
3\int_i^f\triangle^{-1}S_{TS}(t,\bm{x}_f-(t_f-t)\bm{n},\bm{n}) dt
~, \label{genu}
\end{eqnarray}
where $S_{TS}$ denote the portion of the source term due to the
scalar field only:
\begin{equation}
S_{TS}= -3n^i\partial^jE^{(S)}_{ij} +n^kn^j\epsilon_{klj}
\partial_lB^{(S)}_{ij} ~. \label{STS}
\end{equation}
Eq. (\ref{genu}) is much better suited
for numerical investigation than the general expression Eq.~(\ref{chi}).
This can be demonstrated by considering the case of pure CDM without
source term: In this case $\Phi_C=-\Psi_C=$ constant and from
Eq.~(\ref{genu}) we easily recover the well--kown result
\[ {\delta T\over T}(t,\bm{x},\bm{n}) =
{1\over 3}\Psi_C(t_i,\bm{x}-\bm{n}(t-t_i)) ~,\]
whereas Eq.~(\ref{chi}) in this case leads to
\[ {\delta T\over T}(t,\bm{x},\bm{n}) = {\delta T\over T}(t_i,\bm{x}_i,\bm{n})
+ 2\Psi_C(t_i,\bm{x}_i) ~.\]
In other words, the unknown initial condition in Eq.~(\ref{chi})
cancels $5/6$ of the naive result for the case of adiabatic CDM
fluctuations. Even though due to the existence of $\dot{\Psi}_C$ terms.
the cancelation is slightly less substantial in our case, the
assumption of adiabaticity on superhorizon scales is a crucial
ingredient of the model.
The electric and magnetic parts of the Weyl tensor are determined by
the perturbations in the energy momentum tensor via Einstein's
equations.
We assume that the source for the geometric perturbations is given by
the scalar field and dark matter. The contributions from radiation
may be neglected. Furthermore, vector perturbations of dark matter
(which decay quickly) are neglected.
The divergence of $E_{ij}$ is then determined by
({\sl see} Appendix~A)
\begin{equation} 3\partial^jE_{ij} = 8\pi G\rho_{C}a^2D_i +
8\pi G(\partial_i\delta T_{00} +3({\dot{a}\over a})\delta T_{0i}
- (3/2)\partial^j\tau_{ij})
\label{dE} ~,
\end{equation}
where the first term on the r.h.s. is the dark matter source term,
$\rho_{C}$ denoting the dark matter energy density. The second
contribution is due to the scalar field: The energy momentum tensor of
the scalar field
\[ T_{\mu\nu}^{S} = \phi,_\mu\phi,_\nu -
{1\over 2}g_{\mu\nu}\phi^{,\lambda}\phi,_\lambda \]
yields
\[\tau_{ij} \equiv T_{ij} -(a^2/3)\delta_{ij}T^l_l =
\tau_{ij}^{S}
= \phi,_i\phi,_j -(1/3)\delta_{ij}(\nabla\phi)^2 ~,\]
\[ \delta T_{0j}= T_{0j}^{S} = \dot{\phi}\phi,_j ~~, \]
\[ \delta T_{00}= T_{00}^{S} = {1\over 2}((\dot{\phi})^2
+ (\nabla\phi)^2) ~,\]
and $D_j$ is a gauge invariant perturbation variable for the density
gradient. For scalar perturbations $ D_j = \partial_jD$.
The evolution equation for the dark matter density perturbation
is given by ({\sl see} \cite{d90})
\begin{equation}
\ddot{D} + ({\dot{a}\over a})\dot{D} - 4\pi Ga^2\rho_{C}D
= 8\pi G \dot{\phi}^2
\label{C} ~. \end{equation}
During the radiation dominated era $8\pi G\rho_RD_R$ in principle
has to be included in Eq.~(\ref{C}). But since radiation
perturbations quickly decay
on sub-horizon scales, and since dark matter fluctuations cannot grow
in a radiation dominated universe \cite{Me}, their influence is not
relevant.
The equation of motion for $B_{ij}$ is more involved. A somewhat
cumbersome derivation ({\sl see Appendix~A}) yields
\begin{equation}
a^{-1}(aB_{ij})^{\cdot\cd} -\triangle B_{ij}
= 8\pi G{\cal S}^{(B)}_{ij} ~, \label{Bij}
\end{equation}
\[ \mbox{with }~~~ {\cal S}^{(B)}_{ij}= -\epsilon_{lm(i}\delta T_{0l},_{j)m}
+\epsilon_{lm(i}\dot{\tau}_{j)l},_m ~.\]
Here $(i...j)$ denotes symmetrization in the indices $i$ and $j$.
To the source term ${\cal S}^{(B)}$
only vector and tensor perturbations contribute. It is thus entirely
determined by the energy momentum tensor of the scalar field.
\section{The System of equations for Global Scalar Field Induced
Fluctuations}
In this section we collect all the equations which determine the
system under consideration. We also repeat equations which have
already been derived in Section~2. Let us begin with the scalar field
equation of motion.
The energy momentum tensor of the scalar field is a small perturbation.
In first order perturbation theory, we can thus solve the equation
of motion of the scalar field in the background, Friedmann--Lema\^{\i}tre
geometry, neglecting geometric perturbations.
The equation of motion for the scalar field $\phi$ is then given by
\begin{equation}
g^{\mu\nu}\nabla_\mu\nabla_\nu\phi + {\partial V\over \partial\phi} = 0 ~,
\label{phi1}\end{equation}
where $g^{\mu\nu}$ denotes the unperturbed metric and $\nabla_\mu$ is
the covariant derivative with respect to this metric. For our numerical
computations, we consider an $O(4)$ model. In $O(N)$ models the
scalar field,
$\phi\in \bm{ R}^N$ and the zero temperature potential is given by
$V_0={\lambda\over 4}(\phi^2-\eta^2)^2$ for some energy scale $\eta$. At high
temperatures, $T>T_c \sim \eta$, one loop corrections to the effective
potential dominate and the minimum of the effective potential is at
$\phi=0$. Below the critical temperature the minimum is shifted (in the
simplest case) to $<\phi^2> =(1-(T/T_c)^2)\eta^2$ ({\sl see} \cite{Ki,d95}
and references therein). The vacuum manifold, i.e. the space of minima
of the effective potential, then becomes a
$(N-1)$--sphere, $\bm{ S}^{(N-1)}$. Since
\[ \pi_k(\bm{ S}^m) = \left\{ \begin{array}{ll}
0~~, & k<m\\
\bm{ Z}~~, k=m, \end{array}\right. \]
the lowest non--vanishing homotopy group of a $m$--sphere is always
$\pi_m$. Since probably higher defects are unstable and decay
into lower ones\footnote{This is an unproven conjecture, motivated,
e.g., by observations of the density of textures and monopoles in
liquid crystals and by numerical experiments \cite{CDTY,Kinew}}, the
$m$--sphere is a suitable vacuum manifold to study $\pi_m$ defects.
If the system under consideration is at a temperature $T$ much below the
critical temperature, $T\ll T_c$, it becomes more and more improbable for
the field $\phi$ to leave the vacuum manifold. $\phi$ will leave the
vacuum manifold only if it would otherwise be forced to gradients of
order $ (\nabla \phi)^2 \sim \lambda\phi^2\eta^2$, thus only over length scales
of order $l=1/(\sqrt{\lambda}\eta) \equiv m_\phi^{-1}$ ($l$ is the transversal
extension of the defects).
For GUT scale phase transitions $l \sim 10^{-30}$cm
as compared to cosmic distances of the order of Mpc $\sim 10^{24}$cm.
If we are willing to loose the information of
the precise field configuration over these tiny regions,
it seems well justified to fix $\phi$ to the vacuum manifold $\cal N$.
Instead of discussing the field equation Eq.~(\ref{phi1}),
we require $ \phi/\eta \in \bm{ S}^{(N-1)}$. The remaining field
equation, $\Box\phi=0$, then demands that
\[ \phi/\eta\equiv \beta ~:~ {\cal M} \rightarrow \bm{ S}^{(N-1)} \]
is a harmonic map from spacetime $\cal M$ into {\bf S}$^{(N-1)}$.
The topological defects we are interested in are singularities of these
maps. When the gradients of $\phi$ become very large, like, e.g., towards
the center of a global monopole, the field leaves the vacuum manifold
and assumes non vanishing potential energy. If
$\beta\in \bm{ S}^{(N-1)}$ is enforced, a singularity develops by
topological reasons.
In the physics literature harmonic maps are known as $\sigma$--models.
The action of a $\sigma$--model is given by
\begin{equation}
S_\sigma = \int_{\cal M}g^{\mu\nu}\partial_\mu\beta^A\partial_\nu\beta^B\gamma_{AB}(\beta)
\sqrt{|g|}d^4x ~, \label{Ssi}
\end{equation}
where $\gamma_{AB}$ denotes the metric on $\bm{ S}^{N-1}$ and
$g_{\mu\nu}$ is the metric of spacetime.
We now fix $\beta$ to lay in the vacuum manifold, $\bm{ S}^{N-1}$
by introducing a Lagrange multiplier. We then obtain the
following equation of motion for $\beta$:
\begin{equation}
\Box\b - (\b\cdot\Box\b)\b =0 ~, \label{si2}
\end{equation}
which shows that the $\sigma$--model is scale free.
There are thus two possible evolution equations for the scalar field
at low temperature. We call Eq.~(\ref{phi1}) the 'potential model' evolution
equation and Eq.~(\ref{si2}) the $\sigma$--model approach.
The energy momentum tensor of the scalar field perturbs spacetime
geometry and induces perturbations in the dark matter energy density
according to Eq.~(\ref{C})
\begin{equation}
\ddot{D} + ({\dot{a}\over a})\dot{D} - 4\pi Ga^2\rho_{C}D
= 8\pi G \dot{\phi}^2
\label{D..} ~, \end{equation}
where $D$ is a gauge invariant variable for the dark matter
perturbations \cite{d90}. On subhorizon scales $D \sim \delta\rho/\rho$.
In comoving coordinates, the total perturbed energy momentum tensor is
given by
\[ \delta T_\mu^\nu = \phi,_\mu\cdot\phi^{,\nu}-{1\over 2}\delta_\mu^\nu
\phi,_\lambda\cdot\phi^{,\lambda} +\rho_{C}D\delta_\mu^0\delta_0^\nu ~.\]
As already mentioned in section~2, the perturbed Einstein equations to
this energy momentum tensor
yield an algebraic equation for the divergence of the electric part
of the Weyl tensor and an evolution equation for the magnetic part
of the Weyl tensor ({\sl see Appendix~A}):
\begin{equation} \partial^jE_{ij} = -{8\pi\over 3} G\rho_{C}a^2D_i -
8\pi G({1\over 3}\partial_i\delta T_{00} +({\dot{a}\over a})\delta T_{0i}
+ {1\over 2}\partial^j\tau_{ij})
\label{djE} ~,\mbox{ and}
\end{equation}
\begin{equation}
{1\over a}(aB)_{ij})^{\cdot\cd} + -\triangle B_{ij}
= 8\pi G{\cal S}^{(B)}_{ij} ~, \label{B..}
\end{equation}
\[ \mbox{with }~~~ {\cal S}^{(B)}_{ij}=
\epsilon_{lm(i}[T^{S}_{0l},_{j)m} +
\dot{\tau}_{j)l},_m]~, ~~~\mbox{ and }~~~~
\tau_{ij}=\phi,_i\phi,_j-{1\over 3}\delta_{ij}(\nabla\phi)^2 ~.\]
The source term for the perturbation of Liouville's equation is given
by Eq.~(\ref{STS}):
\begin{equation} -3n^i\partial^jE^{(S)}_{ij} + n^kn^j\epsilon_{klj}
\partial_lB^{(S)}_{ij} \equiv S_{ST}(t,\bm{x},\bm{n})~ .
\label{sou} \end{equation}
The CMB fluctuations are then determined according to
\begin{eqnarray}
{\delta T\over T}(t,\bm{x},\bm{n}) =
\int_{t_i}^t S_{ST}(t',\bm{x}+(t'-t)\bm{n}, \bm{n})dt'
~ + ~ \Phi_S(t_i,\bm{x}+(t_i-t)\bm{n}, \bm{n}) ~,
\nonumber \\
+ {1\over 3}\Psi_C(t_i,\bm{x}_i) -
{2\over 3}\dot{\Psi}_C/(\dot{a}/a)(t_i,\bm{x}_i)
+2\int_i^f\dot{\Psi}_C dt
\label{chi2} \end{eqnarray}
Eqs.~(\ref{phi1}) and (\ref{D..}) to (\ref{chi2}) form a
closed, hyperbolic system of partial differential equations. Actually
all except the scalar field equation Eq.~(\ref{phi1}) are linear perturbation
equations with source term and can thus be solved, e.g., by the
Wronskian method, i.e., by some integrals over the source term.
The corresponding solution for $\delta T/T$ is given above in
Eq.~(\ref{chi2}),
the general solution of the dark matter equation is given below in
Eqns.~(\ref{D2}), (\ref{Dgen}) and (\ref{wrons}).
Let us briefly describe the general solution for $B_{ij}$: We switch to
Fourier space, because there the $\triangle$ is a simple
multiplication by $-k^2$ and Eq.~(\ref{B..}) becomes an ordinary
differential equation with scalar homogeneous solutions
\begin{equation}
b^{\pm} = {1\over a}\exp(\pm ikt) ~.
\end{equation}
The general solution to the inhomogeneous equation is given by
\begin{equation}
B_{ij}=(b^{+}C^{+}_{ij} +b^{-}C^{-}_{ij}) + B^{(hom)}_{ij}
~, \label{BW}\end{equation}
where $B^{hom}$ denotes an arbitrary homogeneous solution and
$C^{+}$, $C^{-}$ are given by
\begin{eqnarray}
C^{+}_{ij} &=& -8\pi G\int{\tilde{\cal S}^{(B)}_{ij}b^- \over W}dt\\
C^{-}_{ij} &=& 8\pi G\int{\tilde{\cal S}^{(B)}_{ij}b^+ \over W}dt~.
\end{eqnarray}
Here $W$ denotes the Wronskian determinant of the solutions which
amounts to
\begin{equation}
W=b^+\dot{b}^--b^-\dot{b}^+ = {2ik\over a^2} ~.
\end{equation}
\section{Initial Conditions and Numerical Methods}
\subsection{The scalar field:}
As already shown in the previous section, the equation of motion of the
scalar field is given
by
\begin{equation}
g^{\mu\nu}\nabla_\mu\nabla_\nu\phi + {\partial V\over \partial\phi} = 0 ~,
\label{phi}\end{equation}
where $g^{\mu\nu}$ is the background (unperturbed metric).
With $\beta = \phi/\eta$ and $m=\sqrt{\lambda}\eta$, Eq.~(\ref{phi})
yields for $O(N)$ models in a Friedmann universe
\begin{equation}
\partial_t^2\beta + 2(\dot{a}/a)\partial_t\beta -\nabla^2\beta =
{1\over 2}a^2m^2(\beta^2-1)\beta ~. \label{bb}
\end{equation}
This equation as it stands can not be treated numerically in the regime
which is interesting for large scale structure formation. The two scales
in the problem are the horizon scale $t\sim (\dot{a}/a)^{-1}$ and the
inverse symmetry
breaking scale, the comoving scale $(am)^{-1}$. At recombination, e.g.,
these scales differ by a factor of about $10^{53}$ and can thus not
both be resolved in one computer code.
There are two approximations to treat the scalar field numerically.
As we shall see, they are complementary and thus the
fact that both approximations agree with each other within about 10\%
is reassuring.
The first possibility is to replace $(am)^{-1}$ by $w$, the smallest
scale which can be resolved in a given simulation, typically twice the
grid spacing, $w \sim 2\Delta x$. The time dependence of $(am)^{-1}$ which
results in a steepening of the potential is mimicked by an additional
damping term: $ 2(\dot{a}/a) \rightarrow \alpha \dot{a}/a$, with
$\alpha \sim 3$ \cite{PSR}.
Numerical tests have shown, that this procedure, which usually is
implemented by a modified staggered leap frog scheme \cite{NR}, is not
very sensitive on the values of $\alpha$ and $w$ chosen. With this method we
have replaced the growing comoving mass $am$ by the largest mass which
our code can resolve. For a $(256)^3$ grid which simulates
the evolution of the scalar field until today, we obtain
$256\Delta x \sim t_0 \sim 4\times10^{17}$sec$/a_0$, so that
$w\sim 4\times 10^{15}$sec$/a_0$,
i.e., $am\sim \eta a_{rec}\sim 10^{17}GeV$ is replaced by about
$w^{-1}= a_010^{-39 }GeV \sim 10^{-35}GeV$, where we set
$a_{eq}=1$.
We are confident that this modified equation mimics the behavior of
the field, since the actual mass of the scalar field is irrelevant as
long as it is much larger
than the typical kinetic and gradient energies associated with the field
which are of the order the inverse horizon scale. Therefore, as soon as
the horizon scale is substantially larger than $\Delta x$, the code should
mimic the true field evolution on scales larger then $w$. But, to our
knowledge, there exists no rigorous mathematical approximation scheme
leading to the above treatment of the
scalar field which would then also yield the optimal choice for $\alpha$.
Alternatively, we can treat the scalar field in the $\sigma$--model
approximation given in the previous section. This approach is opposite
to the one outlined above in
which the scalar field mass is much too small, since the $\sigma$--model
corresponds to setting the scalar field mass infinity.
The $\sigma$--model equation of motion cannot be treated numerically
with a leap frog scheme, since it involves non--linear time
derivatives. In this case, a second order accurate integration scheme
has been developed by varying the discretized action with respect to the
field \cite{PST}.
The two different approaches have been extensively tested by us and
other workers in the field, and good
agreement has been found on scales larger than about 3 -- 4 grid sizes
\cite{BCL1,B}. We have compared our potential code with the exact
spherically symmetric scaling solution \cite{TS} and with our old
spherically symmetric $\sigma$--model code \cite{DHZ}. Outside the
unwinding events which extend over approximately 3 grid sizes, the
different approaches agree within about 5\%.
This is very encouraging, especially since the two
treatments are complementary: In the $\sigma$--model, we let the scalar
field mass $m$ go to infinity. In the potential approach, we replace
$m$ by
$\sim 1/\Delta x \sim 200/t_{0} \sim 200a_0/10^{10}y \sim 10^{-35}$GeV.
The integration of the scalar field equation is numerically the hardest
part of the problem, since it involves
the solution of a system of nonlinear partial differential equations.
A good test of our numerical calculations, next to checking the
scaling behavior of $\rho_S$, is energy momentum conservation of the
scalar field, $T^{(S)\mu\nu}_{;\nu}=0$.
Energy momentum conservation in the potential model, about 15\%, is
slightly worse than in the $\sigma$--model, where it is about 5\%
({\sl see Fig.~2}).
Therefore, the final results presented here are all obtained with
the $\sigma$--model approach. Our checks lead us to the
conclusion, that we can calculate the scalar field energy momentum
tensor, which then is the source of dark matter and CMB fluctuations
to an accuracy of about 10\%.
The problem of choosing the correct initial condition may induce
another (systematical) error in our calculations which we hope
to remain below 20\%. Other sources of error are negligible.
\subsection{Dark matter}
Once the scalar field $\beta(\bm{x},t)$ is known, the dark matter
perturbations can easily be calculated by either
using the Wronskian method (see below) or some standard ordinary
differential equation solver. We have performed both methods and they
agree very well. For later use, we briefly describe the Wronskian
method.
We normalize the scale factor by
\[ a= {t\over \tau}(1+{1\over 4}t/\tau) ~~~, \mbox{ with} \]
\[ \tau=1/\sqrt{(4\pi G/3)\rho_{eq}}= {t_{eq}\over 2(\sqrt{2}-1)}~.\]
Here $t_{eq}$ denotes the time of equal matter and radiation density,
$ \rho_{rad}(t_{eq})=\rho_{C}(t_{eq})=(1/2)\rho(t_{eq})$. We have
normalized $a$ such that $a_{eq}=a(t_{eq})=1$.
Transformed to the variable $a$, the dark matter equation Eq.~(\ref{C}) then
yields
\begin{equation}
{d^2 D\over da^2} + {2+3a\over 2a(1+a)}{d D\over d a} -{3\over 2a(1+a)}D
= 2\epsilon \dot{\beta}^2({ da\over dt})^2 = (1+a)S/\tau^2 ~,
\label{da}\end{equation}
\[ S=2\epsilon\dot{\beta}^2 ~ \mbox{ and } \epsilon=4\pi G\eta^2 ~.\]
The homogeneous solutions to this linear differential equation are
well known \cite{GP}:
\begin{eqnarray} D_1 &=& 1+{3\over 2}a ~, \label{D1}\\
D_2 &=& (1+{3\over 2}a)[\ln\left({\sqrt{a+1}+1\over\sqrt{a+1}-1}
\right) -3\sqrt{a+1}]~. \label{D2}
\end{eqnarray}
The general solution to Eq.~(\ref{da}) is given by
\begin{equation}
D(t)=c_1(t)D_1(t) + c_2(t)D_2(t) \label{Dgen}
\end{equation}
with
\begin{equation}
c_1 = -\int(SD_2/W)dt \;\;\mbox{ , }\;\;
c_2 = \int(SD_1/W)dt \;\; .
\label{wrons} \end{equation}
\[ W =D_1\dot{D_2} - \dot{D_1}D_2 =
{\dot{a}(2a-1)\over a\sqrt{a+1}} = {2a-1\over a\tau} \]
is the Wronskian determinant of
the homogeneous solutions. The integrals Eq.~(\ref{wrons}) have to be
performed numerically with $S=\epsilon\dot{\beta}^2$. When discussing the
initial conditions for $D$ in subsection~4.4, we shall present an
analytic approximation for the source term $S$.
\subsection{The CMB anisotropies}
The CMB anisotropies are given by
\[ {\delta T\over T} = \triangle^{-1}\chi \]
up to monopole and dipole contributions which we disregard.
Here $\chi$ is a solution Eq.~(\ref{chi}) of
Eq.~(\ref{Lchi}). The source term $S_T$ is determined via Eq.~(\ref{dE})
and Eq.~(\ref{Bij}). However, using this straightforward approach results
in a big waste of computer memory (which we cannot afford): We would
be satisfied to calculate $\delta T/T$ for about 30 observers in each
simulations, which means we need $\triangle^{-1}\chi$ only at 30
positions {\bf x}. But since we have to perform an inverse Laplacian
which is done by fast Fourier transforms, we have to calculate $\chi$
on the whole grid, which consists of $192^3 \sim 8\cdot 10^6$ positions. In
Addition, to calculate the spherical harmonic amplitudes of $\delta T/T$
up to about $l\sim 40$ (angular resolution of about $4^o$), we need
typically $5\times10^4$ directions $\bm{n}$. The $\chi$ variable alone
(in double precision) would thus require 16 G-bytes of memory, an
amount which is not available on most present day computers. The way
out is to take the inverse Laplacian already in the equation of
motion Eq.~(\ref{Lchi}). This results in
\begin{equation}
(\partial_t +\gamma^i\partial_i){\delta T\over T} = -3\gamma^i\triangle^{-1}(\partial^jE_{ij})
- \gamma^k\gamma^j\epsilon_{kli}\triangle^{-1}(\partial_lB_{ij}) \equiv
\triangle^{-1}S_T(t,\bm{x},\bm{\gamma})~ .
\label{LT} \end{equation}
Here the inverse Laplacian has to be performed for a vector field and a
symmetric traceless tensor field, a total of 8 scalar variables which
only depend on {\bf x} and not on $\bm{\gamma}$. For a $192^3$ grid
this requires nearly 1 Gbyte of memory, no problem for presently
available machines. Eq.~(\ref{LT}) has the general solution ({\sl see}
Eq.~(\ref{genu}))
\begin{eqnarray}
{\delta T\over T}(t_0,\bm{x}_0,\bm{\gamma}) &=&
+\Phi^S(t_i,\bm{x}_i) -
3\int_i^f\triangle^{-1}S_{TS}(t,\bm{x}_0-(t_f-t)\bm{n},\bm{n}) dt
\nonumber \\ &&
+ {1\over 3}\Psi_C(t_i,\bm{x}_i) -
{2\over 3}\dot{\Psi}_C/(\dot{a}/a)(t_i,\bm{x}_i)
+2\int_i^f\dot{\Psi}_C dt
~,\label{dT} \end{eqnarray}
with $S_{TS}$ given in Eq.~(\ref{STS}).
The first term of Eq.~(\ref{dT}) determines the initial condition of
the CMB anisotropies due to the source term. In the numerical simulation
we just set it $ -3n^l\triangle^{-1}(\partial^jE^{(S)}_{lj})(t_i,\bm{x}_i)t_i$.
This assumes that
the source term is approximately constant until $t_i$ and that
magnetic contributions can be neglected. The resulting amplitude is not
very sensitive to this assumption, but changing it can somewhat change
in the spectral index.
We have solved Eq.~(\ref{dT}) numerically by just summing
up the contributions from each time step for 27 observer positions
$\bm{ x}_0$. The value of the source term at position
$\bm{x}_0+(t-t_0)$ is determined by linear interpolation. The
quantity $\partial^iE^S_{ij}$ is determined by Eq.~(\ref{dE}) and its inverse
Laplacian is calculated by fast
Fourier transforms. To obtain $\triangle^{-1}B^S_{ij}$ from equation
Eq.~(\ref{Bij}), we directly calculate
$\triangle^{-1}{\cal S}_B$ in $k$--space, then solve the ordinary, linear
differential equation for $\triangle^{-1}B_{ij}$ in $k$--space
by the Wronskian method. Since, as we shall argue later, all components
$T^{(S)}_{\mu\nu}$ in average scale like $A/\sqrt{t}$ on super horizon
scales, $S^(B) \sim At^{-3/2}$ and therefore $C^{\pm} \propto t^{3/2}$
on super horizon scales. Therefore, we can neglect the contribution
to $C^\pm$ from the lower boundary in the integral. Furthermore, since
the homogeneous solution $B_{ij}^{(hom)}$ is decaying, we drop it
entirely. This procedure corresponds to setting $B(t_{i})=0$ and
calculating $B(t)$ according to Eq~(\ref{BW}).
\subsection{Initial conditions}
Initially, the field $\phi$ itself and/or the velocities $\dot{\phi}$ are
laid down randomly on the grid points. The initial time, $t_{in}$ is
chosen to be the grid size, $t_{in} = \Delta x$, so that the field at
different grid points should not be correlated. The configuration is
then evolved in time with one of the approximation schemes discussed
above.
Because our initial conditions for dark matter and photons very
sensitively depend on the scaling behavior of the dark matter,
we can only start the dark matter or photon simulations when scaling
is fully reached, $t_{in}=8\Delta x$. Starting our simulations, e.g., at
$t=4\Delta x$ changes the results by about a factor of 2.
Further doubling of the initial time, changes our results by less
than 20\%, we thus believe that at $t=8\Delta x$ scaling is sufficiently
accurately. Unfortunately, this late initial time reduces our
dynamical range to about $192/8=24$ for a $192^3$ grid, which is seen
clearly in our results for the CMB anisotropies discussed below.
It is very important to choose the correct initial conditions for the
dark matter and, especially, photon perturbations. Changing them
can change the CMB fluctuation amplitudes by nearly a factor two.
Since these
fluctuations are used to normalize the model, i.e. to determine $\epsilon$,
this reflects in corresponding changes in $\epsilon$. We want to do better
than a factor two by choosing physically plausible initial conditions.
The cleanest way would by to simulate the evolution of perturbations
through the phase transition, assuming that before the phase transition,
the universe was an unperturbed Friedmann universe with $\phi\equiv 0$.
On the other hand, since we want to calculate the perturbation
spectrum on scales of up to 1000Mpc with a $(256)^3$ grid,
we cannot start our dark matter and CMB simulation earlier than at a
time when the horizon
distance is approximately $8\Delta x \sim 30$Mpc. At the beginning of the
scalar field simulation our grid scale $\Delta x\sim 4$Mpc is of the order
of the horizon scale. We therefore have
to decide on the amplitudes of super horizon perturbations. One
possibility is setting all geometrical
perturbations initially to zero. The requirements
\[ \partial^jE_{ij}(t_{i})=0 ~~\mbox{ and }~~~ S^B_{ij}(t_{i})=0\]
then yield initial conditions for the dark matter fluctuations $D$
and the photon variable $\chi$. But these, let us call them 'strict
isocurvature' initial conditions are not natural since they do
not propagate in time: Even if we start with $E$ and $B$ vanishing on
super horizon scales, after some time residual fluctuations have leaked
into these scales and one obtains the white noise fluctuations spectrum
on super horizon scales shown in {\sl Fig.~3}.
This does not violate causality, since white noise is uncorrelated
and just results from the residuals of correlated fluctuations on smaller
scales. The correct initial values for $D$ and $\dot{D}$ would of course
be those obtained by solving the equation of motion Eq.~(\ref{C}) from the
symmetry breaking time until the start of the simulation. We found a
method to incorporate this at least approximately: The spectrum of the
dark matter source term
$8\pi G| \widetilde{\dot{\phi}^2}|^2 $ can be approximated by
\begin{equation}
8\pi G |\widetilde{\dot{\phi}^2}| =2\epsilon \widetilde{\dot{\beta}^2} =
\epsilon\sqrt{1\over V}\int d^3x\dot{\beta}^2(x)e^{ikx} \approx
{\epsilon A\over \sqrt{t}(1+a_1kt +a_2(kt)^2)} ~, \label{fit}\end{equation}
with
\[ A= 3.3 ~,~~~ a_1=-0.7/(2\pi)~,~~~ a_2 = 0.7/(2\pi)^2 ~. \]
This numbers have been obtained by a $\chi^2$--minimization scheme. The
approximation is not very good. It yields a $\chi^2 \approx 2000$
for about 1000 data points. Its comparison with the real data
in {\sl Figs.~4 and 5} shows that Eq.~(\ref{fit}) approximates the
source term
to about 10\% on superhorizon scales, but does not follow the wiggles
present in the data on smaller scales. Since we shall not use the fit
on subhorizon scales, this is not important for our simulations.
However, in general $\widetilde{\dot{\phi}^2}$ is complex and setting it
equal to its absolute value, we neglect the evolution of phases.
Again, by causality, this will not severely affect scales larger than
the horizon, since on these scales the phases are (approximately)
frozen. But on subhorizon scales our fis is nor very useful due to
the incoherent evolution of phases.
Assuming this form of the source term, we can solve Eq.~(\ref{C})
analytically on super horizon scales, where we approximate the
source term by
\begin{equation} 2\epsilon \widetilde{\dot{\beta}^2} = {\epsilon A\over \sqrt{t}},
~~~~~\mbox{ on super horizon scales.} \label{sqrtt}\end{equation}
The homogeneous solutions of Eq.~(\ref{C}) are given by Eq.~(\ref{D1},\ref{D2})
The general inhomogeneous solution, $D=c_1D_1 +c_2D_2$, even with the
simple source term Eq.~(\ref{sqrtt}), becomes rather complicated. But
in the radiation and matter dominated regimes we find the simple
approximations
\begin{eqnarray}
D &=& (4/7)t^2S ~~;~~ \dot{D}=(6/7)t^2S ~~~
\mbox{ radiation dominated} \label{Dinr} \\
D &=& -(4/9)t^2S ~~;~~ \dot{D}=-(2/3)t^2S ~~~\mbox{ matter dominated.}
\label{Dinm}\end{eqnarray}
From $D$ we can calculate $\Psi_C$, leading to the dark matter
contribution to the CMB anisotropies.
As mentioned above, the initial contribution of the scalar field is
approximated by
\[ \Phi^S(t_i,\bm{x}_i) \sim -3t_in^i\triangle^{-1}(\partial^jE^{(S)}_{ij}) ~.\]
The result is very sensitive to initial conditions: If we do not
separate the dark matter and choose some arbitrary, non-adiabatic
initial condition, the resulting $C_\ell$'s increase by nearly a factor
of $10$ and the dark matter induces $80\%$ of the total fluctuation.
However, choosing the adiabatic initial condition discussed in
Section~3, leading to Eq.~(\ref{dT}), dark matter only contributes about
20\% to the $C_\ell$'s and the main contribution is due to the defects.
The dark matter contribution to the CMB anisotropies is not scale
invariant, but white noise. It has spectral index $n=0$. This result
was found numerically ({\sl see Fig.~6}) but, as we argue in Appendix~B, it
can also be understood analytically.
our value of $\epsilon$ obtained with these physical isocurvature and
on super horizon scales adiabatic initial
conditions is in reasonable agreement with the values obtained in
\cite{PST} and \cite{BR}.
Let us also present a heuristic derivation of the numerical finding
Eq.~(\ref{sqrtt}) on superhorizon scales: We know that the average value
$\langle\dot{\beta}^2\rangle \propto 1/t^2$, the usual scaling
behavior. The Fourier transform of $\dot{\beta}^2$ determines the
fluctuations on this 'background' on a given comoving
scale $\lambda=2\pi/k$. As long as this scale is super horizon, $\lambda>t$,
a patch of size $\lambda^3$ consists of $N=(\lambda/t)^3$ independent horizon
size volumes. The fluctuations on this scale should thus be
proportional to
\[\tilde{\dot{\beta}^2}\propto \langle\dot{\beta}^2\rangle /\sqrt{N}
\propto 1/\sqrt{t} ~, \]
which is just the behavior which we have found numerically on
super--horizon scales.
As soon as a given scale becomes sub--horizon, $\lambda\ll t$,
$\tilde{\dot{\beta}^2}$ starts decaying from this large scale value like
$1/t^2$.
\section{Results}
\subsection{CMB anisotropies}
To analyze the CMB anisotropies, we expand $\delta T/T$
in spherical harmonics
\begin{equation} {\delta T\over T}(t_0,\mbox{\boldmath{$x,\gamma$}}) =
\sum_{lm}a_{lm}(\mbox{\boldmath{$x$}})Y_{lm}(\mbox{\boldmath{$\gamma$}})
~. \end{equation}
As usual, we assume that the average over $N_x$ different observer
positions coincides with the ensemble average and define
\begin{equation} C_\ell = {1\over (2\ell+1)N_x} \sum_{m,x}
|a_{\ell m}(\mbox{\boldmath{$x$}})|^2 ~, ~~ \ell\ge 2
\label{cl} ~. \end{equation}
Gaussian fluctuations are characterized by the two point
correlation function. Since the angular two point correlation
function is given by
\begin{equation} \langle {\delta T\over T}(\bm{ n}){\delta T\over T}(\bm{ n}')
\rangle_{(\bm{ n\cdot n}'=\cos\theta)} ={1\over 4\pi} \sum_\ell(2\ell+1)
C_\ell P_\ell(\cos\theta) ,\end{equation}
Gaussian distributed CMB fluctuations are fully
determined by the $C_\ell$'s.
However, as can be seen from Fig.~7, in our case the distribution of
the CMB fluctuations is not quite Gaussian. It is slightly negatively
skewed. We find an average skewness of $-0.5$ and a kurtosis of
$0.7$.
For a simulation on a $192^3$ grid with 27 different observer
positions for each simulation. The harmonic amplitudes with
are shown in Fig.~8.
The low order multipoles depend strongly on the random initial
conditions (cosmic variance), like in the spherically symmetric
simulation \cite{DHZ}.
It is well known, that cold dark matter fluctuations with a power
spectrum of spectral index $n$ gravitationally induce CMB anisotropies
with a spectrum given by \cite{Ef}
\begin{equation} C_\ell = C_2{\Gamma(l+(n-1)/2)\Gamma((9-n)/2)\over \Gamma(l+(5-n)/2)\Gamma((n+3)/2)}
\label{spec} ~.\end{equation}
We have performed a least square fit of our numerical results for
$\log(C_\ell)$ and the $\log(C_\ell)$ obtained from
Eq.~(\ref{spec}).\footnote{In the case of
topological defect induced fluctuations, the $C_\ell$ spectrum does not
have exactly this form, since CBM fluctuations are not only induced
by the dark matter but mainly by the scalar field perturbations and the
assumptions made for the derivation of this formula are not valid.}
If we take into account
all the $C_\ell$'s reliably calculated in our simulations, which
limits us approximately to $\ell\le 22$ we find a very nice scale
invariant spectrum,
\begin{equation} n= 0.9\pm 0.2 \end{equation}
with quadrupole amplitude
\begin{equation} Q =\sqrt{(5/4\pi)C_2}T_{CMB} = 2.8\pm 0.7 K \cdot \epsilon ~. \label{Q}\end{equation}
The 1, 2 and 3 sigma contour plot is shown in Fig.~9. The minimal
$\chi^2$ is 0.56.
It is very interesting, that the dark matter contribution to the CMB
anisotropies does not yield a scale invariant spectrum, but white noise.
This can be understood analytically: The $\dot{\Psi}_C$ contributions
to $\delta T/T$ in Eq.~(\ref{genu}) are not very important and
\[ (\delta T/T)_C(t_0,\bm{k}) \sim
{1\over 3}\Psi_C(t_i,\bm{k})\exp(i\bm{k\cdot nt_0}) =
{\epsilon A\over 6\sqrt{t_i}}{\exp(i\bm{k\cdot nt_0})\over k^2} \]
on super horizon scales. For the second equal sign we used
$k^2\Psi_C=4\pi GD_C\sim {\epsilon A\over 2\sqrt{t_i}}$.
By standard arguments ({\sl see, e.g.} \cite{Ef}) one then finds
\begin{equation}
C_\ell^{(C)} = {\epsilon^2A^2 \over 18\pi t_i}\int{dk\over k^2}j_\ell^2(kt_0)
\propto {\Gamma(\ell-0.5)\over \Gamma(\ell+2.5)} ~,
\end{equation}
corresponding to Eq.~(\ref{spec}) with $n=0$. This is also what we
find numerically ({\sl see Fig.~6}). The dark matter contribution
caused the spectral index $n$ of the total CMB anisotropies to drop
slightly below $n=1$.
To reproduce the COBE amplitude $ Q_{COBE}= (20 \pm 5)\mu K$
\cite{Go}, we have to normalize the spectrum by choosing the phase
transition scale $\eta$ according to
\begin{equation} \epsilon = 4\pi G\eta^2 = (0.8 \pm 0.4)10^{-5}
\label{ep}~. \end{equation}
This value is somewhat smaller, but still comparable with the results
obtained in \cite{BR,PST}.
Another method to determine $\epsilon$ is the following: The total
temperature fluctuation amplitude on a given angular scale $\theta_c$ is
given by
\begin{equation}
\sigma^2_T(\theta_c) = {1\over 4\pi}
\sum_\ell C_\ell(2\ell+1)\exp(-\ell^2\theta_c^2/2)
~. \end{equation}
In Fig.~10 we show $\sigma_T$ as a function of $\theta_c$. In a recent
analysis of the COBE data\cite{WB} $\sigma_T^{(COBE)}(7^o) \sim 44\mu K$ and
$\sigma_T^{(COBE)}(10^o) \sim 40\mu K$ for a spectral index $n\sim 1$,
which leads again to the result given in Eq.~(\ref{ep}).
\subsection{Dark matter fluctuations}
Using fast Fourier transforms we calculate the spectrum
$P(k) = |\delta(k)|^2$ of the dark matter density fluctuations
is shown in Fig.~11.
The fit shown as solid line in Fig.~11 is given by
\begin{equation}
P(k)h^3/(2\pi)^3= {Ck\over (1 +\alpha k + (\beta k)^{1.5} +(\gamma k)^2)^2}
~, \label{Pk} \end{equation}
with $h=0.5$ and
\begin{eqnarray}
C &=& 215h^{-1}Mpc^4 \label{CC}\\
\alpha &=& 10 h^{-2}Mpc ~~(=0.5\tau) \\
\beta &=& 1.25h^{-2} Mpc ~~ \sim \tau/(4\pi) \\
\gamma &=& 2.3 h^{-2} Mpc ~~ \sim \tau/(2\pi) ~,
\end{eqnarray}
where we have used $\tau=19.36h^{-2}$Mpc which is approximately the
comoving time at equal matter and radiation.
The parameter $C$, which is most important to determine the bias factor
can also be obtained by the following rough analytical argument:
On super horizon scales, $|D|^2 \sim (0.5\epsilon A)^2t^3$ according to
Eq.~(\ref{Dinm}). As soon as the perturbation enters the horizon at
$t=2\pi/k$, the
source term disappears and $D$ starts growing like $t^2$, leading to
\begin{equation}
P(k,t_0) \sim {(0.5\epsilon A)^2\over 2\pi}kt_0^4 =
{(2\pi)^3\over h^3}C_{an}k ~.
\end{equation}
Inserting the numbers $\epsilon=0.8\times 10^{-5}$, $A=3.3$,
$t_0^2 =4a_0\tau^2$,
$ a_0\sim 2.5h^2\times 10^4$, we obtain $C_{an}\sim 190$ in excellent
agreement with Eq.~(\ref{CC}).
Figure~11 can be compared directly with the IRAS observation
\cite{IRAS} and it is compatible with a bias factor of order 1.
A more detailed calculation with Gaussian or square hat window function
yields for $\epsilon=0.8\times 10^{-5}$
\begin{eqnarray}
\sigma_{sim}(10MPc)= 1/b_{10} \sim 0.5 - 1 & \sigma_{QDOT}(10MPc)= 1&
\mbox{for $h=0.5$},
\end{eqnarray}
yielding $b_{10}\sim 1 - 2 $ for the value of $\epsilon$ found by
comparison with COBE Eq.~(\ref{ep}) A value even somewhat closer to
1 is found for $b_{20}$.
Observations and simulations of nonlinear clustering of dark matter
and baryons \cite{CO} suggest a bias factor $b_{10}\sim 1 - 2$ which is
compatible with our results. It is remarkable, that unlike in the
simulations by Pen et al. \cite{PST}, our bias factor is approximately
constant and physically acceptable. (To determine our power spectrum,
we have not taken into account
any smoothing which might change the results by at most 15\%.)
In Fig.~12 we have shown the dark matter pixel distribution from a
$100^3$ simulation. It is interesting, that the skewness of the dark
matter distribution is positive, where the $\delta T/T$ skewness
is negative.
\section{Conclusions}
Our simulations show that global texture lead to a scale invariant
spectrum of microwave background fluctuations on large scales like
inflationary models of structure formation. This is one of the main
results of this investigation. It is however interesting
that the dark matter contribution to the CMB anisotropies is not scale
invariant, but white noise. Therefore it is important that the initial
condition for dark matter and radiation are adiabatic in which case
the dark matter contribution to the $C_\ell$'s is small and the
flat spectrum caused by the defects is maintained.
Our second main result is the dark matter fluctuation spectrum.
The spectrum is very close to scale invariant and the bias factor
needed for $\epsilon$ from the CMB anisotropies is around $b\sim$ 1--2.
This value is certainly acceptable and smaller than the bias factor
obtained in previous investigations \cite{CST}.
The deviation from Gaussian statistics seems to us not very
significant ({\sl see Figs.~7 and 12}) and it is thus important
to develop other means to distinguish topological defects from
inflationary scenarios. A clean and promising candidate for this
distinction are the Doppler peaks which are calculated for the
texture scenario in \cite{DGS}.
From our investigations we thus conclude that concerning the large
scale CMB anisotropies and the linear dark matter perturbation spectrum
the texture scenario and probably also other models with global defects
are compatible with present observations.
\vspace{2cm}
\noindent {\large\bf Acknowledgment}
We thank the staff at CSCS, for valuable
support. Especially we want to mention Andrea Bernasconi, Djiordic
Maric and Urs Meier. We acknowledge discussions with Joachim Laukenmann
who carefully read the manuscript and checked the more involved
algebraic derivations with Maple. We profited from many helpful and
encouraging discussions of this work especially we want to mention
Marc Hindmarsh, Philipp Jetzer, Yipeng Jing, Mairi Sakellariadou,
Norbert Straumann, Neil Turok, Simon White and others.
We also thank Neil Turok for providing us their $\sigma$--model code.
|
\section{Introduction}
Let F to be $R,C,H$.The inner product on $F^{n+1}$ is given as;
$$<(z_1,\cdots,z_{n+1}),(w_1,\cdots,w_{n+1})>=
\sum_1^n{z_i\bar{w_i}}-z_{n+1}\overline{w_{n+1}}$$
Let $GL(n+1,F)$ act on $F^{n+1}$ from the right and let $P^n_F$ be the space of
left $F$-lines.
Then $G=O_F(n,1)$ is a subgroup of $GL(n+1,F)$ preserving this inner product.
The induced action of $O_F(n,1)$ on $P^n_F$ gives three invariant subsets
$D_+,D
_0,D_-$ and the action is positive definite on $D_+$.A component of $D_+$ is
called hyperbolic space.An isotropy group of $O_F(n,1)$ is a maximal compact
subgroup
$K=O_F(1) \times O_F(n)$.If we numerate them;
$$ O_R(n,1)=O(n,1),K=O(1)\times O(n)$$
$$ O_C(n,1)=U(n,1),K=U(1)\times U(n)$$
$$ O_H(n,1)=Sp(n,1),K=Sp(1) \times Sp(n)$$
The hyperbolic Cayley plane is similarly realized and the isometry group is the
real form of $F_4$ of real rank one.The maximal compact subgroup is $Spin(9)$.
We can compactify the hyperbolic space by adding $D_0$.If $dim_RF=k$ then $D_0$
is a $(kn-1)$ dimensional sphere denoted by $\partial{H^n_F}$.
Take the Iwasawa decomposition $G=KAN$.Then $A$ and $N$ fix $\xi$ in the sphere
at infinity and $N$ is a nilpotent goup obtained as an extion
$$ 0 \rightarrow ImF \rightarrow N \rightarrow F^{n-1} \rightarrow 0.$$
Horospheres based at $\xi$ are the orbits of $N$.
The stabilizer in $O_F(n,1)$ of $\xi$ is a parabolic subgroup with Langlands
decomposition $MAN$.For $F=R,C,H$ $M$ is a subgroup of $K$ of the form
$O_F(1)\times O_F(n-1)$.For Cayley plane $M$ is $Spin(7)$.
$K$ actually acts transitively on $\partial{H^n_F}$ giving
$K/M=\partial{H^n_F}$
.The centralizer of $M$ in $K$ acts freely on $K/M$ with quotient $P^{n-1}_F$.
This gives Hopf fibration $\partial{H^n_F}\rightarrow P^{n-1}_F$.
Since $K$ acts transitively on $\partial{H^n_F}$ ,killing form on its Lie
algebra determines a Riemannian metric on $\partial{H^n_F}$.Define the
orthogonal complements
to the vertical tangent spaces of the Hopf fibration $\partial{H^n_F}
\rightarrow P^{n-1}_F$.Then there is $K$ invariant Riemannian metric on the
horizontal bundle and if we define the distance of two points
as the minimum of the lengthes of the pathes joining these two points and
staying horizontal bundle it is Carnot-Caratheodary metric.Mitchell[Mi]
calculated
that the housdorff dimension of $\partial{H^n_F}$ in this metric is $k(n+1)-2$.
\section{Boundary of Rank one symmetric Spaces}
The boundary of rank one symmetric space $H^m_F$ is a one point
compactification of nilpotent group $N$ in Iwasawa decomposition,denoted by $N
\cup \infty$.We will introduce nice coordinates and left invariant distance to
introduce cross ratio of four points.We will define everything in Cayley
numbers because that is the most
general case among four.
Cayley number is a pair of quaternions $(q_1,q_2)$.The multiplication is
defined
as:
$$(q_1,q_2)(p_1,p_2)=(q_1p_1-\bar{p_2}q_2,q_2p_1+p_2\bar{q_1})$$
Also we define $\overline{(q_1,q_2)}=(\bar{q_1},-q_2)$.
Then it has the following properties.
$1)q\bar{q}=|q|^2$
$2)|qp|=|q||p|$
$3)q^{-1}=\bar{q}/|q|^2$
$4)\overline{pq}=\bar{q}\bar{p}$
Even though Cayley numbers are not commutative nor associative,using Artin's
theorem saying that the subalgebra generated by two elements
is associative,we have following extra properties.
1)$(xy)y^{-1}=x$
2)$(xy)^{-1}=y^{-1}x^{-1}$
Here is some usuful fact. $Aut(O)=G_2$ and it is compact group.It fixes real
number and acts transitively on unit pure imaginaries.The stabilizer of $i$ is
a
copy of $SU(3)$ and it acts transitively on the unit pure imaginaries
orthogona
l to $i$.The stabilizer of $i,j$ which fixes $k$ since $k=ij$,acts transitively
on the unit pure imaginaries orthogonal to $i,j,k$ and it is a copy of $SU(2)$.
For more extensive informations,see [Fr].
Now an element of $N$ is denoted as $[(t,q),k]$ where
1)$F=R(Real),(t,q)=0,k \in R^{m-1}$.
2)$F=C(Complex),t$ is real,$q=0,k \in C^{m-1}$.
3)$F=H(Quaternion),t \in R^3,q=0,k \in H^{m-1}$.
4)$F=O(Cayley),t \in R^3,q$ is a quaternion,k is a Cayley number.
The multiplication is defined as;
$$[(t,q),k][(t',q'),k']=[(t+t',q+q')+2Im<k,k'>,k+k']$$
We define the gauge of $[(t,q),k]$ as;
$$A([(t,q),k])=(|k|^2+t,q)$$
Then define $|[(t,q),k]|=(|k|^4+|t|^2+|q|^2)^{1/4}$ and $d(g,g')=|g'^{-1}g|$.
This is a left invariant distance and the housdorff dimesion of the boundary
with respect to this metric is $dim_R(F) \times (m+1)-2$ wich agrees with
Mitchell's calculation.Even though this metric is not riemannian if we take the
inner
metric of this it becomes Carnot-Caratheodory metric.For references
see[Gr][Pa][Mi].
The action of $Iso(H^m_F)$ extends continously to $\partial(H^m_F)$.Let
$Sim(N)$ denote the subgroup of $Iso(H^m_F)$ wich fixes $\infty$.Then it is
isomorphic to $ N \rtimes (O_F(m-1)\cdot O_F(1) \times R)$ and $N \rtimes
O_F(m-1)\cdot O_F(1)$ acts as isometries with respect to the metric given
above.Note that $N$ part comes from left action of the group itself and
$O_F(m-1)\cdot O_F(1) \times R$ part comes from
hyperbolic isometries of the form in $O_F(m,1)$
\[
\left[
\begin{array}{ccc}
M & 0 & 0 \\
0 & \nu\cosh(s) & \nu\sinh(s) \\
0 & \nu\sinh(s) & \nu\cosh(s)
\end{array}
\right]
\]
where $O_F(m-1)\cdot O_F(1)=O(m-1),U(m-1),Sp(m-1)\cdot Sp(1),O_O(1)\cdot
O_O(1)$ depending on $F=R,C,H,O$ respectively.These actions will be described
in next
section.
We define the cross ratio of four points as;
$$[g_1,g_2,g_3,g_4]=\frac{|A(g_3^{-1}g_1)||A(g_4^{-1}g_2)|}
{|A(g_4^{-1}g_1)||A(g_3^{-1}g_2)|}$$
Let's diverge to the unit ball model for a while so that we can get some
connetion between the metric of hyperbolic space and the metric defined above.
In Real,complex,Quaternion case for $x,y \in H^m_F$
$$cosh(d(x,y))=\frac{|1-<x,y>|}{(1-<x,x>)^{1/2}(1-<y,y>)^{1/2}}$$
For Cayley hyperbolic case if we define
$$R<v,w>=Re(v_1\bar{v_2})(w_2\bar{w_1})-Re(\bar{v_2}w_2)(\bar{w_1}v_1)$$
Then for $x,y \in H^2_O$ the distance is
$$cosh(d(x,y))=\frac{(|1-<x,y>|^2+2R<x,y>)^{1/2}}
{(1-<x,x>)^{1/2}(1-<y,y>)^{1/2}}$$
For details see[Mo].
Let $<<x,y>>=1-<x,y>$ for $F=R,C,H$ and $<<x,y>>={(|1-<x,y>|^2+2R<x,y>)^{1/2}}$
for Cayley case.Then the cross ratio of the four points in the boudary of
hyperbolic space is defined as
$$[x,y,z,w]=\frac{<<z,x>><<w,y>>}{<<w,x>><<z,y>>}$$
Note that this is a limit of
$$\frac{\cosh(d(z_i,x_i))\cosh(d(w_i,y_i))}
{\cosh(d(w_i,x_i))\cosh(d(z_i,y_i))}$$
where $x_i,y_i,z_i,w_i$ tend to $x,y,z,w$.So it is invariant under isometries
of hyperbolic space.
To make connection between these two definitions of cross ratio we introduce
generalize projection from $N \cup \infty$ to the boundary of unit ball.
$$w_1=2(1+|k|^2-t,-q)^{-1}k$$
$$w_2=(1+|k|^2-t,-q)^{-1}(1-|k|^2+t,q)$$
Note that $(0,0)$ corresponds to $(0,1)$,$\infty$ to $(0,-1)$.
Let $g_i=[(t_i,q_i),(c_i,d_i)]$ for $i=1,2$.
By the general projection these two points correspond to
$$x_i=1/((1+|k_i|^2)^2+|t_i|^2+|q_i|^2)
[2(c_i+|k_i|^2c_i+t_ic_i-\bar{d_i}q_i,q_ic_i+d_i+$$
$$d_i|k_i|^2-d_it_i),(1-|k_i|^4+2t_i-|t_i|^2-|q_i|^2,2q_i)]$$
Since (0,0) and $\infty$ correspond to $(0,1)$ and $(0,-1)$
$$<<x_i,(0,-1)>>=2/((1+|k_i|^2)^2+|t_i|^2+|q_i|^2)$$
$$<<x_i,(0,1)>>=2\frac{\sqrt{(|k_i|^4+|k_i|^2+|t_i|^2+|q_i|^2)^2
+|t_i|^2+|q_i|^2}}{(1+|k_i|^2)^2+|t_i|^2+|q_i|^2}$$
Since
$((|k_i|^4+|k_i|^2+|t_i|^2+|q_i|^2)^2+|t_i|^2+|q_i|^2=(|k_i|^4+|t_i|^2+|q_
i|^2)((1+|k_i|^2)^2+|t_i|^2+|q_i|^2)$ we get
$$[\infty,0,x_1,x_2]=\frac{(k_2|^4+|t_2|^2+|q_2|^2)^{1/2}}
{(k_1|^4+|t_1|^2+|q_1|^2)^{1/2}}$$
$$=|A(g_2^{-1})|/|A(g_1^{-1})|$$
so we are done.
For the case $[0,g_1,\infty,g_2]$,just note that
$[0,g_1,\infty,g_2]=[g_2^{-1},g_2^{-1}g_1,\infty,0]$.By using above fact we are
done again.
In fact for $F=R,C,H$ it is easy to show that two definitions of cross ratio
agree by brutal calculation but in Cayley Hyperbolic case calculation is quite
hard.But to prove our main theorem we will just need what we proved.
\section{Action of Isometries on $\partial{H^m_F}$}
In this section we want to incode the action of hyperbolic isometry on the
boundary.A hyperbolic isometry in $O_F(m,1)$ has the form of ;
\[
\left[
\begin{array}{ccc}
M & 0 & 0 \\
0 & \nu\cosh(s) & \nu\sinh(s) \\
0 & \nu\sinh(s) & \nu\cosh(s)
\end{array}
\right]
\]
where $\nu=1,M \in O_F(m-1)$ for $F=R,C$ and $|\nu|=1,M \in Sp(m-1)$ for $F=H$
and $ |M|=|\nu|=1$ for Cayley hyperbolic case.
Then this element send $[(t,q),k]$ to;
$$(t',q')=e^{-2s}{\nu}^{-1}(t,q)\nu$$
\begin{eqnarray*}
k'=e^{-s}\{[{\nu}^{-1}(e^{2s}+|k|^2-t,-q)^{-1}{\nu}]^{-1}\}& & \\
\{[{\nu}^{-1}((e^{2s}+|k|^2-t,-q)^{-1}(1+|k|^2-t,-q))][((1+|k|^2-t,-q)^{-1}k)A]
\}& &
\end{eqnarray*}
First note that the matrix acts on the unit ball as;
$$w_1'=(w_2{\nu}\sinh(s)+{\nu}\cosh(s))^{-1}(w_1A)$$
$$w_2'=(w_2{\nu}\sinh(s)+{\nu}\cosh(s))^{-1}(w_2\nu\cosh(s)+\nu\sinh(s))$$
using coordinate chage between unit ball model and $F^{m+1}$ sending $w$ to
$(w,1)$.
To check above claim we use Artin's theorem crucially.
By using generalized projection and above equations we can show
\begin{eqnarray*}
w_2'=\{((1+|k|^2-t,-q)^{-1}(1-|k|^2+t,q))\nu\sinh(s)+\nu\cosh(s)\}^{-1}& & \\
\{[(1+|k|^2-t,-q)^{-1}(1-|k|^2+t,q)]\nu\cosh(s)+\nu\sinh(s)\}& &
\end{eqnarray*}
$$=[{\nu}^{-1}((e^{2s}+|k|^2-t,-q)^{-1}(1+|k|^2-t,-q))]
[((1+|k|^2-t,-q)^{-1}(e^{2s}-|k|^2+t,q))\nu]$$ equal to
$$[({\nu}^{-1}(e^{2s}+|k|^2-(t,q))^{-1})\nu]
[{\nu}^{-1}(e^{2s}-|k|^2+(t,q)\nu]$$
To prove this ,set $r_1=e^{2s}+|k|^2,r_2=e^{2s}-|k|^2,r_3=1+|k|^2,(t,q)=Q$ and
using identity $-QQ=|Q|^2$
we can show
$$[{\nu}^{-1}((r_1+Q)^{-1}(r_3-Q))][((r_3+Q)(r_2+Q))\nu]=
[({\nu}^{-1}(r_1-Q)^{-1})\nu][{\nu}^{-1}(r_2+Q)\nu].$$
For $w'_1$,it is easy to check.
\section{Marked length spectrum determines representation}
First we will make very simple but important observation to prove the theorem.
\begin{lemma}Let $a,b$ be two hyperbolic isometries such that $x_1$ is a
repelling
fixed point of $a$ ,$x_3$ is a attracting fixed point of $a$ and $x_4$ is a
attracting fixed point of $b$,$x_2$ is a repelling fixed point of $b$.Then
$$\lim_{n\rightarrow\infty}e^{l(a^n)+l(b^n)-l(a^nb^n)}=|[x_1,x_2,x_3,x_4]|.$$
\end{lemma}
\begin{pf}Choose $x_1^n$ on the axis of $a$,$z_1^n$ on the axis of
$b^na^n$,$x_2
^n$ on the axis of $b$,$w_2^n$ on the axis of $a^nb^n$ so that $d(x_1^n,z_1^n)$
and $d(x_2^n,w_2^n)$ go to zero.If we put
$x_3^n=a^n(x_1^n),z_3^n=a^n(z_1^n),x_4
^n=b^n(x_2^n),w_4^n=b^n(w_2^n)$ and
$d_{13}^n=d(x_1^n,x_3^n),d_{24}^n=d(x_2^n,x_
4^n),d_{23}^n=d(w_2^n,z_3^n),
d_{14}^n=d(z_1^n,w_4^n)$ then
$$\lim \sqrt{\frac{<<x_1^n,x_3^n>><<x_3^n,x_1^n>>
<<x_2^n,x_4^n>><<x_4^n,x_2^n>>}
{<<x_1^n,x_4^n>><<x_4^n,x_1^n>><<x_2^n,x_3^n>><<x_3^n,x_2^n>>}}$$
$$=\lim
\frac{(e^{d^n_{13}}+e^{-d^n_{13}})(e^{d^n_{24}}+e^{-d^n_{24}})}{(e^{d^n_
{14}}+e^{-d^n_{14}})(e^{d^n_{23}}+e^{-d^n_{23}})}$$
$$=\lim \frac{e^{d^n_{13}}e^{d^n_{24}}}{e^{d^n_{14}}e^{d^n_{23}}}+\lim
\frac{e^{
d^n_{13}}}{e^{d^n_{14}}e^{d^n_{23}}}+\lim
\frac{e^{d^n_{24}}}{e^{d^n_{14}}e^{d^n
_{23}}}$$
$$=\lim e^{d^n_{13}+d^n_{24}-d^n_{14}-d^n_{23}}+\lim
e^{d^n_{13}-d^n_{14}-d^n_{2
3}}+\lim e^{d^n_{24}-d^n_{14}-d^n_{23}}$$
$$=\lim e^{l(a^n)+l(b^n)-l(a^nb^n)}+\lim e^{l(a^n)-l(a^nb^n)}+\lim
e^{l(b^n)-l(a
^nb^n)}$$
Since $l(a^n)+l(b^n)-l(a^nb^n)$ always
exists,$l(a^n)-l(a^nb^n),l(b^n)-l(a^nb^n)
$ both go to $-\infty.$ For more general argument see [Kim].
So we finally get
$$\lim e^{l(a^n)+l(b^n)-l(a^nb^n)}=[x_1,x_2,x_3,x_4]$$
\end{pf}
This lemma shows that marked length spectrum determines cross ratio on the
limit set because every two points in the limit set can be approximated by end
points of some hyperbolic isometry.By using limit argument it is clear that
cross ratio of every four points in the limit set is determined by its marked
length spectrum.
Now we are ready to prove the theorem.
\begin{th}Let $\rho:G\rightarrow Iso^{\pm}(H^n_K),\phi:G\rightarrow
Iso^{\pm}(H^
m_F)$ be two nonelementary,irreducible representations having the same marked
le
ngth spectrum such that $n\times dim_R(K)=m \times dim_R(F)$ where $K,F$ is
real
,complex,quaternion or cayley fields.Then $n=m,K=F$ and they are conjugate.
\end{th}
\begin{pf}
After we conjugate representations we may assume that 0,and $\infty$ are in the
limit sets and they are the two fixed points of hyperbolic
isometries $a$ and $a'$ in $\rho$ and $\phi$ respectively.Furthermore we have
$$\frac{|x_i|}{|x_j|}=|[x_i,x_j,[0,0],\infty]|^{1/2}=|[y_i,y_j,[0,0],\infty]|^{
1/2}=\frac{|y_i|}{|y_j|}$$
So after scaling we can assume that $|x_i|=|y_i|$.
Similarly
$$\frac{d(x_i,x_j)}{|x_j|}=|[[0,0],x_i,\infty,x_j]|^{1/2}=|[[0,0],y_i,
\infty,y_j
]|^{1/2}=\frac{d(y_i,y_j)}{|y_j|}$$
and hence $d(x_i,x_j)=d(y_i,y_j).$
Let [0,1],$[0,z]$ are the two points in the limit set of $\rho$ and
$[c,d],[(t,q
),k]$ be the corresponding points in the limit set of $\phi$.
Then $a$ sends $[(t,q),k]$ to
$$(t',q')=e^{-2s}{\nu}^{-1}(t,q)\nu$$
\begin{eqnarray*}
k'=e^{-s}\{[{\nu}^{-1}(e^{2s}+|k|^2-t,-q)^{-1}{\nu}]^{-1}\}& & \\
\{[{\nu}^{-1}((e^{2s}+|k|^2-t,-q)^{-1}(1+|k|^2-t,-q))][((1+|k|^2-t,-q)^{-1}k)M]
\}& &
\end{eqnarray*}
where $\nu=1,M \in O(m-1),U(m-1)$ in real and complex case,$|\nu|=1,M \in
Sp(m-1
),|M|=1$ in quaternion and cayley cases.
Similarly $a'$ sends $[(t,q),k]$ to
$$(t',q')=e^{-2s}{\nu}^{-1}(t,q)\nu$$
\begin{eqnarray*}
k'=e^{-s}\{[{\mu}^{-1}(e^{2s}+|k|^2-t,-q)^{-1}{\mu}]^{-1}\}& & \\
\{[{\mu}^{-1}((e^{2s}+|k|^2-t,-q)^{-1}(1+|k|^2-t,-q))][((1+|k|^2-t,-q)^{-1}k)N]
\}& &
\end{eqnarray*}
Now
$$d([0,z],[0,1])^4=d([(t,q),k],[c,d])^4$$ gives us that $|(t,q)|$ is a
function
of $z,k,c,d$.But
$$d([0,z],a^n([0,1]))^4=d([(t,q),k],a'^n([c,d]))^4$$ gives that $|(t,q)|$ is a
function of $e^{-ns},z,k,c,d$ for all n.This is because right side of equation
has the higher power of $e^{-s}$.This is not possible unless $|(t,q)|=0$.
By this way we can conclude that all the limit sets of $\phi$ corresponding to
$[0,z]$ of $\rho$ are the form of $[0,k]$.Then by further conjugation we can
assume that [0,1] is also the corresponding point of $\phi$ to [0,1] of $\rho$.
Next we want to show that either $M=N,\nu=\mu$ or $M=\bar{N},\nu=\bar{\mu}$.
This can be seen from the equation
$$d(a^n([0,1]),[0,1])^4=d(a'^n([0,1]),[0,1])^4$$ for all n.
Now by conjugating the representation by the map $(k \rightarrow \bar{k})$ we
can assume that $M=N,\nu=\mu$.
If $[0,w]$ is a point in the limit set of $\rho$ and $[0,w']$ is the
corresponding point of $\phi$ then
$$d(a^n([0,w]),[0,1])^4=d(a'^n([0,w']),[0,1])^4$$
gives that $w=w'$.
These all show that $K=F,m=n$.So far we conjugated two representations to get
that two limit sets are equal on the set $K^{m-1}=\{[0,k]:k \in
R^{m-1},C^{m-1},Q^{m-1},
O\}$ and two spaces are actually the same one of the four hyperbolic spaces.
For any point $x$ of the limit set of $\rho$ not in the set $K^{m-1}$
there are only two possible corresponding points $y$(either $x$ itself or
reflection of $x$ along $K^{m-1}$)
such that the distance between $x$ and every point in the limit set of $\rho$
lying on the set $K^{m-1}$ is equal to the
distance between $y$ and
corresponding point of $\phi$.
By taking reflection along this set if necessary,we can assume that $x=y$.
Now for any point in the limit set of $\rho$ off the set $K^{m-1} \cup x$ there
is a unique corresponding point as above.So two limit sets are actually equal.
This shows that we found an isometry such that after we conjugate one
representation by this isometry the limit sets are equal.But this actually
shows that two representations are equal.The reason is the following.
Note that the end points of hyperbolic isometry $aba^{-1}$ are the images of
end
points of $b$ under $a$.But since every isometry is determined by images of
finitely many points on $\partial{H^m_F}$ and since two representations have
the same limit sets after conjugation each element should be the same.
\end{pf}
\begin{co}Let M,N be rank one locally symmetric manifolds of the same dimension
and homotopically equivalent.If none of them have totally geodesically embedded
submanifold of dimension greater than one and if they have the same marked
length spectrum then they are isometric.
\end{co}
Since three manifolds get special attention we will write down the corollary
for hyperbolic three manifolds.Here we do not need the assumption about
irreducibility.
\begin{co}If we have two nonelementary orientable hyperbolic three manifolds
having the same marked length spectrum then they are isometric.
\end{co}
\begin{pf}Since every orientation preserving Mobius map is determined by the
images of three distinct points whether they are on the line or not we don't
have to worry about irreducibility.
\end{pf}
\section{Space of Representations into $SL_2(C)$}
\subsection{Backgroud}
Let G be a finitely generated group.Space of representations from G to
$SL_2(C)$
is a set of homomorphisms from G into $SL_2(C)$.We say that two
representations
are equivalent if they are conjugated by some element of $SL_2(C)$.The
character of a representation $\rho$ is the function $\chi_{\rho}:G
\rightarrow C$ such
that $\chi_{\rho}(g)=tr(\rho(g))$.
If $G=\{g_i,i=1,\cdots,n;r_1=\cdots=r_k=1\}$ then space of representations
$R(G)$ is
a subset of $SL_2(C) \subset C^{4n}$,wich are set of all points $(\rho(g_1),
\cdots,\rho(g_n))$ satisfying the relator relations.It is easy to see that
$R(G)$
is an affine algebraic set in $C^{4n}$.See [CS] for details.
If we have a finite volume complete hyperbolic 3-manifold then its holonomy
representation is discrete ,faithful and irreducible.
For each $g \in G$ we define a regular function $\tau_g$ on $R(G)$ by
$\tau_g(\rho)=tr\rho(g)$.
If T is a ring generated by all functions $\tau_g,g\in G$ then
it is finitely generated.See [CS] for a proof.
Also it is shown in [CS] that character space of an irreducible component of
$R(
G)$ is an affine variety.
Thurston([Th] chapter 5) showed that geometric structure of compact manifold M
are determined by holonomy representations of $\pi_1(M)$ near some geometric
structure
and its complex dimension is at least $3
\times(\#(generators)-\#(relators)-1)$.
\begin{th}([Th],[CS])Let N be a compact orientable 3-manifold.Let
$\rho:\pi_1(N)
\rightarrow SL_2(C)$ be an irreducible representation such that for each torus
component T of $\partial{N}$,$\rho(im(\pi_1(T) \rightarrow \pi_1(N)))$ not
trivial.
Let R be an irreducible component of $R(\pi_1(N))$ containing $\rho$.Then
$X_0=character(R)$ has dimension $\geq s-3\chi(N)$,where s is the number of
tori component of $\partial{N}$.
\end{th}
\subsection{Local smooth coordinate chart of Representation space by lengthes
of
finite number of loops}
Let $G$ be a group and $\Re$ be the representation space of $G$ into $SL_2(C)$
,where $\Re$ is the set of equivalence classes of homomorphisms from $G$ into
$SL_2(C)$ and two representations are equivalent if they are conjugate .
Suppose $\rho$ is a representation sitting smoothly in
$\Re$.We want to find nice coordinate chart around $\rho$ such that each
coordinate is a translation length of some element in $G$.
\begin{lemma}
Let $\alpha$ be an isometry in $SL_2(C)$. Then
$$
|tr(\alpha)-2|+|tr(\alpha)+2|=2(e^{\frac{l(\alpha)}{2}}+e^{-\frac{l(\alpha)}{
2}}) $$
\end{lemma}
\begin{pf}
If $\alpha$ is
\[
\alpha=\left[
\begin{array}{cc}
\lambda & 0 \\
0 & \lambda^{-1}
\end{array} \right]
\]
Put $x=|\lambda+\lambda^{-1}-2|,y=|\lambda+\lambda^{-1}+2|$.
Then $x^2+y^2=2|\lambda|^2+2\lambda\overline
{\lambda^{-1}}+2\lambda^{-1}\bar{\lambda}
+2|\lambda^{-1}|^2+8$ and
$$x^2+y^2+xy=4(|\lambda|^2+2+|\lambda^{-1}|^2).$$
By using $|\lambda|=e^{\frac{l(\alpha)}{2}}$ claim follows.
\end{pf}
Now let $\acute{G}$ be a subset of $G$ consisting of all elements whose image
under $\rho$ is either hyperbolic or elliptic.
Notice that if $a,b \in{\acute{G}}$ then $a^n,b^n,a^nb^n$ are all in
$\acute{G}$
for large n. Also if $a\in G$ and $b\in {\acute{G}}$ then $aba^{-1}$ is in
$\acute{G}$.This implies that marked length spectrum on $G'$ determines
representation up to conjugacy.Choose finite set S of $G$ so that traces of
those
elements determines representation up to conjugation and trace map from the
neighborhood to ${C^*}^S$ where $C^*=C-\{2,-2\}$
is an immersion.This is possible if the representation is not
elementary.See
the arguement after the example below for justification.
\begin{lemma}(Vogt)
Let G be a free group with three generators.Let $trX_i=x_i,trX_iX_j=y_{ij
},trX_iX_jX_k=z_{ijk}$.Define
$$ P=x_1y_{23}+x_2y_{13}+x_3y_{12}-x_1x_2x_3$$
$$
Q=x_1^2+x_2^2+x_3^2+y_{12}^2+y_{13}^2+y_{23}^2+y_{12}y_{13}y_{23}-x_1x_2y_{12
}\\-x_1x_3y_{13}-x_2x_3y_{23}-4$$
$$\Delta(X_1,X_2,X_3)=P^2-4Q$$
Then $z_{123}$ and $z_{213}$ are roots of the quadratic equation for $z$:
$$z^2-Pz+Q=0$$Given prescribed values $x_i,y_{ij}$ there exists only one
conjugacy class if and only if $\Delta(X_1,X_2,X_3)=0.$
\end{lemma}
See [Ma 1] for the proof.
{\bf{Example}}.Let $G$ be a free group with three generators.I want to show
that
trace map is not an immersion sometimes.Let \[ X=\left(
\begin{array}{cc}
\alpha & 0 \\
0 & {\alpha}^{-1}
\end{array} \right) \] and \[ Y=\left(
\begin{array}{cc}
\beta & 0 \\
0 & {\beta}^{-1}
\end{array} \right) \]
Let $Z$ be a third element which does not commute with any of two elements and
$\Delta(X,Y,Z)\neq 0$.
Since $G$ is free neighborhood of this representation is a regular neighborhood
of $SL_2(C) \times SL_2(C) \times SL_2(C) $ and obviously sitting smoothly in
$\Re$.Let tr be a map from this neighborhood to $C^7$ such that
$$tr(\rho)=(trX,trY
,trZ,trXY,trXZ,trYZ,trXYZ)$$Then elementary calculation shows that
\begin{eqnarray*}
dtr_{(X,Y,Z)}(\xi_1,\xi_2,\xi_3)&=&(tr(X\xi_1),tr(Y\xi_2),tr(Z\xi_3),tr(X\xi_2)
\\
& & \mbox{} +tr(Y\xi_1),tr(X\xi_3)+tr(Z\xi_1),tr(Y\xi_3)+tr(Z\xi_2),...)
\end{eqnarray*}
where $\xi_i \in sl_2(C).$
Then dimension of kernel at $(X,Y,Z)$ is 4.To see this note that there are
two degrees of freedom from $tr(X\xi_1)$ and so on ,so there are six degrees of
freedom from $\xi_1,\xi_2,\xi_3$.
$tr(X\xi_1),
tr(Y\xi_2)$ equal to zero implies $tr(X\xi_2)+tr(Y\xi_1)$ is zero since $X$
and $Y$ commute.But it should satisfy $tr(X\xi_3)+tr(Z\xi_1)=0$ and
$tr(Y\xi_3)+tr(Z\xi_2)=0$ so it takes out two degrees of freedom.
If $\xi_1,\xi_2,\xi_3$ satisfy
$tr(X\xi_1)=tr(Y\xi_2)=tr(Z\xi_3)=tr(X\xi_2)+tr(Y\xi_1)
=tr(X\xi_3)+tr(Z\xi_1)=tr(Y\xi_3)+tr(Z\xi_2)=0$ then the last term vanishes
automatically.Here is the reseason.
If $X_i(t)$ are curves passing throug $X,Y,Z$
respectively and with tangent vectors $\xi_1,\xi_2,\xi_3$ set
$$tr(X_i(t))=x_i(t)
,trX_i(t)X
_j(t)=y_{ij}(t),trX_1(t)X_2(t)X_3(t)=z(t).$$
By lemma 3 $z(t)^2 -P(t)z(t)+Q(t)=0.$If we differentiate this equation at 0 we
get $z'(0)(2z(0)-P(0)=0$ by using $P'(0)=Q'(0)=0$.But $2z(0)-P(0)\neq 0$ since
$\Delta$ is not equal to zero.So tr is not an embedding at this point.
This example has an obvious geometric meaning.For $tr(X\xi_1)$ to be zero,
diagonal elements of $\xi_1$ should be equal to zero.This implies that we
deform $X$ by
conjugation.
The same is true for $Y$.Since $X,Y$ have the same axis, conjugating $X$ by
those elements conjugating $Y$ does not change trace of $X$ so infinitesimal
change of $tr(X)$ is equal to zero. Same argument holds for $Y$.
More explicitly let \[ g_t=\left(
\begin{array}{cc}
x_1(t) & x_2(t) \\
x_3(t) & x_4(t)
\end{array}
\right) \] and \[ h_t=\left(
\begin{array}{cc}
y_1(t) & y_2(t) \\
y_3(t) & y_4(t)
\end{array} \right) \]
such that $g_0=h_0=I$ and $x_2'(0)=-y_2'(0)=x_3'(0)=-y_3'(0)$.
Let \[ Z=\left(
\begin{array}{cc}
a & b \\
b & d \end{array} \right) \] for simplicity.
Set $X_t=g_tX{g_t}^{-1},Y_t=h_tY{h_t}^{-1},Z_t=g_tZ{g_t}^{-1}$.
Then ${\xi(t)}=(X_t,Y_t,Z_t)$ is a 1-parameter family of defermation which is
not a conjugation defermation.For \[ X'(0)= \left(
\begin{array}{cc}
0 & -x_2'\alpha+x_2'{\alpha}^{-1}\\
x_3'\alpha-x_3'{\alpha}^{-1} & 0
\end{array} \right) \]
and similarly for $Y'(0)$ there is no conjugation deformation giving tangent
vectors $X'(0)$ and $Y'(0)$.It is easy to see that $dtr_{(X,Y,Z)}(\xi'(0))=0$
.For if we calculate $\frac{dtr(Y_tZ_t)}{dt}$ at $t=0$,it is equal to
$\beta(x_2'(0)b-x_3'(0)b)+
b(-y_2'(0)\beta+x_2'(0){\beta}^{-1})+b(y_3'(0)\beta-y
_3'(0){\beta}^{-1})+{\beta}^{-1}(-x_2'(0)b+x_3'(0)b)$.
If we look at above example carefully,it shows if no two elements commute then
trace map
is an immersion,which implies that every tangent vector in the kernel of $dtr$
comes from some conjugation deformation.
Observe the following calculation.
\[ g_t=\left( \begin{array}{cc}
x_1(t) & x_2(t) \\
x_3(t) & x_4(t)
\end{array} \right) \]
\[X= \left(
\begin{array}{cc}
a & b \\
c & d \end{array} \right) \]
Then conjugation of $X$ by $g_t$ give tangent vector at $X$
\[ \left( \begin{array}{cc}
x_2'c-x_3'b & -x_2'a+2x_1'b+x_2'd \\
x_3'a+2x_4'c-x_3'd & -x_2'c+x_3'b
\end{array} \right) \]
Notice that two entries are linearly independent.So if $g_t,h_t$ give the same
infinitesimal deformation on two different element which do not commute then
they give the same infinitesimal deformation on every element.
By using this fact we can show that if the representation is not elementary
then
we can show that trace map from a neighborhood of a representation $\rho$
sitting
smoothly in representation space to $C^N$ for large enough N is a
smooth
immersion.
To see this if $S=\{X_1,\cdots,X_n\}$ is the finite set containing all
generators of the group such that its traces determine representation uniquely
up to conjugacy(which is possible by [CS]) and their images under $\rho$ are
all hyperbolic isometries.By abusing the notation we will identify $X_i$ with
$\rho(X_i)$.
This can be seen that if $S$ is such a set
possibly
contaning parabolic or elliptic elements then choose a hyperbolic element
and multiply those elements which are not hyperbolic in $S$ by this element and
its inverse to make them hyperbolic and then throw in this element.Then
using the formula $tr(XY)+tr(XY^{-1})=tr(X)tr(Y)$ we see that all the traces of
old elements can be obtained by new elements.This is a new set consisting
entirely of hyperbolic elements.
Since representation is not elementary there is $X_k$ which does not commute
with $X_1$.Choose any $X_i,i\neq1,k$.If it commutes with $X_1$ replace
$X_1,X_i,
X_k
$ by $X_1X_i,X_1X_k,X_k$.It is easy to check that no two of three does commute.
By renaming them we get $(X_1,\cdots,X_n)$ such that no two of $X_1,X_2,X_3$
commute
and $X_i,i\neq 1,2,3$ commutes with at most one of three $X_1,X_2,X_3$.
If $(\xi_1,\cdots,\xi_n)$ is in the kernel of the differential of the trace
map,
there is $g_t$ conjugating $X_1,X_2,X_3$, giving tagent vectors $\xi_1,\xi_
2,\xi_3$ because no two of three commute.Similary $h_t$ conjugating
$X_1,X_2,X_4
$ giving $\xi_1,\xi_2,\xi_4$ since no two of three commute.But $h_t$ can be
replaced by $g_t$ since they give the same infinitesimal deformation on
$X_1,X_2$ which do not commute.In this way we can see that this tangent
vector
comes from conjugation deformation.
By lemma 2 the map $f$ from ${C^*}^S$ to $R^{\acute{G}}$ sending each
component $\alpha$ of an element in ${C^*}^
S$ to $2(e^{\frac{l(\alpha)}{2}}+e^{-\frac{l(\alpha)}{
2}})$ is $C^{\infty}$ since there is no parabolic element in $\acute G$ and
norm
function is smooth at $C^*$.Furthermore by adding more elements in $S$ we can
make this function an immersion.To see this consider
the map $g$ from $R^2$ to $R$ such that $g(x,y)=|(x,y)-2|+|(x,y)+2|$ as in the
above lemma.It is easy to see that if $(x,y)\neq 2,-2$ then dimension of the
kernel of $dg$ is one.Suppose $dg(\xi)=0$ at $\alpha$. Then find an element
$\beta$ so that trace($\beta$) is a
polynomial in $\alpha$ and other elements in $S$.Choosing $\beta$ carefully
it is possible to ensure that $dg(\xi) \neq 0$ at $\beta$.So adding $\beta$ to
$S$ for each such an $\alpha$ we can show the map is an immersion.
\begin{Prop}
Let $\rho$ be an nonelementary representation sitting smoothly inside the
representation space.Then neighborhood of $\rho$,denoted by $N(\rho)$, is
parametrized
by translation lengthes of finitely many hyperbolic elements in $G$.
\end{Prop}
\begin{pf}
So far we got
$$f\circ tr:N(\rho)\rightarrow {C^*}^S \rightarrow R^{\acute{G}}$$ is an
immersion.But since marked length spectrum on $R^{\acute{G}}$ determines
representation it is one to one also.
Since dimension of
representation space is finite around $\rho$ we can find finite subspace $R^N$
such that
projection of the image of the neighborhood to this subspace is injective.
\end{pf}
Specially if the representation is an holonomy of geometrically finite
3-manifold then it is automatically sitting smoothly in representation space
because Teichmuller
space is open in representation space and it satisfies all the necessary
condition for above considerations,so small neighborhood of this
representation
is parametrized smoothly by lengthes of finitely many geodesics.
$$ ACKNOWLEDGEMENT$$
I wish to thank all whom I talked with,among them I want to specially thank
A.Casson as my adviser,C.McMullen,J.P.Otal,F.Bonahon,M.Bourdon,\\
G.Besson for their useful comments.
|
\section{Introduction}
\label{intro}
Since the discovery of the dusty disk around $\beta\:$Pictoris\ (Aumann 1984, Smith \&
Terrile 1984), it has been the subject of extensive investigations with all
possible observational techniques: visible and IR imagery (Smith \& Terrile
1987, Paresce \& Burrows 1987, Lecavelier des Etangs et al. 1993, Golimowski et
al. 1993, Lagage \& Pantin 1994), spectroscopy (Vidal-Madjar et al. 1994,
Vidal-Madjar \& Ferlet 1995 and references therein) and recently photometric
variations (Lecavelier des Etangs et al. 1995). In particular, the stellar
light scattered by dust particles has been well observed by several authors for
many years to follow a power law distribution with a slope in the range -3.6
(Artymowicz et al. 1989, Lecavelier des Etangs et al. 1993) -4.3 (Smith \&
Terrile 1984). However, a convincing explanation for this scattered light
distribution is still needed.
The dust distribution in the $\beta\:$Pictoris\ disk may be a crucial issue for understanding
its origin. We present here a precise dynamical model which will be compared
with the observations for the first time in a quantitative manner.
Another key point revealed by the spectroscopic studies of the gaseous
counterpart of the $\beta\:$Pictoris\ disk is the existence of orbiting kilometer size bodies
which sometimes fall onto the star, giving rise to the observed
spectroscopic redshifted signatures due to ejected gas,
and undoubtedly to dust particles. Already
in 1984, Weissman (1984) questioned whether the material around Vega-like stars
is of cometary or asteroidal origin and suggested that particles could be
continually supplied by sublimation or from collisions between larger bodies.
The time scale evaluation over which dust particles are eliminated by the
Poynting-Robertson effect or collisions (Backman \& Paresce 1993) leads to
durations much shorter than the estimated age of $\beta\:$Pictoris\ ($\sim 2\cdot 10^{8}$
years according to Paresce, 1991). This simple comparison suggests convincingly
that the $\beta\:$Pictoris\ disk is not a remnant of planetary formation but on the contrary
must be continually replenished by secondary sources like evaporation or
collision of small bodies. This seems in fact a very natural hypothesis (see
for example Zuckerman \& Becklin 1993): the presence of small bodies is a
consequence of planetary systems formation (Lissauer 1993).
Furthermore, CO absorptions have been recently observed towards $\beta\:$Pictoris\
(Vidal-Madjar et al. 1994) with the Hubble Space Telescope. The very presence
of CO may also require a permanent source provided by evaporation of comet-like
bodies.
Here we present a new argument: This replenishment is able to reproduce
the main characteristics of the $\beta\:$Pictoris\ dust disk, and in particular
the spatial distribution of dust at large distances. In detail, it gives
possible explanations for the following unexplained issues:
\begin{enumerate}
\item The gradient of the scattered light follows a relatively well-known
but unexplained power law (e.g. Golimowski et al. 1993).
\item The distribution at large distances is obviously not axisymmetric
(Smith \& Terrile 1987, Kallas \& Jewitt 1995).
\item The central part of the disk is relatively clear of dust
(Backman et al. 1992, Lagage \& Pantin 1994).
\item The disk seems to be a ``wedge'' disk: the thickness increases
with radius (Backman \& Paresce 1993, Lecavelier des Etangs et al. 1993).
\item The slope of the scattered light distribution
changes abruptly at about 100~AU from the star.
If confirmed, this fact remains unexplained
(Artymowicz et al. 1990, Golimowski et al. 1993).
\end{enumerate}
Furthermore, the possible explanation of these observational facts could
give new understanding to the following questions:
\begin{itemize}
\item a) What is the mass of the dust disk, since when extrapolated towards
infinite distances the disk mass diverges (Artymowicz et al. 1989)?
\item b) Are there connections between the dust and the gas disks ?
\item c) Can the asymmetry in the observed dust disk be connected with the
asymmetry in the longitude of periastron of the comets in the
Falling-Evaporating-Bodies (FEB) model proposed to explain the
redshifted spectroscopic events
(e.g. Beust et al. 1991)?
\end{itemize}
Therefore, we will consider here a model in which the disk is replenished by a
group of kilometer-size bodies (Section~\ref{model}). Numerical calculations of
the spatial distribution of dust are given in Section~\ref{numerical}. We shall
discuss in Section~\ref{bp disk} the $\beta\:$Pictoris\ disk in such a scenario and summarize
the results in Section~\ref{conclusion}.
\section{The model}
\label{model}
\subsection{Basic concept}
We assume that the dust is produced by cometary or asteroid-like
bodies which create particles through mutual collisions or evaporation
processes. The main assumption is thus that these particles have small initial
velocities relative to the parent bodies: less than 1 km/s like in cometary
production (Gombosi et al. 1985, Sekanina 1987) or in "Chiron burst"
(Luu \& Jewitt 1990). Thus, these velocities
can be considered as negligible in comparison to the orbital velocities.
In our model, this relative velocity is fixed at zero.
However, as soon as a dust particle is ejected, it is perturbed by the
radiation pressure; its orbit is different from the parent body one
and tangent
at the point where it was injected
(Burns et al. 1979). For example, if a
parent body in a circular orbit with a semi-major axis $a_0$ produces
a particle with a ratio $\beta $
of the radiation force to the gravitational force,
the particle orbit has
a semi-major axis $a_{\beta}=a_0(1-\beta )/(1-2\beta )$ and
an eccentricity $e_{\beta}=\beta /(1-\beta )$. The periastron
is $a_{\beta}(1-e_{\beta})=a_0$ and the apoastron
$a_{\beta}(1+e_{\beta})=a_0/(1-2\beta )$.
Therefore, particles can be observed at distances from the central star much
larger than the apoastron of the parent body. Thus, local perturbations on
the distribution of the parent bodies like asymmetries, could produce
observable
signatures on the dust distribution at very large distances.
\subsection{Particle size distribution}
\label{size}
Furthermore, because the parent bodies should not produce single sized
particles,
we introduce in the calculation a size distribution. For
particles with radius $s\ge 1\mu$,
$\beta$ is correlated with the size of the particle by $\beta\sim s^{-1}$
(Artymowicz 1988). We can assume a power
law size distribution as in Solar System cometary
dust (Lien 1990 and references therein),
in collisionally replenished dust (Greenberg \& Nolan 1989),
or in interplanetary medium dust (Le Sergeant \& Lamy 1980).
Thus, if we have a size distribution $dn\propto s^{q}ds$, then we have
$dn\propto \beta^{K}d\beta$ with $K=-q-2$.
\subsection{Analytical consideration}
\label{analytical}
\subsubsection{One point production}
First, let us consider a parent body in a circular orbit
which generates a set of particles
at a given point. Then the particles with $\beta >0.5$ have hyperbolic
orbits and are ejected from the system towards the interstellar medium. The
other particles with $\beta <0.5$ follow orbits within a parabola which
is the orbit of the particles with $\beta=0.5$.
One can evaluate the surface density of particles in the asymptotic
direction of the parabola: for a given particle the true anomaly $\theta$
follows the distribution law
\begin{center}
\begin{equation}
P_{\theta}(\theta )d\theta =\frac{(1-e)^{3/2}}
{2\pi (1+e\cos \theta)^2} d\theta
\end{equation}
\end{center}
In the asymptotic direction $\theta=\pi$, $P_{\theta}(\pi)\propto (1-e)^{-1/2}
\propto r^{1/2} (1+a_0/r)^{1/2}$. The probability that the particle apoastron
is between $r$ and $r+dr$ in the same direction
is $P_r(r)= \beta^K (d\beta /dr) dr$. Since
$\beta=(1-a_0/r)/2$, $P_r(r)\propto (1-a_0/r)^K r^{-2} dr$.
Thus, taking into account the scattering cross section
$\sigma\propto \beta^{-2}$ we obtain the surface
density normal to the plane of the disk
$\sigma \cdot n_s$:
\begin{center}
\begin{equation}
\sigma \cdot 2\pi r\cdot n_s(r)\propto r^{-3/2}(1-a_0/r)^K (1+a_0/r)^{1/2}
\end{equation}
\end{center}
Thus $\sigma \cdot n_s(r)_{r \rightarrow \infty}\propto r^{-2.5}$.
\subsubsection{Parent bodies in circular orbits.}
\label{circular orbits}
However, dust production can take place at every point in the
orbit of the parent body.
Here, we restrict ourselves to a parent body in circular orbit,
thus, producing an axisymmetrical disk.
Each particle has a probability $f_{\beta}(r)$ to be at a distance $r$:
$f_{\beta}(r)\propto ((1-a_0/r)(2\beta -1 +a_0/r))^{-1/2}$.
The surface
density normal to the plane of the disk is then:
\begin{center}
\begin{equation}
\sigma \cdot n_s(r)\propto \int_{(r-a_0)/2r}^{1/2}
\frac{\beta^{K-2} f_{\beta}(r) d\beta }
{2 \pi r\int_{a_0}^{a_0/(1-2\beta)} f_{\beta}(r_1)dr_1}
\label{circular}
\end{equation}
\end{center}
This surface density follows an $r^{-3}$ law and is plotted in
Fig.~\ref{fig1}.
Moreover, the inclination of the particles is the same as the inclination of
the parent bodies. Therefore, the vertical distribution of the particles
depends on the distribution of the inclination of the parent bodies. In the
present
model the thickness of the disk is increasing with radius and the volume
density ($n_v$) is proportional to the surface density ($n_s$) divided by the
distance, i.e. $n_v(r)\propto r^{-4}$.
It has to be noted that
this conclusion is valid at distances larger than the distance of the farthest
parent bodies. Within this zone the opening angle must look smaller.
{}From the $n_v(r)\propto r^{-4}$ law and Nakano's (1990) conclusion
concerning the
connection between the volume density and the scattering light distribution
($F(r)\propto n_v(r)/r $),
one can conclude that a belt of parent bodies in circular orbits should produce
a "wedge" disk with a scattered light distribution $F(r)\propto r^{-5}$.
\begin{figure}[tbp]
\ifthisfig
\psfig{file=Disk.ps,height=\columnwidth,rheight=8.5truecm}
\else\picplace{6truecm}\ifnofig\thisfigfalse\else\thisfigtrue\fi\fi
\caption[]{Plot of the surface density distribution calculated from
Eq.~\ref{circular}. The abrupt density decrease at $r\sim a_0$ is
simply due to the assumed single distance between the central star and
all the parent bodies. It is thus an artificial consequence of an
oversimplified hypothesis.\\
For comparison, the hatched zone represents the distribution range of slopes
deduced by different observers around $\beta\:$Pictoris.}
\label{fig1}
\end{figure}
\section{Numerical results}
\label{numerical}
\subsection{The Monte-Carlo simulation}
The analytical model of the previous section can only
be solved for a limited number of
parameters and parent body distributions.
We have chosen to evaluate the dust distribution by a
Monte-Carlo method.
We shall consider a given family of parent bodies, a dust production
law and a particle size distribution, and we shall integrate the distribution
in the following way:
For each of the $N$ random particles, we first
randomly select a parent body in the family and its position in its
orbit; then we randomly apply the particle size distribution and
we calculate the position $r_i$ of the particle $i$ after a time $t$
where $t$ is randomly chosen between
$0$ and $t_{max}$.
Finally, we evaluate $F(r)$
the surface brightness along the midplane
of the edge-on disk projected on the sky at the projected distance $r$.
If $F_i(r)$ is the contribution of the particle $i$,
from Eq.~2 of Nakano (1990), we have
\begin{center}
\begin{equation}
F_i(r)\propto r^{-1}\int _0 ^{\pi}n_i(\Lambda)\sigma_i(\theta)d\theta
\end{equation}
\end{center}
where $\Lambda$ is the distance to the central
star and $\sigma_i(\theta)$ is the scattering phase function.
Nakano (1990) has already shown that, except if the dust
is much more steeply forward-scattering than in the Solar System,
the isotropic scattering assumption does not
change the observed slope of the scattered light
distribution, thus $\sigma_i(\theta)=\beta _i^{-2}$ with
$\sin \theta = r/\Lambda$.
The volume density is given by
$n_i(\Lambda) = \delta(\Lambda-r_i)\cdot p_{i} /r_i$ where
$p_i$ is the production rate of dust on the parent bodies,
then
$F_i(r)\propto r^{-1}\int \delta(\Lambda-r_i) p_i {\rm H}(\Lambda-r)r
d\Lambda /
(r_i \beta _i^{2} \Lambda \sqrt{\Lambda^2-r^2})$
where
H is the Heavyside function.
Finally, the surface brightness
at a given projected distance $r$ from the star is estimated by
\begin{center}
\begin{equation}
F(r)=\sum_{i=1}^{N}F_i(r)
\propto \sum_{i=1}^{N}\frac{ {\rm H}(r_i-r)\cdot p_{i} }
{r_{i}^{2} \sqrt{r_i^2-r^2}\cdot \beta _{i}^{2}}
\end{equation}
\end{center}
\subsection{Results}
Results of runs with different initial conditions on the family of parent
bodies
are summarized in Tables~\ref{resultat_1} and \ref{resultat_2}.
The production rates per annulus of fixed width are given for each run. Several
configurations have been tried. The first one corresponds to the analytical
model of Sect.~\ref{circular orbits}, we find again $\alpha=5.0$, which
validates the Monte-Carlo model. In runs
\#2 and \#3, we have tried different distances and evolutionary times
which show
that the Poynting-Robertson effect does not change the results: the accuracy
for $\alpha$ is about 0.1. From \#4 to \#11, the parent bodies are in
eccentric orbits with eccentricities uniformly distributed between 0 and 0.5 or
between 0 and 1; it seems that bodies with eccentricities larger than 0.5 do
not contribute much to the distribution since a large fraction of ejected
material is then on hyperbolic orbits. In runs \#4 to \#7,
we have taken parent bodies
with a peculiar longitude of periastron ($\omega$),
namely 155$^o$. This value has been selected according
to the Beust et al. (1991) modelisation of the Falling-Evaporating-Bodies,
the so-called FEB scenario needed to reproduce the spectroscopic redshifted
signatures. In the four last runs of Table~\ref{resultat_1}, $\omega$ was
uniformly distributed between 0$^o$ and 360$^o$.
{}From all the results in
Table~\ref{resultat_1}, it can be seen that the slope $\alpha$ is always
greater than or equal to 5.0. In Table~\ref{resultat_2}, we test other parent
body
distributions in which the dust observed at large distances still is
produced close to the star. Incidentally, another solution would be to assume
that the disk is produced by a distribution of asteroids up to 1000~AU from the
star. However this solution does not allow us to explain the observed
asymmetries.
Runs \# 12 to \# 15 take into account the Epstein gas drag with a gas
density $\rho_{gas}=\rho _0 (100\ {\rm AU}/r)$ and temperature $T_{gas}=20$~K.
Runs \# 16 and \# 17 are with dust size distribution characterized by
$K=5-10$.
\begin{table*}
\caption{Slope $\alpha$ of the gradient of brightness in disks with different
conditions for the parents bodies.
In all these runs we assumed N=10000 and K=1.5. Poynting-Robertson drag
is taken into account but there is no gas drag.
For runs \#~4 to \#~7, the two slopes are for the both sides of the
asymmetrical disk seen from the Earth.}
\label{resultat_1}
\begin{tabular}{|c|cccc|c|c|}
\hline
Run&&Parent Bodies Parameters &&&$t_{max}$ &Result: $\alpha$\\
$\#$ & semi-major axis (AU) & Eccentricity & $\omega$ & Production rate &
(years)&$F(r)\propto r^{-\alpha}$ \\
\hline
1 & 20 & 0. & - & $p=const$ & 10000 & 5.0 \\
2 & 2 & 0. & - & $p=const$ & 10000 & 4.9 \\
3 & 20 & 0. & - & $p=const$ & 100000 & 5.0 \\
4 & 20-30 & 0.-0.5 & 155$^o$ & $p=const$ & 10000 & 5.2-5.6 \\
5 & 20-30 & 0.-1. & 155$^o$ & $p=const$ & 10000 & 5.2-5.7 \\
6 & 20-30 & 0.-0.5 & 155$^o$ & $p\propto r^{-3}$ & 10000 & 5.3-5.5 \\
7 & 20-30 & 0.-1. & 155$^o$ & $p\propto r^{-3}$ & 10000 & 5.3-5.6 \\
8 & 20-30 & 0.-0.5 & 0-360$^o$ & $p=const$ & 10000 & 5.2 \\
9 & 20-30 & 0.-1. & 0-360$^o$ & $p=const$ & 10000 & 5.3 \\
10 & 20-30 & 0.-0.5 & 0-360$^o$ & $p\propto r^{-3}$ & 10000 & 5.2 \\
11 & 20-30 & 0.-1. & 0-360$^o$ & $p\propto r^{-3}$ & 10000 & 5.4 \\
\hline
\end{tabular}
\end{table*}
\begin{table*}
\caption{Same as the previous table, with different $K$; gas drag is
taken into account. The parent bodies have semi-major axis $a=$20~AU and
eccentricity $e=0$. $t_{max}$=10000. For the runs \# 16 the slope
gradually changes from $\alpha=3.7$ at 20 AU to $\alpha=4.7$ at 800 AU; and for
run \# 17 from 2.9 to 4.4 respectively.}
\label{resultat_2}
\begin{center}
\begin{tabular}{|c|c|c|c|}
\hline
Run & $\beta$ distribution:K& gas density at 100 AU & Result: $\alpha$ \\
$\#$ & $dn=\beta ^{K}d\beta$ & (cm$^{-3}$) & $F(r)\propto
r^{-\alpha}$ \\
\hline
12 & 1.5 & 100. & 5.0 \\
13 & 1.5 & 1000. & 5.1 \\
14 & 1.5 & 10000. & 4.6 \\
15 & 1.5 & 100000. & 3.8 \\
16 & 5. & 0. & 3.7-4.7 \\
17 & 10. & 0. & 2.9-4.4 \\
\hline
\end{tabular}
\end{center}
\end{table*}
\section{The $\beta\:$Pictoris\ disk}
\label{bp disk}
\subsection{Orbiting-Evaporating-Bodies (OEB). Towards a solution?}
The observed distribution in the external part of the $\beta\:$Pictoris\ disk ($\alpha \le
4.3$) is less steep than all calculated slopes in Tables~\ref{resultat_1}. It
has to be noted that a solution compatible with the observed slope could be
obtained with a gas density $\rho_0 \sim 10^{5}$~cm$^{-3}$ at 100~AU. However,
this is excluded by the upper limit deduced from spectroscopic or HI
observations (Vidal-Madjar et al. 1986, Freudling et al. 1995).
Runs \#16 and \#17 are also compatible with the dust distribution in $\beta\:$Pictoris\
(Fig.~\ref{16}). But they require dust size distributions ($K=5-10$) very
different from the expected ones if they are produced
through collisions between parent
bodies (Fujiwara 1979 and references therein). However they show us that we
need a relatively larger amount of smaller particles which have larger
apoastrons.
In the evaporation process of large bodies (like Chiron), there is
an upper limit $s_{max}$ to the size of the ejected particles extracted by
gas (Luu \& Jewitt 1990). Following Cowan \& A'Hearn (1982) we have calculated
this upper limit and the evaporation rate $Z$ in molecules per
second per unit area for CO and CO$_2$ as a function
of the distance to $\beta\:$Pictoris, assuming $L_{\beta Pic}=6$L$_{\odot}$
and a typical parent body
radius $R_{body}=100$~km (Figs.~\ref{Z} and \ref{smax}). This upper limit is
inversely proportional to the parent body size $R_{body}$.
We have introduced this effect in our model. The results are summarized in
Table~\ref{resultat_co2} for CO$_2$; they would be similar for CO if the parent
bodies were at larger distances. Here, we have taken the production rate
$p\propto Z\cdot R_{body}^2$
and a parent bodies size distribution
$dn=R_{body}^{-\gamma}dR_{body}$ where $\gamma=3.5$ (uncertainties
deduced from observations in the Solar System with $\gamma$
between 3.2 (Whipple 1975, Hughes \& Daniels 1982) and 3.8 (Fern\'andez 1982)
do not change the present results).
In runs \# 101 to \# 103, we take a wide belt of parent bodies from
15 to 30 AU, without correlation between their radius $R_{body}$ and their
distance from the central star. We
obtain $\alpha \sim 4.8-4.9$; the effect of cutoff in particle size
is not efficient
enough to explain the observed slope $\alpha \sim 3.6-4.3$.
However, if the evaporating bodies are coming from larger distances and their
orbital parameters diffuse from a Kuiper belt-like zone towards a
planetary-like
zone, the smaller parent bodies lost their volatile material very early at
larger
distances whereas the larger bodies can evaporate downward at
smaller distances.
This scenario can be modelized in the following way: a parent body
with a characteristic distance (which can be the periastron or the
semi-major axis) $r(t)=r_0-\dot{r}t$ is considered as evaporating at
$r=r(T)$ only if the total mass of the previously evaporated
gas is smaller
than the available mass of volatile material.
\begin{center}
\begin{equation}
\label{dead comet}
\int_{0}^{T} \mu Z(r(t))4\pi R_{body}^{2}\lambda dt < \frac{4\pi }{3}\rho \xi
R_{body}^3
\end{equation}
\end{center}
where $\lambda$ is the percentage of active surface of the body, $\xi$ the
relative mass of CO$_2$ in the parent body, $\rho$ its
density and $\mu$
the CO$_2$ molecular weight.
Eq.~\ref{dead comet} can be reduced to
$R_{body}\ge R_{ref}\int_{r}^{\infty}Z(l)dl/\int_{r_{ref}}^{\infty}Z(l)dl$
where
$R_{ref}$ is the reference size of the smallest body which can evaporate at
$r=r_{ref}$. For CO$_{2}$ evaporation and $r_{ref}=20$~AU, one obtain
\begin{center}
\begin{equation}
\dot{r}=6\cdot 10^{-7} \frac{\lambda}{\xi \rho}
\left( \frac{10 {\rm km}} {R_{ref}} \right){\rm AU\ year}^{-1}
\end{equation}
\end{center}
In runs \# 104 to \# 115, only those with $R_{ref}\sim 30-40$~km give the
observed slope range. Moreover, due to the particle size cutoff,
$\beta \geq 0.4$ there is an abrupt break
in the slope at
$r_{break}\approx a_0/(1-2\beta_{min}) \approx 100\ {\rm AU}$
(see Fig.~\ref{106} for run \#106). For CO, one can obtain
$\dot{r}=6\cdot 10^{-6} \lambda /(\xi \rho) {\rm AU\ year}^{-1}$, and there
would be also an abrupt break in the slope since CO begins to evaporate
from about 100-150 AU.
\begin{figure}[tbp]
\ifthisfig
\psfig{file=s12.ps,height=\columnwidth,rheight=8.5truecm}
\else\picplace{6truecm}\ifnofig\thisfigfalse\else\thisfigtrue\fi\fi
\caption[]{Plot of the surface brightness (arbitrary unit) of the equatorial
plane of an edge-on
disk seen from the Earth
calculated in run \# 16. The slope varies from -3.7 to -4.7.
For comparison, the dashed lines represent the power laws $r^{-3.7}$
and $r^{-4.7}$.}
\label{16}
\end{figure}
\begin{figure}[tbp]
\ifthisfig
\psfig{file=Zco.ps,height=\columnwidth,rheight=8.5truecm}
\else\picplace{6truecm}\ifnofig\thisfigfalse\else\thisfigtrue\fi\fi
\caption[]{Plot of the evaporation rate $Z$ as a function of the distance
from $\beta\:$Pictoris. The thick lines are for CO$_2$ and the thin ones for CO.
The evaporation
is calculated for a steady-state energy balance and depends on the albedo
$A_v$ of the parent body. The short-dashed lines are for $A_v$=0.1, the
long-dashed are for $A_v$=0.8 and the solid line are for $A_v$=0.5. In all the
simulations we take $A_v$=0.5.}
\label{Z}
\end{figure}
\begin{figure}[tbp]
\ifthisfig
\psfig{file=Amax.ps,height=\columnwidth,rheight=8.5truecm}
\else\picplace{6truecm}\ifnofig\thisfigfalse\else\thisfigtrue\fi\fi
\caption[]{Plot of the radius of the largest particle which can be ejected
from the surface of a parent body with radius $R_{body}$=100~km and
density $\rho_{body}=1$
as a function of $\beta\:$Pictoris\ distance. The thick lines are for CO$_2$ and the thin
ones
for CO. The different albedos $A_v$ are the same as in Fig.~3.}
\label{smax}
\end{figure}
\subsection{Discussion}
We saw in the previous section that the observed gradient in the $\beta\:$Pictoris\ disk can
be
explained by the present model only if there is a larger amount of small
particles. Now, there is an upper limit in the size of particles ejected
by slowly evaporating bodies due to a balance between gas drag and
gravitation. In fact, we obtain dust disks
very similar to the $\beta\:$Pictoris\ one by assuming evaporation
of kilometer size bodies which are diffusing from a Kuiper belt-like zone
inward an evaporation zone.
The time scale $\dot{r}\sim 10^{-7}$~AU~year$^{-1}$ in
the diffusion of comet orbital elements, which is necessarily to explain $\beta\:$Pictoris's
characteristics, is
consistent with the few AU in 10 Myr deduced by Torbett \& Smoluchowski (1990)
and Levison's (1991) results on the Kuiper belt dynamical behavior between 30
and 40 AU around the Sun.
The analysis of the evolution of the orbital elements of the parent bodies
is outside the scope of this paper; moreover,
the motion of such Kuiper belt-like bodies under the influence of
a planetary system is probably chaotic (Scholl 1979, Torbett \& Smoluchowski
1990) and very dependent on this planetary system.
The present model merely shows us that
the dust distribution may be an important clue and may
be explained by the present model if there is
a long time scale perturbation of parent bodies by a planetary system which is
compatible with the one predicted for the Solar System.
Another evolution process for heating the bodies consists in the increase
of the star luminosity ($L$)
during its evolution along the main sequence. For a star of
1.6~M$_{\odot}$, we have $dL/Ldt=2\cdot 10^{-10}$~year$^{-1}$, if $\tilde{r}$
is the distance of equal luminosity, this
corresponds to $\dot{\tilde{r}}=\tilde{r}/2L\cdot dL/dt$.
This solution would have the advantage
of being "planetary system" independent. However, if the primary volatile is
CO,
$\dot{\tilde{r}}=10^{-8}$~AU~year$^{-1}$; even if the
volatile is CO$_2$, $\dot{\tilde{r}}=2.5\cdot 10^{-9}$~AU~year$^{-1}$.
The time scales do not seem to be adequate. This mechanism would be efficient
only if we could have bodies with very large quantities of volatiles and
very small fractions of active surface.
The assumption that CO or CO$_2$ can be primary volatiles is not a problem
since around an AV star they can evaporate at very large
distances where the
presence of young objects without a lag deposit of nonvolatile forming a crust
is likely. On the contrary in the Solar System, we observe bright
comets only when H$_2$O can evaporate, simply because
CO or CO$_2$ evaporate only inside
planetary regions where, for dynamical reasons, the presence of young objects
is unlikely: Chiron seems to be an exception but provides a useful analogy
with what may happen around $\beta\:$Pictoris.
As we shall see below, this model is also able to give answers
to other important issues defined in Section~\ref{intro}.
\begin{table*}
\caption{Same as the previous tables, for particles ejected by evaporating
gas. Here we assume that the gas is CO$_2$ and $p\propto Z\cdot R_{body}^2$.
$R_{ref}$ is the radius of the largest dead
comets at $r_{ref}$=20~AU. For runs \# 105 to \# 115
there is a slope break at $r=r_{break}$, and both inward and outward slopes
$\alpha _{in}$ and
$\alpha _{out}$ are given. For runs
\# 113 to \# 115 the two given slopes $\alpha _{in}$ are for each side
of the disk. We always took $t_{max}$=10000 years.}
\label{resultat_co2}
\begin{tabular}{|c|ccc|c||ccc|c|}
\hline
Run && Parent Bodies Parameters &&
$R_{ref}$& &Result: & $F(r)\propto r^{-\alpha}$ & Brightness\\
&&&& & $r_{break}$ & $\alpha _{in}$ &
$\alpha_{out}$ & ratio\\
$\#$ & Periastron (AU) & Eccentricity & $\omega$ & (km) &(AU)&&&\\
\hline
101 & 20-30 & 0. & - & - & - & - & 4.8 & - \\
102 & 20-30 & 0.-0.5 & 0-360$^o$ & - & - & - & 4.9 & - \\
103 & 20-30 & 0.-0.5 & 155$^o$ & - & - & - & 4.9 & - \\
104 & 15-30 & 0. & - & 10. & - & - & 4.6 & - \\
105 & 15-30 & 0. & - & 20. & 50 & 2.9 & 4.5 & - \\
106 & 15-30 & 0. & - & 30. & 100 & 2.7 & 4.3 & - \\
107 & 15-30 & 0. & - & 40. & 180 & 2.5 & 3.9 & - \\
108 & 15-30 & 0.-0.5 & 0-360$^o$ & 10. & - & - & 4.6 & - \\
109 & 15-30 & 0.-0.5 & 0-360$^o$ & 20. & 50 & 2.8 & 4.5 & - \\
110 & 15-30 & 0.-0.5 & 0-360$^o$ & 30. & 130 & 2.8 & 4.3 & - \\
111 & 15-30 & 0.-0.5 & 0-360$^o$ & 40. & 180 & 2.6 & 3.7 & - \\
112 & 15-30 & 0.-0.5 & 155$^o$ & 10. & - & - & 4.6 & 2.9 \\
113 & 15-30 & 0.-0.5 & 155$^o$ & 20. & 50 & 2.2-4.0 & 4.5 & 2.8 \\
114 & 15-30 & 0.-0.5 & 155$^o$ & 30. & 110 & 2.7-3.7 & 4.3 & 3.5 \\
115 & 15-30 & 0.-0.5 & 155$^o$ & 40. & 180 & 2.7-3.4 & 3.7 & 4.0 \\
\hline
\end{tabular}
\end{table*}
\begin{figure}[tbp]
\ifthisfig
\psfig{file=u06.ps,height=\columnwidth,rheight=8.5truecm}
\else\picplace{6truecm}\ifnofig\thisfigfalse\else\thisfigtrue\fi\fi
\caption[]{Plot of the surface brightness $F(r)$
calculated from run \# 106. There
is an abrupt change in the slope at $r=100$~AU. The slope is
$\alpha=-2.7$ inside and $\alpha=-4.3$ outside this limit.
For comparison, the dashed lines represent the power laws $r^{-2.7}$
and $r^{-4.3}$.}
\label{106}
\end{figure}
\subsection{Asymmetry}
The distribution observed at large distances in $\beta\:$Pictoris\ disk
is obviously not axisymmetric
(Smith \& Terrile 1987, Kallas \& Jewitt 1995).
However,
relative keplerian motions should remove such asymmetries: for two particles
on circular orbits with semi-major axis $a_1$ and $a_2$, the time needed
to put them in opposite side is $t\sim 0.4(a_1^{-3/2}-a_2^{-3/2})^{-1}$~years
if $a_1$ and $a_2$ are in AU. With $a_1=100$~AU and $a_2\ge400$~AU we have
$t\le 450 $~years, an extremely short time compared to the age
of the system. Thus,
any asymmetry should quickly disappear.
As noted above, it is however
possible to find the asymmetry observed between the two
extensions of the
$\beta\:$Pictoris\ disk by assuming a peculiar parent body distribution (see
Fig.~\ref{113}).
Runs
\# 112 to \# 115
have been obtained with a peculiar longitude of periastron of
the parent bodies ($\omega=155^o$).
We found a brightness
ratio between the two extensions of the disk ranging from 2.8 to 4.0.
This result shows that
the observed ratio of SW to NE extension brightness from 1.1
(Lecavelier des Etangs et al. 1993) to 1.5 (Kallas \& Jewitt 1995) can be
reproduced if there is a small additional proportion of bodies
with peculiar longitudes of periastron like in the FEB scenario.
However, there is another way to obtain the observed asymmetry, namely
to assume that bodies
with different longitudes of periastron move inwards with different velocities.
For example, if bodies with a given longitude of periastron move more slowly
inward, it produces a
fainter disk extension with a steeper brightness gradient in one particular
direction (Fig.~\ref{108+115}).
\begin{figure}[tbp]
\ifthisfig
\psfig{file=u13.ps,height=\columnwidth,rheight=8.5truecm}
\else\picplace{6truecm}\ifnofig\thisfigfalse\else\thisfigtrue\fi\fi
\caption[]{Plot of the surface brightness of the two extension
calculated from run \# 113.
In this run all the parent bodies have a longitude of periastron
of $155^{o}$ relative to the line of sight. The brightness ratio between
the two extensions is equal to 2.8.
For comparison, the dashed lines represent the power laws $r^{-2.2}$
and $r^{-4.5}$.}
\label{113}
\end{figure}
\begin{figure}[tbp]
\ifthisfig
\psfig{file=u16.ps,height=\columnwidth,rheight=8.5truecm}
\else\picplace{6truecm}\ifnofig\thisfigfalse\else\thisfigtrue\fi\fi
\caption[]{Plot of the surface brightness calculated from the addition
of runs \# 108 and \#115.
In this combined run the parent bodies with a longitude of periastron
of $155^{o}$ go more slowly inwards. The brightness ratio between
the two extensions is equal to 1.3. The fainter extension has a steeper slope
and is represented by a long-dashed line.\\
As in Fig.~1., the hatched zone represents the distribution range of slopes
deduced by different observers around $\beta\:$Pictoris.}
\label{108+115}
\end{figure}
\subsection{Mass of the disk}
\label{mass}
The total mass of the $\beta\:$Pictoris\ disk calculated from optical observations is
\begin{center}
\begin{equation}
M=\int\int n_s(r,s) \frac{4\pi }{3}\rho s^3 2\pi r dr dn_s
\end{equation}
\end{center}
where $s$ is the particle size and $dn_s$ the size distribution. As discussed
above we can take $dn_s=s^{-3.5}ds$. Previously, authors assumed that
$s$ is independent of $r$ the distance from the star. This assumption
is the simpler one and others must be justified.
This means however that:
\begin{center}
\begin{equation}
M\propto
\frac{8\pi ^{2}\rho }{3} \int_{r_{in}}^{r_{out}}n_s(r)rdr \int s^{-0.5}ds
\end{equation}
\end{center}
The second integral diverges towards the largest bodies; the first
one also diverges for $r_{out} \rightarrow \infty$ if the slope of the
scattering light is $\ge-4$ (that is the case according to Artymowicz et al.
(1989), Lecavelier des Etangs et al. (1993) and for
Golimowski et al.'s (1993) NE
extension). Thus to evaluate the mass, an arbitrary limit for the disk
extension and particle size must be fixed.
In the framework
of our model both problems are directly solved. At a given distance
$r$, we have
$
2\epsilon \le s \le \frac{2\epsilon r}{r-a_0}
$
where $a_0$ is the distance of the parent body to the star and $\epsilon$
is defined by $\epsilon=s\beta $. We saw in Section~\ref{size} that
$\epsilon$ can be assumed to be constant. We obtain
\begin{center}
\begin{equation}
M \propto \frac{4\pi \rho}{3}\int \left( n_s(r)2\pi r dr
\int_{2\epsilon}^{(2\epsilon r)/(r-a_0)}s^{-0.5}ds\right)
\end{equation}
\end{center}
\begin{center}
\begin{equation}
M \propto \frac{16\pi ^2}{3}\rho \sqrt{2\epsilon}
\int_{r_{in}}^{\infty} n_s(r)\frac{a_0}{2}dr
\end{equation}
\end{center}
This integral now converges and the computed mass
is of the order of few lunar masses, i.e.
of the same
order as the lower limit evaluated by Artymowicz et al. (1989) who assumed
an arbitrary outer limit of 500~AU and single-size particles
of $1\mu$. This mass is consistent with the total
mass deduced from submillimeter observations (Zuckerman \& Becklin 1993).
\subsection{Gas-dust ratio and CO detection.}
\label{CO}
One can investigate the connection between the dust observed
around $\beta\:$Pictoris\ since 1984 and the gas disk with its stable component thanks to
the recent detection of CO in UV lines with the HST
($N_{co}\approx 10^{15}$~cm$^{-2}$, Vidal-Madjar et al. 1994).
The very presence of CO may also need a permanent replenishment naturally
provided by evaporation of comet-like bodies.
We can evaluate the total mass of dust $M_d$ which is associated with the
evaporation of CO:
\begin{center}
\begin{equation}
M_d=M_{co} \varphi \psi _{dust/gas} \psi _{gas/CO} \frac{t_d}{t_{CO}}
\end{equation}
\end{center}
where $M_{CO}$ is the total mass of CO in the disk, $\psi_{dust/gas}$ and
$ \psi_{gas/CO}$ are the mass ratios of dust to gas and gas to CO, $\varphi$ is
the mass ratio of the dust produced effectively kept in the disk and
$t_d$ and $t_{CO}$ are the life time of dust and CO: due to photodissociation
by UV interstellar radiation $t_{CO}\sim 300$~years (Vidal-Madjar et al. 1994),
and
at a distance of 100 AU where CO begins to be vaporized, $t_d \sim 10^6$~years
(Backman \& Paresce 1993).
{}From the upper limit to particle size ($\beta \ge 0.4$), one can evaluate
$\varphi\sim 0.1$.
With a simple geometry of a disk with an opening angle of 10$^o$,
we can deduce from the observed CO column density
a total CO mass of about
$M_{co} \approx 2\cdot 10^{20}$~kg.
{}From observations in the Solar System, we can take $\psi_{dust/gas} \ge 0.1$
(Newburn \& Spinrad 1989). Finally, $\psi_{gas/CO}$ depends on the composition
of the
evaporated gas; it is between 1. if only CO is present
and $\sim 10.$ if all volatiles evaporate (Mumma et al. 1993).
Thus, with $\varphi \psi _{dust/gas} \psi _{gas/CO} \approx 0.1$,
$M_d$ is about one lunar mass, that is of the same
order as the total mass of the dust disk independently evaluated!
This rough calculation shows that CO and dust are compatible with a common
origin
and produced by the same process: evaporation of comet-like bodies at large
distances from the star.
\section{Conclusion}
\label{conclusion}
The model we have proposed is able to account for the main characteristics of
the $\beta\:$Pictoris\ disk as natural consequences of the production of secondary origin
particles by small bodies. It is in fact an answer to the question 3 of
Zuckerman \& Becklin (1993): are
the dust grains primordial or continually replenished?
Indeed, we have shown that if bodies produce dust particles
with a given size distribution, these particles follow very eccentric orbits.
Thus, a disk can be seen
at large distances from the star, and the spatial distribution of dust
is very close to a power-law.
Moreover, the particles have the same inclinations as
the parent bodies: the thickness of the disk thus
increases with radius, and the
disk is a "wedge" disk: (issue 4).
The asymmetric spatial distribution can be explained if we assume that
the parent bodies distribution is not axisymmetric (issue 2).
This assumption is not surprising since asymmetry is observed for the FEBs
and is probably the case in the Solar System for Kuiper belt objects trapped
in planetary resonances (Jewitt \& Luu 1995).
Since particles are produced with periastrons greater than or equal to those of
the parent
bodies, the dust present close to the star is either produced there or
brought in by Poynting-Robertson drag: the central part must be relatively
clear (issue 3).
Finally, this model gives connection between the particle size and their
distances from the central star, and enable us to solve the issue
of the disk mass (issue a), Section~\ref{mass}).
Thus, the issues 2), 3), 4) and a) listed in Section~\ref{intro} are
simultaneously solved with the simplest form of the model presented here.
However, in order to explain the flatter gradient observed in the $\beta\:$Pictoris\ disk
than the one obtained without additional hypothesis, we propose to add two
other assumptions:\\
1) The dust is produced by evaporation of bodies of few kilometers in radius,
so that there is a balance between evaporating
gas drag and gravitation of these parent bodies,
and only the smallest particles are extracted.\\
2) These bodies are moving slowly towards $\beta\:$Pictoris, so that they become extinct
before arriving close to the star where the evaporating rate is large enough
to produce the largest particles. \\
This OEB (Orbiting-Evaporating-Bodies)
scenario is proposed to solve issue 1), but surprisingly
it is also able to explain the abrupt change in the slope by the
cutoff in distribution of particle size (issue 5)). It has to be noted that
the cutoff must not necessary be very sharp and in fact is rather smooth
in the model since we assumed a parent body size distribution.
The model gives also a natural
connection between the dust and gas disks (issue b): they may well be produced
by the same bodies in the same process.
The connection between the FEB and asymmetry in the disk seems to be possible
(issue c)), however this point needs further analysis.
This new model is able for the first time to explain simultaneously
these issues, but we must now ask if other models or hypothesis can also
be made with the same consequences. If correct, this OEB scenario allows
to see the $\beta\:$Pictoris\ disk as a gigantic multi-cometary tail with all its
components: gas and dust.
\begin{acknowledgements}
We would like to express our gratitude to
Dr. F.Roques who first recall us that ejected particles follow
peculiar eccentric orbit.
We are particularly indebted to the anonymous referee for his very
useful comments
which substantially improved the paper.
Our thanks go also to Dr. M. Friedjung for improving the
manuscript.
\end{acknowledgements}
|
\section{Basic Method}
Previous numerical simulations have been performed
in a similar way, essentially
aimed at producing catalogues of galaxies.
Chokshi {\it et al.} \cite*{chokshi88} simulate galaxies in
clusters uniformly distributed in space.
In each cluster, about 1,000 galaxies are drawn,
all at the same redshift. The magnitudes are sorted according
to the Schechter Luminosity Function
(LF), brighter than a certain cut-off which of course depends on
the cluster redshift.
Yee \cite*{yee91} first sorts the apparent magnitude according to
the observed distribution, the
redshift is then computed in
bins of apparent magnitude according to a certain Luminosity Function
(LF) based on a given, fixed, cosmological and evolutionary model.
Our method is slightly different: first,
a functional form for the evolution is chosen,
and the cosmological
parameters are varied on a certain grid.
There is then a unique evolution parameter ensuring
the condition $<V/V_{a}> = 1/2$ for a given, actual,
sample, $V_{a}$ being the available
comoving volume according to Avni {\it et
al} \cite*{avni81}. The zero--redshift LF $\Phi $ is fixed by such a procedure,
and fitted according to some previous analytic choice
for the LF. Once the evolution of the LF
is known, $V_{a}(M_0)$ may be computed in terms of the zero--redshift
absolute magnitude $(M_0)$. The resulting probability distribution function
{\it in the sample} of
absolute magnitude is then $\Phi (M_0)\times V_{a}(M_0)$.
It is easy to sort the absolute magnitude of object $i$ according to this PDF.
A second parameter, either the apparent magnitude or the redshift,
is needed to define a QSO. Given the absolute
magnitude $M_0$ and the cosmological and evolutionary model,
$V_a(M_0)$ may be computed. A second random drawing according to a uniform
distribution gives a number $x \in [0,1]$, and the redshift of
object $i$ may be computed from the condition: $V=x \times V_{a}(M_0) $.
$(m,z)$, the apparent magnitude--redshift
pair, is then derived from these two variables. For each cosmological model,
the procedure is repeated until the desired catalogue size is reached.
The null hypotheses are: that PLE applies, that the
luminosities are distributed according
to the observed LF and that, according to
the Cosmological Principle, the QSOs are
uniformly distributed in volume on Gpc scales, so that
the variable $V/V_{a}$ is uniformly spread over $[0,1]$. From the LF
we are able to draw
the two variables $M_0$, the zero--redshift absolute
magnitude, and $x$; and from the assumption of uniformity we draw
the $V/V_{max}$ ratio.
Within this framework, particular
attention must be paid to the computation of individual $V/V_{a}$.
\section{Theoretical background}
\subsection{Cosmology }
The universe is described by a Friedman-Robertson-Walker model which is
defined by three independent fundamental parameters (calculated at
$z=0$): the matter
density $\Omega_{m}$, which may include a non-baryonic component, the
cosmological
constant $\Omega_{\Lambda}$, and the Hubble parameter $H_0$. The value
of $H_0$ is very important for the creation of synthetic catalogues because it
scales the distance and the absolute luminosity of the objects.
Nevertheless, $H_0$ has no influence on individual $V/V_{a}'s$, which
are scale free. The other known cosmological parameters are
related to these three parameters.
No arbitrary choice of a preferred cosmological model is made.
General formulae
for the distance and the volume are given in Caroll {\it et al.}
\cite*{caroll92}.
We define the conformal coordinate $\chi (z)$ as:
\begin{eqnarray}
\chi (z)&=&\, 2\, |\Omega_K|^{1/2}\,\, \times \nonumber\cr
&&\nonumber\cr
&& \quad \, \int_0^z {dz'\over ((1+z')^2(1+\Omega_m z')-z'(2+z')\Omega_{\Lambda})^{1/2}} \quad ;
\label{integ}
\end{eqnarray}
\noindent
The volume out to redshift $z$ is written:
\begin{equation}
V(z)=\left({c\over {H_0}}\right)^3{1\over 2\Omega_K|\Omega_K|^{1/2}}\left[
sinn(\chi(z))-\chi(z) \right],
\label{volu}
\end{equation}
\noindent
where $\Omega_K=1-\Omega_m-\Omega_{\Lambda}$ is the curvature, $sinn$
is $sin$ if $\Omega_K<0$ and $sinh$ if $\Omega_K>0$.
This equation is very useful for numerical computation because
it does not require inverse trigonometric functions. Contrary to
what happens in a closed universe, with the expressions given, e.g.,
in Caroll {\it et al.} \cite*{caroll92},
Eq.(\ref{volu}) is single valued and there is no discontinuity at
the value $\chi =\pi $, as shown in Mathez {\it et al.} \cite*{mathez91}.
\subsection{Evolution}
The quasar population resides in an extremely young universe,
and it has evolved up until the present epoch.
It is likely that individual QSOs do not live for a
long period of time,
so we speak rather about the evolution in a statistical sense.
In any case, assuming a PLE model for quasars
amounts to assuming that the fraction of active quasars is
constant with epoch, their luminosity decreasing with increasing
time, and that they are in sufficient number so that
we may apply this average evolutionary law to individual quasars in
order to derive maximum redshifts. The absolute
magnitude at redshift $z$ is given by:
\begin{equation}
M(z)=M_0-2.5log_{10}(e(z)),
\label{evolu}
\end{equation}
\noindent
where $M_0$ is the absolute magnitude at $z=0$, and $e(z)$ is the
evolutionary law:
\begin{equation}
e(z)\equiv (1+z)^{k_L}
\label{pwle}
\end{equation}
for a Power Law Luminosity Evolution (PWLE), and
\begin{equation}
e(z) \equiv exp\left(k_L t(z)/H_0\right) = exp\left(t(z)/\tau \right)
\label{lexp}
\end{equation}
for an Exponential Luminosity Evolution (LEXP); $k_L$ is the evolutionary
parameter, $t(z)$ the look-back time and $\tau $ the
characteristic evolutionary time. Strictly speaking,
as quoted by Bigot and Triay \cite*{bigot91}, this
evolution law has only a statistical sense and
does not apply to individual QSOs. However, to compute each $V/V_{a}$ value, we
fictitiously move each QSO toward high redshifts, and it is necessary to apply
this evolution law to each QSO.
Assuming a power law spectrum $f_\nu=\nu ^{-\alpha}$ with a spectral
index $\alpha=0.5$, the Mattig relation relates the observables
$(m,z)$ of a quasar to $M_0$, its absolute magnitude at $z=0$:
\begin{equation}
m_{M_0}(z)=M_0-5+2.5log_{10}\left({d_L^2(z) K(z)\over e(z)}\right);
\label{mattig}
\end{equation}
\noindent
here $K(z)= (1+z)^{(\alpha-1)}$ is the K-correction.
\begin{figure*}
\picplace{20cm}
\caption{%
Four possible $apparent\; magnitude-redshift$
plots: monotonicaly increasing, one maximum,
both one maximum and one minimum, and monotonicaly decreasing.
The evolutionary tracks
are drawn according to the cosmology and exponential luminosity
evolution as described by the top labels. The $m=m(z)$ curves (solid lines)
correspond to bins of constant width in absolute magnitude.
The horizontal dashed
line represents a constant limiting magnitude $B=21$; the dashed $m(z)$
line corresponds to the faint separator ${\cal S}_2$ between
magnitude-limited and redshift-limited quasars (see text).
Depending on the parameters (cosmology and evolution),
the same quasar may be
either magnitude-- or redshift--limited.
In models where $m(z)$ shows a maximum (for example in Fig. 1b),
for quasars lying between $\cal S$ and $B=21$ there exists a
redshift slice where $m(z)$ is above $B=21$. The corresponding volume
must be subtracted from the available volume $V_2-V_1$.
Quasars of the complete
Boyle {\it et al.} (1990) sample are superimposed.
}
\end{figure*}
The evolutionary tracks $m(z)$ on the Hubble diagram
are shown in Fig. 1 for a given set of quasars in
four different cosmologies and evolutionary histories.
The curves appear to be quite
different. In particular the quasars in Fig1.b,c,d become brighter and
brighter as the redshift increases, because the evolution
dominates the cosmological dimming at large $z$.
\subsection{Distribution of the QSOs in the Hubble plane}
Available complete QSO samples usually have a well defined
limiting magnitude, towards faint magnitudes. Due to saturation
effects (in the photometry or in the spectroscopy), a bright limiting
magnitude may also apply.
All UVX QSOs are limited to redshifts $z\le 2.2$, precisely because of
their selection as objects with a UV excess, a standard way to detect
quasars in optical astronomy.
As the redshift decreases (below $z=0.3$), there
is another limitation because quasars become spatially
resolved and can no longer be classified as stellar objects.
These redshift limitations prevent the use of maximum (respectively minimum)
redshifts outside the limiting redshifts of the sample.
Assume that we have a QSO sample in the magnitude
and redshift ranges $[m_{1},m_{2}]$ and $[z_1,z_2]$, respectively.
For $UVX$ samples we would have $z_1=$0.3 and $z_2=$2.2.
We define $\Phi (M,z)$ as the LF at redshift $z$ and at
absolute magnitude $M$. However, for the case of PLE it
can be written $\Phi (M(z))$; then all luminosity
functions at different redshifts differ by an additive constant along
the absolute magnitude axis. This
allows us to work with the LF defined at a fixed epoch
($z=0$) where $\Phi (M(z=0))=\phi (M_0)$, whatever the evolutionary law
\cite{kassiola91}. The number
of objects in the range of absolute luminosity $[M_0,M_0+dM_0]$ and
redshift $[z,z+dz]$ writes:
\begin{eqnarray}
d^2N=\phi (M_0) \, dM_0 \, H(m_2-m) \, H(m-m_1)\;\; \times \hfill \nonumber\cr
\hfill \, H(z_2-z) \, H(z-z_1) \, {dV\over dz}, \, \, \, dz \;\;\; ,
\label{dene}
\end{eqnarray}
\noindent
with the following notations:
\begin{itemize}
\item $dV$ is the comoving volume element of
the shell of redshift $z$ to $z+dz$ in the
solid angle $\omega $ of the survey;
\item $\phi (M_0)$ is defined above;
\item The four Heaviside functions are the selection function,
ensuring that apparent magnitudes and redshifts remain inside 'box
$\cal B$', or $[m_1,m_2]\otimes[z_1,z_2]$,
the limitations due to the selection criteria.
\end{itemize}
We now define ${\cal D}$ as the domain in the Hubble
plane where all quasars are redshift--limited.
Let $z_{top}$ (respectively $z_{bot}$) be the redshift in
the range $[z_1,z_2]$
at which the evolutionary track passes through
the maximum (minimum) of the $m(z)$ curve. $z_{top}$ ($z_{bot}$) is
the same for all quasars, since
their $m(z)$ curves are parallel, differing by
an additive constant equal to the absolute magnitude. Consider a given
absolute magnitude $M_0$, and
let $m_{M_0,top}$ ($m_{M_0,bot}$) be the corresponding
faintest (brightest) apparent
magnitude in box $\cal B$:
\begin{eqnarray}
m_{M_0,top}&=\sup_{z\in [z_1,z_2]}\left[ m_{M_0}(z) \right]&= m_{M_0}(z_{top});\nonumber\cr
\\
m_{M_0,bot}&=\inf_{z\in [z_1,z_2]}\left[ m_{M_0}(z) \right]&= m_{M_0}(z_{bot}).
\label{topbot}
\end{eqnarray}
$m_{M_0}(z)$ is the apparent magnitude corresponding to the
absolute magnitude $M_0$ at redshift $z$.
$z_{top}$ is equal to $z_2$ in the case of little or no luminosity
evolution. Consider the $m(z)$ curve
corresponding to $m_{top}=m_{2}$: it is unique; let us
call it the 'faint separator'
${\cal S}_2$. There is also a 'bright separator' -- the unique curve ${\cal S}_1$
crossing $(z_{bot},m_{bot}=m_1)$. ${\cal S}_2$ is drawn on Fig. 1.
Domain ${\cal D}$ is simply
the domain located {\it below } ${\cal S}_2$ and {\it above} ${\cal
S}_1$.
Let $M_0^{S_i}$ be the two absolute magnitudes corresponding to
the curves ${\cal S}_i, \; (i=1,2)$. The domain ${\cal D}$ simply
consists of that part of the Hubble plane brighter than $M_0^{S_2}$ and
fainter than $M_0^{S_1}$. We call $\cal D^*$ the complement of $\cal D$
with respect to the box $\cal B$. We call $\cal D$ and $\cal D^*$
quasars the quasars located in each of these domains.
\subsection{The $V/V_{max}$ test}
Since our method relies on $V/V_{a}$ values, it is necessary
to review the $V/V_{a}$ computation, particulary in the case of strong
evolution.
Consider a quasar in the sample with $(m,z)$ as its representative point
in the Hubble diagram, and define the following quantities:
\begin{itemize}
\item $M_0$: the quasar absolute magnitude at epoch $z=0$;
\item $z_{max}(M_0,m_i) \;(i=1,2)$: the redshift(s)
where the quasar has an apparent magnitude $m_i$;
\item $V_{max}(M_0,m_i)$: the corresponding comoving volume(s);
\item $V_{a}$: the available volume;
\item $V_i \; (i=1,2)$: the comoving volumes enclosed by
the two limiting redshifts $z_i$.
\end{itemize}
The various possible shapes of the evolutionary track (Fig. 1) are
crucial for understanding the $V/V_{a}$ test.
The quasar may leave the sample in one of two ways:
either its magnitude leaves the range $[m_1,m_2]$, or its
redshift leaves the range $[z_1,z_2]$. The available volume
$V_{a}$ does {\it not} reduce to $V_{max}(M_0,m_2)$:
it is the total volume $V_2-V_1$, minus $\sum V_{out}$,
the sum of the volumes of all slices of the universe where
the evolutionary track is outside box $\cal B$,
so that it prevents the quasar from remaining in the sample.
Assuming that volumes are randomly distributed over the range $[V_1,V_2]$,
the variable uniformly distributed over $[0,1]$ is the ratio $x$ defined
below. The number of quasars in the redshift range $[z_1,z_2]$ is:
\begin{eqnarray}
N&=&\int_{(M_0)_1}^{(M_0)_2}dM_0 \Phi (M_0) \int_{sup(z_1,z_{max}(M_0,m_1)}^{inf(z_2,z_{max}(M_0,m_2))}dz{dV\over dz} \nonumber\cr
\\
&=&\int_{(M_0)_1}^{(M_0)_2}dM_0 \Phi (M_0)V_{a}(M_0),
\label{numb}
\end{eqnarray}
where:
\begin{eqnarray}
V_{a}(M_0)&=&inf(V_2,V_{max}(M_0,m_2)) \;\; - \cr
&& sup(V_1,V_{max}(M_0,m_1))\;-\;\sum V_{out} .
\label{available}
\end{eqnarray}
\noindent
Note that for $\cal D$ quasars
$ \; inf(z_2,z_{max})=z_2$ and $sup(z_1,z_{min})=z_1$ {\it
simultaneously}.
It is straightforward to find the new variable $x$:
\begin{equation}
x=(V-V_{sup}-\sum V_{out})/(V_{inf}-V_{sup}-\sum V_{out}).
\label{nvar}
\end{equation}
\noindent
Indeed, one shows that the mean of the ratio $x$ over the sample is
$1/2$:
\begin{eqnarray}
<x>&=&{1\over N} \int_{(M_0)_1}^{(M_0)_2}dM_0 \Phi (M_0)\;\;\times \cr
&& \int_{V_{sup}}^{V_{inf}} dV
{(V-V_{sup}-\sum V_{out})\over (V_{inf}-V_{sup}-\sum V_{out})} \nonumber\cr
&&\nonumber\cr
&&\nonumber\cr
&=&1/2, \qquad \forall \Phi , \forall (M_0)_1, \forall (M_0)_2;
\label{undemi}
\end{eqnarray}
\noindent
moreover, since $dV=V_{a}(M_0) dx$, we have:
\begin{eqnarray}
dN&=&\int_{(M_0)_1}^{(M_0)_2} dM_0 \Phi (M_0)V_{a}(M_0) \; dx\nonumber\cr
&&\nonumber\cr
&&\nonumber\cr
&=&N \;\; dx,
\label{unif}
\end{eqnarray}
\noindent
which characterizes a PDF of $x$ which is uniform
over $[0,1]$. For a redshift--limited sample over $[z_1,z_2]$,
taking $V_{sup}=0$ overestimates $V/V_{a}$ and taking $V_{inf}=V_{a}$
underestimates $V/V_{a}$. In either case, $<x>$ cannot be equal to
$1/2$ {\it even in the correct cosmology and with the correct evolutionary
law.}
Varying the shape of the evolutionary
track changes the expression of the variable $x$.
There may exist zero, one or several maxima of $m(z)$, or
the quasar may become fainter than $m_2$ at low redshift; in fact
there may be several roots of the equation $m(z)=m_2$.
Fig. 1 shows various curves $m(z)$ for various cosmologies and
evolutionary histories, in order to illustrate the existence of
redshift ranges, lying inside the range $[z_1,z_2]$,
but for which $m(z)$ is {\it not} in the range $[m_1,m_2]$. Such ranges
must be excluded from the volume computation.
\section{ Luminosity Functions in real samples}
The LF is not a direct observable -- on the contrary it depends on the
choice of the cosmological model and of the evolutionary law.
For our purpose, it is
necessary to be realistic in reproducing the data; in particular, the artificial
catalogues must have the same LF as the real sample
whatever the choice of their parent cosmological model.
As discussed in Kassiola and Mathez, \cite*{kassiola91},
given the cosmology and the
evolution, there are two ways to derive the LF from data:
firstly one can construct the {\it Global Luminosity
Function} (GLF), by shifting the absolute magnitudes of the whole sample
to the present epoch $z=0$, according to the evolution law, and secondly
one can compute the {\it Restricted Luminosity Functions} (RLFs), i.e.
the LFs in different redshift bins.
A necessary condition for PLE to apply is that the LF in all redshift
bins differ by a simple shift along the absolute magnitude axis.
To do this, one has to
compare the RLFs. Assuming that PLE is correct, the reason for choosing
the GLF to construct the catalogues is to avoid the noise created by
binning the data in redshift. In Section 4.3, we test the
PLE hypothesis by comparing of the RLF and the GLF.
\subsection{ Real data: the Boyle {\it et al.} sample }
The complete sample of $UVX$ quasars of Boyle {\it et al.} (1990) is used.
It contains
383 quasars, if we exclude the Narrow Emission Lines (NL) and
the redshifts outside the range $[0.3,2.2]$.
\subsection{The Global Luminosity Function and its derivation}
We now consider the computation of the GLF for any redshift--and magnitude
limited--QSO sample.
The sample is binned in absolute magnitude $M_0$ with bin width $\Delta M_0$.
Let $N_i$ be the QSO number in the $i^{th}$ absolute magnitude bin.
The GLF in bin $i$ is given by:
\begin{equation}
\phi _{obs}^i (M_0)=\sum_{j=1\atop M_{0j}\in B_i}^{N_i} \left(1/ V_a^j\right),
\label{fctlum}
\end{equation}
\begin{eqnarray}
B_{i}= \left[M_0-{\Delta M_0\over 2},M_0+{\Delta M_0\over 2}\right],
\end{eqnarray}
\noindent
with an error on $\phi$:
\begin{equation}
\sigma_\phi^i=\left(\sum_j^{N_i} {V_a^j}^{-2} \right)^{1/2}.
\label{erreur}
\end{equation}
\noindent
For the case of the
Boyle sample, none of the eight fields intersect; thus, the
available volume follows directly from Eq. (8):
\begin{eqnarray}
V_a&=&\sum_k^{N_{fields}} \omega_k \;\; \times \cr
&& \left[ inf(V_{max}^k,V_2)-sup(V_{min}^k,V_1)- \sum V_{out} \right],
\label{vola}
\end{eqnarray}
\noindent
where $\omega_k$ is the solid angle of the $k^{th}$ field.
The GLF looks similar for all of the cosmologies we tried.
Special attention has been paid to some of the cosmological models
(e.g. $\Omega =0, \Lambda=0;\; \Omega =1, \Lambda=0;\; \Omega =0,
\Lambda=0.8$), by various authors which
found that the LF is well fitted by either a single
\cite{marshall83} or a double \cite{boyle88} power--law. We adopted
the double power law--model in all cosmologies and for the two functional
forms of evolution:
\begin{equation}
\phi _{model} (M_0)={\phi^\star\over
10^{0.4(M_0-M_\star)(\alpha+1)}+10^{0.4(M_0-M_\star)(\beta+1)}};
\label{funcform}
\end{equation}
\noindent
$M_\star$ is the characteristic absolute magnitude of the two power--law
distribution, i.e. the knee of the distribution. Absolute magnitudes
for the quasars are computed according to Eq. (\ref{mattig}). The
evolution parameter $k_L$ is computed by setting $<x>=1/2$ for the
whole sample. Such a procedure differs from current work
\cite{marshall83,boyle88}, in which $k_L$ is
fitted like the other parameters. The free parameters $\phi^\star$, $M_\star$, $\alpha$, and
$\beta$ are derived by a $\chi^2$ minimization of the binned GLF:
\begin{equation}
\chi^2=\sum_{i=1}^{N_{bin}} \left({\phi_{obs}^i-\phi_{model}\over
\sigma_{\phi_{obs}}^i} \right)^2,
\label{chi2}
\end{equation}
\noindent
where $N_{bin}$ is the number of absolute magnitude bins.
An important difference with other LF computations
is that we have only four free parameters.
\begin{figure}
\picplace{8cm}
\caption{Global Luminosity Functions in four cosmologies, for a power--law
luminosity evolution. The models $a,b,c,d$ of Table 1 progress from the
larger to the thicker line.}
\end{figure}
\begin{figure}
\picplace{8cm}
\caption{Same as above, exponential luminosity evolution. The models
$e,f,g,h$ of Table 1 progress from the larger to the thicker line.}
\end{figure}
Table 1 and Figs. 2 and 3 show a panel of GLFs computed for four different
cosmologies, scaled to $H_0=50 \; km s^{-1} Mpc^{-1}$.
The general behaviour of the GLF is identical for the
exponential and power--law evolutions.
The GLF is enlarged and QSOs are brighter for a $\Lambda $--dominated
universe. This is due to the fact that, at fixed redshift, the distance in
this type of model is greater than in other models,
and therefore the absolute luminosities are larger. The effect is in
the opposite sense
for matter--dominated universes where the objects become fainter and the GLF
narrower. The knee of the luminosity distribution evolves according to
the preceding remarks. The faintest, shortest and densest GLF is obtained
for $\Omega=1$ and $\Lambda=0$.
The evolution becomes weaker in $\Lambda$-dominated universes but has
a more standard value ($k_L$ around 3--4 in PWLE models)
if $\Omega$ is greater than 0.4, regardless of the value of $\Lambda$.
Despite our different LF computation method, we find results quite similar to
those of Malhotra {\it et al.} \cite*{malhotra95} and Boyle {\it et al.}
\cite*{boyle90}.
\subsection{ The Restricted Luminosity Functions }
Restricted LFs (RLFs) have been computed from the Boyle sample
in three different redshift bins
($[0.3,1.0]$, $[1.0,1.7]$, $[1.7,2.2]$).
We are interested in checking whether or not the
PLE hypothesis is confirmed. If this is the
case, any RLF may be deduced from the GLF by
a simple shift along the absolute magnitude axis according to
the evolution law (Eq. \ref{evolu}). Since we have computed
the GLF in the last
Section, we are able to compare our RLF to the GLF.
This is done by a $\chi^2$--procedure between the GLF and the three RLFs:
\begin{equation}
\chi^2=\sum_{i=1}^{N_{bin}} \left({\phi_{GLF}-\phi_{RLF}^i\over
\sigma_{\phi_{RLF}}^i} \right)^2.
\label{chi2}
\end{equation}
Here, $\phi_{RLF}^i$ and $\sigma_{\phi_{RLF}}^i$ refer to the RLF, computed
from the data in the limited redshift bins, and $\phi_{GLF}$ is shifted to
this redshift range.
The result is shown in Table 1
for four cosmologies and the two evolutionary models. The probabilities
$P(>\chi^2)$ are high, and, except for one case
(LEXP, $\Omega_{mat}=\Lambda=1$),
the test succeeds to better than the 15\% confidence level.
This result is in condradiction to those of a similar
test done by Kassiola and Mathez (1991), which found strong discrepancies
between the GLF and the RLF. However, they interpret these discrepancies
as the
lack of homogeneity of their composite catalogues.
Since our results are consistent with PLE (favoured by Boyle {\it et al.},
1987), we use this hypothesis to construct our synthetic catalogues.
\section{ Evolution characteristic time in various cosmologies }
In the exponential model, the characteristic evolution time $\tau $
enters explicitely in Eq. (\ref{lexp}), the expression for $e(z)$.
This is not the case for the power law model -- but a characteristic time
may be defined as the look--back time to which all luminosities were
higher by a factor $e$:
\begin{equation}
\tau = t\left(z_k = exp(1/k_L) - 1\right) .
\label{tau}
\end{equation}
Table 2 (3) gives the evolutionary parameters $k_L$ for a grid of
cosmological parameters in PWLE (LEXP).
Table 4 (5) gives the corresponding PWLE (LEXP) characteristic times,
computed according to
Eqs. (\ref{tau}) and (\ref{lexp}), respectively.
These times are in units of the Hubble time $H_0^{-1}$.
As is well known, the quasar evolution time is of the order of
$H_0^{-1}/(7.7\pm3.5)$ for exponential evolution
\cite{mathez76,marshall83,boyle88},
strongly dependent on the cosmological model. For power--law
evolution, the characteristic time is
$H_0^{-1}/(3.7\pm 0.6)$.
Tables 6 and 7 give the same characteristic times as Tables 4 and 5,
but in units of the age of the universe in each cosmological model.
With the exception of the $\Omega =0, \, \Lambda =1$ model,
all ratios in
Table 6 are close to 0.30$\pm 0.02$, implying that the characteristic
evolution
time is about equal to 1/(3.2$\pm 0.2$) of the age of the universe,
whatever the cosmological model, in PWLE models. In
LEXP models, this ratio is 1/(6.7$\pm $0.9).
The characteristic times in Tables 4 and 5, apparently quite
dependent of the cosmology, are indeed surprinsingly
constant when expressed in terms of the age of the universe.
\section{An algorithm for producing quasar catalogues}
\medskip
\subsection{ The algorithm}
Fixing the evolutionary law, the cosmology (including
the choice of $H_0$) and
the catalogue limits (i.e. $m_1,m_2,z_1,z_2$
which allow the box $\cal B$ and domains $\cal D$ and $\cal D^*$
to be defined), each QSO $(m,z)$ is fully
determined by the other pair of variables $(V/V_{a},M_0)$.
This is easy to understand from Fig. 1, where the Mattig functions appear.
All of our previous discussion shows that under the PLE hypothesis
(which was checked on the data), any sample is fully determined from
its limiting magnitudes and redshifts, the GLF
and a uniform distribution of the $V/V_{a}$. From the probability
distribution function (PDF) of absolute magnitudes $M_0$,
which is $\phi (M_0) \times V_{a}(M_0)$,
and from the $V/V_{a}(M_0)$ ratios uniformly spread over $[0,1]$, it
is possible to draw Monte--Carlo catalogues.
The algorithm is shown in Fig. 4.
Choosing the cosmology and a functional form for the evolution, the first
part of the work is to compute the evolutionary parameter
from the constraint $<V/V_{a}>=0.5$, using the Boyle sample.
At the same time, $V_{a}(M_0)$ is computed and used to obtain the GLF.
In the second part,
two series of random numbers, $\epsilon_1$ and $\epsilon_2$, are used to
extract the $V/V_{a}$ and the absolute magnitude $M_0$ of each object in
the synthetic catalogue from the relations:
\begin{eqnarray}
\epsilon_1&=&V/V_{a}={V-V_{sup}-\sum V_{out}\over V_{inf}-V_{sup}-\sum
V_{out}},\nonumber\cr
\epsilon_2&=&{\int_{-\infty}^{M_0} dM \phi (M) V_{a} (M) \over
\int_{-\infty}^{+\infty}dM \phi (M) V_{a} (M) }.
\label{catapar}
\end{eqnarray}
Obtaining the redshifts and apparent magnitudes is straightforward from a
simple inversion procedure. Synthetic catalogues have been produced in
the past, where both a LF {\it plus} a redshift
histogram were jointly fitted. As already explained, this
method suffers from at least two problems:
Essentially, the two distributions may
be incompatible in different cosmologies.
Our method has the advantage of avoiding these problems since we use only one
distribution, the GLF, and the other distributions (redshift, magnitude)
are derived only after the cosmology and the evolution have been fixed.
The algorithm shown in Fig. 4 ensures the coherence of these
distributions, moreover it takes the evolution into account very
accurately.
This is not obviously true in previous methods where it is not clear how
the evolution intervenes. In fact, we used the Cosmological Principle
to replace the redshift
distribution of the standard method by demanding the uniformity of $V/V_{a}$.
Moreover, the limiting magnitude and the redshift range for the synthetic
sample may be different from their values in the input sample.
\begin{figure}
\picplace{10cm}
\caption{Algorithm for the construction of synthetic catalogues}
\end{figure}
\subsection{ Redshift and Magnitude Histogrames}
The input \cite{boyle90} and output (synthetic catalogue with the
same observational biases) redshift histogrames are compared in Fig.
5 for models $b$ and $f$ of Table 1.
Similarly, the histogrames of absolute magnitude are compared in Fig. 6.
The fit is quite satisfactory for both models; this is true for all models
we tried.
It is possible to introduce various biases in the synthetic catalogues,
depending on which sort of catalogue we want to simulate. For example, for
UVX quasars it is well known that the color selection criteria introduce
a deficiency of QSOs in the redshift range $[1.5,1.8]$ \cite{boyle88}.
Moreover, QSOs are
variable objects, which can be simulated by randomly selecting a magnitude
variation
in a pre--defined distribution function. Noise in the magnitude and redshift
measurements may be simulated in the same way.
\begin{figure}
\picplace{8cm}
\caption{Redshift Histogrames: (1): Boyle sample; (2): simulated
catalogue, translated 15 upwards, model $b$; (3): idem, translated 35 upwards,
model $f$ }
\end{figure}
\begin{figure}
\picplace{8cm}
\caption{Absolute Magnitude Histogrames: dashed curve: Boyle sample;
full line: simulated catalogue, model $b$}
\end{figure}
\begin{figure}
\picplace{8cm}
\caption{Absolute Magnitude Histogrames: dashed curve: Boyle sample;
full line: simulated catalogue, model $f$}
\end{figure}
\section{Conclusion}
We have reviewed in detail the $V/V_{max}$ computation in various cosmologies.
The Mattig relation is used to understand how to compute the available volume
$V_a$ of a QSO, even in complicated cases. Both the Pure Luminosity
Evolution (PLE) assumption and the Cosmological Principle allow
the variables $(M_0,V/V_{a})$ to be used instead of $(m,z)$ to construct
synthetic QSO samples in a coherent way.
Some of the previous catalogue constructions were shown to suffer problems.
The Global Luminosity Function used for this purpose is derived in a
slightly different way from previous calculations
since we fixed the evolution parameter by the condition
$<V/V_{a}>=1/2$. This significantly changes the value of this parameter,
and hence of the GLF.
A surprising result is that the QSO characteristic evolution times
are constant when expressed in terms of the age of
the universe, regardless of the cosmology.
The PLE hypothesis was checked on the data, which comfirms previous results
\cite{boyle88}. The advantage of such catalogues is that we tightly control all
aspects of the sample:
the parent cosmology, the strength and the functional form of the
evolution, the luminosity function, the magnitude and redshift depths,
and the effects of magnitude and redshift biases, all of which influence
the results of the new cosmological test introduced in Paper I. In Paper
I we apply the
test to catalogues with different redshift limitations, which
leads us to propose a new observational strategy for the construction of
future QSO samples.
It is possible to test a catalogue with a wrong hypothesis, for
example on the evolution, by assuming a power--law evolution and running the
cosmological test with the exponential hypothesis.
Moreover, our method is tractable with a Pure Density Evolution hypothesis,
provided that density--weighted volume elements $\rho dV$ are substituted
for all volume elements $dV$.
\begin{table*}
\caption[]
The LF (Eq. (\ref{fctlum})) in various redshift ranges
for four arbitrary cosmologies and two evolution laws.
Since we are interested in input parameters to construct catalogues,
confidence intervals are not needed. Only the best
fit parameters are given. We choose $H_0=50 km s^{-1} Mpc^{-1}$
}
\begin{flushleft}
\begin{tabular}{ccccccccccccccc}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
&&&&&&&&&&&\cr
&$\Omega $&$\Lambda $&\hfil $k_L$&\hfil $\alpha$\hfil &\hfil $\beta$\hfi
&\hfil $M_{\star}$& $\phi^{\star}$&\hfil $P(>\chi^2)$\hfi
&\hfil $P(>\chi^2)$\hfil &\hfil $P(>\chi^2)$\hfil &\hfil $P(>\chi^2)
\hfil \cr
& &&&&&&$10^{-6} Mpc^{-3}$&$z_1=$0.3 & 0.3&1.0&1.7 \cr
& &&&&&&&$z_2=$2.2 & 1.0&1.7 &2.2\cr
\hline\noalign{\smallskip}
PWLE&&&&&&&&&&\cr
a&0&0&2.94&-3.32&-1.38&-23.41&1.8$10^{-6}$&\hfil 0.97\hfil &\hfil 0.20\hfil &\hfil 0.78\hfil &\hfil 0.80\hfil \cr
b&1&0&3.60&-3.89&-1.30&-22.10&1.1$10^{-5}$&\hfil 0.99\hfil &\hfil 0.99\hfil &\hfil 0.69\hfil &\hfil 0.92\hfil \cr
c&0&1&1.91&-2.43&-1.23&-23.76&2.6$10^{-6}$&\hfil 0.89\hfil &\hfil 0.15\hfil &\hfil 0.88\hfil &\hfil 0.58\hfil \cr
d&1&1&3.81&-4.22&-1.37&-22.25&3.2$10^{-6}$&\hfil 0.99\hfil &\hfil 0.92\hfil &\hfil 0.66\hfil &\hfil 0.81\hfil \cr
\noalign{\smallskip}
\hline
\noalign{\smallskip}
LEXP&&&&&&&&&&\cr
e&0&0&6.65&-3.51&-1.49&-22.18&1.5$10^{-6}$&\hfil 0.99\hfil &\hfil 0.58\hfil &\hfil 0.78\hfil &\hfil 0.40\hfil \cr
f&1&0&11.8&-4.61&-1.55&-19.84&6.3$10^{-6}$&\hfil 0.97\hfil &\hfil 0.88\hfil &\hfil 0.57\hfil &\hfil 0.24\hfil \cr
g&0&1&1.90&-2.69&-1.49&-24.66&1.2$10^{-6}$&\hfil 0.76\hfil &\hfil 0.16\hfil &\hfil 0.23\hfil &\hfil 0.58\hfil \cr
h&1&1&9.71&-4.62&-1.60&-20.13&2.1$10^{-6}$&\hfil 0.99\hfil &\hfil 0.64\hfil &\hfil 0.57\hfil &\hfil 0.01\hfil \cr
\noalign{\smallskip}
\hline
\end{tabular}
\end{flushleft}
\end{table*}
\begin{table}
\caption[]
Power Law Luminosity Evolution (PWLE) parameter in different cosmologies.
The model $\Omega =0, \, \Lambda =1$, which is marked with an
asterisk, is close to a universe without a Big--Bang, and not
very reliable.%
}
\begin{flushleft}
\begin{tabular}{lllllllllllllll}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
$\Lambda $ &$\Omega $& 0.0 & 0.2 & 0.4 & 0.6 & 0.8 & 1.0 \cr
0.0 && 2.94 & 3.16 & 3.28 & 3.44 & 3.50 & 3.60 \cr
0.2 && 2.88 & 3.13 & 3.28 & 3.47 & 3.50 & 3.63 \cr
0.4 && 2.78 & 3.10 & 3.28 & 3.47 & 3.53 & 3.66 \cr
0.6 && 2.66 & 3.06 & 3.28 & 3.47 & 3.60 & 3.69 \cr
0.8 && 2.47 & 3.00 & 3.31 & 3.47 & 3.63 & 3.75 \cr
1.0 && 1.91(*) & 2.94 & 3.35 & 3.50 & 3.66 & 3.81 \cr
\noalign{\smallskip}
\hline
\end{tabular}
\end{flushleft}
\end{table}
\begin{table}
\caption[]
Same as Table 2 for LEXP, Exponential Luminosity Evolution.}
\begin{flushleft}
\begin{tabular}{lllllllllllllll}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
$\Lambda $ &$\Omega $& 0.0 & 0.2 & 0.4 & 0.6 & 0.8 & 1.0 \cr
0.0 && 6.65 & 8.02 & 9.15 & 10.10 & 11.02 & 11.77 \cr
0.2 && 6.02 & 7.40 & 8.65 & 9.71 & 10.59 & 11.40 \cr
0.4 && 5.28 & 6.84 & 8.15 & 9.28 & 10.15 & 11.03 \cr
0.6 && 4.40 & 6.15 & 7.53 & 8.78 & 9.71 & 10.65 \cr
0.8 && 3.40 & 5.40 & 6.96 & 8.21 & 9.28 & 10.15 \cr
1.0 && 1.90(*) & 4.59 & 6.28 & 7.65 & 8.78 & 9.71 \cr
\noalign{\smallskip}
\hline
\end{tabular}
\end{flushleft}
\end{table}
\begin{table}
\caption[]
Evolution characteristic times: look--back time to which
all luminosities were higher by a factor $e$, given
the PWLE parameter in Table 2. Times are in
units of $H_0^{-1}$, the mean is 0.27 and the dispersion is 0.04.
}
\begin{flushleft}
\begin{tabular}{lllllllllllllll}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
$\Lambda $ &$\Omega $& 0.0 & 0.2 & 0.4 & 0.6 & 0.8 &
1.0 \cr
0.0 && .29 & .27 & .26 & .24 & .23 & .23 \cr
0.2 && .30 & .28 & .26 & .24 & .24 & .23 \cr
0.4 && .32 & .29 & .27 & .25 & .24 & .23 \cr
0.6 && .35 & .30 & .28 & .25 & .24 & .24 \cr
0.8 && .39 & .31 & .28 & .26 & .25 & .23 \cr
1.0 && .52(*) & .33 & .29 & .27 & .25 & .24 \cr
\noalign{\smallskip}
\hline
\end{tabular}
\end{flushleft}
\end{table}
\begin{table}
\caption[]
Same as Table 4, but for LEXP. Mean characteristic time:
0.13; dispersion: 0.05.
}
\begin{flushleft}
\begin{tabular}{lllllllllllllll}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
$\Lambda $ &$\Omega $& 0.0 & 0.2 & 0.4 & 0.6 & 0.8
& 1.0
\cr
0.0 && .15 & .12 & .11 & .10 & .09 & .08 \cr
0.2 && .17 & .13 & .12 & .10 & .09 & .09 \cr
0.4 && .19 & .15 & .12 & .11 & .10 & .09 \cr
0.6 && .23 & .16 & .13 & .11 & .10 & .09 \cr
0.8 && .29 & .18 & .14 & .12 & .11 & .10 \cr
1.0 && .53(*) & .22 & .16 & .13 & .11 & .10 \cr
\noalign{\smallskip}
\hline
\end{tabular}
\end{flushleft}
\end{table}
\begin{table}
\caption[]
PWLE characteristic times (Table 4) in units of the age of
the universe in the
corresponding cosmology. Mean: 0.31; dispersion: 0.02. These
ratios are far more independent of cosmology than the times in Table 4.
}
\begin{flushleft}
\begin{tabular}{lllllllllllllll}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
$\Lambda $ &$\Omega $& 0.0 & 0.2 & 0.4 & 0.6 & 0.8 &
1.0
\cr
0.0 && .29 & .32 & .33 & .33 & .34 & .34 \cr
0.2 && .28 & .31 & .33 & .32 & .33 & .34 \cr
0.4 && .28 & .31 & .32 & .32 & .33 & .33 \cr
0.6 && .27 & .30 & .31 & .31 & .32 & .32 \cr
0.8 && .25 & .29 & .30 & .30 & .31 & .31 \cr
1.0 && .13(*)& .28 & .29 & .30 & .30 & .31 \cr
\noalign{\smallskip}
\hline
\end{tabular}
\end{flushleft}
\end{table}
\begin{table}
\caption[]
Same as Table 6, but for LEXP. Mean: 0.15; dispersion:
0.02. As for PWLE, the characteristic time depends
far less on the age of the universe than on the Hubble time.
}
\begin{flushleft}
\begin{tabular}{lllllllllllllll}
\noalign{\smallskip}
\hline
\noalign{\smallskip}
$\Lambda $ &$\Omega $& 0.0 & 0.2 & 0.4 & 0.6 & 0.8 &
1.0
\cr
0.0 && .15 & .15 & .14 & .14 & .13 & .13 \cr
0.2 && .16 & .15 & .14 & .14 & .13 & .13 \cr
0.4 && .16 & .16 & .15 & .14 & .13 & .13 \cr
0.6 && .17 & .16 & .15 & .14 & .13 & .13 \cr
0.8 && .19 & .17 & .15 & .14 & .13 & .13 \cr
1.0 && .13(*)& .18 & .16 & .14 & .14 & .13 \cr
\noalign{\smallskip}
\hline
\end{tabular}
\end{flushleft}
\end{table}
\paragraph{Acknowledgements}
We thank B. Fort, P.Y. Longarreti, G. Soucail, J.P. Picat, R. Pell\'o
and J.F. Leborgne for support and discussions, and especially Jim
Bartlett for a careful reading of the manuscript and enlighting
suggestions.
L.V.W. thanks the french MESR for grant 93135.
This work was supported by grants from the french CNRS (GdR Cosmologie)
and from the
European Community (Human Capital and Mobility ERBCHRXCT920001).
|
\section{Introduction}
The subject of applications of classical Lie symmetries to boundary value
problems is well studied (see the monography \cite{ovs}). In
contrast, the question of involving higher symmetries to the same problem has
received much less attention, unlike say the Cauchy problem. However, one
should stress that nowadays the higher symmetries' approach becomes the basis
of the modern integrability theory \cite{mss}.
A number of attempts to apply the inverse scattering method (ISM) to the
initial boundary value problem has been undertaken. It turned out that if
both initial data and boundary value are chosen arbitrary then the ISM loses
essentially its efficiency. On the other hand side the investigation by
E.Sklyanin \cite{skl} based on the $R-$ matrix approach demonstrated that
there is a kind of boundary conditions, compatible completely with the
integrability. The analytical aspects of such kind problems were studied in
\cite{tar}, \cite{bikt}. After \cite{hab1} it becomes clear that boundary
value problems found can effectively be investigated with the help of the
B\"acklund transformation.
Below we will discuss a higher symmetry test, proposed in \cite{hab2},
\cite{gur} to verify whether the boundary condition given is compatible with
the integrability property of the equation. It is worthwhile to note that
all known classes of boundary conditions, compatible with integrability occur
to pass this symmetry test. Boundary conditions involving explicit time
dependance for the Toda lattice compatible with higher symmetries has
recently been studied in \cite{adl}. It was established there that finite
dimensional systems obtained from the Toda lattice by imposing at both ends
boundary conditions consistent with symmetries were nothing else but
Painlev\'e type equations.
Let us consider the evolution type equation
\begin{equation}
u_{t}=f(u,u_{1},u_{2},...,u_{n}) \label{eq}
\end{equation}
and a boundary condition of the form
\begin{equation}
p(u,u_{1},u_{2},...,u_{k}) \vert_{x=0}=0, \label{gu}
\end{equation}
imposed at the point $x=0$. Here $u_i$ stands for the partial derivative of
the order $i$ with respect to the variable $x.$ Suppose that the equation
given possesses a higher symmetry
\begin{equation}
u_{\tau}=g(u,u_{1},...,u_{m}). \label{sy}
\end{equation}
We call the problem (\ref{eq})-(\ref{gu}) compatible with the symmetry
(\ref{sy}) if for any initial data prescribed at the point $t=0$ a common
solution to the equations (\ref{eq}), (\ref{sy}) exists satisfying the
boundary condition (\ref{gu}). Let us explain more exactly what we mean.
Evidently one can differentiate the constraint (\ref{gu}) only with respect
to the variables $t$ and $\tau,$ (but not respect to $x$). For instance, it
follows from (\ref{gu}) that
\begin{equation} \sum_{i=0}^{n} {\partial p \over
\partial u_i}(u_i)_{\tau}=0, \label{dgu}
\end{equation}
where one should replace $\tau$-derivatives by means of the equation
(\ref{sy}). The boundary value problem (\ref{eq})-(\ref{gu}) be compatible
with the symmetry (\ref{sy}) if the equation (\ref{dgu}) holds identically by
means of the condition (\ref{gu}) and its consequences obtained by
differentiation with respect to $t.$
To formulate an effective criterion of compatibility of the boundary value
problem with a symmetry it's necessary to introduce some new set of dynamical
variables consisting of the vector $v=(u,u_{1},u_{2}, ... u_{n-1})$ and its
$t$-derivatives: $v_{t}$, $v_{tt}$, ... . Passing to this set of variables
allows one really to exclude the dependance on the variable $x.$
In terms of these variables the symmetry (\ref{sy}) and the constraint
(\ref{gu}) take the form
\begin{equation}
v_{\tau}=G\,(v,v_{t} ,
v_{tt} ,... {\partial^{m_1}v\over \partial t^{m_1}}), \label{sy1}
\end{equation}
\begin{equation}
P(v, {\partial v\over
\partial t},...,{\partial^{k_1}v\over \partial t^{k_1}})=0.
\label{gu1}
\end{equation}
The following criterion of compatibility was established in (\cite{gur}).
\th{Theorem.} The boundary value problem (\ref{eq})-(\ref{gu}) is
compatible with the symmetry (\ref{sy}) if and only if the
differential connection (\ref{gu1}) is consistent with the system
(\ref{sy1}).
We call the boundary condition (\ref{gu}) compatible with the integrability
property of the equation (\ref{eq}), if the problem (\ref{eq})-(\ref{gu}) is
compatible with infinite series of linearly independent higher order
symmetries.
The problem of the classification of integrable boundary conditions is solved
completely for the Burgers equation (see \cite{gur})
\begin{equation}
u_{t}=u_{2}+2\,u\,u_{1}, \label{bur}
\end{equation}
\th{Theorem.} If the boundary condition $p(u,u_{1})\vert_{x=0}=0$ is
compatible at least with one higher symmetry of the Burgers equation
(\ref{bur}) then it is compatible with all even order homogeneous symmetries
and is of the form $c_1(u_{1}+u^{2})+c_2\,u+c_3=0.$
In the Burgers case the boundary conditions of the general form (\ref{gu})
can also be described completely with the help of the "recursion operator for
the boundary conditions" $L={\partial \over\partial x}+u,$ which acts on
the set of integrable boundary conditions (see \cite{svi}). For instance, the
boundary condition
$L(c_1(u_{1}+u^{2})+c_2\,u+c_3)=c_1(u_{2}+3uu_{1}+u^{3})+c_2(u_1+u^2)
+c_3\,u=0$ is also integrable.
Let us describe boundary value problems of the form
\begin{equation}
a(u,u_x)\vert_{x=0}=0 \label{sggu}
\end{equation}
\begin{equation}
u_{tt}-u_{xx}+\sin u=0, \label{sg}
\end{equation}
for the sine-Gordon equation compatible with the third order symmetry.
As it is shown in \cite{jsh} the complete algebra of higher symmetries for
the equation (\ref{sg}) i.e. $u_{\xi\eta}=\sin u,$ where $2\xi=x+t,$
$2\eta=x-t$ splits into the direct sum of two algebras consisting of
symmetries of equations
$
u_{\tau}=u_{\xi\xi\xi}+u_{\xi}^3/2,\,
u_{\tau}=u_{\eta\eta\eta}+u_{\eta}^3/2,
$
correspondingly, which are nothing else but potentiated MKdV equation.
Particularly, the following flow commutes with the sine-Gordon equation
\begin{equation}
u_{\tau}=c_1(u_{\xi\xi\xi}+u_{\xi}^3/2)+c_2(u_{\eta\eta\eta}+
u_{\eta}^3/2).
\label{lcs}
\end{equation}
The symmetry (\ref{lcs}) isn't compatible with any boundary condition of the
form (\ref{sggu}) unless $c_1=-c_2,$ under this constraint the equation
(\ref{sggu}) is of one of the forms
\begin{equation}
u=const, \qquad
v=c_1\cos (u/2)+c_2\sin (u/2). \label{sgco}
\end{equation}
Note that the list of boundary conditions (\ref{sgco}) coincides with that
found by A.Zamolodchikov within the framework of the R matrix approach
\cite{zam}. The latter in (\ref{sgco}) in particular cases was studied
earlier in \cite{skl} and \cite{bikt}. The compatibility of the
former in (\ref{sgco}) with the usual version of ISM was declared earlier in
\cite{bikt}. But the statement was based in a mistake (see
\cite{hab3}). Our requirement of consistency is weaker than that is used
in \cite{bikt}. Applications of these and similar problems for
the sine-Gordon equation and the affine Toda lattice in the quantum field
theory are studied in \cite{sal} and \cite{cor}.
According to the theorem above one reduces the problem of finding integrable
boundary conditions to the problem of looking for differential connections
admissible by the following system of equations, equivalent to (\ref{lcs})
with $c_1=-c_2$ and $v=u_x:$
\begin{equation}
\begin{array}{ll}
u_{\tau}=8u_{ttt}+6u_{t}\cos u+3v^2u_t+u_{t}^3,\\
v_{\tau}=8v_{ttt}+6v_{t}\cos u+6u_{tt}vu_t+3v^2v_t+3u_{t}^2v_t.\label{sgt2}
\end{array}
\end{equation}
One can prove that the boundary conditions (\ref{sgco}) are compatible with
rather large subclass of the sine-Gordon equation such that
\begin{equation}
u_{\tau}=\phi
(u,u_1,...u_{k_1})-\phi (u,\bar u_1,...\bar u_{k_1}), \label{sgsym}
\end{equation}
where $u_j=\partial^ju/\partial\xi^j,$ $\bar
u_j=\partial^ju/\partial\eta^j,$ and the equation
$u_{\tau}=\phi_i(u,u_1,...u_{k_i}),$ $i=1,2$ is a symmetry of the equation
$u_{\tau}=u_{\xi\xi\xi}+u_{\xi}^3/2.$
Another well-known integrable equation of hyperbolic type
\begin{equation}
u_{tt}-u_{xx}=\exp (u) +\exp (-2u) \label{ts}
\end{equation}
has applications in geometry of surfaces. For the first time it was found by
Tzitzeica \cite{tzi}. The presence of higher symmetries for this equation
has been established by A.Jiber and A.Shabat \cite{jsh}. The simplest higher
symmetry of this equation is of the fifth order
\begin{equation}
u_{\tau}=u_{\xi\xi\xi\xi\xi}+5(u_{\xi\xi}u_{\xi\xi\xi}-u_{\xi}^2u_{\xi\xi\xi}
-u_{\xi}u_{\xi\xi}^2)+u^5_{\xi}. \label{tssym}
\end{equation}
It is proved in the article cited that the symmetry algebra for (\ref{ts}) is
the direct sum of the symmetry algebras of (\ref{tssym}) and of the equation
obtained from (\ref{tssym}) by replacing $\xi$ by $\eta.$
Let us look for boundary conditions of the form
\begin{equation}
a(u,u_x)=0, \label{tsgu}
\end{equation}
for the equation (\ref{ts}), compatible with the symmetry
\begin{equation}
\begin{array}{lr}
u_{\tau}=u_{\xi\xi\xi\xi\xi}+5(u_{\xi\xi}u_{\xi\xi\xi}-u_{\xi}^2u_{\xi\xi\xi}
-u_{\xi}u_{\xi\xi}^2)+u^5_{\xi}-u_{\eta\eta\eta\eta\eta}- \\
5(u_{\eta\eta}u_{\eta\eta\eta}-u_{\eta}^2u_{\eta\eta\eta}
-u_{\eta}u_{\eta\eta}^2)-u^5_{\eta}. \label{tst1}
\end{array}
\end{equation}
Rather simple but tediously long computations lead to the following
statement.
\th{Theorem.} Boundary conditions (\ref{tsgu}) for the Jiber-Shabat equation
compatible with the symmetry (\ref{tst1}) (and then compatible with
integrability) are either of the form $u_x+c\exp (-u)\vert_{x=0}=0$ or
$u_x+c\exp (u/2)\pm \exp (-u) \vert_{x=0}=0 ,$ where $c$ is arbitrary.
Notice that all equations above are invariant under the reflection type
symmetry $x\rightarrow-x.$ It is unexpected that equations which don't admit
any reflection symmetry admit nevertheless boundary conditions compatible
with integrability. For instance, the famous KdV equation
\begin{equation}
u_t=u_{xxx}+6u_xu
\end{equation}
is consistent with the boundary condition
\begin{equation}
u=0\vert_{x=0},\, u_{xx}\vert_{x=0}=0. \label{kdvcon}
\end{equation}
It implies immediately that the boundary value problem
$$
u_t=u_{xxx}+6u_xu,\qquad
u=0\vert_{x=0}
$$
with the Dirichlet type condition at the axis $x=0$ admits
an infinite dimensional set of "explicit" finite-gap solutions.
This work was partially
supported by Russian Foundation of Fundamental Researches (grant
93-011-165) and International Scientific Foundation (grant
RK-2000).
|
\section{Introduction}
\label{sec:intro}
The formation of a black hole is, in principle, one of the most
efficient mechanisms for generation of gravitational waves. Such
sources tie together two major research initiatives. Laser
interferometric gravity wave detectors\cite{ligo} hold out a promise
of the detection of gravitational waves from astrophysical events. To
interpret the results of the gravitational wave signals, and to help
find signals in the detector noise, a broad and detailed knowledge
will be needed of astrophysical gravitational waveforms. This is one
of the underlying motivations for the ``grand
challenge''\cite{grandchallenge} in high performance computing, aimed
at computing the coalescence of black hole binaries.
Evolving numerical spacetimes and extracting outgoing radiation
waveforms is indeed a challenge. In a straightforward numerical
approach, a good estimate of the asymptotic waveform requires long
numerical evolutions so that the emitted waves can be propagated far
from the source. The necessary long evolutions are difficult for a
number of reasons. General difficulties include throat stretching when
black holes form, numerical instabilities associated with curvilinear
coordinate systems, and the effects of outer boundary
conditions which are approximate.\cite{ast95}
We suggest here that at least part of the cure for this problem may
lie in the use of the theory and techniques of nonspherical
perturbations of the Schwarzschild spacetime (``NPS''). By this we
mean the techniques for treating spacetimes as deviations, first order
in some smallness parameter, from the Schwarzschild spacetime. These
techniques differ from ``linearized theory'' which treats
perturbations of the spacetime from Minkowski spacetime and which
cannot describe black holes. The basic ideas and methods were set
down by many authors and lead to ``wave equations'' for the even
parity\cite{zerilli} and odd parity\cite{rw} perturbations.
NPS has been used to compute outgoing radiation waveforms from a wide
variety of black hole processes, including the scattering of
waves\cite{waves}, particles falling into a hole\cite{particle}, and
stellar collapse to form a hole\cite{stars}. The general scheme of
NPS also underlies the techniques for extraction of radiation from
numerically evolved spacetimes\cite{ae90}. NPS computations have
recently been used in conjunction with fully numerical evolution, as a
code test\cite{abhss} and as a strong-field radiation extraction
procedure\cite{ast95}.
Here we are interested in another sort of application of NPS
theory. To understand such applications we consider an example: Two
very relativistic neutron stars falling into each other, coalescing
and forming a horizon, as depicted in Fig.~1.
The curve ``hypersurface," in Fig.~1, indicates a spacelike ``initial"
surface. The spacetime can be divided into three regions by this
initial surface and the horizon. The early evolution, in region I,
below the initial hypersurface, is highly dynamical and
nonspherical. Spherical perturbation theory is clearly
inapplicable. Above the initial surface the spacetime remains highly
nonspherical in region II inside the event horizon, but outside the
event horizon, in region III, it may be justified to consider the
spacetime to be a perturbation of a Schwarzschild spacetime. This is
essentially guaranteed if the initial hypersurface is chosen late
enough, in some sense, after the formation of the horizon. The
evolution in region III, then, is determined by cauchy data on the
initial hypersurface exterior to the horizon. It is important to note
that this is made possible by the fact that the horizon is a causal
boundary which shields the outer region from the dynamics of the
highly nonspherical central region.
The scheme inherent in this division of spacetime has the potential
greatly to increase the efficiency of the computation of the radiation
generated when strong field sources form black holes. If one starts
from the cauchy data on the initial hypersurface, one can evolve
forward in time with the linear equations of perturbation theory.
Many of the long-time evolution problems of numerical relativity are
avoided and the interpretation of the computed fields in terms of
radiation is immediate.
The approach suggested would then seem to be: Use numerical relativity
up to the initial hypersurface; use the techniques of nonspherical
perturbations in the future of the initial hypersurface. In fact, the
efficiency that can be achieved may be even greater. In the early,
highly nonspherical, pre-initial hypersurface phase of the development
of the spacetime, there may be relatively little generation of
gravitational radiation. By using a computational technique which
suppresses the radiative degrees of freedom one may be able to compute
the early stages of evolution relatively easily. There are two very
recent examples of just such applications of this viewpoint. Price and
Pullin\cite{price_pullin94} used as initial data the
Misner's\cite{misner} solution to the initial value equations for two
momentarily stationary black holes. Abrahams and
Cook\cite{abrahams_cook94} considered two holes moving towards each
other, and used numerical values of the initial value equations. In
neither case was there {\em any} use of fully nonlinear numerical
evolution. The rather remarkable success of both computations suggests
that there is something robust about the underlying idea of separating
horizon-forming astrophysical scenarios into an early phase with no
radiation and a late phase with small deviations from sphericity
outside the horizon. It is plausible that the bulk of the radiation
in most processes is generated only in the very strong-field
interactions around the time of horizon formation and that radiation
generation in the early dynamics can be ignored. One would, however,
think that strong radiation would be emitted during the stages at
which the early horizon is very nonspherical and at which time
nonspherical perturbation theory would seem to be inapplicable. There
should be a tendency for this ``early'' radiation, produced very close
to the horizon, to go inward into the developing black hole, so that
the application of nonspherical perturbation theory to the exterior
really requires that on the initial spacetime the perturbation are
small only well outside the horizon. It would seem that something of
this sort would have to be happening to explain the accuracy of the
Price-Pullin and Abrahams-Cook results.
Whether or not many problems can be treated with {\em no} use of fully
numerical evolution, it appears clear to us that these perturbation
methods will be applied to a variety of problems in which data on the
initial hypersurface is available numerically. The primary purpose of
this paper is to provide justification and background for earlier work
on this subject and a clear recipe for future applications. In the
next section we discuss the meaning, and limitations, of extracting a
``perturbation'' from this numerical data and computing radiated
energies. The explicit process of extracting the perturbations from
the numerical data is given in Sec.~III. In Sec.~IV we demonstrate
the use of this procedure via application to a specific example, the
Misner initial data.
\section{Initial data as Schwarzschild perturbations}
We outline here the formalism for perturbation theory based on work by
Regge and Wheeler\cite{rw} and by Zerilli\cite{zerilli}, but we will
draw heavily on the gauge invariant reformulation of those earlier
works by Moncrief\cite{moncrief74}. Our starting point is an initial
hypersurface which can be taken as a surface of constant Schwarzschild
time. We assume that the coordinates ${x^i}$ on that surface are
almost Schwarzschild coordinates $r,\theta,\phi$ and we assume that
the values are known, on this hypersurface and in these coordinates,
for the 3-metric $\gamma_{ij}$ and the extrinsic curvature
$K_{ij}$. The conditions for finding such a hypersurface and such
coordinates will be made explicit in Sec.~III.
Underlying perturbation theory is the idea of a family of metric
functions $g_{\mu\nu}(x^\alpha;\epsilon)$, depending on the parameter
$\epsilon$, which satisfy the Einstein equations for all $\epsilon$,
and which, in the limit $\epsilon\rightarrow0$, become the
Schwarzschild metric functions, such as $g_{rr}=S^{-1}$. (Here
$S\equiv 1-2M/r$ and $M$ is the mass of the Schwarzschild spacetime;
we use units throughout in which $c=G=1$.) NPS theory amounts to the
approximation
\begin{equation}\label{pertbasic}
g_{\mu\nu}(x^\alpha;\epsilon)\approx
g_{\mu\nu}(x^\alpha;\epsilon)|_{\epsilon=0}+
\epsilon\frac{\partial}{\partial\epsilon}
g_{\mu\nu}(x^\alpha;\epsilon)|_{\epsilon=0}\ .
\end{equation}
\subsection{Choice of expansion parameter}
It is of some practical importance to realize that the choice of the
expansion parameter can have a considerable effect on the range over
which perturbation theory gives a good approximation. Let us imagine
that we introduce a new parameter $\epsilon'$ which is a function of
$\epsilon$ such that $ d\epsilon'/d\epsilon$ approaches unity as
$\epsilon\rightarrow0$. If we take $\epsilon'$ to be the basis of our
perturbation approach, the approximation becomes
\begin{eqnarray}
g(x^\alpha;\epsilon)&=g(x^\alpha;\epsilon(\epsilon')) =
g_{\mu\nu}(x^\alpha;\epsilon(\epsilon'))|_{\epsilon'=0}+
\epsilon'\frac{\partial}{\partial\epsilon'}
g_{\mu\nu}(x^\alpha;\epsilon(\epsilon'))|_{\epsilon'=0}+{\cal
O}(\epsilon'^2)\nonumber\\ &=
g_{\mu\nu}(x^\alpha;\epsilon(\epsilon'))|_{\epsilon=0}+
\left[\epsilon\frac{\partial}{\partial\epsilon'}
g_{\mu\nu}(x^\alpha;\epsilon(\epsilon'))|_{\epsilon=0}\right]
\left\{\frac{\epsilon'}{\epsilon}\right\}+{\cal O}(\epsilon'^2)\
.\label{prime}
\end{eqnarray}
At $\epsilon=0$ the derivative of $g_{\mu\nu}$ with respect to
$\epsilon$ and with respect to $\epsilon'$ have the same values, so
for a given spacetime --- that is, for a given value of $\epsilon$ ---
the nonspherical perturbation in (\ref{prime}) differs from that in
(\ref{pertbasic}) by the factor
$\left\{\epsilon'/\epsilon\right\}$. Computed energies (which are
quadratic in the nonspherical perturbations) will differ by the square
of this ratio. Different choices of parameterization will change this
factor and affect the accuracy of the linearized approximation.
To show the effects of this parameterization dependence, we take as an
example Misner data \cite{price_pullin94}\cite{misner} for two holes.
The initial separation of the holes, in units of the mass of the
spacetime, is described by Misner's parameter $\mu_0$. The metric
perturbations, however, are not analytic in $\mu_0$ as
$\mu_0\rightarrow0$, so $\mu_0$ cannot be used as the expansion
parameter in (\ref{pertbasic}). The actual expansion parameter used
by Price and Pullin, was a function of $\mu_0$ denoted $\kappa_2$. We
consider here what would be the results of perturbation theory done
with the expansion parameter
\begin{equation}\label{eq.diffk}
\epsilon=\frac{\kappa_2}{1-k\kappa_2}\ .
\end{equation}
Figure 2 shows the results, along with the energies computed by
numerical relativity applied to full nonlinear evolution
\cite{anninos_etal93}. For all choices of $k$ the agreement between
perturbation theory and numerical relativity is good at sufficiently
small initial separation (sufficiently small $\mu_0$), but as $\mu_0$
grows larger, the agreement increasingly depends on the which
parameterization is used. The $k=0$ parameterization, the parameter
of the Price-Pullin paper, is a reasonably good approximation even up
to separations ($\mu_0>1.36$) for which the initial apparent horizon
consists of two disjoint parts. For positive values of $k$ the
agreement is less impressive, while for $k=-4$, it appears that
perturbation theory is giving excellent answers for initial data that
are very nonspherical. Clearly the $k=-4$ parameterization is
``better,'' at least for the purpose of computing radiated
energy. There exist yet better choices; in principle a
parameterization could be found for which the energy computed by
linearized theory is perfect for any initial separation. The crucial
point is that we have no {\em a priori} way of choosing what is and
what is not a good parameterization. The choice of expansion parameter
$\kappa_2$ was made in the Price-Pullin analysis, because it occurred
naturally in the mathematical expressions for the initial
geometry. There was no {\em a priori} reason for believing it to be a
particularly good, or particularly bad parameterization. This point
will be discussed again, in connection with numerical results
presented in Sec.~IV.
The fact, demonstrated in Fig.~2, that the results of linear
perturbation theory are arbitrary may seem to suggest that
perturbation answers, from a formal expansion or numerical initial
data, are of little value. It should be realized, however, that the
arbitrariness exhibited in Fig.~2 is simply a demonstration of the
fact that linearized perturbation results are uncertain to second
order in the expansion parameter. The fact that the results for
different parameterizations start to differ from each other around
$\mu_0\approx1.5$ simply signals that $\kappa_2$ is around unity. (In
fact, $\kappa_2\approx0.24$ for $\mu_0=1.2$.) Higher order uncertainty
is an unavoidable feature in the range where the expansion parameter
is of order unity. But there is a potential misunderstanding about the
meaning of ``expansion parameter around unity.'' To see this consider
a change to a new expansion parameter $\epsilon=10^{-4}*\kappa_2$. The
new expansion parameter $\epsilon$ is of order unity for
$\mu_0\approx7$, yet we know that perturbation fails dramatically for
such a large value of $\mu_0$. The issue here is that we need some way
of ascribing an appropriate ``normalization'' to the expansion
parameter. A sign that the normalization is good is that
physically-based measures of distortion start getting large for
$\epsilon$ around unity. If we had reliable measures of this type
then we could have some confidence about the range of the the
expansion parameter for which we could neglect second order
uncertainty, whether due to parameter arbitrariness or the omission of
higher order terms in the calculation. One can formulate interesting
measures for the normalization of the expansion parameter, such as the
extent to which the linearized initial conditions violates the exact
Hamiltonian constraint \cite{suen}. Most such measures are useful only
for finding a very rough normalization for $\kappa_2$ (equivalently,
for roughly finding the range in which linearized perturbation theory
is reliable). The only reliable procedure for this is to carry out
computations of radiated waveforms and energy to second order in the
expansion parameter. The ratio of second order corrections to first
order results gives the only direct measure of the reliability of
perturbation results. If one computes an energy for which the second
order correction to the first order result is 10\%, then one knows
that the third order correction (due to a change in parameterization
or an inclusion of third order terms in the computation) will be on
the order of 1\%.
\subsection{Treating nonlinear initial data as a perturbation expansion}
We turn now to the central question of this paper: How does one apply
perturbation theory to numerically generated initial data? To do this
we consider our numerical initial data to be initial data for a
solution in a parameterized family $g_{\mu\nu}(x^\alpha;\epsilon)$
corresponding to $\epsilon=\epsilon_{\rm num}$. The application of
perturbation theory is equivalent to replacing
$g_{\mu\nu}(x^\alpha;\epsilon_{\rm num})$ by
\begin{equation}\label{eq.linnum}
g_{\mu\nu}(x^\alpha;\epsilon)|_{\epsilon=0}+
\frac{\partial}{\partial\epsilon}
g_{\mu\nu}(x^\alpha;\epsilon)|_{\epsilon=0}\epsilon_{\rm num}.
\end{equation}
An added familiar complication is that we can introduce a family of
coordinate transformations $x^\alpha=x^\alpha(x^{\mu'};\epsilon)$
which reduces to $x^\alpha =x^{\alpha'}$ for
$\epsilon\rightarrow0$. Such a transformation takes the original
family to a new family $g^{'}_{\mu\nu}(x^{\alpha'};\epsilon)$, which
satisfies the same requirements as the original family. We follow
Moncrief\cite{moncrief74} in constructing, from the 3-metric
$\gamma_{ij}$ on constant-$t$ surfaces, quantities $q_i$, which are
invariant to first order in $\epsilon$ (``gauge invariant''), for
coordinate transformations. The construction of these Moncrief $q_i$
is done in two steps. First, the multipole moments of the metric are
extracted. In practice this is done by multiplying the metric
functions by certain angular factors and integrating over
angles. Since we are only interested in quadrupole and higher order
for radiation, this step also eliminates the spherically symmetric
background parts of the metric function. The second step is to form
linear combinations of these multipoles and of their derivatives with
respect to radius. We symbolically represent the process of forming
these quantities as
\begin{equation}\label{monconstr}
q_i=Q_i(\gamma_{ij}, \partial_r \gamma_{ij})\ .
\end{equation}
Here the symbol ``$Q_i$'' represents the process of multiplying by
angular functions and integrating, then multiplying by certain
functions of $r$ and taking linear combinations of the results. (Our
notation here disagrees with that of Moncrief\cite{moncrief74} in a
potentially confusing way. Moncrief's perturbation quantities are
independent of the size of $\epsilon$. In order to have definitions
that can be applied to numerical data we use quantities that -- to
first order -- are proportional to $\epsilon$.)
The Moncrief gauge invariants play two different roles. For even
parity one of the gauge invariants, $q_2$, is a constraint; it
vanishes in linearized theory as a result of the initial value
equations. In linearized theory, the remaining Moncrief quantities,
denoted $q_1$ here, satisfy wave equations $L(q_1)=0$, the
Regge-Wheeler equation in odd parity and Zerilli equation in even
parity.
{}From our numerical data we construct the quantities $q_i$ precisely
according to (\ref{monconstr}). Our numerically constructed
``perturbation'' quantities will not be invariant under coordinate
transformations, but rather will transform as $q'_i=q_i+{\cal
O}(\epsilon_{\rm num}^2)$. Similarly, the linearized constraint, $q_2$
will not vanish, but will be of order $\epsilon_{\rm num}^2$. The
numerically constructed wavefunctions $q_1$ will satisfy $L(q_1)={\cal
O}((\epsilon_{\rm num})^2)$, where $L$ is the Regge-Wheeler or Zerilli
wave operators.
The use of NPS methods is equivalent to ignoring the second order
terms in the wave equations. The wavefunction $q_i$ can then be
propagated from the initial hypersurface forward and the radiation
waveforms extracted from it. To evolve $q_1$ off the initial
hypersurface, however, requires the initial time derivative $\partial
q_1/\partial t$. This can be computed from the initial extrinsic
curvature, but some care is needed. Indeed, the possible ambiguities
that arise here are the justification for the somewhat protracted
discussion in this section.
If ${\bf n}$ is the future-directed unit normal to the initial hypersurface
then the rate at which the 3-metric is changing is given by
\begin{equation}\label{K=Lng}
K_{ij}=-\frac{1}{2}{\cal L}_{\bf n}\,\gamma_{ij}\ ,
\end{equation}
where $K_{ij}$ is the extrinsic curvature and ${\cal L}_{\bf n}$ is
the Lie derivative along the unit normal. The unit normal is related
to the derivative with respect to Schwarzschild time by
$\partial/\partial t=S^{1/2}\ {\bf n}$. The time derivative of the
Moncrief function then can be written
\begin{displaymath}
\partial q_1/\partial t=S^{1/2} {\cal L}_{\bf n}q_1
\end{displaymath}\begin{equation}
=S^{1/2}{\cal L}_{\bf n}Q_1(\gamma_{ij},
\partial\gamma_{ij}/\partial r)\ .
\end{equation}
To evaluate the right hand side we need to know how $Q_1$ changes when
it is Lie dragged by ${\bf n}$. Since $Q_1$ depends only on
$\gamma_{ij}$ it might appear that one need only Lie drag
$\gamma_{ij}$ to find the change in $Q_1$, and that ${\cal L}_{\bf
n}Q_1=Q_1({\cal L}_{\bf n}\gamma_{ij},\partial{\cal L}_{\bf
n}\gamma_{ij}/\partial r)$. From this it would follow that $\partial
q_1/\partial t=-2S^{1/2} Q_1(K_{ij}, \partial K_{ij}/\partial r)$. It
is important to note that this is {\em not} the correct relationship
between $K_{ij}$ and the cauchy data for the wave equation. The
fallacy in this procedure lies in the fact that $q_1$ must be computed
from the 3-metric on a slice for which Schwarzschild time is constant
(to first order in $\epsilon_{\rm num}$). Lie dragging by ${\bf n}$
moves the 3-metric to a surface that is not (to first order) a
constant time surface. The cure is clearly to compare quantities on
surfaces of constant $t$ by using ${\cal L}_t\equiv S^{1/2}{\cal
L}_{\bf n}$. It is the Schwarzschild time derivative that commutes
with the Schwarzschild radial derivative ${\cal L} _t (\partial
/\partial r)^a=0$. The correct prescription then follows from
\begin{displaymath}
\partial q_1/\partial t=S^{1/2}{\cal L}_{\bf n}q_1
\end{displaymath}\begin{displaymath}
=Q_1(S^{1/2}{\cal L}_{\bf n}\gamma_{ij},
\partial (S^{1/2}{\cal L}_{\bf n}\gamma_{ij})/\partial r)
\end{displaymath}
\begin{equation}\label{qdot}
= -2 Q_1(S^{1/2} K_{ij},
\partial (S^{1/2}K_{ij})/\partial r)\ .
\end{equation}
We note that the perturbed Schwarzschild metric does have a shift
vector $\beta_i$ of order $\epsilon$, and in principle the shift
vector influences the time development of $\gamma_{ij}$ according to
$\partial_{t} \gamma_{ij} = \partial_{t'} \gamma_{ij} + 2
\nabla_{(i}\beta_{j)}$, where $t'$ is a time coordinate in which the
shift vector vanishes. But the shift vector can be considered to be
``pure gauge.'' It is necessary if one wants a complete specification
of the coordinates and the metric components, but its value is a
matter of choice, and is not necessary for a complete specification of
the physics. The initial value, and evolution, of the gauge invariant
quantity $q_1$ is invariant with respect to the choice of $\beta_i$,
and $q_1$ carries all the (physically meaningful) information about
gravitational waves.
The evaluation of $q_1$ from (\ref{monconstr}) and $\partial
q_1/\partial t$ from (\ref{qdot}) completes the extraction, from the
numerical data for $\gamma_{ij}, K_{ij}$ of the cauchy data for the
Regge-Wheeler or Zerilli wave equation. An alternative procedure
arises if one uses the scalar wave-equations derived from the
perturbative reduction of the nonlinear wave-equation for the
extrinsic curvature which arises in a new explicitly hyperbolic form
of the Einstein equations\cite{aacby95}. In this system, the scalar
wave equations are one order lower in time derivative from the usual
Regge-Wheeler and Zerilli equations, so the Cauchy data consists of
the extrinsic curvature and its time-derivative (which involves the
3-dimensional Ricci curvature).
{}From the above it is clear that linearized evolution should give good
accuracy when applied to numerically generated initial data with
sufficiently small deviations from sphericity. For initial data which
are known in analytic form one can, of course, apply linearized theory
even to cases in which initial deviations from sphericity are only
marginally small. The results in Fig.~2, for example, show that the
results of such application of perturbation theory give reasonable
accuracy for values of $\mu_0$ at which an initial horizon is highly
distorted. It is worrisome to apply linearized evolution to marginally
nonspherical initial data, which do not, for example, satisfy the
constraint $q_2=0$ with reasonable accuracy. Such a procedure ---
linear evolution of nonlinear initial data --- has, among other
disadvantages, no clear theoretical framework.
\subsection{Calculating radiated energy by ``forced linearization''}
We wish to point out here that NPS methods can be used more broadly,
and a procedure we call ``forced linearization'' can be applied to
numerically generated initial data in a way that amounts to extracting
the linearized part of the data and evolving linearly. This procedure
circumvents the difficulty of performing formal linearization to data
which is only known numerically. We imagine that we start with an
initial value problem in which there is some adjustable parameter,
call it $\mu$, such that $\mu=0$ corresponds to the Schwarzschild
initial data. There is no requirement that the family of solutions
$g_{\mu\nu}(x^\alpha;\mu)$ be analytic in $\mu$ at $\mu\rightarrow0$.
There may be additional parameters, call them $p_i$, such as the
parameters governing the initial momenta of holes. To apply forced
linearization we fix the values of the $p_i$ and make a choice of
$\mu$ such that the computed initial data $\gamma^{\rm
vns}_{ij},K^{\rm vns}_{ij}$ are ``very nearly spherical.'' One
criterion for this would be that $q_2$ is very small. We then
interpret this initial data as being essentially linearized data, to
which the approximation in (\ref{eq.linnum}) applies. We extract
multipoles, form a gauge invariant wave function $q_1$, and evolve it
with the Zerilli or Regge-Wheeler equation, all as described
above. The result of this will be a late-time waveform $q_1^{\rm
vns}(r,t)$ and the energy $E^{\rm vns}$ that it carries. The next
step is to characterize the results with a well behaved gauge
invariant parameter. To do this we choose some fiducial radius $r_{\rm
fid}$, and evaluate $\epsilon^{\rm vns}\equiv q_1(r_{\rm fid},t=0)$
the gauge invariant wave function of the initial hypersurface at this
radius.
Next, we leave the $p_i$ unchanged, but choose a larger value of $\mu$
for which the numerically generated initial data set $\gamma^{\rm
mrgnl}_{ij},K^{\rm mrgnl}_{ij}$ is ``marginal'' in that it corresponds
to deviations from sphericity large enough so that it differs
significantly form linearized initial conditions; one sign of this
would be that the condition $q_2=0$ is significantly violated. For
this data set we go through the same procedure as above in
characterizing the data set by a parameter $\epsilon^{\rm mrgnl}\equiv
q_1(r_{\rm fid},t=0)$. For this marginally spherical initial data we
take the solution for the wavefunction and energy to be
\begin{equation}
q_1^{\rm mrgnl}(r,t)=\left( \frac{\epsilon^{\rm mrgnl}}{\epsilon^{\rm
vns}}\right)q_1^{\rm vns}(r,t)
\ \ \ \ \
E^{\rm mrgnl}=\left( \frac{\epsilon^{\rm mrgnl}}{\epsilon^{\rm
vns}}\right)^2E^{\rm vns} \ .
\end{equation}
The idea underlying this method is that the very nearly spherical data
give us the solution for for $\partial
g_{\mu\nu}/\partial\epsilon|_{\epsilon=0}$. For the marginal initial
data set we then need only multiply this initial data by the
appropriate factor telling us how much larger is the linear part of
the nonsphericity than that of the very nearly spherical initial
data. The success of forced linearization requires then that
$\epsilon$ evaluated at $r_{\rm fid}$ be a well behaved
parameterization of the linearized part of the nonsphericity in the
numerical data. Since our expansion parameter $\epsilon$ is the
magnitude of the perturbation, it will be a good expansion parameter
as long as it is evaluated in a region where the nonlinear deviations
from sphericity are small, i.e., where (\ref{eq.linnum}) is a good
approximation. For this reason it is important that $r_{\rm fid}$ be
chosen fairly large. For processes of the type pictured in Fig.~1, the
deviations from sphericity fall off quickly in radius, so that at
large enough $r$ one can be certain that the initial data are an
excellent approximation to linearized data. Evidence for this is that
the violations of the $q_2=0$ constraint are always confined to small
radii. One easily implemented check on the forced linearization
procedure is to look at the factor $\epsilon^{\rm mrgnl}/\epsilon^{\rm
vns}$ and confirm that it is independent of $r$ for $r>r_{\rm
fid}$. In Section III we show that this test is easily passed by a
numerical example, and that the results of forced linearization are
essentially the same as those of formal linearized theory.
\section{Extraction of perturbations from numerical data}
Here we assume that the reader has numerical solutions for the
3-metric on an approximately t=const surface. The first step in
applying NPS to numerical results is to transform to coordinates which
are ``almost Schwarzschild'' coordinates. It is assumed that the
numerical $\gamma_{ij}$ and $K_{ij}$ are expressed in a coordinate
system $R,\theta,\phi$ in which the approximate spherical symmetry is
manifest. This means that $K_{ij}$ and ratios like
\begin{equation}
\frac{
\gamma_{R\theta}
}
{\sqrt{\gamma_{\theta\theta}}}
\ \ \ \ \ \
\frac{
\gamma_{R\phi}
}
{\sqrt{\gamma_{\theta\theta}}}
\ \ \ \ \
\frac{
\gamma_{\theta\phi}
}
{\gamma_{\theta\theta}}
\end{equation}
must be small. They all are, in fact, formally of order
$\epsilon_{\rm num}$, so if they are not all reasonably
small compared to unity there is little reason to think that
NPS will work. A Schwarzchild-like areal radial coordinate $r$ needs to be
introduced. This can be defined as a function of $R$ by
\begin{equation}
r\equiv\left( \int\,\gamma_{\theta\theta}\,\gamma_{\phi\phi}\,d\Omega\
\right)^{1/4}/4\pi.
\end{equation}
where the integral is taken on a surface of constant $R$. The metric
component $\gamma_{rr}$, in terms of this quantity, gives us another test
of how close the geometry is to that of a constant time Schwarzschild
slice. The quantity
\begin{displaymath}
r\left(1-1/\gamma_{rr}\right)
\end{displaymath}
should be nearly equal to the constant $2M$, where $M$ is the mass of
the spacetime. The variability of this quantity in $r, \theta$, and $\phi$,
is formally of order $\epsilon_{\rm num}$.
There are, of course, other ways of specifying the Schwarzschild-like
coordinates. We could, for example, have defined
$r^2\equiv\gamma_{\theta\theta}$ All these coordinate choices,
however, should agree to order $\epsilon_{\rm num}$ and are therefore
equivalent within a linearized gauge transformation.
To compute the gauge invariant perturbation functions, we first assume
that an $\ell m$ multipole of the 3-metric may be expanded as
\begin{equation}\label{mncrform}
\gamma_{ij}=c_1 (\hat e_1)_{ij} +c_2 (\hat e_2)_{ij}
+h_1 (\hat f_1)_{ij} + {H_2 \over S} (\hat f_2)_{ij}+
r^2 K (\hat f_3)_{ij}+ +r^2 G (\hat f_4)_{ij}
\end{equation}
where, for clarity, we have suppressed multipole indices and have
replaced Moncrief's $h_1$ and $h_2$ odd parity perturbation functions
with $c_1$, $c_2$. The multipole moments $c_1, c_2, h_1, H_2, K,$ and
$G$ are computed by projection onto the relevant spherical harmonics
which can be found in Moncrief\cite{moncrief74}. Explicit formulas
for the important special case of even parity, axisymmetric
perturbations may be found in Ref.~\cite{abhss}.
For odd parity perturbations, one function can be constructed from the
amplitudes $c_1$ and $c_2$ which is gauge invariant
and satisfies the Regge-Wheeler equation (below),
\begin{equation}
Q^{\times}_{\ell m} = \sqrt{2\frac{(\ell+2)!}{(\ell-2)!}}\,
\left[c_1+{1 \over 2} \left ( {\partial c_2 \over \partial r}
- {2 \over r} c_2 \right ) \right ] {S \over r}.
\end{equation}
The situation for even parity perturbations is more
complicated. Two gauge invariant functions may be formed out of
the multipole moments:
\begin{eqnarray}
k_1&=& K + {S \over r} (r^2 \partial_r G- 2 h_1)
\\
k_2 &=&{1 \over 2S} \left[ H_2 - r \partial_r k_{1}
- \left(1-\frac{M}{rS}\right)k_1+ S^{1/2}\partial_r
(r^2 S^{1/2} \partial_r G - 2 S^{1/2} h_1) \right]
\end{eqnarray}
{}From $k_1$ and $k_2$ it is possible to form two new functions,
one which is radiative and one which is
equivalent to the perturbed hamiltonian constraint
\begin{eqnarray}
q_1 &=& 4 r S^2 k_2 + \ell (\ell+1) r k_1
\\
q_2 &=& \partial_r [ 4 r S^2 k_2 + \ell (\ell+1) r k_1 ]
+ \ell(\ell+1) [2 S k_2 + (1-M/\{rS\}) k_1 ].
\end{eqnarray}
The scaled function
\begin{equation}
Q^+_{\ell m}={ q_1 \over \Lambda}\sqrt{2(\ell-1)(\ell+2)
\over \ell(\ell+1)}\ ,
\end{equation}
with
\begin{displaymath}
\Lambda\equiv(\ell-1)(\ell+2)+6M/r\ ,
\end{displaymath}
satisfies the Zerilli equation (below).
The time derivatives of the radiative gauge invariant
functions $Q^{\times}_{\ell m}$ and $Q^+_{\ell m}$
are found by substituting $\sqrt{1-2M/r}K_{ij}$ for $\gamma_{ij}$
in the multipole moment computation and forming the
same combinations of moments.
The wavefunctions $Q^{\times}_{\ell m}$ and $Q^{+}_{\ell m}$
obey the Regge-Wheeler and Zerilli wave equations respectively:
\begin{eqnarray}
{\bf L} Q^{\times}_{\ell m} +V^{\times}_{\ell} Q^{\times}_{\ell m} &=& 0 \\
{\bf L} Q^{+}_{\ell m} +V^{+}_{\ell} Q^{+}_{\ell m} &=& 0
\label{eq:zer}
\end{eqnarray}
where the wave operator appropriate to Schwarzschild
spacetime is
\begin{equation}
{\bf L} = {\partial ^2 \over \partial t^2} - {\partial ^2 \over \partial
r_*^2}
\end{equation}
in terms of the ``tortoise coordinate'' $r_*=r+2M \ln (r/2M-1)$,
and where the potentials are given by
\begin{equation}
V^{\times}_{\ell} = (1-2M/r) \left [ {\ell(\ell +1) \over r^2}
-{6M \over r^3} \right ]
\end{equation}
and,
\begin{equation}
V^{+}_{\ell}(r) = (1-2M/r) \left[{1\over\Lambda^2} \left (
{72M^2\over r^5}-{12M\over r^3}(\ell-1)(\ell+2)(1-3M/r)\right)
+{\ell(\ell-1)(\ell+1)(\ell+2) \over
r^2 \Lambda } \right].
\end{equation}
Once the Zerilli and Regge-Wheeler equations are integrated for
all the desired $\ell$ and $m$ modes, the total radiated energy
can be calculated from the asymptotic timeseries for
$Q_{\ell m}^+$ and $Q_{\ell m}^{\times}$:
\begin{equation}
{d E \over dt} = {1 \over 32 \pi} \sum_{\ell=2}^\infty
\sum_{m=-\ell}^\ell \left( {d Q_{\ell m}^+ \over dt}^2+
{d Q_{\ell m}^{\times} \over dt}^2 \right)
{}.
\label{eq:power}
\end{equation}
\section{Example of perturbation extraction}
In this section we demonstrate the extraction of a
perturbation from a numerical solution to the nonlinear constraint
equations -- the Misner data representing two black holes
at a moment of time symmetry. The Misner 3-geometry may
be written \cite{price_pullin94} as
\begin{equation}
\label{eq:mismet}
dl^2 = \Phi(r, \theta, \mu_0)^4(S^{-1} dr^2 + r^2 d\Omega^2)
{}.
\end{equation}
The conformal factor $\Phi$ is given
by
\begin{equation}
\label{eq:miscon}
\Phi(r, \theta; \mu_0)=
1 + 2 (1+M/2R)^{-1}\sum_{l=2,4,...}^{\infty} \kappa_\ell \left({M \over R}
\right)^{\ell+1} P_\ell(\cos \theta),
\end{equation}
where
\begin{displaymath}
R\equiv(\sqrt{r}+\sqrt{r-2M}\,)^2/4
\end{displaymath}
and
\begin{equation}
\label{eq:miskap}
\kappa_{\ell}\equiv \left({1 \over 4 \sum_{n=1}^\infty (\sinh{n
\mu_0})^{-1}} \right)^{\ell+1}\, \sum_{n=1}^\infty {(\coth{n
\mu_0})^\ell \over \sinh n \mu_0 }\ .
\end{equation}
For this exercise, we pretend that the initial geometry is known only
numerically, so no explicit formal linearization can be done. The odd
parity perturbations vanish in the Misner solution. We compute the
even parity gauge invariant wavefunction for $\ell=2$ using numerical
evaluations of (\ref{eq:miscon}) - (\ref{eq:miskap}). Specifically, we
compute $K$ and $H_2$ of (\ref{mncrform}) from
\begin{equation}
K= H_2 = \int d \Omega \Phi^4 Y_{20}\ .
\end{equation}
All the other moments in (\ref{mncrform}) vanish for the conformally
Schwarzschild metric of (\ref{eq:mismet}). The function $Q_{20}^+$ is
evaluated at values of $r$ corresponding to the range $r_* = -20M$ to
$r_*=50M$. The initial value of $Q_{20}^+$ (along with its
time-derivative which is zero for the Misner time-symmetric initial data)
provides initial values for integration of (\ref{eq:zer}). At large
radius, $r= 100M$, the value of $\partial Q_{20}^+/\partial t$ is used in
(\ref{eq:power}) to compute the radiated energy.
First, in Fig.~\ref{fig.misnl} we show the result of directly
computing the gauge invariant function $Q_{20}^+$ from the nonlinear
initial data, integrating the Zerilli equation, and computing the
radiated energy.
For small values of $\mu_0$ the agreement with the explicitly
linearized data of Ref.~\cite{price_pullin94} is excellent. At about
$\mu_0 \simeq 1.2$ the agreement breaks down and the qualitative
behavior becomes dramatically different. It is interesting to note
that the apparent horizon encompassing both black holes does not exist
for $\mu_0 >1.36$, close to the dramatic reversal in the energy curve.
In Fig.~\ref{fig.miscon} the violation of the linearized constraint by
the nonlinear data is shown as a function of radius. We plot the
ratio of the constrained gauge invariant function, $q_1$ to the
radiative function $q_2$ scaled in such a way as to compensate for
large violation at $r=2M$. The value of $q_2$ clearly grows much
faster than the radiative variable $q_1$ as the separation is
increased.
As discussed in Sec.~II, it is possible to obtain the results of
formal perturbation theory directly from the numerical data without
ever making reference to the analytic solution. In
Fig.~\ref{fig.misfl} we demonstrate the application of the forced
linearization procedure to the nonlinear Misner data for various
values of the fiducial radius $r_{\rm fid}$. For very small values of
$\mu_0$, such as $\mu_0=0.5$, the geometry outside the event horizon
is everywhere well approximated by (\ref{eq.linnum}) and forced
linearization works even for small values of $r_{\rm fid}/M$. When
$\mu_0$ is larger than around $1.5$, on the other hand, the initial
geometry near the horizon contains significant nonlinear effects, and
large values of $r_{\rm fid}/M$ must be used to get results equivalent
to those of formal linearized theory.
As $r_{\rm fid}$ gets large, the results become indistinguishable from
those of formal perturbation theory reported in
Ref.~\cite{price_pullin94}. For $r_{\rm fid}=30M$ the difference in
radiated energy for $\mu_0=3.0$ is less than $10^{-3} \%$. This
high-accuracy equivalence deserves some explanation. In particular,
why is forced linearization equivalent to formal linearization with
expansion parameter $\kappa_2$? Why is that expansion parameter
singled out? The equivalence is a result of two features of the way in
which the linearizations were done: First, both the formal
linearization of Ref.~\cite{price_pullin94}, and the forced
linearization results in Fig.~\ref{fig.misfl}, use precisely the same
coordinates. (The forced linearization results, in fact, are not based
on initial values that were generated by genuinely numerical
means. Rather, the closed form solutions for the Misner metric
functions were used. The ``almost-Schwarzschild'' coordinates of the
forced linearization, were precisely the same as the
``almost-Schwarzschild'' coordinates in
Ref.~\cite{price_pullin94}). Secondly, in the ``almost-Schwarzschild''
coordinate system, the parameter $\kappa_2$ is, to all perturbation
orders, the coefficient of the dominant nonsphericity at large radius.
Forced linearization (in the limit of large $r_{\rm fid}$) results in
a parameterization based on a gauge invariant measure of nonsphericity
at large radius. It therefore must be proportional to $\kappa_2$ and
produce results equivalent to those of the formal linearization of
Ref.~\cite{price_pullin94}, in which $\kappa_2$ was the expansion
parameter.
It should be understood that this does not imply that the parameter
$\kappa_2$ is {\em physically} singled out. A first order change in
the ``almost-Schwarzschild'' coordinates will change the coefficient
of the dominant large-radius nonsphericity. We might, for example,
transform from the ``almost-Schwarzschild'' radial coordinate $r$ of
(\ref{eq:mismet}) to a new coordinate $r' \equiv r [1 + \kappa_{2}
P_2 (\cos \theta)]$. In this case the coefficient of the leading large
$r'$ term in the metric will be $\kappa_2+{\cal O}(\kappa_2^2)$, and
the results of forced linearization with the resulting ``numerical''
data will differ, when perturbations are large, from the results in
Ref.~\cite{price_pullin94}. The forced linearization will have induced
an expansion parameter different from $\kappa_2$.
AMA was supported by National Science Foundation
grant PHY 93-18152/ASC 93-18152 (ARPA supplemented).
RHP was supported by the National Science Foundation under grants
PHY9207225 and PHY9507719.
|
\section{Introduction}
The aim of this paper is to solve, in a unified way, several
mysteries which have arisen over the past few years in connection
with generalizations of the notion of Lie algebra in differential
geometry.
T. Courant \cite{Courant:1990} introduced the following antisymmetric bracket
operation on the sections of $TP\oplus T^*P$ over a manifold $P$:
\[
[X_1+ \xi_1, X_2 +\xi_2]=
[X_1,X_2]+ (L_{X_1}\xi_2-L_{X_2}\xi_1 +
d({\textstyle{\frac 12}}(\xi_1(X_2)-\xi_2(X_1))).
\]
Were it not for the last term, this would be the bracket for the
semidirect product of the Lie algebra ${\cal X}(P)$ of vector fields
with vector space $\Omega^1(P)$ of 1-forms via the Lie derivative
representation of ${\cal X}(P)$ on $\Omega^1(P)$. The last term, which
was essential for Courant's work (about which more will be said
later) causes the Jacobi identity to fail. Nevertheless, for
subbundles $E\subseteq TP\oplus T^*P$ which are maximally
isotropic for the bilinear form
$(X_1 +\xi_1,X_2+ \xi_2)_+=
\frac 12(\xi_1(X_2)+\xi_2(X_1))$, closure of $\Gamma(E)$ under the
Courant bracket implies that the Jacobi identity {\em does}
hold on $\Gamma(E)$, because of the
maximal isotropic
condition on $E$. These subbundles are called {\em Dirac
structures} on $P$; the notion is a simultaneous generalization
of that of Poisson structure (when $E$ is the graph of a map
$\tilde\pi:T^*P\to TP$) and that of closed 2-form (when $E$ is the
graph of a map $\tilde\omega:TP\to T^{*}P$).
\begin{quote}
{\bf Problem 1}. Since the Jacobi identity is satisfied on certain
subspaces where $( \ , \ )_+$ vanishes, find a formula for the
{\em Jacobi anomaly}\footnote{``$+ c.p.$'' below (and henceforth)
will denote ``plus the other two terms
obtained by circular permutations of $(1,2,3).''$}
\[
[[e_1,e_2],e_3] + c.p.
\]
in terms of $( \ , \ )_+$.
\end{quote}
\bigskip
The vector space $\chi(P)\oplus \Omega^1(P)$ on which the Courant bracket
is defined is also a module over $C^\infty(P)$. Projection on the
first factor defines a map $\rho$ from $\chi(P)\oplus
\Omega^1(P)$ to derivations of $C^\infty(P)$. If one checks the
Leibniz identity which enters in the definition of a {\em Lie
algebroid} \cite{Mackenzie:book},
\[
[e_1,fe_2]=f[e_1,e_2]+(\rho(e_1)f)e_2 \,
\]
It turns out that this is not satisfied in general, but that it is
satisfied for Dirac structures. This suggests:
\begin{quote}
{\bf Problem 2}. Express the {\em Leibniz anomaly}
$[e_1,fe_2]-f[e_1,e_2]-(\rho(e_1)f)e_2$ in terms of $( \ , \ )_+$.
\end{quote}
When one is given an inner product on a Lie algebra, it is natural
to ask whether it is invariant under the adjoint representation.
Here again, a calculation turns up an {\em invariance anomaly}.
We solve problems 1 and 2 in the paper, finding an expression for
the invariance anomaly as well. The formulas obtained are so
attractive as to suggest:
\begin{quote}
{\bf Problem 3}. Generalize the Courant bracket by writing down a set
of axioms for a skew-symmetric
bracket $\cal E\times\cal E\to\cal E$, a linear map
$\cal E\to$ $\mbox{Der}(C^\infty(M))$, and a symmetric inner product
$\cal E\times\cal E\to C^\infty(M)$ on the space $\cal E$ of
sections of a vector bundle over $M$, and find other interesting
examples of the structure thus defined.
\end{quote}
\bigskip
Our solution of Problem 3 begins with the definition of a
structure which we call a {\em Courant algebroid}{\footnote{
We apologize to our French colleagues for possible confusion
with the nearly homonymous and somewhat less synonymous term,
``alg\`ebre de courants".}}.
Among the examples of Courant algebroids which we find are the
direct sum of any Lie bialgebroid \cite{MackenzieX:1994} and its dual, with
the bracket given by a symmetrized version of Courant's
original definition. This structure thus gives an answer as well
to:
\begin{quote}
{\bf Problem 4}. What kind of object is the double of a Lie
bialgebroid?
\end{quote}
\noindent Furthermore, within each Courant algebroid, one can consider
the maximal isotropic subbundles closed under bracket. These more
general Dirac structures are new Lie algebroids (and sometimes Lie
bialgebroids). Constructions in this framework applied to the Lie
bialgebroid of a Poisson manifold \cite{MackenzieX:1994} lead to new
ways of building Poisson structures and shed new light on the theory
of Poisson-Nijenhuis structures used to explicate the hamiltonian
theory of completely integrable systems \cite{K-SM:1990}. In
particular, we find a composition law for certain pairs of (possibly
degenerate) Poisson structures which generalizes the addition of
symplectic structures: namely, if $U:T^*P\to TP$ and $V:T^*P\to TP$
define Poisson structures such that $U+V$ is invertible, then
$U(U+V)^{-1}V$ again defines a Poisson structure.
When the base manifold $P$ is a point, a Lie algebroid is just a
Lie algebra. A Courant algebroid over a point turns out to be
nothing but a Lie algebra equipped with a nondegenerate ad-invariant
symmetric 2-form (sometimes called an orthogonal structure
\cite{me-re:algebres}).
(The formulas for the anomalies all involve derivatives, so they
vanish when $P$ is a point.) Such algebras and their maximal
isotropic subalgebras are the ingredients of the theory of Lie
bialgebras and Manin triples \cite{dr:quantum}. In fact, just as a
complementary pair of isotropic subalgebras in a Lie algebra with
orthogonal structure determines a Lie bialgebra, so a
complementary pair of Dirac structures in a Courant algebroid
determines a Lie bialgebroid. It is this fact, which exhibits our
theory as a generalization of the theory of Manin triples, which is
responsible for the application to Poisson-Nijenhuis pairs
mentioned above.
We mentioned earlier that the notion of Dirac
structures was invented in order to treat in the same framework
Poisson structures, which satisfy the equation $[\pi,\pi]=0$, and
closed 2-forms, which satisfy $d\omega=0$. One could look for a more direct
connection between these equations.
\begin{quote}
{\bf Problem 5}. What is the relation between the equations
$[\pi,\pi]=0$ and $d\omega=0$?
\end{quote}
Our solution to Problem 5 is very simple. In a Courant algebroid of
the form $A\oplus A^*$, the double of a Lie bialgebroid, the
equation which a skew-symmetric operator $\tilde I:A\to A^*$
must satisfy in order for its graph to be a Dirac structure
turns out to be the Maurer-Cartan equation $dI+\frac 12[I,I]=0$
for the corresponding bilinear form $I\in\Gamma(\wedge^2 A^*)$.
The structure of the original Courant algebroid
$TM\oplus TM^*$ (also viewed dually as $T^*M\oplus TM$)
is sufficiently degenerate that one of the terms in the
Maurer-Cartan equation drops out in each of the two cases.
The next problem arises from Drinfeld's study \cite{dr:poisson} of
Poisson homogeneous spaces for Poisson Lie groups. He shows in that
paper that the
Poisson manifolds on which a Poisson Lie group $G$ acts transitively
are essentially (that is, if one deals with local rather than global
objects, as did Lie in the old days) in 1-1 correspondence with Dirac
subspaces of the double of the associated Lie bialgebra $({\mathfrak g} ,
{\mathfrak g}^{*})$. It is natural, then, to look for some kind of
homogeneous space associated to a Dirac subbundle in the double of a
Lie bialgebroid.
The object of which a Lie bialgebroid $E\lon P$ is the infinitesimal
limit is a Poisson groupoid, i.e. a Poisson manifold $\Gamma$ carrying
the structure of a groupoid with base $P$, for which the graph of
multiplication $\{(k,g,h)|k=gh\}$ is a coisotropic submanifold of
$\Gamma \times \overline{\Gamma}\times \overline{\Gamma}$.
($\overline{\Gamma }$ is $\Gamma $ with the opposite Poisson
structure. See \cite{MackenzieX:1994} \cite{MackenzieX:1996}
and \cite{we:coisotropic}.)
Unlike in the case of groups, a Poisson groupoid corresponding to a
given Lie bialgebroid may exist only locally.
\begin{quote}
{\bf Problem 6}. Define a notion of Poisson homogeneous space for a
Poisson groupoid. Show that Dirac structures in the double of a Lie
bialgebroid $(A,A^{*})$ correspond to (local) Poisson homogeneous
spaces for the (local) Poisson groupoid $\Gamma $ associated to $(A,A^{*})$.
\end{quote}
Our solution to Problem 6 will be contained in a sequel to this
paper \cite{LWX}.
Even if we work locally, it is somewhat complicated, since the
``homogeneous spaces'' for groupoids, which are already hard to define
in general (see \cite{br-da-ha:topological}), in this case can involve
the quotient spaces of manifolds by arbitrary foliations.
To give a flavor of our results, we mention here one example. For the
standard Lie bialgebroid $(TM,T^{*}M)$, the associated Poisson
groupoid is the pair groupoid $M\times M$ with the zero Poisson
structure. A Dirac structure transverse to $T^{*}M$ is the graph of a
closed 2-form $\omega $ on $M$. The corresponding Poisson homogeneous
space for $M\times M$ is $M\times (M/{\cal F})$, where the factor
$M$ has the zero Poisson structure, and the factor $M/{\cal F}$ is the
(symplectic) Poisson manifold obtained from reduction of $M$ by the
characteristic foliation ${\cal F}$ of $\omega $. (Of course, the
leaf space $M/{\cal F}$ might not be a manifold in any nice sense.)
Dually, our Dirac structure also defines a Poisson homogeneous space for
the Poisson groupoid of the Lie bialgebroid $(T^{*}M,TM)$, which is
$T^{*}M$ with the operation of addition in fibres and the Poisson
structure given by the canonical 2-form. The homogeneous space is
again $T^{*}M$, with the Poisson structure coming from the sum of the
canonical 2-form and the pullback of $\omega $ by the projection
$T^{*}M\lon M$.
\bigskip
We turn now to some problems which remain unsolved.
The only examples of Courant algebroids which we have given are the
doubles of Lie bialgebroids, i.e. those admitting a direct sum
decomposition into Dirac subbundles. For Courant algebroids over a
point, there are many examples which are not of this type, even when
the symmetric form has signature zero, which is necessary
for such a decomposition. For instance, we may take
the direct sum of two Lie algebras of dimension $k$ with invariant bilinear
forms, one positive definite and one negative definite. Any isotropic
subalgebra of dimension $k$ must be the graph of an orthogonal
isomorphism from one algebra to the other. Such an isomorphism may
not exist. Even if it does, it might be the case that the graphs of
any two such isomorphisms must have a line in common. (For instance,
take two copies of ${\mathfrak su}(2)$ and use the fact that every
rotation of ${\mathbb R}^{3}$ has an axis.) These examples and a further
study of Manin triples from the point of view of Lie algebras with
orthogonal structure may be found in \cite{me-re:lie}.
\begin{quote}
{\bf Open Problem 1}. Find interesting examples of Courant
algebroids which are not doubles of Lie bialgebroids, including
examples which admit one Dirac subbundle, but not a pair of transverse
ones. Are there Courant algebroids which are not closely related to
finite dimensional Lie algebras, for which the bilinear form is
positive definite?
\end{quote}
In his study of quantum groups and the Knizhnik-Zamolodchikov
equation, Drinfeld \cite{dr:quasi} introduced quasi-Hopf algebras, in
which the axiom of coassociativity is weakened, and their classical
limits, the Lie quasi-bialgebras. The latter notion was studied in
depth by Kosmann-Schwarzbach \cite{ko:quasi} (see also
\cite{ba-ko:double}), who defined various structures
involving a pair of spaces in duality carrying skew symmetric brackets
whose Jacobi anomalies appear as coboundaries of other objects. Her
structures are not subsumed by ours, though, since our expression for the
Jacobi anomaly is zero when the base manifold is a point. Jacobi
anomalies as coboundaries also appear in the theory of ``strongly
homotopy Lie algebras'' \cite{la-ma:strongly} and in recent work of
Ginzburg \cite{gi:resolution}.
The relation of these studies to Courant
algebroids is the subject of work in progress with Dmitry Roytenberg.
\begin{quote}
{\bf Open Problem 2}. Define an interesting type of structure which
includes both the Courant algebroids and the Lie quasi-bialgebras
as special cases.
\end{quote}
At the very beginning of our study, we found that if the bracket on
a Courant algebroid is modified by the addition of a symmetric
term, many of the anomalies for the resulting asymmetric bracket
become zero. This resembles the ``twisting'' phenomenon of Drinfeld
\cite{dr:quasi}.
\begin{quote}
{\bf Open Problem 3}. What is the geometric meaning of such
asymmetric brackets, satisfying most of the axioms of a Lie
algebroid?
\end{quote}
The next problem is somewhat vague. The Maurer-Cartan equation
$d\alpha+\frac 12[\alpha,\alpha]$ appears as an integrability condition in the
theory of connections and plays an essential role in modern
deformation theory. (See \cite{mi:rational} and various original
sources cited therein.)
\begin{quote}
{\bf Open Problem 4}. Find geometric or deformation-theoretic
interpretations of the Maurer-Cartan equation for Dirac structures.
\end{quote}
Lie algebras, Lie algebroids and (the doubles of) Lie bialgebras
are the infinitesimal objects corresponding to Lie groups, Lie
groupoids, and (the doubles of) Poisson Lie groups respectively.
Moreover, Kosmann-Schwarzbach \cite{ko:quasi} has
studied the global objects corresponding to Lie
quasi-bialgebras, and Bangoura \cite{ba:quasi} has recently identified
the dual objects. Yet the following problem is unsolved, even for
$TM\oplus T^*M$.
\begin{quote}
{\bf Open Problem 5}. What is the global, groupoid-like object
corresponding to a Courant algebroid? In particular, what is the
double of a Poisson groupoid?
\end{quote}
A solution to Open Problem 4 might come from a solution to the next
problem. When one passes from an object such as a Lie bialgebra or
even a Lie quasi-bialgebra to its double, the resulting object is
frequently ``nicer'' in the sense that some of the anomalies possessed
by the original object now vanish.
\begin{quote}
{\bf Open Problem 6}. What is the double of a Courant algebroid?
\end{quote}
Finally, we would like to
remark that many of the constructions in this paper can be carried out
at a more abstract level, either replacing the sections of a vector
bundle $E$ by a more general module over $C^\infty(P)$, as in
\cite{hu:poisson}, or in the context of local functionals on mapping
spaces as in \cite{Dorfman} by Dorfman.
\bigskip
\noindent
{\bf Acknowledgements.} In addition to the funding sources mentioned
in the first footnote, we like to thank several institutions
for their hospitality while work on this project was being done: the
Isaac Newton Institute (Weinstein, Xu); the Nankai Institute for
Mathematics (Liu, Weinstein, Xu); Peking University (Xu). Thanks go
also to Yvette Kosmann-Schwarzbach, Jiang-hua Lu, Kirill Mackenzie
and Jim Stasheff for their helpful comments.
\section{Double of Lie bialgebroids}
\begin{defi}
\label{def:quasi-algebroid}
A {\em Courant algebroid} is a vector bundle $E\lon P$
equipped with a nondegenerate symmetric bilinear form
$( \cdot , \cdot )$ on the bundle, a skew-symmetric
bracket $[\cdot , \cdot ]$ on $\gm (E)$
and a bundle map $\rho :E\lon TP$ such that the following
properties are satisfied:
\begin{enumerate}
\item For any $e_{1}, e_{2}, e_{3}\in \gm (E)$,
$[[e_{1}, e_{2}], e_{3}]+c.p.={\cal D} T(e_{1}, e_{2}, e_{3});$
\item for any $e_{1}, e_{2} \in \gm (E)$,
$\rho [e_{1}, e_{2}]=[\rho e_{1}, \rho e_{2}];$
\item for any $e_{1}, e_{2} \in \gm (E)$ and $f\in C^{\infty} (P)$,
$[e_{1}, fe_{2}]=f[e_{1}, e_{2}]+(\rho (e_{1})f)e_{2}-
(e_{1}, e_{2}){\cal D} f ;$
\item $\rho \mbox{\tiny{$\circ $}} {\cal D} =0$, i.e., for any $f, g\in C^{\infty}(P)$,
$({\cal D} f, {\cal D} g)=0$;
\item for any $e, h_{1}, h_{2} \in \gm (E)$,
$\rho (e) (h_{1}, h_{2})=([e , h_{1}]+{\cal D} (e ,h_{1}) ,
h_{2})+(h_{1}, [e , h_{2}]+{\cal D} (e ,h_{2}) )$,
\end{enumerate}
where $T(e_{1}, e_{2}, e_{3})$ is the function on the base $P$
defined by:
\begin{equation}
\label{eq:T0}
T(e_{1}, e_{2}, e_{3})=\frac{1}{3} ([e_{1}, e_{2} ], e_{3})+c.p.,
\end{equation}
and
${\cal D} : C^{\infty}(P)\lon \gm (E)$
is the map defined\footnote{In this paper,
$d_{0}$ denotes the usual differential from
functions to 1-forms, while $d$ will denote the
differential from functions to sections of the dual of a Lie
algebroid.} by ${\cal D} = \frac{1}{2} \beta^{-1}\rho^{*} d_{0}$,
where $\beta $ is
the isomorphism between $E$
and $E^*$ given by the bilinear form. In other words,
\begin{equation}
\label{eq:D}
({\cal D} f , e)= \frac{1}{2} \rho (e) f .
\end{equation}
\end{defi}
{\bf Remark.} Introduce a twisted bracket (not antisymmetric!) on $\gm (E)$ by
$$[e ,h \tilde{]}=[e ,h]+{\cal D} (e, h). $$
Then (iii) is equivalent to
\begin{equation}
[e_{1}, fe_{2} \tilde{]}=f [e_{1},e_{2}\tilde{]}+ (\rho (e_{1})f)e_{2};
\end{equation}
(v) is equivalent to
\begin{equation}
\rho (e) (h_{1} , h_{2})=([e ,h_{1}\tilde{]}, h_{2})+ (h_{1} ,
[e, h_{2}\tilde{]} );
\end{equation}
and (ii) and (iv) can be combined into a single equation:
\begin{equation}
\rho [e_{1} , e_{2}\tilde{]}=[\rho e_{1}, \rho e_{2}].
\end{equation}
It would be nice to interpret equation (i) in terms of this
twisted bracket. The geometric meaning of this
twisted bracket remains a mystery to us.
\begin{defi}
Let $E$ be a Courant algebroid. A subbundle $L$ of $E$ is
called {\em isotropic} if it is isotropic under the
symmetric bilinear form $( \cdot , \cdot )$. It is called {\em integrable}
if $\gm (L)$ is closed under the bracket $[\cdot , \cdot ]$.
A {\em Dirac structure}, or {\em Dirac subbundle}, is a subbundle $L$
which is maximally isotropic and integrable.
\end{defi}
The following proposition follows immediately from the definition.
\begin{pro}
\label{prop}
Suppose that $L$ is an integrable isotropic subbundle of
a Courant algebroid \\ $(E, \rho , [\cdot , \cdot ], ( \cdot , \cdot
)) $.
Then $(L, \rho |_{L} , [\cdot , \cdot ])$
is a Lie algebroid.
\end{pro}
Suppose now that both $A$ and $A^{*}$ are Lie algebroids over the base
manifold $P$, with anchors $a$ and $a_{*}$ respectively.
Let $E$ denote their vector bundle direct sum:
$E=A\oplus A^{*}$.
On $E$, there exist two natural nondegenerate
bilinear forms, one symmetric and another antisymmetric, which are
defined as follows:
\begin{equation}
\label{eq:pairing}
(X_{1}+\xi_{1} , X_{2}+\xi_{2})_{\pm}=\frac{1}{2} (\langle \xi_{1}, X_{2} \rangle
\pm \langle \xi_{2} , X_{1}\rangle ).
\end{equation}
On $\gm (E)$, we introduce a bracket by
\begin{equation}
\label{eq:double}
[e_{1}, e_{2}]=([X_{1}, X_{2}]+L_{\xi_{1}}X_{2}-L_{\xi_{2}}X_{1}-d_{*}(e_{1},
e_{2})_{-})
+ ([\xi_{1} , \xi_{2}]+L_{X_{1}}\xi_{2}-L_{X_{2}}\xi_{1} +d(e_{1}, e_{ 2})_{-}),
\end{equation}
where $e_{1}=X_{1}+\xi_{1}$ and $e_{2}=X_{2}+\xi_{2}$.\\\\
Finally, we let $\rho : E\lon TP$ be the bundle map defined by
$\rho =a +a_{*}$. That is,
\begin{equation}
\rho (X+\xi )=a(X)+a_{*} (\xi ) , \ \ \forall X\in \gm (A) \mbox{ and }
\xi \in \gm (A^{*})
\end{equation}
It is easy to see that in this case the operator ${\cal D}$ as
defined by Equation (\ref{eq:D}) is given by
$${\cal D}=d_{*}+d, $$
where $d_{*}: C^{\infty}(P)\lon \gm (A)$ and
$d: C^{\infty}(P)\lon \gm ( A^{*})$ are the
usual differential operators associated to Lie algebroids \cite{MackenzieX:1994}.
When $(A, A^{*})$ is a Lie bialgebra $({\mathfrak g} , {\mathfrak g}^{*})$,
the bracket above reduces to the famous Lie
bracket of Manin on the double ${\mathfrak g} \oplus {\mathfrak g}^{*}$.
On the other hand, if $A$ is the tangent bundle Lie algebroid
$TM$ and $A^{*}=T^{*}M $ with zero bracket,
then Equation (\ref{eq:double}) takes the form:
$$[X_{1}+\xi_{1} , X_{2}+\xi_{2}]=[X_{1} ,X_{2}]
+\{ L_{X_{1}}\xi_{2} -L_{X_{2}}\xi_{1} +d(e_{1}, e_{ 2})_{-} \}.$$
This is the bracket first introduced by Courant \cite{Courant:1990}, then
generalized to the context of the formal variational calculus by Dorfman
\cite{Dorfman}.
Our work in this paper is largely motivated by an attempt to unify the two
examples above, based on the observation that Courant's bracket appears to be some
kind of ``double.'' In order to generalize Manin's construction to Lie
algebroids, it is necessary to have a compatibility condition between Lie
algebroid structures on a vector bundle and its dual. Such a condition, providing
a definition of {\em Lie bialgebroid}, was given in
\cite{MackenzieX:1994}. We quote here an equivalent formulation from
\cite{K-S:1994}.
\begin{defi}
A {\em Lie bialgebroid} is a dual pair $(A,A^*)$ of vector bundles equipped with
Lie algebroid structures such that the differential $d_*$ on $\Gamma(\wedge^*A)$
coming from the structure on $A^*$ is a derivation of the Schouten-type bracket
on $\Gamma(\wedge^*A)$ obtained by extension of the structure on $A$.
\end{defi}
The
following two main theorems of this paper show that we have indeed found the
proper version of the theory of Manin triples for the Lie algebroid case.
\begin{thm}
\label{thm:main1}
If $(A, A^{*})$ is a Lie bialgebroid, then
$E=A \oplus A^{*}$ together with $([\cdot , \cdot ], \rho , (\cdot , \cdot)_{+})$
is a Courant algebroid.
\end{thm}
Conversely, we have
\begin{thm}
\label{thm:main2}
In a Courant algebroid $(E, \rho , [\cdot , \cdot ], ( \cdot , \cdot ))$,
suppose that $L_{1}$ and $L_{2}$ are Dirac subbundles transversal
to each other, i.e., $E=L_{1}\oplus L_{2}$.
Then, $(L_{1}, L_{2} )$ is a Lie bialgebroid, where $L_{2}$ is
considered as the dual bundle of $L_{1}$ under the
pairing $2( \cdot , \cdot )$.
\end{thm}
An immediate consequence of the theorems above
is the following duality property of Lie bialgebroids,
which was first proved in \cite{MackenzieX:1994}
and then by Kosmann-Schwarzbach \cite{K-S:1994} using a simpler method.
\begin{cor}
If $(A, A^{*})$ is a Lie bialgebroid, so is $(A^{*}, A)$.
\end{cor}
\section{Jacobi anomaly}
In this section, we begin the computations leading to the proofs of our
main theorems. Throughout this section, we assume that
$A$ is a Lie algebroid with anchor $a$ and that its dual $A^*$ is also equipped
with a Lie algebroid structure with anchor $a_{*}$.
However, we shall {\em not} assume any compatibility conditions
between these two algebroid structures.
For simplicity, for any $e_{i}=X_{i}+\xi_{i}\in \gm (E), i=1,2,3$,
we let
$$J(e_{1}, e_{2}, e_{3})=[[e_{1}, e_{2} ], e_{3}]+c.p.$$
The main theorem of this section is the following
\begin{thm}
\label{thm:main-jacobi}
Assume that both $(A, a)$ and $(A^{*}, a_{*})$
are Lie algebroids. Then, for $e_{i}=X_{i}+\xi_{i}\in \gm (E), i=1,2,3$,
we have
\begin{equation}
J(e_{1}, e_{2}, e_{3})={\cal D} T( e_{1}, e_{2}, e_{3})-(J_{1} +J_{2} +c.p ),
\end{equation}
\end{thm}
where $$J_{1} =i_{X_{3}}(d [\xi_{1} , \xi_{2} ]-L_{\xi_{1}}d \xi_{2}
+L_{\xi_{2}}d \xi_{1} )
+ i_{\xi_{3}} (d_{*} [X_{1}, X_{2}]-L_{X_{1}}d_{*}X_{2}
+L_{X_{2}}d_{*}X_{1} ), $$
and
$$J_{2} = L_{d_{*}(e_{1}, e_{2})_{-}}\xi_{3}
+ [d(e_{1}, e_{2})_{-},\ \xi_{3}]
+L_{d (e_{1}, e_{2})_{-}}X_{3}+[d_{*}(e_{1}, e_{2})_{-},\ X_{3} ]. $$
We need a series of lemmas before proving this theorem.
\begin{lem}
For $e_{i}=X_{i}+\xi_{i}\in \gm (E), i=1,2,3$,
$T$ is skew-symmetric, and
\begin{equation}
\label{eq:T}
T(e_{1}, e_{2}, e_{3})=\frac{1}{2} \{\langle [X_{1}, X_{2}], \xi_{3} \rangle
+\langle [\xi_{1}, \xi_{2}], X_{3}\rangle
+a(X_{3})(e_{1}, e_{2})_{-}-a_{*}(\xi_{3})(e_{1}, e_{2})_{-}\}+c.p.
\end{equation}
\end{lem}
\noindent{\bf Proof.}\ The first assertion is obvious from the definition of $T$.
For the second one, we first have
\be
&&([e_{1}, e_{2}], e_{3})_{+}\\
&=&\frac{1}{2} \{ \langle [X_{1}, X_{2}], \xi_{3} \rangle + \langle L_{\xi_{1}}X_{2}, \xi_{3}\rangle
- \langle L_{\xi_{2}}X_{1} , \xi_{3}\rangle -a_{*}(\xi_{3}) (e_{1}, e_{2})_{-}\\
&&+\langle [\xi_{1}, \xi_{2}], X_{3}\rangle +\langle L_{X_{1}}\xi_{2} ,X_{3}
\rangle -\langle L_{X_{2}}\xi_{1} , X_{3}\rangle
+a(X_{3})(e_{1}, e_{2})_{-}\}\\
&=&\frac{1}{2} \{[\langle [X_{1}, X_{2}], \xi_{3} \rangle+\langle
[\xi_{1}, \xi_{2}], X_{3}\rangle ]+c.p. \}\\
&&+\frac{1}{2} \{a_{*}(\xi_{1})\langle X_{2} ,\xi_{3} \rangle
-a_{*}(\xi_{2})\langle X_{1}, \xi_{3} \rangle
-a_{*}(\xi_{3})(e_{1}, e_{2})_{-}\\
&&\ \ \ \ \ \ \ \ +a(X_{1})\langle \xi_{2} ,X_{3}\rangle -a(X_{2})\langle \xi_{1} ,X_{3}\rangle +a(X_{3})(e_{1}, e_{2})_{-} \}\\
&=&\frac{1}{2} [\{\langle [X_{1}, X_{2}], \xi_{3} \rangle
+\langle [\xi_{1}, \xi_{2}], X_{3}\rangle
+a(X_{3})(e_{1}, e_{2})_{-}-a_{*}(\xi_{3})(e_{1}, e_{2})_{-}\}+c.p.]\\
&&+\frac{1}{2} \rho (e_{1}) (e_{2}, e_{3})_{+}-\frac{1}{2} \rho (e_{2}) (e_{3}, e_{1})_{+}
\ee
Therefore, by taking the sum of its cyclic permutations, one obtains
\be
T(e_{1}, e_{2}, e_{3})&=&\frac{1}{3} ([e_{1}, e_{2}], e_{3})_{+}+c.p.\\
&=& \frac{1}{2} \{\langle [X_{1}, X_{2}], \xi_{3} \rangle
+\langle [\xi_{1}, \xi_{2}], X_{3}\rangle
+a(X_{3})(e_{1}, e_{2})_{-}-a_{*}(\xi_{3})(e_{1}, e_{2})_{-}\}+c.p.
\ee \qed
As a by-product, we obtain the following identity
by substituting Equation (\ref{eq:T}) into the last step of
the computation of $([e_{1}, e_{2}], e_{3})_{+}$ in the
proof above. This formula will be useful
later.
\begin{equation}
\label{eq:T1}
([e_{1}, e_{2}], e_{3})_{+}=T(e_{1}, e_{2}, e_{3})
+\frac{1}{2} \rho (e_{1}) (e_{2}, e_{3})_{+}-\frac{1}{2} \rho (e_{2}) (e_{3}, e_{1})_{+}
\end{equation}
\begin{lem}
\label{lem:4.3}
\begin{equation}
i_{X}L_{\xi}d\eta =[\xi, L_{X}\eta ]-L_{L_{\xi}X} \eta
+[d\langle \eta , X\rangle ,\ \xi ]
+d (a_{*}(\xi)\langle \eta , X\rangle ) -d\langle [\xi , \eta ] ,X\rangle .
\end{equation}
\end{lem}
\noindent{\bf Proof.}\ For any $Y\in \gm (A)$,
\be
\langle i_{X}L_{\xi} d \eta , Y\rangle &=& (L_{\xi }d\eta )(X, Y)\\
&=& a_{*}(\xi )[d\eta (X, Y)]-d\eta (L_{\xi }X, Y)-
d\eta (X, L_{\xi }Y)\\
&=&a_{*}(\xi ) a(X) \langle \eta , Y\rangle -a_{*}(\xi )a(Y)\langle \eta , X\rangle
-a_{*}(\xi )\langle \eta , [X, Y]\rangle \\
&&-a(L_{\xi }X )\langle \eta , Y\rangle +a(Y)\langle \eta , L_{\xi}X\rangle
+\langle \eta , [L_{\xi} X , Y] \rangle \\
&&-a(X) \langle \eta , L_{\xi }Y\rangle
+a(L_{\xi }Y)\langle \eta , X\rangle +\langle \eta , [X, L_{\xi}Y]\rangle\\
&=&a_{*}(\xi )\langle L_{X}\eta , Y\rangle -a_{*}(\xi )a(Y)\langle \eta , X\rangle +a(Y)a_{*}(\xi ) \langle \eta , X\rangle\\
&&-a(Y)\langle [\xi , \eta ], X\rangle -\langle L_{L_{\xi}X}\eta , Y\rangle
-\langle L_{X}\eta , L_{\xi }Y\rangle +\langle L_{\xi}Y, d\langle \eta ,X\rangle \rangle \\
&=&\langle [\xi , L_{X}\eta ], Y\rangle +\langle [d \langle \eta , X\rangle ,\xi ], Y\rangle \\
&&+a(Y)a_{*}(\xi )\langle \eta , X\rangle -a(Y) \langle
[\xi ,\eta ], X\rangle -\langle L_{L_{\xi }X}\eta , Y\rangle.
\ee
The lemma follows immediately. \qed
\begin{lem}
\label{lem:T1}
\begin{equation}
\label{eq:14}
([e_{1}, e_{2}], e_{3})_{-}+c.p.=T(e_{1}, e_{2}, e_{3})+
[ \{a(X_{3})(e_{1}, e_{2})_{-}+2a_{*}(\xi_{3})(e_{1}, e_{2})_{-}
- \langle [\xi_{1}, \xi_{2}], X_{3}\rangle \} +c.p. ]
\end{equation}
\end{lem}
\noindent{\bf Proof.}\ By definition,
$$([e_{1}, e_{2}], e_{3})_{-}+([e_{1}, e_{2}], e_{3})_{+}
=\langle [e_{1}, e_{2}]^{*}, X_{3}\rangle, $$
where $[e_{1}, e_{2}]^{*}$ refers to the component of $[e_{1}, e_{2}]$
in $\gm (A^{*})$.
It thus follows that
\be
&&\{([e_{1}, e_{2}], e_{3})_{-}+ c.p. \}+ 3T(e_{1}, e_{2}, e_{3})\\
&=& \langle [e_{1}, e_{2}]^{*} , X_{3} \rangle +c.p.\\
&=& \langle [\xi_{1}, \xi_{2}]+L_{X_{1}}\xi_{2}-L_{X_{2}}\xi_{1} +d(e_{1}, e_{ 2})_{-},
\ \ X_{3} \rangle +c.p.\\
&=& \{\langle [\xi_{1}, \xi_{2}], X_{3}\rangle + a(X_{1})\langle \xi_{2}, X_{3}\rangle
-\langle \xi_{2} , [X_{1}, X_{3}]\rangle\\
&&+a(X_{3})(e_{1}, e_{2})_{-}-a(X_{2})\langle \xi_{1} , X_{3} \rangle
+\langle \xi_{1} , [X_{2}, X_{3}]\rangle \} +c.p.\\
&=& \{\langle [\xi_{1}, \xi_{2}], X_{3}\rangle
+2 \langle [X_{1}, X_{2}], \xi_{3}\rangle
+3a(X_{3})(e_{1}, e_{2})_{-} \}+c.p.\\
&=&4T(e_{1}, e_{2}, e_{3})+[\{a(X_{3})(e_{1}, e_{2})_{-} +2a_{*}(\xi_{3})(e_{1},
e_{2})_{-}-\langle [\xi_{1}, \xi_{2}], X_{3}\rangle \}+c.p. ],
\ee
where the second from the last step follows essentially from
reorganizing cyclic permutation terms and the last
step uses Equation (\ref{eq:T}).
Equation (\ref{eq:14}) thus follows immediately. \qed
{\bf Proof of Theorem \ref{thm:main-jacobi}} We denote by $I_{1}$ and $I_{2}$ the
components of $J(e_{1}, e_{2}, e_{3})$ on $\gm (A^{*})$ and
$\gm (A)$ respectively.
Thus,
by definition,
\be
I_{1} &=&\{ [[\xi_{1}, \xi_{2}], \xi_{3}]+[L_{X_{1}}\xi_{2}-L_{X_{2}}\xi_{1} ,\xi_{3}]+[d(e_{1}, e_{2})_{-}, \xi_{3}]\\
&&\ \ +L_{ [X_{1}, X_{2}]+L_{\xi_{1}}X_{2}-L_{\xi_{2}}X_{1}-d_{*}(e_{1}, e_{
2})_{-}
}\xi_{3}\\
&&\ \ \ \ +L_{X_{3}}L_{X_{2}}\xi_{1}-L_{X_{3}}L_{X_{1}}\xi_{2}-L_{X_{3}}[\xi_{1}, \xi_{2}]\\
&&\ \ \ \ \ \ -d[ a(X_{3})(e_{1}, e_{ 2})_{-}] +d([e_{1}, e_{2}], e_{3})_{-} \}+ c.p.
\ee
By using the Jacobi identity:
$[[\xi_{1}, \xi_{2}], \xi_{3}]+c.p. =0$ and the relation: $L_{[X_{1}, X_{2}]}=[L_{X_{1}}, L_{X_{2}}]$,
we can write
\begin{eqnarray}
I_{1}&=&\{[L_{X_{1}}\xi_{2}-L_{X_{2}}\xi_{1} ,\xi_{
3}]+ L_{L_{\xi_{1}}X_{2}-L_{\xi_{2}}X_{1}}\xi_{3} \nonumber \\
&&-L_{d_{*}(e_{1}, e_{2})_{-}}\xi_{3}
+ [d(e_{1}, e_{2})_{-},\ \xi_{3}]-L_{X_{3}}[\xi_{1}, \xi_{2}] \nonumber\\
&&+d([e_{1}, e_{2}], e_{3})_{-}-d(a(X_{3})(e_{1}, e_{2})_{-})\}+c.p.
\label{eq:11}
\end{eqnarray}
Now,
\be
L_{X_{3}}[\xi_{1}, \xi_{2}] &=& (di_{X_{3}}+i_{X_{3}}d )[\xi_{1}, \xi_{2}]\\
&=&d\langle X_{3}, [\xi_{1}, \xi_{2}] \rangle +
i_{X_{3}}L_{\xi_{1}}d\xi_{2}-i_{X_{3}}L_{\xi_{2}}d\xi_{1}\\
&&\ \ \ \ \ \ +i_{X_{3}}(d[\xi_{1}, \xi_{2}]-L_{\xi_{1}}d\xi_{2}+L_{\xi_{2}}d\xi_{1}).
\ee
By using Lemma \ref{lem:4.3} twice, we can write
\begin{eqnarray}
&&L_{X_{3}}[\xi_{1}, \xi_{2}]+c.p. \nonumber \\
&=&\{ [L_{X_{1}}\xi_{2}-L_{X_{2}}\xi_{1}, \xi_{3}]+
L_{L_{\xi_{1}}X_{2}-L_{\xi_{2}}X_{1}}\xi_{3}+2[d(e_{1}, e_{2})_{-},\ \xi_{3}]
\nonumber \\
&&+2d( a_{*}(\xi_{3} )(e_{1}, e_{2})_{-}) -d\langle X_{3}, [\xi_{1}, \xi_{2}]
\rangle
+i_{X_{3}}(d[\xi_{1}, \xi_{2}]-L_{\xi_{1}}d\xi_{2}+L_{\xi_{2}}d\xi_{1})\}+c.p.
\label{eq:12}
\end{eqnarray}
Substituting Equation (\ref{eq:12}) into Equation (\ref{eq:11}),
we have
$$I_{1}=\{ d [ ([e_{1}, e_{2}], e_{3})_{-}-a(X_{3})(e_{1}, e_{2})_{-}
-2a_{*}(\xi_{3})(e_{1}, e_{2})_{-} +\langle [\xi_{1}, \xi_{2}], X_{3} \rangle ]
-K_{1} -K_{2} \}+c.p.,
$$
where $$ K_{1} =i_{X_{3}}(d [\xi_{1} , \xi_{2} ]-L_{\xi_{1}}d \xi_{2}
+L_{\xi_{2}}d \xi_{1} )$$
and
$$K_{2} = L_{d_{*}(e_{1}, e_{2})_{-}}\xi_{3}
+ [d(e_{1}, e_{2})_{-},\ \xi_{3}].$$
It follows from Lemma \ref{lem:T1} that
$$I_{1}=dT( e_{1}, e_{2}, e_{3})-\{K_{1} +K_{2} +c.p.\}. $$
A similar formula for $I_{2}$ can be obtained in
a similar way.
The conclusion follows immediately. \qed
\section{Proof of Theorem 2.5}
This section is devoted to the proof of Theorem \ref{thm:main1}.
Throughout the section, we assume that $(A, A^{*})$ is a Lie bialgebroid and
$E=A\oplus A^{*}$ as in Theorem \ref{thm:main1}.
We also let ${\cal D}: C^{\infty}(P)\lon \gm (E)$ and $\rho :E\lon TP$ be
defined as in Section 2.
To prove Theorem \ref{thm:main1}, it suffices to verify
all the five identities in Definition \ref{def:quasi-algebroid}.
First, Equation (i) follows directly from Theorem
\ref{thm:main-jacobi} and properties of Lie bialgebroids.
Equation (iv) is equivalent to saying that $aa_{*}^{*}$ is
skew symmetric, which is again a property of a Lie bialgebroid
(see Proposition 3.6 in \cite{MackenzieX:1994}).
Below, we shall split the rest of the proof into several propositions.
\begin{pro}
For any $f\in C^{\infty}(P)$ and $e_{1}, e_{2}\in \gm (E)$, we
have
\begin{equation}
[e_{1}, fe_{2}]=f[e_{1}, e_{2}]+(\rho (e_{1})f)e_{2}-
(e_{1}, e_{2})_{+} {\cal D} f .
\end{equation}
\end{pro}
\noindent{\bf Proof.}\ Suppose that $e_{1}=X_{1}+\xi_{1}$ and $e_{2}=X_{2}+\xi_{2}$.
Then, we have
$$[e_{1}, fe_{2}]=[X_{1}, fX_{2}]+[X_{1}, f\xi_{2}]+[\xi_{1}, fX_{2}]
+[\xi_{1} , f\xi_{2}].$$
Here, we have
\be
&&[X_{1}, fX_{2}] =f [X_{1},X_{2}]+(a(X_{1})f)X_{2};\\
&&[ \xi_{1} , f\xi_{2}]=f[\xi_{1}, \xi_{2}]+((a_{*}\xi_{1} )f)\xi_{2};\\
&&[X_{1}, f\xi_{2}]=f[X_{1}, \xi_{2}]+((aX_{1})f)\xi_{2}
-\frac{1}{2} \langle X_{1}, \xi_{2} \rangle {\cal D} f ;\\
&&[\xi_{1}, fX_{2}]=f[\xi_{1},X_{2}] +((a_{*}\xi_{1} ) f)X_{2}-
\frac{1}{2} \langle X_{2} , \xi_{1} \rangle {\cal D} f .
\ee
The conclusion follows from adding up
all the equations above. \qed
\begin{pro}
\label{pro:hom}
For any $e_{1}, e_{2}\in \gm (E)$, we have
$$\rho [e_{1}, e_{2} ]=[\rho e_{1}, \rho e_{2}]. $$
\end{pro}
We need a lemma first.
\begin{lem}
\label{lem:4.2}
If $(A, A^{*})$ is a Lie bialgebroid with anchors $(a, a_{*})$,
then for any $X\in \gm (A) $ and $\xi \in \gm (A^{*})$,
$$ [a(X), a_{*}(\xi )] =
a_{*}(L_{X}\xi )-a (L_{\xi }X) + a a_{*}^{*}d_{0} \langle \xi , X \rangle . $$
\end{lem}
\noindent{\bf Proof.}\ For any $f\in C^{\infty}(P)$,
\be
&&(a a_{*}^{*}d_{0} \langle \xi , X \rangle ) f\\
&=& \langle d_{*}\langle \xi , X \rangle , df \rangle\\
&=&L_{df}\langle \xi , X \rangle\\
&=&\langle L_{df}\xi, X \rangle +\langle \xi, L_{df} X \rangle\\
&=& -\langle L_{\xi} df , X\rangle + \langle \xi, [X, d_{*}f]\rangle\\
&=& -a_{*}(\xi )\langle df , X\rangle +\langle df , L_{\xi} X \rangle
+ a(X)a_{*}(\xi )f-\langle L_{X}\xi , d_{*}f\rangle\\
&=&[a(X) , a_{*}(\xi )]f-a_{*}(L_{X}\xi )f+ a(L_{\xi} X )f,
\ee
where in the fourth equality we have used the fact that $L_{df}X=[X , d_{*}f]$,
a property of a general Lie bialgebroid (see Proposition 3.4
of \cite{MackenzieX:1994}).
\qed
{\bf Proof of Proposition \ref{pro:hom}} Let
$e_{1}=X_{1}+\xi_{1}$ and $e_{2}=X_{2}+\xi_{2}$.
\be
\rho [e_{1}, e_{2}]&=&
a\{ [X_{1}, X_{2}]+L_{\xi_{1}}X_{2}-L_{\xi_{2}}X_{1}-
d_{*}(e_{1}, e_{2} )_{-}\}\\
&&+a_{*}\{[\xi_{1} ,\xi_{2} ]+L_{X_{1}}\xi_{2}-L_{X_{2}}\xi_{1} +
d(e_{1}, e_{2})_{-}\}\\
&=&a[X_{1}, X_{2}]+a(L_{\xi_{1}}X_{2})-a(L_{\xi_{2}}X_{1})-
\frac{1}{2} a a_{*}^{*}d_{0} (\langle \xi_{1}, X_{2}\rangle- \langle\xi_{2},
X_{1} \rangle )\\
&&+a_{*}[ \xi_{1} , \xi_{2} ]+a_{*}(L_{X_{1}}\xi_{2})-a_{*}(L_{X_{2}}\xi_{1})
+\frac{1}{2} a_{*}a^{*} d_{0} (\langle \xi_{1}, X_{2}\rangle -\langle \xi_{2},
X_{1} \rangle )\\
&=&a[X_{1}, X_{2}]+[a(L_{\xi_{1}}X_{2})-a_{*}(L_{X_{2}}\xi_{1})
-aa_{*}^{*} d_{0} \langle \xi_{1}, X_{2} \rangle ]\\
&&-[ a(L_{\xi_{2}}X_{1})-a_{*}(L_{X_{1}}\xi_{2})-aa_{*}^{*}d_{0}
\langle \xi_{2} , X_{1} \rangle ]
+a_{*}[\xi_{1}, \xi_{2}]\\
&=&[aX_{1}, aX_{2}]+[a_{*}\xi_{1}, a_{*}\xi_{2}]
+[aX_{1}, a_{*}\xi_{2}]+[a_{*}\xi_{1}, aX_{2}]\\
&=&[\rho (e_{1}) , \rho (e_{2})],
\ee
where in the third equality we have used
the skew-symmetry of the operator $aa_{*}^{*}$, and
the second from the last follows from Lemma \ref{lem:4.2}.
\qed
\begin{pro}
For any $e, h_{1}, h_{2}\in \gm (E)$, we have
\begin{equation}
\label{eq:17}
\rho (e) (h_{1}, h_{2})_{+}=([e , h_{1}]+{\cal D} (e ,h_{1})_{+} , \ h_{2})_{+}+
(h_{1}, \ [e , h_{2}]+{\cal D} (e ,h_{2})_{+} )_{+}
\end{equation}
\end{pro}
\noindent{\bf Proof.}\ According to Equation (\ref{eq:T1}),
$$([e ,h_{1}], h_{2})_{+}=T(e, h_{1}, h_{2})+\frac{1}{2} \rho(e) (h_{1}, h_{2})_{+}
-\frac{1}{2} \rho (h_{1})(e, h_{2})_{+}$$
and
$$(h_{1}, [e, h_{2}])_{+}=T(e, h_{2}, h_{1})+\frac{1}{2} \rho(e) (h_{2}, h_{1})_{+}
-\frac{1}{2} \rho ( h_{2}) (e, h_{1})_{+}.$$
By adding these two equations, we obtain
Equation (\ref{eq:17}) immediately
since $T(e, h_{1}, h_{2})$ is skew-symmetric with respect to
$h_{1}$ and $h_{2}$.
\qed
\section{Proof of Theorem 2.6}
This section is devoted to the proof of Theorem \ref{thm:main2}.
We denote sections of $L_{1}$ by letters $X, Y$, and
sections of $L_{2}$ by $\xi , \eta $ etc.. For any
$X\in \gm (L_{1})$ and
$\xi \in \gm (L_{2})$, we define their pairing by
\begin{equation}
\langle \xi , X \rangle =2( \xi , X ).
\end{equation}
Since $(\cdot , \cdot )$ is nondegenerate,
$L_{2}$ can be considered as the dual bundle of
$L_{1}$ under this pairing. Moreover, the symmetric
bilinear form $(\cdot , \cdot )_{+} $ on $E$ defined by Equation (\ref{eq:pairing})
coincides with the original one.
By Proposition \ref{prop}, both $L_{1}$ and $L_{2}$ are Lie algebroids, and
their anchors are given by $a=\rho |_{L_{1}}$ and $a_{*}=\rho |_{L_{2}}$
respectively.
We shall use $d: \gm (\wedge^{*}L_{2})\lon \gm (\wedge^{*+1} L_{2} )
$ and $d_{*}: \gm (\wedge^{*}L_{1})\lon \gm (\wedge^{*+1} L_{1} )$
to denote their induced de-Rham differentials as usual.
Equation (v) in Definition \ref{def:quasi-algebroid} implies immediately
that the bracket between $ X\in \gm (L_{1}) $ and $ \xi \in \gm (L_{2}) $
is
given by
\begin{equation}
[X ,\xi ]= (-L_{\xi } X +\frac{1}{2} d_{*} \langle \xi , X \rangle )
+( L_{X}\xi -\frac{1}{2} d\langle \xi , X \rangle )
\end{equation}
Thus we have
\begin{pro}
Under the decomposition $E=L_{1}\oplus L_{2}$, for sections $e_{i}\in
\gm (E)$, $i=1, 2$ if
we write $e_{i}=X_{i}+\xi_{i}$,
then the bracket $[e_{1}, e_{2}]$ is given by
Equation (\ref{eq:double}).
\end{pro}
Before proving Theorem \ref{thm:main2}, we need the following lemma.
\begin{lem}
\label{lem:exchange}
Under the assumption of Theorem \ref{thm:main2}
we have
\be
&&L_{d_{*}f}\xi =-[df , \xi]; \\
&&L_{df}X=-[d_{*}f , X],
\ee
for any $f\in C^{\infty}(P)$, $X\in \gm (L_{1})$ and
$\xi\in \gm (L_{2})$.
\end{lem}
\noindent{\bf Proof.}\ Clearly, Equation (iv) in Definition
\ref{def:quasi-algebroid}
yields that $a\mbox{\tiny{$\circ $}} d_{*}=-a_{*} \mbox{\tiny{$\circ $}} d$.
Therefore,
\begin{eqnarray}
[a_{*} \xi , a X]&=&[\rho \xi , \rho X ] \nonumber\\
&=&\rho [\xi , X] \nonumber \\
&=&\rho ( L_{\xi } X -\frac{1}{2} d_{*} \langle \xi , X \rangle -L_{X}\xi +\frac{1}{2}
d\langle \xi , X \rangle ) \nonumber\\
&= &a( L_{\xi } X -\frac{1}{2} d_{*} \langle \xi , X\rangle )+ a_{*}(-L_{X}\xi +
\frac{1}{2} d\langle \xi , X \rangle) \nonumber\\
&=&a( L_{\xi } X)-a_{*}(L_{X}\xi )+
( a_{*} d ) \langle \xi , X \rangle . \label{eq:15}
\end{eqnarray}
On the other hand,
\begin{eqnarray}
((a_{*}d )\langle \xi , X \rangle ) f &=&
(a (d_{*}f))\langle \xi ,X \rangle \nonumber \\
&=&\langle L_{d_{*}f}\xi , X\rangle +\langle \xi , [d_{*}f , X]\rangle \nonumber \\
&=&\langle [\xi , df], X \rangle -\langle \xi , L_{X}d_{*}f\rangle
+\langle L_{d_{*}f}\xi +[df , \xi ], X\rangle \nonumber \\
&=&a_{*}(\xi )a (X)f- \langle df , L_{\xi }X\rangle -
a(X)a_{*}(\xi )f +\langle L_{X}\xi , d_{*}f\rangle
+\langle L_{d_{*}f}\xi +[df , \xi ], X\rangle \nonumber \\
&=&[a_{*}(\xi ), a(X)]f-a(L_{\xi }X)f+a_{*} (L_{X}\xi )f
+\langle L_{d_{*}f}\xi +[df , \xi ], X\rangle . \label{eq:16}
\end{eqnarray}
Comparing Equation (\ref{eq:15}) with ( \ref{eq:16}),
we obtain
$$\langle L_{d_{*}f}\xi +[df , \xi ] , X\rangle
=0.$$
Therefore, $L_{d_{*}f}\xi =-[df , \xi ]$. The other equation
can be proved similarly.
\qed
{\bf Proof of Theorem \ref{thm:main2} }
It follows from Theorem \ref{thm:main-jacobi} that
$J_{1} +J_{2} +c.p. =0$, for any $e_{1}, e_{2}$ and $e_{3}\in \gm (E)$.
Using Lemma \ref{lem:exchange}, we have $J_{1} + c.p. =0$.
In particular, if we take $e_{1}=X_{1}, \ e_{2}=X_{2}$ and
$e_{3}=\xi_{3}$, we obtain that $i_{\xi_{3}} (d_{*} [X_{1}, X_{2}]-L_{X_{1}}d_{*}X_{2}
+L_{X_{2}}d_{*}X_{1} )=0$,
which implies the compatibility condition between
$A$ and $A^{*}$. \qed
\section{Hamiltonian operators}
Throughout this section, we will assume that $(A, A^{*})$ is a Lie bialgebroid.
Suppose that $H: A^{*}\lon A$ is a bundle map. We denote by
$A_{H}$ the graph of $H$, considered as a subbundle of $E=A\oplus A^{*}$.
I.e., $A_{H}=\{ H\xi +\xi |\xi\in A^{*}\}$.
\begin{thm}
$A_{H}$ is a Dirac subbundle iff $H$ is skew-symmetric and
satisfies the following Maurer-Cartan type equation:
\begin{equation}
\label{eq:Maure}
d_{*}H+\frac{1}{2} [H, H]=0.
\end{equation}
In this equation, $H$ is considered as a section of $\wedge^{2}A$.
\end{thm}
In the sequel, we shall use the same symbol to denote a section
of $\wedge^{2}A$ and its induced bundle map if
no confusion is caused. \\
\noindent{\bf Proof.}\ It is easy to see that $A_{H}$ is isotropic
iff $H$ is skew-symmetric.
For any $\xi , \eta \in \gm (A^{*})$,
let
\begin{equation}
[\xi , \eta ]_{H}=L_{H\xi}\eta -L_{H\eta }\xi +d \langle \xi , H\eta \rangle.
\end{equation}
Since
\be
&&[H\xi , \eta ]=L_{H\xi }\eta -L_{\eta }H\xi - \frac{1}{2} (d-d_{*})\langle \eta ,
H\xi \rangle \ \ \mbox{and} \\
&&[\xi , H\eta ]=L_{\xi}H\eta -L_{H\eta }\xi +\frac{1}{2} (d-d_{*})\langle \xi ,
H\eta \rangle ,
\ee
then,
$$[H\xi , \eta ]+ [\xi , H\eta ]=[\xi , \eta ]_{H}+L_{\xi}H\eta -L_{\eta}H\xi
+d_{*}\langle \eta , H\xi \rangle .$$
On the other hand, we have the following
formula (see \cite{K-SM:1990}):
\begin{equation}
[H\xi , H\eta ]=H[\xi , \eta ]_{H}-\frac{1}{2} [H , H](\xi ,\eta ).
\end{equation}
Therefore, we have,
\be
[H\xi +\xi , H\eta +\eta ]&=&[H\xi , H\eta ]+[\xi , H\eta ]+[H\xi , \eta ]+
[\xi , \eta ]\\
&=&(L_{\xi}H\eta -L_{\eta}H\xi +d_{*}\langle \eta , H\xi \rangle
+H[\xi , \eta ]_{H}-\frac{1}{2} [H, H](\xi , \eta ))\\
&& +([\xi , \eta ]+[\xi , \eta ]_{H} ).
\ee
It thus follows that $A_{H}$ is integrable iff
for any $\xi , \eta \in \gm (A^{*})$
\begin{equation}
\label{eq:25}
H[\xi , \eta ]= L_{\xi}H\eta -L_{\eta}H\xi +d_{*}\langle \eta , H\xi \rangle
-\frac{1}{2} [H, H](\xi , \eta ).
\end{equation}
On the other hand,
\be
(d_{*}H ) (\xi , \eta , \zeta )&=&a_{*}(\xi )\langle \eta , H\zeta \rangle
-a_{*}(\eta )\langle \xi , H\zeta \rangle
+a_{*}(\zeta )\langle \xi , H\eta \rangle\\
&&-\langle [\xi ,\eta ], H\zeta \rangle +\langle [\xi , \zeta ], H\eta \rangle
-\langle [\eta , \zeta ], H\xi \rangle\\
&=&\langle H[\xi , \eta ]+L_{\eta }H\xi -L_{\xi }H\eta +d_{*}\langle \xi , H\eta \rangle , \zeta \rangle .
\ee
Hence,
\begin{equation}
(d_{*}H ) (\xi , \eta )=H[\xi ,\eta ]+L_{\eta }H\xi -L_{\xi }H\eta
-d_{*}\langle \eta , H\xi \rangle .
\end{equation}
This implies that Equation (\ref{eq:25}) is equivalent to
$$(d_{*}H )(\xi , \eta ) +\frac{1}{2} [H, H](\xi , \eta )=0, $$
or equivalently
$$d_{*}H+\frac{1}{2} [H, H]=0. $$ \qed
{\bf Remark} (1). Because of the symmetric role of $A$ and $A^*$,
we have the following assertion: the graph $A_{I}=\{ X+I X|X\in \gm ( A )\}$
of a bundle map $I: A\lon A^{*}$
defines a Dirac subbundle iff
$I$ is skew-symmetric and
satisfies the following Maurer-Cartan type equation:
\begin{equation}
dI+\frac{1}{2} [I, I]=0.
\end{equation}
(2). For the canonical Lie bialgebroid $(TM, T^{*}M)$ where $M$
is equipped with the zero Poisson structure, Equation (\ref{eq:Maure}) becomes
$[H,H]=0$, which is the defining equation for a Poisson structure. On the other
hand, if we exchange $TM$ and $T^{*}M$, and consider the Lie bialgebroid $(T^{*}M
, TM )$, the bracket term drops out of Equation (\ref{eq:Maure}), whose solutions
correspond to a presymplectic structures. Encompassing these two
cases into a general framework was indeed the main motivation for Courant
\cite{Courant:1990} to define and study Dirac structures.\\
(3). The Maurer-Cartan equation is a kind of
integrability equation. It is also basic in deformation theory, where it may
live on a variety of differential graded Lie algebras. It would be interesting
to place the occurrence of this equation in our theory in a more general context.
\begin{defi}
Given a Lie bialgebroid $(A, A^{*})$, a section
$H\in \gm (\wedge^{2} A)$ is called a {\em hamiltonian
operator } if $A_{H}$ defines a Dirac structure.
$H$ is called a {\em strong hamiltonian operator} if
$A_{\lambda H}$ are Dirac subbundles for all $\lambda \in {\mathbb R}$.
\end{defi}
\begin{cor}
For a Lie bialgebroid $(A, A^{*})$, $H\in \gm (\wedge^{2} A)$
is a hamiltonian operator
if Equation (\ref{eq:Maure}) holds. It is a strong hamiltonian operator
if $d_{*}H=[H, H]=0. $
\end{cor}
For a hamiltonian operator $H$, $A_{H}$ is a Dirac subbundle
which is transversal to $A$ in $A\oplus A^{*}$. Therefore, $(A, A_{H})$ is a
Lie bialgebroid according to Theorem \ref{thm:main2}.
In fact, $A_{H}$ is isomorphic to $A^{*}$,
as a vector bundle, with the anchor and bracket of its Lie algebroid
structure
given respectively by $\hat{a}_{*}=a_{*}+a\mbox{\tiny{$\circ $}} H$
and $[\xi , \eta \hat{]}=[\xi , \eta ]+[\xi , \eta]_{H}$, for all
$\xi , \eta \in \gm (A^{*})$. In particular, if
$H$ is a strong hamiltonian operator, one obtains
a one parameter family of Lie bialgebroids transversal
to $A$, which can be considered as a deformation
of the Lie bialgebroid $(A, A^{*})$.
Conversely, any Dirac subbundle
transversal to $A$ corresponds to a hamiltonian
operator in an obvious way. For example,
consider the standard Lie bialgebra $(\mathfrak{k}, \mathfrak{b})$
arising from the Iwasawa decomposition of $\mathfrak{k}^{{\mathbb C}}$
\cite{LuW}, where $\mathfrak{k} $ is a compact semi-simple Lie algebra and
$\mathfrak{b}$ its corresponding dual Lie algebra.
Then any real form of $\mathfrak{k}^{{\mathbb C}}$
which is transversal to $\mathfrak{b}$ will correspond
to a hamiltonian operator (see \cite{LiuQ} for a complete
list of such real forms for simple Lie algebras).
It is straightforward to check that such a hamiltonian operator is not strong.
On the other hand, for the Cartan subalgebra $\mathfrak{h}$ of
$\mathfrak{k}$, every element in $\wedge^{2}h$
gives rise to a strong hamiltonian
operator $H: \mathfrak{k}^{*} \lon \mathfrak{k}$,
This gives rise to
a deformation of
the standard Lie bialgebra $(\mathfrak{k}, \mathfrak{b})$ (see \cite{LS}).
These examples can be generalized to
any gauge algebroid
associated to
a principal $K$-bundle.
Below, we will look at hamiltonian operators in two special
cases, each of which corresponds to some familiar
objects.
\begin{ex}
Let $P$ be a Poisson manifold with Poisson tensor $\pi$.
Let $(TP, T^* P)$ be the canonical
Lie bialgebroid associated to the Poisson manifold $P$
and $E=TP \oplus T^* P$ equipped with the
induced Courant algebroid structure.
It is easy to see that a bivector field $H$ is a hamiltonian operator
iff $H+\pi$ is a Poisson tensor. $H$ is a strong hamiltonian operator
iff $H$ is a Poisson tensor Schouten-commuting with $\pi$.
\end{ex}
\begin{ex}
Similarly, we may switch $TP$ and $T^{*}P$, and
consider the Lie bialgebroid $(T^{*}P, TP )$ associated to
a Poisson manifold $P$ with Poisson structure $\pi$.
Let $E= T^{*}P \oplus TP$ be equipped with its
Courant algebroid structure. In this case, a
hamiltonian operator corresponds to a two-form
$\omega \in \Omega^{2}(P)$ satisfying
$d\omega + \frac{1}{2} [\omega , \omega ]_{\pi }=0$.
Here, $[\cdot , \cdot ]_{\pi}$ refers to the Schouten bracket
of differential forms on $P$ induced by the Poisson structure
$\pi$.
Given a hamiltonian operator $\omega$, its
graph $A_{\omega }$ defines a Dirac subbundle transversal to
$T^{*}P$, the first component of $E$ also being considered as
a Dirac subbundle. Therefore, they form
a Lie bialgebroid. Their induced Poisson structure
on the base space can be easily checked to be given
by
$-2(\pi^{\#}+N\pi^{\#}) $, where
$N: TP \lon TP$ is the composition
$\pi^{\#}\mbox{\tiny{$\circ $}} \omega^{b}$ and $\omega^{b} : TP \lon T^{*}P$
is the bundle map induced by the two-form $\omega$.
If $\omega $ is a strong hamiltonian operator,
then $N \pi^{\#} $ defines a Poisson structure compatible with
$\pi$. In fact, in this case, $(\pi , N)$ is
a Poisson-Nijenhuis structure
in the sense of \cite{K-SM:1990}.
\end{ex}
We note that Vaisman \cite{va:complementary} has studied 2-forms on
Poisson manifolds satisfying the condition $[\omega , \omega
]_{\pi}=0$. Such forms, called {\em complementary} to the Poisson
structure, also give rise to new Lie algebroid structures on $TM$.
To end this section, we
describe a example of Lie bialgebroids, where both the
algebroid and its dual arise from
hamiltonian operators.
\begin{pro}
Let $U$ and $V$ be Poisson tensors over a manifold $M$
and denote by $T^{*}M_{U}$ and $T^{*}M_{V}$ their associated
canonical cotangent Lie algebroids on $T^{*}M$.
Assume that $U-V$ is nondegenerate. Then $(T^{*}M_{U}, T^{*}M_{V})$
is a Lie bialgebroid, where their pairing is given
by $(\xi , \eta )=(U-V)(\xi , \eta )$ for any
$\xi \in T^{*}M_{U}$ and
$\eta \in T^{*}M_{V}$. Their induced Poisson tensor on the
base space $M$ is given by
$-2U(U-V)^{-1}V$.
\end{pro}
\noindent{\bf Proof.}\ Let $E=TM\oplus T^* M$ be equipped with the usual
Courant bracket.
Since both $U$ and $V$ are Poisson tensors,
their graphs $A_{U}$ and $A_{V}$ are Dirac subbundles.
They are transversal since $U-V$ is nondegenerate.
Therefore, $(A_{U}, A_{V})$ is a Lie bialgebroid,
where their pairing is given by
\begin{equation}
\langle\!\langle U\eta +\eta , V\xi +\xi \rangle\!\rangle = \frac{1}{2} \langle \xi , (U-V)\eta \rangle .
\end{equation}
On the other hand, as Lie algebroids,
$A_{U}$ and $ A_{V}$ are clearly isomorphic to
cotangent Lie algebroids $T^{*}M_{U}$ and $ T^{*}M_{V}$
respectively.
Moreover, their anchors $a_{U}: A_{U}\lon TM $ and $a_{V}: A_{V}\lon TM $
are given respectively by
$a_{U} (U\xi +\xi )=U\xi $ and
$a_{V} (V\xi +\xi )=V\xi $.
This proves the first part of the proposition.
To calculate their induced Poisson structure on the base $M$,
we need to find out
the dual map $a_{V}^{*}: T^{*}M\lon A_{V}^{*}\cong A_{U}$.
For any $\xi \in T^{*}M$,
we assume that $a_{V}^{*} (\xi )=U\eta +\eta \in A_{U}$ via the
identification above.
For any $\zeta \in T^{*}M$,
\be
( a_{V}^{*} \xi , V\zeta +\zeta )&=& \langle \xi , a_{V}
(V\zeta +\zeta )\rangle \\
&=& \langle \xi , V\zeta \rangle
\ee
On the other hand, $( a_{V}^{*} \xi , V\zeta +\zeta )=
\langle\!\langle U\eta +\eta , V\zeta +\zeta \rangle\!\rangle
=\frac{1}{2} \langle \zeta , (U-V)\eta \rangle$.
It thus follows that $\eta =-2(U-V)^{-1}V\xi $.
Therefore, according to Proposition 3.6 in
\cite{MackenzieX:1994}, the induced
Poisson structure $a_{U}\mbox{\tiny{$\circ $}} a_{V}^{*}:
T^{*}M\lon TM$ is given by $(a_{U}\mbox{\tiny{$\circ $}} a_{V}^{*})\xi
=-2U(U-V)^{-1}V\xi $. \qed
Replacing $V$ by $-V$ in the proposition above, we obtain the following
``composition law'' for Poisson structures.
\begin{cor}
\label{cor-composition}
Let $U$ and $V$ be Poisson tensors over manifold $M$
such that $U+V$ is nondegenerate.
Then, $U(U+V)^{-1}V$ also defines a Poisson tensor on $M$.
\end{cor}
Note that, if $U$ and $V$ are nondegenerate, then
$U(U+V)^{-1}V=(U^{-1}+V^{-1})^{-1}$ is the Poisson tensor corresponding to the
sum of the symplectic forms for $U$ and $V$. Since the sum of closed forms is
closed, it is obvious in this case that $U(U+V)^{-1}V$ is a Poisson tensor. We do
not know such a simple proof of Corollary \ref{cor-composition} in the general
case.
\section{Null Dirac structures and Poisson reduction}
In this section, we consider another class of Dirac structures related
to Poisson reduction and dual pairs of Poisson manifolds.
\begin{pro}
\label{pro:null}
Let $(A, A^{*} )$ be a Lie bialgebroid, and $h \subseteq A$ a subbundle of $A$.
Then $L=h\oplus h^{\perp} \subseteq E=A\oplus A^* $
is a Dirac structure iff $h$ and $h^{\perp}$ are, respectively,
Lie subalgebroids of $A$ and $A^* $.
\end{pro}
\noindent{\bf Proof.}\ Obviously, $L=h\oplus h^{\perp}$ is a maximal isotropic subbundle of $E$.
If $L$ is a Dirac structure, clearly $h$ and $h^{\perp}$
are Lie subalgebroids of $A$ and $A^*$ respectively.
Conversely, suppose that both $h$ and $h^{\perp}$ are
Lie subalgebroids of $A$ and $A^*$ respectively. To prove
that $L$ is a Dirac structure, it suffices to show that
$[X, \xi ]$ is a section of $L$ for any
$X \in \gm (h) $ and $\xi \in \gm (h^{\perp})$. Now
$$[X,\xi]=L_{X}\xi-L_{\xi}X .$$
For any section $Y\in \gm (h)$,
$$<L_{X}\xi, \ Y>=a(X)<\xi,Y>-<\xi,[X,Y]>=0. $$
Therefore, $L_{X}\xi $ is still a section of $h^{\perp}$.
Similarly, $L_{\xi}X $ is a section of $h$.
This concludes the proof of the proposition. \qed
It is clear that a subbundle $L\subseteq E $ is of the form $L=h\oplus h^{\perp}$
iff the minus two-form $(\cdot,\cdot)_{-}$ on $E$, as defined
by Equation (\ref{eq:pairing}), vanishes on $L$.
For this reason, we call a Dirac structure of this
form a {\em null Dirac structure}.
An immediate consequence of Proposition \ref{pro:null}
is the following:
\begin{cor}
\label{cor:reduction}
Let $(P, \ \pi )$ be a Poisson manifold, and $D$ a subbundle of $TP$.
Let $T^{*}P$ be equipped with the cotangent Lie
algebroid structure so that $(TP, \ T^{*}P)$
is a Lie bialgebroid. Then $L=D\oplus D^{\perp} $
is a Dirac structure in $E= TP\oplus T^{*}P$
iff $D$ is an integrable distribution and
the Poisson structure on $P$ descends to
a Poisson structure on the quotient space $P/D $ \footnote{When the
quotient space is not a manifold,
this means that at each point there is a local
neighborhood $U$ such that the Poisson structure
on $U$ descends to its quotient.} such that the natural projection
is a Poisson map.
\end{cor}
\noindent{\bf Proof.}\ This follows directly from the following
\begin{lem}
\label{lem:reduction}
Let $D$ be an integrable distribution on a Poisson
manifold $P$.
$P/D$ has an induced Poisson structure (in the above general
sense) iff $D^{\perp}\subset T^{*}P$ is
a subalgebroid\footnote{Such a foliation is also called
cofoliation by Vaisman \cite{vaisman:book}}.
\end{lem}
\noindent{\bf Proof.}\ For simplicity, let us assume that $P/D$ is a manifold.
The general case will follow from the same principle.
It is clear that a function $f$ is constant along leaves of $D$ iff
$df $ is a section of $D^{\perp}$. If $D^{\perp}$ is a
subalgebroid, it follows from the equation
\begin{equation}
\label{eq:dfg}
d\{f,\ g\}=[df, dg]
\end{equation}
that $C^{\infty}(P/D)$ is a Poisson algebra.
Conversely, a local one-form $fdg$ is in $D^{\perp}$
iff $g$ is constant along $D$. The conclusion thus
follows from Equation (\ref{eq:dfg}) together with
the Lie algebroid axiom relating the bracket and
anchor. \qed
{\bf Remark.} \ Poisson reduction was considered by Marsden
and Ratiu in \cite{MR}. Lemma \ref{lem:reduction} can be
considered as a special case of their theorem when $P=M$
in the Poisson triple $(P, \ M, \ E)$ (see \cite{MR}).
It would be interesting to interpret their general
reduction theorem in terms of Dirac structures
as in Corollary \ref{cor:reduction}.
The rest of the section is devoted to several examples
of Corollary \ref{cor:reduction}, which will lead to some
familiar results in Poisson geometry.
Recall that, given a
Poisson Lie group $G$ and a Poisson manifold $M$, an action
$$ \sigma : G\times M \lon M$$
is called a Poisson action if $\sigma$ is a Poisson map.
In this case, $M$ is called a Poisson $G$-space.
Now consider $P=G \times M$ with the product Poisson structure and
diagonal $G$-action. Then $P/G$ is isomorphic
to $M$, and the projection from $P$ to $P/G=M$
becomes the action map $\sigma$, which is a
Poisson map when $P/G$ is equipped with the given
Poisson structure on $M$.
By Corollary \ref{cor:reduction}, we obtain
a null Dirac structure $L=D\oplus D^{\perp}$ in
$TP\oplus T^*P$.
Clearly, $L$ is a Lie algebroid over $P$, which is $G$-invariant.
It would be interesting to explore the relation between this
algebroid and the one defined on
$ (M\times {\mathfrak g})\oplus T^*M $, which was studied by Lu in \cite{Lu}.
For a Poisson Lie group $G$ with tangential Lie bialgebra
$({\mathfrak g} ,{\mathfrak g}^{*})$, the Courant algebroid double $E=
TG \oplus T^{*}G$ can be identified, as a vector
bundle, with the trivial product
$G \times ({\mathfrak g} \oplus {\mathfrak g}^{*}) $ via left translation.
Under such an identification, a left invariant
null Dirac structure has the form $L=G\times (h\oplus h^{\perp})$,
where $h$ is a subalgebra of ${\mathfrak g}$ and $h^{\perp}$ is
a subalgebra of ${\mathfrak g}^{*}$. Thus, one obtains
the following reduction
theorem:
for a connected closed subgroup $H$ with Lie algebra $h$, $G/H$
has an induced Poisson
structure iff $h^{\perp}$ is a subalgebra of ${\mathfrak g}^*$.
More generally, let $G$ be a Poisson group, $M$ a Poisson $G$-space.
Suppose that $H\subseteq G$ is a closed subgroup with Lie algebra $h$.
Assume that $M/H$ is a nice manifold such that the projection
$p: M\lon M/H$ is a submersion. Then the $H$-orbits define an
integrable distribution ${\mathfrak h}$ on $M$. According to Corollary \ref{cor:reduction},
the Poisson structure on $M$ descends to $M/H$
iff ${\mathfrak h}^{\perp}$ is a subalgebroid
of the cotangent algebroid $T^{*}M$ of the Poisson manifold $M$.
On the other hand, we have
\begin{pro}
\label{pro-perp}
If $h^{\perp}$ is a subalgebra of ${\mathfrak g}^*$, then ${\mathfrak h}^{\perp}$
is a subalgebroid of $T^{*}M$. Conversely, if the
isotropic subalgebra at each point is a subalgebra of $h$,
and in particular if the action is locally free, then
that ${\mathfrak h}^{\perp}$ is a subalgebroid implies
that $h^{\perp}\subseteq {\mathfrak g}^*$ is a subalgebra.
\end{pro}
\noindent{\bf Proof.}\ It is easy to see that ${\mathfrak h}^{\perp}\cong \phi^{-1}(h^{\perp})$, where
$\phi : T^{*}M\lon {\mathfrak g}^{*}$ is the momentum mapping for the
lifted $G$-action on $T^{*}M$, equipped with the canonical
cotangent symplectic structure.
According to Proposition 6.1 in \cite{Xu},
$\phi : T^{*}M\lon {\mathfrak g}^{*}$ is a Lie algebroid morphism. Before
continuing, we need the following
\begin{lem}
Let $A\lon M$ be a Lie algebroid with anchor $a$, ${\mathfrak g}$ a Lie algebra, and $\phi :A\lon
{\mathfrak g}$ an algebroid morphism. Suppose that $h\subseteq {\mathfrak g}$ is a subalgebra
such that $\phi^{-1}h \subseteq A$ is a subbundle. Then $\phi^{-1}h$ is a
subalgebroid.
Conversely, given a subalgebroid $B\subseteq A$, if $\phi (B|_{m})$ is
independent of $m\in M$, then it is a subalgebra of ${\mathfrak g}$.
\end{lem}
\noindent{\bf Proof.}\ This follows directly from the following equation
(see \cite{Mackenzie:1992}):
$$
\phi [X,Y]=(aX)(\phi Y)-(aY)(\phi X)+{[\phi X, \ \phi Y]}^{.}, \ \ \forall
X, Y\in \gm (A),
$$
where $\phi X$, $\phi Y$ and $\phi [X, Y]$ are considered as
${\mathfrak g}$-valued functions on $P$, and $[\cdot , \cdot]^{\cdot}$
refers to the pointwise bracket. \qed
Now, the first part of Proposition \ref{pro-perp}
is obvious according to the lemma above. For the
second part, we only need to note that
$\phi ({\mathfrak h}^{\perp} ) = h^{\perp}\cap Im \phi $, and the assumption that
the isotropic subalgebra at each point is a subalgebra of $h$ is equivalent
to that $ h^{\perp}\subseteq Im \phi $.
This concludes the proof of the Proposition. \qed
From the above discussion, we have the following conclusion:
if $h^{\perp}$ is a subalgebra of ${\mathfrak g}^*$, then the Poisson structure
on $M$ descends to $M/H$. This is a well-known reduction theorem of
Semenov-Tian-Shansky \cite{STS} (see also \cite{we:coisotropic}).
Conversely, if the
isotropic subalgebra at each point is a subalgebra of $h$,
and in particular if the action is locally free, the converse
also holds.
Another interesting example arises when $P$ is a symplectic manifold
with an invertible Poisson tensor $\pi$.
In this case, $\pi^{\#}: \ T^{*}P\lon TP$ is a Lie algebroid isomorphism.
Given a null Dirac structure $L=D\oplus D^{\perp}$,
$\bar{D}= \pi^{\#}(D^{\perp})$ is a subalgebroid of
$TP$. It is simple to see that $(\bar{D})^{\perp}=(\pi^{\#})^{-1}(D)$, and
is therefore a subalgebroid of $T^{*}P$.
Thus, $\bar{L} \stackrel{def}{=}\bar{D}\oplus (\bar{D})^{\perp}$
defines another null Dirac structure.
It is easy to see that $D$ and $\bar{D}$ are symplectically
orthogonal to each another. Thus $P/\bar{D}$ is a Poisson
manifold (assume that it is a nice manifold) so that
$P/D$ and $P/\bar{D}$ constitute a full dual pair,
which is a well known result of Weinstein \cite{we:1983}.
Conversely, it is clear that a full dual pair corresponds to
a null Dirac structure.
|
\section{Introduction}
\mbox{}
The Haldane Shastry Model [H,S] has been a subject of much attention in the
past years. This Model is a version of the XXX Heisenberg Spin Chain where
spins interact with a potential inversely proportional to the squared distance
between the spins. Like the XXX Chain with nearest-neighbour interaction, the
Haldane Shastry Model is Integrable in the sense that its Hamiltonian is a
member of a family of mutually commuting, independent Integrals of Motion. The
Haldane Shastry Model, however, has been solved in much greater detail than the
nearest-neighbour XXX Model. The reason for this is a remarkably large algebra
of symmetries found in the former Model. For the Haldane Shastry Spin Chain
with the spins taking values in the fundamental representation of ${\frak
g}{\frak l}_n$ the symmetry algebra is the ${\frak g}{\frak l}_n$-Yangian. This
infinite-dimensional algebra of symmetries facilitated computation of explicit
expressions for the energy levels and the eigenvectors!
[HHTBP,BPS,H2]. Some of the corre
lation functions have been found as well [HZ].
An important reason for the attention attracted by the Haldane Shastry Model
is the fractional statistics exhibited by elementary excitations over the
antiferromagnetic ground state present in this Model [H2]. In the case of
${\frak g}{\frak l}_2$-Haldane Shastry Chain these excitations are spin-1/2
particles that are ``semions'' -- particles with statistics exactly half-way
between bosons and fermions in the sense of Haldane. Being exactly solvable the
Haldane Shastry Model provides a valuable laboratory for a study of physical
implications of the fractional statistics.
A remarkable connection exists between the Haldane Shastry Spin Chain in the
limit of infinite number of sites and WZNW Conformal Field Theory at level 1
[BPS2,BLS]. This connection has led to a novel description of the space of
states in WZNW level-1 CFT, where the states are organized into irreducible
multiplets of Yangian symmetry algebra inhereted from the Haldane Shastry
Model. This, in turn, provided an explanation for the Fermionic Virasoro
character formulas for the level-1 integrable representations of $\hat{{\frak
s}{\frak l}}_2$ that were earlier derived by [DKKMM].
In the paper [BGHP] it was realized that the Haldane Shastry Spin Chain is
related to a more general class of Integrable Models -- the so-called Dynamical
Models that describe quantum particles with spin moving along a circle. These
Dynamical Models can be thought of as generalzations of the Calogero-Sutherland
Model with inverse squared sine potential. The precise way in which the Haldane
Shastry Hamiltonian and associated higher conserved charges are obtained from
the hierarchy of Dynamical Models was recently explained by Polychronakos in
[P], and by Talstra and Haldane in [TH]. In the last paper the authors have
shown how the hierarchy of Integrable Spin Models including the Haldane Shastry
Model appears in the static limit of the Dynamical Models in which the
coordinates of the particles are ``frozen'' along the circle in an
equidistantly spaced lattice.
In the present paper we define trigonometric counterparts of the spin-1/2
Haldane Shastry Hamiltonian and the associated higher conserved charges. The
hierarchy of Integrable Spin Chain Hamiltonians that we obtain has an
infinite-dimensional symmetry algebra $U_q'(\hat{{\frak g}{\frak l}}_2)$ that takes over the role played by
the Yangian in the Haldane Shastry Model. We compute the eigenvalue spectrum of
the hierarchy and find that it has an additive, particle-like form. The space
of states is decomposed into eigenspaces of the operators that form the
hierarchy. Each of the eigenspaces is a highest-weight irreducible
representation of $U_q'(\hat{{\frak g}{\frak l}}_2)$ parametrized by a sequence of integers --``magnon
quasimomenta'' or a ``motif'' in terminology of [HHTBP,BPS]. The eigenvalue
that corresponds to such eigenspace is a $q$-deformation of the eigenvalue of
the Haldane Shastry Hamiltonian parametrized by the same magnon quasimomenta.
The procedure that is used to derive the trigonometric hierarchy has been
inspired by the Talstra-Haldane approach -- we extract the Integrable Spin
Models from a static limit of the trigonometric $U_q'(\hat{{\frak g}{\frak l}}_2)$-invariant Dynamical
Models that were defined in [BGHP]. The spectrum of the Spin Models is obtained
by a descent from the spectrum of these Dynamical Models.
\subsection{A survey of the method and results}
In this subsection we highlight the main steps of the procedure that is used to
define the trigonometric hierarchy of Integrable Spin Models and formulate the
results of this article. The details and proofs of the statements are contained
in the main body of the paper starting with sec.1.
We derive the hierachy of trigonometric Spin Models by a two-step reduction
from the trigonometric Dynamical Models that were introduced in [BGHP]. First
of all we recall the definition of these trigonometric Dynamical Models.
\subsubsection{Trigonometric Dynamical Models of [BGHP]}
The trigonometric Dynamical Models are defined starting with two
representations of the finite-dimensional Hecke Algebra $H_N(q)$ (Cf. {\bf
1.1}). The first of these representations is defined in the ring of polynomials
in $N$ variables ${\Bbb C}[z_1,\dots,z_N]$. The generators $g_{i,i+1} \quad (i=1,\dots,N-1)$ of $H_N(q)$ in
this representation have the following form
\begin{equation}
g_{i,i+1} :=\frac{q^{-1}z_i - qz_{i+1}}{z_i-z_{i+1}}(K_{i,i+1} - 1) + q
\qquad (i=1,\dots,N-1) .
\end{equation}
Where $ K_{i,j} $ is the exchange operator for variables $ z_i , z_j $. \\
The second representation of $H_N(q)$ is defined in the $N$-fold tensor product
of two-dimensional vector spaces $H := V^{\otimes N},\; V:= {\Bbb C}^2 = {\Bbb C}\{
v^+ , v^-\}$. The Hecke generator $t_{i,i+1} \quad (i=1,\dots,N-1)$ is a matrix acting in $H$
according to the formula
\begin{equation}
t_{i,i+1} = I\otimes \dots \otimes I\otimes\underset{i,i+1}{t}\otimes I \otimes \dots
\otimes I \quad (i=1,\dots,N-1),
\end{equation}
where $t$ is the matrix which acts in $ V\otimes V$:
\begin{eqnarray}
t v^+\otimes v^- & = & (q-q^{-1})v^+\otimes v^- + v^-\otimes v^+ \\
t v^-\otimes v^+ & = & v^+\otimes v^- \\
t v^\pm\otimes v^\pm & = & qv^\pm\otimes v^\pm.
\end{eqnarray}
The two of these representations naturally extend to $ {\Bbb C}[z_1,\dots,z_N]\otimes H$.
The Hecke representation generated by $g_{i,i+1} \quad (i=1,\dots,N-1)$ is enlarged to a
representation of the Affine Hecke Algebra $\widehat{H_N(q)}$ by adjoining
affine generators $ Y_i \quad (i=1,\dots,N)$ :
\begin{eqnarray}
Y_i & := & g_{i,i+1}^{-1}K_{i,i+1}\dots
g_{i,N}^{-1}K_{i,N}p^{D_i}K_{1,i}g_{1,i}\dots K_{i-1,i}g_{i-1,i}.
\end{eqnarray}
Where $ p $ is a $c$-number , $ D_i := z_i \frac{\partial}{\partial z_i} $ and
\begin{equation}
g_{i,j} :=\frac{q^{-1}z_i - qz_j}{z_i-z_j}(K_{i,j} - 1) + q \qquad (i,j =
1,\dots,N-1) .
\end{equation}
After introducing the objects just described the authors of [BGHP] define the
hierarchy of Dynamical Models. The operators $ \Delta^{(n)} \quad (n=1,\dots,
N)$ that constitute the hierarchy are coefficients of the polynomial $\Delta (u) =
\sum_{n=0}^N u^n \Delta^{(n)}$ which generates elementary symmetric functions
of $ Y_1,Y_2,\dots,Y_N $ :
\begin{equation}
\Delta (u) : = \prod_{i=1}^N ( 1 + u Y_i ) .
\end{equation}
Due to the mutual commutativity of the operators $ Y_1,Y_2,\dots,Y_N $ the
hierachy of Dynamical Models is integrable:
\begin{equation}
[ \Delta(u) , \Delta(v) ] = 0 .
\end{equation}
\mbox{}
In the space $ {\Bbb C}[z_1,\dots,z_N]\otimes H $ one defines a representation of the algebra $
U_q'(\hat{{\frak g}{\frak l}}_2) $ (Cf. {\bf 1.2}). In this representation the generators of $ U_q'(\hat{{\frak g}{\frak l}}_2) $ are
obtained by expanding in the parameter $u$ the monodromy matrix $ T_a(u) $
which is defined in a standard way as a product of elementary $L$-operators
[BGHP]:
\begin{equation}
T_a(u) : = L_{a1}(uY_1)L_{a2}(uY_2)\dots L_{aN}(uY_N).
\end{equation}
Where the elementary $L$-operator $L_{ai}(uY_i)\quad (i=1,\dots,N)$ acts in the tensor
product of an auxiliary copy of the two-dimensional vector space denoted by
$V_a$ and the space $ {\Bbb C}[z_1,\dots,z_N]\otimes H$:
\begin{eqnarray}
L_{ai}(uY_i) & := & \frac{uY_it_{a,i} - t_{a,i}^{-1}}{uY_i - 1}P_{a,i} \quad
(i=1,\dots,N).
\end{eqnarray}
Here $P$ is the permutation operator in $ V\otimes V$ and $ t_{a,i} \, , \,
P_{a,i} $ are the usual extensions of $ t \, , \, P$ as operators in $
V_a\otimes H$.
The fact that $ T_a(u) $ defines a representation of $ U_q'(\hat{{\frak g}{\frak l}}_2) $ follows from the
$RTT = TTR$ relation which involves the trigonometric $R$-matrix:
\begin{align}
\bar{R}_{ab}(u/v)T_a(u)T_b(v) & = T_b(v)T_a(u))\bar{R}_{ab}(u/v), \\
\bar{R}(z) & := \frac{ zt - t^{-1} }{qz - q^{-1}}P.
\end{align}
The hierarchy of Dynamical Models defined by $\Delta (u)$ is $U_q'(\hat{{\frak g}{\frak l}}_2)$-invariant, that is
\begin{equation}
[ \Delta (u) , T_a(v) ] = 0 .
\end{equation}
This again follows from the mutual commutativity of the operators
$Y_1,Y_2,\dots,Y_N.$
Both $\Delta (u)$ and $ T_a(u) $ act in the ``bosonic'' subspace of $ {\Bbb C}[z_1,\dots,z_N]\otimes H $ as
explained in [BGHP]. This subspace which we denote by ${\cal B}$ is defined by the
requirement of the Hecke-invariance:
\begin{equation}
{\cal B} := \{ b \in {\Bbb C}[z_1,\dots,z_N]\otimes H | (g_{i,i+1} - t_{i,i+1})b = 0 \quad (i=1,\dots,N-1) \}.
\end{equation}
In any operator $O$ acting in ${\cal B}$ one can eliminate the coordinate exchange
operators $K_{i,j}$ by carrying them one-by-one to the right of any expression
in $O$ and replacing a $ K_{i,j} $ standing on the right of an expression in
accordance with the rule: $ g_{i,i+1} \rightarrow t_{i,i+1} \quad (i=1,\dots,N-1) $. This leads to a
uniquely defined operator $\widehat{O}$ which does not contain coordinate
permutations, such that
\begin{equation}
O{\cal B} = \widehat{O}{\cal B} .
\end{equation}
Where the notation means that the equality holds for any vector in ${\cal B}$.\\
With this definition the eq. (0.0.9,.12,.14) lead to
\begin{align}
[ \widehat{\Delta(u)} , \widehat{\Delta(v)}]{\cal B} & = 0 , \\
[ \widehat{\Delta (u)} , \widehat{T_a(v)} ]{\cal B} & = 0 , \\
(\bar{R}_{ab}(u/v)\widehat{T_a(u)}\widehat{T_b(v)} & -
\widehat{T_b(v)}\widehat{T_a(u)}\bar{R}_{ab}(u/v)){\cal B} = 0 .
\end{align}
The set of relations (0.0.17-.19) constitutes the result of [BGHP] concerned
witn the trigonometric Dynamical Models.
\subsubsection{The hierarchy of trigonometric Dynamical Models at $p=1$}
The first step of reduction from the Dynamical to the Spin Models consists in
taking the static limit $ p\rightarrow 1$ in the construction described in the
previous subsection. Following the idea of [TH] we expand the generating
function for the commuting charges of the hierarchy around the point $p=1$ up
to the linear term in $p-1$:
\begin{equation}
\Delta (u) = \Delta_0(u) + (p-1) \Delta_1(u) + O ( (p-1)^2).
\end{equation}
In the symmetry generator $ T_a(u) $ we retain the leading term only:
\begin{equation}
T_a(u) = T_a^0(u) + O(p-1).
\end{equation}
The operators $ \Delta_0(u) $ and $ \Delta_1(u) $ turn out to have a very special form.
First of all $ \Delta_0(u) $ is a {\em constant }:
\begin{equation}
\Delta_0(u) = \prod_{i=1}^N(1 + uq^{2i - N -1}).
\end{equation}
({\bf Remark} In the paper [TH] which deals with the rational case, the leading
term of the generating function
\begin{align*}
\Delta_{rational}(u) & := \prod_{i=1}^N(u - d_i) , \\
(d_i & := hD_i +
\sum_{j>i}\frac{z_i}{z_i-z_j}K_{i,j}-\sum_{j<i}\frac{z_j}{z_j-z_j}K_{i,j} \quad
(i=1,\dots,N) )
\end{align*}
in the static limit $ h\rightarrow 0 $ is not a constant but a function of
$z_1,\dots,z_N$. This is due to a choice of the Dunkl operators $d_i$ (above)
which do not act in the space of polynomials. The rational limit of the affine
Hecke generators $ Y_i $ which we use gives the gauge transformed Dunkl
operators acting in the space of polynomials ( Cf. [BGHP]).)
Since $\Delta_0(u)$ is a constant, the first-order differential operator $ \Delta_1(u) $
takes over the role of the generating function for Integrals of Motion:
\begin{align}
[\Delta_1(u) , \Delta_1(v) ]& = 0, \\
[\Delta_1(u) , T^0_a(v)] & = 0 .
\end{align}
Secondly, we find that $\Delta_1(u)$ has the following structure:
\begin{equation}
\Delta_1(u) = \sum_{i=1}^N \theta(u;z)_iD_i + \Xi(u;z) .
\end{equation}
Where $ \Xi(u;z) $ is a function of the operators $ z_1,\dots,z_N $ and $
K_{i,j} \quad (i,j = 1,\dots,N ) $ only. On the other hand the coefficients $
\theta(u;z)_i $ do not depend on the operators of coordinate permutation and
are functions of the coordinates $ z_1,\dots,z_N \quad (i=1,\dots,N) $ only. This kind
of separation of the differentials $ D_i $ and the operators of coordinate
permutation was first observed by [TH] in the rational case.
For the differential part $ {\cal D}(u) := \sum_{i=1}^N \theta(u;z)_iD_i $ of
$\Delta_1(u) $ we obtain an explicit expression in terms of the generating function $
D(u;p,t)$ for the Macdonald operators [M],[JKKMP] (Cf. {\bf 2.1}):
\begin{equation}
{\cal D}(u) = D_1(q^{N-1}u;q^{-2}) .
\end{equation}
Where $ D_1(u;t)$ is the linear term in the expansion of the generating
function $ D(u;p,t)$
around the point $ p=1$ :
\begin{equation}
D(u;p,t) = \Delta_0(u;t) + (p-1)D_1(u;t) + O((p-1)^2).
\end{equation}
The zero-order part $\Xi(u;z)$ of the differential operator $ \Delta_1(u) $ is the
object which we use to define the hierarchy of Integrable Spin Models.
First of all we observe that $ \Delta_1(u) $ lies in the centre of the Affine Hecke
Algebra generated by
$ g_{i,i+1} \quad (i=1,\dots,N-1) $ and $ y_i := Y_i|_{p=1} \quad (i=1,\dots,N) $. Therefore $ \Delta_1(u) $ acts
in the bosonic subspace ${\cal B}$ defined in the previous subsection, and in this
subspace we have analogs of the relations (0.0.17,-.19):
\begin{align}
[ \widehat{\Delta_1(u)} , \widehat{\Delta_1(v)}]{\cal B} & = 0 , \\
[ \widehat{\Delta_1(u)} , \widehat{T^0_a(v)} ]{\cal B} & = 0 , \\
(\bar{R}_{ab}(u/v)\widehat{T^0_a(u)}\widehat{T^0_b(v)} & -
\widehat{T^0_b(v)}\widehat{T^0_a(u)}\bar{R}_{ab}(u/v)){\cal B} = 0 .
\end{align}
The operator $\widehat{\Delta_1(u)}$ is a sum of two parts: the $ {\cal D}(u) $ which
is a first order differential operator and $ \widehat{\Xi(u;z)}$ which is a
matrix acting on $H$ whose entries are rational functions of the coordinates $
z_1,\dots,z_N .$ At the second step of the reduction from the Dynamical Models
to the Spin Models we shall eliminate the differential part of $\widehat{\Delta_1(u)}$
by restricting the coordinates to special values in a way which leaves the
Integrability and the $ U_q'(\hat{{\frak g}{\frak l}}_2) $-invariance intact.
\subsubsection{Definition of the hierarchy of trigonometric Spin Models}
The way to ``freeze'' the coordinates while keeping the spins as dynamical
variables goes trough the use of the {\em evaluation map } $Ev(v): {\Bbb C}[z_1,\dots,z_N]\otimes H
\mapsto H $. This map is parametrized by complex numbers $ v_1,\dots, v_N $ and
works by taking values of functions of $z_1,\dots,z_N $ at the point $ z_1 =
v_1 ,\dots,z_N = v_N $. We use this map at the special point $z = \omega : z_1
= \omega^1,\dots,z_N = \omega^N $ where $ \omega = e^{2\pi i /N}.$
In order to explain the relevance of this point we first of all observe, that
at this point the coefficients of the differential operator $ {\cal D}(u) $ are
all equal one to another:
\begin{equation}
\theta(u;\omega)_i = \theta(u) \qquad (i=1,\dots,N) .
\end{equation}
Where the constant $\theta(u)$ is given by eq. (2.2.26) in the main text. Next,
we observe, that in the expansion of $\widehat{\Delta_1(u)}$ in $u$: $ \widehat{\Delta_1(u)}:=
\sum_{n=1}^N u^n \Hat{\Delta}_1^{(n)} $ the term $\Hat{\Delta_1^{(N)}}$ is the
scale operator: $\Hat{\Delta}_1^{(N)} = D_1 + D_2 + \dots + D_N$. We can modify
the generating function $ \widehat{\Delta_1(u)}$ by subtracting from it the product of
the constant $\theta(u)$ and the operator $\Hat{\Delta}_1^{(N)}$. The equations
(0.0.28 -.30) clearly still hold for this modified generating function $
\widehat{\Delta_1(u)} - \theta(u)\Hat{\Delta}_1^{(N)} $. Moreover due to (0.0.31) for
any vector $ b \in {\cal B} $ we have:
\begin{equation}
Ev(\omega) (\widehat{\Delta_1(u)} - \theta(u)\Hat{\Delta}_1^{(N)})b = Ev(\omega)
\widehat{\Xi(u;z)}b .
\end{equation}
The map $Ev(\omega)$ naturally pulls through the operators $\widehat{\Xi(u;z)}$
and $ \widehat{T_a^0(u)} $ since these operators are matrices acting in $H$,
and the entries of these matrices are rational functions of $z \equiv (
z_1,\dots,z_N )$ non-singular at the point $z = \omega$. By pulling
$Ev(\omega)$ through $\widehat{\Xi(u;z)}$ and $ \widehat{T_a^0(u)}$ we define
operators $ \Xi(u;\omega) $ and $ T_a^0(u;\omega) $ acting in the image of $ Ev(\omega)
$ in $H$ by :
\begin{align}
Ev(\omega)\widehat{\Xi(u;z)}{\cal B} & = \Xi(u;\omega) Ev(\omega){\cal B} , \\
Ev(\omega)\widehat{T_a^0(u)}{\cal B} & = T_a^0(u;\omega) Ev(\omega){\cal B} .
\end{align}
Taking (0.0.31) and (0.0.33,.34) into account we apply the evaluation map to
the relations (0.0.28 - .30) and get:
\begin{align}
[ \Xi(u;\omega) , \Xi(v;\omega)]H_{{\cal B}}(\omega) & = 0 , \\
[ \Xi(u;\omega) , T^0_a(v;\omega) ]H_{{\cal B}}(\omega) & = 0 , \\
(\bar{R}_{ab}(u/v)T^0_a(u;\omega)T^0_b(v;\omega) & -
T^0_b(v;\omega)T^0_a(u;\omega)\bar{R}_{ab}(u/v))H_{{\cal B}}(\omega) = 0 .
\end{align}
Where $ H_{{\cal B}}(\omega):= Ev(\omega){\cal B} \subset H .$
Next, we prove, that $H_{{\cal B}}(\omega) = H $ .Therefore the relations (0.0.35
-.37) express Integrability and $ U_q'(\hat{{\frak g}{\frak l}}_2)$-invariance of the hierarchy of Spin
Chain Models. The Hamiltonians of these Models act in $H$ and are obtained by
expanding the generating function $ \Xi(u;\omega) $ in the parameter $u$:
\begin{equation}
\Xi(u;\omega) = \sum_{n=1}^{N-1} u^n \Xi^{(n)}(\omega).
\end{equation}
While the generating function $ \Xi(u;\omega) $ is completely defined by the relation
(0.0.25), the computation of an explicit expression is still quite a difficult
task. We have computed the explicit expression only for the first member of the
hierarchy -- the operator $ \Xi^{(1)}(\omega)$. To give this expression we
introduce several notations. For $ i\neq j \in \{1,\dots,N-1\}$ define the
rational functions:
\begin{equation}
a_{i,j} := \frac{q^{-1}z_i - qz_j}{z_i-z_j}\; , \quad b_{i,j} :=
\frac{(q-q^{-1})z_i}{z_i-z_j}.
\end{equation}
Define also the matrices:
\begin{multline*}
Y_{i,i+1}(w) := \frac{ wt_{i,i+1} - t_{i,i+1}^{-1} }{qw - q^{-1}}, \\
M^{(j,i)}(x,y) := \left( Y_{i+1,i+2}(y/z_{i+1})\dots
Y_{j-1,j}(y/z_{j-1})\right)^{-1}Y_{i,i+1}(x/y) \times \\ \times \left(
Y_{i+1,i+2}(x/z_{i+1})\dots Y_{j-1,j}(x/z_{j-1})\right) \quad (j > i).
\end{multline*}
With these notations we have:
\begin{multline}
\Xi^{(1)}(\omega) = \\
\sum_{M=2}^N \frac{(-1)^M}{\d} \sum_{N \geq i_M > \dots > i_1 \geq 1} {\cal
H}(\omega)_{i_M,i_{M-1},\dots,i_1} R(\omega)_{i_2,i_1}R(\omega)_{i_3,i_2 \dots
}R(\omega)_{i_M,i_{M-1}} + cI ; \\
{\cal H}(z)_{i_M,i_{M-1},\dots,i_1}:= \left( \prod_{i_1 < f <
i_2}a_{i_1,f}\right)\left( \prod_{i_2 < f < i_3}a_{i_2,f} \right)\dots
\left(\prod_{i_M < f < N + i_1}a_{i_M,f (\mod N)}\right) \times \\
\times b_{i_M,i_{M-1}} b_{i_{M-1},i_{M-2}}\dots b_{i_2,i_1} b_{i_1,i_M}, \\
R(z)_{j,i} := M^{(j,i)}(z_{i_M},z_i) \qquad (N \geq j > i \geq 1).
\end{multline}
Where $c$ is unimportant constant.
In the limit $q \rightarrow 1 $ we recover the Haldane Shastry Hamiltonian:
\begin{equation}
\lim_{q\rightarrow 1}\frac{\Xi^{(1)}(\omega)}{\d} = H_{HS} := -\sum_{N \geq j
> i \geq 1} \frac{\omega^i \omega^j}{ (\omega^i - \omega^j)^2 } (P_{i,j} - 1).
\end{equation}
While the Hamiltonian $\Xi^{(1)}(\omega)$ looks rather intimidating, its
spectrum, as well as the spectrum of the whole hierarchy generated by $\Xi(u;\omega)$
has a remarkably simple additive form. We describe this spectrum in the next
subsection.
\subsubsection{Eigenvalue spectrum of the hierarchy of Spin Hamiltonians}
As in the case of the Haldane Shastry Model the common eigenspaces of the
operators $\Xi^{(n)}(\omega)\quad (n=1,\dots,N-1)$ are in one-to-one
correspondence with ${\frak sl}_2$ motifs (Cf. [HHTBP] or [BPS]). A sequence
of integers $(m_1,m_2,\dots,m_M) $ is called an ${\frak sl}_2$ motif iff :
\begin{align}
& 1 \leq m_1 < m_2 < \dots < m_M \leq N-1 \, ; \tag{\theequation a.} \\
& m_{i+1} - m_i \geq 2 \quad ( i = 1,\dots , M-1 ) \tag{\theequation b.}.
\end{align}
For a fixed $N$ we denote the set of all $\frak{sl}_2$ motifs including the
empty one by ${\frak M}_N$. For the eigenspace of $\Xi(u;\omega)$ which correspons to a motif
$(m_1,m_2,\dots,m_M) $ we use the notation $ H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega)$. We have:
\begin{equation}
\Xi(u;\omega) H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega) = \left( \sum_{i=1}^M \xi^{(m_i)}(u)
\right)H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega).\end{equation}
Where the sum is understood to be zero for $ M=0$. \\
The elementary eigenvalue $\xi^{(m)}(u) \quad (m \in \{1,\dots,N-1\})$ is
\begin{equation}
\xi^{(m)}(u) = u\prod_{k=1}^N(1 + u q^{2k-N-1})\left\{\sum_{i=1}^m \frac{q^{2i
- N-1}}{1+ uq^{2i - N-1}} -\frac{m}{N} \sum_{i=1}^N \frac{q^{2i - N-1}}{1+
uq^{2i - N-1}} \right\}.
\end{equation}
In particular the elementary eigenvalue of the Hamiltonian $\Xi^{(1)}(\omega)$
is given by
\begin{equation}
\xi^{(m),(1)} = \frac{q^{-N}}{N}( Nq^m[m]_q - mq^N[N]_q ).
\end{equation}
Where we used the usual notation $ [x]_q \equiv \frac{q^x - q^{-x}}{\d}$. In
the limit $ q\rightarrow 1 $ we recover the elementary eigenvalue of the
Haldane Shastry Model [HHTBP,BPS]:
\begin{equation}
\lim_{q\rightarrow 1}\frac{\xi^{(m),(1)}}{\d} = m(m - N).
\end{equation}
The space of states of the Spin Models is represented as a direct sum of the
eigenspaces $ H^{(m_1,m_2,\dots,m_M)}(\omega)$:
\begin{equation}
H = \bigoplus_{(m_1,m_2,\dots,m_M) \in {\frak M}_N }H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega).
\end{equation}
\subsubsection{Structure of the common eigenspaces of the hierarchy of Spin
Hamiltonians.}
Each of the eigenspaces $ H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega)$ is an irreducible highest-weight
representation of $U_q'(\hat{{\frak g}{\frak l}}_2)$. The Drinfeld polynomial [CP] $Q^{(m_1,m_2,\dots,m_M)}(u)$ of $
H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega)$ is
\begin{equation}
Q^{(m_1,m_2,\dots,m_M)}(u) = \prod\begin{Sb}1\leq k \leq N \\ k \not\in \{m_i,m_i+1\} \end{Sb}
(1 - q^{-2k+N+1}u).
\end{equation}
(Cf. {\bf 1.2 } for our conventions about Drinfeld polynomials ).
To describe an eigenspace $ H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega) \quad (m_1,m_2,\dots,m_M) \in {\frak M}_N $ in a more
explicit way we give some facts (Cf. {\bf 4.1}) about eigenvectors of the
operator $ \Delta_1(u) $ which defines the hierarchy of Dynamical Models at $p=1$.
The linear space of polynomials ${\Bbb C}[z_1,\dots,z_N]$ is represented as a direct sum of
eigenspaces of the operator $\Delta_1(u) $. There is a one-to-one correspondence
between these eigenspaces and partitions with $ N$ parts. In an eigenspace
which corresponds to a partition $ \l := ( \l_1 \geq \l_2 \geq \dots \l_N \geq
0 ) $ there exists a basis $\{ \varphi_{\sigma}^{\l}(z) \}_{\sigma \in \S}$. The polynomials forming
this basis are indexed by elements from symmetric group $S_N$. There is a
one-to-one correspondence between the elements of the basis $\{ \varphi_{\sigma}^{\l}(z) \}_{\sigma
\in \S}$ and elements of the set $ \S \quad (\S \subset S_N )$ defined in {\bf
1.4}. A polynomial $ \varphi_{\sigma}^{\l}(z) \quad (\sigma \in \S) $ is completely specified by the
three conditions:\begin{align}
\Delta_1(u) \varphi_{\sigma}^{\l}(z) & = \left(\prod_{k=1}^N(1 + u q^{2k-N-1})\left\{\sum_{i=1}^N
\frac{uq^{2i-N-1}}{1 + uq^{2i-N-1}}\l_i \right\} \right) \varphi_{\sigma}^{\l}(z) ,
\tag{\theequation a.} \\
y_i \varphi_{\sigma}^{\l}(z) & = q^{2\sigma_i - N -1} \varphi_{\sigma}^{\l}(z) \quad (i=1,\dots,N) , \tag{\theequation b.} \\
\varphi_{\sigma}^{\l}(z) & = z_1^{\l_{\sigma_1}}z_2^{\l_{\sigma_2}} \dots z_N^{\l_{\sigma_N}} + \text{smaller
monomials}.\tag{\theequation c.}
\end{align}
We remind, that $y_i := Y_i|_{p=1} \quad (i=1,\dots,N) $. \\
The ``smaller monomials `` means a linear combination of monomials that are
smaller than the monomial $z_1^{\l_{\sigma_1}}z_2^{\l_{\sigma_2}} \dots
z_N^{\l_{\sigma_N}}$ in the ordering described in {\bf 1.4(2)} (Cf. [BGHP]).
With any motif $ (m_1,m_2,\dots,m_M) \in {\frak M}_N $ associate the partition
\begin{multline}
(\underset{1}{M},\dots,\underset{m_1}{M},\underset{m_1+1}{M-1},\dots
,\underset{m_2}{M-1},\underset{m_2+1}{M-2}, \ldots \\ \ldots,
\underset{m_M}{1},\underset{m_M+1}{0},\dots ,\underset{N}{0} )
\end{multline}
We use the same notation $ (m_1,m_2,\dots,m_M) $ for a motif and the associated partition.
For an element $\sigma $ of the symmetric group $S_N$ define
\begin{align}
\{\sigma_1 , \sigma_2 , \dots ,\sigma_N \} & := \sigma.\{1,2,\dots,N \} \, ; \tag{\theequation a.} \\
i & := \sigma_{p^{\sigma}_i} \qquad (i=1,\dots,N) . \tag{\theequation b.}
\end{align}
For any $(m_1,m_2,\dots,m_M) \in {\frak M}_N $ define the subset $S^{(m_1,m_2,\dots,m_M)}_{N,(m_1,m_2,\dots,m_M)}$ of $S_N$ (Cf. {\bf
4.3(2)}):
\begin{multline}
S^{(m_1,m_2,\dots,m_M)}_{N,(m_1,m_2,\dots,m_M)} := \\ \left\{ \sigma \in S_N \quad \begin{array}{| c c c } p^{\sigma}_i
< p^{\sigma}_{i+1} & ( m_k < i < m_{k+1}) & \text{for all}\; k \in
\{0,1,\dots,M\} \\
p^{\sigma}_{m_k} > p^{\sigma}_{m_k+1} & & \text{for all} \; k \in \{1,\dots,M-1\}
\end{array} \right\}.
\end{multline}
Where we adopt the convention: $ m_0 := 0 \, , m_{M+1} := N+1 .$
Next, for any motif $ (m_1,m_2,\dots,m_M) \in {\frak M}_N $ we define the subspace $W^{(m_1,m_2,\dots,m_M)}$ of the
space of states $H$ (Cf. {\bf 6.1(2) }) as follows:
\begin{multline}
W^{(m_1,m_2,\dots,m_M)}:= \\ S_q(\underset{1}{V}\otimes\dots\otimes\underset{m_1-1}{V})\otimes
A_q(\underset{m_1}{V}\otimes\underset{m_1+1}{V})\otimes
S_q(\underset{m_1+2}{V}\otimes\dots\otimes\underset{m_2-1}{V})\otimes
A_q(\underset{m_2}{V}\otimes\underset{m_2+1}{V})\otimes \ldots \\ \ldots
\otimes
S_q(\underset{m_{M-1}+2}{V}\otimes\dots\otimes\underset{m_M-1}{V})\otimes
A_q(\underset{m_M}{V}\otimes\underset{m_M+1}{V})\otimes
S_q(\underset{m_M+2}{V}\otimes\dots\otimes\underset{N}{V}) \\
W^{(m_1,m_2,\dots,m_M)} \subset H:= \underset{1}{V}\otimes\underset{2}{V}\otimes \ldots
\underset{N}{V}.
\end{multline}
Where $ S_q $ and $ A_q$ mean $q$-symmetrization and $q$-antisymmetrization as
defined in (1.1.14,.15,5.5.17).
The space $W^{(m_1,m_2,\dots,m_M)} \quad (m_1,m_2,\dots,m_M) \in {\frak M}_N$ is an irreducible highest weight
$U_q'(\hat{{\frak g}{\frak l}}_2)$-module. The Drinfeld polynomial of this module is given by (0.0.47). The
$U_q'(\hat{{\frak g}{\frak l}}_2)$-action on $W^{(m_1,m_2,\dots,m_M)}$ is given by
\begin{equation}
{\cal L}(u;\{q^{2\sigma[0]_i - N -1}\}):= L_{a1}(uq^{2\sigma[0]_1 - N
-1})L_{a2}(uq^{2\sigma[0]_2 - N -1})\dots L_{aN}(uq^{2\sigma[0]_N - N -1}).
\end{equation}
Where for a fixed $(m_1,m_2,\dots,m_M)$ we have introduced the notation
\begin{equation}
\sigma[0] := (m_1,m_1+1)(m_2,m_2+1)\dots(m_M,m_M+1) \in S_{N,(m_1,m_2,\dots,m_M)}^{(m_1,m_2,\dots,m_M)} \subset S_N.
\end{equation}
The eigenspace $ H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega) \quad ((m_1,m_2,\dots,m_M) \in {\frak M}_N) $ is isomorphic to $
W^{(m_1,m_2,\dots,m_M)}$ as $U_q'(\hat{{\frak g}{\frak l}}_2)$-module. This isomorphism is given by an invertible
intertwiner $ \check{\BU }^{(m_1,m_2,\dots,m_M)}(\omega)$
\begin{equation}
H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega) = \check{\BU }^{(m_1,m_2,\dots,m_M)}(\omega)W^{(m_1,m_2,\dots,m_M)}.
\end{equation}
The intertwiner $\check{\BU }^{(m_1,m_2,\dots,m_M)}(\omega)$ is defined by the expression
\begin{equation}
\check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega) := (-(q^2 + 1))^M\sum_{\sigma \in
S_{N,(m_1,m_2,\dots,m_M)}^{(m_1,m_2,\dots,m_M)}}\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(\omega){\Bbb Y}(\sigma) .
\end{equation}
In this definition
\begin{equation}
\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(\omega) = \varphi_{\sigma}^{\l}(z)|_{z_1=\omega^1,\dots,z_N=\omega^N}.
\end{equation}
Where the partition $ \l$ is the one specified by $ (m_1,m_2,\dots,m_M) $ in accordance with
(0.0.48).
The matrix ${\Bbb Y}(\sigma) \quad ( \sigma \in S_{N,(m_1,m_2,\dots,m_M)}^{(m_1,m_2,\dots,m_M)})$ is an intertwiner which is
defined by the recursion relations (Cf. {\bf 5.2(4)} ):
\begin{gather*}
{\Bbb Y}(\sigma[0]):= \text{Id} \; , \\
{\Bbb Y}((i,i+1)\sigma) := \begin{cases} Y^+_{i,i+1}(q^{2\sigma_i-2\sigma_{i+1}}){\Bbb Y}(\sigma) &
\text{ if $ \sigma_i - \sigma_{i+1} \geq 2 $}, \\
Y^-_{i,i+1}(q^{2\sigma_i-2\sigma_{i+1}}){\Bbb Y}(\sigma) & \text{ if $ \sigma_i - \sigma_{i+1} \leq -2
$ }.
\end{cases}
\end{gather*}
Where the matrices $ Y^{\pm}_{i,i+1}(w) $ are
\begin{align}
Y^{\pm}_{i,i+1}(w) & := \varrho^{\pm}(w)\frac{wt_{i,i+1} - t_{i,i+1}^{-1}}{q^{-1}w - q } \,
, \\
\varrho^+(w) & := \frac{w -1}{q^2 w -1} \;, \qquad \varrho^-(w) := \frac{w -
q^2}{ w -1}. \notag
\end{align}
Notice that ${\Bbb Y}(\sigma) \quad ( \sigma \in S_{N,(m_1,m_2,\dots,m_M)}^{(m_1,m_2,\dots,m_M)})$ is an invertible
intertwiner.
In general we cannot claim to know the explicit expression for the intertwiner
$\check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega)$ since we have not found the eigenvectors $
\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(z) $ explicitely. One exception is the case $q=0$ when
$\check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega)$ becomes very simple. In this case we have
\begin{equation}
\check{\BU}^{(m_1,m_2,\dots,m_M),q=0}(\omega)|_{W^{(m_1,m_2,\dots,m_M),q=0}} =
\omega^{\frac{1}{2}\sum_{i=1}^Mm_i(m_i + 1)}{\mathrm {Id}}.
\end{equation}
Therefore at $q=0$ the eigenspace $H_{{\cal B}}^{(m_1,m_2,\dots,m_M)}(\omega)$ is the linear span of
the following vectors in $H$:
\begin{multline*}
\left|\left\{\begin{array}{c c c c c
}\overset{1}{+}&\overset{}{+}&\overset{}{+}&\dots &\overset{m_1-1}{+} \\
-& +& +& \dots & + \\
-& -& +& \dots & + \\
\vdots & \vdots & \vdots & & \vdots \\
- & - & - & \dots & - \end{array} \right\} \right.
\overset{m_1}{+}\overset{m_1+1}{-}
\left\{\begin{array}{c c c c c
}\overset{m_1+2}{+}&\overset{}{+}&\overset{}{+}&\dots &\overset{m_2-1}{+} \\
-& +& +& \dots & + \\
-& -& +& \dots & + \\
\vdots & \vdots & \vdots & & \vdots \\
- & - & - & \dots & - \end{array} \right\}
\overset{m_2}{+}\overset{m_2+1}{-} \ldots \\ \ldots
\overset{m_M}{+}\overset{m_M+1}{-}
\left.\left\{\begin{array}{c c c c c
}\overset{m_M+2}{+}&\overset{}{+}&\overset{}{+}&\dots &\overset{N}{+} \\
-& +& +& \dots & + \\
-& -& +& \dots & + \\
\vdots & \vdots & \vdots & & \vdots \\
- & - & - & \dots & - \end{array} \right\} \right\rangle
\end{multline*}
Where we use the notation:
\begin{equation}
|\epsilon_1 \epsilon_2 \dots \epsilon_N \rangle := v^{\epsilon_1}\otimes
v^{\epsilon_2}\otimes \dots \otimes v^{\epsilon_N} \quad ( \epsilon_i = \pm )
{}.
\end{equation}
For example when $N$ is even, $H_{{\cal B}}^{(1,3,\dots,N-1)}(\omega)$ is
one-dimensional and at $q=0$ it is spanned by the vector (antiferromagnetic
ground state):
\begin{equation*}
|+ - + - \dots + - \rangle
\end{equation*}
\mbox{}
\noindent In the rest of the paper we give a detailed exposition of the matters
which were briefly recounted in this introduction. In the sec.1 we gather
predominantly known facts about the trigonometric Dynamical Models of [BGHP]
and explain the conventions about the algebra $ U_q'(\hat{{\frak g}{\frak l}}_2) $ that we use. In the sec.
2 we discuss the hierarchy of Dynamical Models in the static limit $ p=1$. In
the sec. 3 we define the hierarchy of the Spin Models. The sec. 4. is concerned
with properties of the eigenvectors of the hierarchy of Dynamical Models in the
limit $p = 1$. In the sec. 5 we construct the Hecke-invariant ``bosonic''
eigenspaces for the Dynamical Models with spin at $p=1$. The eigenvalue
spectrum of the Spin Models and the $U_q'(\hat{{\frak g}{\frak l}}_2) $-representation content of their
eigenspaces are derived in sec. 6.
\mbox{}
\noindent \begin{Large}{\bf Acknowledgments}\end{Large}\\ I am most grateful
to Drs. R. Kedem and R. Weston and to Professors M.Jimbo, M.Kashiwara and
T.Miwa for numerous discussions and support.
\section{The $U_q'(\hat{{\frak g}{\frak l}}_2)$-invariant Dynamical Models}
In this section we summarize largely known facts about the trigonometric
Dynamical Models defined by [BGHP]. We also recount several facts about the
algebra $U_q'(\hat{{\frak g}{\frak l}}_2)$ and explain our notations.
\subsection{The representations of the Affine Hecke Algebra}
\subsubsection{The representation of $ \widehat{H_N(q)} $ in the space of
polynomials}
Following [BGHP] define the operators $ g_{i,j} \in End( {\Bbb C}[z_1,\dots,z_N]) \quad (i,j
=1,\dots, N) $:
$$ g_{i,j} := a_{i,j}K_{i,j}+ b_{i,j} $$
where $$ a_{i,j} = \frac{q^{-1}z_i - qz_j}{z_i-z_j}\;, \qquad b_{i,j} =
\frac{(q-q^{-1})z_i}{z_i-z_j} = q - a_{i,j}. $$ $ K_{i,j} $ is the interchange
operator for variables $ z_i , z_j $.
For $ p \in {\Bbb C} $ and $ D_i := z_i \frac{\partial}{\partial z_i} $ define
operators $ Y_i \in End( {\Bbb C}[z_1,\dots,z_N]) (i=1,\dots,N) $:
\begin{eqnarray}
Y_i & := & g_{i,i+1}^{-1}K_{i,i+1}\dots
g_{i,N}^{-1}K_{i,N}p^{D_i}K_{1,i}g_{1,i}\dots K_{i-1,i}g_{i-1,i}
\end{eqnarray}
Taken together with $ { g_{i,i+1} }\; (i=1,\dots, N-1 )$ these operators satisfy the
relations of the Affine Hecke Algebra $ \widehat{H_N(q)} $ [BGHP] :
\begin{eqnarray}
g_{i,i+1}^2 & = & (q-q^{-1})g_{i,i+1} + 1 \\
g_{i,i+1} g_{k,k+1} & = & g_{k,k+1} g_{i,i+1} \; , \qquad |i-k| \geq 2 \\
g_{i,i+1} g_{i+1,i+2}g_{i,i+1} & = & g_{i+1,i+2}g_{i,i+1} g_{i+1,i+2} \\
Y_k g_{i,i+1} & = & g_{i,i+1} Y_k \; , \qquad k \neq i,i+1 \\
g_{i,i+1} Y_i & = & Y_{i+1} g_{i,i+1}^{-1} \\
Y_i Y_j & = & Y_j Y_i \: .
\end{eqnarray}
Any symmetric polynomial in $ Y_i \; (i=1,\dots,N) $ belongs to the center of $
\widehat{H_N(q)} $. All symmetric polynomials in $ Y_i \; (i=1,\dots,N) $
are generated by the elementary symmetric polynomials which are obtained by
expanding in the parameter $u$ the generating function $ \Delta(u) $ :
\begin{eqnarray}
\Delta(u) & = & \prod_{(i=1,\dots,N) }(1 + uY_i)
\end{eqnarray}
\subsubsection{The representation of the Hecke Algebra $ H_N(q) $ in $
({\Bbb C}^2)^{\otimes N} $ }
Let $ V := {\Bbb C}^2 = {\rm span} \{ v^+,v^- \} $. Define $ t \in End(V\otimes V)
$ by
\begin{eqnarray}
t v^+\otimes v^- & = & (q-q^{-1})v^+\otimes v^- + v^-\otimes v^+ \\
t v^-\otimes v^+ & = & v^+\otimes v^- \\
t v^\pm\otimes v^\pm & = & qv^\pm\otimes v^\pm
\end{eqnarray}
The operators $ \Pi^\pm(q) $:
\begin{eqnarray}
\Pi^+(q) & := & \frac{ q^{-1} + t }{q + q^{-1} } \\
\Pi^-(q) & := & \frac{ q - t }{q + q^{-1} }
\end{eqnarray}
are orthogonal projectors on the subspaces:
\begin{eqnarray}
S_q( V\otimes V ) & := & {\Bbb C} \{ v^+\otimes v^+ , v^-\otimes v^- , qv^+\otimes
v^- + v^-\otimes v^+ \} \\
A_q( V\otimes V ) & := & {\Bbb C} \{ v^+\otimes v^- - q v^-\otimes v^+ \}
\end{eqnarray}
respectively.
Let $ H := V^{\otimes N}$. For an $ O \in End(V\otimes V) $ denote by $
O_{i,j} \in End(H) $ the standard injection $ End(V\otimes V) \rightarrow
End(H) $ which acts trivially on all the factors except the $i$-th and $j$-th
ones.
The matrices $ t_{i,i+1} \quad (i=1,\dots, N-1) $ satisfy the defining
relations (1.1.2-.4) of the finite-dimensional Hecke Algebra $ H_N(q) $.
\subsection{The Algebra $ U_q'(\hat{{\frak g}{\frak l}}_2) $ at level 0 and some of its representations}
In this subsection we summarize several facts about the algebra $U_q'(\hat{{\frak g}{\frak l}}_2)$ and its
representations. Our conventions and notations mainly follow [JKKMP].
\subsubsection{The Algebra $ U_q'(\hat{{\frak g}{\frak l}}_2) $ }
In the $L$-operator formalism $ U \equiv U_q'(\hat{{\frak g}{\frak l}}_2) $ at zero level is defined to be
the associative algebra with unit generated by
elements $ l_{ij}^\pm[\pm n]\;(i,j = 1,2 \:; n = 0,1,\dots) $.The $L$-operators
$ L^\pm(u) \in End(V\otimes U) $ are the generating series in the spectral
parameter $u$:
\begin{eqnarray}
L^\pm(u) & := & \sum_{\pm n \geq 0 } u^n \left( \begin{array}{c c} l_{11}[n] &
l_{12}[n] \\
l_{21}[n] &
l_{22}[n]
\end{array}
\right)
\end{eqnarray}
The defining relations of $U$ are written in the the form:
\begin{eqnarray}
\bar{R}_{ab}(u/v)L^\pm_a(u)L^\pm_b(v) & = &
L^\pm_b(v)L^\pm_a(u)\bar{R}_{ab}(u/v) \\
\bar{R}_{ab}(u/v)L^+_a(u)L^-_b(v) & = & L^-_b(v)L^+_a(u)\bar{R}_{ab}(u/v)
\\
l_{ii}^+[0]l_{ii}^-[0]\; = \; 1 \quad (i=1,2)\:, & & l_{21}^+[0] \: = \:
l_{12}^-[0] \: = \: 0
\end{eqnarray}
where the $R$-matrix $ \bar{R}_{ab}(z) \in End(V\otimes V := V_a \otimes V_b )
$ is defined as follows:
\begin{eqnarray}
\bar{R}(z) & = & \frac{ zt - t^{-1} }{qz - q^{-1}}P
\end{eqnarray}
by $P$ we denote the permutation operator in $ V\otimes V $.
\subsubsection{Some representations of $U$}
A finite-dimensional highest weight module $W$ of $U$ contains non-zero vector
$\Omega$ which satisfies the condition:
\begin{eqnarray}
L^\pm(u)\Omega & = & \left( \begin{array}{c c}
A^\pm(u) & * \\
0 & D^\pm(u)
\end{array} \right) \Omega
\end{eqnarray}
where $ A^\pm(u)$ and $ D^\pm(u)$ are ${\Bbb C}$-valued series in $u$.
If $W$ is irreducible, it is specified up to equivalence by its Drinfeld
polynomial $Q(u)$ whch is determined by the conditions: $ Q(0) \: = \: 1 $ and
\begin{eqnarray}
q^{degQ}\frac{Q(q^{-2}u)}{Q(u)} & = & \frac{A^+(u^{-1})}{D^+(u^{-1})}
\qquad ( u \rightarrow 0 ) \\
& = & \frac{A^-(u^{-1})}{D^-(u^{-1})}
\qquad ( u \rightarrow \infty )
\end{eqnarray}
The example of such $W$ is the 2-dimensional evaluation module $ W(a) $ where $
a \in {\Bbb C} \backslash \{0\} $ is the parameter. As a vector space $W(a)$ is
isomorphic to $V$. The generators $ l_{ij}^\pm[\pm n]\;(i,j = 1,2 ; n =
0,1,\dots) $ are defined by expanding the $L$-operator:
\begin{eqnarray}
L(ua) & := & \frac{uat - t^{-1}}{ua - 1}P \; \in End(V_a \otimes V)
\end{eqnarray}
into power series in $ u^\pm $ around zero and infinity respectively.
The Drinfeld polynomial of $W(a)$ is: $ Q(u;a) \; = \; 1 - a^{-1}u .$
A pair of tensor products $W(a)\otimes W(b)$, $W(b)\otimes W(a)$ is intertwined
by the matrix
\begin{eqnarray}
\bar{Y}(z) & = & zt - t^{-1} \quad ( z \in {\Bbb C} ) \; \in End(V\otimes V )
\end{eqnarray}
i.e.:
\begin{eqnarray}
\bar{Y}_{12}(a/b)L_{a1}(ua)L_{a2}(ub) & = &
L_{a1}(ub)L_{a2}(ua)\bar{Y}_{12}(a/b)
\end{eqnarray}
this relation holds in the tensor product of an auxiliary copy of $V$ indicated
by the subscript $a$ and $ V \otimes V $ indicated by subscripts 1 and 2.
The intertwiner $\bar{Y}(a/b)$ is invertible unless either $ a = q^2b $, in
which case
\begin{eqnarray}
\bar{Y}(q^2) & = & (q^2 - 1)(q+q^{-1}) \Pi^+(q)
\end{eqnarray}
or $ a = q^{-2}b $, in which case
\begin{eqnarray}
\bar{Y}(q^{-2}) & = & (1-q^{-2})(q+q^{-1}) \Pi^-(q)
\end{eqnarray}
Together with (26) this leads to the invariance relations:
\begin{eqnarray}
L_{a1}(q^2u)L_{a2}(u):A_q(V\otimes V) & \subset & A_q(V\otimes V) \\
L_{a1}(q^2u)L_{a2}(u):S_q(V\otimes V) & \subset & A_q(V\otimes V)\oplus
S_q(V\otimes V)\\
L_{a1}(u)L_{a2}(q^2u):S_q(V\otimes V) & \subset & S_q(V\otimes V) \\
L_{a1}(u)L_{a2}(q^2u):A_q(V\otimes V) & \subset & A_q(V\otimes V)\oplus
S_q(V\otimes V)
\end{eqnarray}
\subsection{The hierarchy of $U$-invariant Dynamical Models}
The central elements of $\widehat{H_N(q)}$ generated by $\Delta(u)$ were
proposed in [BGHP] to define the hierarchy of integrable Dynamical Models which
are trigonometric - that is $U$-invariant - generalizations of the
Yangian-invariant Dynamical Models found by the same authors.
Define $ T_a(u) \in End({\Bbb C}[z_1,\dots,z_N]\otimes H) $ by taking the tensor product of the
$L$-operators (24):
\begin{eqnarray}
T_a(u)&=& L_{a1}(uY_1)L_{a2}(uY_2)\dots L_{aN}(uY_N)
\end{eqnarray}
After expansion in $u^\pm$, $ T_a(u) $ gives rise to a representation of $U$ in
$ {\cal P} \: := \: {\Bbb C}[z_1,\dots,z_N] \otimes H $.
The action of $\Delta(u)$ naturally extends to $ {\cal P}$. Retain the same
notation $\Delta(u)$ for this extension.
The $U$-invariance and integrability relations for $\Delta(u)$ are immediate:
\begin{eqnarray}
[ \Delta(u) , T_a(v) ] & = & 0 \\
\mbox{} [ \Delta(u) , \Delta(v) ] & = & 0
\end{eqnarray}
Both $\Delta(u)$ and $ T_a(v)$ act in the (``bosonic'') subspace ${\cal B}$:
\begin{eqnarray}
{\cal B} & := & \{ b \in {\cal P} | g_{i,i+1} b = t_{i,i+1}b \; (i=1,\dots,N-1) \}
\end{eqnarray}
This allows to restrict $\Delta(u)$ and $ T_a(v)$ on ${\cal B}$ where both of these
operators can be rewritten in such a way that they do not depend explicitely on
the operators $K_{i,j}$.
\subsection{Eigenvalue spectrum of the operators $ Y_i \; (i=1,\dots,N) $ }
{\bf 1.} We shall work in the monomial basis of $ {\Bbb C}[z_1,\dots,z_N] $.
Introduce a convenient parametrization of the elements of this basis. With a
monomial $ z^{\nu} := z_1^{\nu_1}z_2^{\nu_2}\dots z_N^{\nu_N} \quad (\nu_i \in
{\Bbb Z}^+ (i=1,\dots,N) ) $ associate a partition $ \l := (\l_1 \geq \l_2 \geq \dots
\geq \l_N \geq 0 ) $ such that $ \{\nu_1,\nu_2,\dots,\nu_N \} = \{
\l_{\sigma_1},\l_{\sigma_2},\dots,\l_{\sigma_N} \}:= \sigma.\{ \l_1,\l_2,\dots,\l_N \} $ for
some $ \sigma \in S_N = $ symmetric group with $ N-1 $ generators. A partition
$ \l $ is uniquely specified by $ \nu $, while in general $ \sigma $ is not.
For $ \sigma \in S_N : \{ \sigma_1,\sigma_2,\dots , \sigma_N \} = \sigma.\{1,2,\dots,N \} $ define
$ p^{\sigma}_i \; : \; i = \sigma_{p^{\sigma}_i} \quad (i=1,\dots,N) .$ Let $ \Lambda_N $ be the
set of all {\it N-}member partitions. For any $\l \in \Lambda_N $ write: $\l =
(\l_1=\l_2=\dots =\l_{m_1} > \l_{m_1+1}=\l_{m_1+2}=\dots =\l_{m_2} > \ldots
\\ \ldots > \l_{m_M+1}=\l_{m_M+2}=\dots=\l_N )$.
\begin{df}
For any $ \l \in \Lambda_N $:
\begin{align*}
\S := \{ \sigma \in S_N \; | \; p^{\sigma}_1 < p^{\sigma}_2 < \dots p^{\sigma}_{m_1} , \;
p^{\sigma}_{m_1+1} < p^{\sigma}_{m_1+2} < \dots p^{\sigma}_{m_2}, \ldots \\ \ldots ,
p^{\sigma}_{m_M+1} < p^{\sigma}_{m_M+2} < \dots p^{\sigma}_N \}.
\end{align*}
\end{df}
For a given $\l$ the elements of the set $ \S $ are in one-to-one
correspondence with distinct rearrangements of the sequence $ \{
\l_1,\l_2,\dots,\l_N \}.$ If $ \sigma \in \S $ then $ (i\,i+1)\sigma \in \S $ iff $
\l_{\sigma_i} \neq \l_{\sigma_{i+1}} \quad (i=1,\dots,N-1) .$ The particular choice of
$ \S $ as a subset of $ S_N $ parametrizing distinct rearrangements of a
partition will be explained by the Proposition 1.
Denote $ z^{\l_{\sigma}} := z_1^{\l_{\sigma_1}}z_2^{\l_{\sigma_2}}\dots z_N^{\l_{\sigma_N}}
\quad (\l \in \Lambda_N\:, \; \sigma \in \S ).$ With this notation:
\begin{eqnarray}
{\Bbb C}[z_1,\dots,z_N] = \bigoplus_{ \l \in \Lambda_N } \oplus_{ \sigma \in \S } {\Bbb C} z^{\l_{\sigma}}.
\end{eqnarray}
{\bf 2.} Introduce an ordering on the set of monomials $ \{ z^{\l_{\sigma}} \} \quad ( \l
\in \Lambda_N,\; \sigma \in \S ) .$
We say that $ \l > \tilde{\l} \quad ( \l,\tilde{\l} \in \Lambda_N ) $ iff the
first ( counting from left ) non-vanishing element of the sequence $ \{ \l_1 -
\tilde{\l}_1, \l_2 - \tilde{\l}_2, \dots, \l_N - \tilde{\l}_N \} $ is
positive. Fix $ \l \in \Lambda_N .$ We say that $ \sigma > \tilde{\sigma}\quad ( \sigma,
\tilde{\sigma} \in \S )$ iff the last non-vanishing element of the sequence $ \{
\l_{\sigma_1} - \l_{\tilde{\sigma}_1}, \l_{\sigma_2} - \l_{\tilde{\sigma}_2}, \dots \l_{\sigma_N}
- \l_{\tilde{\sigma}_N} \}$ is negative.
For $ \l , \tilde{\l} \in \Lambda_N \:;\; \sigma \in \S , \: \tilde{\sigma} \in
S^{\tilde{\l}}_N $ define $ \l_{\sigma} > \tilde{\l}_{\tilde{\sigma}} $ iff either $ \l
> \tilde{\l} \:,$ or $ \l = \tilde{\l}\:, \sigma > \tilde{\sigma} $. The ordering on
monomials $ z^{\l_{\sigma}} $ is induced by the ordering on the exponents $ \l_{\sigma} \quad
( \l \in \Lambda_N,\; \sigma \in \S ) .$
\mbox{}
\noindent {\bf 3.} The action of $ H_N(q) $ in the monomial basis is found by a
straightforward computation to be as follows:
\begin{equation} g_{i,j} z^{\l_{\sigma}} = ( i < j ) =
\begin{cases}
(q-q^{-1}) z^{\l_{\sigma}} + qz^{\l_{(ij)\sigma}} + \text{``s.p.''} & \text{ if $\l_{\sigma_i} >
\l_{\sigma_j}$ }, \\
qz^{\l_{\sigma}} & \text{ if $\l_{\sigma_i} = \l_{\sigma_j}$ }, \\
q^{-1}z^{\l_{(ij)\sigma}} + \text{``s.p.''} & \text{ if $\l_{\sigma_i} < \l_{\sigma_j}$ }.
\end{cases}
\end{equation}
And
\begin{equation} g_{i,j}^{-1} z^{\l_{\sigma}} = (i < j ) =
\begin{cases}
qz^{\l_{(ij)\sigma}} + \text{``s.p.''} & \text{ if $\l_{\sigma_i} > \l_{\sigma_j}$ },\\
q^{-1}z^{\l_{\sigma}} & \text{ if $\l_{\sigma_i} = \l_{\sigma_j}$ }, \\
(q-q^{-1}) z^{\l_{\sigma}} + q^{-1}z^{\l_{(ij)\sigma}} + \text{``s.p.''} & \text{ if
$\l_{\sigma_i} < \l_{\sigma_j}$ }.
\end{cases}
\end{equation}
Where ``s.p.'' means a linear combination of monomials with smaller partitions.
It follows that:
\begin{equation} K_{i,j}g_{i,j} z^{\l_{\sigma}} = ( i < j ) =
\begin{cases}
qz^{\l_{\sigma}} + \text{``s.m.''} & \text{ if $\l_{\sigma_i} \geq \l_{\sigma_j}$ }, \\
q^{-1}z^{\l_{\sigma}} + \text{``s.m.''} & \text{ if $\l_{\sigma_i} < \l_{\sigma_j}$ }.
\end{cases}
\end{equation}
And
\begin{equation} g_{i,j}^{-1}K_{i,j} z^{\l_{\sigma}} = ( i < j ) =
\begin{cases}
q^{-1}z^{\l_{\sigma}} + \text{``s.m.''} & \text{ if $\l_{\sigma_i} \geq \l_{\sigma_j}$ }, \\
qz^{\l_{\sigma}} + \text{``s.m.''} & \text{ if $\l_{\sigma_i} < \l_{\sigma_j}$ }.
\end{cases}
\end{equation}
Where ``s.m.'' signifies a linear combination of smaller monomials.
\mbox{}
\noindent {\bf 4.} The formulas of the preceding paragraph lead to the
following proposition:
\begin{prop}
The operators $ Y_i \; (i=1,\dots,N) $ are triangular in the monomial basis of $ {\Bbb C}[z_1,\dots,z_N] $.
The action of these operators on monomials is given by:
\begin{equation*}
Y_iz^{\l_{\sigma}} = p^{\l_{\sigma_i}}q^{l_{\sigma_i}}z^{\l_{\sigma}} + \text{{\em``s.m.''}} \quad ( \l \in
\Lambda_N,\; \sigma \in \S,\; (i=1,\dots,N) )
\end{equation*}
where $ l_i := 2i-N-1 \quad (i=1,\dots,N) $.
\end{prop}
This proposition shows that $ \zeta_i^{\l}(\sigma) := p^{\l_{\sigma_i}}q^{l_{\sigma_i}}
\quad (\l \in \Lambda_N,\; \sigma \in \S) $ constitute a complete set of
characteristic numbers of the operator $ Y_i \quad (i \in \{ 1,2,\dots,N \}).$
In order to prove that the operators $ Y_i \; (i=1,\dots,N) $ are simultaneously
diagonalizable and that $ \{\zeta_i^{\l}(\sigma)\}\quad (\l \in \Lambda_N,\; \sigma \in
\S) $ form the complete set of eigenvalues of $ Y_i \quad (i \in \{ 1,2,\dots,N
\}) $ we shall make use of Lemma 1 discussed in the next paragraph.
\mbox{}
\noindent {\bf 5.} The aim of this paragraph is to recall the following (
presumably well-known ) result:
\begin{lemma}
Let $ {\cal V} = {\Bbb C}\{f_a\}_{a = 1,2,\dots,d={\mathrm {dim}}{\cal V}} $ be a
finite-dimensional vector space. Let $ {\cal Z}_i \in End({\cal V})\quad (i=1,2,\dots ,
{\cal N}) $ and :
\begin{align}
[{\cal Z}_i,{\cal Z}_j] & = 0 \quad (i,j = 1,2,\dots, {\cal N}), \tag{a} \\
{\cal Z}_i \quad (i=1,2,\dots , {\cal N}) \quad & \text{are simultaneously triangular
in the basis}\: \{f_a\}_{a = 1,2,\dots,d}: \notag \\
{\cal Z}_i f_a & = \xi_i^a f_a + \sum_{b < a} m_i^{b\,a} f_b \quad (i=1,2,\dots
, {\cal N}), \tag{b} \\
& \text{where $ m_i^{b\,a} $ are coefficients.} \notag
\end{align}
{\em (c)} The joint set of characteristic numbers $ \{ \xi_i^a \} \quad
(i=1,2,\dots,{\cal N} \:;\; a = 1,2,\dots,d ) $ is multiplicity-free:
\begin{align}
\forall\; a \neq b \; ( a,b = 1,2,\dots,d )\; & \; \exists \; I(a,b) \subset
\{1,2,\dots,{\cal N}\}: \notag \\
\forall \; i \in I(a,b) \; & \xi_i^a - \xi_i^b \neq 0 . \notag
\end{align}
\noindent Then $ \exists $ a basis $ \{ \phi_a \}_{a=1,2,\dots,d} $ :
\begin{align}
{\cal Z}_i \phi_a & = \xi_i^a \phi_a \quad ( a=1,2,\dots,d ;\: i=1,2,\dots,{\cal N} )
\notag \\
\phi_a & = f_a + \sum_{b < a } \phi_{b\,a} f_b \quad ( a=1,2,\dots,d ) \notag
\end{align}
Where the coefficients $ \phi_{b\,a} $ are recursively defined as follows:
\begin{align}
\phi_{b\,a} & = \frac{1}{ \xi^a(w) - \xi^b(w) } ( m_{b\,a}(w) + \sum_{b < c <
a}m_{b\,c}(w)\phi_{c\,a}), \notag \\
\text{here} \quad \xi^a(w) & := \sum_{i=1}^{{\cal N}}w^{i-1}\xi_i^a \,, \qquad
m_{a\,b}(w) \; := \; \sum_{i=1}^{{\cal N}}w^{i-1} m_i^{a\,b}\,. \notag
\end{align}
$ w \in {\Bbb C} $ and $ \phi_{b\,a} $ does not depend on w.
\end{lemma}
\mbox{}
\noindent {\bf 6.} The joint characteristic number spectrum $ \{ \zeta_i^{\l}(\sigma) =
p^{\l_{\sigma_i}}q^{l_{\sigma_i}} \}\quad ((i=1,\dots,N) ; \; $ \mbox{} $ \l \in \Lambda_N ; \;
\sigma \in \S ) $ of the operators $ Y_i \quad (i=1,\dots,N) $ is explicitely
multiplicity-free.The operators $ g_{i,j} \quad ( i,j = 1,2,\dots,N-1 ) $ preserve
the finite-dimensional subspaces of $ {\Bbb C}[z_1,\dots,z_N] $ formed by homogeneous polynomials
of any total degree. Therefore $ g_{i,j} \quad ( i,j = 1,2,\dots,N-1 ) $ and
consequently $ Y_i \quad (i=1,\dots,N) $ are direct sums of finite dimensional
operators. So we can apply the result of Lemma 1 and arrive at the following
proposition:
\begin{prop}
There exist polynomials $ \Phi^{\l}_{\sigma}\quad (\l \in \Lambda_N ; \; \sigma \in \S
) $ s.t.:
\begin{align}
{\Bbb C}[z_1,\dots,z_N] & = \bigoplus_{\l \in \Lambda_N}E^{\l} \; , \;
E^{\l}\::=\;\oplus_{\sigma \in \S} {\Bbb C}\Phi^{\l}_{\sigma} \, , \tag{i} \\
Y_i \Phi^{\l}_{\sigma} & = \zeta_i^{\l}(\sigma) \Phi^{\l}_{\sigma} \quad (i=1,\dots,N) , \tag{ii} \\
\Delta(u) \Phi^{\l}_{\sigma} & = \prod_{i=1}^N (1 + up^{\l_i}q^{l_i})
\Phi^{\l}_{\sigma} \,, \tag{ii'} \\
\Phi^{\l}_{\sigma} & = z^{\l_{\sigma}} + \text{{\em``s.m.''}}. \tag{iii}
\end{align}
\end{prop}
\subsection{Action of the Hecke Algebra in the eigenspaces $E^{\l}\quad ( \l
\in \Lambda_N )$ of the operator $\Delta (u)$.}
\noindent {\bf 1.} Fix any $ \l \in \Lambda_N $ . Since $[\widehat{H_N(q)}\:,\:
\Delta (u)]\:=\: 0$ with the action of $\widehat{H_N(q)}$ defined in {\bf 1.1.1}, we
have:\begin{equation*}
\widehat{H_N(q)}: \: E^{\l} \rightarrow E^{\l}
\end{equation*}
The affine generators of $ \widehat{H_N(q)} $ act in $ E^{\l} $ as given by
(ii) in Proposition 2. In this section we find the action of the
finite-dimensional Hecke Algebra generated by $ g_{i,i+1} \quad (i=1,...,N-1) $ on
the polynomials $\Phi_{\sigma}^{\l} \quad ( \sigma \in \S ) $ forming a basis in $ E^{\l} \quad (
\l \in \Lambda_N ). $
\noindent {\bf 2.} Introduce operators $ T_{i,i+1}\; \in \; \widehat{H_N(q)}
\quad ( i=1,\dots,N-1 ) $ :
\begin{equation}
T_{i,i+1} := g_{i,i+1} (Y_{i+1} - Y_i ) - (q-q^{-1}) Y_{i+1} \quad (i=1,\dots,N-1).
\end{equation}
These operators satisfy the following relations:
\begin{align}
T_{i,i+1}Y_i \; = \; Y_{i+1} T_{i,i+1}\:, & \quad T_{i,i+1}Y_{i+1} \; = \;
Y_i T_{i,i+1} \quad (i=1,\dots,N-1) \\ \mbox{} [ T_{i,i+1}, Y_j ]& = 0 \qquad ( j
\neq i,i+1)
\end{align}
Since $ T_{i,i+1}\; \in \; \widehat{H_N(q)} $, we have: $ T_{i,i+1}: E^{\l}
\rightarrow E^{\l} \quad ( i=1,\dots,N-1 ). $
\mbox{}
\noindent {\bf 3.} Consider the vector $ T_{i,i+1}\Phi_{\sigma}^{\l} \in E^{\l} \quad ( \sigma \in
\S \; , \; i \in {1,2,\dots,N-1} ). $ Due to (1.1.43,44):
\begin{eqnarray*}
Y_k T_{i,i+1}\Phi_{\sigma}^{\l} & = & \xi_k^{\l}((i,i+1)\sigma) T_{i,i+1}\Phi_{\sigma}^{\l} \quad ( k =
1,\dots,N ).
\end{eqnarray*}
Since the spectrum of $ Y_k \quad ( k = 1,\dots,N ) $ on $ E^{\l}$ does not
contain $ \xi_k^{\l}(\sigma) $ s.t. $ \sigma \not\in \S $, we have:
\begin{align}
T_{i,i+1}\Phi_{\sigma}^{\l} & = 0 \quad \text{when $ \sigma \in \S\,, (i,i+1)\sigma \not\in \S $}
\quad (i=1,\dots,N-1).
\end{align}
Since the joint spectrum of $ Y_i \; (i=1,\dots,N) $ on $ E^{\l}$ is multiplicity-free,
we have:
\begin{align}
T_{i,i+1}\Phi_{\sigma}^{\l} & = \tau_{i,i+1}^{\l}(\sigma) \Phi_{(i,i+1)\sigma}^{\l} \quad \text{when $ \sigma , (i,i+1)\sigma
\in \S $} \quad (i=1,\dots,N-1).
\end{align}
Where $\tau_{i,i+1}^{\l}(\sigma)$ is a coefficient.
Let $ \sigma \in \S $ and $ i $ be s.t. $ (i,i+1)\sigma \not\in \S $. Then $ \l_{\sigma_i}
= \l_{\sigma_{i+1}} \;,\; \sigma_{i+1} = \sigma_i + 1 $, and (1.1.45) gives:
\begin{align}
(g_{i,i+1} - q )\Phi_{\sigma}^{\l} & = 0 .
\end{align}
Let $ \sigma \in \S $ and $ i $ be s.t. $ (i,i+1)\sigma \in \S $. Then we can recast
(1.1.46) as follows:
\begin{align}
g_{i,i+1}\Phi_{\sigma}^{\l} & = \frac{(q-q^{-1})\zeta_{i+1}^{\l}(\sigma)}{\zeta_{i+1}^{\l}(\sigma) -\zeta_i^{\l}(\sigma)}\Phi_{\sigma}^{\l} + \frac{
\tau_{i,i+1}^{\l}(\sigma)}{\zeta_{i+1}^{\l}(\sigma)-\zeta_i^{\l}(\sigma)}\Phi_{(i,i+1)\sigma}^{\l}\, .
\end{align}
\noindent In order to find $\tau_{i,i+1}^{\l}(\sigma)$ we shall equate
coefficients standing in front of monomials $ z^{\l_{\sigma}} \, ,\; z^{\l_{(i\,i+1)\sigma}} $ in the both
sides of eq. (1.1.48). Recall that $ \sigma ,\, (i,i+1)\sigma \in \S $ entails in
particular $ \l_{\sigma_i} \neq \l_{\sigma_{i+1}} $.
\noindent Let $ \l_{\sigma_i} > \l_{\sigma_{i+1}}\,\Rightarrow\, z^{\l_{\sigma}} > z^{\l_{(i\,i+1)\sigma}} $.
\noindent According to (1.1.38) the monomial $ z^{\l_{\sigma}} $ which is the maximal
monomial in $\Phi_{\sigma}^{\l}$ appears in $g_{i,i+1}\Phi_{\sigma}^{\l}$ from two sources: from $\giz^{\l_{\sigma}}$ and
from $\giz^{\l_{(i\,i+1)\sigma}}$. Denote by $x$ the coefficient at the monomial $z^{\l_{(i\,i+1)\sigma}}$ in
$\Phi_{\sigma}^{\l}$.
\noindent In the RHS of (1.1.48) $z^{\l_{\sigma}}$ appears as the maximal monomial in
$\Phi_{\sigma}^{\l}$ and does not appear in $\Phi_{(i,i+1)\sigma}^{\l}$.
Using (1.1.38) to compute the contributions from $g_{i,i+1}\Phi$ we equate the
coefficients in front of $z^{\l_{\sigma}}$ in the both sides of (1.1.48):
\begin{align}
\d + q^{-1}x & = \frac{\d\zeta_{i+1}^{\l}(\sigma)}{\zeta_{i+1}^{\l}(\sigma) - \zeta_i^{\l}(\sigma)} .
\end{align}
Computing the contribution from $z^{\l_{(i\,i+1)\sigma}}$ we find that in the LHS of (1.1.48)
this monomial appears only in $\giz^{\l_{\sigma}}$, while in the RHS it appears with
coefficient $x$ in $\Phi_{\sigma}^{\l}$ and as the maximal monomial in $\Phi_{(i,i+1)\sigma}^{\l}$. Equating the
coefficients we get:
\begin{align}
q & = \frac{\d\zeta_{i+1}^{\l}(\sigma)}{\zeta_{i+1}^{\l}(\sigma) - \zeta_i^{\l}(\sigma)} x + \frac{\tau_{i,i+1}^{\l}(\sigma)}{\zeta_{i+1}^{\l}(\sigma) -
\zeta_i^{\l}(\sigma)} \,.
\end{align}
Combining (1.1.49) and (1.1.50) we obtain:
\begin{align}
\tau_{i,i+1}^{\l}(\sigma) & = q\frac{(q^{-1}\zeta_{i+1}^{\l}(\sigma) - q\zeta_i^{\l}(\sigma))(q\zeta_{i+1}^{\l}(\sigma) -
q^{-1}\zeta_i^{\l}(\sigma))}{\zeta_{i+1}^{\l}(\sigma) - \zeta_i^{\l}(\sigma)} \quad ( \l_{\sigma_i} > \l_{\sigma_{i+1}})
\end{align}
\mbox{}
\noindent Let $ \l_{\sigma_i} < \l_{\sigma_{i+1}}\,\Rightarrow\, z^{\l_{\sigma}} < z^{\l_{(i\,i+1)\sigma}} $.
Equate the coefficients in front of the monomial $z^{\l_{(i\,i+1)\sigma}}$ in the both sides of
(1.1.48).
\noindent In the LHS the contribution comes only from $\giz^{\l_{\sigma}}$. In the RHS
only $\Phi_{(i,i+1)\sigma}^{\l}$ contributes $z^{\l_{(i\,i+1)\sigma}}$ as its maximal monomial. Application of
(1.1.38) leads to:
\begin{align}
q^{-1} & = \frac{\tau_{i,i+1}^{\l}(\sigma)}{\zeta_{i+1}^{\l}(\sigma) - \zeta_i^{\l}(\sigma)} \quad ( \l_{\sigma_i} <
\l_{\sigma_{i+1}}).
\end{align}
\mbox{}
\noindent {\bf 4.} To summarize, we have obtained the following proposition:
\begin{prop}
Let $ \sigma \in \S \quad (\l \in \Lambda_N)$, then $H_N(q)$ acts in
$E^{\l}\:=\:{\Bbb C}\{\Phi_{\sigma}^{\l} \}_{\sigma \in \S}$ as follows: \begin{equation} g_{i,i+1}\Phi_{\sigma}^{\l} =
\frac{(q-q^{-1})\zeta_{i+1}^{\l}(\sigma)}{\zeta_{i+1}^{\l}(\sigma) -\zeta_i^{\l}(\sigma)}\Phi_{\sigma}^{\l} +
\begin{cases}
q\frac{(q^{-1}\zeta_{i+1}^{\l}(\sigma) - q\zeta_i^{\l}(\sigma))(q\zeta_{i+1}^{\l}(\sigma) - q^{-1}\zeta_i^{\l}(\sigma))}{(\zeta_{i+1}^{\l}(\sigma) - \zeta_i^{\l}(\sigma))(\zeta_{i+1}^{\l}(\sigma)
- \zeta_i^{\l}(\sigma))}\Phi_{(i,i+1)\sigma}^{\l} & \\ \text{{\em when} $\l_{\sigma_i} > \l_{\sigma_{i+1}}\,\Rightarrow\,
(i,i+1)\sigma \in \S $}, & \\
0 \quad \text{{\em when} $\l_{\sigma_i} = \l_{\sigma_{i+1}}\,\Leftrightarrow\,
(i,i+1)\sigma \not\in \S $}, & \\
q^{-1}\Phi_{(i,i+1)\sigma}^{\l} & \\ \text{{\em when} $\l_{\sigma_i} < \l_{\sigma_{i+1}}\,\Rightarrow\,
(i,i+1)\sigma \in \S $}. &
\end{cases}
\end{equation}
\end{prop}
\section{The limit $p \rightarrow 1$ of the hierarchy of Dynamical Models.}
\subsection{Few facts about Macdonald operators.}
{\bf 1.}The Macdonald operators $ D_N^n(p,t) \quad (n=0,\dots,N)$ [M] act in
the subspace of ${\Bbb C}[z_1,\dots,z_N]$ formed by symmetric polynomials. In notation of [JKKMP]
these operators are defined as follows:
\begin{equation}
D_N^n(p,t) := t^{n(n-1)/2}\sum_{I_n}\prod\begin{Sb}
i\in I_n \\
j\not\in I_n
\end{Sb}
\frac{tz_i - z_j}{z_i - z_j}\prod_{k\in I_n} p^{D_k} \quad (n=0,\dots,N),
\end{equation}
where the summation is over all subsets $ I_n $ of $\{1,2,\dots,N \}$ which contain $n$
elements. Using the formula:
\begin{equation*}
det\left\| \frac{(t-1)w_i}{tw_i-w_j}\right\|_{1\leq i,j\leq m} = \quad
t^{m(m-1)/2}\prod_{1\leq i\neq j\leq m}\frac{w_i-w_j}{tw_i-w_j},
\end{equation*}
where $w_i \quad (i=1,\dots,m)$ are numbers; we can rewrite the definition of
the operators $ D_N^n(p,t) \quad (n=0,\dots,N)$ in another form:
\begin{equation}
D_N^n(p,t) = \sum_{I_n}detA_{I_n}(t)\prod_{k\in I_n} p^{D_k} \quad
(n=0,\dots,N),
\end{equation}
where $A_{I_n}(t)$ is a submatrix of the matrix:
\begin{equation}
A(t) = \| A_{ij}(t)\|_{1\leq i,j\leq N}\;,\;A_{ij}(t):=\frac{(t-1)z_i}{tz_i -
z_j} \prod\begin{Sb}
1\leq k\leq N \\
k\neq i
\end{Sb}\frac{tz_i - z_k}{z_i -
z_k} ;
\end{equation}
which is defined as follows:$ A_{I_n}(t):=\| A_{ij}(t)\|_{i,j\in I_n}.$
The Macdonald polynomials $P_{\l}(p,t) \quad (\l \in \Lambda_N)$ are
eigenfunctions of the operators $ D_N^n(p,t) \quad (n=0,\dots,N)$:
\begin{equation}
D(v;p,t)P_{\l}(p,t) = \prod_{i=1}^N (1 + t^{N-i}p^{\l_i}v)\, P_{\l}(p,t) \quad
(\l \in \Lambda_N),
\end{equation}
here $D(v;p,t)$ is the generating function of Macdonald operators:
\begin{equation*}
D(v;p,t):= \sum_{n=0}^N v^n D_N^n(p,t) \;.
\end{equation*}
\noindent {\bf 2.} In the limit $p \rightarrow 1$ one finds [M]:
\begin{equation}
P_{\l}(p,t) = e_{\l '} + O(p-1) \quad (\l \in \Lambda_N),
\end{equation}
where $\l '$ is the conjugate partition of $\l$ and for a partition $\pi \: :
(\pi_1 \geq \pi_2 \geq \dots ) \; , \; (\pi_1 \leq N) \quad e_{\pi}:=
e_{\pi_1}e_{\pi_2}\ldots \quad ;$ where $ e_r $ is the elementary symmetric
polynomial:
\begin{equation*}
e_r := \sum_{1\leq i_1 \leq\dots\leq i_r\leq N} z_{i_1}z_{i_2}\ldots z_{i_r}.
\end{equation*}
Consider the limit $p \rightarrow 1$ of the generating function $D(v;p,t)$:
\begin{equation}
D(v;p,t) \overset{p \rightarrow 1}{=} D_0(v;t) + (p-1) D_1(v;t) + O((p-1)^2).
\end{equation}
{}From (2.2.1,.4,.5) it follows that $D_0(v;t)$ is a multiplication by a
constant:
\begin{equation}
D_0(v;t) = \prod_{i=1}^N (1 + t^{N-i}v) \;.
\end{equation}
The first-order term $D_1(v;t)$ is a differential operator:
\begin{equation}
D_1(v;t) = \sum_{i=1}^N \left(\sum_{n=1}^N v^n \sum_{I_n: i\in I_n}
detA_{I_n}(t)\right)D_i
\end{equation}
Expanding the eq. (2.2.4) up to the first order in $ p-1 $ and using (2.2.5,.7
) we get:
\begin{equation}
D_1(v;t)e_{\l '} = \left(v\sum_{j=1}^N\prod\begin{Sb} 1\leq k\leq N \\
k\neq j \end{Sb} (1 + v t^{N-k})
t^{N-j}\l_j\right)\,e_{\l '} \quad ( \l \in \Lambda_N ).
\end{equation}
\subsection{Taking the limit $p \rightarrow 1$ in the hierarchy of Dynamical
Models}
{\bf 1.} The following fact was established in the paper [JKKMP]. Let $S$ be
any symmetric polynomial $ ( S \in {\Bbb C}[z_1,\dots,z_N] )$. The action of the operator $\Delta (u)$ (cf.
{\bf 1.1}) on such $S$ coincides with the action of the generating function of
Macdonald operators:
\begin{equation}
\Delta (u) S = D(q^{N-1}u;p,q^{-2})S \;.
\end{equation}
There is another connection between $\Delta (u)$ and Macdonald operators.
For an operator $O$ which is a function of operators $ D_i , z_i \; (i=1,\dots,N) ,
K_{i,i+1} $ $ (i = 1,\dots,N-1) $ introduce a normal ordering $:\; :$. The
normal ordering is described as follows: in $O$ bring all the operators $ D_i $
to the right {\em without} taking commutators between $ D_i $ and $ z_j $, but
taking into account the commutation relations between $ D_i $ and $ K_{j,j+1}
$. For instance:
\begin{equation*}
:\; p^{D_1}\, \frac{q^{-1}z_1 - qz_2}{z_1-z_2}(K_{12} - 1)\; : \;=
\frac{q^{-1}z_1 - qz_2}{z_1-z_2}(K_{12} \,p^{D_2}- p^{D_1}) .
\end{equation*}
\noindent We formulate the following Lemma:
\begin{lemma}
\begin{equation}
:\; \Delta (u) \; : \; = D(q^{N-1}u;p,q^{-2}) \; .
\end{equation}
\end{lemma}
\begin{pf}
To facilitate the proof we introduce an extension of the algebra generated by $
z_1,z_2,\dots,z_N $ and $ K_{1,2},K_{2,3},\dots,K_{N-1,N}$ by symbols $
\xi_1,\xi_2,\dots,\xi_N $. These symbols are defined by the commutation
relations
\begin{align*}
[\xi_i,\xi_j] & = 0 \; , \; [\xi_i,z_j] = 0 \qquad ( i,j \in \{1,2,\dots,N \}), \\
K_{i,j}\xi_j & = \xi_i K_{i,j}\;,\; [K_{i,j},\xi_k] = 0 \quad (k\neq i,j)
\qquad ( i\neq j \in \{1,2,\dots,N \}).
\end{align*}
One can take $\xi_i := f(z_i) $, where $f$ is any function of one variable, as
a realization for $ \xi_i$. \\
Together with $ g_{i,i+1} $ the operators
\begin{equation*}
Y_i(\xi) := g_{i,i+1}^{-1}K_{i,i+1}\dots g_{i,N}^{-1}K_{i,N}\xi_i
K_{1,i}g_{1,i}\dots K_{i-1,i}g_{i-1,i}
\end{equation*}
still satisfy the defining relations of $ \widehat{H_N(q)}$. \\
This implies in particular that for the operator
\begin{equation*}
\Delta(u;\xi) := \prod_{i=1}^N(1 + u Y_i(\xi))
\end{equation*}
we have
\begin{equation}
[\Delta(u;\xi),g_{i,i+1}] = 0 \qquad (i=1,\dots,N-1). \tag{i}
\end{equation}
Using the commutation relations with $ z_1,z_2,\dots,z_N $ and $ K_{i,j} $ we
can bring the symbols $ \xi_i $ to the right of all expressions in
$\Delta(u;\xi)$. Denote $\Delta(u;\xi)$ with all $ \xi_i $ brought to the right
by $\Delta(u;\xi)'.$ \\
We have
\begin{equation*}
: \Delta(u) : = \Delta(u;\xi)' |_{\xi_i \rightarrow p^{D_i}}.
\end{equation*}
Therefore in order to prove the statement of the lemma we compute the
coefficients standing in front of monomials $ \xi_{i_1}\xi_{i_2}\dots \xi_{i_n}
\quad (1\leq i_1 \leq i_2 \leq \dots \leq i_n \leq N)$ in the symmetric
functions
\begin{equation*}
\Delta(\xi)^{(n) } := \sum_{N \geq k_1 > k_2 > \dots > k_n \geq 1 }
Y_{k_1}(\xi)Y_{k_2}(\xi)\dots Y_{k_n}(\xi) \quad ( 1 \leq n \leq N ).
\end{equation*}
With notation of {\em {\bf 1.1.1}} we have
\begin{align*}
r_{i,j} & := K_{i,j}g_{i,j} = a_{j,i} + b_{j,i}K_{i,j} , \\
r_{i,j}^{-1} & = a_{i,j} - b_{j,i}K_{i,j}, \\
Y_i(\xi) & = r_{i,i+1}^{-1}\dots r_{i,N}^{-1}\xi_i r_{1,i}\dots r_{i-1,i}.
\end{align*}
\mbox{}
\noindent Let us compute the terms in $\Delta(\xi)^{(n)'} \quad ( 1\leq n \leq
N)$ which contain symbols $ \xi_1 ,\xi_2, \dots ,\xi_n $ only. By inspection we
find that such terms can appear only in $ Y_n(\xi)Y_{n-1}(\xi)\dots Y_1(\xi)
$. Furthermore the relevant contributions from the individual factors in the
last expression are
\begin{align*}
Y_1(\xi) & \rightarrow r_{1,2}^{-1}\dots r_{1,n}^{-1}a_{1,n+1}\dots a_{1,N}
\xi_1 , \\
Y_2(\xi) & \rightarrow r_{2,3}^{-1}\dots r_{2,n}^{-1}a_{2,n+1}\dots a_{2,N}
\xi_2 r_{1,2}, \\
& \vdots \\
Y_k(\xi) & \rightarrow r_{k,k+1}^{-1}\dots r_{k,n}^{-1}a_{k,n+1}\dots a_{k,N}
\xi_k r_{1,k}\dots r_{k-1,k}, \\
& \vdots \\
Y_{n-1}(\xi) & \rightarrow r_{n-1,n}^{-1}a_{n-1,n+1}\dots a_{n-1,N} \xi_{n-1}
r_{1,n-1}\dots r_{n-2,n-1}, \\
Y_n(\xi) & \rightarrow a_{n,n+1}\dots a_{n,N} \xi_n r_{1,n}\dots r_{n-1,n}.
\end{align*}
Multiplying these contributions we find that there is only one term in
$\Delta(\xi)^{(n)'}$ which contains $ \xi_1 ,\xi_2, \dots ,\xi_n $ only; and
this term is
\begin{multline*}
(a_{1,n+1}\dots a_{1,N})(a_{2,n+1}\dots a_{2,N})\ldots (a_{n,n+1}\dots
a_{n,N})\xi_1\xi_2\dots\xi_n .
\end{multline*}
Next we use the Hecke-invariance relation {\em (i) } and find
\begin{equation*}
\Delta(\xi)^{(n)} = \sum_{I_n} \left(\prod\begin{Sb} i \in I_n \\ j \not\in I_n
\end{Sb} a_{i,j}\right) \prod_{i \in I_n } \xi_i.
\end{equation*}
Where the summation is over all $n$-element subsets of $\{1,2,\dots,N \}$. Comparing this
expression with {\em (2.2.1)} we obtain the statement of the lemma.
\end{pf}
\mbox{}
\noindent {\bf 2.} Let us take the limit $p \rightarrow 1$ in the operators $
\Delta (u) , T_a(u), Y_i \quad (i=1,\dots,N).$ Expanding around $ p = 1 $ and keeping the first
two terms of the expansion we write:
\begin{align}
Y_i &\overset{p \rightarrow 1}{=} y_i + (p-1)x_i + O((p-1)^2)\quad ((i=1,\dots,N)), \\
\Delta (u) &\overset{p \rightarrow 1}{=} \Delta_0(u) + (p-1)\Delta_1(u)+ O((p-1)^2)\:,
\\
T_a(u) &\overset{p \rightarrow 1}{=} T_a^0(u) + (p-1) T_a^1(u)+ O((p-1)^2)\: .
\end{align}
Here we introduced the operators:
\begin{align}
y_i & := g_{i,i+1}^{-1}K_{i,i+1}\dots g_{i,N}^{-1}K_{i,N} K_{1,i}g_{1,i}\dots
K_{i-1,i}g_{i-1,i} \quad ((i=1,\dots,N)) , \\
x_i & := g_{i,i+1}^{-1}K_{i,i+1}\dots g_{i,N}^{-1}K_{i,N}D_i
K_{1,i}g_{1,i}\dots K_{i-1,i}g_{i-1,i} \quad ((i=1,\dots,N)) , \\
& \Delta_0(u) := \prod_{i=1}^N(1 + uy_i) \;, \\
& \Delta_1(u) := u\sum_{j=1}^N \prod_{1\leq i<j}(1 + uy_i)\, x_j\prod_{j<k\leq
N}(1 + uy_k)\;,\\
T_a^0(u) & := L_{a1}(uy_1)L_{a2}(uy_2)\ldots L_{aN}(uy_N) \quad \in
End{\Bbb C}[z_1,\dots,z_N]\otimes H\;.
\end{align}
The operators $y_i , g_{i,i+1}$ satisfy the relations (cf. {\bf 1.1}) of Affine
Hecke Algebra and $ T_a^0(u)$ defines a representation of $U$.
\noindent {\bf 3.} From Lemma 2 it follows that the operators $ \Delta_0(u)\,
, \, \Delta_1(u) $ have a rather special form. We have the proposition:
\begin{prop}
The following statements hold:
\noindent Let $D_0(v;t)\, , \; D_1(v;t)$ be those defined in {\em (2.2.6,.8)};
then:
\begin{align}
\Delta_0(u) & = D_0(q^{N-1}u;q^{-2}) \, = \, \prod_{i=1}^N (1 + u q^{2i-N-1})I
\, . \tag{i}
\end{align}
I.e. $\Delta_0(u)$ is a multiplication by a constant.
\begin{align}
\Delta_1(u) & = D_1(q^{N-1}u;q^{-2}) + \Xi(u) \,. \tag{ii}
\end{align}
Where operator $ \Xi(u)$ is a function of operators $ z_i \,,\, K_{i,j} \quad
(i,j = 1,\dots, N) $ only ( and {\em not} of $ D_i $).
\end{prop}
\mbox{}
\noindent Let $ \Delta (u)_1 := \sum_{i=1}^{N}u^i \Delta_1^{(i)} $ and $ \Xi(u) :=
\sum_{i=1}^{N}u^i \Xi^{(i)} $. We have computed explicit expressions for the
operators $ \Delta_1^{(N)} $ and $ \Delta_1^{(1)} $. In notation of section
{\bf 1.1} one has:
\begin{align}
\Delta_1^{(N)} & = D_1 + D_2 + \dots + D_N \, , \\
\Delta_1^{(1)} & = \sum_{i=1}^N \left( \prod\begin{Sb} 1\leq k\leq N \\ k\neq i
\end{Sb} a_{i,k} \right)D_i \, + \, \Xi^{(1)}\,
\end{align}
where:
\begin{multline}
\Xi^{(1)} = \\ \sum_{M=2}^N \frac{(-1)^M}{\d}\sum_{N\geq i_M > \dots >i_1\geq
1} {\cal A}_{i_M,i_{M-1},\dots,i_1}{\cal B}_{i_M,i_{M-1},\dots,i_1}
K_{i_M,i_{M-1}}\dots K_{i_2,i_1} \; + \; \varphi^{(1)} \:, \notag \\
{\cal A}_{i_M,i_{M-1},\dots,i_1} = \left(\prod_{i_1<f<i_2}
a_{i_1,f}\right)\left(\prod_{i_2<f<i_3} a_{i_2,f}\right)\ldots
\left(\prod_{i_M<f<N+i_1} a_{i_M,f\pmod{N}}\right) \:, \\
{\cal B}_{i_M,i_{M-1},\dots,i_1} = b_{i_M,i_{M-1}}b_{i_M-1,i_{M-2}}\ldots
b_{i_2,i_{1}}b_{i_1,i_M} \:, \\
\varphi^{(1)} = -\sum_{1\leq k< i\leq N}\frac{a_{i,i+1}\dots a_{i,N}
a_{i,1}\dots a_{i,k-1}a_{i,k+1}\dots a_{i,i-1} b_{k,i} b_{i,k}}{\d} \: . \notag
\end{multline}
\mbox{}
\noindent {\bf 4.} Let us fix the notation:
\begin{equation}
{\cal D}(u) := D_1(q^{N-1}u;q^{-2}) = \sum_{i=1}^N \theta_i(u)\,D_i\, .
\end{equation}
Where according to (2.2.8) :
\begin{equation}
\theta_i(u) := \sum_{n=1}^N u^n q^{n(N-1)}\sum_{I_n : i\in I_n} det
A_{I_n}(q^{-2}) \quad ((i=1,\dots,N)).
\end{equation}
(Cf. sec. {\bf 2.1} for the definition of $A_{I_n}(t)$).
The functions $\theta_i(u)\quad ((i=1,\dots,N))$ can be written in the following form:
\begin{equation}
\theta_i(u) = \frac{\partial}{\partial \gamma_i}det(I + u\Gamma q^{N-1}
A(q^{-2}))|_{\Gamma = I} \, ,
\end{equation}
where we have introduced an auxiliary matrix: $ \Gamma := {\mathrm
{diag}}\{\gamma_1,\ldots, \gamma_N \} .$
Using this representation we compute $\theta_i(u)\quad ((i=1,\dots,N))$ at the point $
z_1 = \omega^1,\dots,z_N = \omega^N $, where $\omega := {\mathrm {exp}}(2\pi
i/N).$ The computation yields:
\begin{gather}
\theta_i(u)|_{z_1=\omega^1,\dots,z_N=\omega^N} = \frac{1}{N}\prod_{k=1}^N
(1+uq^{l_k})\,\sum_{n=1}^N \frac{uq^{l_n}}{1+uq^{l_n}} \; \equiv \theta(u)
\quad (i=1,\dots,N). \\
l_i := 2i - N - 1 \notag
\end{gather}
Thus the point $ z_1 = \omega^1,\dots,z_N = \omega^N $ (or any point obtained
from it by a permutation of coordinates) is special in that at this point all
the $\{z_i\}$-dependent coefficients $\theta_i(u)$ of the first-order
differential operator ${\cal D}(u)$ become equal one to another.
\section{Definition of the Hierarchy of Integrable, $U$-invariant Spin Models}
\subsection{Preliminaries}
{\bf 1.} Let us expand the relations (1.1.34,.35) around the point $p=1$ using
the definitions (2.2.12-.14) and the fact that $\Delta_0(u)$ is a constant. For
$\Delta_1(u),\,T_a^0(v) \in End({\Bbb C}[z_1,\dots,z_N]\otimes H)$ we obtain:
\begin{align}
[\Delta (u)_1,T_a^0(v)] & = 0 \; , \\
\mbox{} [\Delta_1(u),\Delta_1(v)] & = 0.
\end{align}
Expanding the relations:
\begin{equation}
[\Delta (u),Y_i] = 0 \quad (i=1,\dots,N),\quad [\Delta (u),g_{i,i+1}] = 0 \quad (i=1,\dots,N-1),
\end{equation}
we get:
\begin{equation}
[\Delta_1(u),y_i] = 0 \quad (i=1,\dots,N),\quad [\Delta_1(u),g_{i,i+1}] = 0 \quad
(i=1,\dots,N-1).
\end{equation}
Due to (3.3.4) and the fact that $y_i \quad (i=1,\dots,N) ,\; g_{i,i+1} \quad
(i=1,\dots,N-1)$ satisfy the Affine Hecke Algebra relations, the operators
$\Delta_1(u), T_a^0(v)$ act in the ``bosonic'' subspace ${\cal B}$ (cf. 1.1.36 and
3.3.7).
\mbox{}
\noindent {\bf 2.} Introduce several definitions.
Let
\begin{align}
{\cal R} & := {\Bbb C} [\{\frac{1}{z_i - z_j}\}_{1\leq i\neq j\leq N}\,
,\{\frac{1}{z_i - q^{2}z_j}\}_{1\leq i\neq j\leq N}\,,\: z_1,\dots,z_N ]\otimes
H.
\end{align}
Let ${\cal P}$ and ${\cal B}$ be the subspaces of $ {\cal R}$:
\begin{align}
{\cal P} & := {\Bbb C}[z_1,\dots,z_N]\otimes H \, , \\
{\cal B} & := \{b\in {\cal P} | (g_{i,i+1} - t_{i,i+1})b = 0 \quad (i=1,\dots,N-1)\}.
\end{align}
For $v_1,\dots,v_N \, \in {\Bbb C} $ such that $ v_i \neq v_j\,,\,q^{2} v_j \quad
(i\neq j\,; i,j =1,\dots,N)$
define the evaluation map: $ Ev(v): {\cal R} \mapsto H $ by taking values of
rational functions at the point $
z_1 = v_1,\dots, z_N = v_N.\; ( \Leftrightarrow z = v ) $.
\noindent For any $ O \in End({\cal R}) $ define an operator $ \hat{O} \in
End({\cal R}) $ by the rule [BGHP]:
Using the commutation relations bring all the permutation operators
$K_{i,j}\quad (i,j \in \{1,2,\dots,N \})$
to the right of an expression in $O$; replace the rightmost of $K_{i,j}$ using
the substitution:
\begin{equation}
g_{i,i+1} \rightarrow t_{i,i+1} \Rightarrow K_{i,i+1} \rightarrow
\frac{z_it_{i,i+1}^{-1} - z_{i+1}t_{i,i+1}}{q^{-1}z_i - qz_{i+1}} \quad
(i=1,\dots,N-1).
\end{equation}
Repeat the procedure until there are no operators $K_{i,j}$ left. The result
is $ \hat{O}$.
In what follows we adopt the following notational convention: if ${\cal L}$ is
a linear space and $A,B$ are linear operators defined on ${\cal L}$, we write:
\begin{gather*}
A{\cal L} = B{\cal L}\quad \text{meaning } \quad Al = Bl \quad \forall l \in
{\cal L}.
\end{gather*}
In particular for $O , \hat{O} \quad \in End({\cal R})$ defined above we have:
\begin{equation}
O{\cal B} = \hat{O}{\cal B}.
\end{equation}
Let $O,O' \in End({\cal P})$ be s.t.:
\begin{gather}
[O,O']{\cal P} = 0 \quad \text{and}\quad O ,O': {\cal B}\mapsto {\cal B},
\end{gather}
then
\begin{equation}
[\hat{O},\hat{O'}]{\cal B} = 0.
\end{equation}
\noindent {\bf 3.}With notation of (2.2.20,.23,.26) let us consider the
following differential operator $\widetilde{{\cal D}(u)} \in End({\cal R})$:
\begin{equation}
\widetilde{{\cal D}(u)} := {\cal D}(u) - \theta(u)\Delta_1^{(N)}.
\end{equation}
Let $Ev(\omega)$ be the evaluation map $Ev(v)$ taken at the special point $ v_1
= \omega^1,\dots,v_N = \omega^N \quad ( \Leftrightarrow v = \omega ) $. Then in
virtue of (2.2.26) we obtain the following property of $ \widetilde{{\cal
D}(u)} $:
\begin{equation}
Ev(\omega)\widetilde{{\cal D}(u)}{\cal R} = 0.
\end{equation}
Let us introduce the modified generating function $ \widetilde{\Delta_1(u)} $ by
subtracting the product of the constant $\theta(u)$ and the operator
$\Delta_1^{(N)}$:
\begin{equation}
\widetilde{\Delta_1(u)} := \Delta_1(u) - \theta(u)\Delta_1^{(N)} = \widetilde{{\cal D}(u)} +
\Xi(u) .
\end{equation}
Since $\Delta_1^{(N)}$ is a member of the hierarchy of commuting operators
defined by $\Delta_1(u)$, the equations (3.3.1,.2),(3.3.4) still hold if we replace in
these equations $\Delta_1(u)$ by $\widetilde{\Delta_1(u)}$:
\begin{align}
[\widetilde{\Delta_1(u)},\widetilde{\Delta_1(v)}]{\cal P} & = 0 , \\
[\widetilde{\Delta_1(u)}, T_a^0(v) ]{\cal P} & = 0 , \\
\widetilde{\Delta_1(u)} : {\cal B} & \mapsto {\cal B} .
\end{align}
\subsection{Definition of the hierarchy of Spin Models}
{\bf 1.} Let $ H_{{\cal B}}(\omega)$ be the image of ${\cal B}$ under the action of the
evaluation map $Ev(\omega)$:
\begin{equation}
Ev(\omega){\cal B} = H_{{\cal B}}(\omega) \subset H .
\end{equation}
Since $ \widehat{T_a^0(u)}\;,\,\widehat{\Xi(u)} $ do not depend on the diffrential
operators $ D_i \; (i=1,\dots,N) $ and the operators of coordinate permutation, we can
define the operators $T_a^0(u;\omega) , \Xi(u;\omega)$ as follows:
\begin{equation}
Ev(\omega)\widehat{T_a^0(u)} = T_a^0(u;\omega) Ev(\omega) , \quad
Ev(\omega)\widehat{\Xi(u)} = \Xi(u;\omega) Ev(\omega).
\end{equation}
Applying $Ev(\omega)$ to the relations :
\begin{align}
\widehat{T_a^0(u)}:{\cal B} & \mapsto {\cal B} , \\
\widehat{\widetilde{\Delta_1(u)}}:{\cal B} & \mapsto {\cal B},
\end{align}
and using (3.3.13) we get:
\begin{align}
T_a^0(u;\omega) : H_{{\cal B}}(\omega) & \mapsto H_{{\cal B}}(\omega) , \\
\Xi(u;\omega) : H_{{\cal B}}(\omega) & \mapsto H_{{\cal B}}(\omega).
\end{align}
\noindent {\bf 2.} Apply the evaluation map $Ev(\omega)$ to the relations
(3.3.15,.16) taking (3.3.13) and (3.3.23) into account. As the result we find
that the operator $\Xi(u;\omega)$ is a generating function of the commuting,
$U$-invariant integrals of motion which are operators in $ H_{{\cal B}}(\omega)$:
\begin{gather}
[\Xi(u;\omega), \Xi(v;\omega) ] H_{{\cal B}}(\omega) = 0 \, , \\
[\Xi(u;\omega) , T_a^0(v;\omega) ] H_{{\cal B}}(\omega) = 0 , \\
\left(\bar{R}_{ab}(u/v)T_a^0(u;\omega) T_b^0(v;\omega) -
T_b^0(v;\omega)T_a^0(u;\omega)\bar{R}_{ab}(u/v)\right) H_{{\cal B}}(\omega) = 0 .
\end{gather}
In sec. 6 we shall show that $H_{{\cal B}}(\omega)=H$. This completes the definition
of the hierarchy $ \Xi(\omega)^{(1)},\dots,\Xi(\omega)^{(N-1)} \; ( \Xi(u;\omega) =
\sum_{n=1}^{N-1} u^n \Xi^{(n)}(\omega) ) $ of Spin Models.
\section{Eigenvalue spectrum of the operators $\Delta_1(u)\;,
\protect\newline y_i \; (i=1,\dots,N) $}
\subsection{Characteristic numbers and eigenvalues of $\Delta_1(u)\;,\;y_i $}
{\bf 1.} To find the action of the operators $\Delta_1(u)\;,\;y_i \; (i=1,\dots,N)$ in the
monomial basis of ${\Bbb C}[z_1,\dots,z_N]$ we can take the limit $p\rightarrow 1$ in the formulas
of Proposition 1. This gives the following proposition:
\begin{prop}
The operators $\Delta_1(u)\;,\;y_i \; (i=1,\dots,N)$ are triangular in the monomial basis of
${\Bbb C}[z_1,\dots,z_N]$. The action of these operators on monomials is given by:
\begin{align}
y_iz^{\l_{\sigma}} & = q^{l_{\sigma_i}}z^{\l_{\sigma}} + \text{{\em``s.m''}}\quad (\l \in \Lambda_n, \sigma
\in \S, \; (i=1,\dots,N) ), \tag{i} \\
\Duz^{\l_{\sigma}} & = \delta^{\l}(u)z^{\l_{\sigma}} + \text{{\em``s.m''}}\quad (\l \in \Lambda_N, \sigma
\in \S ) \tag{ii}.
\end{align}
where $l_i := 2i - N -1 \quad (i=1,\dots,N) $ and
\begin{equation*}
\delta^{\l}(u) := u\sum_{j=1}^N\left(\prod\begin{Sb} 1\leq k\leq N \\ k\neq j
\end{Sb}(1 + uq^{l_k})\right)q^{l_j}\l_j .
\end{equation*}
\end{prop}
Since $\Delta_1(u)\;$ and $\;y_i \; (i=1,\dots,N)$ commute among themselves (cf. 3.3.4) and the
joint spectrum of characteristic numbers of $\Delta_1(u)\;$ and $\;y_i \; (i=1,\dots,N)$ given
by Proposition 5 is explicitely multiplicity-free, we apply Lemma 1 and claim
that $\Delta_1(u)\;,\;y_i \; (i=1,\dots,N)$ are simultaneously diagonalizable:
\begin{prop} There exist polynomials $\varphi_{\sigma}^{\l}\quad (\l \in \Lambda_N, \sigma \in \S )$
s.t.:
\begin{align}
{\Bbb C}[z_1,\dots,z_N] & = {\bigoplus}_{\l \in \Lambda_N}{\cal E}^{\l}\;,\;{\cal
E}^{\l}:={\oplus}_{\sigma\in\S}{\Bbb C}\varphi_{\sigma}^{\l}\,,\tag{i}\\
y_i\varphi_{\sigma}^{\l} & = q^{l_{\sigma_i}}\varphi_{\sigma}^{\l} \quad (i=1,\dots,N) ,\tag{ii}\\
\Delta_1(u)\varphi_{\sigma}^{\l} & = \delta^{\l}(u)\varphi_{\sigma}^{\l}\,,\tag{ii'}\\
\varphi_{\sigma}^{\l} & = z^{\l_{\sigma}} + \text{{\em``s.m''}}\,.
\end{align}
\end{prop}
\mbox{}
\noindent {\bf 2.} Let us show that $ \varphi_{\sigma}^{\l} = {\lim}_{p\rightarrow 1} \Phi_{\sigma}^{\l}\quad
(\l \in \Lambda_N, \sigma \in \S )$ where $\Phi_{\sigma}^{\l}\quad (\l \in \Lambda_N, \sigma \in \S )$
are eigenfunctions of the operators $Y_i\quad(i=1,\dots,N)$ (Cf. Proposition 2).
Expanding $\Phi_{\sigma}^{\l}\quad (\l \in \Lambda_N, \sigma \in \S )$ around $p = 1$ let us
write:
\begin{equation}
\Phi_{\sigma}^{\l} \overset{p\rightarrow 1}{=} (p - 1)^{s} \psi^{\l}_{\sigma} + O((p-1)^{s+1}),
\end{equation}
where $ \psi^{\l}_{\sigma}$ is a non-zero polynomial. Since $ \Phi_{\sigma}^{\l} = z^{\l_{\sigma}} +
\text{{\em `` s.m.''}}
$, we have: $s \leq 0$.
Let us show that $s=0$. Suppose $s<0$. The satement (c) of Lemma 1 when applied
to the eigenvectors $\Phi_{\sigma}^{\l}$ enables us to detect which coefficients in the
decomposition of $\Phi_{\sigma}^{\l}$ into monomials are potentially singular in the limit $p
\rightarrow 1$. The singularities may arise because of the presence of
denominators of the form
\begin{equation}
\frac{1}{p^{\l_{\sigma_i}}q^{l_{\sigma_i}} - p^{\mu_{\sigma_i}}q^{l_{\sigma_i}}}
\end{equation}
where $\mu$ is a partition {\em smaller} than $\l$ and $\sigma \in \S , S^{\mu}_N
$.Therefore if $s$ in (4.4.2) is negative, the maximal monomial in $
\psi^{\l}_{\sigma}$ is {\em smaller } than any of the monomials $z^{\l_{\sigma}} \quad (\sigma \in
\S)$.
Expanding the equations (ii),(ii') of the Proposition 2 around the point $p=1$
and taking into account that $\Delta_0(u) = \Delta (u) |_{p=1}$ is a constant, we
arrive at the following equations:
\begin{align*}
\Delta_1(u)\psi^{\l}_{\sigma} & = \delta^{\l}(u)\psi^{\l}_{\sigma}\,, \\
y_i\psi^{\l}_{\sigma} & = q^{l_{\sigma_i}} \psi^{\l}_{\sigma}.
\end{align*}
Since the joint spectrum of the operators $\Delta_1(u)$ and $y_i \; (i=1,\dots,N)$ is
multiplicity-free, we must have:
\begin{equation*}
\psi^{\l}_{\sigma} \propto \varphi_{\sigma}^{\l}\,.
\end{equation*}
This contradicts the observation that the maximal monomial of $\psi^{\l}_{\sigma}$
is smaller than any of the monomials $z^{\l_{\sigma}} \quad (\sigma \in \S)$.
Thus $s=0$ therefore $ \psi^{\l}_{\sigma} = \varphi_{\sigma}^{\l} $ and consequently $ \varphi_{\sigma}^{\l} =
{\lim}_{p\rightarrow 1}\Phi_{\sigma}^{\l}\quad (\l \in \Lambda_N, \sigma \in \S )$.
\subsection{Action of the Hecke Algebra in the eigenspaces ${\cal E}^{\l}\;(\l
\in \Lambda_N)$ of the
operator $\Delta_1(u)$ .}
{\bf 1.} According to (3.3.4) the Hecke Algebra $H_N(q)$ generated by $ g_{i,i+1} \;
(i=1,\dots,N-1) $ acts in each eigenspace ${\cal E}^{\l}\;(\l \in \Lambda_N)$
of the operator $\Delta_1(u)$. To compute this action explicitely in the basis $\{ \varphi_{\sigma}^{\l}
\}_{\sigma \in \S}$ we can either repeat almost word-by-word the derivation
described in {\bf 1.5 } or take the limit $p \rightarrow 1$ in the result of
Proposition 3. Either way we arrive at the following proposition:
\begin{prop} Let $\sigma \in \S \quad (\l\in \Lambda_N)$, then $H_N(q)$ generated
by $ g_{i,i+1} \; (i=1,\dots,N-1) $ acts in ${\cal E}^{\l} = {\Bbb C} \{ \varphi_{\sigma}^{\l} \}_{\sigma \in
\S}$ as follows:
\begin{equation}
g_{i,i+1}\varphi_{\sigma}^{\l} = \frac{\d q^{l_{\sigma_{i+1}}}}{q^{l_{\sigma_{i+1}}}-q^{l_{\sigma_i}}}\varphi_{\sigma}^{\l} +
\begin{cases}
& q\frac{(q^{-1}q^{l_{\sigma_{i+1}}}-qq^{l_{\sigma_i}})(qq^{l_{\sigma_{i+1}}}-q^{-1}q^{l_{\sigma_i}})}{(q^{l_{\sigma_{i+1}}}-q^{l_{\sigma_i}})(q^{l_{\sigma_{i+1}}}-q^{l_{\sigma_i}})}\varphi_{(i,i+1)\sigma}^{\l} \\
& \text{{\em when} $ \l_{\sigma_i} > \l_{\sigma_{i+1}} \Rightarrow (i,i+1)\sigma \in \S $}
, \\
& 0 \quad \text{{\em when} $\l_{\sigma_i} = \l_{\sigma_{i+1}} \Leftrightarrow (i,i+1)\sigma
\not\in \S $}, \\
& q^{-1}\varphi_{(i,i+1)\sigma}^{\l} \\
& \text{{\em when} $ \l_{\sigma_i} < \l_{\sigma_{i+1}} \Rightarrow (i,i+1)\sigma \in \S $}.
\end{cases}
\end{equation}
\end{prop}
\subsection{``Motifs'' and associated partitions}
{\bf 1.} Following [HHTBP,BPS] introduce the definition:
\begin{df}
Call a sequence of $M$ integers $(m_1,m_2,\dots,m_M)$ a {\em motif} iff:
\begin{align}
1 \leq m_1 < m_2 < \ldots < m_M \leq N-1 \, , \tag{i} \\
m_{i+1} \geq m_i + 2 \quad (i=1,\dots,M-1). \tag{ii}
\end{align}
\end{df}
With any motif $(m_1,m_2,\dots,m_M)$ associate a partition $\mu$:
\begin{equation*}
\mu = (M,\dots,\underset{m_1}{M},M-1,\dots,\underset{m_2}{M-1}, \; \ldots \;
,\underset{m_{M-1}+1}{1},\dots,\underset{m_M}{1},0,\dots,\underset{N}{0}) .
\end{equation*}
The subscripts in the last equation indicate positions of numbers in the
partition.
In what follows we shall identify motifs with the partitions they define. We
shall indiscriminately use the notation $(m_1,m_2,\dots,m_M)$ for both a motif
and the corresponding partition. Let $ {\frak M}_N $ be the set of all motifs for a
fixed $N$. We use the same notation for the corresponding subset of all
partitions.
\mbox{}
\noindent {\bf 2.} Let $(m_1,m_2,\dots,m_M)$ be a partition from the set $ {\frak M}_N $. We subdivide
the set $S_N^{\m}$ (cf. {\bf 1.4 }) into disjoint subsets:
\begin{df}
For any subset $\{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \quad (0\leq L\leq M) $ define $ S_{N,\mi}^{\m} \subset S_N^{\m}
\subset S_N $ as follows:
\begin{eqnarray*}
\lefteqn{S_{N,(\emptyset)}^{(m_1,m_2,\dots,m_M)} := \{ {\mathrm {id}} \} , } \\
\lefteqn{\text{{\em for } $ 1 \leq L \leq M $ } \qquad S_{N,\mi}^{\m} := } \\
& & \left\{ \sigma \in S_N^{\m} \quad \begin{array}{|c c} p^{\sigma}_{m_{i_k}} >
p^{\sigma}_{m_{i_k + 1}} & \forall\; 1\leq k \leq L \\
p^{\sigma}_{m_j} < p^{\sigma}_{m_j + 1 } & \forall\; j \in \{1,2,\dots,M \}\setminus\{i_1,i_2,\dots ,i_L\}
\end{array} \right\}
\end{eqnarray*}
\end{df}
Recall ({\bf 1.4}) that for $ \sigma \in S_N $ we define:
\begin{equation*}
\{ \sigma_1,\dots,\sigma_N \} := \sigma.\{1,2,\dots,N \} \; ,\quad i = \sigma_{p^{\sigma}_i} \quad (i=1,\dots,N) .
\end{equation*}
\noindent {\bf Example} Let $N=4\;,\; M=2$ and the motif is: $ (m_1 , m_2) = (1
, 3)$. The corresponding partition is: $ (2,1,1,0) $. In this case the set $
S_N^{(m_1,m_2)} = S_4^{(1,3)} $ contains altogether twelve elements. This set
is subdivided into four subsets: $ S_{4,(\emptyset)}^{(1,3)}\;,\;
S_{4,(1)}^{(1,3)}\;,\; S_{4,(3)}^{(1,3)}\;,\; S_{4,(1,3)}^{(1,3)} $ :
\begin{align*}
S_{4,(\emptyset)}^{(1,3)} & = \{ \{ 1234 \} \} \,; \\
S_{4,(1)}^{(1,3)} & = \{ \{ 2134\} ,\{ 2314 \} ,\{ 2341 \} \} \, ; \\
S_{4,(3)}^{(1,3)} & = \{ \{ 1243 \} ,\{ 1423 \} ,\{ 4123 \} \} \, ; \\
S_{4,(1,3)}^{(1,3)} & = \{ \{ 2143\} ,\{ 2413 \} ,\{ 4213 \} ,\{ 2431 \} ,\{
4231 \} \} .
\end{align*}
\mbox{}
\noindent {\bf 3.} Let us describe several properties of the sets $S_{N,\mi}^{\m} \quad (
(m_1,m_2,\dots,m_M) \in {\frak M}_N \;, \{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \quad (0\leq L\leq M) )$. Throughout this paragraph we
fix such a set $S_{N,\mi}^{\m} $. Let $\mu$ be the partition that corresponds to $(m_1,m_2,\dots,m_M)$.
\begin{lemma}
Let $\sigma \in S_{N,\mi}^{\m} $.
then either:
\begin{align*}
\exists \: i \in \{1,2,\dots,N-1 \} \, , & \; \sigma\prime \in S_{N,\mi}^{\m} \quad
\text{{\em s.t.}} \tag{i} \\
& \sigma = (i,i+1)\sigma\prime \: ,\; \mu_{\sigma_i} < \mu_{\sigma_{i+1}} . \\
& \text{{\em and therefore} $ \sigma\prime > \sigma \; ( \Leftrightarrow \mu_{\sigma\prime}
> \mu_{\sigma})$};
\end{align*}
or
\begin{align*}
\sigma & = \sigma[0]:= (m_{i_1},m_{i_1}+1)(m_{i_2},m_{i_2}+1)\ldots (m_{i_L},m_{i_L}+1)
. \tag{ii}
\end{align*}
\end{lemma}
\begin{pf}
\noindent Let $ \sigma = \{\sigma_1,\sigma_2,\dots,\sigma_N\} $. Examine the pairs $ \sigma_i ,
\sigma_{i+1} \; (i =1,\dots,N-1) $ step by step starting with $i=1$ and increasing
$i$ by $1$ at every next step.
\noindent At each step $i$ one has the two possibilities:
\begin{equation*}
{\mathrm I.}\quad \mu_{\sigma_i} \geq \mu_{\sigma_{i+1}} \qquad \qquad {\mathrm II.}
\quad \mu_{\sigma_i} < \mu_{\sigma_{i+1}}
\end{equation*}
\noindent If {\em I.}, go to the next step. If at each step $ (i
=1,\dots,N-1)$ holds {\em I.}, then $\sigma = {\mathrm id}$ and therefore $ L=0 \,
, \, \sigma[0] = {\mathrm id} $. The poof is finished.
\noindent If {\em II.}, then one has the further two possibilities:
\begin{equation*}
1. \; (i,i+1)\sigma \not\in S_{N,\mi}^{\m} \qquad \qquad 2. \; (i,i+1)\sigma \in S_{N,\mi}^{\m} .
\end{equation*}
\noindent If {\em 1.} go to the next step. If for all cases where {\em II.}
holds we have {\em 1.}, then $ \sigma = \sigma[0]$. The poof is finished.
\noindent If {\em 2.} denote $\sigma\prime := (i,i+1)\sigma $. Since $ \mu_{\sigma_i} <
\mu_{\sigma_{i+1}} $, we have $\mu_{\sigma\prime} > \mu_{\sigma} \Leftrightarrow
\sigma\prime > \sigma .$ The poof is finished.
\end{pf}
\noindent The element $\sigma[0]: = (m_{i_1},m_{i_1}+1)(m_{i_2},m_{i_2}+1)\ldots
(m_{i_L},m_{i_L}+1) \\ \in S_{N,\mi}^{\m} $ is the maximal element in the set $S_{N,\mi}^{\m}$
(Cf. {\bf 1.4.2} for the definition of the ordering of elements of $S_N^{\m}$).
\noindent We summarize the properties of the subset $S_{N,\mi}^{\m} $ in the following
proposition:
\begin{prop}
Let $\{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \;,\; (0\leq L\leq M) $.
\noindent Then:
\begin{equation*}
S_{N,\mi}^{\m} = S_{N,(m_{i_1},\dots,m_{i_L})}^{(m_{i_1},\dots,m_{i_L})}\; ,\tag{i}
\end{equation*}
\begin{gather*}
\forall \; \sigma \in S_{N,\mi}^{\m} \quad \exists \; \{ j_1,j_2,\dots,j_r\} \subset \{
1,\dots,N-1\} \quad ( r \geq 0 ) \; \text{{\em s.t.}} \tag{ii}
\end{gather*}
the elements $ \sigma[r] , \sigma[r-1],\dots , \sigma[1] $ defined by
\begin{align*}
\sigma[0] & := (m_{i_1},m_{i_1}+1)(m_{i_2},m_{i_2}+1)\dots (m_{i_L},m_{i_L}+1), \\
\sigma[k] & := (j_k,j_k+1)\sigma[k-1] \quad ( k=1,2,\dots,r )
\end{align*}
belong to the set $ S_{N,\mi}^{\m} $, \\
satisfy
\begin{equation*}
\sigma[k] < \sigma[k-1] \qquad ( k=1,2,\dots,r ),
\end{equation*}
and $ \sigma[r] = \sigma $. \\
\mbox{}
\noindent {\em (iii)} \\
If there exists $i \in \{1,2,\dots,N\}$ such that
\begin{align*}
\sigma \; , \; (i,i+1)\sigma \; & \in \; S_{N,\mi}^{\m} , \\
\text{then} & \quad |\sigma_{i+1} - \sigma_i | \geq 2 .
\end{align*}
If there exists $i \in \{1,2,\dots,N\}$ such that
\begin{align*}
\sigma \;\in \; S_{N,\mi}^{\m} , & \; (i,i+1)\sigma \; \in \;
S_{N,(m_{i_2},\dots,m_{i_L})}^{(m_1,m_2,\dots,m_M)} , \\
\text{then} & \quad \sigma_i \sigma_{i+1} + 1 .
\end{align*}
\end{prop}
\begin{pf}
\noindent {\em (i)} is a direct consequence of the {\em Definitions 1 ({\bf
1.4}) and 3}.
\noindent {\em (ii)} follows from Lemma 3 and the observation that $\sigma[0] =
(m_{i_1},m_{i_1}+1)(m_{i_2},m_{i_2}+1)\ldots (m_{i_L},m_{i_L}+1) $ is the
maximal element in $S_{N,\mi}^{\m} $.
\noindent {\em (iii)} For $\sigma \in S_{N,\mi}^{\m} $ assume that $ \sigma_{i+1} = \sigma_i +1 $.
One has the two possibilities:
\begin{align*}
1. & \quad i = p^{\sigma}_j \; , \; i+1 = p^{\sigma}_{j+1} \quad \text{where}\;
m_{s-1}+1 \leq j,j+1\leq m_s \\
& \text{for some}\; s \in \{1,2,\dots,M \} \; (m_0 := 0). \\
2. & \quad i = p^{\sigma}_{m_s}\; , \; i+1 = p^{\sigma}_{m_s+1} \\
& \text{for some} \; s \in \{1,2,\dots,M \} .
\end{align*}
In the case {\em 1.} $ (i,i+1)\sigma \not\in S_N^{\m} .$ \\
In the case {\em 2.} $ (i,i+1)\sigma \in S_{N,(m_s,m_{i_1},\dots,m_{i_L})}^{(m_1,m_2,\dots,m_M)} .
$
\noindent Assume that $ \sigma_i = \sigma_{i+1} + 1 .$
\noindent In this case $ \exists \; s \in \{i_1,i_2,\dots ,i_L\} $ {\em s.t. }
\begin{equation*}
\sigma_i = m_s + 1 \; , \; \sigma_{i+1} = m_s \quad \Leftrightarrow \quad i =
p^{\sigma}_{m_s+1} \; , \;i +1 = p^{\sigma}_{m_s}.
\end{equation*}
Since by definition of $S_{N,\mi}^{\m}$ :
\begin{gather*}
i < p_{m_s+2} < \ldots < p_{m_{s+1}} \, , \\
p_{m_{s-1}} < \ldots < p_{m_s-2} < p_{m_s-1}< i+1 ;
\end{gather*}
we have:
\begin{equation*}
(i,i+1)\sigma \in S_{N,(m_{i_1},\dots,m_{i_L})\setminus m_s}^{(m_1,m_2,\dots,m_M)}.
\end{equation*}
The first statement in {\em (iii)} is proven.
\noindent The second statement in {\em (iii)} is proven in a similar way.
\end{pf}
\subsection{A property of the eigenvectors $\varphi_{\sigma}^{\l}$ for $\l \in {\frak M}_N$}
{\bf 1.} In this section we shall derive a certain proprerty of the eigenspaces
${\cal E}^{\mu}\quad (\mu \in {\frak M}_N)$ of the operators $\Delta_1(u) , y_i (i=1,\dots,N). $
First of all, we notice that the eigenvalue $ \delta^{\mu}_1(u) $ of $ \Delta_1(u) $
associated with ${\cal E}^{\mu}$ can be represented in the additive
``particle'' form:
\begin{equation}
\delta^{\mu}_1(u) := \delta^{(m_1,m_2,\dots,m_M)}_1(u) = \sum_{k=1}^M \delta^{(m_k)}_1(u),
\end{equation}
where $ \delta^{(m)}_1(u) \quad m \in \{1,\dots,N-1\}$ is the one-particle
eigenvalue:
\begin{equation}
\delta^{(m)}_1(u) := u\sum_{i=1}^m \prod\begin{Sb} 1\leq j\leq N \\ j\neq i
\end{Sb} (1 + u q^{l_j})q^{l_i} .
\end{equation}
(Cf. 2.2.26 for the definition of $ l_i $ ).
\noindent {\bf 2.} Let $ \mu = (m_1,m_2,\dots,m_M) \in {\frak M}_N$ then the conjugate partition
$\mu\prime$ is: $\mu\prime = (m_M,m_{M-1},\dots,m_1)$.
Due to (2.2.9,.10) we have:
\begin{equation}
\Delta_1(u) e_{m_1}e_{m_2}\dots e_{m_M} = \left(\sum_{k=1}^M
\delta^{(m_k)}_1(u)\right)e_{m_1}e_{m_2}\dots e_{m_M}.
\end{equation}
While for any symmetric polynomial $S$ one has:
\begin{equation}
y_i S = q^{l_i} S \quad (i=1,\dots,N) .
\end{equation}
\mbox{}
\noindent Now we have the following proposition:
\begin{prop}
Let $ (m_1,m_2,\dots,m_M) \in {\frak M}_N\:,\; \{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \;\\ (0\leq L\leq M), $ and $
\{j_1,\dots,j_{M-L}\} = \{1,2,\dots,M \}\setminus\{i_1,i_2,\dots ,i_L\}. $
\noindent Then:
\begin{equation}
\varphi^{(m_1,m_2,\dots,m_M)}_{\sigma} \; (\sigma \in S_{N,\mi}^{\m})\; = \varphi^{(m_{i_1},\dots,m_{i_L})}_{\sigma}
e_{m_{j_1}}e_{m_{j_2}}\dots e_{m_{j_{M-L}}}.
\end{equation}
\end{prop}
\begin{pf}
Compute the action of the operators $\Delta_1(u) , y_i \; (i=1,\dots,N). $ on the polynomial $
\varphi^{(m_{i_1},\dots,m_{i_L})}_{\sigma} e_{m_{j_1}}e_{m_{j_2}}\dots
e_{m_{j_{M-L}}}.$
\begin{align}
y_i\varphi^{(m_{i_1},\dots,m_{i_L})}_{\sigma}S & =
q^{l_{\sigma_i}}\varphi^{(m_{i_1},\dots,m_{i_L})}_{\sigma}S \quad (i=1,\dots,N). \tag{a}
\end{align}
For any symmetric polynomial $ S $.
\begin{multline}
\Delta_1(u)\varphi^{(m_{i_1},\dots,m_{i_L})}_{\sigma}e_{m_{j_1}}e_{m_{j_2}}\dots
e_{m_{j_{M-L}}} = \\
= [\Delta_1(u)\,,\,\varphi^{(m_{i_1},\dots,m_{i_L})}]e_{m_{j_1}}e_{m_{j_2}}\dots
e_{m_{j_{M-L}}} + \tag{b} \\
+ \varphi^{(m_{i_1},\dots,m_{i_L})}\Delta_1(u) e_{m_{j_1}}e_{m_{j_2}}\dots
e_{m_{j_{M-L}}}.
\end{multline}
Due to the special form of $\Delta_1(u)$ (Cf. {\em Proposition 4}) the commutator in
the last formula is a zero-order differential operator. Therefore if $ S $ is a
symmetric polynomial we have:
\begin{equation}
[\Delta_1(u)\,,\,\varphi^{(m_{i_1},\dots,m_{i_L})}]S = f(u)S,
\end{equation}
where $f(u)$ is a function of $ z_i \; (i=1,\dots,N) $ independent of $S$.
Furthermore:
\begin{equation*}
\Delta_1(u) 1 = 0,
\end{equation*}
and therefore:
\begin{align*}
[\Delta_1(u)\,,\,\varphi^{(m_{i_1},\dots,m_{i_L})}]1 & =\Delta_1(u)
\varphi^{(m_{i_1},\dots,m_{i_L})} = \\
& = \sum_{1\leq k\leq L}^M
\delta^{(m_{i_k})}_1(u)\varphi^{(m_{i_1},\dots,m_{i_L})}.
\end{align*}
Taking {\em (b),(4.4.7,.11)} into account we get:
\begin{multline*}
\Delta_1(u)\varphi^{(m_{i_1},\dots ,m_{i_L})}_{\sigma}e_{m_{j_1}}e_{m_{j_2}}\dots
e_{m_{j_{M-L}}} = \\
= \left(\sum_{1\leq k\leq L} \delta^{(m_{i_k})} + \sum_{1\leq s\leq M - L}
\delta^{(m_{j_s})} \right) \varphi^{(m_{i_1},\dots
,m_{i_L})}_{\sigma}e_{m_{j_1}}e_{m_{j_2}}\dots e_{m_{j_{M-L}}}= \\
= \sum_{1\leq i\leq M} \delta^{(m_i)} \varphi^{(m_{i_1},\dots
,m_{i_L})}_{\sigma}e_{m_{j_1}}e_{m_{j_2}}\dots e_{m_{j_{M-L}}}.
\end{multline*}
\noindent Since the joint spectrum of $ \Delta_1(u) \,,\,y_i \; (i=1,\dots,N) $ is
multiplicity-free, we conclude from {\em (a)} and the last equation that:
\begin{equation}
\varphi^{(m_{i_1},\dots,m_{i_L})}_{\sigma}e_{m_{j_1}}e_{m_{j_2}}\dots
e_{m_{j_{M-L}}} = {\mathrm {const}}\, \varphi^{(m_1,m_2,\dots,m_M)}_{\sigma} \quad ( \sigma \in S_{N,\mi}^{\m} ) .
\end{equation}
By comparison of the maximal monomials in the both sides of the equation {\em
const} $ =1$.
\end{pf}
\mbox{}
\noindent Since $ e_r ( z_1=\omega^1,\dots,z_N = \omega^N ) = 0 \quad (1\leq r
\leq N-1)$
we have the following corollary to Proposition 9:
\begin{cor}
Let $ (m_1,m_2,\dots,m_M) \in {\frak M}_N\:,\; M > 0 \:,\; \sigma \in S_N^{\m} $.
\noindent Then:
\begin{align*}
\varphi^{(m_1,m_2,\dots,m_M)}_{\sigma}( z_1=\omega^1,\dots,z_N = \omega^N )& = 0 \\
\text{{\em unless}} & \quad \sigma \in S^{(m_1,m_2,\dots,m_M)}_{N,(m_1,m_2,\dots,m_M)}.
\end{align*}
\end{cor}
\subsection{Limit $q\rightarrow 0$ of the eigenfunctions of $\Delta_1(u) , y_i\quad
(i=1,\dots,N)$}
{\bf 1.} The aim of this subsection is to compute some of the eigenfunctions
$\varphi_{\sigma}^{\l}\quad (\l\in {\frak M}_N,\sigma \in \S)$ in the limit $q\rightarrow 0$ {\em under
assumption} that this limit is well-defined. We do not have a proof of the last
statement. In particular examples where $N$ is small this statement holds.
\mbox{}
\noindent Introduce operators:
\begin{equation}
\gamma_{i,j} := qg_{i,j}|_{q=0} = \frac{z_i}{z_i -z_j}(K_{i,j}- 1) \quad (i\neq
j\in \{1,\dots,N\} ).
\end{equation}
\begin{lemma}
Let $(m_1,m_2,\dots,m_M) \in {\frak M}_N$, and $\psi\in{\Bbb C}[z_1,\dots,z_N]$ satisfies the equations:
\begin{align}
(\gamma_{m_i,m_i+1} + 1 )\psi & = 0 \quad ( i=1,\dots,M), \\
\gamma_{n,n+1}\psi & = 0 \quad ( n\in
\{1,\dots,N-1\}\setminus\{m_1,\dots,m_M\}).
\end{align}
Then
\begin{equation}
\psi = (z_1\dots z_{m_1})^M(z_{m_1+1}\dots
z_{m_2})^{M-1}\ldots(z_{m_{M-1}+1}\dots z_{m_M}) S(z_1,\dots,z_N) ,
\end{equation}
where $S$ is a symmetric polynomial.
\end{lemma}
\begin{pf}
Let $N=2$ , $\psi = \psi(z_1,z_2)$.
The equation $ (\gamma_{1,2}+ 1)\psi = 0 $ implies that
\begin{equation}
\frac{z_1\psi(z_2,z_1) - z_2\psi(z_1,z_2)}{z_1 - z_2} = 0.
\end{equation}
This leads to $ \psi(z_1,z_2) = z_1 S(z_1,z_2)$, where $S$ is a symmetric
polynomial.
The equation
\begin{equation}
\gamma_{1,2}\psi = \frac{z_1(\psi(z_2,z_1) - \psi_(z_1,z_2))}{z_1 - z_2} = 0 ,
\end{equation}
yields $ \psi(z_1,z_2) = S(z_1,z_2) $ where $ S$ is a symmetric polynomial.
The case of arbitrary $N$ reduces to the case $ N=2$ by consideration of
consequtive pairs of coordinates.
\end{pf}
\mbox{}
\noindent {\bf 2.} We {\em conjecture} that the limit $q \rightarrow 0$ of
$\varphi_{\sigma}^{\l}$ is well-defined:
\begin{equation}
\varphi_{\sigma}^{\l}|_{q=0}:= \varphi_{\sigma}^{\l,0} = z^{\l_{\sigma}} + \text{``s.m.''} \qquad (\l \in
{\frak M}_N ,\; \sigma\in \S).
\end{equation}
In what follows we assume that the statement of the conjecture is valid.
Fix $(m_1,m_2,\dots,m_M) \in {\frak M}_N $. Our purpose is to find $ \varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0}$ for $
\sigma[0]:= (m_1,m_1+1)\dots(m_M,m_M+1).$ Take the limit $q\rightarrow 0$ in the
eq. (4.4.4) of Proposition 9. This yields
\begin{align}
(\gamma_{m_i,m_i+1} + 1) \varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0} & =
\varphi_{(m_i,m_i+1)\sigma[0]}^{(m_1,m_2,\dots,m_M),0}\quad (i = 1,\dots,M),\\ \gamma_{n,n+1}
\varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0} & = 0 \quad (n \in
\{1,\dots,N-1\}\setminus\{m_1,\dots,m_M\}).
\end{align}
According to Proposition 12, we have
\begin{multline}
\varphi_{(m_i,m_i+1)\sigma[0]}^{(m_1,m_2,\dots,m_M),0} =
\varphi_{(m_1,m_1+1)\dots\widehat{(m_i,m_i+1)}\dots (m_M,m_M+1)}^{(m_1,m_2,\dots,m_M),0} = \\
\varphi_{(m_1,m_1+1)\dots\widehat{(m_i,m_i+1)}\dots
(m_M,m_M+1)}^{(m_1,\dots,\hat{m_i},\dots,m_M ),0} e_{m_i} \quad (i =1,\dots,M).
\end{multline}
Where we put a hat over terms that are omitted.
The pair of equations (4.4.20,.21) provides a set of recurrent relations for
the eigenfunctions $\varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0}$. Notice that when $M=0$ we have $
\varphi_{\sigma[0]}^{(\emptyset),0} = 1$. Taking into account Lemma 4 we write the
general solution of these recurrent relations:
\begin{align}
\varphi_{\sigma[0]}^{(\emptyset),0} & = 1, \\
\varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0} & = (e_{m_1} - z_1\dots z_{m_1})(e_{m_2} - z_1\dots
z_{m_2})\ldots (e_{m_M} - z_1\dots z_{m_M})+ \notag\\ & + (z_1\dots
z_{m_1})^M(z_{m_1+1}\dots z_{m_2})^{M-1}\ldots(z_{m_{M-1}+1}\dots
z_{m_M})\times \notag \\ & \times S^{(m_1,m_2,\dots,m_M)}(z_1,\dots,z_N) \quad ( M \geq 1).
\notag
\end{align}
Where $ S^{(m_1,m_2,\dots,m_M)}(z_1,\dots,z_N) $ is an arbitrary symmetric polynomial.
Observe now that the total degree of the homogeneous polynomial $
\varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0} $ is equal to $ m_1 + \dots + m_M \quad ((m_1,m_2,\dots,m_M) \in {\frak M}_N) $.
Therefore we must have $ S^{(m_1,m_2,\dots,m_M)}(z_1,\dots,z_N) = {\mathrm {const}} $. Observe
further that
\begin{multline}
(z_1\dots z_{m_1})^M(z_{m_1+1}\dots z_{m_2})^{M-1}\ldots(z_{m_{M-1}+1}\dots
z_{m_M}) > \\
> {\mathrm {max}}( \varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0} ) \quad ( \sigma[0]:=
(m_1,m_1+1)\dots(m_M,m_M+1) ).
\end{multline}
Where $ {\mathrm {max}}( \varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0} ) $ is the maximal monomial of
$ \varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0}. $
Therefore we must have: $ S^{(m_1,m_2,\dots,m_M)}(z_1,\dots,z_N) = 0 $.
Thus for $ (m_1,m_2,\dots,m_M) \in {\frak M}_N $ :
\begin{multline}
\varphi_{\sigma[0]}^{(\emptyset),0} = 1, \\
\varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0} = (e_{m_1} - z_1\dots z_{m_1})(e_{m_2} - z_1\dots
z_{m_2})\ldots (e_{m_M} - z_1\dots z_{m_M})\notag \\ ( M \geq 1 ) \notag
\end{multline}
Notice that
\begin{multline}
\varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M),0}(z_1=\omega^1,\dots,z_N=\omega^N ) = (-1)^M
\omega^{\frac{1}{2}\sum_{i=1}^M m_i(m_i+1)} , \\ \quad ( (m_1,m_2,\dots,m_M) \in {\frak M}_N , \; \sigma[0]=
(m_1,m_1+1)\dots(m_M,m_M+1)).
\end{multline}
\section{Hecke-invariant (``bosonic'') subspaces of ${\cal E}^{\l}\otimes H$
for ${\l} \in {\frak M}_N $}
\subsection{Preliminaries}
{\bf 1.} Let ${\cal E}^{\l} \subset {\Bbb C}[z_1,\dots,z_N]$ be the eigenspace of the operators $
\Delta_1(u)\,,\,y_i\quad(i=1,\dots,N)$ parametrized by a partition $\l$ and $ H :=
({\Bbb C}^2)^{\otimes N}$. The bosonic subspace ${\cal B}^{\l}$ of ${\cal E}^{\l}\otimes
H $ is defined as follows:
\begin{equation}
B^{\l}:= \{ b \in {\cal E}^{\l}\otimes H | (g_{i,i+1} -t_{i,i+1})b = 0\quad (i=1,\dots,N-1) \}.
\end{equation}
Since $ {\cal P}:={\Bbb C}[z_1,\dots,z_N]\otimes H = \oplus_{\l}({\cal E}^{\l}\otimes H)$ and $ g_{i,i+1}
:{\cal E}^{\l}\mapsto {\cal E}^{\l}\quad(i=1,\dots,N-1) ;$ we have: $ {\cal B}
=\oplus_{\l}{\cal B}^{\l}. $ (Cf. (1.1.36), (3.3.7) for the definition of ${\cal B}$).
\noindent {\bf 2.} Any vector $\psi$ from ${\cal E}^{\l}\otimes H$ is
represented as follows:
\begin{equation}
\psi = \sum_{\sigma\in\S}\varphi_{\sigma}^{\l} \chi_{\sigma} ,
\end{equation}
where $\chi_{\sigma} \quad (\sigma \in \S) \in H $.
\mbox{}
\noindent The condition $ (g_{i,i+1} -t_{i,i+1})\psi = 0\quad(i=1,\dots,N-1) $ gives a set of linear
equations which must be satisfied by the vectors $\chi_{\sigma} \quad (\sigma \in \S) \in H
.$ In order to derive these equations we can apply the result of Proposition 9
to find out the action of $g_{i,i+1}\quad(i=1,\dots,N-1) $ on $\psi$, and then use the
linear-independence of the polynomials $ \varphi_{\sigma}^{\l}\quad (\sigma\in\S)$. In this way we
arrive at the following proposition:
\begin{prop}
A vector $ \psi \in {\cal E}^{\l}\otimes H ;\; \psi = \sum_{\sigma\in\S}\varphi_{\sigma}^{\l} \chi_{\sigma} $
belongs to $ {\cal B}^{\l} $ iff $\chi_{\sigma} \quad (\sigma \in \S) \in H $ satisfy the following
set of equations $ (i=1,\dots,N-1) $:
\begin{gather*}
q\frac{(q^{-1}q^{l_{\sigma_{i+1}}} - qq^{l_{\sigma_i}})(q^{-1}q^{l_{\sigma_i}} -qq^{l_{\sigma_{i+1}}})}{q^{l_{\sigma_i}} - q^{l_{\sigma_{i+1}}}}\chi_{(i,i+1)\sigma} =\tag{a} \\ =
\left( (q^{l_{\sigma_{i+1}}}-q^{l_{\sigma_i}})t_{i,i+1} - \dq^{l_{\sigma_{i+1}}}\right)\chi_{\sigma} \\
\text{{\em when}}\quad \l_{\sigma_{i+1}} > \l_{\sigma_i }\;, \\
(t_{i,i+1} - q ) \chi_{\sigma} = 0 \tag{b} \\
\text{{\em when}}\quad \l_{\sigma_{i+1}} = \l_{\sigma_i } \Leftrightarrow (i,i+1)\sigma
\not\in \S \Rightarrow \sigma_{i+1} = \sigma_i+1 \; , \\
\chi_{(i,i+1)\sigma} = \frac{ (q^{l_{\sigma_{i+1}}} -q^{l_{\sigma_i}})t_{i,i+1} - \dq^{l_{\sigma_{i+1}}}}{q^{-1}(q^{l_{\sigma_{i+1}}} - q^{l_{\sigma_i}})}\chi_{\sigma} \tag{c}\\
\text{{\em when}}\quad \l_{\sigma_{i+1}} < \l_{\sigma_i } \; .
\end{gather*}
\end{prop}
\subsection{ Spaces ${\cal B}^{\mu}$ for $ \mu \in {\frak M}_N $ }
{\bf 1.} Let us fix a partition $\mu \in {\frak M}_N$ parametrized by a motif $ (m_1,m_2,\dots,m_M) $.
We use the same notation $(m_1,m_2,\dots,m_M)$ for both the motf and the partition.
\noindent For $\mu \in {\frak M}_N$ let us further analyse the equations (a)-(c)
obtained in the Proposition 10. In section {\bf 4.3} (Definition 3) we
introduced the decomposition of the set $ S_N^{\m} $ into disjoint subsets:
\begin{equation}
S_N^{\m} = \bigsqcup_{\{i_1,i_2,\dots ,i_L\} \in \{1,2,\dots,M \}} S_{N,\mi}^{\m} \qquad (0\leq L\leq M)
\end{equation}
This decomposition is reflected in the equations (a)-(c) of the Proposition 10.
\noindent {\bf 2.} Let in these equations $ \sigma\,,\, i $ be such that $ \sigma \,
,\, (i,i+1)\sigma \in S_{N,\mi}^{\m} $ for some $ \{i_1,i_2,\dots ,i_L\} \in \{1,2,\dots,M \} $. According to Proposition
8,(iii) $ |\sigma_i - \sigma_{i+1}| \geq 2.$
Let $ \sigma_i - \sigma_{i+1} \geq 2 $ and consequently $ \mu_{\sigma_i} < \mu_{\sigma_{i+1}}
$. Since $ \sigma_i - \sigma_{i+1} \geq 2 $ the coefficient in front of $ \chi_{(i,i+1)\sigma} $ in
Pr.10(a) is not equal to zero. Therefore we have:
\begin{multline}
\chi_{(i,i+1)\sigma} = \frac{(q^{l_{\sigma_i}}-q^{l_{\sigma_{i+1}}})((q^{l_{\sigma_{i+1}}}-q^{l_{\sigma_i}})t_{i,i+1}
-\dq^{l_{\sigma_{i+1}}})}{q(q^{-1}q^{l_{\sigma_{i+1}}}-qq^{l_{\sigma_i}})(q^{-1}q^{l_{\sigma_i}}-qq^{l_{\sigma_{i+1}}})}\chi_{\sigma} \\
( \sigma , (i,i+1)\sigma \in S_{N,\mi}^{\m}\, ,\, \sigma_i - \sigma_{i+1} \geq 2 ).
\end{multline}
\mbox{}
\noindent Let $\sigma_i - \sigma_{i+1} \leq -2$ and consequently $ \mu_{\sigma_i} >
\mu_{\sigma_{i+1}} $. Pr.10(c) gives:
\begin{multline}
\chi_{(i,i+1)\sigma} = \frac{((q^{l_{\sigma_{i+1}}}-q^{l_{\sigma_i}})t_{i,i+1} -\dq^{l_{\sigma_{i+1}}})}{q^{-1}(q^{l_{\sigma_{i+1}}}-q^{l_{\sigma_i}})}\chi_{\sigma} \\
( \sigma , (i,i+1)\sigma \in S_{N,\mi}^{\m}\, ,\, \sigma_i - \sigma_{i+1} \leq -2 ).
\end{multline}
\mbox{}
\noindent Introduce a pair of $U$-intertwiners:
\begin{equation}
Y^{\pm}(z) : = \varrho^{\pm}(z)\frac{zt - t^{-1}}{q^{-1}z - q} \; \in
End(V\otimes V),
\end{equation}
where
\begin{equation}
\varrho^+(z) := \frac{z - 1}{q^2z - 1 }\; , \qquad \varrho^-(z) := \frac{z -
q^2}{z - 1 }.
\end{equation}
The eq. (5.5.4,.5) can be reformulated as follows:
Let $ \sigma , (i,i+1)\sigma \in S_{N,\mi}^{\m} \\ \quad (\{i_1,i_2,\dots ,i_L\} \in \{1,2,\dots,M \} \quad (0\leq L\leq M) ) $ then:
\begin{equation}
\chi_{(i,i+1)\sigma} =
\begin{cases}
Y^+_{i,i+1}(q^{l_{\sigma_i}-l_{\sigma_{i+1}}}) \chi_{\sigma} & \text{when $ \sigma_i - \sigma_{i+1} \geq
2 \Rightarrow \mu_{\sigma_i} < \mu_{\sigma_{i+1}}$ }, \\
Y^-_{i,i+1}(q^{l_{\sigma_i}-l_{\sigma_{i+1}}}) \chi_{\sigma} & \text{when $ \sigma_i - \sigma_{i+1} \leq
-2 \Rightarrow \mu_{\sigma_i} > \mu_{\sigma_{i+1}}$}.
\end{cases}
\end{equation}
Notice that when $ \sigma , (i,i+1)\sigma \in S_{N,\mi}^{\m} $ the intertwiners
$Y^{\pm}_{i,i+1}(q^{l_{\sigma_i}-l_{\sigma_{i+1}}})$ are invertible.
\mbox{}
\noindent {\bf 3.} Now consider the situation when $ \sigma$ and $(i,i+1)\sigma$ in
Proposition 10 belong to different subsets $S_{N,\mi}^{\m}$. According to Proposition 8
this situation takes place when $|\sigma_i - \sigma_{i+1} | = 1.$
\noindent Let $ \sigma_{i+1}= \sigma_i + 1$ and consequently $ \mu_{\sigma_i} >
\mu_{\sigma_{i+1}}.$ In this case $ \exists \; s \in \{1,2,\dots,M \} $ s.t. $ \sigma_i = m_s \, ,
\: \sigma_{i+1} = m_s + 1 $ or , equivalently , $ i = p^{\sigma}_{m_s} \,,\: i+1 =
p^{\sigma}_{m_s+1}$; and if $ \sigma \in S_{N,\mi}^{\m} $ then $ s \not\in \{i_1,i_2,\dots ,i_L\} .$ Since $
((i,i+1)\sigma)_i = m_s+1\;,\; ((i,i+1)\sigma)_{i+1} = m_s ;$ we have $ (i,i+1)\sigma \in
S_{N,(m_s,m_{i_1},\dots,m_{i_L})}^{(m_1,m_2,\dots,m_M)} $.
\noindent Substituting $ \sigma_{i+1}= \sigma_i + 1$ into Pr.10(c) we find:
\begin{multline}
\chi_{(i,i+1)\sigma} = q(t_{i,i+1} - q)\chi_{\sigma} = -(q^2 + 1)\Pi_{i,i+1}^-(q)\chi_{\sigma} \\
( \sigma \in S_{N,\mi}^{\m} \;,\; (i,i+1)\sigma \in S_{N,(m_s,m_{i_1},\dots,m_{i_L})}^{(m_1,m_2,\dots,m_M)}\;,\;
\\ \sigma_i = m_s , \sigma_{i+1} = m_s + 1 ( s \in \{1,2,\dots,M \}\setminus\{i_1,i_2,\dots ,i_L\} ) ).
\end{multline}
(Cf. (1.1.12,.13) for the definition of projectors $\Pi_{i,i+1}^{\pm}(q)$).
\mbox{}
\noindent Let $ \sigma_i= \sigma_{i+1} + 1$ and consequently $ \mu_{\sigma_i} <
\mu_{\sigma_{i+1}}.$ In this case $ \exists \; s \in \{1,2,\dots,M \} $ s.t. $ \sigma_i = m_s+1 \,
, \: \sigma_{i+1} = m_s $ ; and since $ \sigma \in S_{N,\mi}^{\m} \: ; \; s \in \{i_1,i_2,\dots ,i_L\} .$ On the
other hand $ (i,i+1)\sigma \in S_{N,(m_{i_1},\dots,m_{i_L})\setminus m_s}^{(m_1,m_2,\dots,m_M)}.$
\\ Here $(m_{i_1},\dots,m_{i_L})\setminus m_s$ signifies the motif obtained
from $( m_{i_1},\dots,m_{i_L})$ by removing $m_s$.
\noindent Substituting $ \sigma_i= \sigma_{i+1} + 1$ into Pr.10(a) we find:
\begin{multline}
(t_{i,i+1} + q^{-1})\chi_{\sigma} = 0 \quad \Rightarrow \quad \Pi_{i,i+1}^+(q)\chi_{\sigma} = 0 \\
( \sigma \in S_{N,\mi}^{\m} \;,\; (i,i+1)\sigma \in S_{N,(m_{i_1},\dots,m_{i_L})\setminus
m_s}^{(m_1,m_2,\dots,m_M)}\;,\; \\ \sigma_i = m_s+1 , \sigma_{i+1} = m_s ( s \in \{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} )
).
\end{multline}
\mbox{}
\noindent {\bf 4.} Now we are in position to reformulate Propositon 10 in a
more suitable form in the case when $ \l \equiv \mu \equiv (m_1,m_2,\dots,m_M) \in {\frak M}_N $.
\begin{prop}
Let $ \psi \in {\cal E}^{\mu}\otimes H$, i.e.
\begin{equation}
\psi = \sum_{\sigma \in S_N^{\m}}\varphi^{(m_1,m_2,\dots,m_M)}_{\sigma} \chi_{\sigma} \qquad ( \chi_{\sigma} \in H ).
\end{equation}
Then $ \psi \in {\cal B}^{(m_1,m_2,\dots,m_M)} $ iff $ \chi_{\sigma} \quad (\sigma \in S_N^{\m}) $ satisfy the
following linear relations:
\begin{multline}
\forall \; \{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \quad (0\leq L\leq M) \\
\chi_{(m_{i_1},m_{i_1}+1)\dots (m_{i_L},m_{i_L}+1)} = \\
= -(q^2 + 1) \Pi_{m_{i_k},m_{i_k}+1}^-(q) \chi_{(m_{i_1},m_{i_1}+1)\dots
\widehat{ (m_{i_k},m_{i_k}+1)}\dots (m_{i_L},m_{i_L}+1)} \\
(k = 1,2,\dots,L).
\end{multline}
Where $\widehat{\mbox{}}$ means that the corresponding factor is omitted from
the product.
\begin{multline}
\forall \; \{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \quad (0\leq L\leq M) \\
\text{{\em and }} \; j \in \{1,\dots,N-1\} \; \text{{\em s.t.}} \\
\{j,j+1\}\cap \{m_1,m_1+1,m_2,m_2+1,\dots,m_M,m_M+1\} = \emptyset \; ; \\
\Pi_{j,j+1}^-(q)\chi_{(m_{i_1},m_{i_1}+1)\dots (m_{i_L},m_{i_L}+1)} = 0.
\end{multline}
\begin{gather}
\forall \; \sigma \in S_{N,\mi}^{\m} \quad ( \{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \quad (0\leq L\leq M) ) \\
\chi_{\sigma} = {\Bbb Y}(\sigma)\chi_{(m_{i_1},m_{i_1}+1)\dots (m_{i_L},m_{i_L}+1)}. \notag
\end{gather}
Where invertible ${\Bbb Y}(\sigma)\;\in\; End(H)$ is recursively defined as follows:
\begin{gather*}
{\Bbb Y}((m_{i_1},m_{i_1}+1)\dots (m_{i_L},m_{i_L}+1)) := {\mathrm {Id}} \, , \\
\text{{\em for}} \; (i,i+1)\sigma \, \in \, S_{N,\mi}^{\m} \\
{\Bbb Y}((i,i+1)\sigma) =
\begin{cases}
Y^+_{i,i+1}(q^{l_{\sigma_i}-l_{\sigma_{i+1}}}){\Bbb Y}(\sigma) & \text{{\em if}}\; \sigma_i -
\sigma_{i+1} \geq \mbox{} 2\, , \\
Y^-_{i,i+1}(q^{l_{\sigma_i}-l_{\sigma_{i+1}}}){\Bbb Y}(\sigma) & \text{{\em if}}\; \sigma_i -
\sigma_{i+1} \leq -2\, .
\end{cases}
\end{gather*}
\end{prop}
\mbox{}
\noindent It is possible to give more explicit expression for the matrix $
{\Bbb Y}(\sigma) $ that appears in (5.5.14). In notation of Proposition 8 (ii) we have:
\begin{align}
\text{For} & \; \sigma \in S_{N,\mi}^{\m} \qquad ( \{i_1,i_2,\dots ,i_L\} \subset \{1,2,\dots,M \} \; (0\leq L\leq M) ) \notag \\
{\Bbb Y}(\sigma) & = Y^-_{j_r,j_r+1}(q^{l_{\sigma [r-1]_{j_r}}-l_{\sigma [r-1]_{j_r+1}}}) \dots
Y^-_{j_1,j_1+1}(q^{l_{\sigma [0]_{j_1}}-l_{\sigma [0]_{j_1+1}}}).
\end{align}
Where:
\begin{align}
\sigma&[k] := (j_k,j_k+1)\dots (j_1,j_1+1)\sigma[0] \in S_{N,\mi}^{\m} \; ( 1\leq k \leq r),
\notag \\
\sigma&[0] := (m_{i_1},m_{i_1}+1)\dots (m_{i_L},m_{i_L}+1) , \notag \\
\sigma&[r] = \sigma \notag \\
\sigma&[k-1]_{j_k} - \sigma[k-1]_{j_k+1} \leq -2 \quad (1 \leq k \leq r ).
\end{align}
\mbox{}
\noindent {\bf 5.} Introduce two definitions.\\
For $ (m_1,m_2,\dots,m_M) \in {\frak M}_N $ :
\begin{equation}
{\cal Z}^{(m_1,m_2,\dots,m_M)}:= \left\{ v \in H \; \begin{array}{|c } \Pi_{j,j+1}^-(q)v = 0
\quad \text{for all}\quad j\in\{1,\dots,N-1\}\;\text{s.t.} \\
\{j,j+1\}\cap\{m_1,m_1+1,\dots,m_M,m_M+1\}=\emptyset \end{array} \right\}
\end{equation}
More explicitely:
\begin{multline*}
{\cal Z}^{(m_1,m_2,\dots,m_M)} =
S_q(\underset{1}{V}\otimes\dots\otimes\underset{m_1-1}{V})\otimes\underset{m_1}{V}\otimes\underset{m_1+1}{V}\otimes S_q(\underset{m_1+2}{V}\otimes\dots\otimes\underset{m_2-1}{V})\otimes\underset{m_2}{V}\otimes\underset{m_2+1}{V}\otimes\ \\ \ldots \otimes S_q(\underset{m_M+2}{V}\otimes\dots\otimes\underset{N}{V}) \; \subset \; H := \underset{1}{V}\otimes\underset{2}{V}\otimes\ldots\otimes\underset{N}{V}.
\end{multline*}
Where $S_q$ means $q$-symmetrization and the subscripts indicate positions of
the factors in the tensor product $ V^{\otimes N}.$ \\
\mbox{} \\
For any $ \{i_1,i_2,\dots ,i_L\} \in \{1,2,\dots,M \} \quad (0\leq L\leq M) $ define the following projector:
\begin{equation}
\Pi^{-,(m_1,m_2,\dots,m_M)}_{(m_{i_1},\dots,m_{i_L})}:=
\begin{cases}\Pi^-_{m_{i_1},m_{i_1}+1}(q)\dots\Pi^-_{m_{i_L},m_{i_L}+1}(q) &
\quad (1\leq L \leq M), \\
{\mathrm I} & (L =0) .\end{cases}
\end{equation}
\mbox{}
\noindent {\bf 6.} Proposition 11 yields the following expression for $
{\cal B}^{\mu}\: \equiv \:{\cal B}^{(m_1,m_2,\dots,m_M)} $:
\begin{equation}
{\cal B}^{(m_1,m_2,\dots,m_M)} = \BU^{(m_1,m_2,\dots,m_M)}(z){\cal Z}^{(m_1,m_2,\dots,m_M)}.
\end{equation}
Where
\begin{multline}
\BU^{(m_1,m_2,\dots,m_M)}(z) := \\ \sum\begin{Sb} I \subset \{1,2,\dots,M \} \\ I:=\{i_1,i_2,\dots ,i_L\}
\end{Sb}(-(q^2+1))^L\left\{\sum_{\sigma\inS_{N,\mi}^{\m}}\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(z){\Bbb Y}(\sigma)\right\}\Pi^{-,(m_1,m_2,\dots,m_M)}_{(m_{i_1},\dots,m_{i_L})} \; , \\
\BU^{(m_1,m_2,\dots,m_M)}(z) : H \mapsto {\Bbb C}[z_1,\dots,z_N]\otimes H .
\end{multline}
In the last formula we explicitely indicated $z$-dependence of the polynomials
$\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}$.
\mbox{}
\noindent {\bf 7.} The space $ {\cal B}^{(m_1,m_2,\dots,m_M)}$ is a $U$-module where the action of $U$
is given by (2.2.19):
\begin{equation}
T_a^0(u) = L_a(u;\{y_i\}):= L_{a1}(uy_1)L_{a2}(uy_2)\dots L_{aN}(uy_N) \; \in
\; End({\Bbb C}[z_1,\dots,z_N]\otimes H).
\end{equation}
The space ${\cal Z}^{(m_1,m_2,\dots,m_M)}$ is a (reducible,indecomposable) $U$-module as well,
with the $U$-action:
\begin{equation}
L_a(u;\{q^{l_i}\}):= L_{a1}(uq^{l_1})L_{a2}(uq^{l_2})\dots L_{aN}(uq^{l_N}) \;
\in \; End(H).
\end{equation}
The operator $\BU^{(m_1,m_2,\dots,m_M)}(z)$ in (5.5.19,.20) is an $U$-intertwiner of these two
modules. To see this let us consider the product $T_a^0(u)\BU^{(m_1,m_2,\dots,m_M)}(z)$.Since
$\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(z)$ are eigenvectors of $ y_i \; (i=1,\dots,N) $ (Cf. Proposition 6)
we get:
\begin{multline}
T_a^0(u)\BU^{(m_1,m_2,\dots,m_M)}(z) = \\
\sum\begin{Sb} I \subset \{1,2,\dots,M \} \\ I:=\{i_1,i_2,\dots ,i_L\} \end{Sb}(-(q^2+1))^L \; \times \\
\times \; \left\{\sum_{\sigma\inS_{N,\mi}^{\m}}\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(z)
L_a(u;\{q^{l_{\sigma_i}}\}){\Bbb Y}(\sigma)\right\}\Pi^{-,(m_1,m_2,\dots,m_M)}_{(m_{i_1},\dots,m_{i_L})}.
\end{multline}
It follows from the recursive definition of ${\Bbb Y}(\sigma)\quad (\sigma \in S_{N,\mi}^{\m})$ given
in Proposition 11 that ${\Bbb Y}(\sigma)$ is an $U$-intertwiner:
\begin{equation}
L_a(u;\{q^{l_{\sigma_i}}\}){\Bbb Y}(\sigma) = {\Bbb Y}(\sigma)L_a(u;\{q^{l_{\sigma[0]_i}}\}).
\end{equation}
Where $ \sigma[0] = (m_{i_1},m_{i_1}+1)\dots (m_{i_L},m_{i_L}+1)$.\\
Furthermore the projector $\Pi^{-,(m_1,m_2,\dots,m_M)}_{(m_{i_1},\dots,m_{i_L})}$ is an
interwiner as well (Cf. 1.1.28):
\begin{equation}
L_a(u;\{q^{l_{\sigma[0]_i}}\}\Pi^{-,(m_1,m_2,\dots,m_M)}_{(m_{i_1},\dots,m_{i_L})} =
\Pi^{-,(m_1,m_2,\dots,m_M)}_{(m_{i_1},\dots,m_{i_L})}L_a(u;\{q^{l_i}\}).
\end{equation}
Hence we obtain:
\begin{equation}
T_a^0(u)\BU^{(m_1,m_2,\dots,m_M)}(z) = \BU^{(m_1,m_2,\dots,m_M)}(z)L_a(u;\{q^{l_i}\}).
\end{equation}
Thus the intertwinig property of $ \BU^{(m_1,m_2,\dots,m_M)}(z)$ is established.
\section{Spectrum and eigenspaces of the operators $\Xi^{(n)}(\omega) $ $
(n=1,\dots,N-1)$ forming the
hierarchy of $U$-invariant Spin Models.}
\subsection{Eigenspaces $H^{\mu}_{\bf{B}}(\omega)$ of the generating function
$\Xi(u;\omega) $}
{\bf 1.} Let $\mu \equiv (m_1,m_2,\dots,m_M) \in {\frak M}_N .$ Define a subspace $H^{\mu}_{{\cal B}}(\omega)
\subset H_{{\cal B}}(\omega) \subset H $ by applying the evaluation map
$Ev(\omega)$ (Cf. {\bf 3.1}) to the ``bosonic'' subspace $ {\cal B}^{(m_1,m_2,\dots,m_M)} $ introduced
in the previous section:
\begin{equation}
H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) := Ev(\omega){\cal B}^{(m_1,m_2,\dots,m_M)}\quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N).
\end{equation}
{}From (3.3.13,.14,.19) we obtain
\begin{multline}
\Xi(u;\omega) H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) = \left(\sum_{i=1}^M (\delta^{(m_i)}_1(u) -
\theta(u)m_i)\right)H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) \\ ((m_1,m_2,\dots,m_M) \in{\frak M}_N).
\end{multline}
Where $ \delta^{(m_i)}_1(u) $ was defined in (4.4.6) and $\theta(u)$ was
defined in (2.2.6). The last equation says that $ H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) $ is an
eigenspace of $\Xi(u;\omega)$ unless $ H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) \equiv 0 $.
\mbox{}
\noindent {\bf 2.} Application of the evaluation map $ Ev(\omega) $ to
${\cal B}^{(m_1,m_2,\dots,m_M)}$ (5.5.19) yields the explicit expression for the space
$H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)$:
\begin{equation}
H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) = \BU^{(m_1,m_2,\dots,m_M)}(\omega){\cal Z}^{(m_1,m_2,\dots,m_M)}\quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N).
\end{equation}
Where
\begin{multline}
\BU^{(m_1,m_2,\dots,m_M)}(\omega) := \BU^{(m_1,m_2,\dots,m_M)}(z_1 =\omega^1,\dots,z_N = \omega^N) = \\
= (-(q^2+1))^M \sum_{\sigma\in S^{(m_1,m_2,\dots,m_M)}_{N,(m_1,m_2,\dots,m_M)}}\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(\omega){\Bbb Y}(\sigma)
\Pi^{-,(m_1,m_2,\dots,m_M)}_{\{1,2,\dots,M \}} \; , \\
\BU^{(m_1,m_2,\dots,m_M)}(\omega) : H \mapsto H .
\end{multline}
To get the expression for $\BU^{(m_1,m_2,\dots,m_M)}(\omega)$ we used the Corollary 1 to
Proposition 9.
\mbox{}
\noindent Let us introduce the notation
\begin{equation}
W^{(m_1,m_2,\dots,m_M)} := \Pi^{-,(m_1,m_2,\dots,m_M)}_{\{1,2,\dots,M \}}{\cal Z}^{(m_1,m_2,\dots,m_M)}\quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N).
\end{equation}
The definition of ${\cal Z}^{(m_1,m_2,\dots,m_M)}$ given in (5.5.17) leads to a more explicit
form of $W^{(m_1,m_2,\dots,m_M)}$:
\begin{multline}
W^{(m_1,m_2,\dots,m_M)} = \\ S_q(\underset{1}{V}\otimes\dots\otimes\underset{m_1-1}{V})\otimes
A_q(\underset{m_1}{V}\otimes\underset{m_1+1}{V})\otimes
S_q(\underset{m_1+2}{V}\otimes\dots\otimes\underset{m_2-1}{V})\otimes
A_q(\underset{m_2}{V}\otimes\underset{m_2+1}{V})\otimes\ \\ \ldots \otimes
S_q(\underset{m_M+2}{V}\otimes\dots\otimes\underset{N}{V}) \; \subset \; H :=
\underset{1}{V}\otimes\underset{2}{V}\otimes\ldots\otimes\underset{N}{V}.
\end{multline}
Where $A_q$ signifies $q$-antisymmetrization (1.1.15).
Observe that $W^{(m_1,m_2,\dots,m_M)}\quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N)$ is an irreducible highest-wight
$U$-module with the $U$-action is given by
\begin{equation} L_a(u;\{q^{l_{\sigma[0]_i}}\}):=
L_{a1}(uq^{l_{\sigma[0]_1}})L_{a2}(uq^{l_{\sigma[0]_2}})\dots L_{aN}(uq^{l_{\sigma[0]_N}})
\; \in \; End(H).
\end{equation}
Where we used the notation
\begin{equation}
\sigma[0] := (m_1,m_1+1)(m_2,m_2+1)\ldots (m_M,m_M+1) \quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N).
\end{equation}
The Drinfeld polynomial of this module is
\begin{equation}
Q^{(m_1,m_2,\dots,m_M)}(u) = \prod\begin{Sb}1\leq n \leq N \\ n \neq m_i,m_i+1 \end{Sb}(1 -
q^{-l_n}u).
\end{equation}
\mbox{}
\noindent According to (6.6.3) we have
\begin{multline}
H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) = \Check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega)W^{(m_1,m_2,\dots,m_M)}\quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N), \\
\Check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega):= (-(q^2+1))^M \sum_{\sigma\in
S^{(m_1,m_2,\dots,m_M)}_{N,(m_1,m_2,\dots,m_M)}}\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(\omega){\Bbb Y}(\sigma).
\end{multline}
\mbox{}
\noindent {\bf 3.} The space $ H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) \quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N)$ is a
$U$-module with the $U$-action given by $T_a^0(u;\omega)$ defined in (3.3.19). Explicitely
(Cf. 5.5.23):
\begin{multline}
T_a^0(u;\omega) H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) = T_a^0(u;\omega)\Check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega)W^{(m_1,m_2,\dots,m_M)} = \\
=(-(q^2+1))^M \sum_{\sigma\in S^{(m_1,m_2,\dots,m_M)}_{N,(m_1,m_2,\dots,m_M)}}\varphi_{\sigma}^{(m_1,m_2,\dots,m_M)}(\omega)
L_a(u;\{q^{l_{\sigma_i}}\}){\Bbb Y}(\sigma), \\
L_a(u;\{q^{l_{\sigma_i}}\}):= L_{a1}(uq^{l_{\sigma_1}})L_{a2}(uq^{l_{\sigma_2}})\dots
L_{aN}(uq^{l_{\sigma_N}}).
\end{multline}
Applying (5.5.24) we find that $ \Check{\BU}^{(m_1,m_2,\dots,m_M)}$ is an intertwiner of the
modules $H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)$ and $ W^{(m_1,m_2,\dots,m_M)}$:
\begin{multline}
T_a^0(u;\omega) H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) = T_a^0(u;\omega)\Check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega)W^{(m_1,m_2,\dots,m_M)} = \\
\Check{\BU}^{(m_1,m_2,\dots,m_M)}(\omega)L_a(u;\{q^{l_{\sigma[0]_i}}\})W^{(m_1,m_2,\dots,m_M)} \quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N).
\end{multline}
Since $ W^{(m_1,m_2,\dots,m_M)}$ is irreducible so is $H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)$. The highest-weight
vector of $ W^{(m_1,m_2,\dots,m_M)}$ is
(Cf. 6.6.6)
\begin{multline}
\tilde{\Omega}^{(m_1,m_2,\dots,m_M)} := \\ v^+\otimes\dots\otimes
v^+\otimes(\underset{m_1}{v^+}\otimes\underset{m_1+1}{v^-}-q v^-\otimes
v^+)\otimes \quad\ldots \quad\otimes v^+\otimes\dots\otimes v^+.
\end{multline}
\noindent If the vector $\Omega^{(m_1,m_2,\dots,m_M)} := \Check{\BU}^{(m_1,m_2,\dots,m_M)}\tilde{\Omega}^{(m_1,m_2,\dots,m_M)} \in
H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)$ is not zero, it is the highest-weight vector, and the
modules $H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)$ and $ W^{(m_1,m_2,\dots,m_M)}$ are isomorphic , with
$\Check{\BU}^{(m_1,m_2,\dots,m_M)}$ defining the isomorphism explicitely.
\mbox{}
\noindent {\bf 4.} In order to show that $\Omega^{(m_1,m_2,\dots,m_M)} \quad ((m_1,m_2,\dots,m_M) \in{\frak M}_N) $ is
not zero we compute this vector at $q =0$.
\noindent Consider the matrix ${\Bbb Y}(\sigma)\quad (\sigma \in S_{N,(m_1,m_2,\dots,m_M)}^{(m_1,m_2,\dots,m_M)})$ that enters
the definition of $\Check{\BU}^{(m_1,m_2,\dots,m_M)}$. According to (5.5.15,.16):
\begin{align}
{\Bbb Y}(\sigma) & = {\mathrm {Id}} \quad ( \sigma = \sigma[0]:= (m_1,m_1+1)\dots(m_M,m_M+1)),
\\
{\Bbb Y}(\sigma) & = {\Bbb Y}(\sigma)' Y^-_{i,i+1}(q^{l_{\sigma[0]_i}-l_{\sigma[0]_{i+1}}}) \quad ( \sigma
\neq \sigma[0]).
\end{align}
Where depending on $\sigma$ , $i$ takes one of the values in the set $
\{m_k-1,m_k+1\}_{k\in\{1,2,\dots,M \}}$. For any such $i$ we have $ \sigma[0]_i - \sigma[0]_{i+1}
= -2 .$ ${\Bbb Y}(\sigma)'$ is either identity or a product of intertwiners of the form
$ Y^-_{j,j+1}(q^{-2r}) \quad j\in (\{1,\dots,N-1\})$ where $ r \geq 2 . $
\noindent For $r \geq 2$ we find
\begin{equation}
Y^-_{j,j+1}(q^{-2r})|_{q = 0} = -\Pi_{j,j+1}^-(0) = -(|+-><+-|)_{j,j+1} \quad
(j=1,\dots,N-1).
\end{equation}
Thus
\begin{multline}
\lim_{q\rightarrow 0}{\Bbb Y}(\sigma) = -{\Bbb Y}(\sigma)'|_{q=0}\Pi_{i,i+1}^-(0) \quad (i\in
\{m_k-1,m_k+1\}_{k\in\{1,2,\dots,M \}}), \\ \quad (\sigma \in S_{N,(m_1,m_2,\dots,m_M)}^{(m_1,m_2,\dots,m_M)},\; \sigma\neq \sigma[0] ).
\end{multline}
The highest-weight vector $ \tilde{\Omega}^{(m_1,m_2,\dots,m_M)}$ in the limit $ q\rightarrow 0
$ is
\begin{multline}
\tilde{\Omega}^{(m_1,m_2,\dots,m_M),q=0} := \\
v^+\otimes\dots\otimes\underset{m_1}{v^+}\otimes\underset{m_1+1}{v^-}\otimes
v^+\dots \otimes\underset{m_2}{v^+}\otimes\underset{m_2+1}{v^-}
\otimes\quad\ldots \quad\otimes v^+\otimes\dots\otimes v^+.
\end{multline}
Therefore taking into account (6.6.17) and (4.4.26) we arrive at the following
expression for $\Omega^{(m_1,m_2,\dots,m_M)} \in H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)$ at $q=0$:
\begin{multline}
\Omega^{(m_1,m_2,\dots,m_M),q=0} = (-1)^M \varphi_{\sigma[0]}^{(m_1,m_2,\dots,m_M)}(\omega) \tilde{\Omega}^{(m_1,m_2,\dots,m_M),q=0} =
\\
=\omega^{\frac{1}{2}\sum_{i=1}^M m_i(m_i+1)} \tilde{\Omega}^{(m_1,m_2,\dots,m_M),q=0}.
\end{multline}
Since $\Omega^{(m_1,m_2,\dots,m_M),q=0}$ is not zero we can argue that same holds for any
generic value of $q$ (not a root of unity) and therefore $ H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)
$ and $ W^{(m_1,m_2,\dots,m_M)} $ are isomorphic $U$-modules for any generic $q$ and $ (m_1,m_2,\dots,m_M) \\
\in {\frak M}_N .$
\mbox{}
\noindent {\bf 5.} From consideration of the case $q=0$ we deduce that
\begin{equation}
H = \bigoplus_{(m_1,m_2,\dots,m_M) \in {\frak M}_N} W^{(m_1,m_2,\dots,m_M)}.
\end{equation}
Since the $U$-modules $ H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega)$ have different Drinfeld
polynomials (6.6.9) for different motifs $(m_1,m_2,\dots,m_M)$ any two of these modules do not
intersect except at zero vector. Therefore we can take the direct sum of all
these modules. Since ${\mathrm{ dim}}H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) = {\mathrm
{dim}}W^{(m_1,m_2,\dots,m_M)} $, we conclude from (6.6.20) that
\begin{equation}
H = \bigoplus_{(m_1,m_2,\dots,m_M) \in {\frak M}_N} H^{(m_1,m_2,\dots,m_M)}_{{\cal B}}(\omega) .
\end{equation}
Thus we have found the complete decomposition of the space of states into
eigenspaces of the operator $\Xi(u;\omega)$, as well as the $U$-representation content
of this decomposition.
|
\section{Introduction}
Supersymmetry \cite{BaWe} provides a unified framework for the
discussion of bosons and fermions and has been proposed as a mechanism
to improve the behavior of quantum field theory divergences. It may
offer the only viable alternative at present for grand unified models
\cite{Ca}.
Because it allows to write the Hamiltonian in terms of its ``square
roots'' it may provide a new perspective on the constraint equations
\cite{DeHaOb,GrCs}.
Supergravity was cast in a canonical fashion several years ago by
Deser, Kay and Stelle \cite{DeKaSt} , Pilati \cite{Pi} and D'Eath
\cite{De}. Teitelboim \cite{TeTa} pointed out that it
provided a square root of the Wheeler-DeWitt equation. Jacobson
\cite{Ja} showed that new canonical variables similar to the ones
introduced by Ashtekar in general relativity could be used in
supergravity. Here we will follow this latter approach.
In this paper we will show that several ideas that have appeared in the
context of the loop quantization of general relativity in terms of the
Ashtekar new variables \cite{As} find a natural counterpart in supergravity.
There are already partial results concerning the use of new variables
and loops in the context of connection representations for
supergravity. F\"ul\"op \cite{Fu} first noticed that the theory could
be cast in terms of an $GSU(2)$ connection and discussed the
Chern-Simons form as a possible state. This state had already been
considered in a semi-classical context by Sano and Shiraishi
\cite{SaSh}. Matschull \cite{Ma} noted that bosonic Wilson loops were
spurious solutions to all the constraints of supergravity.
In this paper we will use $GSU(2)$ Wilson loops with nontrivial
fermionic content as an overcomplete basis in terms of which we can
expand quantum state. We
discuss the exponential of the $GSU(2)$ Chern-Simons form as a
diffeomorphism invariant solution to all the constraints of supergravity
with a cosmological constant with non-trivial fermionic content. We
will also build a loop representation for the theory and discuss the
kinematical state space. We will see that the loop counterpart of the
Chern-Simons solution is associated with the Dubrovnik version of
the Kauffman polynomial, which is therefore
compatible with the $GSU(2)$ Mandelstam identities.
\section{Quantum supergravity in terms of a $GSU(2)$ connection}
When cast in a canonical form in terms of the Ashtekar new variables
\cite{As}, supergravity is described by variables
$\tilde{E}^a_i,A_a^i,\psi_a^A,\tilde{\pi}^a_A$ where lowercase indices from
the beginning of the alphabet are spatial tensor indices, lowercase
indices from the middle of the alphabet are $SU(2)$ indices and
uppercase indices are spinor indices. The $E$'s are densitized triads,
the $A$'s are the Sen connection, the $\psi$'s are Grassman-valued
Rarita-Schwinger fields and the $\pi$'s are their canonically
conjugate momenta. We refer the reader to the paper by Jacobson
\cite{Ja} for details. We will follow the notation of F\"ul\"op
\cite{Fu}.
The canonical framework (with a cosmological constant) has the
following constraints,
\begin{eqnarray}
{\cal G}_i &=&
D_a \tilde{E}^a_i + {i\over \sqrt{2}} \pi^a_A \psi_{aB}
\sigma^{AB}_i=0
\label{gauss}\\
{\cal S}_A &=& D_a \pi^a_A -i \alpha \tilde{E}^a_i \sigma_{iA}^B \psi_{aB}\\
\bar{\cal S}^A &=& \epsilon^{ijk} \tilde{E}^a_i \tilde{E}^b_j
{\sigma_k}^A_B (-4 i D_{[a} \psi_{b]}^B+\sqrt{2}\bar{\alpha}
\ut{\epsilon}_{abc}
\pi^{cB}) =0
\end{eqnarray}
where $\sigma^{AB}_i$ are usual Pauli matrices and $\alpha$ and
$\bar{\alpha}$ are such that the cosmological constant is given by
$\Lambda = -\alpha \bar{\alpha}$.
The usual Hamiltonian and diffeomorphism constraints of general
relativity (with the corresponding fermionic extra terms) are obtained
by taking Poisson brackets of the ${\cal S}$ and $\bar{\cal S}$
constraints.
One could construct a quantum representation by considering wavefunctions
$\Psi(A,\psi)$ and promoting the above constraints to wave equations.
Matschull observed that if one does so and considers purely bosonic
wavefunctions consisting of the Wilson loops built with the bosonic
connection $A$ along smooth non-intersecting loops, all the contraints
are solved. This presented a puzzle, since such wavefunctions are
evidently not diffeomorphism invariant. How could they therefore
provide a solution to the diffeomorphism constraint? The answer is
given by the fact that the commutator of ${\cal S}$ with $\bar{\cal
S}$ give expressions for the usual Hamiltonian and diffeomorphism
constraint multiplied by the determinant of the spatial metric. Since
Wilson loops based on smooth loops are annihilated by the determinant
of the metric, they automatically solved the resulting constraints.
Here we will proceed in a different way. F\"ul\"op noticed that if one
considers the Gauss law and the right supersymmetry generator, they
form under Poisson brackets a graded Lie algebra associated with the
$GSU(2)$ group \cite{PaRi},
\begin{eqnarray}\label{4}
\{{\cal G}_i,{\cal G}_j\} &=& \sqrt{2} \epsilon_{ijk} {\cal G}_k\\
\{{\cal G}_j, {\cal S}_A\} &=&-
{i\over \sqrt{2}} (\sigma_j)^B{}_A
{\cal S}_B\\\label{6}
\{{\cal S}_A,{\cal S}_B\} &=& -i \alpha (\sigma^j)_{AB} {\cal G}_j.
\end{eqnarray}
In view of this, one can define new $GSU(2)$ variables $\tilde{\bf
E}^a_I$, ${\bf A}_a^J$, ${\bf F}_{ab}^K$,
\begin{eqnarray}
\tilde{\bf E}^a_I &=& ( \tilde{E}^a_i,\tilde{\pi}^a_A)\\
{\bf A}_a^I &=& (A_a^i,\psi_a^A)\\
{\bf F}_{ab}^I &=& (F_{ab}^i,2 D_{[a} \psi_{b]}^A)
\end{eqnarray}
The uppercase indices from the middle of the alphabet range from 1 to
5 and refer to a basis for the fundamental representation of
$GSU(2)$, given by matrices ${\bf G}_I$,
\begin{equation}
{\bf G}_1 = -{i\over \sqrt{2}}
\left(\begin{array}{ccc} 0&1&0\\1&0&0\\0&0&0\end{array}\right),\quad
{\bf G}_2 = -{i\over \sqrt{2}}
\left(\begin{array}{ccc} 0&-i&0\\i&0&0\\0&0&0\end{array}\right),\quad
{\bf G}_3 = -{i\over \sqrt{2}}
\left(\begin{array}{ccc} 1&0&0\\0&-1&0\\0&0&0\end{array}\right),
\end{equation}
\begin{equation}
{\bf G}_4 =
\sqrt{\alpha\over \sqrt{2}}
\left(\begin{array}{ccc} 0&0&1
\\0&0&0\\0&1&0\end{array}\right),\quad
{\bf G}_5 =
\sqrt{\alpha\over \sqrt{2}}
\left(\begin{array}{ccc} 0&0&0\\0&0&1\\-1
&0&0\end{array}\right),\quad
{\bf e} =
\left(\begin{array}{ccc} 1&0&0\\0&1&0\\0&0&0\end{array}\right),
\end{equation}
where we have included the matrix ${\bf e}$ such that
$({\bf G}_i)^2=-{\bf e}/2$ for future purposes. The matrices satisfy
the commutation relations of the algebra (\ref{4}-\ref{6}).
The $GSU(2)$ group has a Killing-Cartan metric associated with
the orthosymplectic form
\begin{equation}
g_{IJ} \lambda^I \lambda^J =
\delta_{ij} \lambda^i \lambda^j -\sqrt{2} \alpha
\eta_{AB} \lambda^A \lambda^B
\end{equation}
where the $\lambda$'s are parameters in
the group with bosonic components $\lambda_i$ and anticommuting
components $\lambda_{A}$ and $\eta_{AB}$ is the antisymmetric symbol in
two dimensions. One has the usual normalization of the generators of
$GSU(2)$ in the case $\alpha =\sqrt{2}$, the other cases correspond to
a constant rescaling of the generators.
A remarkable fact is that
in terms of the $GSU(2)$ connection one
can introduce a covariant derivative ${\bf D}_a$ such that the Gauss
law and the right supersymmetry constraint can be written as a single
$GSU(2)$ Gauss law \cite{Fu},
\begin{equation}
{\bf D}_a \tilde{\bf E}^a_I =0.
\end{equation}
The left supersymmetry constraint can be written in terms of these
variables but it is not invariant under $GSU(2)$ transformations. This
is reasonable since it has a nontrivial Poisson bracket with the right
generator, giving rise to the usual Hamiltonian and diffeomorphism
constraints. In terms of these variables the left supesymmetry
constraint can be written as,
\begin{equation}
\bar{\cal S}_A = {\bf g}_A{}^{IJ}{}_K (\tilde{\bf E}^a_I \tilde{\bf E}^b_J
{\bf F}_{ab}^K +i \Lambda
\ut{\epsilon}_{abc}
\tilde{\bf E}^a_I \tilde{\bf E}^b_J \tilde{\bf E}^{cK})
\end{equation}
where
\begin{equation}
{\bf g}_A{}^{ij}{}_B \equiv -2 i \epsilon^{ij}_k \sigma^k{}_{AB}
\end{equation}
and all other components vanish. The definition of $\bf g$ is
obviously non $GSU(2)$ covariant.
The point of writing the equations in terms of these variables is that
it allows to find in a straightforward manner several solutions to
all the quantum constraint equations. Some of these solutions have a quite
nontrivial form when decomposed in terms of the original $SU(2)$ variables.
In order to see this let us consider a quantum representation in which
wavefunctions are functionals of the $GSU(2)$ connection $\Psi({\bf A})$.
The operator $\hat{\tilde{\bf E}}^a_I$ is represented as a functional
derivative whereas the operator $\hat{\bf A}_a^I$ is multiplicative.
It is immediate to find solutions to the $GSU(2)$ Gauss law. Any
$GSU(2)$ invariant expression will do. In particular one can
consider Wilson loops constructed with the $GSU(2)$
connection,
\begin{equation}
W_\gamma ={\rm STr P}\exp\left(\oint_\gamma dy^a {\bf A}_a\right)
\end{equation}
where $\gamma$ is a loop on the spatial manifold. The trace taken in
the above expression is the supertrace \cite{DeWi}, which for any $GSU(2)$
matrix $A$ is given by ${\rm STr}(A)=A_{11}+A_{22}-A_{33}$.
Since it is $GSU(2)$ invariant, the Wilson loop is annihilated by the Gauss
law. These states are therefore invariant under
right supersymmetry transformations and triad
rotations. If one studies the action of the left supersymmetry
constraint on them it is immediate to notice that if one
considers loops $\gamma$ that are smooth (they do not have kinks or
intersections), the constraint annihilates these states. The
reasons are the same as the ones that made the usual Ashtekar Hamiltonian
constraint annihilate a Wilson loop based on a smooth loop \cite{JaSm}.
The states
have a quite nontrivial fermionic content, as we will discuss in the
next section (they correspond to an infinite superposition of terms
built with holonomies with arbitrarily high number of fermionic
insertions along the loop, if $\alpha\neq0$). Notice that these states
are different from those of Matschull \cite{Ma} which were purely
bosonic. They reduce to them in the case $\alpha=0$. They also share
with those states the pathology that Matschull pointed out:
although they solve all the
constraints they are not diffeomorphism invariant. We will however
find that
related states are useful for the construction of a loop
representation in the next section.
It is possible to find another exact state that solves all the
constraint equations. This state is genuinely diffeomorphism invariant
and is associated with a non-degenerate spatial metric (with non-vanishing
determinant). If one
considers the state built by taking the exponential of the
Chern-Simons form of the $GSU(2)$ connection,
\begin{equation}
\Psi_\Lambda({\bf A}) = \exp({i\over 2 \Lambda}
\int d^3x {\rm STr}({\bf A}_a \partial_b {\bf A}_c +
{\sqrt{2} i \over 3}{\bf A}_a {\bf A}_b {\bf A}_c) \tilde{\epsilon}^{abc}
\end{equation}
it is immediate to see that it satisfies all the supersymmetry
constraints. It is annihilated by the $GSU(2)$ gauss law since it is
$GSU(2)$ invariant. It is annihilated by the left supersymmetry
constraint for the same reason the corresponding state was annhilated
by the Hamiltonian constraint of quantum gravity with a cosmological
constant \cite{BrGaPu}: the state has the property $\hat{\tilde{\bf
E}}^a_I \Psi(A) ={i \over 2\Lambda} \tilde{\epsilon}^{abc}{\bf F}_{bcI}
\Psi(A)$ and therefore the two terms in the left supersymmetry
constraint are identical and of opposite sign, cancelling
each other. Being the integral of a three form, the state is naturally
diffeomorphism invariant. Some concerns may be raised about factor
ordering since in the ordering with the triads to the left (where the
state is a solution) the constraint that usually corresponds to
diffeomorphisms formally does not generate that symmetry. As was
discussed in \cite{BrGaPu}
using a symmetric regularization for the constraint
it actually generates diffeomorphisms and annihilates the state and
similar considerations apply here. This state was first introduced by
Kodama \cite{Ko} for usual gravity and was discussed in a semiclassical
supergravity context by Sano and Shiraishi \cite{SaSh}.
F\"ul\"op \cite{Fu} noticed that it
was an exact solution of the Wheeler-DeWitt equation of full supergravity.
All these discussions were in terms of the $SU(2)$
variables. By writing the state and the constraints in terms of the
$GSU(2)$ variables we notice here that it is very easy to see why it
is annihilated by the constraints. It will also be the key to finding
a counterpart of this state in the loop representation.
The Chern-Simons form has been observed to be a solution of all the
constraints of Maxwell theory, Yang-Mills \cite{Ja83}, Einstein
\cite{BrGaPu}, Einstein-Maxwell \cite{GaPu} and also supergravity. It
is remarkable that so different theories have a similar state in
common.
\section{Loop representation}
The usual starting
step to construct a loop representation \cite{GaTr,RoSm90} for
a theory based on a connection is to expand the states of the theory
in terms of a basis of functions given by the Wilson loops constructed
with the connection. Such basis
is gauge invariant and the resulting representation (the loop
representation) is therefore well suited for the description of gauge
invariant operators. Crucial to this construction is the knowledge
that every gauge invariant quantity can be expanded in terms of Wilson
loops. For the bosonic case this last statement is the content of a
theorem by Giles \cite{Gi}. For the supersymmetric case we do not know if a
similar theorem holds. Even if a theorem like the one above holds, in
general one can only expand wavefunctions in terms of products of
Wilson loops. In many theories one can re-express an arbitrary product
of Wilson loops as a linear combination of products of a fixed minimal
number of Wilson loops. The resulting loop representation is based on
wavefunctions depending on multiloops of at most that minimal number,
through the relation,
\begin{equation}
\Psi(\gamma_1,\gamma_2,\ldots,\gamma_n) =
\int dA W_{\gamma_1}[A] W_{\gamma_2}[A]\cdots W_{\gamma_n}[A] \Psi[A]
\end{equation}
where $n$ is the minimal number of Wilson loops in terms of which one
can express an arbitrary product of Wilson loops.
To address the two issues mentioned above, namely if Wilson loops are
enough to represent any state and which is the minimal number of Wilson
loops needed, let us discuss some properties of the supesymmetric Wilson
loops.
Because they are traces of group elements, the Wilson loops satisfy
certain identities called Mandelstam identities \cite{GaTr86}
which reflect particular properties of the group in question. For the
group we are considering they are rather nontrivial, so we start with a
discussion for the case with $\Lambda=0$. They are given by,
\begin{eqnarray}
W_{\gamma_1\circ \gamma_2} &=& W_{\gamma_2\circ \gamma_1},\\
W_{\gamma_1} W_{\gamma_2} &=&
W_{\gamma_1\circ\gamma_2} +
W_{\gamma_1\circ\gamma_2^{-1}} -
W_{\gamma_1}-W_{\gamma_2}+1.\label{mandelfund}
\end{eqnarray}
These identities allow to express any product of Wilson loops as a
linear combination of Wilson loops. They can be combined in many
nontrivial ways\footnote{For the usual loop representation of quantum
gravity, which is based on an $SU(2)$ group, the identity
(\ref{mandelfund}) reads $
W_{\gamma_1} W_{\gamma_1} =
W_{\gamma_1\circ\gamma_2} +
W_{\gamma_1\circ\gamma_2^{-1}}$.}. For instance, it follows from
(\ref{mandelfund}) taking $\gamma_1$ to a point that,
\begin{equation}
W(\gamma) = W(\gamma^{-1}).
\end{equation}
In order to derive these identities one needs to recall that a generic
element of the group in the case $\Lambda=0$
is written as ${\bf 1} +\phi_0 {\bf e} +
\phi_I {\bf G}^I$ with $(1+\phi_0)^2+\sum_{I=1}^3 \phi_I^2 =1$.
For the case $\Lambda\neq 0$ it is harder to find the identities. We
have only suceeded in finding part of them. In order to do this we make
use of the techniques introduced in \cite{UrWaZe} that related the
Mandelstam identities with the Cayley-Hamilton theorem. This theorem
states that any matrix is a root of its characteristic polynomial. For
supermatrices a generalization of this statement has been discussed in
reference \cite{UrMo}. There it was established that a supermatrix is a
root of a polynomial associated with the characteristic polynomial,
which acts as a generalization of the Cayley-Hamilton polynomial to the
case of supermatrices. For a generic matrix of the orthosymplectic
group (which includes as a subgroup $GSU(2)$) the Cayley-Hamilton
equation is \cite{UrWaZe},
\begin{equation}\label{21}
{\bf H}^3(\gamma) -({\rm STr}({\bf H}(\gamma))+2)
({\bf H}^2(\gamma)-{\bf H}(\gamma))-1 =0,
\end{equation}
If one now
multiplies it by ${\bf H}(\eta) {\bf H}(\gamma)^{-1}$ and takes the
trace one gets,
\begin{eqnarray}
&&{\rm STr}({\bf H}(\eta){\bf H}(\gamma)^2))-
({\rm STr}({\bf H}(\gamma))+2)
{\rm STr}({\bf H}(\eta){\bf H}(\gamma))+\nonumber\\
&&\quad({\rm STr}({\bf H}(\gamma))+2)
({\rm STr}({\bf H}(\eta)))
-{\rm STr}({\bf H}(\eta){\bf H}(\gamma)^{-1})=0.\label{mandid}
\end{eqnarray}
This is a ``Mandelstam identity'' in the sense that it is satisfied by
the traces of the holonomy we are considering. However one is usually
interested in identities that allow to reduce the number of traces.
{}From this point of view, the above identity does not help. One
can derive another identity that allows to reduce the number of traces
in a product by considering the generalization of formula (21) to the
case of generic $(2,1)$ matrices\footnote{An
arbitrary $3\times 3$ matrix in which the $13$, $23$, $31$ and $32$
components are Grassmanian.} \cite{BeUr}.
One then considers a $(2,1)$
matrix given by a sum of $GSU(2)$ matrices and inserts it in the
identity.
The resulting expression relates products of traces of the $GSU(2)$
matrices. and allows to expand an arbitrary product of Wilson loops as a
linear combination of products of three Wilson
loops. When we build the loop representation it will therefore be
based on wavefunctions depending at most on three loops. There is yet
another Mandelstam identity that has to be considered, that which
reflects the fact that the matrices are of unit determinant. We will not
consider it here.
The use of Cayley-Hamilton techniques allows us to state that in
general, for any theory based on a group or supersymmetric group the
Cayley-Hamilton theorem can be used to prove that the loop
representation will involve wavefunctions of a finite number of loops.
Let us address the question of up to what extent one can use
combinations of these Wilson loops to represent any gauge invariant
state. As we mentioned before, in the bosonic case this was ensured by
Giles' theorem. Here we do not have a similar theorem. However it is
easy to see that it is likely to be a problem. Consider the case
$\Lambda=0$. As we saw, the supersymmetric Wilson loops reduce to a
purely bosonic expression coinciding with the Wilson loop constructed
with the bosonic $SU(2)$ connection. Therefore they clearly fail to
capture any fermionic information. For the $\Lambda\neq 0$ case we do
not know what the situation exactly is. It should be noted that the
$\Lambda=0$ case is somewhat pathological from the point of view of the
$GSU(2)$ symmetry we are exploiting since the invariant orthosymplectic
form collapses to the usual Euclidean form. We will proceed to build a
pure loop representation to highlight other aspects of the construction
but it should be forewarned that it is possible that the resulting
quantum representation only captures part of the information present in
the theory.
Apart from the above mentioned peculiarities of the supersymmetric case,
there is yet another difference with the usual loop representation of a
bosonic theory. The $GSU(2)$ Wilson
loops naturally implement the symmetry of the theory under triad
rotations and right supersymmetry transformations. If we construct a
loop representation using them as a basis for states it will be
difficult to write an expression in such a representation for the left
supersymmetry generator $\bar{\cal S}$, which is not invariant under
$GSU(2)$ rotations. It is like trying to represent in the usual loop
representation for a Yang-Mills theory
a non-gauge invariant quantity. A similar
situation arises in gravity when one considers the diffeomorphism
constraint. The space of solutions to this constraint is given by
functions of knots. In this space we cannot represent the Hamiltonian
constraint, since it is not diffeomorphism invariant. This forces us to
define the Hamiltonian constraint on functions of loops rather than on
functions of knots, in spite of the fact that the space of solutions to
the constraint that are physical will be given by functions of knots.
Here we will proceed in a similar way. We will build a representation in
terms of loops and an extra structure, in order to be able to represent
the left supersymmetry constraint. The solutions to the constraint
however, will be given by pure functions of loops.
In order to accomplish this let us
examine in detail the expression for the $GSU(2)$
Wilson loop we introduced above. If one expands out the expression for
the path ordered exponential one has that,
\begin{equation}
W_\gamma({\bf A}) = 1 +\sum_{n=1}^\infty \oint_\gamma dy_1^{a_1}
\cdots \int_o^{y_{n-1}} dy_n^{a_n} {\rm Tr}({\bf A}_{a_1}(y_1)\ldots
{\bf A}_{a_n}(y_n)).
\end{equation}
To further discuss this expression we introduce the following
notation: the non-boldface matrix $A$ will represent the combination
${\bf A}_a^i
G_i$, $i=1\ldots 3$ and the matrix $\psi_a$ will represent the
combination ${\bf A}_a^B G_B$, $i=4,5$. Both are $GSU(2)$
matrices. In terms of these quantities the above expression can be
written as,
\begin{eqnarray}
W_\gamma({\bf A}) &=& 1+\sum_{n=1}^\infty
\oint_\gamma dy_1^{a_1}
\cdots \int_o^{y_{n-1}} dy^{a_n}_n {\rm STr}(A_{a_1}(y_1)\ldots
\ldots A_{a_n}(y_n))\nonumber\\
&&+\sum_{n=2}^\infty \sum_{i=1}^n
\oint_\gamma dy_1^{a_1}
\cdots \int_o^{y_{n-1}} dy^{a_n}_n{\rm STr}(A_{a_1}(y_1)\ldots
A_{a_{i-1}}(y_{i-1}) \psi_{a_i}(y_i) \ldots A_{a_n}(y_n))\nonumber\\
&&+\sum_{n=2}^\infty \sum_{j=2}^n \sum_{i=1}^{j-1}
\oint_\gamma dy_1^{a_1}
\cdots \int_o^{y_{n-1}} dy^{a_n}_n\\&&\quad\times{\rm
STr}(A_{a_1}(y_1)\ldots
A_{a_{i-1}}(y_{i-1}) \psi_{a_i}(y_i) \ldots
A_{a_{j-1}}(y_{j-1}) \psi_{a_j}(y_j) \ldots A_{a_n}(y_n)). \nonumber
\end{eqnarray}
The above expression can be rearranged as,
\begin{eqnarray}
W_\gamma({\bf A}) &=& \sum_{n=0}^\infty
GT^n_{\gamma}({\bf A})\label{wila}\\
GT^n_{\gamma}({\bf A})&\equiv& \oint_\gamma dy_1^{a_1}
\cdots \int_o^{y_{n-1}} dy^{a_n}_n
{\rm Tr}(\psi_{a_1}(y_1) U(\gamma_{y_1}^{y_2})
\psi_{a_2}(y_2) \ldots U(\gamma_{y_{n}}^{y_1})).
\end{eqnarray}
Each of the terms in the summation in (\ref{wila}) is $SU(2)$ invariant.
In order to see this notice that the Gauss law (\ref{gauss}) acts
homogeneously in $\psi$ and therefore does not mix the different terms
in (\ref{wila}). Since the sum is invariant, each term has to be
invariant as well. This can be explicitly checked and one notices that
it occurs in a rather nontrivial fashion, using explicitly the fact that
the $\psi$'s are Grassman-valued. These invariants can be viewed as
holonomies that one breaks at certain points and inserts a $\psi$ and
then integrates the point along the loop. These invariants are quite
different than the ones one usually considers when building loop
representations for theories coupled to fermions, where the fermions can
only appear at the ends of open paths and is directly connected with the
$GSU(2)$ invariance that is characteristic of this theory.
One can build a loop representation based on the invariants
introduced above either through a transform or through the
introduction of further invariants that involve the momenta ${\bf E}$.
We will choose the first approach for brevity. Given a wavefunction in
the connection representation $\Psi({\bf A})$ one constructs a
wavefunction in the loop representation by,
\begin{equation}
\Psi(\epsilon,\gamma) = \sum_{i=1}^\infty \int d {\bf A} \epsilon^n
GT^n_{\gamma}({\bf A}) \Psi({\bf A})
\end{equation}
Wavefunctions in the loop representation are characterized by a loop
and an real parameter $\epsilon$.
The $GT^n_{\gamma}$'s satisfy a series of identities similar to the
Mandelstam identities for the Wilson loops we introduced at the
beginning of this section, we will not discuss them here since we will
not need them for the issues addressed in this paper.
In this representation the left supersymmetry constraint cannot be
directly represented since it is not $SU(2)$ invariant. However one
can build very easily $SU(2)$ invariant expressions that are equivalent
to the left supersymmetry constraint, for instance, by contracting it
with the spin $3/2$ field.
The fact that the representation is not completely cast in terms of
loops appears to detract from the geometric nature of the usual loop
representation. However, as we pointed out, one needs the extra
parameter to represent non $GSU(2)$ invariant quantities and states. The
solutions to the constraint are $GSU(2)$ invariants and therefore are
expressible purely in terms of loops. We will analyze in the following
section one such solution, the one that is obtained by transforming into
the loop representation the Chern-Simons state we discussed in the
previous section.
\section{The Chern-Simons state and the
Dubrovnik version of the Kauffman Polynomial
as a solution to the
super-Wheeler--DeWitt equation}
Let us consider the expression in the loop representation of the
solution to all the constraint equations with a cosmological constant
given by the exponential of the super Chern-Simons form,
\begin{equation}
\Psi(\gamma_1,\ldots,\gamma_n) = \int dA e^{{i\over 2 \Lambda} S_{CS}}
W_{\gamma_1}[A] W_{\gamma_2}[A]\cdots W_{\gamma_n}[A]
\end{equation}
where $S_{CS}$ is the super Chern-Simons form we introduced above. Since
it is a $GSU(2)$ invariant state we only need to introduce Wilson loops
in the transform.
Assuming that one is using a diffeomorphism invariant measure in the
transform, the resulting state in the loop representation has to be a
knot polynomial. We will show that it is the Kauffman Polynomial,
Dubrovnik version.
The derivation is analogous to the
one giving the Kauffman bracket as a state of ordinary bosonic gravity
in the loop representation and was discussed in reference
\cite{BrGaPu}.
We will first study the transform in the case of a one component loop
\begin{equation}
\Psi(\gamma) = \int dA e^{{i\over 2 \Lambda} S_{CS}} W_\gamma[A].\label{30}
\end{equation}
In the case of bosonic gravity, the transform of the
Chern-Simons state is given by the HOMFLY polynomial.
The skein relations satisfied by a regular-isotopic\footnote{As in the
case of the bosonic transform of the Chern-Simons state the resulting
invariant is a function of framed links. In the bosonic case that is
the reason the result is the HOMFLY polynomial rather than the Jones
polynomial.} HOMFLY polynomial $F(z,t,\alpha)_\gamma$ are,
\begin{eqnarray}
F_U&=&1\\
F_{\hat{L}_+} &=& \alpha F_{\hat{L}_0}\\
F_{\hat{L}_-} &=& \alpha^{-1} F_{\hat{L}_0}\\
t F_{L_+} -t^{-1} F_{L_-} &=& z F_{L_0}
\label{skein3}
\end{eqnarray}
where these relations are to be understood as follows. Given a knot,
pick a crossing in its planar diagram and replace it with either
$L_+$, $L_-$ or $L_0$ as depicted in figure \ref{skein} or figure
\ref{hats} for the hatted elements. Evaluate the polynomial on the
resulting links. The resulting polynomials are related by the above
expressions. The first equation is a normalization condition stating
that the polynomial evaluated on the unknot is equal to one.
\begin{figure}
\hspace{3.7cm}\epsfxsize=300pt \epsfbox{fig1.eps}
\caption{The different crossings involved in the skein relations}
\label{skein}
\end{figure}
\begin{figure}
\hspace{2.7cm}\epsfxsize=400pt \epsfbox{fig4.eps}
\caption{Regular isotopy invariants are not invariant under
the addition and removal of a curl. This is determined by the skein relations
involving the elements shown}
\label{hats}
\end{figure}
We will show that the expression of the transform of the super Chern-Simons
state satisfies related but different skein relations. In order to do
this we will perform the following
computation first suggested by Smolin \cite{Sm} and Cotta-Ramusino et
al \cite{CoGuMaMi}. Starting from the expression of the transform
evaluated at an intersection we will append to it an infinitesimal
loop in such a way as to turn the intersection into an over
crossing. We will then repeat the same procedure to turn it into an
under crossing. We see that the difference of the resulting
expressions is related to the value of the transform evaluated at
$L_0$ through a skein relation of the same nature of (\ref{skein3}), but
for a different polynomial.
The variation of the loop transform of the state when a small loop of
element of area $\sigma^{ab}$ is appended to a single-component loop
$\gamma$ at an intersection as shown in figure \ref{inter} is given
by,
\begin{figure}
\hspace{5.7cm}\epsfxsize=100pt \epsfbox{fig2.eps}
\caption{The addition of a small loop at an intersection}
\label{inter}
\end{figure}
\begin{equation}
\sigma^{ab} \Delta_{ab}(y) \Psi[\gamma] = -2 \Lambda i (-1)^P
\int dA \sigma^{ab}
\epsilon_{dab} {\rm STr}[{\bf G}_I {\bf H}_{23}(\gamma_y^y)
{\bf H}_{41}(\gamma_y^y)] {\delta \over \delta
A_d^I(y)} g^{IJ} {\rm exp}({\textstyle {i\over 2\Lambda}} S_{CS})
\end{equation}
where $\Delta_{ab}$ is the loop derivative \cite{GaPubook}, $G_I$ is
one of the generators of $GSU(2)$ and we have used $\Delta_{ab}(x)
{\rm STr}[{\bf H}(\gamma)] = {\bf F}_{ab}^I(x)
{\rm STr}[{\bf G}_I {\bf H}(\gamma_x^x)]$ and $\gamma_x^x$
is the loop with origin at the point $x$. The labels on the holonomy
indicate the connectivity at the intersection, for instance $H_{23}$
is the loop that begins at $2$ and ends at $3$ as indicated in the
figure. We have also used the fundamental property of the Chern-Simons
state that we discussed in section 3,
\begin{equation}
{\delta \over \delta {\bf A}_a^I} \Psi_{\Lambda}[A] =
{i \over 2 \Lambda}
\tilde{\epsilon}^{abc} {{\bf F}_{bc}}_I \Psi_{\Lambda}[A]
\end{equation}
to convert the $F_{ab}$ factor due to the loop derivative into a
functional derivative acting on the exponential of the Chern-Simons
action. It should be recalled that indices like $I$ are raised and lowered
with the orthosymplectic metric $g^{IJ}$. The factor $(-1)^P$ is
introduced to take care of the flips in sign that take place when
$F_{ab}$ is moved from the left of the supertrace to the right of it
before converting it into a functional derivative and
is determined by the odd/even nature of each of the components of
$F_{ab}$ and the supertrace.
Integrating by parts and
choosing the element of area $\sigma^{ab}$ parallel to segment 1-2 so that the
contribution of the functional derivative corresponding to the action on the
segment 1-2 vanishes (since the volume element is zero) we get,
\begin{eqnarray}
\sigma^{ab} \Delta_{ab} \Psi[\gamma] &=&
2 i \Lambda \int dA
\sigma^{ab} \epsilon_{abc} \int dv^c \delta(y-v) \times \nonumber\\
&& {\rm STr}[ {\bf G}_I {\bf H}_{23}(\gamma_y^y)
{\bf G}_J {\bf H}_{41}(\gamma_y^y) ] g^{IJ}
{\rm exp} ({\textstyle {i \over 2 \Lambda}} S_{CS})(-1)^{JH},
\end{eqnarray}
and in the integration by parts the $(-1)^P$ factor is cancelled but a
factor $(-1)^{JH}$ is introduced when the functional derivative is
``introduced'' in the supertrace at the end of the holonomy going from
$2$ to $3$ and is determined by the odd/even nature of the connection in
the functional derivative and the components of the
holonomy going from $2$ to $3$. As is usual in these kinds of
variational derivations \cite{Br}, a regularization of the volume element
determined by the element of area of the loop derivative and the tangent
to the loop is
needed, we take it in such a way that the volume is normalized to be
$\pm1$ depending on the orientation.
We now make use of the Fierz identity for $GSU(2)$,
\begin{eqnarray}
g^{IJ} {({\bf G}_I)^{\kappa}}_\beta {({\bf G}_J)^{\gamma}}_\delta &=&
{{\bf e}^\kappa}_\delta {{\bf e}^\gamma}_\beta-{1\over 2}
{{\bf e}^\kappa}_\beta {{\bf e}^\gamma}_\delta \nonumber\\
&&+{\textstyle {1\over 2}}(\delta^\kappa_1 \delta^3_\beta +\delta^\kappa_3
\delta^2_\beta)(\delta^\gamma_2 \delta^3_\delta-\delta^\gamma_3
\delta^1_\delta) \nonumber\\
&&-{\textstyle {1\over 2}}
(\delta^\kappa_2 \delta^3_\beta -\delta^\kappa_3
\delta^1_\beta)(\delta^\gamma_1 \delta^3_\delta+\delta^\gamma_3
\delta^2_\delta)
\end{eqnarray}
and taking into account the explicit general form of an element of
$GSU(2)$ \cite{To},
\begin{equation}
\left(\begin{array}{ccc}a (1-\alpha p q/2)&b (1-\alpha p
q/2)&\sqrt{\alpha} p\\c (1-\alpha p q/2)&d (1-\alpha p
q/2)&\sqrt{\alpha} q\\\sqrt{\alpha} (-a q +c p)&
\sqrt{\alpha} (d p -b q)& (1+\alpha p q)\end{array}\right)\label{explicit}
\end{equation}
where $ad-bc=1$ and $p$ and $q$ are Grassmanian variables one finally gets,
\begin{eqnarray}
&&\sigma^{ab} \Delta_{ab} \Psi[\gamma] = \label{potskein}\\
&=& i \Lambda \int dA \;
\sigma^{ab} \epsilon_{abc} \int dv^c \delta(y-v)
{\rm STr}[{\bf H}_{23}(\gamma_y^y)]
{\rm STr}[{\bf H}_{41}(\gamma_y^y)]
{\rm exp}({\textstyle {i \over 2\Lambda}}
S_{CS}) \nonumber\\
&+& i \Lambda
\int dA \; \sigma^{ab} \epsilon_{abc} \int dv^c \delta(y-v)
{\rm STr}[{\bf H}_{23}(\gamma_y^y) {\bf H}^{-1}_{41}(\gamma_y^y)]
{\rm exp}( {\textstyle {i \over 2\Lambda}} S_{CS}).\nonumber
\end{eqnarray}
At this point we may proceed as in the bosonic case and reinterpret
relation (
\ref{potskein}) as a skein relation. The form of the expression we got
suggests that the skein relation is,
\begin{equation}
\Psi[L_\pm] -\Psi[L_I] =\pm i \Lambda ( \Psi[L_0] +
\Psi[L_W] ).\label{skeinbuena}
\end{equation}
There is a subtle difference, however, between expression (\ref{potskein})
and the skein relation (\ref{skeinbuena}).
The first involves a rerouting of a portion of
the loop. The implication of that rerouting for the connectivity of the
loop at the intersection (the only ingredient that participates in the skein
relation) is only determined given an initial connectivity of the loop at
the intersection. The above skein relation corresponds to starting with
a loop with a connectivity at the intersection corresponding to a single
loop. One could keep the same intersection and reconnect the strands in
such a way that the initial loop has two components. That requires a
separate derivation for the addition of an infinitesimal element of area.
Let us therefore perform such a calculation. We consider a two
component loop with an intersection as shown in figure \ref{multi},
\begin{equation}
\Psi(\gamma_1,\gamma_2) = \int dA e^{{i\over 2 \Lambda} S_{CS}}
W_{\gamma_1}[A] W_{\gamma_2}[A].\label{42}
\end{equation}
The addition of a small loop at the intersection gives, through a
calculation very similar to the one performed in the case of a single
component loop,
\begin{eqnarray}
\sigma^{ab} \Delta_{ab} \Psi(\gamma_1, \gamma_2) &=&
2 i \Lambda \int dA
\sigma^{ab} \epsilon_{abc} \int dv^c \delta(y-v) \times \nonumber\\
&& {\rm STr}[ {\bf G}_I {\bf H}_{12}({\gamma_1}_y^y)] \times {\rm STr}[
{\bf G}_J {\bf H}_{43}({\gamma_2}_y^y) ] g^{IJ}
{\rm exp} ({\textstyle {i \over 2 \Lambda}} S_{CS})(-1)^{JH_{12}},
\end{eqnarray}
and the labels refer to the figure \ref{multi}.
\begin{figure}
\hspace{3.7cm}\epsfxsize=300pt \epsfbox{fig5.eps}
\caption{The same ``straight-through'' crossing can correspond to a two loop
link or a single loop depending on the connectivity of the diagram.}
\label{multi}
\end{figure}
Using the Fierz identity this expresion can be written as
\begin{eqnarray}
&&\sigma^{ab} \Delta_{ab} \Psi(\gamma_1,\gamma_2) = \label{44}\\
&=& i \Lambda \int dA \;
\sigma^{ab} \epsilon_{abc} \int dv^c \delta(y-v)
{\rm STr}[{\bf H}_{12}({\gamma_1}_y^y)
{\bf H}_{43}({\gamma_2}_y^y)]
{\rm exp}({\textstyle {i \over 2\Lambda}}
S_{CS}) \nonumber\\
&-& i \Lambda
\int dA \; \sigma^{ab} \epsilon_{abc} \int dv^c \delta(y-v)
{\rm STr}[{\bf H}_{12}({\gamma_1}_y^y) {\bf H}^{-1}_{43}({\gamma_2}_y^y)]
{\rm exp}( {\textstyle {i \over 2\Lambda}} S_{CS}).\nonumber
\end{eqnarray}
These results can be interpreted as the following skein relation
for the intersection,
\begin{equation}
\Psi[L_\pm] -\Psi[L_I] =\pm i \Lambda ( \Psi[L_0]- \Psi[L_W])
\end{equation}
for the case of a two component link.
If we now replace $W(\gamma)$ by $\hat{W(\gamma)}= -W(\gamma)$ in the
above expression (\ref{30},\ref{potskein}) and in (\ref{42},\ref{44}), signs
change in such a way that both expressions can be associated with a
single skein relation,
\begin{equation}
\Psi[L_{+}] -\Psi[L_{-}] = -2 i \Lambda
( \Psi[L_0] - \Psi[L_W] ).
\end{equation}
Some comments are in order. The study of the variational relation
obtained by adding an infinitesimal element of area to a crossing with
kinks proceeds along similar lines as the one we exhibited above. The
main difference is that the volume term has two contributions that add
up with opposite signs, corresponding to the addition of a small area
along the two different tangents that enter the kink. With a suitable
regularization, the whole contribution can be taken to be zero. This
implies that an intersection with a kink can be taken as equivalent to
no intersection at all, as shown in the figure \ref{kinks}, and this
is what allow us to replace the kinks by the $L_0$ and $L_W$
in equations (\ref{potskein},\ref{44}) to obtain the
skein relations.
\begin{figure}
\hspace{3.7cm}\epsfxsize=300pt \epsfbox{fig3.eps}
\caption{Intersections without (a) and with (b) kinks.
Under the present regularization intersections of type (b) are
equivalent to no intersection (c).}
\label{kinks}
\end{figure}
To completely characterize the polynomial, we need the value of
the polynomial for the
unknot, which is -1, to be consistent with
the redefinition of the holonomy as minus the supertrace we introduced
above. Also, since the polynomial turns out to be
a regular isotopy invariant, we need to study the effect of
the addition of a ``curl'' at a point with no intersection. The
details are exactly the same as those in reference \cite{BrGaPu} so we
will omit them here. The only new ingredient needed is a contraction
of the Fierz identity,
\begin{equation}
g^{IJ} {({\bf G}_I)^\kappa}_\beta {({\bf G}_J)^\beta}_\gamma=
\delta^\kappa_\gamma,
\end{equation}
and one gets as result,
\begin{equation}
\Psi[\hat{L}_\pm] = (1 \pm 2 \Lambda i) \Psi[\hat{L}_0]
\end{equation}
where the meaning of the hatted elements is described in figure \ref{hats}.
Let us considered now the Dubrovnik version of the Kauffman Polynomial; it is
defined by the following relations \cite{onknots},
\begin{eqnarray}
\Psi[L_{+}] -\Psi[L_{-}] &=& z ( \Psi[L_0]- \Psi[L_W]) \\
\Psi[{\rm unknot}] &=& (a-a^{-1})/z + 1 \\
\Psi[\hat{L}_{+}] &=& a \Psi[L_0] \\
\Psi[\hat{L}_{-}] &=& a^{-1}\Psi[L_0] \newline
\end{eqnarray}
One can check that this correponds exactly to the results we obtained
to first order provided $z= -2i\Lambda$ and $a = 1+2i\Lambda $ and $
\Psi[L_0] = -1 $ which is exactly the case.
We therefore see that the
Dubrovnik version of the Kauffman Polynomial bears the same relationship
with a $GSU(2)$ Chern--Simons theory as the Kauffman bracket has with
the usual $SU(2)$ theory. It is also
important to remember that we have shown that the Dubrovnik version of the
Kauffman polynomial is the transform into the loop representation of an
exact quantum state of supergravity. It therefore should follow (as it
happens in ordinary gravity) that the polynomial should be annihilated by
all the constraints of quantum supergravity.
It is also worthwhile pointing out that one could explore
states without cosmological constant based on the ambient invariant
polynomial associated with the one we found, as happens in the bosonic
case \cite{essay}.
An interesting aspect is that Saleur and Zhang \cite{Sa,Zh} have
considered representations of the braid group associated with graded
groups and found associated knot polynomials. It would be interesting
to check if these have a relation with the Dubrovnik version of the
Kauffman polynomial, as the calculation we have performed suggests. It
would also be important to find ways to compute the skein relations we
found to first order in exact form. This could be accomplished via a
supersymmetric version of the Moore-Seiberg-Witten \cite{MoSe,Wi}
construction based on conformal field theory. Finally, one could
evisage computing explicit expressions for each coefficient of the
Dubrovnik version of the Kauffman polynomial by perturbatively
evaluating the expectation value of a Wilson loop in Chern--Simons theory.
For the bosonic case this was first studied by Guadagnini, Martellini,
and Mintchev \cite{GuMaMi} and although more complex, the supersymmetric
version of this calculation is completely feasible. This will be the first
time that explicit expressions for this polynomial have been found.
\section{Discussion}
This paper explored several issues that arise when trying to construct
a loop representation for supergravity using the fact that the theory
can be cast as a gauge theory of the $GSU(2)$ group. There are many
detailed results that are yet to be derived, as the explicit form of
the left supersymmetry constraint in the representation constructed, a
complete set of Mandelstam identities and a suitable regularization
for the constraint. Yet, we are already able to see the emergence of a
rich mathematical structure of the representation to be constructed,
in particular concerning the space of physical states of the
theory. It is remarkable that a gauge theory with fermions yields a
state space that only includes closed loops, contrary to what happens
in other cases \cite{GaFo,MoRo}. This may be related to the extra
symmetry present in supersymmetric theories equating bosons to
fermions. As in the non supersymmetric case it is expected that one
could find a basis of gauge invariant states that are free of
Mandelstam constraints through the use of spin networks. In this case
they would be spin networks associated with a graded group. The
properties of such objects are yet to be explored. One could also
complete the quantization of the theory in the Euclidean sector using
rigorous measure theory, as has been done in the non supersymmetric
case. Finally, as a by product we have showed that the knot polynomial
associated with Chern-Simons theory based on a graded group is the
Dubrovnik version of the Kauffman Polinomial. This is a remarkable
result; it allows new insights into that polynomial
and opens new perspectives in the search for the conjectured
Link Polinomial \cite{onknots2} which has the HOMFLY and the Kauffman
Polinomials as particular cases.
\acknowledgements
We wish to thank Luis Urrutia and Leonardo Setaro for discussions and
John Baez for pointing out several references. This work was supported
in part by grants NSF-PHY-9423950, NSF-PHY-9396246, NSF-INT-9406269,
research funds of the Pennsylvania State University, the Eberly Family
research fund at PSU and PSU's Office for Minority Faculty
development. JP acknowledges support of the Alfred P. Sloan foundation
through a fellowship. We acknowledge support of Conicyt (Uruguay) and
Conacyt (Mexico), through grant 4862-E9406.
|
\section{Introduction}
The connection-triad variables introduced by Ashtekar
\cite{Ash87} have
simplified the constraint equations of quantum gravity; further,
these variables suggest that in the future we may be able to
reformulate gravity in terms of non-local holonomies rather than
local field operators \cite{RovSmo, GambTri}. However, the new
variables are
unfamiliar, and it is not always clear what they mean physically
and geometrically. For example, it is not clear what operators
or structures correspond to gravity waves. On the other hand,
there has been great progress in constructing diffeomorphism-
invariant volume and area operators; these operators turn out to
be essentially counting operators for numbers of loops and loop
intersections \cite{AV}. The quantum constraint equations are
much
simpler
in the new variables, and solutions to these equations have been
found \cite{RovSmo, knotsol} . However, these solutions
correspond
to
some metric, and it is not always clear what that metric is (the
problem of physical interpretation again). In order to
investigate the metric, one needs a measure, so as to be able to
form dot products and take expectation values. Some progress has
been made in constructing a measure \cite{measure}.
The new approach is not yet truly non-local, and it shares
the same renormalization and regularization difficulties which
plague other local field theories such as QCD. For gravity,
there are some new twists to the old story. The theory is hard
to regulate, because the regulators do not always respect
diffeomorphism invariance \cite{Blen, OptrOrd, Bor}. However,
the
theory is astoundingly
easy to renormalize \cite{RovSmo}. (Compare to QCD, which is
easy
to regulate,
whereas renormalization is difficult.)
Also, it is not always possible to operator-order the
gravitational constraints so that both the constraint algebra
closes (commutator of two constraints = sum of constraints) and
the vector constraints generate spatial diffeomorphisms
\cite{knotsol}. (This
is a difficulty which the Ashtekar approach shares with more
traditional approaches.) The constraint closure is not only
essential to the consistency of the Dirac quantization procedure;
closure is important even classically. When the 3 + 1 splitup is
integrated forward in time, to construct the entire spacetime,
the theory will not be invariant under the full four-dimensional
diffeomorphism group unless the constraint algebra closes
\cite{HKTei}. Thus
an invariance of the classical theory is lost if commutator
closure is neglected.
In this paper we consider the application of the Ashtekar
formalism to the problem of plane gravitational waves. "Plane"
means the metric possesses two commuting spacelike Killing
vectors, and we shall choose coordinates so that these vectors
are unit vectors pointing in the x and y directions.
\begin{equation}
k^{(x)} = \partial _x; k^{(y)} = \partial _y.
\label{1.0}
\end{equation}
We begin the quantization procedure in section 2 by
choosing a factor ordering and verifying closure of the
constraint algebra. We find a rather surprising result: in the
plane wave case, one and the same factor ordering makes the
vector constraint into a diffeomorphism generator and allows the
algebra to close. This result requires the high symmetry (the
two spatial Killing vectors).
In section 3 we tackle the problem of constructing a
measure. Here we attain some partial success. Our measure
respects the reality constraints obeyed by the Ashtekar
connections. However, our measure is not invariant under the
constraints.
The literature on plane waves is vast, but we single out two
papers which are especially close to the present paper. Husain
and Smolin \cite{HSm} were the first to apply the Ashtekar
formalism to the two-Killing vector case.
Neville \cite{Nev}, working with the traditional geometrodynamic
variables,
found the transformations which reduce the Hamiltonian to
parameterized free field form and constructed the classical
constants of the motion. This paper will be referred to as I.
Reference I studied waves which were unidirectional as well
as plane, where "unidirectional" in the present coordinates means
the waves are moving only in the +z direction. Since
unidirectional waves are known to obey a superposition principle,
they do not scatter, except off waves moving in the -z direction.
The present paper does not use the unidirectional assumption.
All results on closure and the measure hold in the presence of
scattering.
Our notation is typical of papers based upon the Hamiltonian
approach with concomitant 3 + 1 splitup. Upper case indices A,
B, $\ldots $,I, J, K, $\ldots$ denote local Lorentz indices
("internal" SU(2) indices) ranging over X, Y, Z only. Lower case
indices a, b, $\ldots $, i, j, $\ldots $ are also three-
dimensional and denote global coordinates on the three-manifold.
Occasionally the formula will contain a field with a superscript
(4), in which case the local Lorentz indices range over X, Y, Z,
T and the global indices are similarly four-dimensional; or a
(2), in which case the local indices range over X, Y (and global
indices over x, y) only. The (2) and (4) are also used in
conjunction with determinants; e.\ g., g is the usual 3x3
spatial determinant, while\ $^{(2)}e$ denotes the determinant of
the 2x2 X, Y subblock of the triad matrix $e^A_a$. We use Levi-
Civita symbols of various dimensions: $\epsilon _{TXYZ} =
\epsilon _{XYZ} = \epsilon _{XY} = +1$. The basic variables of
the Ashtekar approach are an inverse densitized triad \E{a}{A}
and a complex SU(2) connection \A{A}{a}.
\begin {eqnarray}
\E{a}{A}& =& \sqrt g e^a_A; \\
\label{eq:1.1}
[\E{a}{A},\A{B}{b}]&=& -\hbar \delta (x-x') \delta ^B_A
\delta ^a_b.
\label{eq:1.2}
\end{eqnarray}
The local Lorentz indices are vector rather than spinor; strictly
speaking the internal symmetry is O(3) rather than SU(2), gauge-
fixed to O(2) rather than U(1).
\section{Closure of Constraints}
Since the result that the constraints do close is
surprising, and since any proof of closure is bound to be
detailed, it would be helpful if we could state some simple
reason why the constraints close, before becoming enmeshed in the
details. This is not too hard to do. We consider both the
geometrodynamical and Ashtekar approaches, since closure should
be
independent of choice of basis variables.
We begin by establishing some notation. In the coordinate
system (1.1), metric components are independent of x and y, and
it is possible to bring the metric to a block diagonal form
\cite{EhlK}. One
2x2 subblock connects only t and z components of the metric;
another 2x2 subbolck connects only x and y components. The x,y
subblock
may be parameterized using variables suggested by
Szekeres \cite{Sz}:
\[
ds^2 = e^A [dx^2 e^B \cosh{W} + dy^2 e^{-B} \cosh{W} -
2dxdy
\sinh{W} ]
\]
\begin {equation}
+ e^{D-A/2} \{ [ -{(\mbox{${\rm N}'$})}^{2} + {(\rm N ^z)}^{2}]dt^2 +2 \rm N^z dzdt
+
dz^2 \}.
\label{eq:2.1}
\end{equation}
\mbox{${\rm N}'$}\ is not quite the usual lapse N.
\begin{equation}
\mbox{${\rm N}'$} = \rm N /\sqrt{g_{zz}} .
\label{eq:2.2}
\end{equation}
Now recall the usual reason why the constraints do not
close. Let the Hamiltonian density be $\rm N H_0 + \rm N ^i H_i$, with
$H_0$ the scalar constraint and $H_i$ the vector constraints.
Then the commutator of two scalar constraints,
\begin{equation}
[\rm N H_0,\rm M H_0] = \delta(x - x')[\rm M\partial_i\rm N - \rm N
\partial_i\rm M ]H^j g^{ij},
\label{eq:2.3}
\end{equation}
does not close because the final inverse metric factor prevents
$H_j$ from annihilating the wave funtional, implicitly assumed to
stand to
the right on both sides of equation~(\ref{eq:2.3}). In the
present case,
from equation~(\ref{eq:2.1}), the inverse spatial metric is
block diagonal, with a 1x1 subblock containing only
$g^{zz}$. Also, x and y vector constraints have been gauged away
in the process of bringing the metric to block diagonal form, so
that the problem contains only a z vector constraint, and the
$g^{ij}$ in equation ~\ref{eq:2.3} collapses to $g^{zz} =
1/g_{zz}$. Thus we can completely eliminate the $g^{ij}$ problem
by absorbing a factor of $\sqrt{1/g_{zz}} $ into each of the
lapse functions N and M, as at equation~(\ref{eq:2.2}). This is
the simple
reason why the constraints close. The innocuous-looking
renormalization~(\ref{eq:2.2}) is the key step.
What would the renormalization~(\ref{eq:2.2}) look like in
the Ashtekar notation? In that formalism the 3-metric and its
conjugate momenta are replaced by complex connection fields
\A{A}{a} and densitized inverse triads \E{a}{A}, where A = X,Y,Z
is a local Lorentz index. Besides the vector and scalar
constraints, there are three new constraints, which generate
local SU(2) rotations (or in our case, local O(3) rotations,
since A is a vector rather than a spinor index.), The 3x3
\E{a}{A}
matrix may be brought to block diagonal form, with one 1x1
subblock
plus one 2x2
subblock, exactly as for the 3x3 spatial subblock of the metric
$g^{ij}$ The \E{a}{A} matrix is not symmetric, however, so
contains
more independent components than the metric. To bring the
inverse
triad
matrix to block diagonal form , one must gauge away $9 - (1 + 4)
= 4$ inverse triad components, whereas in the metric case one
removes
only $6 - (1 + 3) = 2$ components. To remove the extra two
components one must fix (the x and y vector constraints, as
before, plus) two more constraints, the X and Y SU(2) or O(3)
generators. The 1x1 subblock of the \E{a}{A} matrix is occupied
by
\E{z}{Z} , while the 2x2 subblock contains all \E{a}{A} with a =
x,y and
A = X,Y. The Ashtekar scalar constraint H is density weight 2,
and the new Lagrange multiplier for
the scalar constraint is a weight -1 lapse \ut{N} which is
related
to \mbox{${\rm N}'$} as follows:
\begin{eqnarray}
\ut{N} \E{z}{Z} &=& (\rm N /\sqrt{g})
(\sqrt{g} e^z_Z) \nonumber \\
&=& \rm N \sqrt{g^{zz}} \nonumber \\
&=& \mbox{${\rm N}'$} .
\label{eq:2.4}
\end{eqnarray}
Hence in the Ashtekar approach we will be absorbing factors of
\E{z}{Z} into the densitized lapse.
Since $\ut{N} H = \mbox{${\rm N}'$} (H/\E{z}{Z}) $, the new scalar
constraint
$H/\E{z}{Z} $ becomes a rational function of the basic fields,
rather than a polynomial function. This complication is a price
which must be paid in order to secure closure. We consider it a
small
price. Firstly, the constraints should close as a matter of
principle, in order to have a consistent quantization and full
four-dimensional diffeomorphism invariance.
Secondly, the presence of the $1/\E{z}{Z} $ factor actually
makes the Hamiltonian \underline{less} singular. $H$ contains
products of several operators, all evaluated at the same point z,
which can lead to undefined $\delta (z-z)$ factors when $H$ acts
upon a wave functional. To avoid such factors, Husain and Smolin
\cite{HSm}
must regulate $H$ by point-splitting. However, most of the terms
in $H$ contain a factor of $\E{z}{Z} $, which is removed by the
$1/\E{z}{Z} $, leaving behind a simpler operator which cannot
produce a $\delta (z-z)$ and does not have to be regulated. To
study the term where the $1/\E{z}{Z} $ does not cancel, we recall
that \E{z}{Z} is conjugate to the complex Ashtekar connection
\A{A}{a} ,
\begin{equation}
[\E{a}{A} ,\A{B}{b} ] = -\hbar \delta (z - z')\delta ^a_b
\delta ^B_A,
\label{eq:2.7}
\end{equation}
Therefore the action of \E{z}{Z} on a wave functional $\psi =
\psi [{\rm A}]$ is $\E{z}{Z} \psi = -\hbar \delta \psi / \delta
\A{Z}{z} $. We expect the \A{Z}{z} dependence of $\psi $ to be
holonomic.
\begin{eqnarray}
\psi & =& \cdots X(z_3) \exp [i\int_{z2}^{z3} \A{Z}{z} S_Z
dz]X(z_2)\exp [i\int_{z1}^{z2} \A{Z}{z} S_Z dz] \cdots \nonumber
\\
&:=& \cdots X(z_3)U(z_3, z_2)X(z_2)U(z_2,z_1)\cdots .
\label{eq:2.8}
\end{eqnarray}
The X are assumed to be operators in the Lie algebra of SU(2),
and independent of \A{Z}{z} . Explicit factors of i are present,
because we use the usual, Hermitian SU(2) generators $S_I$. Then
the direct action of \E{z}{Z} on the wave functional is
\begin{equation}
\E{z}{Z} (z)\psi = \cdots + \cdots X(z_3)[\hbar \theta (z_3-z)
\theta (z-z_2) U(z_3, z_2)(-iS_Z)]X(z_2)U\cdots + \cdots ,
\label{eq:2.8a}
\end{equation}
one such term for each U. To find the action of
$[1/\E{z}{Z}]$, multiply both sides of equation~(\ref{eq:2.8a})
by $\theta (z_3-z)\theta (z-z_2)$, in order to project out the
term
exhibited
explicitly on the right; then multiply both sides of
equation~(\ref{eq:2.8a})
by $\hbar /\E{z}{Z}$.
\begin{eqnarray}
[\hbar /\E{z}{Z} (z)] &\cdot &\theta (z_3-z)\theta (z-z_2)[\cdots
X(z_3)U(z_3, z_2)(-iS_Z)X(z_2)
\cdots ] \nonumber \\
& = & \theta (z_3-z)\theta (z-z_2)[\cdots X(z_3)U(z_3,
z_2)X(z_2)\cdots ] + \nonumber \\
& & + const.
\label{eq:2.8c}
\end{eqnarray}
In order to make this look like $\hbar /\E{z}{Z}$ acting upon
$\psi$,
multiply $\psi$ by the following partition of unity:
\begin{equation}
1 = [\theta (z-z_n) + \theta (z_n-z)\theta(z-z_{n-1}) + \cdots +
\theta (z_3-z)\theta (z-z_2) + \cdots ].
\label{eq:2.8d}
\end{equation}
Therefore
\begin{eqnarray}
[\hbar /\E{z}{Z} (z)]\psi & = & \cdots + \cdots X(z_3)[\theta
(z_3-
z) \theta (z-z_2) U(z_3, z_2)(-iS_Z)^{-1}]X(z_2)U\cdots +
\nonumber
\\
& & + \cdots + {\rm const.}.
\label{eq:2.8b}
\end{eqnarray}
Evidently the action of $\hbar /\E{z}{Z}$ on $\psi$ is quite
mild.
There
is not even a $\delta (z-z')$ factor, let alone a $\delta (z -
z)$
{}.
If the X are helicity-changing operators,
the eigenvalue of $S_Z$ in equation~(\ref{eq:2.8b}) will vary
from one U to the next. If we use the 2x2 Pauli representation,
the eigenvalue never vanishes and $(S_Z)^{-1}$ is always
finite. However, in future work we shall use the (2j+1)x(2j+1)
representation, where $S_Z$ can have a zero
eigenvalue if j is integer. When $S_Z$ vanishes, so that
$U(z_3,z_2)$ is unity, one may
replace the square bracket in equation~(\ref{eq:2.8b}) by
\begin{equation}
[-\theta (z_3-z) \theta (z-z_2) U(z_3, z_2)\int_2^3 \A{Z}{z} dz].
\label{eq:2.8e}
\end{equation}
As a check, when equation~(\ref{eq:2.8b}) is inverted by
multiplying both sides
by $\E{z}{Z}/\hbar $, the square bracket~(\ref{eq:2.8e}) gives
the
same answer
for vanishing $S_Z$ as the square bracket~(\ref{eq:2.8b}) gives
for
finite $S_Z$.
We should also check that \E{z}{Z} is not a gauge artifact
(which can be gauged to zero!). Despite its contravariant z
index, \E{z}{Z} is a scalar function under diffeomorphisms,
\begin{equation}
\delta \E{z}{Z} = \rm N ^z \partial _z \E{z}{Z}
\label{eq:2.5}
\end{equation}
The \E{z}{Z} field is both contravariant and weight 1, and in
effect the two transformation properties cancel each other in one
space dimension, leaving an ordinary scalar. Another way to see
the scalar
behavior is to relate \E{z}{Z} to the metric variables A, B, W,
and D introduced at equation~(\ref{eq:2.1}). Using the same
relationships
as at equation~(\ref{eq:2.4}), we find
\begin{equation}
\E{z}{Z} = exp(A).
\label{eq:2.6}
\end{equation}
Since A occurs in the x,y sector of the metric~(\ref{eq:2.1}),
and
this sector transforms as a scalar under diffeomorphisms, the
function A is a scalar. Hence $\E{z}{Z} $ cannot be gauged away.
We now pass to the details. In a coordinate system where
both \E{a}{A} and \A{A}{a} fields are block diagonal, the total
Hamiltonian reduces to
\begin{eqnarray}
H_T &=& \mbox{${\rm N}'$} [\epsilon _{MN}\E{x}{M} \E{y}{N} (\E{z}{Z}
)^{-1}
\epsilon _{AB} \A{A}{x} \A{B}{y} + \epsilon _{MN} \E{b}{M} {\rm
F}^N_{zb}]
\nonumber \\ & & + i\rm N ^z\E{b}{M} {\rm F}^M_{zb} \nonumber \\
& & -i{\rm N_G} [\partial _z\E{z}{Z} - \epsilon
_{IJ}\E{a}{I} \A{J}{a} ] \nonumber \\
&\equiv & \mbox{${\rm N}'$} H_S + \rm N^zH_z + {\rm N_G}H_G,
\label{eq:2.9}
\end{eqnarray}
where\begin{equation}
{\rm F}^N_{zb} = \partial _z\A{N}{b}
-\epsilon _{NQ}\A{Z}{z} \A{Q}{b},
\label{eq:2.10}
\end{equation}
and the ${\rm N_G}$ term (the Gauss constraint) is the generator
of local SU(2) rotations around the Z axis. The lapse has been
renormalized as at equation~(\ref{eq:2.4}), by removing a factor
\E{z}{Z}
from the scalar constraint. From now on the "scalar constraint"
will mean the expression $H_S$ multiplying \mbox{${\rm N}'$}\ in
equation~(\ref{eq:2.9}), namely, the usual Ashtekar scalar
constraint H
divided by \E{z}{Z} . We have operator-ordered
equation~(\ref{eq:2.9}) in
a way which anticipates the following section, where we shall
consider solutions $\psi $ which depend on \A{Z}{z} and the four
\Etld\ in the 2x2 sector. If we call these five commuting
variables
the Q variables ($\psi = \psi (Q)$) and the five conjugate
variables the P variables, then we have ordered P's to the right,
Q's to the left in $H$. We shall carry out the proof of closure
for this specific choice of Q's and this specific ordering, but
the proof would also go through for the other popular choice of
Q's. in which the five A's are chosen as Q's (and the A's are
ordered to the left in $H$).
Now let us ask which commutators, or which parts of which
commutators, are likely to give trouble. First of all, it is
easy to check that the \underline{classical} commutators (or
rather, the classical Poisson brackets) close on pure
constraints, with no undesirable factors of $g^{zz}$ or the
Ashtekar analog of $g^{zz}$. Since the quantum commutators are
designed to reproduce exactly the same fields as the classical
Poisson brackets, there will be no factor of $g^{zz}$ in the
quantum case either. We will get the same fields; and the only
thing which can go wrong is that the commutator yields a P field
to the left of a Q field. Remembering that each constraint is a
sum of terms of the form f(Q)g(P), we want, schematically,
\begin{equation}
[f_1(Q)g_1(P),f_2(Q)g_2(P)] = f_3(Q)g_3(P).
\label{eq:2.11}
\end{equation}
There would be trouble if P's occured to the left of Q's on the
right-hand side, for example $Pf_3(Q)g_3(P)$.
We now show that almost all the terms in a typical
constraint commutator $[H_i,H_j]$ will give no trouble. Each
term in this commutator will look like the left-hand side of
equation~(\ref{eq:2.11}).
Write this term out using the identity
\begin{eqnarray} \label{eq:2.12}
[AB,CD]& =& AC[B,D]+A[B,C]D+ \nonumber \\
& & +C[A,D]B+[A,C]DB. \\
\label{eq:2.13}
[f_1g_1,f_2g_2]&=& 0 + f_1[g_1,f_2]g_2 + f_2[f_1,g_2]g_1 +
0.
\end{eqnarray}
On the right-hand side, the f and g factors outside the
commutators
are in the correct order (Q's to the left). Therefore if
the commutators on the right yield fields which commute among
themselves, the entire expression on the right can be ordered
correctly, and the term gives no trouble. In particular, if
either f or g is a monomial, say $g\sim $P, then the commutator
of any f with g yields only Q fields, and the term can be ordered
correctly. Examination of equation~(\ref{eq:2.9}) shows that all
the terms in $H_T$, except the $(\E{z}{Z} )^{-1}$ term, are
monomials in the P's, of the form $Q^2P$ or $QP$; and one term is
independent of the P's (pure Q) which is even better. Therefore
the only commutators we have to check are the ones involving the
$(\E{z}{Z} )^{-1}$ term,
\begin{eqnarray}
H_E &:= &\mbox{${\rm N}'$} [\epsilon _{MN}\E{x}{M} \E{y}{N} (\E{z}{Z} )^{-
1}\epsilon _{AB} \A{A}{x} \A{B}{y} \nonumber \\
&\sim &Q^2(1/P)P^2.
\label{eq:2.14}
\end{eqnarray}
We now investigate commutators of $H_E$ with Q terms,
QP terms, $Q^2P $ terms, and finally commutators of
$H_E$ with itself. (a). Commutators of $H_E$ with Q and
QP terms. For this case, in commutator~(\ref{eq:2.11}),
\underline{both} $f_2$ and $g_2$ are monomials or constants. It
follows immediately that both commutators on the right-hand side
of equation~\ref{eq:2.13} involve at least one monomial and can
be
correctly ordered. Even if $f_2$ is \A{Z}{z} , there is no
problem, since the commutator with (\E{z}{Z} )$^{-1}$,
\begin{equation}
[(\E{z}{Z} )^{-1},\A{Z}{z} ] = \hbar \delta (z-z') (\E{z}{Z}
)^{-2},
\label{eq:2.14a}
\end{equation}
yields factors which commute among themselves.
Equation~(\ref{eq:2.14a})
may be proven by multiplying it on left and right
by \E{z}{Z} .
(b). Commutators of $H_E$ with $Q^2P$ terms. These
are terms of the form \E{b}{M} \A{Z}{z} \A{Q}{b} coming from the
F's
(field strengths) in the scalar and vector constraints, equations
{}~(\ref{eq:2.9})-(\ref{eq:2.10}). (b1). When commuting the
scalar
constraint with itself, for example, we get commutators of the
form
\begin{eqnarray*}
[Q_E^2(1/P_E)P_E^2(z),Q^2P(z')] &+&
[Q^2P(z),Q_E^2(1/P_E)P_E^2(z')] \\
&=&\cdots + Q_E^2[(1/P_E)P_E^2(z),Q^2(z')]P
+ \\ & & +Q_E^2[Q^2(z),(1/P_E)P_E^2(z')]P,
\end{eqnarray*}
where subscripts E denote fields coming from $H_E$, and $\cdots $
indicates harmless commutators involving the monomial P. The two
commutators on the last line differ by a minus sign and an
interchange of z and $z'$. The interchange affects nothing,
since
the commutator is proportional to $\delta (z-z')$. Therefore the
two commutators cancel each other, and the term is harmless.
(b2). When commuting the scalar constraint with the vector
constraint, we get
\begin{eqnarray*}
[Q^2P(z),Q^2(1/P)P^2(z')]&=& \cdots +
Q_2(z')[Q^2(z),(1/P)P^2(z')]P(z) \\
&=& \cdots + \mbox{$^{(2)}\!\tilde{\rm E} $} [\E{b}{M} \A{Z}{z} ,(\E{z}{Z} )^{-
1}\epsilon _{AB} \A{A}{x} \A{B}{y}]i\A{N}{b} \epsilon _{MN} \\
&=& \cdots + \mbox{$^{(2)}\!\tilde{\rm E} $} \E{b}{M} [\A{Z}{z} ,(\E{z}{Z} )^{-
1}]\epsilon _{AB} \A{A}{x} \A{B}{y} i\A{N}{b} \epsilon _{MN} + \\
& &
+ \mbox{$^{(2)}\!\tilde{\rm E} $} (\E{z}{Z} )^{-1}[\E{b}{M} ,\epsilon _{AB} \A{A}{x}
\A{B}{y}]\A{Z}{z} i\A{N}{b} \epsilon _{MN} \\
&=& \cdots - \mbox{$^{(2)}\!\tilde{\rm E} $} \E{b}{M} [\hbar \delta (z-z')(\E{z}{Z}
)^{-2}]\epsilon _{AB} \A{A}{x} \A{B}{y} i\A{N}{b} \epsilon _{MN}
+0. \end{eqnarray*}
On the third line we have used the ABCD identity,
equation~(\ref{eq:2.12}). On the last line, we can commute the
i\A{N}{b} factor to the left, until it forms the expression
$i\E{b}{M} \A{N}{b} \epsilon _{MN}$. We can subtract from this
expression the Gauss constraint $H_G$, equation~(\ref{eq:2.9}).
This causes no problems, since the Gauss constraint commutes with
everything to the right of $i\E{b}{M} \A{N}{b} \epsilon _{MN}$,
hence can be commuted to the far right where it will eventually
annihilate the wave functional. Since $i\E{b}{M} \A{N}{b}
\epsilon _{MN} - H_G = i\partial _z\E{z}{Z} $, the last line
becomes
\begin{displaymath}
\cdots +i\mbox{$^{(2)}\!\tilde{\rm E} $} [\hbar \delta (z-z')\partial _z(\E{z}{Z} )^{-
1}]\epsilon _{AB} \A{A}{x} \A{B}{y}.
\end{displaymath}
This now has the same form as the corresponding term in the
classical calculation, and moreover the operators are correctly
ordered (Q's to the left).
The calculation just completed, however, suggests a new way
in which the constraints might fail to close. Suppose that at
some point in the calculation it is necessary to insert the Gauss
constraint in the middle of a term (as was done just above); if
the $H_G$ factor cannot be commuted to the far right, then
closure will be spoiled. Fortunately, the classical calculation
once again comes to the rescue. Since the classical calculation
has the same pattern of fields as the quantum calculation, a
Gauss insertion is necessary in the quantum calculation if and
only if a Gauss insertion is necessary at the same point in the
classical calculation. It turns out that the \underline{only}
point where a Gauss insertion occurs, classically, is in the
[vector,scalar] commutator term just considered, and this
insertion is harmless.
(c). Finally, we consider the commutator of $H_E$ with
itself. Using the ABCD identity~(\ref{eq:2.12}), we get
\[
[H_E(z),H_E(z')] = \mbox{$^{(2)}\!\tilde{\rm E} $} (z)(\E{z}{Z} )^{-1}[\epsilon _{AB}
\A{A}{x} \A{B}{y} (z),\mbox{$^{(2)}\!\tilde{\rm E} $} (z')(\E{z}{Z} )^{-1}]\epsilon _{CD}
\A{C}{x} \A{D}{y} (z')+
\]
\[
\mbox{$^{(2)}\!\tilde{\rm E} $} (z')(\E{z}{Z} )^{-1}[\mbox{$^{(2)}\!\tilde{\rm E} $}
(z)(\E{z}{Z} )^{-1},\epsilon _{CD} \A{C}{x} \A{D}{y}
(z')]\epsilon _{AB} \A{A}{x} \A{B}{y} (z).
\]
The two commutators cancel, after a relabeling AB$\leftrightarrow
$CD in the second commutator. This completes the proof that the
constraints close.
It is also easily verified that the constraint $H_z$
generates
diffeomorphisms in the z direction. (In fact $H_z$ fails to
generate diffeomorphisms
only when the P's are ordered to the left \cite{knotsol}.)
$H_z$
is not quite the
diffeomorphism generator, but differs from that generator by a
term
of the
form $N^z\A{Z}{z} H_G$, $H_G$ the Gauss constraint. A term of
this
form can
be added to $H_z$ without affecting closure, since the added term
is linear
in P and therefore harmless.
Since there is no factor of $g^{zz}$ or other fields in the
constraint algebra, it is a true Lie algebra (structure
``constants'' are at most $\delta $ functions, not functions of
fields). The structure of this Lie algebra is very
simple. It breaks up into two commuting subalgebras generated by
$(H_z \pm H_S)/2$,
\begin{equation}
\int dz\int dz'[M(z)(H_z \pm H_S)/2,N(z')(H_z \pm H_S)/2] \\ =
i\hbar \int dz(M\partial _zN -N\partial _zM)(H_z \pm H_S)/2.
\label{eq:2.15}
\end{equation}
Presumably these generators may be interpreted physically as
displacements along the light cone, in directions (z $\pm $ct).
\section{The Reality Constraints}
Since the Ashtekar connections are complex, they obey
reality constraints of the form A + A* = 2 Re A = known function
of the \Etld. For completeness, and to establish certain
detailed formulas which we will need later, we sketch a
derivation of these constraints. The derivation ends at
equation~(\ref{eq:5.12}). At equation~(\ref{eq:5.13}) we propose
a measure
to enforce these constraints.
We start from the four-dimensional connection
\begin{equation}
2{\rm G}\, ^{(4)}\!\A{IJ}{a} = \omega ^{IJ}_a +i\epsilon
^{IJ}_{MN}\omega ^{MN}_a,
\label{eq:5.1}
\end{equation}
where G is the Newtonian constant and $\omega $ is the SL(2,C)
Lorentz connection. After the 3 + 1 splitup \cite{Alect, Rovlect},
one obtains the SU(2)
connection which is canonically conjugate to $\E{a}{S}$
(equation~(\ref{eq:1.2})).
\begin{eqnarray}
2{\rm G}\A{S}{a} &\:= & \epsilon _{MSN}\, ^{(4)}\!\A{MN}{a}
\nonumber \\
&=&\epsilon _{MSN}\omega ^{MN}_a - 2i\omega ^{TS}_a.
\label{eq:5.2}
\end{eqnarray}
{}From this, the real part of A is
\begin{equation}
{\rm G}[\A{S}{a} + \A{S}{a} *] = \epsilon _{MSN}\omega
^{MN}_a ,
\label{eq:5.3}
\end{equation}
or when these equations are written out for the 1x1 and 2x2
subblocks,
\begin{eqnarray}
{\rm G}[\A{Z}{z} + \A{Z}{z} *] &=&-2\omega ^{XY}_z; \\
\label{eq:5.3a}
{\rm G}[\A{I}{i} + \A{I}{i} *]&=&2\epsilon _{IJ} \omega
^{ZJ}_i .
\label{eq:5.3b}
\end{eqnarray}
These ``reality constraints'' relate the A's to the \Etld 's ,
since $\omega = \omega (\tilde {\rm E} )$. The next step is to
exhibit
this $\tilde {\rm E}$ dependence. This is done by first relating
the
$\omega$ 's to the triads $e^I_i $ and inverse triads $e^i_I$, then
relating the $ e^I_i$ and $ e^i_I$ to the $ \tilde {\rm E}$ . The
requirement that
the triads have zero covariant derivative leads to
\begin{equation}
\omega ^{IJ}_a = [\Omega _{i[ja]} + \Omega _{j[ai]} - \Omega
_{a[ij]}]e^{iI} e^{jJ},
\label{eq:5.4}
\end{equation}
where
\begin{equation}
\Omega _{i[ja]} = e_{iM}[\partial _je^M_a - \partial
_ae^M_j]/2.
\label{eq:5.5}
\end{equation}
In the present case the triad matrix is block diagonal, with 2x2
and 1x1 subblocks, and these equations simplify considerably.
Also, from equations ~(\ref{eq:5.3a}) and ~(\ref{eq:5.3b}) we shall
need only
\begin{eqnarray}
\omega^ {XY}_z&=& [e^X_i\partial _ze^{Yi} - e^Y_i\partial
_ze^{Xi}]; \label{eq:5.6} \\
\omega ^{ZJ}_i &=& -\partial _zg_{ij}e^{zZ}e^{jJ}/2,
\label{eq:5.7}
\end{eqnarray}
where the indices i,j range over x,y only.
Now we replace metric, triads, and inverse triads by \Etld\
fields. The inverse triad fields are easiest to replace. From
equation~(\ref{eq:1.1})
\begin{eqnarray}
e^a_A &=& \E{a}{A} /\sqrt g \nonumber \\
&=&\E{a}{A} /\sqrt{\E{z}{Z}\mbox{$^{(2)}\!\tilde{\rm E} $} }.
\label{eq:5.8}
\end{eqnarray}
For the metric and triad fields, the strategy is to express them
in terms of the inverse triads, then replace the latter. For
example in the 2x2 subblock,
\begin{eqnarray}
^{(2)}g_{ab}& =& \epsilon _{am}\epsilon
_{bn}{}^{(2)}\!g^{mn}\, ^{(2)}\!g\nonumber \\
&=&\epsilon _{am}\epsilon _{bn} \E{m}{M} \E{n}{M} ^{(2)}\!g/g
\nonumber \\
&=&\epsilon _{am}\epsilon _{bn} \E{m}{M} \E{n}{M} \E{z}{Z}
/ \mbox{$^{(2)}\!\tilde{\rm E} $};
\label{eq:5.9}
\end{eqnarray}
\begin{eqnarray}
e^M_m &=& \epsilon _{MN}\epsilon _{mn}e^n_N\, ^{(2)}e
\nonumber \\
&=& \epsilon _{MN}\epsilon _{mn} \E{n}{N} \sqrt{\E{z}{Z}
/\mbox{$^{(2)}\!\tilde{\rm E} $} } .
\label{eq:5.10}
\end{eqnarray}
We make these replacements in equations~(\ref{eq:5.6}) and
{}~(\ref{eq:5.7}) and obtain
\begin{eqnarray}
\omega^ {XY}_z&=&-[\epsilon _{mn} \E{m}{M} \partial
_z\E{n}{M}]/2 \mbox{$^{(2)}\!\tilde{\rm E} $} ; \\
\label{eq:5.11}
\omega ^{ZJ}_i &=& -[\E{j}{J} /2\mbox{$^{(2)}\!\tilde{\rm E} $} ]\partial _z[\epsilon
_{im}\epsilon _{jn} \E{m}{M} \E{n}{M} \E{z}{Z} /\mbox{$^{(2)}\!\tilde{\rm E} $} ].
\label{eq:5.12}
\end{eqnarray}
After these algebraic preliminaries, we are ready to
consider the measure. Since our complete set of commuting
observables are the four \Etld\ in the 2x2 X,Y sector, plus the
complex connection \A{Z}{z}, we try a dot product of the form
\begin{equation}
<\phi \mid \psi > = \int \mbox{$\phi\!*$}\psi \mu d^4\tilde {\rm E} d^2A,
\label{eq:5.13}
\end{equation}
where $d^2A \equiv dRe\A{Z}{z} dIm\A{Z}{z} $. The measure $\mu $
must satisfy several requirements. (i) It must guarantee the
quantum form of the reality constraints.
\begin{equation}
<\phi \mid A\psi > + <A\phi\mid \psi > = 2<\phi\mid {\rm
ReA}\psi >.
\label{eq:5.14}
\end{equation}
(ii) It must guarantee the invariance of $\mu d^4\tilde {\rm E}
d^2A$
under transformations generated by the scalar, vector, and Gauss
constraints. (iii) It must contain enough gauge-fixing delta
functions to remove the usual unbounded integrations over
infinite numbers of gauge copies. Note that (ii) requires only
invariance under the constraints, not invariance under four-
dimensional diffeomorphisms. In a 3 + 1 formalism, one does not
have the proper set of fields to implement the latter invariance,
essentially because all fields are evaluated on a constant time
hyperslice, whereas four-dimensional diffeomorphisms move
fields off the hyperslice \cite{IK}.
As yet we do not know how to satisfy requirements (ii)-
(iii). We shall, however, propose a $\mu $ which will satisfy
requirement (i). We set
\begin{equation}
\mu = \delta [{\rm G}(\A{Z}{z} + \A{Z}{z}*) + 2\omega
^{XY}_z].
\label{eq:5.15}
\end{equation}
{}From equation~(\ref{eq:5.3a}), this delta function enforces the
\A{Z}{z}
reality constraint. The surprising fact is that it also enforces
the remaining reality constraints~(\ref{eq:5.3b}) as well, as we
shall prove now.
First we should clarify our notation. Since our integration
variables are Re\A{Z}{z} and Im\A{Z}{z} , not A and A*, the
\A{Z}{z} functional derivative really means
\begin{equation}
\delta /\delta \A{Z}{z} := [\delta /\delta Re\A{Z}{z} +
(1/i)\delta /\delta Im\A{Z}{z} ]/2,
\label{eq:5.16}
\end{equation}
which follows from ReA = (A + A*)/2, etc. We can then check
that, since the $\phi$ in the bra, equation~(\ref{eq:5.14}) , is
complex
conjugated,
\begin{eqnarray}
\delta \mbox{$\phi\!*$} /\delta \A{Z}{z} &=& \delta \mbox{$\phi\!*$}[{\rm A*}]/
\delta {\rm A} \nonumber \\
&=&\int dz'[ \delta \mbox{$\phi\!*$}/\delta \A{Z}{z} (z')\!*][\delta
\A{Z}{z}\!
*/\delta Re\A{Z}{z} + (1/i)\delta \A{Z}{z}\! */\delta Im\A{Z}{z}
]/2 \nonumber \\
&=& 0,
\label{eq:5.17}
\end{eqnarray}
as expected. Now write
\begin{eqnarray}
<\phi \mid \A{I}{i} (z) \psi >& =& -\hbar \int \mbox{$\phi\!*$}\mu
\delta \psi /\delta \E{i}{I} (z) \nonumber \\
&=&+\hbar \int [\delta \mbox{$\phi\!*$}/\delta \E{i}{I} \mu \psi +
\mbox{$\phi\!*$}\delta \mu /\delta \E{i}{I} \psi]\nonumber \\
&=&-<\A{I}{i} \phi \mid \psi > + \hbar \int \mbox{$\phi\!*$}
\delta \delta [{\rm G(A + A*)} + 2\omega ^{XY}_z]/\delta
\A{Z}{z} (z') \times \nonumber \\
& & \times [2\delta \omega ^{XY}_z(z')/\delta \E{i}{I} (z)]
dz'\psi \nonumber \\
&=&-<\A{I}{i} \phi \mid \psi > - (\hbar /{\rm G})\int \mbox{$\phi\!*$}
\mu [2\delta \omega ^{XY}_z(z')/\delta \E{i}{I} (z)] dz' \delta
\psi /\delta \A{Z}{z} (z') \nonumber \\
&=&-<\A{I}{i} \phi \mid \psi > - (1/{\rm G})\int \mbox{$\phi\!*$}
\mu [2\delta \omega ^{XY}_z(z')/\delta \E{i}{I} (z)] dz'\E{z}{Z}
(z') \psi .
\label{eq:5.18}
\end{eqnarray}
Again, the $\delta /\delta \A{Z}{z}$ is really a sum of ReA and ImA
functional derivatives, as at equation~(\ref{eq:5.16}), with $
\delta
\mu /\delta Im\A{Z}{z} $ vanishing. The last square bracket can be
rewritten using equation~(\ref{eq:5.11}) and $^{(2)} \tilde {\rm E}
= \epsilon
_{ij}\epsilon _{IJ}\E{i}{I} \E{j}{J}/2$.
\begin{eqnarray}
-\E{z}{Z} (z')2\delta \omega ^{XY}_z(z')/\delta \E{i}{I} (z) &=&
\E{z}{Z} [(\delta \epsilon _{in}\partial _{z'}\E{n}{I} + \epsilon
_{mi}\E{m}{I} \partial _{z'}\delta )/\mbox{$^{(2)}\!\tilde{\rm E} $} \nonumber \\
& & - \epsilon _{mn} \E{m}{M} \partial _{z'}\E{n}{M} \delta
\epsilon _{ij}\epsilon _{IJ}\E{j}{J} /(\mbox{$^{(2)}\!\tilde{\rm E} $} )^2],
\label{eq:5.19}
\end{eqnarray}
where $\delta = \delta (z - z')$. Now use $\epsilon _{mn}
\epsilon _{ij} = \epsilon _{mi} \epsilon _{nj} +\epsilon _{mj}
\epsilon _{in}$ in the last term of equation~(\ref{eq:5.19}). The
two terms involving $ \epsilon _{in}$ cancel. After integrating by
parts to remove the derivative from the delta function in the
second term, one may replace the
$\E{m}{I}$ by $\E{m}{M}\delta^M_I = \E{m}{M} [\epsilon
_{nj}\epsilon
_{IJ}\E{n}{M} \E{j}{J} /\mbox{$^{(2)}\!\tilde{\rm E} $} ]$. We then have
\begin{eqnarray}
-\E{z}{Z}2 \delta \omega ^{XY}_z(z')/\delta \E{i}{I} &=&-\partial
_{z'}[\E{z}{Z} \epsilon _{mi}\epsilon _{nj}\E{m}{M} \E{n}{M}
/\mbox{$^{(2)}\!\tilde{\rm E} $} ]\delta (z - z')\epsilon _{IJ}\E{j}{J} /\mbox{$^{(2)}\!\tilde{\rm E} $} \nonumber \\
&=& 2\omega ^{ZJ}_i \epsilon _{IJ}\delta (z - z') \nonumber
\\
&=& 2G {\rm Re}\A{I}{i} \delta (z - z'),
\label{eq:5.20}
\end{eqnarray}
where the third line uses equation~(\ref{eq:5.12}) and the last
line uses equation~(\ref{eq:5.3b}). Inserting the
result~(\ref{eq:5.20}) into equation~(\ref{eq:5.18}), we obtain the
reality condition for the \A{I}{i} field, QED.
\section{Directions for Further Research.}
For the plane wave problem, we have constructed a constraint
algebra which closes after a simple renormalization of the lapse
function. We have argued that the cost of this renormalization
(rational, rather than polynomial constraints) is small compared to
the benefits (consistent constraints in the quantum-mechanical
theory; full diffeomorphism invariance in the classical theory).
We have also made modest progress toward constructing a measure.
It is a standard result that the Hamiltonian has surface terms
whenever the spatial manifold is non-compact \cite{deW}. In future
work, we intend to describe these surface terms. They are
surprising: in the plane wave case, it is not automatically true
that the Gauss constraint term in the Hamiltonian falls to zero at
infinity.
Because of these additional terms at infinity, one must
exercise
care when interpreting any wavefunctional involving
holonomies defined over open contours. In the planar case a
holonomy $\exp (i\int \A{Z}{z}S_z)$ integrated over a closed
contour is necessarily zero, since the z contour must retrace
itself. Therefore it is natural to consider open contours and
extend the endpoints to $z = \pm \infty$, to respect spatial
diffeomorphism invariance. It is possible to enrich this elemental
holonomic structure in two ways. First, insert \Etld\ operators at
various points along the holonomy; the wavefunctional then looks
like a Rovelli-Smolin $T^n$ operator \cite{RovSmo} (defined over an
open rather than closed contour). Second, replace the usual 2x2
$S_z$ matrices in the holonomy and in \Etld\ by the (2j + 1) x (2j
+ 1) spin-j representation. The resultant structure is
reminiscent of a symmetric state, or spin network state
\cite{symmsta} , with the holonomies corresponding to flux lines of
spin j and the \Etld\ operators corresponding to vertices. For
appropriate choice of the \Etld, a wavefunctional constructed in
this manner is annihilated by the Hamiltonian at {\it finite}
values of z.
In three spatial dimensions, this would be essentially the end
of the story: the flux exiting at infinity is irrelevant, since
Gauss rotations at infinity are not allowed. The surviving
surface terms in the Hamiltonian simply give the ADM energy. In
one spatial dimension, however, the Gauss term contributes at
infinity. One could add Fermionic matter to the theory
\cite{Romano, Morales} and terminate the flux lines on Fermions at
$\pm \infty$. However, adding Fermions probably complicates the
theory unnecessarily. In one-dimensional QED, for example, one can
learn quite a bit by studying electromagnetic plane waves at finite
z, while ignoring what happens to the wave at infinity. In a
future publication we will adopt this philosophy and study the
finite z properties of the open flux line solutions. Work is also
in progress on solutions involving closed flux lines.
|
\section{Background}
\label{sec:Background}
A current trend in logic is to attempt to incorporate semantic
information into the domain of
deduction, \cite{gabbay94}, \cite{barwiseetal95}.
An area for which this strategy is particularly useful is the problem of
categorial
grammar parsing. The categorial grammar research programme requires
the use of a range of logical calculi for linguistic
description. Some researchers have considered labelled deduction as a
tool for implementing categorial parsers \cite{moortgat92},
\cite{morrill95}, and this paper can be seen as a new contribution to
this field.
In this paper we aim for a modular approach, in which the basic
grammar is kept constant, while different calculi can be implemented
and experimented with by constraining the derivations produced by the
theorem prover. At present, our system covers the classical Lambek
Calculus, \calcL{}, as well as the non-associative Lambek calculus
\calcNL{}, \cite{lambek61}, and variants such as Van Benthem's
\cite{vanbenthem88} \calcLP{}, \calcLPC{}, \calcLPE{} and \calcLPCE{},
and their non-associative counterparts. The system is based on
labelled analytic deduction, particularly on the LKE method, developed
by D'Agostino and Gabbay \cite{dagostino94}. LKE is similar to a
Smullyan-style tableau system, in which the derivations obey the
sub-formula principle, but it improves on efficiency by restricting
the number of branching rules to just one. Different categorial logics
are handled by assigning different properties to the labelling
algebra, while the basic syntactic apparatus remains the same. This
allows the user to experiment with various linguistic properties
without having in principle to modify the grammar itself.
The basic structure of the paper is as follows. In section
\ref{sec:Family}, we introduce the family of categorial calculi, and
discuss some of the linguistic arguments which have been put
forward in the literature with regard to these calculi.
In section \ref{sec:Framework}, we introduce the logical
apparatus on which the system is based, describe the algorithm and
prove some of the properties mentioned in section \ref{sec:Family}
within this framework. We also show how different grammars can be
characterised and present a worked example.
In section \ref{sec:Comparison} the system is compared with other
strategies for dealing with multiple categorial logics, such as
hybrid formalisms and unification-based Gentzen-style deduction. In
this section, we also suggest some ways to improve the efficiency of
the system, and strategies for dealing with the complexity of
labelled unification.
\subsection{A Family of Categorial Calculi and Their Linguistic Applications}
\label{sec:Family}
Categorial Grammars can be formalised in terms of a hierarchy of well
understood and mathematically transparent logics, which yield as
theorems a range of combinatorial operations. However the precise
nature of the
combinatorial power required for an adequate characterisation of
natural language is still very much a matter of debate. For this
reason, it is desirable to have a means of systematically testing the
linguistic consequences of adopting various calculi. In this section
we give an overview of the linguistic applications of some of the
calculi in the hierarchy, with a view towards motivating the
usefulness of a generic categorial theorem prover as a tool for
linguistic study.
The combinatorial
possibilities of expressions in general can be characterised in terms
of {\em reduction laws}. In {\em R1}-{\em R6} below, we give some reduction
laws discussed in \cite{moortgat88}, which have been found to be
linguistically useful.\footnote{We adopt the Lambek notation, in which
X$/$Y is a function which ``takes'' a Y to its right to yield an X,
and Y$\backslash$X is a function which ``takes'' a Y to its left to
yield an X.}.
{\small
\begin{center}
\begin{tabular}[ht]{llllll}
{\em R1:} & {\em Application} & {\em R2:} & {\em Composition} & {\em
R3:} & {\em Associativity} \\
& X$/$Y, Y \EN X & & X$/$Y, Y$/$Z \EN X$/$Z &
& (Z$\backslash$X)$/$Y \EN Z$\backslash$(X$/$Y) \\
& Y, Y$\backslash$X \EN X & & Z$\backslash$Y,
Y$\backslash$X \EN Z$\backslash$X & &
Z$\backslash$(X$/$Y) \EN (Z$\backslash$X)$/$Y \\ & & & & & \\
{\em R4:} & {\em Lifting} & {\em R5:} & {\em Division (main functor)}
& {\em R6:} & {\em Division (sub. functor)} \\
& X \EN Y$/$(X$\backslash$Y) & & X$/$Y \EN
(X$/$Z)$/$(Y$/$Z) & & X$/$Y \EN (Z$/$X)$\backslash$(Z$/$Y)\\
& X \EN (Y$/$X$)\backslash$Y & & Y$\backslash$X
\EN (Z$\backslash$Y)$\backslash$(Z$\backslash$X) & &
Y$\backslash$X \EN (Y$\backslash$Z)$/$(X$\backslash$Z)\\
\end{tabular}
\end{center}}
It is possible to define a hierarchy of logical calculi, each of which
admits one or more of {\em R1}-{\em R6} as theorems; from the purely
applicative calculus {\sf AB}, of Ajdukiewicz and Bar-Hillel,
\cite{ajduk35}, which supports only {\em R1}, to the full Lambek
calculus {\sf L}, which supports all the above laws. Calculi
intermediate in power between {\sf AB} and {\sf L} have been
explored (e.g. Dependency Categorial Grammar \cite{pickering93}), as
well as stronger calculi which extend the power of {\sf L} through
the addition of structural rules.
Much of the interest in using categorial grammars for
linguistic research derives from the possibilities they offer for
characterizing a flexible notion of constituency. This has been found
particularly useful in the development of theories of coordination,
and incremental interpretation.
For example, assuming standard lexical type assignments, the following
right node raised sentence cannot be derived in {\sf AB}, but does
receive a derivation in a system which includes {\em R3}, with
each conjunct assigned the type indicated.
\begin{examples}
\item \label{ex:rnr}
{} [John resents $_{S/NP}$] and [Peter envies $_{S/NP}$] Mary
\end{examples}
A calculus which includes composition, {\em R2}, will allow a
function to apply to an unsaturated argument, and it is this property
which allows Ades and Steedman \cite{steedman82} to treat long
distance dependencies, and motivates much of Steedman's later work on
incremental interpretation.
Dowty \cite{dowty88} uses the combination of composition, {\em R2}
and lifting, {\em R4},
to derive examples of non-constituent
coordination such as {\em John gave mary a book and Susan a record}.
We can increase the power of {\sf L} by adding the
structural transformations {\em Permutation}, {\em Contraction} and
{\em Expansion}, to derive the calculi {\sf LP}, {\sf LPC}, {\sf LPE}
and {\sf LPCE}. The structural transformation {\em Permutation}, which
removes the restrictions on the linear order of types, allows us
to go beyond the purely concatenative derivations of {\sf L}.
This allows us to deal with sentences exhibiting non-standard constituent
order. For example, Moortgat suggests using permutation for dealing
with {\em heavy NP-shift} in examples similar to the following
\cite{moortgat88}:
\begin{examples}
\item \label{ex:comics}
John gave [to his nephew $_{PP}$]
[all the old comic books which he'd collected in his troubled
adolescence $_{NP}$].
\end{examples}
In (\ref{ex:comics}), the bracketed constituents can be ``rearranged'' via
permutation so that a derivation is possible that employs the standard
type ((NP$\backslash$ S)/PP)/NP for the ditransitive verb {\em
gave}.
In {\sf L}, while it is possible to specify a type
missing an argument on
its left or right periphery, it is not possible to specify a type
missing an argument ``somewhere in the middle'', making it
impossible to deal with non-peripheral extraction. However,
as Morrill et al show, permutation provides the additional power
necessary to account for this phenomenon \cite{morrilletal90}.
In addition to permutation, there are also linguistic examples which
motivate contraction (e.g. gapping, \cite{moortgat88}) and expansion
(e.g. right dislocation, \cite{moortgat88}). However it is universally
recognized that a system employing the unrestricted use of structural
transformations would be far too powerful for any useful
linguistic application, since it would allow arbitrary word order
variation, copying and deletion. For this reason, a goal of current
research is to build a system in which the resource freedom of the
more powerful calculi can be exploited when required, while the basic
resource sensitivity of {\sf L} is retained in the general case. One
such approach is to employ structural modalities
\cite{morrilletal90}, which are operators that explicitly mark those
types which are permitted to be manipulated by specific structural
transformations.\footnote{Systems which allow the selective use of structural
transformations may be implemented in the general framework presented
here, although we do not address this issue.}
\section{A framework for Categorial Deduction}
\label{sec:Framework}
In this section we describe the theorem proving framework for
categorial deduction. We start by setting up basic ideas of categorial
logic, giving formal definitions of the core logical language. Then we
move on to the theorem proving strategy, introducing the LKE approach
\cite{dagostino94} and the algebraic apparatus used to characterise
different calculi.
\subsection{The core syntax}
We assume that there is a finite set of atomic grammatical categories
which will be represented by special symbols: NP for
noun phrases, S for sentences, etc. So, the set of
well-formed categories can be defined as below.
\DEF{The set of well-formed categories, \cC is the smallest set which
contains every basic category and which is closed under the following
rule: \begin{itemize}
\item[(i)] If X \EL \cC and Y \EL \cC, then X/Y, X\BS Y and X \BU Y \EL \cC
\end{itemize}
}{def:wff}
Our purpose in this section is to define a prcedure
which will enable us to verify, given an entailment relation \EN,
whether or not such a relation holds for the logic
being considered.
Many proof procedures for classical logic have been proposed: natural
deduction, Gentzen's sequents, analytic (Smullyan style) tableaux,
etc. Among these, methods which conform to the sub-formula principle
are particularly interesting, as far as automation is concerned. See
\cite{fitting90a} for a survey. Most of these methods, along with
proof methods developed for resource logics, such as Girard's proof
nets (a variant of Bibel's connection method), can be used for
categorial logic. Leslie \cite{leslie90} presents and compares some
categorial versions of these procedures for the standard Lambek
calculus L, taking into account complexity and proof presentation
issues. Although tableau systems are not discussed in
\cite{leslie90}, a close relative, the cut-free sequent calculus is
presented as being the one which represents the best compromise
between implementability and display of the proof.
Smullyan style tableau systems, however, have been shown to be inherently
inefficient \cite{dagostino94a}. They cannot even simulate
truth-tables in polynomial time. The main reason for this is the fact
that many of the Smullyan tableau expansion rules cause the proof tree to
branch, thus increasing the complexity of the search. Moreover,
keeping track of the structure of the derivations represents an extra
source of complexity, which in most categorial parsers
\cite{moortgat92,morrill95} is reflected in expensive unification
algorithms employed for dealing with substructural implication. In
order to cope with efficiency and generality, we have chosen the LKE
system \cite{dagostino94} as the proof theoretic basis of our
approach\footnote{For standard propositional logic, it has been shown
that LKE can simulate standard tableau in polynomial time, but the
converse is not true.}. LKE is an analytic (its derivations exhibit
the sub-formula property) method of proof by refutation\footnote{A
formula is proved by building a counter-model for its negation.}
which has only one branching rule. In addition, its formulae are
labelled according to a labelling algebra which will determine the
closure conditions for the proof trees\footnote{See section
\ref{sec:Comparison} for a discussion of how LKE, unlike proof nets
or standard tableaux, enables us to reduce the computational cost of
label unification.}. In what follows, we shall concentrate on
explaining our version of the system, the heuristics that we have found
useful for dealing with particularities of the calculi covered, and
the relevant results for these calculi. The usual completeness and
soundness results (with respect to the algebraic semantic provided)
are already given in \cite{dagostino94}, so we will not discuss them
here.
We have mentioned that the condition for a branch to be considered
closed in a standard
tableau is that both a formula and its negation occur on it. The
calculus defined above presents no negation, though. So, we have to
appeal to some extrinsic mechanism to express
contradiction. In Smullyan's original formulation, the formulae
occurring in a derivation were all preceded by {\em signs}: T or F.
For instance, assume that we want to prove A \IM A in classical logic.
We start by saying that the formula is false, prefixing it by F, and
try to find a refutation for F A \IM A. For this to be the case both
T A (the antecedent) and F A (the consequent) have to be the case,
yielding a contradiction. In classical logic we can interpret T and F
as assertion and denial respectively, and so we can incorporate F into
the language as negation,
obtaining uniform notation by eliminating the need for signed
formulae. In our approach, since negation is not defined in the
language, we shall make use of signed formulae as proof theoretic
devices. T and F will be used to indicate whether or not a certain string
available for combination to produce a new one.
\subsection{The generalised parsing strategy}
If we had restricted the system to dealing with signed formulae, we
would have a proof procedure for an implicational fragment of standard
propositional logic enriched with backwards implication and
conjunction. However, we have seen that the Lambek calculus does not
exhibit any of the structural properties of standard logic, and that
different calculi may be obtained by varying structural
transformations. Therefore, we need a mechanism for keeping track of
the structure of our proofs. This mechanism is provided by labelling
each formula in the derivation with {\em information
tokens}\footnote{See \cite{gabbay94} for a proof theoretic
motivated, LDS approach, and \cite{barwiseetal95} for an approach
based on a finer-grained, semantically motivated information
structure.}.
Labels will act not only as mechanisms for encoding the structure of
the proof, from a proof-theoretic perspective, but will also serve as
means to propagate semantic information through the derivation. A
label can be seen as an information token supporting the information
conveyed by the signalled formula it labels. Tokens may convey
different degrees of informativeness, so we shall assume that they are
ordered by an anti-symmetric, reflexive and transitive relation, \PO,
so that an expression like x \PO y asserts that y is at least as
informative as x (i.e. verifies at least as many sentences as x). We
also assume that this semantic relation, ``verifies'', is closed under
deductibility.
It is natural to suppose that, as well as categories, information
tokens can be composed. We have seen that a type S/NP
can combine with a type NP to produce an S. If we assume that there
are tokens x and y verifying respectively S/NP and NP, how would we
represent the token that verifies S? Firstly, we define a token
composition operation \CI. Then, we assume that, a priori, the order
in which the categories appear in the string matters. So, a minimal
information token verifying S would be x \CI y. As we shall see below,
the constraints we impose on \CI will ultimately determine which
inferences will be valid. For instance, if we assume that the order in
which the types occur is not relevant, then we may allow permutation
on the operands, so that x \CI y \PO y \CI x; if we assume that
contraction is a structural property of the calculus then the string
[S/NP, NP, NP] will also yield an S, since y \CI y \PO y, etc.
Let's formalise these notions by defining an algebraic structure,
called {\em Information frame}.
\DEF{An Information Frame is a structures \cL = \OA\cP,\CI,\ID{},\PO\CA,
where (i) \cP is a non-empty set of information tokens; (ii) \cP is
a complete lattice under \PO; (iii) \CI is an order-preserving,
binary operation on \cP which satisfies continuity, i.e., for {\em every}
directed family \{z\emth{_{i}}\},
\emth{\bigsqcup} \{z\emth{_{i}} \CI x\} = \emth{\bigsqcup} \{z\emth{_{i}}\}
\CI x
{\em and} \emth{\bigsqcup} \{x \CI z\emth{_{i}}\} = x \CI
\emth{\bigsqcup}\{z\emth{_{i}}\}; and (iv) \ID{} is an identity element in \cP.
}{def:InformationFrame}
Combinations of types are accounted for in the labelling algebra by
the composition operator. Now, we need to define an algebraic
counterpart for syntactic composition, \BU, itself. When a formula
like S/NP \BU NP is verified by a token x, this is because its
components were available for combination, and consequently were
verified by some other tokens. Now, suppose S/NP was verified by a
token, say a. What would be the appropriate token for NP, such that
S/NP combined with NP would be verified by x? It certainly would not
be more informative than x. Moreover, if the expression S/NP \BU NP
were to stand for the composition of the (informational) meanings of
its components, then the label for NP would have to verify, when
combined with a, at most as much information as x. In order to express
this, we define the label for NP as being {\em the greatest y s.t.
x is at least as informative as a combined with y}. This token will be
represented by x \opRD a. In general, x \opRD y \ED \emth{\bigsqcup}
\{z\emth{\mid} y \CI z \PO x\}. An analogous operation, \opLD, can be defined
to cope with cases in which it is necessary to find the
appropriate label for the first operand by reversing the
order of the tokens in the definition above. Some properties of
\opRD \cite{dagostino94}:
\begin{minipage}[b]{.48\linewidth}
\FOR{Property1}{y \CI (x \opRD y) & \PO & x}
\end{minipage}\hfill
\begin{minipage}[b]{.48\linewidth}
\FOR{Property2}{\ID{} & \PO & x \opRD x}
\end{minipage}
\begin{minipage}[b]{.48\linewidth}
\FOR{Property3}{(x \opRD y) \CI z & \PO & (x \CI z) \opRD y}
\end{minipage}\hfill
\begin{minipage}[b]{.48\linewidth}
\FOR{Property4}{(x \opRD y) \opRD z & \PO & x \opRD (y \CI z) }
\end{minipage} \\
Having set the basic elements of our proof-theoretic apparatus, we are
now able to define the components of a derivation as follows:
\DEF{Signed labelled formulae (SLF) are expressions of the form S Cat
: L, where S \EL \{T,F\}, Cat \EL \cC and L \EL \cL}
{def:SignedFormula}
A derivation, or proof will be a tree structure built according to
certain syntactic rules. These rules will be called {\em expansion}
rules, since their application will invariably expand the tree
structure. There are three sorts of expansion rules: those which
expand the tree by generating two formulae from a single one occurring
previously in the derivation, those which expand the tree by combining
two formulae into a third one which is then added to the tree, and the
branching rule. The first kind of rule corresponds to what is called
\ALPHA{}-rule in Smullyan tableaux; these rules will be called
\ALPHA{}-rules here as well. The second and third kinds have no
equivalents in standard tableau systems. We shall refer to the
second kind as \SIGMA{}-rules, and to the branching rule as
\BETA{}-rule -- after Smullyan's, even though his branching rules are
different. \REF{fig:ExpansionRules} summarises the expansion rules to
be employed by the system. A deduction bar says that if the formula(e)
appearing above it occurs in the tree, then the formula(e) below it
should be added to the tableau. The rules are easily interpreted
according to the intuitions assigned above to signs, formulae and
information tokens. A rule like \ALPHA{(i)}, for example, says that if
A\BS B is not available for combination and x verifies such
information, then this is because there is an A available at some token a,
but the combination of a and x (notice that the order is relevant) does
not produce B.
\noindent
\begin{figure}[!h]
\centering\begin{tabular}{|l|l|l|l|l|l|l|}
\hline
\small{{\bf \ALPHA{}-rules}} & (i) & (ii) &
(iii) & \multicolumn{2}{c}{\small{{\bf
\BETA{}-rule}}} &\\
\hline
\underline{(\ALPHA{1})} & \underline{F A\BS B : x} &
\underline{F A/B : x} & \underline{\raisebox{-.1ex}[.05ex][.6ex]{ T A
\BU B : x}} &
\multicolumn{3}{c|}{} \\
(\ALPHA{2}) & T A : a$^{*}$
& T B : a$^{*}$ & T A : a$^{*}$ &
\multicolumn{3}{c|}{\emth{\overline{(\BETA{1}) \hspace{0.2cm} T A : x
\emth{\mid}
(\BETA{2})\hspace{0.2cm} F A : x}}} \\
(\ALPHA{3}) & F B : a \CI x & F A : x \CI a & T B : x \opRD a &
\multicolumn{3}{c|}{} \\
\hline
\small{{\bf \SIGMA{}-rules}} & (i) & (ii) & (iii) & (iv) & (v) & (vi) \\
\hline
(\SIGMA{1}) & T A\BS B : x & T A\BS B : x & T A/B : x & T A/B : x &
F A \BU B : x & F A \BU B : x \\
\underline{(\SIGMA{2})} & \underline{T A : y} & \underline{F B : y \CI x} &
\underline{F A: x \CI y} & \underline{T B : y} &
\underline{T A : y} & \underline{T B : y} \\
(\SIGMA{3}) & TB : y \CI x & F A : y & F B : y & T A : x \CI y & F B :
x \opRD y & F A : x \opLD y \\
\hline
\end{tabular}\\
\emth{\ast}:{\em a new label {\em a} (not occurring previously in the
derivation)
must be introduced.}
\caption{Tableau expansion rules}\label{fig:ExpansionRules}
\end{figure}
\noindent Given the expansion rules, the definition of the main data
structure to be manipulated by the theorem proving (parsing) algorithm
is straightforward: a derivation tree, \cT, is simply a binary tree
built from a set of given formulae by applying the rules. The next
step is to define the conditions for a tree to be regarded as
complete. Completion along with inconsistency are the notions upon
which the algorithm's termination depends. It can be readily seen on
\REF{fig:ExpansionRules} that for a finite set of formulae, the number
of times \ALPHA{} and \SIGMA{} rules can be applied increasing the
number of SLFs (nodes) in \cT$\;$ is finite. Unbounded application
of \BETA{}, however, might expand the tree indefinitely. In order to
assure termination, applications of \BETA{} will be restricted to
sub-formulae of formulae in \cT. These notions are formalised in
\RED{def:Completion}.
\DEF{(Tree Completion) Given \cT, a tree for a set of SLFs S, we say
that a binary tree \cT * is a tableau for S if \cT * results from
\cT$\;$ through the application of an expansion rule. A tableau \cT
* is {\em linearly complete} if it satisfies the following
conditions: (i) if \ALPHA{1} \EL \cT, then \ALPHA{2} and \ALPHA{3}
\EL \cT; (ii) if \SIGMA{1} and \SIGMA{2} \EL \cT, then \SIGMA{3} \EL
\cT. A tree \cT * is {\em complete} iff for every A \EL \cT * and
every sub-formula A' of A, both F A' : x and T A' : x have been
added to \cT* by an application of the
\BETA{}-rule.}{def:Completion}
Now, the first step towards building a counter-model for the denial of
the formula to be proved is the search for a tree containing {\em potential}
contradictions. Whether or not a potentially inconsistent tree is a
counter-model for the formula will depend ultimately upon the
constraints on the labelling algebra. This form of inconsistency is
defined below.
\DEF{(Branch and Tree Inconsistency) A branch is {\em
inconsistent} iff for some type X both T X and F X, labelled by
any information token, occur in the branch. A tree is inconsistent
iff its branches are all inconsistent.}{def:TreeInconsistency}
Given the definitions above, we are ready to define an algorithm for
expanding linearly the derivation tree. For efficiency reasons
non-branching rules will be exhaustively applied before we move on to
employing \BETA{}-rules. \RED{def:LinearCompletion} presents the basic
procedure for generating linear expansion for a branch.\footnote{The
reader will notice that if we had allowed \SIGMA{2} formulae to
search for \SIGMA{1} types for combination, in the same way that \SIGMA{1}s
search for \SIGMA{2}s,
then linear expansion would not terminate for some cases.
Consider for example the infinite sequence of \SIGMA{} applications:
T A / B: x, T B / A : y, T A : z, T B : y \CI z, T A : x \CI (y \CI
z), T B : y \CI (x \CI (y \CI z)),... A
strategy to allow unrestricted \SIGMA{}-application without running
into non-terminating procedures, as well as
other practical and computational issues is discussed in
\cite{luz95c}.} The complete LKE algorithm, \RED{def:Expansion},
which uses the procedure below, will be presented after we have
discussed tableau closure from the information frame perspective.
\DEF{(Algorithm: Linear Completion) Given \cT, a LKE-tableau
structure, we define the procedure: \vspace{-1.3ex}\textnormal{
\begin{tabbing}
\mbox{{\sf Linear-Completion(\cT)}} \\
99 \= \kill
1 \> \aDO \= \cT{} \MI \ALPHA{}-completion(\cT) \\
2 \> \> formula \MI head[\cT] \\
3 \> \aWHILE \= ( \NE completed(\cT) \= \aOR \= consistent(\cT) ) \\
4 \> \> \aDO \= \aIF \= \SIGMA{1}-type(formula) \\
5 \> \>\> \aTHEN \aDO \= formula\emth{_{aux}} \MI
search(\cT,\SIGMA{2}) \=
$\;\;$ \= \aCOMMENT{formula\emth{_{aux}} is a set of \SIGMA{2}-type slf's}\\
6 \> \>\>\> \aIF \= formula\emth{_{aux}} \emth{\not = \emptyset}
\> \aCOMMENT{\SIGMA{3}-set results from combining \SIGMA{1} to each
\SIGMA{2} }\\
7\> \>\>\>\> \aTHEN \aDO \= \SIGMA{3}-set \MI
combine-labels(\SIGMA{1},formula\emth{_{aux}}) \\
8\> \>\>\>\>\> \SIGMA{3}-expansion \MI \ALPHA{}-completion(\SIGMA{3}-set) \\
9\> \>\>\>\>\> \cT \MI append(\cT,\SIGMA{3}-expansion) \\
10\> \> \aDO formula \MI next[\cT] \\
11\> \aRETURN \cT
\end{tabbing}
}}{def:LinearCompletion}
We have seen above that the labels are means to propagate information
about the formulae through the derivation tree. From a semantic
viewpoint, the calculi addressed in this paper are obtained by varying
the structure assigned to the set of formulae in the
derivation\footnote{For instance, resource sensitive logics such as
linear logic are frequently characterised in terms of multisets to
keep track of the ``use'' of formulae throughout the derivation.}.
Therefore, in order to verify whether a branch is closed for a
calculus one has to verify whether the information frame satisfies the
constraints which characterise the calculus. For instance, the
standard Lambek calculus \calcL{} does not allow any sort of structural
manipulation of formulae apart from associativity; \calcLP{} allows
formulae to be permuted; \calcLPE{} allows permutations and expansion
(i.e. if B can be proved from the sequent $\Delta$, A, $\Gamma$, then B can
be proved from $\Delta$, A, A, $\Gamma$); \calcLPC{} allows permutation and
contraction; etc. The definition below sets the algebraic counterparts
of these properties.
\DEF{An information frame is: (i) {\em associative} if x \CI (y \CI z)
\PO (x \CI y) \CI z and (x \CI y) \CI z \PO x \CI (y \CI z); (ii)
{\em commutative} if x \CI y \PO y \CI x; (iii) {\em contractive} if
x \CI x \PO x; (iv) {\em expansive} if x \PO x \CI x; (v) {\em
monotonic} if x \PO x \CI y, for all x, y, z \EL
\cP.}{def:Constraints}
Now, we say that a branch is {\em closed} with respect to the
labelling algebra if it contains SLFs of the form T X : x and F X : y,
where x \PO y. Likewise, a tree is closed if it contains only closed
branches. Checking for label closure will depend on the calculus
being used, and consists basically of reducing information token
expressions to a {\em normal form}, via properties
(\ref{Property1})--(\ref{Property4}), and then matching tokens and/or
variables that might have been introduced by applications of the
\BETA{}-rule according to the properties or combination of properties
(\RED{def:Constraints}) that characterise the calculus considered. It
should be noticed that, in addition to the basic algorithm, heuristics
might be employed to account for specific linguistic aspects. Some
examples: (a) it could be assumed that all the bracketing for the
strings is to the right thus favouring an incremental approach; (b)
type reuse could be blocked at the level of the formulae, reducing the
the computational cost of searches for label closure, since most of
the calculi in the family covered by the system are resource
sensitive; (c) priority could be given to juxtaposed strings for
\SIGMA{}-rule application, etc. \RED{def:Expansion} gives the general
procedure for tableau expansion, abstracted from the heuristics
mentioned above.
\DEF{(Algorithm: LKE-completion) The complete tableau expansion for a
LKE-tree \cT is given by the following procedure:
\textnormal{
\begin{pseudocode}{expansion(\cT)}
1 \> \aDO closure-flag \MI no \\
2 \> \aWHILE \= \NE ( completed(\cT) \= \aOR \= closed-\cT = yes) \\
3 \>\> \aDO \= \cT \MI linear-completion(\cT) \\
4 \> \>\> \aIF \= \NE consistent(\cT) \aAND label-closure(\cT) \\
5 \> \>\>\> \aTHEN \aDO \= closure-flag \MI yes \\
6 \> \>\>\> \aELSE \aDO \= subformula \MI select-subformula(\cT) \\
7 \> \>\>\>\> subformula\emth{_{T}} \MI assign-label-T(subformula)\\
8 \> \>\>\>\> subformula\emth{_{F}} \MI assign-label-F(subformula) \\
9 \> \>\>\>\> \cT\emth{_{1}} \MI append(\cT,\{subformula\emth{_{T}}\}) \\
10\> \>\>\>\> \cT\emth{_{2}} \MI append(\cT,\{subformula\emth{_{F}}\}) \\
11\> \>\>\>\> \aIF \= ( expansion(\cT\emth{_{1}}) = yes \aAND
expansion(\cT\emth{_{2}}) = yes ) \\
12\> \>\>\>\>\> \aTHEN \aDO closure-flag \MI yes \\
13\> \aRETURN closure-flag
\end{pseudocode}}}{def:Expansion}
As it is, the algorithm defined above constitutes a semi-decision
procedure. This is due to the fact that even though the search space
for signed formulae is finite, the search space for the labels is
infinite. The labels introduced via \BETA{}-rules are in fact
universally quantified variables which must be instantiated during the
label unification step. This represents no problem if we are dealing
with theorems, i.e. trees which actually close. However, for completed
trees with an open branch, the task might not terminate. In order to
overcome this problem and bind the unification procedure we restrict
label (variable) substitutions to the set of tokens occurring in the
derivation --- similarly to the way parameter instantiation is dealt
with by liberalized quantification rules for first-order logic
tableaux. In practice, the strategy adopted to reduce label
complexity also employs the following refinements: (i) the tableau is
linearly expanded keeping track of the choices made when
\SIGMA{}-rules are applied (the options are kept in a stack); (ii)
once this first step is finished, if the tableau is still open, then
backtrack is performed until either the choices left over are
exhausted or closure is achieved; (iii) only then is the \BETA{}-rule
applied. This explains the role played by the heuristics mentioned
above.\footnote{Furthermore, as we will show in section \ref{sec:Comparison},
if associativity is allowed at the syntactic level then it is possible
to eliminate the branching rule for the class of calculi discussed
here.} We are now able to establish some results regarding the
reduction laws mentioned in section \ref{sec:Family}.
\begin{proposition1}[Reduction Laws]
Let X, Y and Z be types, and \cL an information frame. The
properties R1--R5 hold:
\end{proposition1}
\begin{proof}
\label{prf:ReductionLaws}
The proofs are obtained by straightforward application of
\RED{def:LinearCompletion} and \RED{def:Expansion}. Below we
illustrate the method by proving (R1) and
(R2):
\begin{itemize}
\item[(R1)] To prove right application we start by
assuming that it is verified by the identity token \ID{}. From this we
have: 1- T X/Y \BU Y : m, 2- F X : 1 \CI m = m. Then, we apply
\ALPHA{(iii)} to 1 obtaining 3- T X/Y : n and 4- T Y : m \opRD n.
The next step is to combine 3 and 4 via \SIGMA{(iv)} getting 5- T X
: n \CI (m \opRD n). Now we have a potential closure caused by 5 and
2. If we apply property (\ref{Property1}) to the label for 5 we find
that n \CI (m \opRD n) \PO m, which satisfies the closure condition
thus closing the tableau.
\item[(R2)] Let's prove left composition. As we did
above, we start with: 1- T Z\BS Y \BU Y\BS X : m and 2- F Z\BS X : 1
\CI m. Applying \ALPHA{(iii)} to 1 we get: 3- T Z\BS Y : a and 4- T
Y\BS X : m \opRD a. Now, we may apply \ALPHA{(i)} to 2 and get: 5- T Z
: b and 6- F X : b \CI m. Then, combining 3 and 5 via \SIGMA{(i)}:
7- T Y : b \CI a. And finally 4 and 7 through the same rule: 8- T X
: (b \CI a) \CI (m \opRD a). The closure condition for 8 and 6 is
achieved as follows: \vspace{-1ex}
\[
\begin{array}{llll}
(b \CI a) \CI (m \opRD a) & \PO b \CI (a \CI (m \opRD a)) &
\textit{by associativity} \nonumber \\ & \PO b \CI m & \textit{by
(\ref{Property1}) and \CI being order-preserving} &\nonumber
\end{array}
\]
\end{itemize} \vspace{-4ex}\hfill\QED
\end{proof}
Even though \calcL{} does not enjoy finite axiomatizability, the results
above suggest that the calculus finds a natural characterization in LKE
for associative information frames. In particular,
the Division Rule ({\em R6}) can be regarded as \calcL{}'s
characteristic theorem, since it is not derivable in weaker calculi
such as \calcAB{}, \calcNL{}, and \calcF{}. If we do not allow associative
frames, we get \calcNL{}. Stronger calculi such as \calcLP{},
\calcLPE{}, \calcLPC{} and \calcLPCE{} \cite{moortgat93}
can be obtained for the same general
framework by assigning further properties to \CI in the labelling
algebra. Frames exhibiting combinations of monotonicity, expansivity,
commutativity and contraction allow us to characterise these
substructural calculi. Algebras that are both associative and
commutative describe \calcLP{}. Adding expansivity (weakening)
to \calcLP{} results in \calcLPE{}. Associativity,
commutativity and contraction describe \calcLPC{} frames. \calcLPCE{} is
obtained by combining the properties of \calcLPC{} and \calcLPE{}
algebras.
We end this section with a simple example requiring associativity:
show, in \calcL{}, that an NP ({\sf John}), combined with a type
(NP\BS S)/NP ({\sf
likes}) yields S/NP, i.e a type which combined with a NP will result
in a sentence (Proof \ref{prf:JLBBHM}).
\noindent\begin{proof}Let's assume the following type--string
correspondence: NP for {\sf John}, (NP\BS S)/NP for {\sf likes}.
The expression we want to
find a counter-model for is: 1- F NP \BU (NP\BS S)/NP \ENI{\! L}
S/NP. Therefore, the following has to be
proved: 2- T NP \BU (NP\BS S)/NP : m and 3- F S/NP : m.
We proceed by breaking 2 and 3 down
via \ALPHA{(iii)}, obtaining:
4- T NP : a,
5- T (NP\BS S)/NP : m\opRD a,
6- T NP : b, and
7- F S : (m \CI b). \\
\noindent Now we start applying \SIGMA{}-rules (annotated on the
right-hand side of each line):\\
\noindent\begin{tabular}{llll}
8- T & NP\BS S & : (m \opRD a) \CI b & 5,6 \SIGMA{(i)} \\
9- T & S & : a \CI ((m \opRD a) \CI b) & 4,9 \SIGMA{(i)} \\
\end{tabular} \\
We have derived a potential inconsistency between 7 and 9. Turning
our attention to the information tokens, we verify closure for \calcL{}
as follows:
\begin{center} \begin{tabular}{lllll}
a \CI ((m \opRD a) \CI b) & \PO & (a \CI (m \opRD a)) \CI b & by
associativity \\
& \PO & m \CI a & by property (\ref{Property1})
\end{tabular}\end{center} \vspace{-3ex} \hfill \QED
\label{prf:JLBBHM}
\end{proof}
\subsection{Comparison with Existing Approaches}
\label{sec:Comparison}
Early implementations of CG parsing relied on cut-free Gentzen
sequents implemented via backward chaining mechanisms
\cite{moortgat88}. Apart from the fact that it lacks generality, since
implementing more powerful calculi would involve modifying the code in
order to accommodate new structural rules, this approach presents
several sources of inefficiency. The main ones are: the
generate-and-test strategy employed to cope with associativity, the
non-determinism in the branching rules and in rule application itself.
The impact of the latter form of non-determinism over efficiency can
be reduced by testing branches for {\em count invariance} prior to
their expansion and by performing sequent proof normalisation.
However, non-determinism due to splitting in the proof structure still
remains. As we move on to stronger logics and incorporate structural
modalities such problems tend to get even harder.
An improved attempt to deal uniformly with multiple calculi is
presented in \cite{moortgat92}. In that paper, the theorem prover
employed is based on proof nets, and the characterisation of different
calculi is taken care of by labelling the formulae. For substructural
calculi stronger than \calcL{}, much of the complexity (perhaps too
much) is shifted to the label unification procedures. A strategy for
improving such procedures by compiling labels into higher-order logic
programming clauses is presented in \cite{morrill95} for \calcNL{} and
\calcL{}. However, a comprehensive solution to the problem of binding
label unification, a problem which arises as we move from sequents
to labelled proof nets, has not been presented yet. Moreover, as
discussed in \cite{leslie90}, if we consider that the system is to be
used as a parser, as a tool for linguistic study, the proof net style
of derivation does not provide the clearest or most intuitive display of
the proofs.\footnote{Proof nets and sequent
normalisation have also been employed to get around
spurious ambiguity (i.e. multiple proof for the same
sentence, with the same semantics). Our approach does not exhibit
this problem.}
In our approach, the burden of parsing is not so concentrated in label
unification but is more evenly divided between the theorem prover and
the algebraic checker. This is mainly due to the fact that the system
allows for a controlled degree of non-determinism, present in the
\SIGMA{}-rules, which enables us to reduce the introduction of
variables in the labelling expressions to a minimum. We believe this
represents an improvement on previous attempts. Besides this,
controlling composition via bounded backtrack opens the possibility of
implementing heuristics reflecting linguistic and contextual
knowledge. In fact, we verify that, under the
appropriate application of rules, we are able to eliminate the
\BETA{}-rule for a class of theorems.
\begin{proposition1}[Elimination Theorem]
All closed LKE-trees derivable by the application of the set
of rules \cR =
\SE{\ALPHA{(i)},...,\ALPHA{(iii)},\SIGMA{(i)},...,\SIGMA{(vi)},\BETA{}} can
be also derived from \cR $-$ \{\BETA{}\} + \{ assoc \}.
\end{proposition1}
The proof of this proposition can be done by defining an {\em abstract
Gentzen relation}, proving a substitution lemma with respect to
the labelling algebra (as in \cite{dagostino94}), and showing that our
consequence relation is closed under the relevant Gentzen conditions
even if no \BETA{} rule is employed. The proof appeals to the fact that
no formula signed by F can occur in the sequents on the left-hand side
of the entailment relation, since the calculi presented here do not
have negation.\footnote{Of course, without the \BETA{} rule not all open
trees generated will constitute downward saturated sets, since they
might contain formulae which are not completely analysed.} We
believe that this result shows that, even though LKE label unification might
be computationally expensive for substructural logics in general,
the system seems to be well suited for categorial logics. We
refer the reader to \cite{luz95c} for a more comprehensive discussion
of these issues.
\section{Conclusions and Further Work}
\label{sec:Conclusion}
We have described a framework for the study of categorial logics with
different degrees of expressivity on a uniform basis, providing a tool
for testing the adequacy of different CGs to a variety of linguistic
phenomena. From a practical point of view, we have investigated
the effectiveness and generality issues of a parsing strategy for CG
opening an avenue for future developments. Moreover, we have pointed out
some strategies for improving on efficiency and for dealing with more
expressive languages, including structural modalities.
The architecture proposed seems promising. Its flexibility with
respect to the variety of logics it deals with, and its modularity
suggest some natural extensions to the present work. Among them:
implementing a semantic module based on Curry-Howard correspondence
between type deduction and $\lambda$-terms, adding local control of
structural transformations (structural modalities) to the language,
increasing expressivity in the information frames for covering calculi
weaker than \calcL{} (e.g. Dependency Categorial Grammar
\cite{pickering93}), exploiting the derivational structure encoded in
the labels to define heuristics for models of human
attachment preferences etc. Problems for further investigation might
include: the treatment of polymorphic types (by incorporating rules
for dealing with quantification analogous to Smullyan's \gD and \gG
rules \cite{fitting90a} \cite{smullyan68}), and complexity issues
regarding how the general architecture proposed here would behave
under more standard theorem proving methods.
|
\chapter{Introduction}
Despite their central role in string compactifications, chiral $(0,q)$
supersymmetric sigma models have received less attention in comparison to
the non chiral models due to difficulties involved with their
construction and quantization. Recently there has been considerable
interest in the use of massive linear sigma models to construct string vacua
as the infrared conformal fixed point of the renormalization group flow. This
method allows one to construct a large class of string vacua including those
with chiral supersymmetry. In particular an interesting paper by Witten
[\Wittena] discusses a class of massive linear sigma models possessing
on-shell (0,4) supersymmetry which flow in the infrared to conformally
invariant sigma models describing ADHM instantons [\ADHM,\WW].
Previous work on the ADHM sigma model
has focused primarily on classical aspects of the $(0,4)$ supersymmetry
multiplet used and in particular the construction of off-shell superfield
formalisms [\GR,\GS,\GSt]. Here we will study the models quantum properties
and it's rich interplay between geometry and field theory in detail. The
general $(p,q)$ supersymmetric massive sigma model has been constructed
before [\HPT,\PT] and it's quantization is discussed to two loop order in
[\NDL]. We will show here that the ADHM sigma model is ultraviolet finite to
all orders of perturbation theory and integrate out the massive fields to
obtain the low energy effective theory. Due to anomalies this theory has
interesting non trivial properties and we obtain the quantum corrections to
order $\alpha'^2$ by requiring that the anomalies are appropriately canceled. We
conclude by making some comments about the case where the instanton size
vanishes.
\chapter{The ADHM Sigma Model}
In [\Wittena] Witten constructs an on-shell (0,4) supersymmetric linear sigma
model which parallels the ADHM construction of instantons [\ADHM]. The model
consists of $4k$ bosons $X^{AY}$, $A=1,2,\ Y=1,2...,2k$ with
right handed superpartners $\psi^{A'Y}_-$, $A'=1,2$. There is also a
similar multiplet of fields $\phi^{A'Y'},\ \chi_-^{AY'}$ $Y'=1,2...,2k'$. In
addition there are $n$ left handed fermions $\lambda^a_+$, $a=1,2...,n$. The
$A,B...$ and $A',B'...$ indices are raised (lowered) by the two
by two antisymmetric tensor $\epsilon^{AB}$ ($\epsilon_{AB}$),
$\epsilon^{A'B'}$ ($\epsilon_{A'B'}$). The $Y,Z...$ and $Y',Z'...$ indices are
raised (lowered) by the invariant tensor of $Sp(k)$, $Sp(k')$ respectively
which are also denoted by $\epsilon^{YZ}$ ($\epsilon_{YZ}$),
$\epsilon^{Y'Z'}$ ($\epsilon_{Y'Z'}$).
The interactions are provided for by a tensor $C^a_{AA'}(X,\phi)$ in a similar
manner
to the construction of the general models [\HPT,\PT]. The action for the
theory is given by
$$\eqalign{
S = \int\! d^2x & \left\{ \epsilon_{AB}\epsilon_{YZ} {\partial}_= X^{AY} {\partial}_{\ne}
X^{BZ} +i\epsilon_{A'B'}\epsilon_{YZ} \psi_-^{A'Y} {\partial}_{\ne} \psi_-^{B'Z}
\right. \cr & \left.
+\epsilon_{A'B'}\epsilon_{Y'Z'}{\partial}_= \phi^{A'Y'} {\partial}_{\ne} \phi^{B'Z'}
+i\epsilon_{AB}\epsilon_{Y'Z'}\chi_-^{AY'} {\partial}_{\ne} \chi_-^{BZ'} \right. \cr
& \left.
+i\lambda_+^a {\partial}_= \lambda^a_+
-{im\over2}\lambda_+^a
\left(\epsilon^{BD}{\partial C^a_{BB'}\over \partial X^{DY}}\psi_-^{B'Y}
+\epsilon^{B'D'}{\partial C^a_{BB'}\over \partial
\phi^{D'Y'}}\chi_-^{BY'}\right)
\right. \cr
& \left.
-{m^2\over8}\epsilon^{AB}\epsilon^{A'B'}C^a_{AA'}C^a_{BB'} \right\}\ , \cr}
\eqn\action
$$
where
$$
{\partial}_{\ne} = {1\over{\sqrt{2}}}(\partial_0 + \partial_1 )\ \ \ \ \ \ \ \ \ \
{\partial}_= = {1\over{\sqrt{2}}}(\partial_0 - \partial_1 )\ ,
$$
and $m$ is an arbitrary mass parameter. Note the
twisted form of the Yukawa interactions in \action\ in comparison to the
models
of [\HPT,\PT]. The free field theory ($m$ =0) possesses an
$SU(2)\times Sp(k)\times SU(2) \times Sp(k')$ rigid symmetry acting on the $AB,
YZ,A'B', Y'Z'$ indices respectively which is generally broken by the potential
terms.
Provided that $C^a_{AA'}$ takes the simple form
($M^a_{AA'}$,$N^a_{A'Y}$,$D^a_{AY'}$ and $E^a_{YY'}$ are constant tensors)
$$
C^a_{AA'} = M^a_{AA'} + \epsilon_{AB}N^a_{A'Y} X^{BY} +
\epsilon_{A'B'}D^a_{AY'}\phi^{B'Y'} + \epsilon_{AB}\epsilon_{A'B'}E^a_{YY'}
X^{BY} \phi^{B'Y'} \ ,
\eqn\onesusy
$$
subject to the constraint
$$
C^a_{AA'}C^a_{BB'}+C^a_{BA'}C^a_{AB'} = 0 \ ,
\eqn\foursusy
$$
then the action \action\ has the on-shell (0,4) supersymmetry
$$\eqalign{
\delta X^{AY} &= i\epsilon_{A'B'}\eta^{AA'}_+\psi_-^{B'Y}\cr
\delta \psi^{A'Y}_- &= \epsilon_{AB}\eta^{AA'}_+{\partial}_= X^{BY}\cr
\delta \phi^{A'Y'} &= i\epsilon_{AB}\eta^{AA'}_+\chi_-^{BY'}\cr
\delta \chi^{AY'}_- &= \epsilon_{A'B'}\eta^{AA'}_+{\partial}_= \phi^{B'Y'} \cr
\delta \lambda^{a}_+ &= \eta^{AA'}_+C^a_{AA'} \ ,\cr}
\eqn\susy
$$
where $\eta^{AA'}_+$ is an infinitesimal anticommuting spinor
parameter. As is discussed by Witten [\Wittena], the above construction of
models
with (0,4) supersymmetry can be interpreted as a string theory analogue of the
ADHM construction of instantons with instanton number $k'$ in a target space
dimension of $4k$.
The general form of massive $(p,q)$ supersymmetric sigma models has been
discussed in terms of $(0,1)$ superfields in [\HPT,\PT] and we now provide such
a formulation of the ADHM model. To this end we introduce a tensor $I^A_{\
A'}$ satisfying
$$
\epsilon_{AB}I^A_{\ A'}I^B_{\ B'} = \epsilon_{A'B'}
\eqn\Idef
$$
which can be interpreted as a complex structure in the sense that
$I^{AB'}I_{AC'}=-\delta^{B'}_{C'}$, $I^{BA'}I_{CA'}=-\delta^{B}_{C}$.
The 'twisted' superfields are
$$\eqalign{
{\cal X} ^{AY} &= X^{AY} + \theta^-I^A_{\ A'}\psi_-^{A'Y} \cr
\Phi^{A'Y'} &= \phi^{A'Y'} + \theta^-I^{\ A'}_{A}\chi_-^{AY'}\ \cr
\Lambda^a_+ &= \lambda^a_+ + \theta^-F^a \ ,\cr}
\eqn\superfields
$$
where $\theta^-$ is the anticommuting spinorial (0,1) superspace coordinate
with the associated superspace covariant derivative
$$
D_- = {\partial\over\partial \theta^-} + i\theta^-{\partial}_=
$$
and $F^a$ is an auxiliary field. After removing $F^a$ by it's equation of
motion and using the constraint \foursusy, the action \action\ can be seen to
have the superspace form
$$\eqalign{
S_{effective} = -i\int\! d^2xd \theta^- & \left\{
\epsilon_{AB}\epsilon_{YZ} D_-{\cal X}^{AY}{\partial}_{\ne}
{\cal X}^{BZ} +\epsilon_{A'B'}\epsilon_{Y'Z'} D_-\Phi^{A'Y'}{\partial}_{\ne} \Phi^{B'Z'}
\right. \cr
&\left.
- i \delta_{ab}\Lambda^a_+ D_- \Lambda^b_+ - mC_a\Lambda^a_+
\right\} \ , \cr}
\eqn\suspaction
$$
where $C^a = I^{AA'}C^a_{AA'}$. Since the vector field $C^a$ is harmonic
and the target space is ${\bf R}^{4(k+k')}$, the model satisfies the
requirements of $(0,4)$ supersymmetry found in [\PT]. The inclusion of the
auxiliary field allows one to close a $(0,1)$ part of the supersymmetry
algebra off-shell. As with the component field formulation \action\ the
full $(0,4)$ supersymmetry is only on-shell. A manifestly off-shell form
requires harmonic superfields with an infinite number of auxiliary fields
[\GS].
Lastly we outline the $k=k'=1,\ n=8$ case (ie. a single instanton in ${\bf
R}^4$) analyzed by Witten which will be of primary interest here. The right
handed fermions are taken to be $\lambda^a_+ =
(\lambda^{AY'}_+,\lambda^{YY'}_+)$ and the tensor $C^a_{AA'}$ takes the form
$$\eqalign{
C^{YY'}_{BB'} &= \epsilon_{BC}\epsilon_{B'C'}X^{CY}\phi^{C'Y'} \cr
C^{AY'}_{BB'} &= {\rho\over\sqrt2}\epsilon_{B'C'}\delta^A_B\phi^{C'Y'}\ ,
\cr}
$$
where $\rho$ is an arbitrary constant to be interpreted as the instanton size.
The bosonic potential for this theory is easily worked out as
$$
V = {m^2\over8}(\rho^2 + X^2)\phi^2 \ ,
\eqn\V
$$
where $X^2 = \epsilon_{AB}\epsilon_{YZ}X^{AY}X^{BZ}$ and similarly for
$\phi^2$. Thus, for $\rho \ne 0$, the vacuum states of the theory are
defined by $\phi^{A'Y'}=0$, and parameterize ${\bf R}^4$. The
$X^{AY}$ and $\psi_-^{A'Y}$ are massless fields while $\phi^{A'Y'}$ and
$\chi_-^{AY'}$ are massive. This yields exactly 4 of the
$\lambda^a_+$ massive and 4 massless.
\chapter{Quantization}
\subsection{Renormalization}
It is not hard to see that the model described above is superrenormalizable
in two dimensions as the interaction vertices do not carry any momentum
factors and have at most three legs. In fact a little inspection reveals that
the only possible divergences of the theory are the one loop graphs
contributing to the potential. This can also been seen from the following
simple superspace argument, valid for any $(0,1)$ supersymmetric
linear massive sigma model. The superspace measure $d^2xd\theta^-$ has mass
dimension $-3/2$ while all vertices contribute a factor of $m$ to the
effective action. Thus by power counting, only graphs with a single vertex
can yield divergent contributions to the effective action. Of these, only the
one loop (tadpole) graphs are relevant, all the higher loop divergences are
removed by the renormalization procedure.
Although the potential provides masses for some
of the fields there will in general be massless fields which may cause
infrared divergences. We must therefore add an infrared regulator in the
form of a mass $M$ to the propagator and treat any mass terms in \action\ as
interactions, taking the limit $M\rightarrow 0$ in the final expressions.
Using dimensional regularization in $D=2+\epsilon$ dimensions and the
background field method [\Back] the bosonic graphs are readily calculated to
be
$$\eqalign{ \Gamma_{Div}({\rm bosons})=-&{m^2\over4}\Delta(0)
\left[\epsilon^{AB}\epsilon^{CD}\epsilon^{C'D'}\epsilon^{YZ} {\partial
C^a_{CC'}\over\partial X^{AY}} {\partial C^a_{DD'}\over\partial X^{BZ}}
\right.\cr & \left. \ \ \ \ \ \
+\epsilon^{A'B'}\epsilon^{C'D'}\epsilon^{CD}\epsilon^{Y'Z'} {\partial
C^a_{CC'}\over\partial \phi^{A'Y'}} {\partial C^a_{DD'}\over\partial
\phi^{B'Z'}} \right] \ , \cr} \eqn\bosons
$$
while the fermionic graphs are
$$\eqalign{
\Gamma_{Div}({\rm fermions}) =&{m^2\over4}\Delta(0)
\left[{1\over2}\epsilon^{AC}\epsilon^{BD}\epsilon^{C'D'}\epsilon^{YZ}
{\partial C^a_{CC'}\over\partial X^{AY}}
{\partial C^a_{DD'}\over\partial X^{BZ}} \right. \cr
& \left. \ \ \ \ \ \
+{1\over2}\epsilon^{A'C'}\epsilon^{B'D'}\epsilon^{CD}\epsilon^{Y'Z'}
{\partial C^a_{CC'}\over\partial \phi^{A'Y'}}
{\partial C^a_{DD'}\over\partial \phi^{B'Z'}} \right]\ , \cr}
\eqn\fermions
$$
where all the tensor expressions are evaluated at the background fields and
the bosonic propagator at zero momentum and renormalization scale $\mu$ is
$$
\Delta (0) = -{1\over 2\pi\epsilon}-{\rm ln}\left({M^2\over\mu^2}\right)
+{\rm finite} \ .
$$
One can see that the epsilon tensor terms in \bosons\ and \fermions\ are
different as a result of the twisted form of the Yukawa interactions. It is
not immediately obvious then that the bosonic and fermionic divergences
cancel. However by substituting in \onesusy\ it is not much trouble to see
that they do and hence $\Gamma_{Div}=0$. Thus Witten's ADHM model is
ultraviolet finite to all orders of perturbation theory. Therefore there is
no renormalization group flow in these models. This result may be expected,
but is not guaranteed, by supersymmetry as there is a general argument for
finiteness only for off-shell $(0,4)$ sigma models, with some modifications
required due to anomalies [\HPb].
\subsection{Integrating the Massive Modes}
In this section we will integrate out the massive modes. We shall postpone
the problem of anomalies in chiral supersymmetric models until
the next section. We assume for simplicity here that
$M^a_{AA'}=N^a_{AY}=0$, $D^a_{A'Y'} \ne 0$ so that the $X^{AY}$ and
$\psi^{A'Y}_-$ fields are massless, the $\phi^{A'Y'}$ and $\chi_-^{AY'}$
fields massive and the vacuum is at $\phi^{A'Y'}=0$. The theory is then only
quadratic in the massive fields and integrating over them is therefore exact
at the one loop level. This assumption also ensures that the interacting
theory breaks the $SU(2) \times Sp(k) \times SU(2) \times Sp(k')$ symmetry of
the free theory down to $SU(2) \times Sp(k')$ [\Wittena]. We may therefore
write
$$\eqalign{
C^a_{AA'} &=\epsilon_{A'B'}D^a_{AY'}\phi^{B'Y'}
+ \epsilon_{AB}\epsilon_{A'B'}E^a_{YY'} X^{BY}\phi^{B'Y'} \cr
&\equiv \epsilon_{A'B'}B^a_{AY'}(X)\ \phi^{B'Y'} \ . \cr}
$$
At this point it is necessary to split up the left handed fermions into
there massive and massless parts. If we introduce the zero modes $v^a_i(X)$,
$i=1,2,...,n-4k'$ of the fermion mass matrix, defined such that
$$
v^a_iB^a_{AY'} = 0, \ \ \ \ v^a_iv^a_j = \delta_{ij}
$$
and a similar set of massive modes $u^a_I(X)$, $I=1,2,...,4k'$ satisfying
$$
u^a_Iu^a_J = \delta_{IJ}, \ \ \ \ u^a_Iv^a_j =0
$$
then we may separate the $\lambda^a_+$ as
$$
\lambda^a_+ = v^a_i\zeta^i_+ + u^a_I\zeta^I_+ \ .
\eqn\lamodes
$$
so that the $\zeta^i_+$ are massless and the $\zeta^I_+$ massive.
We now rewrite the action \action\ in terms of the massless and massive
fields
$$
S = S_{massless} + S_{massive}
\eqn\S
$$
where $S_{massless}$ is the part of \action\ which only contains the
massless fields. Explicitly
$$
\eqalign{
S_{massless} = \int\! d^2x & \left\{ \epsilon_{AB} \epsilon_{YZ} {\partial}_= X^{AY} {\partial}_{\ne} X^{BZ}
+ i \epsilon_{A'B'} \epsilon_{YZ} \psi^{A'Y}_- {\partial}_{\ne} \psi^{B'Z}_-
\right. \cr
&\left.
+i\zeta^i_+ (\delta_{ij}{\partial}_=\zeta^j_+ + A_{ijAY}{\partial}_= X^{AY}\zeta^j_+)
\right\} \ ,\cr}
\eqn\Smassless
$$
where
$$
A_{ijAY}=v^a_i{\partial v^a_j\over\partial X^{AY}}
\eqn\Adef
$$
is the induced $SO(n-4k')$ connection and
$$\eqalign{
S_{massive} = \int\! & d^2x \left\{ \epsilon_{A'B'} \epsilon_{Y'Z'}
{\partial}_=\phi^{A'Y'}{\partial}_{\ne}\phi^{B'Z'} + i\epsilon_{AB} \epsilon_{Y'Z'} \chi^{AY'}_- {\partial}_{\ne} \chi^{BZ'}_-
\right. \cr
& \left.
+i\delta_{IJ}\zeta^I_+{\partial}_= \zeta^J_+ + iA_{IJAY} {\partial}_= X^{AY}
\zeta^I_+ \zeta^J_+
+2iA_{iJAY} {\partial}_= X^{AY}
\zeta^i_+\zeta^J_+ \right. \cr
& \left.
-im\epsilon_{B'C'}v^a_iE^a_{YY'}\zeta^i_+\phi^{C'Y'}\psi^{B'Y}_-
-im\epsilon_{B'C'}u^a_IE^a_{YY'}\zeta^I_+\phi^{C'Y'}\psi^{B'Y}_-
\right. \cr
& \left.
-imu^a_{I}B^a_{BY'}\zeta^I_+\chi_-^{BY'}
-{m^2\over8}\epsilon^{AB}\epsilon_{C'D'}
B^a_{AY'}B^a_{BZ'}\phi^{C'Y'}\phi^{D'Z'} \right\} \ ,\cr}
\eqn\Smassive
$$
where $A_{IJAY}=u^a_I{\partial u^a_J}/ {\partial X^{AY}}$ and
$A_{iJAY}=v^a_i{\partial u^a_J}/ {\partial X^{AY}}$.
The classical low energy effective action is simply obtained by considering
the most general action possible which is compatible with all of the
symmetries of the theory. To calculate the effective action quantum
mechanically we will integrate over the massive fields and discard any higher
derivative terms. The presence of higher derivative terms in the effective
action, which are suppressed by powers $p/m$ where $p$ is the low energy
momentum scale, would ruin the renormalizability and prevent a simple
geometrical sigma model interpretation of the effective theory.
First we notice that because of the nontrivial definition of the massless
left handed fermions \lamodes , $S_{massless}$ is not (0,4) supersymmetric by
itself as it is missing a four fermion interaction term. The problem is
rectified by noting that there is a tree graph, with a single internal
$\phi^{A'Y'}$ field propagating, which contributes to the low energy effective
action. In order to avoid the singular behaviour of the propagator at zero
momentum, when calculating this graph it is helpful to use the massive
propagator for $\phi^{A'Y'}$, obtained from the last term in \Smassive .
At this point it is useful to write, using \foursusy\ ,
$$
B^a_{AY'}B^a_{BZ'} = \epsilon_{AB}\epsilon_{Y'Z'}\Omega
$$
where
$$
\Omega(X)={1\over4k'}\epsilon^{AB}\epsilon^{Y'Z'}B^a_{AY'}B^a_{BZ'} \ .
\eqn\Om
$$
The last term in \Smassive\ becomes
$$
-{m^2\over4}\Omega(X)\ \phi^2 \ ,
$$
and hence can be interpreted as the $X^{AY}$
dependent mass term for $\phi^{A'Y'}$. The tree graph can then be seen to
contribute the four fermion term
$$
-{1\over2}\zeta^i_+ \zeta^j_+F^{ij}_{A'YB'Z}\psi^{A'Y}_- \psi^{B'Z}_-
$$
where
$$
F^{ij}_{A'YB'Z} =
2 \epsilon_{A'B'} \epsilon^{Y'Z'}\Omega^{-1}
v^a_i E^a_{(Y|Y'} v^b_j E^b_{|Z)Z'}
\ ,
\eqn\Fdef
$$
which we will later relate to the field strength tensor.
We may now discard all vertices with only one massive field in \Smassive\ and
examine the one loop contributions to the effective action. Inspection of the
quadratic terms in $S_{massive}$ shows there are no contributions to the
gauge connection in \Smassless. Furthermore, inspection shows that of all
the other possible contributions only those corresponding to the effective
potential do not involve higher order derivatives of the massless fields.
A check on this is to note that any terms which are second
order in the derivatives are logarithmically divergent, and by finiteness
of the model, must therefore vanish.
To calculate the effective potential we simply
set ${\partial}_= X^{AY} = {\partial}_{\ne} X^{AY}= \psi^{A'Y}_-=0$. Thus only the last two terms
in \Smassive\ need be considered (we no longer use a massive propagator for
$\phi^{A'Y'}$). The effective potential then receives the standard bosonic
and fermionic contributions (in Euclidean momentum space)
$$
V_{eff}({\rm bosons}) = {\alpha'\over4\pi}\sum_{n=1}^{\infty}{1\over n} \int\!
d^2 p\ {\rm Tr}\left[{\epsilon_{C'D'} \epsilon^{AB}B^a_{AY'}B^a_{BZ'}\over
4p^2/m^2}\right]^n \eqn\Veffbose
$$
and
$$
V_{eff}({\rm fermions}) = -{\alpha'\over4\pi} \sum_{n=1}^{\infty}{1\over n}
\int\! d^2 p\ {\rm Tr}\left[{ u^a_Iu^{bI}B^a_{CY'}B^b_{DZ'}\over 2 p^2/m^2}
\right]^n \ .
\eqn\Vefffermi
$$
Now the definition \Om\ yields the following expressions
$$
\epsilon_{C'D'}\epsilon^{AB}B^a_{AY'}B^a_{BZ'} =
2\epsilon_{C'D'}\epsilon_{Y'Z'}\Omega
$$
and
$$
u^a_Iu^{bI}B^a_{CY'}B^b_{DZ'} =
\epsilon_{CD}\epsilon_{Y'Z'}\Omega \ .
$$
Therefore \Veffbose\ is completely canceled by \Vefffermi\ and there is no
contribution to the effective potential.
{}From the above analysis we conclude that the effective quantum action of the
massless fields is
$$\eqalign{
S_{effective} & = \int\! d^2x \left\{ \epsilon_{AB} \epsilon_{YZ} {\partial}_= X^{AY} {\partial}_{\ne} X^{BZ}
+ i\epsilon_{A'B'} \epsilon_{YZ} \psi^{A'Y}_- {\partial}_{\ne} \psi^{B'Z}_-
\right. \cr
&\left. \ \ \
+i\zeta^i_+ (\delta_{ij}{\partial}_= \zeta^j_+ + A_{ijAY}{\partial}_= X^{AY}\zeta^j_+)
- {1\over2}\zeta^i_+ \zeta^j_+F^{ij}_{A'YB'Z}\psi^{A'Y}_- \psi^{B'Z}_-
\right\} \ .\cr}
\eqn\effective
$$
This is simply the action of the general (0,4) supersymmetric nonlinear sigma
model [\HPT,\PT], although the right handed superpartners of $X^{AY}$ are
'twisted'. As with the original theory \action , the low energy effective
theory \effective\ admits a $(0,1)$ superfield form. Introducing the
superfield $\Lambda^i_+ =\zeta^i_+ + \theta^-F^i$ then allows us (after
removing $F^i$ by it's equation of motion) to express \effective\ as
$$\eqalign{
S_{effective} = -i\int\! d^2xd & \theta^- \left\{
\epsilon_{AB}\epsilon_{YZ} D_-{\cal X}^{AY}{\partial}_{\ne} {\cal X}^{BZ} \right. \cr
& \left.\ \ \ \ \
- i\Lambda^i_+ (\delta_{ij} D_- \Lambda^j_+
+ A_{ijAY} D_-{\cal X}^{AY}\Lambda^j_+) \right\} \ , \cr}
\eqn\superaction
$$
provided that $F^{ij}_{A'YB'Z}$ satisfies
$$
F^{ij}_{A'YB'Z} = I^A_{\ A'}I^B_{\ B'}F^{ij}_{AYBZ}
\eqn\constraint
$$
where $F^{ij}_{AYBZ}$ is the curvature of the connection \Adef ,
$$
F^{ij}_{AYBZ} = \partial_{AY} A_{ijBZ} - \partial_{BZ} A_{ijAY}
+ A_{ikAY}A_{kjBZ}- A_{ikBZ} A_{kjAY}\ .
$$
This is just the familiar constraint on (0,4) models that the field strength
be compatible with the complex structure [\HPT,\PT]. Furthermore it is not hard
to check that $S_{effective}$ does indeed possess the
full on-shell (0,4) supersymmetry (the superspace formulation \superaction\
ensures only off-shell (0,1) supersymmetry) precisely when \constraint\ is
satisfied.
For the $k=k'=1,\ n=8$ model above it is straightforward to determine
the non zero components of $v^a_i$ and $u^a_I$ as
$$
\eqalign{
v^{YY'}_{ZZ'}& =
{\rho\over\sqrt{\rho^2+X^2}}\delta^Y_Z \delta^{Y'}_{Z'}\ \ \ \ \ \
v^{AY'}_{ZZ'} =
-{\sqrt{2}\over\sqrt{\rho^2+X^2}}X^{A}_{\ Z}\delta^{Y'}_{Z'}
\cr
u^{YY'}_{BZ'}& = {\sqrt{2}\over\sqrt{\rho^2+X^2}}X_{B}^{\ Y}\delta^{Y'}_{Z'}
\ \ \ \ \ \
u^{AY'}_{BZ'} = {\rho\over\sqrt{\rho^2+X^2}}\delta^A_{B}\delta^{Y'}_{Z'}
\ ,\cr}
\eqn\uv
$$
and the mass term \Om\ is
$$
\Omega ={1\over2}(X^2 + \rho^2) \ .
$$
The gauge field $A_{ijAY}$ obtained from \uv\ is simply that of
a single instanton on the manifold ${\bf R}^4$
$$
A^{YY'ZZ'}_{AX} = -
\epsilon^{Y'Z'}
{(\delta^{Z}_{X}X_{A}^{\ \ Y}
+ \delta^{Y}_{X}X_{A}^{\ \ Z})\over\rho^2+X^2} \ ,
\eqn\A
$$
and the four fermion vertex \Fdef\ is
$$
F^{TT'UU'}_{A'YB'Z} = {4\rho^2\over(X^2 + \rho^2)^2}
\epsilon_{A'B'} \epsilon^{T'U'} \delta^{T}_{(Y}\delta^{U}_{Z)}\ ,
\eqn\F
$$
which is precisely the field strength of an instanton, justifying our
presumptuous notation, and can be easily seen to satisfy \constraint .
\subsection{Anomalies}
So far we have ignored the possibility of anomalies in the quantum theory.
While the original theory \action\ is simply a linear sigma model and
therefore possesses no anomalies, this is not the case for the effective theory
\effective. It is well known that off-shell $(0,4)$ supersymmetric sigma models
suffer from chiral anomalies which break spacetime gauge and coordinate
invariance, unless the gauge field can be embedded in the spin connection of
the target space. In addition, working in $(0,1)$ superspace
only ensures that $(0,1)$ supersymmetry is preserved and there are
also extended supersymmetry anomalies as the $(0,4)$ supersymmetry is not
preserved. We therefore expect that we will have
to add finite local counter terms to \effective\ at all orders of perturbation
theory so as to cancel these anomalies. This requires that the spacetime metric
and antisymmetric tensor fields become non trivial at higher orders of
$\alpha'$, while on the other hand the gauge connection is unaffected [\HPa].
An alternative way of viewing this is to note that although the action
\effective\ is classically conformally invariant, when quantized it may not be
ultraviolet finite and hence break scale invariance. There is a power counting
argument which asserts that off-shell $(0,4)$ supersymmetric models are
ultraviolet finite to all orders of perturbation theory [\HPb]. This
argument is further complicated by sigma model anomalies and it has been stated
that only the non chiral models are ultraviolet finite. Indeed while
off-shell $(0,4)$ supersymmetric theories are one loop finite,
there is a two loop contribution of the form ${\rm Tr}(R^2 -F^2)$ [\NDL] which
certainly does not vanish in general. Normally this leads
to non vanishing $\beta$-functions which we must then take into account when
determining the conformal fixed point of the renormalization group flow.
However, in models with off-shell $(0,4)$ supersymmetry the non vanishing
$\beta$-functions can be canceled by redefining the spacetime fields at each
order of $\alpha'$, in such a way as to ensure that supersymmetry is
preserved in perturbation theory [\HPa]. This has been well studied and
verified up to three loops. Thus the ultraviolet divergences which arise in
the quantization of off-shell $(0,4)$ models are really an artifact of the
use of a renormalization scheme which does not preserve the supersymmetry.
The off-shell $(0,4)$ models are ultraviolet finite in an appropriate
renormalization scheme.
However the model here has only on-shell $(0,4)$ supersymmetry and these
finiteness arguments do not immediately apply. At least in the $k=k'=1$
case however the gauge group $SO(4)\cong SU(2)\times SU(2)$ contains a subgroup
$Sp(1)\cong SU(2)$ which admits three complex structures obeying the
algebra of the quaternions. This endows the target space of the left handed
fermions with a hyper Kahler structure and facilitates an off-shell
formulation using constrained superfields [\HPc]. We may therefore expect
that it is ultraviolet finite in the same manner as the off-shell models
described above.
In [\HPa] the necessary field redefinitions were derived to order $\alpha'^2$ for
$(0,4)$ supersymmetric sigma models. Both the target space metric and
antisymmetric tensor field strength receive corrections to all orders in
$\alpha'$. Howe and Papadopoulos found that in order to maintain $(0,4)$
supersymmetry in perturbation theory the target space metric (which is flat
here at the classical level) must receive corrections in the form of a
conformal factor
$$
\epsilon_{AB} \epsilon_{YZ} \rightarrow
\left(1 - {3\over2}\alpha' f - {3\over16}\alpha'^2 \triangle f + ... \right)
\epsilon_{AB} \epsilon_{YZ}\ .
\eqn\conf
$$
They also showed, up to three loop order, that these redefinitions cancel the
ultraviolet divergences which arise when one renormalizes \effective\
using standard $(0,1)$ superspace methods, which do not ensure
(0,4) supersymmetry is preserved perturbatively. In addition, the
antisymmetric field strength tensor becomes $H = -{3\over4}\alpha'*df$ so as to
cancel the gauge anomaly
$dH=-{3\alpha'\over4}{\rm Tr} F \wedge F$. Furthermore there are no
corrections to the instanton gauge field.
For the instanton number one model considered here, Howe and Papadopoulos give
the
function $f$ as
$$
f = -\triangle {\rm ln}(X^2 + \rho^2)\ ,
$$
where $\triangle$ is the flat space Laplacian. It is a simple matter to
calculate the conformal factor \conf\ and hence the target space metric as
$$
g_{AYBZ} = \left(1 + 6\alpha'{X^2+2\rho^2 \over(X^2+\rho^2)^2}
- 18\alpha'^2 {\rho^4 \over(X^2+\rho^2)^4}+...
\right)\epsilon_{AB} \epsilon_{YZ} \ . \eqn\metric
$$
To order $\alpha'$ this is the solution of Callan, Harvey and Strominger
[\CHS] obtained by solving the first order equations of motion of the
10 dimensional heterotic string (although with $n=6$ rather than $n=8$
in their notation). Thus the target space has been curved around the instanton
by
stringy effects but remains non singular so long as $\rho \ne 0$. The case
$\rho=0$ is of great interest as it provides a string theoretic
compactification of instanton moduli space. We will briefly discuss this in
the next section.
\chapter{Concluding Remarks}
In the above we found the order $\alpha'^2$ corrections to the low energy
effective action of the ADHM sigma model. Such solutions have been discussed
before [\CHS] and we agree with their solution to first order. In our
calculations we have expanded in the parameter
$$
\alpha'\Omega^{-1}={2\alpha'\over{X^2 +\rho^2}}
$$
and hence our approximations are valid for all $X$ if $\rho^2 \gg \alpha'$ and for
$X^2 \gg \alpha'$ even when $\rho^2$ is small. An interesting question
raised is what are the stringy corrections to the classical instanton in the
extreme case that it's size vanishes? One can see from \metric\ that the
order $\alpha'$ corrections persist when $\rho=0$ so the effective theory is non
trivial. It has been conjectured that there should be a $(4,4)$
supersymmetric sigma model for instantons of zero size [\GS,\GSt] which could
be constructed from a massive linear $(4,4)$ supersymmetric model. In [\GSt]
the conditions for the ADHM model to possess full (4,4) supersymmetry in the
infrared limit were derived. There it was found that the metric must be
conformally flat, with the metric satisfying Laplace's equation. This is in
agreement with what we have found here in the $\rho=0$ case above (see \conf\
and \metric ) and lends some additional support to this conjecture.
If we start with the linear sigma model \action\ with $\rho=0$ the
$\lambda_+^{AY}$ fields are massless and decouple from the theory. The vacuum
states are defined by $X^{AY}=0$ or $\phi^{A'Y'}=0$ and there is a symmetry
between $X^{AY}$ and $\phi^{A'Y'}$. Let us assume we choose the
$\phi^{A'Y'}=0$ vacuum. Then as before the fields $X^{AY}$ and $\psi_-^{AY'}$
are massless and the $\phi^{A'Y'}$, $\chi_-^{AY'}$ and $\lambda_+^{YY'}$
fields all have masses $ m\sqrt{X^2/2}$. Upon integrating out the massive
fields we would simply obtain a free field theory, which trivially
possesses $(4,4)$ supersymmetry. At the degenerate vacuum $X^{AY}=0$
however, all fields are massless and there is a single interaction term
$m\lambda_{+YY'}\phi^{\ \ Y'}_{A'}\psi_-^{A'Y}$. Thus the moduli space of
vacua does not have a manifold structure. For $X^{AY} \ne 0$ the vacuum
states are simply ${\bf R}^4$ but at the point $X^{AY} = 0$ lies another
entire copy of ${\bf R}^4$ (parameterized by the $\phi^{A'Y'}$). This odd
state of affairs is smoothly resolved if we first construct the effective
theory and then take the limit of vanishing instanton size.
Let us now take the limit $\rho \rightarrow 0$ of the effective action
\effective . It should be noted that the Yang-Mills instanton has shrunk to
zero size but it has not disappeared in the sense that the topological charge
remains equal to one. Unfortunately our expressions are not a priori valid
near $X=0$. Nevertheless we will try to shed some light about what the
complete string theory solution could be in that region. When $\rho$ vanishes
both the field strength \Fdef\ and the $O(\alpha')$ sigma model anomaly vanish. We
are however, still left with a non trivial metric and anti symmetric tensor.
It seems reasonable to assume then that all the anomalies are canceled by
these. The metric then has the exact conformal factor
$$
g_{\mu\nu} = \left(1 - {3\alpha'\over2}f\right)\delta_{\mu\nu} \ ,
\eqn\cfact
$$
and antisymmetric field
$$
H_{\mu\nu\rho} = -{3\alpha'\over4}\epsilon_{\mu\nu\rho\lambda}\partial^{\lambda}f \
,
\eqn\H
$$
where $f = -4/X^2$ and we have switched to a more convenient notation. This
geometry is similar to the one discussed by Callan, Harvey and
Strominger [\CHS], although the anti symmetric field is not the same
and leads to a different interpretation in the limit $X^2 \rightarrow 0$ as we
will shortly see. The target space is non singular, asymptotically flat and has
a
semi-infinite tube with asymptotic radius $\sqrt{6\alpha'}$, centered around the
instanton. That is to say the apparent singularity at $X^2=0$ in \metric\ is
pushed off to an "internal infinity" down the infinite tube. Thus the
problematic
$X^{AY}=0$ vacua are pushed an infinite distance away and the manifold
structure
is preserved. The resolution of this description with the non manifold picture
described above has been discussed by Witten [\Wittenb].
In the limit $X^2 \ll \alpha'$ the modified spin connection with torsion becomes,
where $\alpha,\beta$ are vierbein indices,
$$
\eqalign{
\omega_{\mu}^{(-)\alpha\beta} &\equiv \omega_{\mu}^{\alpha\beta}
+ H_{\mu}^{\ \alpha\beta} \cr
&=
-(\delta^\alpha_{\mu} \delta^\beta_{\nu}
-\delta^\alpha_{\nu} \delta^\beta_{\mu}
+ \epsilon_{\mu\nu}^{\ \ \ \alpha\beta}){X^{\nu}\over X^2}
\cr } \ ,
\eqn\ominus
$$
which is a flat connection! That is to say far down the infinite tube
the torsion parallelizes the manifold (which is asymptotically $S^3 \times {\bf
R}$ and is indeed parallelizable). The gauge connection is also flat (for
$X^{AY}
\ne 0$) and can therefore be embedded into the generalized spin connection
\ominus\ (they are both $so(4)$ valued). The low energy effective
theory therefore possesses $(4,4)$ supersymmetry in the region $X^2
\rightarrow 0$ and is free of anomalies there. This supports our assumption
that the anomalies are canceled and the expressions \cfact\ and \H\ are
exact, at least in this region. For the region $X^2 \gg \alpha'$ our perturbative
expansion is valid and the the theory only possesses $(0,4)$ supersymmetry
since the gauge connection can not be embedded into the spin connection.
Although in a similar spirit in the limit $X^2 \rightarrow \infty$ the
curvatures vanish and the theory is free and again has $(4,4)$
supersymmetry. In a sense then the $\rho=0$ ADHM instanton can be viewed as
a soliton in the space of string vacua interpolating between two $(4,4)$
supersymmetric sigma models, just as the target space can be viewed as
interpolating between two supersymmetric ground states of supergravity
[\GT].
However, since we still have instanton
number one, the vector bundle is non trivial whereas the tangent bundle is
trivial. Thus in the region $X^2 \rightarrow 0$ we can only identify the spin
connection with the gauge connection locally. $(4,4)$ supersymmetry is then
broken by global, non perturbative effects back to $(0,4)$ supersymmetry.
This is reflected by the observation [\GT] that the $S^3 \times M^7$
compactification of $D=10$ supergravity breaks half the supersymmetry.
I would like to thank G. Papadopoulos and P.K. Townsend for their advice and
Trinity College Cambridge for financial support.
\refout
\end
. Therefore, in the region $X^2 \ll \alpha'$, the torsion
parallelizes the manifold
|
\section{Introduction}
This talk is organized as follows. First, we briefly
discuss the supersymmetric standard model, in order to introduce
the parameter space we are considering.
Next, we discuss the prediction for the strong coupling constant
and the bottom pole mass with and without including GUT threshold
effects, in the small $\tan\beta$ case. Lastly we discuss our results
in the large $\tan\beta$ regime.
The minimal supersymmetric model is attractive in that it solves the
hierarchy problem. It does this, however, at the expense of doubling
the number of degrees of freedom of the standard model. This is
problematic, as supersymmetry must be broken and hence the number of
new parameters needed to describe the model is in general quite large.
An organizing principle is needed in order to reduce the number of
parameters. In a minimal supergravity scenario the number of
supersymmetry breaking parameters
needed to describe the supersymmetric model are few: a universal
scalar mass $M_0$, a universal gaugino mass $M_{1/2}$, a universal
trilinear scalar coupling $A_0$, and a bilinear scalar coupling $B$.
In addition there is a supersymmetric Higgs mass term, $\mu$.
Given values for these five parameters at the GUT scale, we use the
renormalization group equations~\cite{RGEs} (RGE's)
to determine the various
particle masses and couplings at the weak scale. For a large top-quark
mass, one of the Higgs boson masses is driven negative, and the
radiative breaking of $SU(2)\times U(1)$ symmetry becomes manifest.
The Higgs bosons obtain vev's $v_1$ and $v_2$, and we can determine
the $Z$-boson mass. In practice it is convenient to assume
electroweak symmetry breaks radiatively and to take $M_Z$ as an input
parameter, as well as the ratio of vev's $\tan\beta\equiv v_2/v_1$.
We then determine $\mu^2$ and $B$ from the symmetry breaking conditions.
To summarize, then, the supersymmetric model is parametrized by
$$M_0,\ \ M_{1/2},\ \ A_0,\ \ \tan\beta,\ \ {\rm sign}(\mu).$$
The exact one-loop corrections to the masses, gauge couplings and
Yukawa couplings of the minimal supersymmetric model are described
in Ref. [2]. These corrections are essential ingredients
for accurate tests of grand unification. They allow one to extract
the underlying \mbox{\footnotesize$\overline{\it DR}$}~ parameters from a given set of measured observables.
The \mbox{\footnotesize$\overline{\it DR}$}~ parameters can then be run up to a high scale to explore
the consequences of different unification hypotheses.
Alternatively, the radiative corrections can be used to translate
various limits into excluded regions of the \mbox{\footnotesize$\overline{\it DR}$}~ parameter space. This
is illustrated in Fig.~1, where we show the excluded region of the $M_0,
\ M_{1/2}$ parameter space at the one-loop level, from current
experimental constraints.
\begin{figure}[t]
\epsfysize=3in
\epsffile[30 495 600 755]{f1.ps}
\begin{center}
\parbox{5.5in}{
\caption[]{\footnotesize Excluded region (shaded)
of the $M_0,\ M_{1/2}$ plane,
for $\tan\beta=2,\ m_t=175$ GeV, $A_0=0$, and $\alpha_s=0.117$.
All masses are evaluated at one-loop.
The symbols indicate which experimental constraint is relevant:
$\chi^+\Rightarrow m_{\chi^+} > 47$ GeV; $\tilde g\Rightarrow
m_{\tilde g}>125$ GeV; $\tilde\nu\Rightarrow m_{\tilde\nu}>42$ GeV;
$h\Rightarrow m_h>60$ GeV.}}
\end{center}
\end{figure}
In the following, we treat all supersymmetric threshold corrections
in a complete one-loop analysis.\footnote{See Chankowski
et al.~\cite{Chankowski} for a similar
treatment of finite corrections to $\sin^2\theta_W$.} Our work stands
in contrast to most previous studies, which are based on the ``leading
logarithm approximation." For the gauge coupling threshold corrections,
this approximation involves taking the
standard-model value of $\sin^2\theta_W$ and adding the logarithmic
parts of the SUSY threshold corrections. The approximation works well
if all of the SUSY particle masses are much greater than $M_Z$, in which
case the decoupling theorem implies that the finite effects of the
SUSY particles are negligible for all low-energy observables.
However, in realistic models it is not unusual for the supersymmetric
spectrum to contain light particles of order the $Z$-mass. In this
case the leading logarithm approximation breaks down.
This is illustrated in Fig.~2, where we compare the value of
$\alpha_s$ and $m_b$ in the leading logarithm approximation (LLA) with
the value obtained in the full calculation.
In this talk we use the full set of one-loop radiative
corrections to evaluate the \mbox{\footnotesize$\overline{\it DR}$}~ gauge and Yukawa couplings. The \mbox{\footnotesize$\overline{\it DR}$}~
couplings serve as the boundary conditions for the two-loop gauge and
Yukawa coupling renormalization group equations, which
determine the couplings at very high scales.
\begin{figure}[t]
\epsfysize=3in
\epsffile[35 475 600 735]{f2.ps}
\begin{center}
\parbox{5.5in}{
\caption[]{\footnotesize Comparison of (a) $\alpha_s$
and (b) $m_b$ in the leading logarithm
approximation (LLA) versus the full one-loop calculation, for $M_0=60$
GeV, $\tan\beta=2, \ m_t=175$ GeV, $A_0=0$, and $\mu>0$.
The dashed line shows the result if
the non-universal corrections are neglected.}}
\end{center}
\end{figure}
In what follows, we have converted the strong coupling
to the \mbox{\footnotesize$\overline{\it MS}$}~ scheme so that by $\alpha_s$ we refer to the
standard \mbox{\footnotesize$\overline{\it MS}$}~ value evaluated at the scale $M_Z$.
By $m_b$ we refer to the bottom-quark pole mass.
\section{Small $\tan\beta$}
As a reference point, we show in Fig.~3 contours of
$m_b$ and $\alpha_s$ in the $M_0,\ M_{1/2}$ plane, with no GUT
thresholds, $\tan\beta=2$, $m_t=175$ GeV, and $A_0$=0.
We confine our attention to the region of the theory
which is more natural, i.e. to the region where the
superpartner masses are less than about 1 TeV.
We find the strong coupling is large ($\alpha_s>0.127$) compared to
the PDG value \cite{pdg} $\alpha_s=0.117\pm0.005$. Similarly,
the bottom mass is quite large ($m_b>6$ GeV), far outside the preferred
region which we take to be $4.7 < m_b < 5.2$ GeV.
\begin{figure}[t]
\epsfysize=3in
\epsffile[30 468 600 728]{f3.ps}
\begin{center}
\parbox{5.5in}{
\caption[]{\footnotesize Contours of $\alpha_s$ (solid lines)
and $m_b$ (dashed) in the $M_0$, $M_{1/2}$
plane, with $\tan\beta=2$, $m_t=175$ GeV, and $A_0=0$.
The upper left hand corners are excluded on cosmological grounds
(charged LSP) and the lower shaded regions are excluded by
particle searches. The $m_b$ contours are labeled in GeV.}}
\end{center}
\end{figure}
The experimental uncertainty in the determination of $\alpha_s$ is
primarily due to the uncertainty in determining the electromagnetic
coupling at the $Z$-scale. We use the value recently determined by
Eidelman and Jegerlehner \cite{Eidelman}.
The one-sigma uncertainty in our input
$\alpha_{EM}(M_Z)$ results in a one-sigma uncertainty
in our output $\alpha_s$ of about $\pm0.001$, and an uncertainty
of typically 0.07 GeV in our bottom mass prediction.
Martin and Zeppenfeld \cite{Martin} and Swartz \cite{Swartz}
have also performed analyses to determine $\alpha_{EM}(M_Z)$.
The central value for $\alpha_s$ increases by about 0.001
if we use the value of $\alpha_{\rm EM}(M_Z)$ as determined by
Martin and Zeppenfeld, and it increases by about 0.002 if
we use the value of $\alpha_{EM}(M_Z)$ from Swartz.
The bottom mass is reduced if we go further into the small
$\tan\beta$ region, due to the large top-quark Yukawa coupling
in the bottom mass renormalization group equation.
We consider the case where the top Yukawa is as large as possible;
we set $\lambda_t(M_{\rm GUT})=3$, which is on the verge of the
nonperturbative regime. In this case we will obtain the smallest
possible bottom mass. As seen in Fig.~4,
for $m_t=180$ GeV the bottom mass is less than 5.2 GeV
and the strong coupling is less than 0.127 only if
the squark masses are in the TeV region.
\begin{figure}[t]
\epsfysize=3.5in
\epsffile[25 398 600 718]{f4.ps}
\begin{center}
\parbox{5.5in}{
\caption[]{\footnotesize The bottom-quark mass and $\alpha_s$
vs. $m_t$ for the case of no GUT-scale thresholds,
for various values of $\tan\beta$, with $A_0=0$, $\mu>0$, and
$\lambda_t(M_{\rm GUT})=3$. The circles on the right (solid) leg in
each pair of lines corresponds to $M_{1/2}$ equal to (from top to
bottom) 60, 100, 200, 500 GeV, with $M_0$ fixed at 100 GeV.
The circles on the left (dashed) leg corresponds to $M_0$ equal to
100, 200, 400 and 1000 GeV, with $M_{1/2}=100$ GeV.
The horizontal dot-dashed lines indicate
$m_b=5.2$ GeV and $\alpha_s=0.127$. The $\times$'s mark points
with one-loop Higgs mass $m_h<60$ GeV.}}
\end{center}
\end{figure}
If we consider particular GUT models, we can determine whether the GUT
threshold corrections to the gauge and Yukawa couplings can help improve
the situation. We parametrize
the GUT threshold corrections by $\varepsilon_g$ and
$\varepsilon_b$, where
$$ g_3(M_{\rm GUT}) = g_{\rm GUT}(M_{\rm GUT})\left(
1+\varepsilon_g\right)\ ,$$
$$ \lambda_b(M_{\rm GUT})=\lambda_\tau(M_{\rm GUT})\left(
1+\varepsilon_b\right)\ ,$$
where $\lambda_b$ and $\lambda_\tau$ are the $b$- and $\tau$-Yukawa
couplings, and $M_{\rm GUT}$ is defined as the scale at which $g_1$
and $g_2$ meet, $g_{\rm GUT}\equiv g_1(M_{\rm GUT})=g_2(M_{\rm GUT})$.
A smaller value of $\alpha_s$ requires $\varepsilon_g<0$.
In the small $\tan\beta$ region the bottom mass is tightly correlated
with the value of the strong coupling. Hence, setting
$\varepsilon_g<0$ reduces both $\alpha_s$ and $m_b$. In fact,
the bottom mass is an order of magnitude more sensitive to
$\varepsilon_g$ than to $\varepsilon_b$.
In what follows we examine the GUT corrections
in two SU(5) GUT models.
\begin{figure}[t]
\epsfysize=3in
\epsffile[30 468 600 728]{f5.ps}
\begin{center}
\parbox{5.5in}{
\caption[]{\footnotesize Contours of the smallest possible $\alpha_s$
(solid lines) and $m_b$ (dashed) consistent with nucleon decay
in (a) minimal SU(5), and (b) missing doublet SU(5), with
$m_t=175$ GeV, $A_0=0$ and $\mu>0$. In (a) $\lambda_t(M_{\rm GUT})=3$
($\tan\beta\simeq1.4$) and in (b) $\tan\beta=2$.
The shaded regions are phenomenologically excluded.
The $m_b$ contours are labeled in GeV.}}
\end{center}
\end{figure}
\begin{figure}[t]
\epsfysize=3in
\epsffile[30 468 600 728]{f6.ps}
\begin{center}
\parbox{5.5in}{
\caption[]{\footnotesize Contours of $\alpha_s$ (solid)
and $m_b$ (dashed, labeled in GeV) with no GUT thresholds,
$m_t=175$ GeV, $A_0=0$ and $\tan\beta=30$.}}
\end{center}
\end{figure}
\begin{figure}[t]
\epsfysize=3in
\epsffile[30 468 600 728]{f7.ps}
\begin{center}
\parbox{5.5in}{
\caption[]{\footnotesize Contours of $\alpha_s$ (solid)
and $m_b$ (dashed, labeled in GeV) in missing doublet SU(5), with
$m_t=175$ GeV, $A_0=0$ and $\tan\beta=30$.}}
\end{center}
\end{figure}
In the minimal SU(5) model \cite{minSUfive}, the gauge coupling
threshold correction $\varepsilon_g'$ is given by
\cite{minSUfive epsg}
\begin{equation}
\varepsilon'_g\ =\ {3g_{\rm GUT}^2\o40\pi^2}\,\log\left(M_{H_3}\over
M_{\rm GUT}\right)\ , \label{epsprime}
\end{equation}
where $M_{H_3}$ is the mass of the color-triplet Higgs particle that mediates
nucleon decay. From this expression, we see that $\varepsilon'_g < 0$ whenever
$M_{H_3} < M_{\rm GUT}$. However, $M_{H_3}$ is bounded from below by proton
decay experiments. The $M_{H_3}$ mass limit is of the form \cite{Hisano}
$$ M_{H_3} > {\cal M}\ {|1+y^{tK}|\over\sin2\beta} f(\tilde w, \tilde d,
\tilde u,\tilde e) $$
where ${\cal M}$ is a nuclear matrix element, $y^{tK}$ parametrizes
the amount of third generation mixing, and $f$ is a function
of the wino, squark and slepton masses.
For the conservative choices ${\cal M}=0.003$ GeV$^3$
and $|1+y^{tK}|=0.4$ we find that $M_{H_3}^{\rm min}>M_{\rm GUT}$
unless $M_0>500$ GeV and $M_{1/2} \ll M_0$. Thus, in most of the
parameter space, $\varepsilon_g'>0$.
For this reason, in minimal SU(5), $\alpha_s$ is typically even
larger than in the case of no GUT thresholds, as illustrated in
Fig.~5(a). In order to obtain the smallest possible $\alpha_s$ and
$m_b$, we have set $\lambda_t(M_{\rm GUT})=3$ in Fig.~5(a).
Thus we end up with rather small values for $\tan\beta$
($\sim 1.3$-1.6), and the Higgs mass constraint
rules out a large part of parameter space.
Only in the region $M_0 \gg M_{1/2}$, where the proton
decay amplitude is suppressed, is the strong coupling
reduced relative to the case with no GUT thresholds.
The smallest
value of $\alpha_s$ occurs in this region, a somewhat large value of
0.124. The bottom-quark mass is similarly
on the high side of the preferred region.
We have applied the most favorable Yukawa correction $\varepsilon_b$
given in Wright~\cite{wright}, subject to Yukawa coupling perturbativity
constraints (see Bagger et al.~\cite{bmp}).
The missing-doublet model is an alternative SU(5) theory in which
the heavy color-triplet Higgs particles are split naturally
from the light Higgs doublets \cite{missing-doublet}. In this
model the GUT gauge threshold correction is given by \cite{Yamada}
\begin{equation}
\varepsilon_g^{\prime\prime}\ =
\ {3g_{\rm GUT}^2\o40\pi^2}\,\Biggl\{\log\left(M_{H_3}^{\rm eff}
\over M_{\rm GUT}\right) - {25\o2}\log5 + 15\log2\Biggr\}\ \simeq
\ \varepsilon'_g - 4\%\ .
\label{mdmodel}
\end{equation}
Thus, for fixed $M_{H_3}$, the missing-doublet model has the same
threshold correction as the minimal SU(5) model, minus 4\%. In
eq.~(\ref{mdmodel}), $M_{H_3}^{\rm eff}$ is the effective mass that enters
into the proton decay amplitude, so the bounds on $M_{H_3}$ in the minimal
SU(5) model also apply to $M_{H_3}^{\rm eff}$ in the missing-doublet
model.
The large negative correction in eq.~(\ref{mdmodel}) is due to the mass
splitting in the {\bf 75} representation, and gives rise to much smaller
values for $\alpha_s$ and $m_b$. This is illustrated in
Fig.~5(b), where we show contours of $\alpha_s$ and $m_b$
in the $M_0,\ M_{1/2}$
plane, with $M_{H_3}^{\rm eff}=M_{H_3}^{\rm min}$, at $\tan\beta=2$.
We find (even without going into the far infrared top Yukawa fixed point
region) values of both the strong coupling and the bottom mass near
their central values.
\section{Large $\tan\beta$}
Again, as a reference point, we show in Fig.~6 the strong coupling
and bottom mass
prediction with no GUT corrections, for $A_0=0,\ m_t=175$ GeV,
at $\tan\beta=30$. The $\alpha_s$ predictions are not significantly
different from the small $\tan\beta$ case. The bottom mass
is significantly influenced by the large finite corrections~\cite{mb},
$${\Delta m_b\over m_b}\sim -{\mu\tan\beta\over16\pi^2m_{\tilde q}^2}
\left({8\over3}g_3^2 m_{\tilde g} + \lambda_t^2 A_t\right)\ ,$$
so much so, that for the case $\mu<0$ the bottom mass
is hopelessly large. However, these help to reduce $m_b$
when $\mu>0$. Even so, we see in Fig. 6(a) that $m_b > 5.5$ GeV.
GUT threshold corrections are needed to reduce both $\alpha_s$ and
$m_b$ further.
In minimal SU(5) the proton decay rate is enhanced at large $\tan\beta$,
so the triplet Higgs mass $M_{H_3}$ is forced to be quite large.
Hence, we have a large and positive GUT correction $\varepsilon_g'$
(Eq. (\ref{epsprime})) which forces the strong coupling
to unacceptably large values (\roughly{>}0.14).
Hence we conclude that minimal SU(5) is ruled out at large $\tan\beta$.
In the missing doublet model the large triplet Higgs mass
correction is adequately compensated by the constant $-4\%$ correction
of Eq. (\ref{mdmodel}). As shown in Fig. 7, this cancellation results in
near central values for $\alpha_s$ and $m_b$.
However, $\mu>0$ is clearly
required when $\tan\beta$ becomes large, as the large finite corrections
have the wrong sign in the $\mu<0$ case, yielding values of $m_b$
which are unacceptably large.
\section{Conclusion}
\vspace{-.25cm}
In this talk we have presented results from a complete
calculation of the one-loop corrections to the masses, gauge, and
Yukawa couplings in the MSSM. We have seen that such a calculation
allows us to reliably investigate various unified models to
determine whether they are compatible with current experimental data.
In particular, we found that the finite SUSY corrections, which are neglected
in the leading logarithm approximation, can substantially increase the
prediction for $\alpha_s$ and $m_b$ when some of the SUSY partner
masses are lighter than or of order $M_Z$.
For small $\tan\beta$, we found that in the minimal SU(5) model,
$\alpha_s$ was somewhat large ($\alpha_s>0.124$ with $m_{\tilde q}<1$ TeV)
and $m_b$ was larger than 5 GeV. The missing doublet model
gave much lower values of $\alpha_s$ and $m_b$, near their central
values.
For large $\tan\beta$, the minimal SU(5) model GUT threshold correction
became quite large and positive, and this resulted in unacceptably large
values of the strong coupling ($\alpha_s\roughly{>}0.14$).
The missing doublet model,
on the other hand, had no trouble in accomodating the central values of
the strong coupling and $m_b$. The values of $m_b$ were largely
determined by important finite corrections, and we required $\mu>0$ in order
for these corrections to result in acceptably small values of the
bottom-quark mass.
\section{Acknowledgements}
I would like to thank my collaborators Jonathan Bagger and Konstantin
Matchev. This work was supported
by the U.S. National Science Foundation under grant NSF-PHY-9404057.
\section{References}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.