1: \documentclass[12pt,a4]{article}
2: \usepackage{psfig,epsfig,graphicx}
3: \usepackage{amsmath}
4: \usepackage{amssymb}
5: \usepackage{cite}
6:
7: %\renewcommand{\theequation}{\thesection.\arabic{\equation}}
8:
9: \begin{document}
10:
11: \title{Shear Viscosity of Hot QED at Finite Density from Kubo Formula}
12: \author{Liu Hui \thanks{liuhui@iopp.ccnu.edu.cn} \ \ Hou Defu \thanks{hdf@iopp.ccnu.edu.cn}
13: \ \ Li Jiarong \thanks{ljr@iopp.ccnu.edu.cn} \\[0.5cm] {\it\small Institute of Particle Physics, Central China Normal University,}\\ {\it
14: \small Wuhan(430079), P.R.China}}
15: \date{}
16: \maketitle
17: \begin{abstract}
18: Within the framework of finite temperature field theory this
19: paper discusses the shear viscosity of hot QED plasma through Kubo
20: formula at one-loop skeleton diagram level with a finite chemical
21: potential. The effective widths(damping rates) are introduced to
22: regulate the pinch singularities. The finite chemical potential,
23: which enhances the contributions to the shear viscosity from the
24: electrons while suppresses those from the photons, finally gives a
25: positive contribution compared to the pure temperature
26: environment. The result agrees with that from the kinetics theory
27: qualitatively.
28: \end{abstract}
29:
30:
31:
32: \section{Introduction}
33: Transport properties of relativistic plasma are of great interest
34: both experimentally and theoretically. Taking the signal of quark
35: gluon plasm(QGP) for example, the so-called 'strong-coupled
36: matter' formed in the relativistic heavy ion collider(RHIC) was
37: well described by an ideal hydrodynamics model in the region of
38: $p_T<2GeV$ to fit the data of the elliptic flow
39: $v_2$\cite{star,phenix}. But it is pointed out that the
40: over-prediction in high-$p_t$ region of such model is due to the
41: neglect of dissipation and viscosity\cite{Molnar,Teaney}. Thereby
42: the transport coefficients need to be considered. In addition, the
43: non-ideal fluid properties of viscosity and dissipation may
44: influence some other important equations and quantities such as
45: the equation of state(EoS), the formation(thermalization) time and
46: the sound attenuation length
47: etc.\cite{Teaney,Blaizot,Moore,shuryak}.
48:
49: From the theoretical points of view, two formalisms are usually
50: developed to calculate the transport coefficients. One is kinetic
51: theory which starts from expanding the distribution function near
52: the local equilibrium when dealing with the Boltzmann equation.
53: The transport coefficients, especially the shear viscosity, were
54: discussed by many authors in this
55: formalism\cite{Hosoya,Gavin,Danielewicz,Oertzen,Baym,Arnold1,Arnold2}.
56: The other way of obtaining the transport coefficients is to
57: implement the Kubo formulae within the framework of the finite
58: temperature field theory. The Kubo formulae, which work in both
59: weak coupling and strong coupling regime, have been applied for
60: computing the transport coefficients of strong-coupled theory in
61: lattice Monte Carlo simulations\cite{Karsch,Gupta,Nakamura}, and
62: for analytical calculation in scalar theory\cite{jeon1,wang,wang2}
63: and gauge theory\cite{Thoma,defu,Aarts,Aarts2,Basagoiti,defu2},
64: early with relaxation time approximation and later with ladder
65: diagram resummation and large $N_f$ expansion technique. However
66: these two approaches are not irrelevant. Their inner relations are
67: attractive theoretically. Some literatures have shown that the
68: Kubo formulae with ladder resummation are equivalent to the
69: kinetics theory in scalar field and gauge
70: field\cite{jeon1,Jeon,Carrington,Basagoiti}. In principle, one
71: should resum various diagrams with ladder vertices and effective
72: propagators when evaluating the correlation functions in Kubo
73: formulae, which leads to a set of complicated integral equations.
74: In this paper, we develop another approach, distinguished from the
75: ladder resummation and the naive one-loop calculation which gives
76: an inconvincible result, trying to demonstrate that even at
77: one-loop skeleton diagram level the Kubo formula is still
78: consistent qualitatively with kinetics theory if only the proper
79: effective photon and fermion width are chosen as the infrared
80: regulators in which the main physical characters are involved.
81:
82: In heavy-ion physics, the chemical potential in the central fire
83: ball is not zero thought small compared to the extremely high
84: temperature environment. And most of the existing publications
85: concentrated only on the high temperature case but neglected the
86: small chemical potential in the heavy ion collision. To involve
87: the chemical potential effect is our main motivation. It makes
88: sense to introduce a small chemical potential $\mu(\mu\ll T)$ to
89: study that to what extent it can affect the properties of the
90: plasma. The transport coefficients of QGP at finite density has
91: been discussed in Ref. \cite{defu} with the relaxation-time
92: approximation, and early estimated by Danielewicz and Gyulassy to
93: scale the hydrodynamical equations\cite{Danielewicz}. We estimated
94: the shear viscosity of hot QED plasma at finite temperature and
95: chemical potential, referring to the results in the kinetics
96: framework, since the infrared pinch singularities will appear and
97: long distant interaction should be considered in the transport
98: processes.
99:
100:
101: The whole context will be arranged as follows: In section 2, we
102: review the Kubo formula relevant to shear viscosity in the finite
103: temperature theory. The contributions from the photons and the
104: electrons in the QED plasma are studied respectively in section 3
105: and 4, in which one can see the transport damping rates play
106: important roles in regulating the pinch singularities. Conclusions
107: and discussions are presented in section 5.
108:
109: Here are some notations in this paper: (a) the capital letter
110: stands for the four-momentum, $K=(k_0, {\mathbf{k}})$ with
111: $k=|{\mathbf{k}}|$. (b) $\int \frac{d^4 K}{(2\pi)^4}$ is
112: abbreviated to $\int dK$. (c) When we mention 'hard', it means the
113: momentum is much larger than $T$, or at least at the same order;
114: 'soft' means much smaller than $eT$ or at least at the same order.
115:
116:
117: \section{Shear Viscosity and Two-point Retarded Green Function}
118:
119: Generally speaking, the longitudinal expansion domains in heavy
120: ion collision, therefore the shear viscosity is more important
121: than the bulk viscosity which is proved much smaller than the
122: former in lattice calculation\cite{Nakamura}. In this paper, we
123: ignore the bulk viscosity and just take the shear viscous one into
124: account.
125:
126: In a near equilibrium system with linear response theory, Kubo
127: formula tells us the shear viscosity\cite{Carrington} is
128: determined by
129: \begin{equation}
130: \eta=\frac{1}{10}\int d^3x'\int^0_{-\infty}dt'G_R(0;x',t'),
131: \end{equation}
132: where $G_R(0;x',t')$ is the two-point retarded Green function
133: defined as:
134: \begin{equation}
135: G_R(x,t;x',t')=-i\theta(t-t')<[\pi_{ij}(x,t),\pi_{ji}(x',t')]>,\label{rgf}
136: \end{equation}
137: and $\pi_{ij}$ is the spacial part of the dissipative
138: energy-momentum tensor
139: \begin{equation}
140: \pi_{ij}=(\delta_{ik}\delta_{jl}-\frac{1}{3}\delta_{ij}\delta_{kl})T^{kl}=-\eta\left(\partial_{i}u_{j}-\partial_{j}u_{i}+\frac{2}{3}\delta_{ij}\partial_k
141: u^k\right),
142: \end{equation}
143: where $T^{kl}$ is the spacial energy momentum. The coefficients
144: $\eta$ is the shear viscosity. The four-velocity $u^{\mu}(x)$ in
145: the local rest frame is $(1,0,0,0)$.
146:
147: Transforming Eq.(\ref{rgf}) into momentum space, one obtains
148: \begin{equation}
149: \eta=-\frac{i}{10}\frac{d}{dq_0}[\ \lim\limits_{q\rightarrow
150: 0}G_R(Q)\ ]_{q_0=0},\label{v}
151: \end{equation}
152: with $G_R(Q)=\int dX' e^{-iQ\cdot X'}G_R(0;t',x')$ and
153: $X'=(t',x')$.
154:
155: Through the Kubo formula we know that the shear viscosity is
156: determined by the two-point retarded Green function. Now we start
157: to evaluate it in an explicit dynamics. In the relativistic heavy
158: ion collision, people are interested in the extremely high
159: temperature environment, therefore one can use hard thermal
160: loop(HTL) approximation theoretically. We here choose QED as a
161: sample model and calculate the contributions both from the bosons
162: and from the fermions.
163:
164: The QED Lagrangian reads:
165: \begin{equation}
166: \mathcal{L}_{QED}=\bar\Psi(i\gamma^\mu\cdot\partial_\mu-m)\Psi-\frac{1}{4}F^{\mu\nu}F_{\mu\nu}-e\bar\Psi\gamma^\mu\Psi
167: A_\mu,
168: \end{equation}
169: where $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$ and
170: $\Psi, \bar\Psi, A_\mu$ are the electron, anti-electron and photon
171: fields respectively.
172:
173: With this Lagrangian, one can read the viscous stress tensors in
174: the momentum space,
175: \begin{equation}
176: \pi^e_{ij}(k)=i \bar\Psi(\gamma_i
177: k_j-\frac{1}{3}\delta_{ij}{\vec\gamma}\cdot{\mathbf{k}})\Psi,\label{vf}
178: \end{equation}
179: for electrons and
180: \begin{equation}
181: \pi^\gamma_{ij}(k)=-(k_i k_j-\frac{1}{3}\delta_{ij}k^2)A^2\label{vb}
182: \end{equation}
183: for gauge boson with the Lorentz gauge condition.
184:
185: Combining Eqs.(\ref{rgf}), (\ref{v}), (\ref{vf}) and (\ref{vb}),
186: with the standard definitions of propagator in quantum field
187: theory and transport vertex defined behind, one can figure out the
188: one-loop skeleton diagrams of the retarded Green function
189: presented in Fig.1.
190: \begin{figure}
191: %\begin{minipage}[t]{0.5\linewidth}
192: %\centering
193: \begin{center}
194: \resizebox{8cm}{!}{\includegraphics{fig1.eps}}
195: \end{center}
196: \caption{Two point retarded Green function of boson(a) and fermion(b). The dashed line represents the
197: spacial stress tensor operator $\pi_{ij}$, the solid line for fermions and
198: the wiggle line for bosons.}
199: %\end{minipage}%
200: \end{figure}
201: The main difference between the propagators of the normal field
202: and the thermal field rest on the fact that the latter is a
203: $2\times2$ matrix due to the doubled Hilbert
204: space\cite{Carrington2,Kapusta2,LeBellac}.
205:
206: In a high-temperature environment, the loop momenta are hard
207: therefore the mass of fermions can be omitted and bare thermal
208: propagators are sufficient. Yet in the straightforward calculation
209: pinch singularities occur when the external four-momentum
210: approaches to zero, which is required by Eq.(\ref{v}). To avoid
211: such divergences, we must keep the imaginary part of the
212: propagator, namely, the effective damping rate(widths), as
213: infrared regulators, which can be obtained analytically at
214: leading-log order.
215:
216:
217:
218: \section{Boson Contributions}
219: The diagram of the boson two-point Green function is demonstrated
220: in Fig.1(a), with which one can expressed it as
221: \begin{equation}
222: G^{ph}_{ab}(Q)=i\int dK D^{\mu\nu}_{ab}(K)I_{ij}^\gamma \tau_a
223: D_{ba,\mu\nu}(K+Q)I^\gamma_{ji}\tau_b,
224: \end{equation}
225: where $D_{ab}^{\mu\nu}(Q)$ is the thermal gauge boson propagator,
226: and a, b=1, 2 are the doubled Hilbert space indices;
227: $\tau_a,\tau_b=\pm 1$ represent the type-1 and type-2 vertices
228: respectively. The notation $I^\gamma_{ij}=-k_i
229: k_j+\frac{1}{3}\delta_{ij}k^2$ is the transport vertex of the
230: photon viscosity. Summing over the Lorentz indices and using the
231: reverse relation of
232: Keldysh representation\cite{Carrington2,Keldysh} we obtain:
233: \begin{eqnarray}
234: G^{ph}_R(K)&=&G^{ph}_{11}(Q)-G^{ph}_{12}(Q)\label{bs}\\
235: &=&\frac{32i}{3}\int dK
236: k^4[n_B(k_0)-n_B(k_0+q_0)]D_A(K)D_R(K+Q),\nonumber
237: \end{eqnarray}
238: in which $n_B(k_0)=(e^{\beta |k_0|}-1)^{-1}$ is the Bose-Einstein
239: distribution. $D_{R,A}(K)$ is the retarded(advanced) gauge boson
240: propagator in the Feynman gauge with
241: $D^{\mu\nu}_{R,A}(K)=g^{\mu\nu}D_{R,A}(K)$ and
242: \begin{equation}
243: D_{R,A}(K)=\frac{1}{K^2\pm i \varepsilon sgn(k_0)}.
244: \end{equation}
245: In the last line of Eq.(\ref{bs}), the fluctuation-dissipation
246: theorem has been employed:
247: \begin{equation}
248: G_S(P)=[1+2n_B(p_0)]sgn(p_0)[D_R(P)-D_A(P)].
249: \end{equation}
250:
251: Here we encounter the pinch problem of $D_A(K)D_R(K+Q)$ in
252: Eq.(\ref{bs}) when the external momentum $Q$ approaches to zero
253: which is required by the definition of viscosity. The physical
254: reason responsible for this divergence is the naive adoption of
255: bare propagators. Therefore the so-called HTL resummed propagator
256: developed by Pisarski {\it et al.} needs to be considered to
257: involve the medium effect, namely, the scheme of eliminating the
258: pinch is to endow the finite particle life-time or width
259: consistently. Mathematically, we insert a small $\gamma_{ph}$ into
260: the imaginary part of the propagator which physically means the
261: damping rate. The real part of self-energy(thermal mass term),
262: which is in $gT$ order, can be neglected compared with the hard
263: loop momentum. We drop out $Q$ in the propagators directly since
264: Eq.(\ref{v}) requires $q_0=0$ after
265: derivation on $q_0$, and thus the thermal distribution in Eq.(\ref{bs}) has the form
266: \begin{equation}
267: n_B(k_0)-n_B(k_0+q_0)\approx q_0\beta n_B(k_0)[1+n_B(k_0)]+\cdots
268: \end{equation}
269: which preserves the first order in $q_0$. This procedure can be
270: written manifestly as
271: \begin{eqnarray}
272: D_A(K)D_R(K+Q)&&\xrightarrow{Q\rightarrow
273: 0}\frac{1}{(k_0+i\gamma_{ph})^2-k^2}\cdot\frac{1}{(k_0-i\gamma_{ph})^2-k^2}\\
274: &&\longrightarrow
275: \frac{\pi}{4k^2\gamma_{ph}}[\delta(k_0-k)+\delta(k_0+k))].\label{delta}
276: \end{eqnarray}
277: Notice that the photon damping rate is of order $e^4 T$
278: , much smaller than the momentum $K\gtrsim T$, which implies the width of the
279: photon can be substituted by a delta function approximately in
280: expression (\ref{delta}). Integrating off the delta function and the
281: azimuth angles in the spherical coordinates, the retarded Green
282: function becomes
283: \begin{equation}
284: G^{ph}_R(Q)=\frac{8i\beta q_0}{3\pi^2}\int dk k^4
285: n_B(k)[1+n_B(k)]\frac{1}{\gamma_{ph}}.\label{bosongreen}
286: \end{equation}
287:
288: Now we are in the position to calculate the photon damping rate
289: consistently at a finite temperature and density. There is a fact
290: one should notice that the longitudinal contribution is suppressed
291: compared to the transverse one for hard momenta \cite{LeBellac}.
292: To the leading order evaluation, one should only consider the
293: latter. It was calculated by Thoma with vanishing chemical
294: potential \cite{Thoma3}, and we will generalize it to the finite
295: density case.
296:
297: The photon damping in the plasma owes to two processes, Compton
298: scattering and pair production(Fig.2). The two diagrams in the
299: first line of fig.2 stand for Compton scattering and the last line
300: is for pair production processes.
301:
302: \begin{figure}
303: %\begin{minipage}[t]{0.5\linewidth}
304: %\centering
305: \begin{center}
306: \resizebox{10cm}{!}{\includegraphics{fig2.eps}}
307: \end{center}
308: \caption{The physical processes contribute to the photon damping rate. The first line is for the Compton scattering and the last line for
309: the pair production.}
310: %\end{minipage}%
311: \end{figure}
312: The Braaten-Pisarski method\cite{Bratten2} has been used to
313: calculate the photon damping rate in the QED plasma, where the
314: total effect is divided into two parts by a separating scale. The
315: soft contribution is obtained by cutting the photon self-energy
316: diagram with the HTL resummed fermion propagator(Fig.3) and the
317: hard contribution can be evaluated most conveniently from the
318: matrix elements. Adding up the two contributions one can cancel
319: the separating scale.
320:
321:
322: \subsection{Soft Contributions}
323: The soft contribution of a real hard photon damping rate follows
324: from the the imaginary part of transverse self-energy in Fig.3,
325: where the blob denotes the full fermion propagator.
326:
327: \begin{figure}
328: %\begin{minipage}[t]{0.5\linewidth}
329: %\centering
330: \begin{center}
331: \resizebox{10cm}{!}{\includegraphics{fig3.eps}}
332: \end{center}
333: \caption{Hard photon self-energy with the full propagators denoted by a blob.}
334: %\end{minipage}%
335: \end{figure}
336: The damping rate of a photon is defined as
337: \begin{equation}
338: \gamma=-\frac{1}{2p}Im\Pi_T(p_0=p,p),
339: \end{equation}
340: in the case of non-overdamping. The self-energy is given by
341: \begin{equation}
342: \Pi_{\mu\nu}=2ie^2\int dK Tr[S^*(K)\gamma_\mu
343: S(P-K)\gamma_\nu]\label{photonselfenergy}
344: \end{equation}
345: where $S^*(K)$ is the effective fermion propagator, the factor 2
346: accounts for both diagrams in Fig.3. We take $Q\equiv P-K$ and
347: adopt the helicity representation of the fermion propagator,
348: \begin{eqnarray}
349: S^*(K)&=&\frac{1}{D_+(K)}\frac{\gamma_0-\hat{\mathbf{k}}\cdot \vec
350: \gamma}{2}+\frac{1}{D_-(K)}\frac{\gamma_0+\hat{\mathbf{k}}\cdot
351: \vec
352: \gamma}{2}\\
353: S(Q)&=&\frac{1}{d_+(Q)}\frac{\gamma_0-\hat{\mathbf{q}}\cdot \vec
354: \gamma}{2}+\frac{1}{d_-(Q)}\frac{\gamma_0+\hat{\mathbf{q}}\cdot
355: \vec \gamma}{2},
356: \end{eqnarray}
357: where $\hat{\mathbf{k}} =\mathbf{k}/k$ and
358: \begin{eqnarray}
359: D_{\pm}(K)&=&-k_0\pm k+\frac{m^2}{2k}[(1\mp\frac{k_0}{k})\ln\frac{k_0+k}{k_0-k}\pm 2 ] \\
360: d_{\pm}(Q)&=&-q_0\pm q,
361: \end{eqnarray}
362: with $z=k_0/k$. Here the full fermion propagator is the result
363: from HTL resummation. Projecting the self-energy on the transverse
364: direction, one finds
365: \begin{equation}
366: \Pi_T(p_0,p)=\frac{1}{2}(\delta_{ij}-\frac{p_ip_j}{p^2})\Pi_{ij}(p_0,p).
367: \end{equation}
368: Evaluating the trace over $\gamma$ matrices, one can transform the
369: transverse part of photon self-energy\cite{Kapusta} into
370: \begin{equation}
371: \Pi_T(P)=2ie^2\int dK \left[\frac{1}{D_+(K)}\left
372: (\frac{1-V}{d_+(Q)}+\frac{1+V}{d_-(Q)}\right
373: )+\frac{1}{D_-(K)}\left
374: (\frac{1+V}{d_+(Q)}+\frac{1-V}{d_-(Q)}\right)\right ]
375: \end{equation}
376: where $V=(\mathbf{\hat p}\cdot\mathbf{\hat k})(\mathbf{\hat p}
377: \cdot \mathbf{\hat q})$.
378:
379: To sum the frequencies over the fourth momenta with a chemical
380: potential, we first prove an auxilary formula which is
381: useful to sum over the Matusubara frequencies in a typical
382: one loop diagram at finite chemical potential. The zero chemical
383: potential case was first developed by Braaten, Pisarski and
384: Yuan\cite{Braaten}.
385:
386: The typical one loop diagram one encounters in the temperature
387: field theory may take the form of
388: \begin{equation}
389: O(i\omega)=T\sum\limits_n
390: \triangle(i\omega_n+\mu)\triangle(i\omega
391: -i\omega_n-\mu)\label{oneloop}
392: \end{equation}
393: where $i\omega$ is the external line momentum and $i\omega_n=2\pi
394: (n+1)/\beta$ are the fermion Matusubara frequencies. Employing the
395: spectral representation\cite{LeBellac},
396: \begin{equation}
397: \triangle(i\omega_n+\mu)=-\int dk_0
398: \frac{\rho(k_0)}{i\omega_n+\mu-k_0},
399: \end{equation}
400: one recasts Eq.(\ref{oneloop}) into
401: \begin{equation}
402: O(i\omega)=\int dk_0 \int dq_0 \rho(k_0)\rho'(q_0)T\sum\limits_n
403: \frac{1}{i\omega_n-k_0+\mu} \cdot
404: \frac{1}{i\omega-i\omega_n-k_0+\mu}.
405: \end{equation}
406: Summing over the Matusubara frequencies $n$, one obtains
407: \begin{equation}
408: O(i\omega)=-\int dk_0 \int dq_0 \rho(k_0)\rho'(q_0)
409: \frac{1-f_+(k_0)-f_-(q_0)}{i\omega-k_0-q_0}
410: \end{equation}
411: with $f_{\pm}(k_0)=[e^{\beta(k_0\mp\mu)}+1]^{-1}$.
412:
413: Let $i\omega$ be analytically continued to $p_0$ which are not the
414: Matusubara frequencies and the imaginary part of $O(p_0)$ is
415: \begin{eqnarray}
416: \mbox{Im} O(p_0)&=&\frac{1}{2i}\mbox{Disc}
417: O(p_0)=\frac{1}{2i}[O(p_0+i\epsilon)-O(p_0-i\epsilon)] \\
418: &=&\pi(1-e^{\beta p_0})\int dk_0 \int dq_0
419: f_+(k_0)f_-(q_0)\rho(k_0)\rho'(q_0) \delta(p_0-k_0-q_0)).\nonumber
420: \end{eqnarray}
421: When $T\gg K\gg\mu$, $f_+(k_0)f_-(q_0)=f_+(k_0)f_-(p_0-k_0)\sim
422: e^{-\beta p_0}/2$,
423: \begin{equation}
424: \mbox{Im}O(p_0)=\frac{\pi}{2}(1-e^{\beta p_0})e^{-\beta p_0}\int
425: dk_0 \int dq_0 \rho(k_0)\rho(q_0)\delta(p_0-k_0-q_0).\label{imsum}
426: \end{equation}
427:
428: Using the generalized Eq.(\ref{imsum}) and adopting the similar
429: steps in Ref. \cite{Kapusta}, we finally obtain
430: \begin{equation}
431: \gamma_{soft}=\frac{\pi}{4p}\alpha^2\left(T^2+\frac{\mu^2}{\pi^2}\right)\ln\frac{\Lambda^2}{\pi\alpha(T^2+\frac{\mu^2}{\pi^2})}\label{soft}
432: \end{equation}
433: where $\alpha=e^2/4\pi$ is the fine structure constant, and
434: $\Lambda$ is the separation scale, namely, the upper limit of
435: three-momentum integration.
436:
437: \subsection{Hard Contributions} The hard contribution, when the
438: exchanged fermion carries momentum much larger than the separating
439: scale $\Lambda$, can be evaluated through the matrix elements. The
440: damping rate relevant to the Compton scattering process is
441: \begin{eqnarray}
442: \gamma^{comp}_{hard}=&&\frac{1}{4p}\int \frac{d^3k}{(2\pi)^3
443: 2k}n_F(k)\frac{d^3p}{(2\pi)^3 2p}[1+n_B(p')]\int
444: \frac{d^3k'}{(2\pi)^3
445: 2k'}[1-n_F(k')]\nonumber\\
446: &&\times (2\pi)^4\delta^4(P+K-P'-K')4<|\mathcal M|^2>_{comp},
447: \end{eqnarray}
448: where $<|M|>_{comp}$ is the scattering amplitude and
449: $n_F(k)=(e^{\beta |k_0|}+1)^{-1}$ is the Fermi-Dirac distribution.
450: Similarly, the damping rate for pair creation process is
451: \begin{eqnarray}
452: \gamma^{pair}_{hard}=&&\frac{1}{4p}\int \frac{d^3k}{(2\pi)^3
453: 2k}n_B(k)\frac{d^3p}{(2\pi)^3 2p}[1-n_F(p')]\int
454: \frac{d^3k'}{(2\pi)^3
455: 2k'}[1-n_F(k')]\nonumber\\
456: &&\times (2\pi)^4\delta^4(P+K-P'-K')2<|\mathcal M|^2>_{pair}.
457: \end{eqnarray}
458:
459: Using Mandelstam variables $s=(P+K)^2, t=(P-P')^2$ and $u=-s-t$,
460: one finds the matrix elements are given by\cite{Halzen}
461: \begin{eqnarray}
462: <|\mathcal M|^2>_{comp}&=&-2e^4\left
463: (\frac{u}{s}+\frac{s}{u}\right ),\\
464: <|\mathcal M|^2>_{pair}&=&2e^4\left (\frac{u}{t}+\frac{t}{u}\right
465: ).
466: \end{eqnarray}
467:
468: In the hard external lines assumption, that is $P'\gg T$ and
469: $K'\gg T$, the phase space for the outside region is unfavorable,
470: thus the distribution functions can be simplified as $1\pm
471: n_{F,B}\approx 1$. This considerable simplification may bring us
472: much convenience when using the Lorentz invariant phase space
473: factor\cite{Halzen}
474: \begin{eqnarray}
475: dL&=&(2\pi)^4\delta^4(P+K-P'-K')\frac{d^3p'}{(2\pi)^3
476: 2p'}\frac{d^3k'}{(2\pi)^3 2k'}\nonumber \\
477: &=&\frac{dt}{8\pi s}.
478: \end{eqnarray}
479:
480: In our approximation, the density effect, namely the chemical
481: potential, modifies only the Compton scattering process due to the
482: initial fermion distribution, while those in pair production are
483: kept unchanged as in the pure temperature environment. And the
484: result is already given by Ref. \cite{Thoma3}
485: \begin{equation}
486: \gamma^{pair}_{hard}=\frac{\pi}{6}\frac{\alpha^2
487: T^2}{p}(\ln\frac{4pT}{\Lambda^2}-2.1472).\label{pair}
488: \end{equation}
489: In the Compton process, considering the initial fermion
490: distribution function with a small chemical potential, we can
491: obtain the corresponding damping rate
492: \begin{equation}
493: \gamma^{comp}_{hard}=\frac{\pi}{12p}\alpha^2(T^2+\frac{3\mu^2}{\pi^2})(\ln\frac{4pT}{\Lambda^2}-0.0772)-\frac{\alpha^2T^2}{2\pi
494: p}A(\mu)\label{comp}
495: \end{equation}
496: where
497: \begin{equation}
498: A(\mu)=S_2(e^{-\beta\mu})+S_2(e^{\beta\mu}),
499: \end{equation}
500: \begin{equation}
501: S_n(z)=\sum\limits_{k=1}\limits^{\infty}\frac{z^k}{k^n}\ln k, \ \
502: \ \ \ \ n=0,1,2 \cdots.
503: \end{equation}
504: Adding up Eqs.(\ref{pair}) and (\ref{comp}), one obtains the hard
505: contribution,
506: \begin{equation}
507: \gamma_{hard}=\frac{\pi}{4p}\alpha^2\left
508: [(T^2+\frac{\mu^2}{\pi^2})\ln\frac{4pT}{\Lambda^2}-(1.4572+\frac{2A(\mu)}{\pi^2})T^2-0.0078\mu^2
509: \right ].\label{hard}
510: \end{equation}
511:
512: When the soft and the hard contributions are combined together,
513: the separation scale $\Lambda$ is to be cancelled thus yielding
514: \begin{eqnarray}
515: \gamma&=&\gamma_{soft}+\gamma_{hard}\nonumber \\
516: &=&\frac{\pi}{4p}\alpha^2(T^2+\frac{\mu^2}{\pi^2})F(T,\mu)\label{photondamping}
517: \end{eqnarray}
518: with
519: \begin{equation}
520: F(T,\mu)=\ln\frac{\langle p\rangle
521: T}{\pi\alpha\left(T^2+\frac{\mu^2}{\pi^2}\right)}-\frac{1}{T^2+\frac{\mu^2}{\pi^2}}\left
522: [\left(1.4572+\frac{2A(\mu)}{\pi^2}\right)T^2+0.0078\mu^2 \right
523: ].
524: \end{equation}
525: In the above expression of $F(T,\mu)$ we replaced the momentum in
526: the logarithm by its average value $\langle p\rangle$ to simplify
527: the future integration, which is defined as
528: \begin{equation}
529: \langle p\rangle =\frac{\int d^3k \ k \ n_B(k)}{\int d^3k\
530: n_B(k)}=2.701T.
531: \end{equation}
532:
533: Inserting the Eq.(\ref{photondamping}) into (\ref{bosongreen}),
534: completing the three momentum integration and with Eq.(\ref{v}),
535: we finally obtain:
536: \begin{equation}
537: \eta_{ph}=\frac{1.072T^5}{\alpha^2(T^2+\frac{\mu^2}{\pi^2})F(T,\mu)}\approx
538: 84.64\frac{T^3}{e^4\ln e^{-1}}(1-0.101\frac{\mu^2}{T^2}).
539: \end{equation}
540: where the last approximation is obtained with expansion in terms
541: of $\mu^2/T^2$ and keeping solely the leading-log accuracy.
542:
543: \iffalse
544: If we set $\mu=0$, the photon damping rate is
545: \begin{equation}
546: \eta_{ph}(\mu=0)=84.52\frac{T^3}{e^4\ln e^{-1}},
547: \end{equation}
548: which is coincident with the result given by Arnold {\it et
549: al.}\cite{Arnold}, except for a smaller constant factor. The main
550: reason responsible for the difference is that they considered
551: completely leading series of possible $2\rightarrow2$ processes in
552: the collision term, which is equivalent to resum the infinite
553: ladder diagrams in the language of thermal field theory.
554: \fi
555:
556: \section{Fermion Contribution} In this section we are going to
557: calculate the shear viscosity contributed from the electrons. With
558: Eq. (\ref{vf}) and Fig. 1(b), one can write
559: \begin{equation}
560: G^e_{ab}(Q)=i\int dK Tr[S_{ab}(K\!\!\!\!\! \slash\ )I_{ij}^e
561: \tau_b S_{ab}(K\!\!\!\!\! \slash+Q\!\!\!\! \slash\
562: )I_{ji}^e\tau_a],
563: \end{equation}
564: where $I^e_{ij}=\gamma_i
565: k_j-\frac{1}{3}\delta_{ij}\vec\gamma\cdot\mathbf{k}$ and
566: $S_{ab}(K\!\!\!\! \slash)$ is the element of $2\times2$ fermion
567: propagator matrix. Summed over the thermal indices, the retarded
568: Green function then becomes:
569: \begin{eqnarray}\label{fermiongreen}
570: G^e_R(Q)&=&G_{11}(Q)-G_{12}(Q)\\
571: &=&i\int dK Tr[S_{11}(K\!\!\!\!\! \slash\ )I_{ij}^e S_{11}(K\!\!\!\!\! \slash+Q\!\!\!\! \slash\ )I_{ji}^e
572: -S_{12}(K\!\!\!\!\! \slash\ )I_{ij}^e S_{21}(K\!\!\!\!\! \slash+Q\!\!\!\! \slash\ )I_{ji}^e]\nonumber .
573: \end{eqnarray}
574: Using the inverse relation of Keldysh representation and the
575: fluctuation-dissipation theorem, the last line of Eq.
576: (\ref{fermiongreen}) recasts into
577: \begin{equation}
578: i\int dK Tr[K\!\!\!\!\! \slash\ I_{ij}^e(K\!\!\!\!\!
579: \slash+Q\!\!\!\! \slash\ )I_{ji}^e)][\tilde n_F(k_0)-\tilde
580: n_F(k_0+q_0)] \tilde{S_R}(K)\tilde{S_A}(K+Q),\label{dr}
581: \end{equation}
582: where $\tilde n_F(k_0)=[e^{\beta(|k_0|\pm \mu)}+1]^{-1}$ and
583: $S(K\!\!\!\!\! \slash\ )=K\!\!\!\!\! \slash \ \tilde{S}(K)$.
584:
585: We notice that
586: \begin{equation}
587: Tr[K\!\!\!\!\! \slash\ I_{ij}^e(K\!\!\!\!\! \slash+Q\!\!\!\!
588: \slash\ )I_{ji}^e)]=\frac{8}{3}k^4,
589: \end{equation}
590: and
591: \begin{equation}
592: \tilde{S_R}(K)\tilde{S_A}(K+Q)\xrightarrow{Q\rightarrow
593: 0}\frac{1}{K^2+i\varepsilon sgn(k_0)}\cdot\frac{1}{
594: K^2-i\varepsilon sgn(k_0)}. \label{fs}
595: \end{equation}
596: Similarly, considering the fermion damping in the plasma and using
597: the electron damping rate $\gamma_e$ to improve the infrared
598: behavior, one obtains
599: \begin{equation}
600: \tilde{S_R}(K)\tilde{S_A}(K+Q)\xrightarrow{Q\rightarrow
601: 0}\frac{1}{(k_0+i\gamma_e)^2-k^2}\cdot\frac{1}{(k_0-i\gamma_e)^2-k^2}.\label{fs2}
602: \end{equation}
603:
604: \begin{figure}
605: %\begin{minipage}[t]{0.5\linewidth}
606: %\centering
607: \begin{center}
608: \resizebox{6cm}{!}{\includegraphics{fig4.eps}}
609: \end{center}
610: \caption{Hard fermion damping in the QED plasma. The blob on the photon line stands by
611: a resumed propagator.}
612: %\end{minipage}%
613: \end{figure}
614:
615: The electron damping rate, defined by the imaginary part of the
616: electron self-energy on mass shell with full photon
617: propagator(Fig.4), is just one half of the interaction rate
618: $\Gamma$. According to the cutting rules\cite{Weldon,Braaten3},
619: the interaction rate can be described by the photon longitudinal
620: and transverse spectrum functions $\rho_l$ and $\rho_t$ defined in
621: the HTL approximation for space-like frequencies as
622: \begin{equation}
623: \Gamma(E)=\frac{e^2}{2\pi v }\int^\infty_0 dq\ q
624: \int^{vq}_{-vq}\frac{d\omega}{\omega}\left[\rho_l(\omega,q)+(1-\frac{\omega^2}{q^2})\rho_t(\omega,q)\right]
625: \end{equation}
626: with $v\equiv k/E$ and
627: \begin{eqnarray}
628: \rho_l(\omega,q)&=&\frac{3m_\gamma^2\omega}{2q}\left[\left(q^2+3m_\gamma^2-\frac{3m_\gamma^2\omega}{2q}\log\frac{q+\omega}{q-\omega}\right)^2+\left(\frac{3\pi
629: m_\gamma^2\omega}{2q}\right)^2\right]^{-1}\\
630: \rho_t(\omega,q)&=&\frac{3m_\gamma^2\omega(q^2-\omega^2)}{4q^3}\left\{\left[q^2-\omega^2+\frac{3m_\gamma^2\omega}{2q^2}\left(1+\frac{q^2-\omega^2}{2\omega
631: q}
632: \log\frac{q+\omega}{q-\omega}\right)\right]^2\right.\nonumber\\
633: && \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
634: \left.\left(\frac{3m_\gamma^2\omega(q^2-\omega^2)}{4q^3}\right)^2\right\}^{-1}
635: \end{eqnarray}
636: where
637: $m_\gamma^2=\frac{1}{9}e^2T^2(1+\frac{3}{\pi^2}\frac{\mu^2}{T^2})$
638: is the photon thermal mass. Performing integrals of $dq$ and
639: $d\omega$ with quasi-static approximation, i.e. $v\ll 1$, one
640: finds that the longitudinal part is screened by the
641: thermal mass while the transverse part suffers a logarithmic
642: infrared divergency due to the absence of magnetic screening.
643: Introducing an infrared cutoff $\epsilon$ by hand, one can count
644: the power dependency of the coupling constant
645: \begin{equation}\label{fermiondamping1}
646: \Gamma(E)\sim e^2T(\ln \epsilon^{-1}+constant),
647: \end{equation}
648: which makes the viscosity $\eta\sim e^{-2}$, namely, the
649: reciprocal of quadratic coupling constant, but not quartic as
650: kinetic theory suggusts\cite{Hosoya,Arnold1,Liu}. This
651: disagreement hints us to discover the defects in our previous
652: calculation.
653:
654: Returning back to Fig.(1), one would find that only involving
655: effective propagator is not enough, which means the rungs between
656: the two propagators have been omitted. To include such diagrams we
657: must sum up all these ladder diagrams. While that is a rather
658: complex procedure and a set of integral equations is obtained as a
659: result\cite{jeon1,Jeon,Carrington,Basagoiti}. Here we do not
660: employ this scheme, instead we replace the fermion damping rate,
661: namely the infrared regulator, with a more physical transport
662: damping rate, which has been applied even in the discussion of
663: classic system and Abelian or non-Abelian plasma transport
664: coefficients\cite{Danielewicz,Thoma,defu,Aarts2,xuzhe}.
665:
666: For a Rutherford-like scattering in classic plasma, namely, the
667: elastic scattering with the same type and momentum of incoming and
668: outgoing particles, small angle scattering domains in the
669: dynamical cross section. Whereas, as we know, the large angle
670: scattering, which is caused by a series of accumulated multiple
671: small angle scattering in a long distance, is the most effective
672: to the shear viscous process. Thus an additional factor of
673: $1-\cos\theta$ is multiplied in front of the ordinary cross
674: section\cite{lifshitz}, where $\theta$ is the scattering angle. In
675: this procedure the so-called transport cross section is defined.
676: Furthermore, when discussing the transport process in QED or QCD
677: plasma with the Boltzmann equation in kinetics theory, one will
678: find that the collision term is the convolution integral of
679: scattering amplitudes and their correspondent statistical weights,
680: where if the interaction vertex connects the two particles of the
681: same type, the fluctuation part of the statistic distribution
682: functions will contribute an extra small $q^2$\cite{Arnold1},
683: \begin{equation}\label{kinetics factor q}
684: [\chi_{ij}(\mathbf{p+q})-\chi_{ij}(\mathbf{p})]^2=[\mathbf{q}\cdot\nabla\chi_{ij}(\mathbf{p})+\cdots]^2
685: \end{equation}
686: which is equivalent to multiply the cross section by the factor of
687: $1-\cos\theta$, since
688: \begin{equation}
689: 1-\cos\theta=\frac{2q^2}{s}(1-\frac{\omega^2}{q^2}).
690: \end{equation}
691:
692: Weldon investigated the decay rate in thermal field
693: theory\cite{Weldon}, pointing out that the statistical
694: distributions involving both direct and inverse reactions(detailed
695: balance) should be considered simultaneously, which naturally
696: demanded for introducing the transport cross section when
697: calculating the fermion damping.
698:
699: Now we are aware of what has been lost in the straightforward
700: computation: that is $q^2$ in Eq.(\ref{kinetics factor q}), i.e.,
701: the statistical weights do not reflect the detailed balance, which
702: leads to an incorrect result. Therefore the right way to obtain
703: the effective fermion regulator(width) in transport process is to
704: employ the transport interaction rate and yields,
705: \begin{equation}
706: \gamma_e=\frac{1}{2}\Gamma_{trans}=\frac{1}{2}\int d\Gamma
707: (1-\cos\theta)
708: \end{equation}
709: and
710: \begin{equation}
711: \Gamma_{trans}(E)=\frac{e^2}{\pi v s}\int^\infty_0 dq\ q^3
712: \int^{vq}_{-vq}\frac{d\omega}{\omega}(1-\frac{\omega^2}{q^2})\left[\rho_l(\omega,q)+(1-\frac{\omega^2}{q^2})\rho_t(\omega,q)\right].
713: \end{equation}
714:
715: Changing integration variables from $\omega$ to $x=\omega/q$, we
716: can evaluate the integral over $dq$ analytically and find out the
717: logarithmic dependence on the integral limits. This is again the
718: category which can be treated by employing the Braaten-Pisarski
719: method for the consistent leading order quantities, i.e. employing
720: a separating scale $\Lambda$ to divide the integral into a soft
721: contribution and a hard contribution, then adding the two
722: contributions which leads to the cancellation of $\Lambda$. One
723: would notice in the previous calculation on the photon damping
724: rate that the coefficients in front of the log-terms in the soft
725: and hard contributions are identical so as to guarantee the
726: separating scale to be cancelled. Thereby we can play a little
727: trick to obtain the leading-log result by just calculating the
728: soft contributions and replacing the separating scale with the
729: upper $T$, ignoring the constant factor added to the logarithm
730: term which is less important in the weak coupling limit.
731: Completing the integral over $dq$ we obtain\footnote{Since the
732: plasma is in a high temperature but small chemical potential
733: environment, i.e. $T\gg\mu$, the fermion lines in figure 4 should
734: be hard and the Pauli blocking has tiny effect on the final
735: result. Compared with the contribution from Bose-Einstein
736: distribution function of photon, the chemical potential in the
737: Fermi-Dirac distribution can be neglected. As to the density
738: effect in the full photon propagator, the small chemical potential
739: just modifies the constant in the logarithm but preserves the
740: order in $\alpha T$. Qualitatively one can adopt this result.}
741: \begin{equation}
742: \gamma_{e}=\frac{1}{2}\Gamma_{trans}(E)=\frac{3e^2T}{4\pi
743: s}m_\gamma^2\log\frac{T}{m_\gamma}.
744: \end{equation}
745: with $s=(P+K)^2=(2E_p)^2=4p^2$ in central mass system.
746:
747: Combining ({\ref{fs2}}) with (\ref{dr}) and noticing that
748: $\gamma_e\sim e^4T \ll T$, one adopts a delta function as
749: approximation and finds:
750: \begin{equation}
751: G^e_R(Q)=\frac{2i\pi}{3}\int dK \frac{k^2}{\gamma_e}[\tilde
752: n_F(k_0)-\tilde n_F(k_0+q_0)][\delta(k_0-k)+\delta(k_0+k))].
753: \end{equation}
754: Integrating over $dk_0$ and then expanding the $\tilde n_F$ in
755: terms of small $q_0$, ignoring the terms without $q_0$(because Eq.
756: (\ref{v}) needs $q_0=0$ after differentiation on $q_0$), we obtain
757: \begin{equation}
758: G^e_R(Q)=\frac{2i\beta^2 q_0}{3\pi e^2 m^2_\gamma\ln e^{-1}}\int
759: dk k^6\left[\frac{e^{\beta(k+\mu)}}{(e^{\beta\mu}+e^{\beta k})^2}+
760: \frac{e^{\beta(k+\mu)}}{(e^{\beta(k+\mu)}+1)^2}\right].\label{fermionv}
761: \end{equation}
762:
763:
764: Noticing $T\gg\mu$, one can expand the integrand of
765: (\ref{fermionv}) as Taylor series with respect to $\mu/T$ and
766: complete the final integration. The leading-log shear viscous
767: coefficient contributed by the fermions is:
768: \begin{equation}
769: \eta_e=361.4\frac{T^3}{e^4\ln\frac{1}{e}}(1+0.1765\frac{\mu^2}{T^2}).\label{fermionresult}
770: \end{equation}
771:
772: On the contrary, if an interaction vertex connects the two
773: particles of different types, like Compton scattering or pair
774: production we have discussed in the previous section, no kinetics
775: factor $q^2$ will be isolated from the statistical weight, i.e.,
776: $1-\cos\theta$ will not appear in the photon width
777: expression\cite{Thoma,defu,Aarts3}. That is why we got the right
778: dependence of the coupling constant by just adopting the ordinary
779: photon damping rate as the infrared regulator in the photon
780: contribution. \iffalse One thing we should mention is when this
781: method is generated to QCD case, bosons transport damping rate
782: should be taken into account due to the self-coupling of gluon
783: field, which is different with the photon behavior.\fi
784:
785: \section{Conclusions and Discussion}
786: Adding up the contributions from both the electron and photon
787: sectors, the shear viscous coefficient of QED plasma is obtained,
788: \begin{equation}
789: \eta_{QED}=\eta_e+\eta_{ph}=446.0\frac{T^3}{e^4\ln
790: e^{-1}}(1+0.1234\frac{\mu^2}{T^2}).
791: \end{equation}
792: One may find that the density-relevant part of viscosity coming
793: from the photons is negative but the one from electrons is
794: positive. They compete with each other and leave a positive
795: contribution of chemical potential, which demonstrates an
796: enhancement effect of density.
797:
798: The above result is consistent, in the order accuracy of coupling
799: constant, but differing by a factor with those obtained in kinetic
800: theory at the finite density\cite{Liu} and the pure temperature
801: case\cite{Arnold1} when $\mu$ goes to zero. What we mean
802: consistency here is just at the qualitative level. More accurate
803: effect like ladder resummation, in which none damping rate should
804: be replaced by the transport one since it is self-consistently
805: including the main characters of the transport processes, should
806: be involved in the calculation. The one-loop approximation may
807: partly account for the discrepancy of the factors scaled by the
808: logarithms.
809:
810: From the proceeding analysis we can conclude that the estimation
811: of the shear viscosity of QED plasma calculated at one-loop
812: skeleton diagram level by choosing proper fermion width in the
813: framework of the finite temperature field theory, is economical
814: and reliable. Other transport coefficients of the relativistic
815: plasma at finite temperature and density could be discussed
816: similarly. Using this method helps us to handle the pinch
817: singularity easily if the transport hard fermion and boson damping
818: rates are known. \iffalse For non-Abelian gauge QCD plasma, we can
819: follow the similar steps in this paper with QED case to estimate
820: the transport coefficients at one-loop skeleton diagram level by
821: using the transport damping rates for both gluon and quark
822: sectors.\fi Nevertheless, we may point out that some further
823: improvements could be introduced to obtain more precise results
824: such as the Landau-Pomeranchuk-Migdal(LPM) effects due to the
825: interference of multiple scattering process\cite{Berges,Aurenche}
826: for the complete leading-order contribution.\\[1cm]
827:
828: \centerline{\bf Acknowledgement} This work is partly supported by
829: the National Natural Science Foundation of China under project
830: Nos. 90303007, 10135030 and 10575043, the Ministry of Education of
831: China with Project No. CFKSTIP-704035. We thank M.H. Thoma for his
832: valuable comments.
833:
834:
835:
836: \begin{thebibliography}{99}
837: \bibitem{star} STAR collaboration, Phys. Rev. Lett.
838: 90, 032301 (2003)
839: \bibitem{phenix} PHENIX collaboration, Phys. Rev. Lett.
840: 91, 182301 (2003)
841: \bibitem{Molnar} D. Molnar and P. Huovinen, Phys. Rev.
842: Lett. 94, 012302 (2005) and references
843: therein.
844: \bibitem{Teaney} D. Teaney, Phys. Rev. C 68, 034913 (2003)
845: and references therein.
846: \bibitem{Blaizot} J.P. Blaizot and E. Iancu, Phys. Rept. 359, 355 (2002)
847:
848: \bibitem{Moore} G.D.Moore, hep-ph/0412346
849: \bibitem{shuryak} E.Shuryak, hep-ph/0312227 and references
850: therein.
851: \bibitem{Hosoya} R. Hosoya and K. Kajantie, Nucl. Phy. B 250, 666 (1985)
852: \bibitem{Gavin} S. Gavin, Nucl. Phys. A 435, 826 (1985)
853: \bibitem{Danielewicz} P. Danielewicz and M. Gyulassy, Phys. Rev. D
854: 31, 53 (1985)
855: \bibitem{Oertzen} D.W.von Oertzen, Phys. Lett. B 280, 103 (1992)
856:
857: \bibitem{Baym} G. Baym, H. Monien, C.J. Pethick and D.G. Ravenhall, Phy. Rev. Lett. {64}, 1867 (1990);
858: E. H. Heiselberg, Phys. Rev.{ D 49}, 4739 (1994).
859: \bibitem{Arnold1} P. Arnold, G.D. Moore and L.G. Yaffe, JHEP 0011, 001 (2000)
860: \bibitem{Arnold2} P. Arnold, G.D. Moore and L.G. Yaffe, JHEP 0305, 051 (2003)
861: \bibitem{Karsch} F. Karsche and H.W. Wyld, Phys. Rev. D 35, 2518 (1987)
862: \bibitem{Gupta} S. Gupta, Phys. Lett. B 597, 57 (2004)
863: \bibitem{Nakamura} A. Nakamura and S. Sakai, Phys. Rev. Lett.
864: 94, 072305 (2005)
865: \bibitem{jeon1} S. Jeon, Phys. Rev. D 52, 3591 (1995).
866: \bibitem{wang} E. Wang, U. Heinz and X. Zhang, Phys. Rev.
867: D 53,5978 (1996)
868: \bibitem{wang2} E. Wang and U. Heinz, Phys. Lett. B 471, 208
869: (1999)
870: \bibitem{Thoma} M.H. Thoma, Phy. Lett. B 269, 144 (1991)
871: \bibitem{defu} H. Defu and L. Jiarong, Nucl. Phys. A 618,
872: 371 (1997)
873: \bibitem{Aarts} G. Aarts and J.~M.~Mart\'\i nez Resco, JHEP {0211},
874: 022(2002); {\it ibid.} 0402, 061 (2004);
875: \bibitem{Aarts2} G. Aarts and J.~M.~Mart\'\i nez Resco, JHEP 0503, 074 (2005)
876: \bibitem{Aarts3} G. Aarts and J.~M.~Mart\'\i nez Resco,
877: hep-ph/0409090
878: \bibitem{xuzhe} Xu Zhe and C. Greiner, Phys. Rev. C
879: 71,064901 (2005)
880: \bibitem{Basagoiti} M. Basagoiti, Phys. Rev. D 66, 045005 (2002)
881: \bibitem{defu2} H. Defu, hep-ph/0501284.
882: \bibitem{Jeon} S. Jeon and L.G. Yaffe, Phys. Rev. D 53,
883: 5799 (1996)
884: \bibitem{Carrington} M.E. Carrington, H. Defu and R. Kobes, Phys.
885: Rev. D 62, 025010 (2000); {\it ibid}. 64, 025001 (2001)
886: %\bibitem{Thoma4} C.T. Traxler, H. Vija, and M.H. Thoma, Phy. Lett. B 346, 329
887: % (1995)
888: \bibitem{Carrington2} M.E. Carrington, H. Defu, and M.H. Thoma,
889: Eur. Phys. J. C7, 347 (1999)
890: \bibitem{Kapusta2} J.I. Kapusta, {\it Finite Temperature Field
891: Theory} (Cambridge Univ. Press, Cambridge, 1989)
892: \bibitem{LeBellac} M. Le Bellac, {\it Thermal Field Theory}
893: (Cambridge Univ. Press, Cambridge, 1996)
894: \bibitem{Keldysh} L.V. Keldysh, Zh. Eksp. Teor. Fiz. 47, 1515
895: (1964) [Sov. Phys. JETP 20, 1018 (1965)]
896: \bibitem{Thoma3} M.H. Thoma, Phys. Rev. D 51, 862 (1995)
897: \bibitem{Bratten2} E. Braaten, and R.D. Pisarski, Nucl.
898: Phys. B 337, 569 (1990)
899: \bibitem{Kapusta} J. Kapusta, P. Lichard and D. Seibert, Phys.
900: Rev. D 44, 2774 (1991)
901: \bibitem{Braaten} E. Braaten, R.D. Pisarski and T.C. Yuan,
902: Phys. Rev. Lett. 64, 2242 (1990)
903: \bibitem{Halzen} F. Halzen and A.D. Martin, {\it Quarks and
904: Leptons} (Wiley, New York, 1984)
905: \bibitem{Weldon} H.A. Weldon, Phys. Rev. D 28, 2007 (1983)
906: \bibitem{Braaten3} E. Braaten, and M.H. Thoma, Phys. Rev. D
907: 44, 1298 (1991)
908: \bibitem{Liu} Liu Hui, Hou Defu and Li Jiarong, Eur. Phys. J.
909: C, 45, 459 (2006)
910: \bibitem{lifshitz} E.M. Lifshitz and L. P. Pitaevskii, {\it Physical
911: Kinetics} (Pergamon, New York, 1981)
912: %\bibitem{Baier} R. Baier and R. Kobes, Phys. Rev. D 50,
913: % 5944 (1994)
914: \bibitem{Berges} J. Berges, Phys. Rev. D 70, 105010 (2004)
915: \bibitem{Aurenche} P. Aurenche, F. Geils and H. Zaraket,
916: Phys. Rev. D 62, 096012 (2000)
917: \end{thebibliography}
918:
919:
920: \end{document}
921: