nucl-th0604014/VIS.TEX
1: \documentclass[eqsecnum,twocolumn,showpacs,aps]{revtex4}
2: \usepackage{graphicx}
3: \begin{document}
4: \title{Dissipative hydrodynamics in 2+1 dimension}
5: \author{A. K. Chaudhuri}
6: \email[E-mail:]{akc@veccal.ernet.in}
7: \affiliation{Variable Energy Cyclotron Centre, 1/AF, Bidhan Nagar, 
8: Kolkata 700~064, India}
9: 
10: \begin{abstract}
11: In 2+1 dimension, we have simulated the hydrodynamic evolution of 
12: QGP fluid with dissipation due to shear viscosity. Comparison of 
13: evolution of ideal and viscous fluid, both initialised
14: under the same conditions e.g. same equilibration time, energy density
15: and velocity profile, reveal that the dissipative fluid evolves slowly, 
16: cooling at a slower rate. Cooling get still slower for higher viscosity. 
17: The fluid velocities on  the otherhand evolve faster in a dissipative fluid 
18: than in an ideal fluid.  The transverse expansion is also enhanced in dissipative evolution.
19: For the same decoupling temperature, freeze-out surface for a dissipative fluid is more extended
20: than an ideal fluid.
21: Dissipation produces entropy as a result of which particle production is increased. Particle production is
22: increased due to (i) extension of the freeze-out surface and (ii) change of the equilibrium distribution
23: function to a non-equilibrium one, the last effect being prominent at large transverse momentum. Compared
24: to ideal fluid, transverse momentum distribution of pion production is considerably enhanced. Enhancement
25: is more at high $p_T$ than at low $p_T$. Pion production also increases with viscosity, larger the viscosity,
26: more is the pion production. Dissipation also modifies the elliptic flow. Elliptic flow is reduced in viscous
27: dynamics. Also, contrary to ideal dynamics where elliptic flow continues to increase with transverse momentum,
28: in viscous dynamics, elliptic flow tends to saturate at large
29: transverse momentum. The analysis suggest that initial conditions of the 
30: hot, dense matter produced in Au+Au collisions at RHIC, as extracted from
31: ideal fluid analysis can be changed significantly if the QGP fluid is viscous.  
32: \end{abstract}
33: 
34: \pacs{47.75.+f, 25.75.-q, 25.75.Ld} 
35: 
36: \date{\today}  
37: 
38: \maketitle
39: 
40:  
41: \section{Introduction}
42:  
43: Lattice  QCD predicts that under certain conditions (sufficiently
44: high energy density and temperature),  ordinary  hadronic  matter
45: (where  quarks  and  gluons  are  confined),  can undergo a phase
46: transition to a deconfined matter, commonly known as Quark  Gluon
47: Plasma (QGP).  Nuclear physicists are trying to produce and detect
48: this  new  phase  of matter at RHIC, BNL.  Recent Au+Au collisions at RHIC,
49: indicate that dense, color opaque medium of deconfined matter is created in 
50: very central collisions
51: \cite{BRAHMSwhitepaper,PHOBOSwhitepaper,PHENIXwhitepaper,
52: STARwhitepaper}.  It is now also understood that the quarks and gluons strongly interact, giving rise to  the notion of sQGP
53: (strongly interacting QGP).
54: The experimental data are successfully analysed in a {\em ideal} fluid dynamic model \cite{QGP3}. Hydrodynamic evolution
55:  of ideal QGP, thermalised at $\tau_i$=0.6 fm, with central entropy density of
56: 110 $fm^{-3}$, or energy density of 35 $GeV/fm^3$ can explain a large volume of RHIC data, the $p_T$ spectra of identified
57: particles, the elliptic flow etc \cite{QGP3}. However, experimental data do show deviation from ideal behavior.  The ideal
58: fluid description works well in almost central Au+Au collisions near  mid-rapidity at top RHIC energy, but gradually breaks
59: down in more  peripheral collisions, at forward rapidity, or at lower collision energies \cite{Heinz:2004ar}, indicating the
60: onset of dissipative effects. To describe such deviations from ideal fluid dynamics quantitatively,  requires the  numerical
61: implementation of {\em dissipative} relativistic fluid dynamics.
62: 
63: Eckart \cite{Eckart} and Landau and Lifshitz \cite{LL63} formulated the theory  of dissipative relativistic fluid. Their
64: theories are called first order theories and suffer from the problem of causality; the signal can travel faster than light.
65: 1st order theories  assume that the entropy 4-current contains terms upto linear order in dissipative quantities. This
66: restriction results in  parabolic equations, which leads to causality problem.   The causality problem is removed in the
67: 2nd-order theories, formulated
68: by Israel and Stewart \cite{IS79}. In 2nd order theories,  dissipative fluxes are treated as (extended) thermodynamic
69: variables and the linear relation
70: between entropy 4-current and dissipative quantities is extended to include 
71: quadratic terms. The entropy 4-current contains terms upto 2nd order in 
72: dissipative forces. The resulting equations are hyperbolic in nature and the causality problem is removed. Naturally 2nd
73: order theories are more complicated. In addition to the usual energy-momentum conservations equations, relaxation equations
74: for dissipative fluxes are required to be
75: solved simultaneously. 
76:  
77: Though the theories of dissipative hydrodynamics \cite{Eckart,LL63,IS79} 
78: has been known for more than 30 years , significant progress towards its numerical implementation  has only been made very
79: recently \cite{Muronga:2001zk,Teaney:2004qa,MR04,CH05,Heinz:2005bw}.
80: Earlier attempts were restricted to simple one-dimensional Bjorken 
81: expansion \cite{vis_bj}. Since the hot dense matter, produced in RHIC collisions, undergoes large transverse expansion,
82: such models are not
83: of much help in extracting initial conditions of the fluid from the experiment.
84: Recently Teaney \cite{Teaney:2004qa} solved the hydrodynamic equations in a 1st order dissipative theory. He used the blast
85: wave model and calculated the corrections to thermal distribution functions considering shear viscosity only. Blast wave
86: models lack dynamics. Freeze-out surface is parameterised and one fit the freeze-out parameters with experimental observables.
87:  Information about the initial conditions of the hot dense matter is not available in blast wave models. In \cite{MR04,CH05},
88:  2nd
89: order theories for dissipative fluid, is  solved for QGP fluid. While the models
90:  are dynamic, (energy-momentum conservation equations are solved)
91:  energy-momentum conservation equations  are solved in 1+1 dimension,
92:  assuming longitudinal  boost-invariance and cylindrical symmetry. Thus
93:  these models do not give any information on the most sensitive experimental
94:  observable, the elliptic flow.  Recently in \cite{Heinz:2005bw} explicit
95:  equations describing the space-time evolution of non-ideal relativistic
96:  fluid, undergoing boost-invariant  longitudinal and arbitrary transverse
97:  expansion, are given. Equations are written for both the 1st order and
98:  2nd order theory.
99: 
100: At the Variable Energy Cyclotron Centre, Kolkata, we have developed a  numerical code  (AZHYDRO-KOLKATA) to solve, both the 1st
101: order and 2nd order dissipative hydrodynamics in 2+1 dimension.  In the present paper we will present  AZHYDRO-KOLKATA results
102: for  the 1st order dissipative hydrodynamics. Results for 2nd order dissipative hydrodynamics will be presented in a later
103: publication.
104: %%%%%%%%% ADDED AS PER REFEREE REQUEST %%%%%
105: As mentioned earlier, 1st order theories are acausal, signal can travel faster
106: than light.   In one dimension, 1st order theories can be solved analytically. Analytical solutions indicate that at early
107: time, an unphysical reheating of the fluid can occur \cite{MR04,Baier:2006um}. Even though
108: in the present simulation, we donot find any indication of early reheating,
109: nevertheless, we maintain that only 2nd order results,  which do not have the unphysical causality the problem are more
110: reliable.
111: %%%%%%%%%%%%%%%%%%%%%%%%%%
112: In this paper we will be concerned mainly with the effect of viscosity on fluid evolution and its affects the particle
113: production; $p_T$ distribution and elliptic flow. No attempt will be  made to explain experimental data.
114:  
115: 
116: Paper is organised as follows: in section II, we describe briefly the
117: hydrodynamic equations needed for 1st order dissipative hydro-dynamics.
118: In section III, we describe the equation of state, shear viscosity
119: coefficient and initial conditions used in the present study. With dissipation,
120: equilibrium distribution function is changed. Correction to equilibrium
121: distribution function due to non-equilibrium effects are described in 
122: section IV. We have studied particle (pion) production in the model. The relevant equations for pion production with
123: the equilibrium distribution
124: function and its correction due to non-equilibrium effects, are discussed in section V. Results of the numerical
125: simulations are shown in
126: section VI.   Lastly, in section VII, summary and conclusions are drawn.
127: 
128:  
129: \section{1st order dissipative fluid dynamics}
130:  
131: Relativistic dissipative hydrodynamics and associated equations
132: in 2+1 dimensions has been discussed in detail in ref.\cite{Heinz:2005bw}. 
133: Any fluid dynamical approach starts from the
134: conservation laws for the conserved charges and for energy-momentum.
135: For a singly charged fluid, the conservation laws are:
136: %
137: \begin{eqnarray}
138: \label{eq1}
139: \partial_\mu N^\mu &=&0\\
140: \label{eq2}
141: \partial_\mu T^{\mu\nu} &=& 0.
142: \end{eqnarray}
143: %
144: It must also ensure the second law of thermodynamics
145: %
146: \begin{equation}
147: \label{eq3}
148: \partial_\mu S^\mu \geq 0,
149: \end{equation}
150: %
151: where $S^\mu$ is the entropy current. 
152: 
153: In the present paper we restrict ourselves to central rapidity region,
154: where the QGP fluid is essentially baryon free. We thus neglect 
155: Eq.\ref{eq1}. To keep the calculations simple, we consider the most
156: important dissipative term, the shear viscosity
157: and neglect the other dissipative terms, e.g. heat conduction, bulk viscosity.
158: For a baryon free fluid, the heat conduction is zero. Bulk
159: viscosity is zero for QGP (point particles). Bulk viscosity will
160: be non-zero in the hadronic phase but presently we neglect it.
161: 
162: With the help of hydrodynamic 4-velocity $u^\mu$ (normalised as
163: $u^\mu u_\mu=1$) and projector $\Delta^{\mu\nu}=g^{\mu\nu}-u^\mu u^\nu$,
164: with only shear viscosity as the dissipative flux, in Landau's frame, 
165:  the energy momentum tensor, 
166: and the entropy 4-current, can be decomposed as,
167: 
168: \begin{eqnarray}
169: \label{eq4}
170:  &&T^{\mu\nu} = T_{\rm eq}^{\mu\nu} + \delta T^{\mu\nu}
171:            = \varepsilon\,u^\mu u^\nu -p\Delta^{\mu\nu}
172:                + \pi^{\mu\nu} \\
173: \label{eq5}
174: &&S^\mu = S^\mu_{\rm eq} +\delta S^\mu = s\,u^\mu + \Phi^\mu.
175: \end{eqnarray}
176:  
177: \noindent where    
178: $\varepsilon=u_\mu T^{\mu\nu} u_\nu$ is the energy density, $p$ is the local
179: pressure. $\pi^{\mu\nu}$ is the non-ideal part of the energy-momentum
180: tensor, the    stress tensor due to shear viscosity.   
181: $s=u_\mu S^\mu$ is the entropy density and $\Phi^\mu$ is the entropy
182: flux due to dissipation.
183:  
184: In the first order theories, the shear stress tensors are written as,
185:  
186: \begin{equation}
187: \label{eq6}
188: \pi^{\mu\nu} = 2\eta  \nabla^{<\mu}u^{\nu>}
189: \end{equation}
190: 
191: \noindent where $\eta$ is the shear viscosity coefficient and 
192: $\nabla^{<\mu}u^{\nu>}$ is a trace-less symmetric tensor defined as,
193: 
194: \begin{equation} \label{7}
195:  \nabla^{<\mu}u^{\nu>}=
196:   \left[\frac{1}{2}\left(\Delta^{\mu\sigma}
197: \Delta^{\nu\tau}{+}\Delta^{\nu\sigma}\Delta^{\mu\tau}\right)-\frac{1}{3}
198: \Delta^{\mu\nu}\Delta^{\sigma\tau}\right] 
199: \end{equation}
200: 
201: Viscous pressure $\pi^{\mu\nu}$  is symmetric ($\pi^{\mu\nu}=\pi^{\nu\mu}$),
202: traceless ($\pi^\mu_\mu=0$) and transverse to hydrodynamic velocity,
203: ($u_\mu \pi^{\mu\nu}=0)$. The 16-component $\pi^{\mu\nu}$ has only 5
204: independent components. As mentioned in the beginning, we have solved the
205: equations with the assumption of longitudinal boost-invariance. With boost-invariance
206: the number of independent shear-stress tensors further reduces to 3. 
207: 
208: Heavy ion collisions are best described in terms of proper time $\tau=\sqrt{t^2-z^2}$ and
209: rapidity $\eta_s=\frac{1}{2}\ln \frac{t+z}{t-z}$ (we use the subscript $s$ to distinguish spatial rapidity from the
210: viscous coefficient $\eta$). 
211: In $(\tau,x,y,\eta_s)$ coordinates, with longitudinal boost-invariance,
212: the hydrodynamic 4-velocity can be written as, 
213: 
214: \begin{eqnarray}
215: u^\mu=&&(u^\tau,u^x,u^y,u^{\eta_s}) \nonumber\\
216:           =&&(\gamma_\perp,\gamma_\perp v_x, \gamma_\perp v_y,0),
217: \end{eqnarray}
218: 
219: 
220: %$\gamma_\perp(1,v_x,v_y,0)$, with
221: \noindent with $\gamma_\perp=1/\sqrt{1-v_x^2-v_y^2}$.
222: Explicit equations for energy-momentum conservation,
223: in $(\tau,x,y,\eta_s)$ coordinate system has been developed in ref.\cite{Heinz:2005bw}. 
224: Here we rewrite the results in a form suitable for numerical algorithm.
225: The 
226: energy-momentum conservation equations are,
227: 
228: \begin{widetext}
229: \label{eq8}
230: \begin{eqnarray}
231: \label{eq8a}
232: %&&\frac{1}{\tau}\partial_\tau\Bigl(\tau T^{\tau\tau}\Bigr) 
233: % +\partial_x T^{\tau x}+\partial_y T^{\tau y} = 
234: %  -\,\frac{p+\tau^2 \pi^{\eta\eta}}{\tau},
235: %\\
236: && \partial_\tau\Bigl(\tilde{T}^{\tau\tau}\Bigr) 
237:  +\partial_x \Bigl(\tilde{T}^{\tau \tau} \overline{v}_x \Bigr)+\partial_y \Bigl(\tilde{T}^{\tau \tau} \overline{v}_y \Bigr)= 
238:   -\,(p+\tau^2 \pi^{\eta\eta}),
239: \\
240: \label{eq8b}
241: %&&\frac{1}{\tau}\partial_\tau\Bigl(\tau T^{\tau x}\Bigr) 
242: % +\partial_x \Bigl((T^{\tau x}{-}\pi^{\tau x})v_x\Bigr)
243: % +\partial_y \Bigl((T^{\tau x}{-}\pi^{\tau x})v_y\Bigr) 
244: % = -\partial_x(p{+}\pi^{xx}) - \partial_y\pi^{xy},
245: %\\
246: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
247: &&\partial_\tau\Bigl(\tilde{T}^{\tau x}\Bigr) 
248:  +\partial_x \Bigl(\tilde{T}^{\tau x}v_x\Bigr)
249:  +\partial_y \Bigl(\tilde{T}^{\tau x}v_y\Bigr) 
250:  = -\partial_x(\tilde{p} + \tilde{\pi}^{xx}-\tilde{\pi}^{\tau x} v_x) - \partial_y(\tilde{\pi}^{xy}-\tilde{\pi}^{\tau x}v_y),
251: \\
252: \label{eq8c}
253: %&&\frac{1}{\tau}\partial_\tau\Bigl(\tau T^{\tau y}\Bigr) 
254: % +\partial_x \Bigl((T^{\tau y}{-}\pi^{\tau y})v_x\Bigr)
255: % +\partial_y \Bigl((T^{\tau y}{-}\pi^{\tau y})v_y\Bigr) 
256: % = - \partial_x\pi^{xy}-\partial_y(p{+}\pi^{yy}).\\
257: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
258: &&\partial_\tau\Bigl(\tilde{T}^{\tau y}\Bigr) 
259:  +\partial_x \Bigl(\tilde{T}^{\tau y}v_x\Bigr)
260:  +\partial_y \Bigl(\tilde{T}^{\tau y}v_y\Bigr) 
261:  = -\partial_x(\tilde{\pi}^{xy}-\tilde{\pi}^{\tau y} v_x) - \partial_y(\tilde{p} + \tilde{\pi}^{yy}-\tilde{\pi}^{\tau y}v_y),
262: \end{eqnarray}
263: 
264:  
265: \noindent where $\overline{v}_x=T^{\tau x}/T^{\tau\tau}$ and $\overline{v}_y=T^{\tau y}/T^{\tau\tau}$, and we have used
266: the notation
267: "tilde" to represent quantities multiplied by the factor $\tau$, 
268: $\tilde{p}=\tau p$ and similarly   $\tilde{T}^{ij}=\tau T^{ij}$. 
269: We note that unlike in ideal fluid, in viscous fluid dynamics 
270: conservation equations contain additional pressure gradients containing
271: the dissipative fluxes.  
272: Both $T^{\tau x}$ and $T^{\tau y}$ components of energy-momentum
273: tensors now evolve under
274: the influence of additional pressure gradients.  
275: 
276: In 1st order theory, the shear stress tensor components required  in the 
277: preceding equations are,
278: 
279: \begin{eqnarray} \label{9}
280: &&\pi^{\tau x} =2\eta[ - \frac{1}{2}\partial_x\gamma_\perp 
281:                     + \frac{1}{2}\partial_\tau(\gamma_\perp v_x)
282:                     - \frac{1}{2} D(\gamma_\perp^2 v_x)
283:                     + \frac{\theta}{3}\,\gamma_\perp^2v_x],
284: \\
285: &&\pi^{\tau x} =2\eta[ - \frac{1}{2}\partial_y\gamma_\perp 
286:                     + \frac{1}{2}\partial_\tau(\gamma_\perp v_y)
287:                     - \frac{1}{2} D(\gamma_\perp^2 v_y)
288:                     + \frac{\theta}{3}\,\gamma_\perp^2v_y],\\
289: &&\pi^{\tau\tau} =2\eta[ \frac{\theta}{3}(\gamma_\perp^2-1) 
290:   + \partial_\tau\gamma_\perp-\frac{1}{2}D(\gamma_\perp^2)],
291: \\
292: &&\pi^{\eta\eta} =2\eta[ \frac{1}{\tau^2}
293:   \left(\frac{\theta}{3}-\frac{\gamma_\perp}{\tau}\right)],
294: \\
295: &&\pi^{xx}=2\eta[-\partial_x(\gamma_\perp v_x)-\frac{1}{2}D(\gamma_\perp^2)+
296:  \frac{\theta}{3}(1+\gamma_\perp^2 v_x^2)], \\
297: &&\pi^{yy}=2\eta[-\partial_x(\gamma_\perp v_y)-\frac{1}{2}D(\gamma_\perp^2)+
298:  \frac{\theta}{3}(1+\gamma_\perp^2 v_y^2) ]
299: \end{eqnarray}
300: \end{widetext}
301: 
302: \noindent where $D=u^\mu \partial_\mu$ is the convective time derivative,
303: 
304: \begin{equation}
305: D=\gamma_\perp(\partial_\tau + v_x \partial_x + v_y \partial_y),
306: \end{equation}
307: 
308: \noindent and $\theta$ is the  local expansion rate, given by,
309: 
310: \begin{equation}
311: \theta=\frac{\gamma_\perp}{\tau}+\partial_\tau \gamma_\perp +
312: \partial_x(v_x\gamma_\perp)+\partial_y(v_y \gamma_\perp)
313: \end{equation}
314:  
315: %As noted earlier, of 7, only three stress-tensors are independent. 
316: %It is very important
317: %in 2nd order theories where, for each dissipative flux, one solves a %relaxation simultaneously.  
318: 
319: Given an equation of state, if
320: energy density ($\varepsilon$) and fluid velocity ($v_x$ and $v_y$) distributions, at any time
321: $\tau_i$ are known, Eqs.\ref{eq8a},\ref{eq8b} and \ref{eq8c} can be  integrated
322: to obtain $\varepsilon$,$v_x$ and $v_y$ at the next time step $\tau_{i+1}$. 
323: While for ideal hydrodynamics, this procedure works perfectly,   
324:   viscous hydrodynamics poses a problem
325: that shear stress-tensor components contains 
326: time derivatives, $\partial_\tau \gamma_\perp$, $\partial_\tau u^x$,
327: $\partial_\tau u^x$ etc. Thus at time step $\tau_i$ one needs the  
328: still unknown time derivatives. 
329: In 1st order theories, this problem is circumvented by calculating the
330: time derivatives from the ideal equation of motion,
331: 
332: \begin{eqnarray}
333: \label{eq220}
334: Du^\mu &=&\frac{\nabla^\mu p}{\varepsilon+p},\\
335: \label{eq221}
336: D\varepsilon&=&-(\varepsilon+p)\nabla_\mu u^\mu.
337: \end{eqnarray}
338: 
339: With the help of these two equations all the time derivatives can be 
340: expressed entirely in terms of spatial gradients. 
341: 1st order theories are restricted to contain terms at most linear
342: in dissipative quantities.  Neglect of viscous terms can contribute only
343: in 2nd order corrections, which is neglected in 1st order theories.
344: 
345: \section{Equation of state, viscosity coefficient and initial conditions}
346: 
347: \subsection{Equation of state}
348: 
349: One of the most important inputs of a hydrodynamic model is  
350: the equation of state. Through this input, the macroscopic hydrodynamic models make contact 
351: with the microscopic world. 
352: In the present calculation we have used the equation of state,
353: EOS-Q, developed in ref.\cite{QGP3}.  It is a two-phase equation of
354: state. The hadronic phase of EOS-Q is modeled
355: as a non-interacting gas of hadronic resonance. As the temperature is increased,
356: larger and larger fraction of available energy goes into production of heavier and heavier resonances. This results into a
357: soft equation of state,
358: with small speed of sound, $c^2_s \approx 0.15$. With increasing temperature, the available volume is filled up with
359: resonances and the
360: hadronic states starts to overlap, and microscopic degrees of freedom
361: are changed from   hadron to deconfined quarks and gluons. The
362: QGP phase  is modeled as that of a 
363: non-interacting quark (u,d and s) and gluons, confined by a bag pressure
364: B. Corresponding equation of state, $p=\frac{1}{3}e -\frac{4}{3}B$ is 
365: stiff with a speed of sound $c_s^2 =\frac{1}{3}$. The two phases 
366: are matched by Maxwell construction at the critical temperature, $T_c=164 MeV$, adjusting the 
367: Bag pressure $B^{1/4}$=230 MeV. As discussed in \cite{QGP3}, ideal
368: hydrodynamics explain a large volume of RHIC Au+Au 
369: data with EOS-Q.
370: 
371: \subsection{Shear viscosity coefficient}
372: 
373: Shear viscosity coefficient ($\eta$) of dense nuclear (QGP or 
374: resonance hadron gas) is quite uncertain.
375: In perturbative regime, shear viscosity of a QGP is
376: estimated \cite{Arnold:2000dr,Baym:1990uj},
377: 
378: \begin{equation}
379: \eta=86.473 \frac{1}{g^4}\frac{T^3}{log g^{-1}},
380: \end{equation}
381: 
382: With entropy of QGP, $s=37\frac{\pi^2}{15}T^3$ and 
383: $\alpha_s \approx$0.5, the ratio of viscosity over the entropy, in the perturbative regime is estimated as,
384: 
385: \begin{equation}
386: \left (\frac{\eta}{s} \right )_{pert} \approx 0.135,
387: \end{equation}
388: 
389: However, QGP  produced in nuclear collisions is non-perturbative. It is
390:  strongly interacting QGP. Recently, using the ADS/CFT correspondence
391: \cite{Policastro:2001yc,Policastro:2002se},
392: shear viscosity of a strongly coupled gauze theory, N=4 SUSY YM,
393: has been evaluated, $\eta=\frac{\pi}{8}N^2_cT^3$ and the    entropy
394: is given by $s=\frac{\pi^2}{2}N^2_cT^3$. Thus in the strongly coupled field theory,
395: 
396: \begin{equation}
397: \left ( \frac{\eta}{s} \right )_{ADS/CFT} = \frac{1}{4\pi}\approx0.08,
398: \end{equation}
399: 
400: \noindent  
401: which  is 2 times larger than the perturbative estimate. In the
402: present paper, we treat the shear viscosity as a parameter of the
403: model. 
404: To demonstrate the effect of viscosity on flow and subsequent particle
405:  production, we use both the perturbative and ADS/CFT estimate of viscosity.
406: 
407: 
408:  
409: 1st order theories are acuasal. As mentioned earlier, unphysical effects
410: like reheating of the fluid, early in the evolution, can occur. In one dimension
411: energy-momentum conservation equation can be solved analytically.
412: If initial fluid temperature is $T_i$ at initial time $\tau_i$, for constant 
413: $\eta/s$, the fluid temperature at time $\tau$ can be obtained as \cite{Baier:2006um},
414: 
415: \begin{equation}
416: T(\tau)=T_i \left (\frac{\tau_i}{\tau} \right )^{1/3} \left [1+
417: \frac{2}{3 \tau_i T_i} \frac{\eta}{s} \left(1- \left(\frac{\tau_i}{\tau}\right)^{2/3}\right)
418: \right ]
419: \end{equation}
420: 
421: For early times, $\tau < \tau_{max}$,
422: 
423: \begin{equation}
424: \tau_{max}=\tau_i\left (\frac{1}{3}+\frac{s}{\eta}
425: \frac{\tau_i T_i}{2}\right)^{-3/2},
426: \end{equation}
427: 
428: \noindent the solution shows an unphysical reheating. The unphysical reheating is minimised if $\tau_{max}$ is small or
429: $\eta/s << \tau_i T_i$.
430: As will be explained below, we have used an initial time $\tau_i$=0.6 
431: fm/c and initial temperature of the fluid, $T_i$=0.35 GeV . For both the values of viscosity, $\eta/s << \tau_i T_i$, the
432: unphysical reheating is minimised. 
433: 
434: Shear viscosity can also be expressed in terms of sound attenuation
435: length, $\Gamma_s$, defined as,
436: 
437: \begin{equation}
438: \Gamma_s=\frac{\frac{4\eta}{3}}{sT}
439: \end{equation}
440: 
441: $\Gamma_s$ is equivalent to mean free path and for a 
442: valid hydrodynamic description $\Gamma_s/\tau << 1$, i.e. mean
443: free path is much less than the system size.  With the present choice of
444: equilibration time and temperature, both for ADS/CFT and perurbative estimate of viscosity,
445: at initial time, $\Gamma_s/\tau$ is much less
446: than unity and hydrodynamics remains a valid description.
447: At later time the validity
448: condition becomes even better. 
449: 
450: \subsection{Initial conditions}
451: 
452: As discussed earlier, ideal hydrodynamics has been very successful
453: in explain a large volume of data in RHIC 200AGeV Au+Au collisions 
454:  \cite{QGP3}. 
455: In the present demonstrative calculations,  we have used the 
456: similar initial conditions as in ref.\cite{QGP3}. 
457: Details of the initial conditions can be found
458: in \cite{QGP3}. We just mention that
459: in ref.\cite{QGP3}, initial
460: transverse energy is parameterised geometrically.
461: At an impact parameter $\vec{b}$, transverse distribution of
462: wounded nucleons $N_{WN}(x,y,\vec{b})$    and of binary NN collisions 
463: $N_{BC}(x,y,\vec{b})$ to are calculated in a Glauber model. 
464: A collision at impact parameter $\vec{b}$ is assumed to contain 25\% hard scattering
465: (proportional to number of binary collisions) and 75\% soft scattering
466: (proportional to number of wounded nucleons).  
467: Transverse energy density 
468: profile at impact parameter $\vec{b}$ is then obtained as,
469: 
470: \begin{equation}
471: \varepsilon(x,y,\vec{b})=\varepsilon_0(0.75\times N_{WN}(x,y,\vec{b})+0.25 \times N_{BC}(x,y,\vec{b}))
472: \end{equation}
473: 
474: The parameter $\varepsilon_0$ and the initial equilibration time
475: $\tau_i$ are fixed to reproduce the 
476: experimental transverse momentum distribution of pions in central Au+Au collisions. 
477: STAR and PHENIX data are fitted to obtain initial
478: equilibrium time $\tau_i$=0.6 fm and central  entropy density of $s=110 fm^{-3}$. 
479: This corresponds energy density of the
480: fluid as 25 $GeV/fm^3$, or 
481: initial temperature of 350 MeV.  Apart from initial energy density, 
482: initial velocity distribution is also required in hydrodynamic
483: calculation. In the present
484: calculation it is assumed that the at the initial time $\tau_i$, fluid
485: velocities are zero, $v_x(x,y)=v_y(x,y)=0$.
486:  
487: In dissipative hydrodynamics, additionally, initial conditions 
488: for the dissipative fluxes needs to be specified. In the present paper 
489: we assume that by the equilibration time $\tau_i$, the dissipative fluxes
490: attained their longitudinal  boost-invariant values. 
491:  
492: \begin{eqnarray}
493: \pi^{\tau x} =&&0\\
494: \pi^{\tau x} = &&0\\
495: \pi^{\tau\tau} =&&0\\
496: \tau^2 \pi^{\eta\eta} =&&-4\eta/\tau_i \\ 
497: \pi^{xx}=&&2\eta/\tau_i\\ 
498: \pi^{yy}=&&2\eta/\tau_i 
499: \end{eqnarray}
500: 
501: \section{Non-equilibrium distribution function}
502: 
503: With dissipation the system is not in equilibrium and the equilibrium 
504: distribution function,
505: 
506: \begin{equation}
507: f^{(0)}(x,p)=\frac{1}{exp[\beta(u_\mu p^\mu -\mu)] \pm 1},
508: \end{equation} 
509: 
510: \noindent with inverse temperature $\beta=1/T$ and chemical potential
511: $\mu$
512:  can no longer describe the system. In a highly
513: non-equilibrium system,   distribution
514: function is unknown. If the system is slightly off-equilibrium,  then
515: it is possible to calculate correction to equilibrium distribution 
516: function due to   (small) non-equilibrium effects. Slightly
517: off-equilibrium distribution function can be  approximated  as, 
518: 
519: \begin{equation}
520: F(x,p)=f^{(0)}(x,p) [1+\phi(x,p)],
521: \end{equation}
522: 
523: \noindent  $\phi(x,p)$ is the deviation from equilibrium distribution 
524: function  $f^{(0)}$. With shear viscosity as the only dissipative forces,
525: $\phi(x,p)$ can be locally approximated by a quadratic function 
526: of 4-momentum,
527: 
528: \begin{equation}
529: \phi(x,p)=\varepsilon_{\mu\nu} p^\mu p^\nu.
530: \end{equation}
531: 
532: Without any loss of generality $\varepsilon_{\mu\nu}$ can be written as
533: as,
534: 
535: \begin{equation} \label{eq4_4}
536: \varepsilon_{\mu\nu}=C \pi^{\mu\nu}, C=\frac{\beta^2}{2(\varepsilon+p)},
537: \end{equation}
538: 
539: \noindent completely specifying the non-equilibrium distribution function.
540: 
541: 
542: \section{particle spectra}
543: 
544: With the non-equilibrium distribution function thus specified, it can be used to 
545: calculate the particle spectra from the freeze-out surface. In the standard 
546: Cooper-Frye
547: prescription, particle distribution is obtained as,
548: 
549: \begin{equation} \label{eq5_1}
550: E\frac{dN}{d^3p}=\frac{dN}{dyd^2p_T} =\int_\Sigma d\Sigma_\mu p^\mu f(x,p)
551: \end{equation}
552: 
553: In $(\tau,x,y,\eta_s)$ coordinate, the freeze-out surface is parameterised as,
554: 
555: \begin{equation}
556: \Sigma^\mu=(\tau_f(x,y)\cosh \eta_s, x, y, \tau_f(x,y) \sinh \eta_s),
557: \end{equation}
558: 
559: \noindent and the normal vector on the hyper surface is,
560: 
561: \begin{equation}
562: d\Sigma_\mu=(\cosh \eta_s, -\frac{\partial \tau_f}{\partial x_f}, 
563:                         -\frac{\partial \tau_f}{\partial y_f}, -\sinh \eta_s)
564: \tau_f dx dy d\eta_s
565: \end{equation}
566: 
567: At the fluid position $(\tau,x,y,\eta_s)$ the particle 4-momenta are parameterised as,
568: 
569: \begin{equation}
570: p^\mu=(m_T cosh (\eta_s-Y), p^x, p^y, m_T sinh (\eta_s-Y))
571: \end{equation}
572: 
573: The volume element $p^\mu d\Sigma_\mu$ become,
574: 
575: \begin{equation}
576: p^\mu d\Sigma_\mu=(m_T cosh(\eta-Y)-\vec{p}_T. \vec{\nabla}_T \tau_f) \tau_f dx dy d\eta
577: \end{equation}
578: 
579: Equilibrium distribution function involve the term $\frac{p^\mu u_\mu}{T}$ which can be evaluated as,
580: 
581: \begin{equation}
582: \frac{p^\mu u_\mu}{T}=\frac{\gamma(m_T cosh(\eta-Y) -\vec{v}_T.\vec{p}_T -\mu/\gamma)}{T}
583: \end{equation}
584: 
585: The non-equilibrium distribution function require the sum
586: $p^\mu p^\nu \pi_{\mu\nu}$,
587: 
588: \begin{equation}
589: p_\mu p_\nu \pi^{\mu\nu}=a_1 cosh^2 (\eta-Y) +a_2 cosh(\eta-Y) + a_3
590: \end{equation}
591: 
592: with
593: \begin{eqnarray}
594: a_1=&&m_T^2(\pi^{\tau \tau} +\tau^2 \pi^{\eta \eta})\\
595: a_2=&&-2m_T(p_x \pi^{\tau x} + p_y \pi^{\tau y})\\
596: a_3=&&p_x^2 \pi^{xx} +p_y^2 \pi^{yy} +2p_x p_y \pi^{xy} - m_T^2 \tau^2 \pi^{\eta \eta}
597: \end{eqnarray}
598: 
599: Inserting all the relevant formulas in Eq.\ref{eq5_1} and integrating over 
600: spatial rapidity one obtains,
601:  
602: \begin{equation}
603: \frac{dN}{dyd^2p_T} =\frac{dN^{eq}}{dyd^2p_T}+\frac{dN^{neq}}{dyd^2p_T}
604: \end{equation}
605:  
606: 
607: with,
608: 
609: \begin{widetext}
610: \begin{eqnarray} \label{eq5_12}
611: \frac{dN^{eq}}{dyd^2p_T}=\frac{g}{(2\pi)^3} 
612: \int dx dy \tau_f [m_T K_1(n\beta) - p_T \vec{\nabla}_T \tau_f K_0(n\beta)]\\
613: \frac{dN^{neq}}{dyd^2p_T}=\frac{g}{(2\pi)^3} 
614: \int dx dy \tau_f [m_T\{ {\frac{a_1}{4} K_3(n\beta)+\frac{a_2}{2} K_2(n\beta)+(\frac{3a_1}{4}+a_3+1)K_1(n\beta)
615: +\frac{a_2}{2} K_0(n\beta)} \} \nonumber\\
616: - \vec{p}_T. \vec{\nabla}_T \tau_f \{\frac{a_1}{2} K_2(n\beta)+
617:  a_2 K_1(n\beta)+( \frac{a_1}{2}+a_3+1)K_0(n\beta)\}]
618: \end{eqnarray}
619: \end{widetext}
620: 
621: \noindent where  $K_0$, $K_1$, $K_2$ and $K_3$ are the modified Bessel functions.  
622: 
623: We will also show results for elliptic flow $v_2$. It is defined as,
624: 
625: \begin{equation}
626: V_2=\frac
627: {\int_0^{2\pi} \frac{dN}{dyd^2p_T} \cos(2\phi) d\phi}
628: {\int_0^{2\pi} \frac{dN}{dyd^2p_T}  d\phi}
629: \end{equation}
630: 
631: \begin{figure}[h]
632: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig1.eps}
633: \vspace{-4cm}
634: \caption{(color online). Contour plots of energy density at  (proper) time $\tau$=5.6 fm.
635: The black  lines are for ideal fluid ($\eta/s$=0). The red and blue
636: lines are for viscous fluid with ADS/CFT ($\eta/s$=0.08) and perturbative ($\eta/s$=0.135) estimate of viscosity.}
637: \label{F1}
638: \end{figure}
639: 
640: \section{Results}
641: 
642: \subsection{Evolution of the viscous fluid}
643: 
644: The energy-momentum conservation equations \ref{eq8a},\ref{eq8b},\ref{eq8c} are solved using the SHASTA-FCT algorithm. We
645: have made extensive changes to the publicly available
646: code AZHYDRO (described in \cite{QGP3}) for simulation of ideal fluid.
647: The modified code, called "AZHYDRO-KOLKATA",  simulate the evolution of dissipative fluid, both in 1st and 2nd order theory.
648:   In this section we present
649: results of AZHYDRO-KOLKATA for the 1st order theory of dissipative fluid.
650: In the following we will show the results obtain in Au+Au  collision at impact parameter $b=$ 6.8 fm, which approximately
651: corresponds
652: to 16-24\% centrality Au+Au collisions.  With the same initial conditions,
653: we have solved the energy-momentum conservation equations for
654: ideal fluid and viscous fluid.  
655: 
656:  In Fig.1, we have shown
657: the constant energy density contour plot in x-y plane, after an evolution of 
658: 5 fm.
659: The black lines are for ideal fluid evolution. The  red and blue lines are for
660: viscous fluid with ADS/CFT ($\eta/s$=0.08) and perturbative ($\eta/s$=0.135) estimate of viscosity. Constant energy density
661: contours,
662: as depicted in Fig.1, indicate that with viscosity fluid cools slowly. Cooling
663: gets slower as viscosity increases.
664: Thus at any point in the x-y plane, viscous fluid temperature is higher
665: than that of the ideal fluid.  The results are in accordance with our
666: expectation. For dissipative fluid, ideal equation of motion Eq.\ref{eq221} is changed to,
667: 
668: \begin{equation}
669: \label{eq61a}
670: D\varepsilon=-(\varepsilon+p)\nabla_\mu u^\mu + \pi^{\mu\nu} \nabla_{<\mu} u_{\nu>}\\
671: %\label{eq61b}
672: %&&(\varepsilon+p)Du^\mu =\nabla^\mu p 
673: %- \Delta^{\mu\nu} \nabla^\sigma \pi_{\nu\sigma}
674: %+\pi^{\mu\nu} Du_\nu
675: \end{equation}
676: 
677: Due to viscosity, evolution of energy density is slowed down.   
678: 
679: \begin{figure}%[h]
680: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig2.eps}
681: \vspace{-4cm}
682: \caption{(color online). Contour plots of x-component of fluid velocity $v_x$ at 
683: $\tau$=5.6 fm. The black  lines are for ideal fluid ($\eta/s$=0). The 
684: red and blue lines are for viscous fluid with ADS/CFT and perturbative estimate of viscosity, $\eta/s$=0.08 and 0.135.}
685: \label{F2}
686: \end{figure}
687: 
688: In Fig.2, we have shown the constant $v_x$ contour plot in x-y plane
689: again at $\tau$=5.6 fm. As before the black lines
690: are for the ideal fluid evolution. The red and blue lines are for viscous fluid
691: with $\eta/s$=0.08 and 0.135 respectively. In the central region of the
692: fluid, viscous fluid has more velocity than its ideal counterpart. With
693: viscosity  while the energy density evolve slowly, the fluid velocity
694: evolve faster. Contour plot of the y-component of fluid velocity also
695: indicate similar results.
696: 
697: To obtain an idea of transverse expansion of viscous fluid, as opposed
698: to ideal fluid, in Fig.3, we have shown the constant temperature
699: contours in $\tau-x$ plane ,  at a fixed value of y=0 fm. Transverse expansion
700: is substantially enhanced in a viscous fluid. More the viscosity, more
701: is the transverse expansion. The plot also indicate
702: that at late time, fluid at x=y=0 behaves similarly to a ideal fluid.  
703: 
704: \begin{figure}%[h]
705: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig3.eps}
706: \vspace{-4cm}
707: \caption{(color online). Contour plots of temperature at y=0 fm in $x-\tau$ plane. 
708: The black  lines are for ideal fluid ($\eta/s$=0). The red and blue
709: lines are for viscous fluid with $\eta/s$=0.08 and 0.135 respectively.}
710: \label{F3}
711: \end{figure}
712: 
713: 1st order dissipative theories are acausal. As mentioned earlier, acausality can lead to unphysical behavior like reheating
714:  of the fluid in the early stage of evolution \cite{MR04,Baier:2006um}. Do we see any reheating?
715: In Fig.\ref{F3a}, the evolution of temperature in viscous dynamics,
716: with perturbative estimate of viscosity ($\eta/s$=0.135) is
717: shown. We have shown the temperature at two positions of the fluid,
718: x=y=0  (the solid line) and x=0,y=3 fm  (the dashed line).  
719: In both the positions of the fluid, with time as the fluid expands, temperature  
720: decreases (as it should be). We find no evidence of reheating. Reheating is not seen also  
721: with ADS/CFT estimate of viscosity. 
722: 
723: 
724: \begin{figure}%[h]
725: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig3a.eps}
726: \vspace{-4.5cm}
727: \caption{Evolution of the temperature in viscous dynamics with perturbative estimate of viscosity, $\eta/s$=0.135. The
728: solid and dashed
729: lines are for fluid  at x=y=0 and x=0,y=3 fm respectively. The x=0,y=3 fm curve is plotted with the right side scale.  } 
730: \label{F3a}
731: \end{figure}
732: 
733: 
734: In Fig.\ref{F4}, in 4-panels we have shown    shear stress tensors $\pi^{\tau\tau}(x,y=0)$, $\tau^2 \pi^{\eta\eta}(x,y=0)$,
735: $\pi^{xx}(x,y=0)$ and $\pi^{yy}(x,y=0)$ as a function of x.
736: $\eta/s$ =0.135.
737: The solid, long-dashed, dashed and short-dashed lines
738: are for time 0.6, 2.2, 3.2 and 4.2 fm respectively.
739: Initially at $\tau$=0.6 fm, $\pi^{\tau\tau}$ is zero. As the fluid evolve, $\pi^{\tau\tau}$ increases rapidly to a maximum
740: and
741:  then decreases. By 4 fm of evolution, 
742: it decreases to very small values. We also note that $\pi^{\tau\tau}$
743: is never very large. The viscous pressures
744:  $\tau^2 \pi^{\eta\eta}$, $\pi^{xx}$
745: and $\pi^{yy}$ are non-zero at initial time $\tau_i$=0.6 fm
746: As the fluid evolve these viscous fluxes rapidly decreases to
747: very small values.
748:    
749: 
750: \begin{figure}%[h]
751: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig4.eps}
752: \vspace{-2cm}
753: \caption{Shear stress tensor  $\pi^{\tau\tau}(x,y=0)$, $\tau^2\pi^{\eta\eta}(x,y=0)$,
754: $\pi^{xx}(x,y=0)$ and $\pi^{yy}(x,y=0)$ at 
755: $\tau$=0.6, 2.2, 3.2 and 4.2   fm are shown in 4 panels.
756:  }
757: \label{F4}
758: \end{figure}
759: 
760: Viscosity generates entropy. In the model entropy generation
761: due to dissipation can be calculated as,
762: 
763: \begin{equation}
764: \partial_\mu S^\mu = \frac{\pi^{\mu\nu} \pi_{\mu\nu}}{2\eta T} 
765: \end{equation}
766: 
767: Evolution of spatially average entropy is shown in Fig.\ref{F5}.   Entropy generation saturates after $\sim$ 2 fm of evolution.
768: It is expected.
769: As seen in Fig.\ref{F4}, viscous fluxes rapidly decreases and
770: by 2 fm of evolution, the viscous fluxes are decreased 
771: sufficiently
772: and  do not contribute significantly to the entropy.
773: 
774: \begin{figure}%[h]
775: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig5.eps}
776: \vspace{-4.5cm}
777: \caption{Evolution of average entropy with proper time,
778: for two values of $\eta/s$ is shown.}
779: \label{F5}
780: \end{figure}
781: 
782: \subsection{Particle spectra}
783: 
784: In this exploratory calculations we have not attempted to fit experimental
785: data. We just exhibit the effect of viscosity on (i) transverse momentum
786: distribution and (ii) elliptic flow of pions.   Viscosity
787: influences the particle production by (i) changing the freeze-out surface (freeze-out surface is extended) and (ii) by introducing
788: a correction to the
789: equilibrium distribution function.  
790: Non-equilibrium correction to equilibrium distribution function
791:   depend, quadratically on the momentum and linearly on the 
792: viscous fluxes. 
793: %At large $p_T$, particle spectra
794: %are changed more due to non-equilibrium distribution function than due %to extended freeze-out surfaces.
795: 
796: \begin{figure}%[h]
797: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig6.eps}
798: \vspace{-4.5cm}
799: \caption{$P_T$ distribution of pions. The solid line is for ideal fluid.
800: The long-dashed and medium-dashed  
801: lines are for viscous fluid with ADS/CFT ($\eta/s$=0.08) and perturbative ($\eta/s$=0.135) estimate of viscosity. 
802: Non-equilibrium correction to equilibrium distribution function
803: is included.}
804: \label{F6}
805: \end{figure}
806: 
807: In Fig.\ref{F6}, we have shown the transverse momentum distribution of 
808: pions obtained in the Cooper-Frye formalism. Freeze-out temperature is $T_F$=0.158 GeV. In this calculation resonance contribution
809:  to pion spectra is neglected. Pion production is increased in viscous dynamics.   We also note that effect of viscosity
810: is more prominent at large $p_T$ than at low $p_T$. $p_T$ spectra of pions are flattened with viscosity. Particle production
811: increases if viscosity
812: increases. Thus while with ADS/CFT estimate of viscosity,
813: $\eta/s$=0.08, at   $p_T$ =3 GeV, pion production is increased by a factor 3,
814: with the perturbative estimate of viscosity, $\eta/s$=0.135, the production
815: is increased by a factor of 5. Increase is even more at
816: larger $p_T$.
817: 
818: \begin{figure}%[h]
819: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig7.eps}
820: \vspace{-4.5cm}
821: \caption{Ratio of correction to particle production due to  non-equilibrium
822: distribution to equilibrium distribution function.}  
823: \label{F7}
824: \end{figure}
825: 
826: We have obtained the non-equilibrium distribution as a correction to the
827: equilibrium distribution function. It is implied that non-equilibrium effects
828: are small and  the ratio
829: 
830: \begin{equation}
831: \frac{dN^{neq}}{dN^{eq}}=\frac
832: {\frac{dN^{neq}}{dyd^2p_T}}{\frac{dN^{eq}}{dyd^2p_T}},
833: \end{equation}
834: 
835: \noindent is less than 1. 
836: In Fig.\ref{F7}, the ratio 
837: is shown as a function of $p_T$. With ADS/CFT estimate of viscosity, 
838: $\eta/s$=0.08, non-equilibrium correction to particle 
839: production become comparable to equilibrium contribution beyond 
840: $p_T$=5 GeV. However, with perturbative estimate, $\eta/s$=0.135,
841: non-equilibrium correction become comparable to or exceeds the equilibrium contribution at $p_T$=4.5 GeV. Thus with perturbative
842: estimate of viscosity, hydrodynamic description
843: break down above $p_T \sim 4.5 GeV$. 
844: The blast wave model analysis \cite{Teaney:2004qa} on the otherhand
845: indicated that viscous dynamics get invalidated beyond 
846: $p_T \sim$1.7 GeV. The results are not contradictory. In the blast wave model, at the freeze-out,
847: viscosity is quite large, sound attenuation length $\Gamma_s\sim$ 1.4 fm.
848: In the present simulation, even for perturbative estimate of viscosity,
849: sound attenuation length at the freeze-out is $\Gamma_s \sim 0.2 fm$,
850: 7 times smaller than the sound attenuation length used in the blast wave analysis. Naturally, non-equilibrium corrections to
851: equilibrium distribution
852: function remains small over an extended $p_T$ range.  
853:   
854:  
855: \begin{figure}%[h]
856: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig8.eps}
857: \vspace{-4.5cm}
858: \caption{Elliptic flow as a function of transverse momentum. The solid
859: line is for ideal fluid. The long-dashed, medium-dashed and short-dashed lines are for
860: viscous fluid with $\eta/s$=0.08 and 0.135 respectively.
861: Non-equilibrium correction to equilibrium distribution function is included.}
862: \label{F8}
863: \end{figure}
864: 
865: We have also calculated the elliptic flow in the model. Being a ratio, 
866: elliptic flow is very sensitive to the model. Experimentally, elliptic flow
867: saturates at large $p_T$. It is known
868: that ideal fluid does not explain the saturation of elliptic flow. 
869: In contrast to experiment, with ideal fluid, elliptic flow continues to increase
870: with $p_T$. In Fig.\ref{F8}, we have compared the elliptic flow in ideal and viscous
871: fluid.
872: The solid line is $v_2$ for the ideal fluid. The long-dashed and medium-dashed 
873: lines are  for viscous fluid with ADS/CFT ($\eta/s$=0.08) and 
874: perturbative ($\eta/s$=0.135) estimated viscosity.
875: Elliptic flow decreases with viscosity. As viscosity increases, elliptic flow
876: is also reduced. We also note that both for ADS/CFT and perturbative estimate of viscosity, elliptic flow indicate
877: saturation at large $p_T$.
878: The result is very encouraging, as experimentally also elliptic flow
879: tends to saturate at large $p_T$.
880:  
881: As discussed earlier, ideal fluid dynamics,   can explain a large volume of data in Au+Au 
882: collisions at RHIC.  Our present knowledge about the hot dense matter
883: produced in central Au+Au collisions are obtained from the ideal
884: fluid analysis.
885: As shown in the present paper, QGP fluid,  even with 
886: ADS/CFT estimate of viscosity $\eta/s$=0.08, generate enough entropy to enhance particle production by a factor of 3 at $p_T$=3
887: GeV. Naturally,
888: if QGP fluid is viscous,  initial conditions
889: as required to explain RHIC data with ideal fluid dynamics will overpredict
890: the experimental $p_T$ distribution. Viscous fluid dynamics will require
891: much less initial temperature than an ideal fluid to explain the same
892: $p_T$ spectra.
893:  As an example, in Fig.10, we have compared the pion spectra  obtained 
894: in  viscous dynamics with ADS/CFT estimate of
895: viscosity ($\eta/s$=0.08), initialised with  entropy density of 110,80,
896: and 60 $fm^{-3}$ with the pion spectra obtained in an 
897: ideal fluid dynamics, initialised with entropy density s=110 $fm^{-3}$. 
898: For all the fluids, the initial time is $\tau_i$=0.6 fm and freeze-out temperature
899: is 158 MeV.
900: Viscous fluid initialised with entropy density between 60-80 $fm^{-3}$,
901:  compare well with the pion spectra from ideal fluid initialised
902: at much higher entropy density. 
903: To produce the same pion spectra, while ideal
904: fluid require a initial temperature of 350 MeV, the viscous fluid require
905: much less temperature between 270-290 MeV. The ideal fluid dynamics
906: can overestimate the initial temperature of fluid produced in
907: Au+Au collisions at RHIC by 20-30\%.  
908: 
909: \begin{figure}%[h]
910: \includegraphics[bb=14 13 581 829,width=0.99\linewidth,clip]{fig9.eps}
911: \vspace{-4.5cm}
912: \caption{The solid circle is the $p_T$ distribution obtained 
913: in ideal fluid dynamics, with initial entropy density s=110$ fm^{-3}$. 
914: The solid, long dashed and dashed lines are for viscous fluid with ADS/CFT estimate of viscosity, $\eta/s$=0.08, initialised
915: at entropy density
916: s=60,80 and 110 $fm^{-3}$ respectively. Non-equilibrium correction to equilibrium distribution function is included.}
917: \label{F10}
918: \end{figure}
919:  
920: \section{Summary and conclusions}
921: 
922: We have studied the boost-invariant hydrodynamic evolution of 
923: QGP fluid with dissipation due to shear viscosity. In this study we 
924: have employed the 1st order theory of dissipative relativistic fluid. 
925: 1st order theories suffer from the problem of causality, signal can 
926: travel faster than light.  Unphysical effects like
927: reheating of the fluid, early in the evolution, can occur. However, for
928: a fluid like QGP, where viscosity is small, with appropriate initial
929: conditions, effects of causality violation can be minimised.
930: In this model study, we have 
931: considered two values of viscosity, the ADS/CFT motivated value,
932: $\eta/s \approx$0.08 and perturbatively estimated viscosity, $\eta/s \approx$0.135.
933: Both the ideal and viscous fluids are initialised similarly. At the initial 
934: time $\tau_i$=0.6 fm, initial central entropy density is 110 $fm^{-3}$, 
935: with transverse profile taken from a Glauber model calculation. 
936: Viscous hydrodynamics require initial conditions  for the shear-stress tensor components. It is assumed that at the equilibration
937: time, the shear stress
938: tensors components have reached their boost-invariant values. 
939: The initial conditions of
940: the fluid are such that for both the values of viscosity ($\eta/s$=0.08 and 0.135), the condition of validity of viscous
941: hydrodynamics, $\Gamma_s/\tau << 1$ is satisfied all through the evolution.
942: Explicit simulation of ideal and viscous fluids confirms that energy density of a 
943: viscous fluid, evolve slowly than its ideal counterpart. The fluid velocities
944: on the other hand evolve faster in viscous dynamics than in ideal 
945: dynamics. Transverse expansion is also more in viscous dynamics.
946: For a similar freeze-out condition freeze-out surface is extended in
947: viscous fluid.
948: 
949: We have also studied the effect of viscosity on particle production.
950: Viscosity generates entropy leading to enhanced particle production.
951: Particle production is  increased due to (i) extended freeze-out surface and (ii) non-equilibrium
952: correction to equilibrium distribution function. Non-equilibrium
953: correction to equilibrium distribution function is a dominating
954: factor influencing the particle production
955: at large $p_T$.
956:   With ADS/CFT (perturbative) estimate of viscosity, at $p_T$=3 GeV,
957: pion production is increased by a factor 3 (5) . Increase is even more at large $p_T$. While viscosity enhances particle
958: production,
959: it reduces the elliptic flow. At $p_T$=3 GeV, for ADS/CFT(perturbative) estimate of viscosity,
960: elliptic flow is reduced by a factor of 2(3). We also find that at large $p_T$
961: elliptic flow tends to saturate.
962: 
963: To conclude, present study shows viscosity, even if small, can be
964: very important in analysis of RHIC Au+Au collisions. Currently accepted
965: initial temperature of hot dense matter produced in RHIC Au+Au collisions,
966: obtained from ideal fluid analysis
967: can be changed by 20\% or more with dissipative dynamics. 
968:  
969: \acknowledgements
970: Prof. U. Heinz initiated this programme of numerical 
971: simulation of dissipative hydrodynamics in 2+1 dimension.  
972: The author would like to thank Prof.  Heinz for several discussions
973: and suggestion.
974: %%%%%%%%%%%%%%%%%%%%% Bibliography %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
975: 
976: \begin{thebibliography}{99}
977: 
978: \bibitem{BRAHMSwhitepaper} 
979: BRAHMS Collaboration, I. Arsene {\it et al.},  
980: %"Quark-gluon plasma and the color glass condensate at RHIC?  
981: % The prespective from the BRAHMS experiment" 
982: Nucl. Phys. A {\bf 757}, 1 (2005). 
983: %nucl-ex/0410020 
984:  
985: \bibitem{PHOBOSwhitepaper} 
986: PHOBOS Collaboration,  B. B. Back {\it et al.},  
987: %"The PHOBOS Perspective on Discoveries at RHIC'' 
988: Nucl. Phys. A {\bf 757}, 28 (2005). 
989: %nucl-ex/0410020 
990:  
991: \bibitem{PHENIXwhitepaper} 
992: PHENIX Collaboration, K.~Adcox {\it et al.}, 
993: %``Formation of dense partonic matter in relativistic nucleus nucleus 
994: %collisions at RHIC: Experimental evaluation by the PHENIX  collaboration,'' 
995: Nucl. Phys. A {\bf 757} (2005), in press [arXiv:nucl-ex/0410003]. 
996: %arXiv:nucl-ex/0410003. 
997: %%CITATION = NUCL-EX 0410003;%% 
998:   
999: \bibitem{STARwhitepaper} 
1000: STAR Collaboration, J. Adams {\it et al.}, 
1001: %''Experimental and theoretical challenges in the search for the quark  
1002: %  gluon plasma: The STAR Collaboration's critical assessment of the  
1003: %  evidence from RHIC collisions'' 
1004: Nucl. Phys. A {\bf 757} (2005), in press [arXiv:nucl-ex/0501009]. 
1005: %%CITATION = NUCL-EX 0501009;%% 
1006: 
1007: \bibitem{QGP3}
1008: P.~F. Kolb and U. Heinz,
1009: in {\it Quark-Gluon Plasma 3}, edited by R.~C. Hwa and 
1010: X.-N. Wang (World Scientific, Singapore, 2004), p.~634.
1011: %%CITATION = NUCL-TH 0305084;%%
1012: 
1013: \bibitem{Heinz:2004ar}
1014: U.~Heinz,
1015:   %``Hydrodynamics at RHIC: How well does it work, where and how does it break
1016:   %down?,''
1017: J.\ Phys.\ G {\bf 31}, S717 (2005).
1018:   %  [arXiv:nucl-th/0412094].
1019:   %%CITATION = NUCL-TH 0412094;%%
1020: 
1021: \bibitem{Eckart}
1022: C.~Eckart, Phys. Rev. {\bf 58}, 919 (1940).
1023: 
1024: \bibitem{LL63}
1025: L.~D.~Landau and E.~M.~Lifshitz, {\it Fluid Mechanics}, Sect. 127,
1026: Pergamon, Oxford, 1963.
1027: 
1028: \bibitem{IS79}
1029: W. Israel, Ann. Phys. (N.Y.) {\bf 100}, 310 (1976);
1030: W.~Israel and J.~M.~Stewart, Ann. Phys. (N.Y.) {\bf 118}, 349 (1979).
1031: 
1032: \bibitem{Teaney:2004qa}
1033:   D.~A.~Teaney,
1034:   %``Viscosity and thermalization,''
1035:   J.\ Phys.\ G {\bf 30}, S1247 (2004).
1036: %  [arXiv:nucl-th/0403053].
1037:   %%CITATION = NUCL-TH 0403053;%%
1038: %\cite{Teaney:2003kp}
1039: %\bibitem{Teaney:2003kp}
1040: %  D.~Teaney,
1041:   %``Effect of shear viscosity on spectra, elliptic flow, and Hanbury
1042:   %Brown-Twiss radii,''
1043:   Phys.\ Rev.\ C {\bf 68}, 034913 (2003)
1044:   [arXiv:nucl-th/0301099].
1045:   %%CITATION = NUCL-TH 0301099;%%
1046: 
1047: \bibitem{Muronga:2001zk}
1048:   A.~Muronga,
1049:   %``Second order dissipative fluid dynamics for ultra-relativistic nuclear
1050:   %collisions,''
1051:   Phys.\ Rev.\ Lett.\  {\bf 88}, 062302 (2002)
1052:   [Erratum {\it ibid.}\  {\bf 89}, 159901 (2002)];
1053: %  [arXiv:nucl-th/0104064].
1054:   %%CITATION = NUCL-TH 0104064;%%
1055:   %``Causal Theories of Dissipative Relativistic Fluid Dynamics for Nuclear
1056:   %Collisions,''
1057:   and Phys.\ Rev.\ C {\bf 69}, 034903 (2004).
1058: %  [arXiv:nucl-th/0309055].
1059:   %%CITATION = NUCL-TH 0309055;%%
1060: 
1061: 
1062: \bibitem{MR04}
1063:   A.~Muronga and D.~H.~Rischke,
1064:   %``Evolution of hot, dissipative quark matter in relativistic nuclear
1065:   %collisions,''
1066:   nucl-th/0407114\,(v2).
1067:   %%CITATION = NUCL-TH 0407114;%%
1068: 
1069: \bibitem{CH05}
1070:   A.~K.~Chaudhuri and U.~Heinz,
1071:   %``Hydrodynamical evolution of dissipative QGP fluid,''
1072:   nucl-th/0504022.
1073:   %%CITATION = NUCL-TH 0504022;%%
1074: 
1075: %\cite{Heinz:2005bw}
1076: \bibitem{Heinz:2005bw}
1077:   U.~W.~Heinz, H.~Song and A.~K.~Chaudhuri,
1078:   %``Dissipative hydrodynamics for viscous relativistic fluids,''
1079:   Phys.\ Rev.\ C {\bf 73}, 034904 (2006)
1080:   [arXiv:nucl-th/0510014].
1081:   %%CITATION = NUCL-TH 0510014;%%
1082: %%%%%%%%%%%%%%%%%%%%%%%%
1083: \bibitem{vis_bj} 
1084: %\cite{Kajantie:1985tj}
1085: %\bibitem{Kajantie:1985tj}
1086:   K.~Kajantie,
1087:   %``Hydrodynamics And Approach To Equilibrium In Quark - Gluon Plasma,''
1088:   Nucl.\ Phys.\ A {\bf 418}, 41C (1984).
1089:   %%CITATION = NUPHA,A418,41C;%%
1090: %\cite{Baym:1985tn}
1091: %\bibitem{Baym:1985tn}
1092:   G.~Baym,
1093:   %``Entropy Production And The Evolution Of Ultrarelativistic Heavy Ion
1094:   %Collisions,''
1095:   Nucl.\ Phys.\ A {\bf 418}, 525C (1984).
1096:   %%CITATION = NUPHA,A418,525C;%%
1097: %\cite{Hosoya:1983xm}
1098: %\bibitem{Hosoya:1983xm}
1099:   A.~Hosoya and K.~Kajantie,
1100:   %``Transport Coefficients Of QCD Matter,''
1101:   Nucl.\ Phys.\ B {\bf 250}, 666 (1985).
1102:   %%CITATION = NUPHA,B250,666;%%
1103: %\cite{Chaudhuri:1998kv}
1104: %\bibitem{Chaudhuri:1998kv}
1105:   A.~K.~Chaudhuri,
1106:   %``Chemical equilibration in viscous quark-gluon plasma and  electromagnetic
1107:   %signals,''
1108:   J.\ Phys.\ G {\bf 26}, 1433 (2000)
1109:   [arXiv:nucl-th/9808074].
1110:   %%CITATION = NUCL-TH 9808074;%%
1111: %\cite{Chaudhuri:1997zs}
1112: %\bibitem{Chaudhuri:1997zs}
1113:   A.~K.~Chaudhuri,
1114:   %``Effect of viscosity on one dimensional hydrodynamic flow and direct
1115:   %photons from 200-A-GeV S + Au collisions at CERN SPS,''
1116:   Phys.\ Scripta {\bf 61}, 311 (2000)
1117:   [arXiv:nucl-th/9705047].
1118:   %%CITATION = NUCL-TH 9705047;%%
1119: %\cite{Chaudhuri:1995se}
1120: %\bibitem{Chaudhuri:1995se}
1121:   A.~K.~Chaudhuri,
1122:   %``Thermal photons from S + Au collisions at 200-A/GeV: A Hadron gas
1123:   %picture,''
1124:   Phys.\ Rev.\ C {\bf 51}, 2889 (1995).
1125:   %%CITATION = PHRVA,C51,2889;%%
1126: 
1127: %\cite{Baier:2006um}
1128: \bibitem{Baier:2006um}
1129:   R.~Baier, P.~Romatschke and U.~A.~Wiedemann,
1130:   %``Dissipative hydrodynamics and heavy ion collisions,''
1131:   arXiv:hep-ph/0602249.
1132:   %%CITATION = HEP-PH 0602249;%%
1133: 
1134: %\cite{Arnold:2000dr}
1135: \bibitem{Arnold:2000dr}
1136:   P.~Arnold, G.~D.~Moore and L.~G.~Yaffe,
1137:   %``Transport coefficients in high temperature gauge theories. I:  Leading-log
1138:   %results,''
1139:   JHEP {\bf 0011}, 001 (2000)
1140:   [arXiv:hep-ph/0010177].
1141:   %%CITATION = HEP-PH 0010177;%%
1142: 
1143: %\cite{Baym:1990uj}
1144: \bibitem{Baym:1990uj}
1145:   G.~Baym, H.~Monien, C.~J.~Pethick and D.~G.~Ravenhall,
1146:   %``Transverse Interactions And Transport In Relativistic Quark - Gluon And
1147:   %Electromagnetic Plasmas,''
1148:   Phys.\ Rev.\ Lett.\  {\bf 64}, 1867 (1990).
1149:   %%CITATION = PRLTA,64,1867;%%
1150: 
1151: 
1152: 
1153: %\cite{Policastro:2001yc}
1154: \bibitem{Policastro:2001yc}
1155:   G.~Policastro, D.~T.~Son and A.~O.~Starinets,
1156:   %``The shear viscosity of strongly coupled N = 4 supersymmetric Yang-Mills
1157:   %plasma,''
1158:   Phys.\ Rev.\ Lett.\  {\bf 87}, 081601 (2001)
1159:   [arXiv:hep-th/0104066].
1160:   %%CITATION = HEP-TH 0104066;%%
1161: 
1162:  %\cite{Policastro:2002se}
1163: \bibitem{Policastro:2002se}
1164:   G.~Policastro, D.~T.~Son and A.~O.~Starinets,
1165:   %``From AdS/CFT correspondence to hydrodynamics,''
1166:   JHEP {\bf 0209}, 043 (2002)
1167:   [arXiv:hep-th/0205052].
1168:   %%CITATION = HEP-TH 0205052;%%
1169: 
1170: \end{thebibliography}
1171: \end{document}
1172: 
1173: %\cite{Paech:2006st}
1174: %\bibitem{Paech:2006st}
1175: %  K.~Paech and S.~Pratt,
1176:   %``Origins of Viscosity at RHIC,''
1177:  % arXiv:nucl-th/0604008.
1178: %  %%CITATION = NUCL-TH 0604008;%%
1179: 
1180: %\cite{Pratt:2006ss}
1181: \bibitem{Pratt:2006ss}
1182: %  S.~Pratt and K.~Paech,
1183:   %``Viscosity at RHIC,''
1184:  % arXiv:nucl-th/0604007.
1185: %  %%CITATION = NUCL-TH 0604007;%%
1186: 
1187: %\cite{Baier:2006sr}
1188: %\bibitem{Baier:2006sr}
1189: %  R.~Baier, P.~Romatschke and U.~A.~Wiedemann,
1190: %  %``Transverse flow in relativistic viscous hydrodynamics,''
1191: %  arXiv:nucl-th/0604006.
1192: %  %%CITATION = NUCL-TH 0604006;%%
1193: 
1194: 
1195: