nucl-th0607009/fl.tex
1: \documentclass[twocolumn,showpacs,superscriptaddress,preprintnumbers,
2: amsmath,amssymb]{revtex4}
3: \usepackage{graphicx}
4: \usepackage{dcolumn}
5: \usepackage{bm}
6: 
7: \newcommand{\mean}[1]{\left\langle #1 \right\rangle}
8: 
9: \def\simge{\mathrel{
10:      \rlap{\raise 0.511ex \hbox{$>$}}{\lower 0.511ex \hbox{$\sim$}}}}
11: \def\simle{\mathrel{
12:      \rlap{\raise 0.511ex \hbox{$<$}}{\lower 0.511ex \hbox{$\sim$}}}}
13: 
14: \begin{document}
15: \preprint{SPhT-T06/077; TIFR/TH/06-20}
16: \title{Eccentricity fluctuations and elliptic flow at RHIC}
17: \author{Rajeev S. Bhalerao}
18: \affiliation{Department of Theoretical Physics, TIFR,
19:    Homi Bhabha Road, Colaba, Mumbai 400 005, India}
20: \author{Jean-Yves Ollitrault}
21: \affiliation{Service de Physique Th\'eorique, CEA/DSM/SPhT, Unit\'e de
22: recherche associ\'ee au CNRS,\\ F-91191 Gif-sur-Yvette Cedex, France.}
23: 
24: \date{\today}
25: \begin{abstract}
26: Fluctuations in nucleon positions can affect the spatial 
27: eccentricity of the overlap zone in nucleus-nucleus collisions. 
28: We show that elliptic flow should be scaled by different 
29: eccentricities depending on which method is used for the 
30: flow analysis.
31: These eccentricities are estimated semi-analytically.
32: When $v_2$ is analyzed from 4-particle cumulants, or using the 
33: event plane from directed flow in a zero-degree calorimeter, 
34: the result is shown to be insensitive to eccentricity fluctuations. 
35: \end{abstract}
36: 
37: \pacs{25.75.Ld, 24.10.Nz}
38: 
39: \maketitle
40: 
41: 
42: \medskip
43: \noindent 1. Introduction
44: 
45: 
46: Elliptic flow, $v_2$, is one of the key observables in 
47: nucleus-nucleus collisions at RHIC~\cite{Ackermann:2000tr}. 
48: It originates from the almond shape of the overlap zone 
49: (see Fig.~\ref{fig:ellipse}) which produces, through unequal 
50: pressure gradients, an anisotropy in the transverse momentum 
51: distribution~\cite{Ollitrault:1992bk}, the so-called 
52: $v_2\equiv\mean{\cos 2\phi}$, where $\phi$'s are the azimuthal
53: angles of the detected particles with respect to the
54: reaction plane. 
55: 
56: Preliminary analyses of $v_2$ in Cu-Cu collisions at RHIC
57: ~\cite{Masui:2005aa,Manly:2005zy,Wang:2005ab},
58: presented at the QM'2005 conference, 
59: reported values surprisingly large compared to 
60: theoretical expectations,
61: almost as large as in Au-Au collisions. 
62: It was shown by the PHOBOS collaboration~\cite{Manly:2005zy}
63: that fluctuations in nucleon positions provide a natural explanation 
64: for this large magnitude. 
65: The idea is the following:
66: The time scale of the nucleus-nucleus collision at RHIC is so 
67: short that each nucleus sees the nucleus coming in the opposite 
68: direction in a frozen configuration, with nucleons located at positions 
69: whose probabilities are determined according to the nuclear wave
70: function. 
71: Fluctuations in the nucleon positions result in fluctuations 
72: in the almond shape and orientation (see Fig.~\ref{fig:ellipse}), 
73: and hence in larger values of $v_2$. 
74: 
75: In this Letter, we discuss various definitions of the 
76: eccentricity of the overlap zone. 
77: We show that estimates of $v_2$ using different 
78: methods should be scaled by appropriate choices of the 
79: eccentricity.
80: We then compute the effect of fluctuations 
81: on the eccentricity semi-analytically to leading order in $1/N$, 
82: where $N$ is the mean number of participants at a given centrality.
83: A similar study was recently performed by S. Voloshin 
84: on the basis of Monte-Carlo Glauber calculations~\cite{Voloshin:2006gz}. 
85: 
86: \medskip
87: \noindent 2. Eccentricity scaling and fluctuations
88: 
89: Elliptic flow 
90: is determined by the initial density profile. 
91: Although its precise value depends on the detailed shape of the
92: profile, most of the relevant information is encoded in 
93: three quantities: 
94: 1) the initial 
95: eccentricity of the overlap zone, $\varepsilon$, which will be defined 
96: more precisely below;
97: 2) the density $n$, which determines 
98: pressure gradients through the equation of state 
99: (by density, we mean the particle density, $n$, 
100: at the time when elliptic flow develops; this time is of 
101: the order of the transverse size $R$. Quite remarkably, 
102: the density thus defined varies little with centrality, 
103: and has almost the same value in Au-Au and Cu-Cu collisions 
104: at the same colliding energy per nucleon~\cite{Bhalerao:2005mm});
105: 3) the system transverse size $R$, which determines the number 
106: of collisions per particle. $v_2$ scales like $\varepsilon$ for 
107: small $\varepsilon$, that is, $v_2=\varepsilon f(n, R)$. 
108: 
109: This proportionality relation is only approximate. However, 
110: hydrodynamical calculations~\cite{Bhalerao:2005mm} show that it is a
111: very good approximation in practice for nucleus-nucleus collisions. 
112: Eccentricity scaling holds for integrated flow 
113: as well as for the differential flow of identified particles. 
114: In the latter case, the function $f(n,R)$ also depends on 
115: the mass, transverse momentum and rapidity of the particle. 
116: 
117: 
118: Eccentricity scaling of $v_2$ is generally believed to be
119: a specific prediction of relativistic 
120: hydrodynamics. In the form above, the scaling is expected 
121: to be more general: it does not require 
122: thermalization, as implicitly assumed by hydrodynamics. 
123: If thermalization is achieved, that is, if the system size $R$ 
124: is much larger than the mean free path $\lambda$, then the scaling
125: is stronger: $v_2/\varepsilon$ no longer depends on $R$, 
126: but only on the density $n$~\cite{Bhalerao:2005mm}. 
127: 
128: \begin{figure}
129: \includegraphics*[width=0.7\linewidth]{ellipse.eps}
130: \caption{Schematic view of a collision of two identical
131: nuclei, in the plane transverse to the beam direction ($z$-axis). 
132: The $x$- and $y$-axes are drawn as per the standard convention. 
133: The dots indicate the positions of 
134: participant nucleons. Due to fluctuations, the 
135: overlap zone could be shifted and tilted with respect to the $(x,y)$
136: frame. $x'$ and $y'$ are the principal axes of inertia of the dots.}
137: \label{fig:ellipse}
138: \end{figure}
139: 
140: The standard definition of the eccentricity is~\cite{Sorge:1998mk} 
141: \begin{equation} 
142: \label{epsstandard}
143: \varepsilon_{\rm s}=
144: \frac{\mean{y^2-x^2}} {\mean{y^2+x^2}},
145: \end{equation}
146: where $(x,y)$ is the position of a participant nucleon in 
147: the coordinate system defined in Fig.~\ref{fig:ellipse}.
148: Throughout this paper, $\mean{...}$ denotes an ensemble average: 
149: here, it means an average over participant nucleons and 
150: over many collision events of the same impact parameter. 
151: This standard eccentricity applies to most hydrodynamic 
152: calculations: indeed, most hydrodynamic calculations 
153: (with the exception of Ref.~\cite{Socolowski:2004hw}) use smooth,  
154: event-averaged initial conditions, with an initial entropy density 
155: proportional to the density of participants~\cite{Kolb:2003dz}. 
156: The charged multiplicity then scales like the number of 
157: participants, as observed experimentally. 
158: (Adding a component proportional to the number of 
159: binary collisions, as argued in~\cite{Kharzeev:2000ph}, does not
160: change the eccentricity significantly). 
161: 
162: It was recently argued~\cite{Drescher:2006pi}
163:  that the Color Glass Condensate picture of heavy-ion 
164: collision leads to a different definition of the eccentricity, which
165: may be significantly larger than the standard eccentricity. This 
166: interesting possibility will not be considered further here. 
167: 
168: Now, because of the event-by-event fluctuations in the participant 
169: nucleon positions~\cite{Miller:2003kd}, the eccentricity driving 
170: elliptic flow in a given event is that defined by the principal 
171: axes $(x',y')$ of the distribution of participant nucleons, 
172: see Fig.~\ref{fig:ellipse}. 
173: This ``participant eccentricity'' $\varepsilon_{\rm part}$ 
174: can be written as~\cite{Manly:2005zy}
175: \begin{equation}
176: \label{epspart}
177: \varepsilon_{\rm part}=\frac{
178: \sqrt{(\sigma_y^2-\sigma_x^2)^2+4 \sigma_{xy}^2}}
179: {\sigma_y^2+\sigma_x^2},
180: \end{equation}
181: where 
182: \begin{eqnarray}
183: \label{defsigma}
184: \sigma_x^2&=&\{x^2\}-\{x\}^2\cr
185: \sigma_y^2&=&\{y^2\}-\{y\}^2\cr
186: \sigma_{xy}&=&\{xy\}-\{x\}\{y\},
187: \end{eqnarray}
188: and $\{...\}$ denotes the average over all participants in one 
189: collision event (sample average). 
190: Our basic assumption in this paper is that {\it in each event\/},
191: the elliptic flow $v_2$ is proportional to the participant
192: eccentricity $\varepsilon_{\rm part}$. 
193: 
194: Experimentally, elliptic flow is analyzed by selecting events in a
195: centrality class. The quantities which drive elliptic flow, 
196: namely, the density $n$, the transverse size $R$, the eccentricity 
197: $\varepsilon_{\rm part}$, fluctuate from one event to the other.
198: This causes dynamical fluctuations of $v_2$. 
199: The effect of impact parameter fluctuations was  carefully 
200: studied in~\cite{Adler:2002pu} and was shown to be small. 
201: In this paper, we focus on fluctuations in the positions of 
202: participant nucleons. 
203: It will be shown below that fluctuations in the eccentricity 
204: $\varepsilon_{\rm part}$ dominate over fluctuations in size and density.
205: 
206: 
207: \medskip
208: \noindent 3. To each method its own eccentricity 
209: 
210: 
211: Since the reaction plane is not known exactly on an event-by-event basis, 
212: $v_2$ is measured indirectly using azimuthal correlations. 
213: Several methods have been used, which yield different estimates
214: of $v_2$. 
215: We argue that these estimates are affected in different ways
216: by fluctuations of $v_2$:
217: \begin{itemize}
218: \item The event-plane method~\cite{Poskanzer:1998yz} has been 
219: implemented by the STAR~\cite{Ackermann:2000tr}, 
220: PHOBOS~\cite{Back:2002gz} and PHENIX~\cite{Adler:2003kt} 
221: collaborations at RHIC. The event plane is an estimate of the 
222: reaction plane; it is defined as the plane spanned by the collision 
223: axis and the major axis of the ellipse formed by the transverse momenta of 
224: outgoing particles. 
225: It corresponds to the $x'$ axis defined by the participants 
226: (Fig.~\ref{fig:ellipse}), up to statistical
227: fluctuations which are taken care of by the analysis 
228: (this is the so-called ``event-plane resolution''). 
229: Outgoing particles are then individually correlated to this event
230: plane. The corresponding estimate of $v_2$ will be denoted by 
231: $v_2\{ {\rm EP2}\}$. 
232: Another method determines $v_2$ from two-particle azimuthal 
233: correlations between outgoing particles~\cite{Adcox:2002ms}. 
234: The corresponding estimate is usually denoted by $v_2\{2\}$, and 
235: is defined by an equation of the type 
236: $v_2\{2\}=\sqrt{\mean{\cos 2(\phi_1-\phi_2)}}$. 
237: Both methods are essentially equivalent~(see Sec.~III.C.2 of 
238: Ref.~\cite{Borghini:2000sa}). 
239: If $v_2$ fluctuates, both yield the rms value of 
240: $v_2$~\cite{Adler:2002pu}: 
241: $v_2\{2\}\simeq v_2\{{\rm EP2}\}=\sqrt{\mean{v_2^2}}$. 
242: Since $v_2$ in each event scales with 
243: $\varepsilon_{\rm part}$, one expects 
244: $v_2\{2\}\simeq v_2\{{\rm EP2}\}\propto\varepsilon\{2\}$, where 
245: $\varepsilon\{2\}$ is defined by~\cite{Miller:2003kd}
246: \begin{equation}
247: \label{v2std}
248: \varepsilon\{2\}\equiv\sqrt{\mean{\varepsilon_{\rm
249:     part}^2}}. 
250: \end{equation}
251: This scaling differs from that proposed in Ref.~\cite{Manly:2005zy}, 
252: $v_2\propto\mean{\varepsilon_{\rm part}}$. 
253: \item Four-particle cumulants of azimuthal 
254: correlations~\cite{Borghini:2001vi} can be used to 
255: reduce the bias from nonflow correlations. The corresponding 
256: estimate of $v_2$ is denoted by $v_2\{4\}$. It involves a combination
257: of 2-particle and 4-particle correlations, 
258: i.e., the 2nd and 4th moments of the distribution of 
259: $v_2$, $\mean{v_2^2}$ and $\mean{v_2^4}$. Assuming again that 
260: $v_2$ scales with $\varepsilon_{\rm part}$ in each event, 
261: one obtains $v_2\{4\}\propto\varepsilon\{4\}$, where 
262: $\varepsilon\{4\}$ is defined by~\cite{Miller:2003kd}: 
263: \begin{equation}
264: \label{v24}
265: \varepsilon\{4\}\equiv\left(2\mean{\varepsilon_{\rm part}^2}^2
266: -\mean{\varepsilon_{\rm part}^4}\right)^{1/4}.
267: \end{equation}
268: 
269: \item An alternative method determines the event-plane from directed 
270: flow~\cite{Adams:2003zg}, determined in a zero-degree calorimeter
271: (ZDC) which detects spectator neutrons. Particles in the central
272: rapidity region are then correlated to this event plane. This 
273: last estimate of $v_2$ is denoted by 
274: $v_2\{ {\rm ZDC}\}$~\cite{Wang:2005ab}. 
275: It differs from the previous ones in that the reference direction
276: (the event plane) is determined by spectator neutrons from the 
277: projectile, rather than by participants. The direction defined by 
278: spectator neutrons is the $x$-axis, not the $x'$-axis 
279: (see Fig.~\ref{fig:ellipse}), up to fluctuations in spectator 
280: positions which are taken care of by the analysis. 
281: Therefore, the relevant eccentricity for this analysis 
282: is the reaction-plane eccentricity
283: \begin{equation}
284: \varepsilon_{\rm RP}\equiv \frac{\sigma_y^2-\sigma_x^2}
285: {\sigma_y^2+\sigma_x^2},
286: \end{equation}
287: and $v_2\{ {\rm ZDC} \}$ should scale correspondingly like 
288: \begin{equation}
289: \label{v2ZDC}
290: v_2\{ {\rm ZDC}\}\propto \mean{\varepsilon_{\rm RP}}. 
291: \end{equation}
292: \end{itemize}
293: This estimate of $v_2$ is also unbiased by nonflow correlations 
294: because it involves a 3-particle correlation (instead of 
295: a 2-particle correlation in the standard event-plane method), 
296: and also because the ZDC calorimeter has a wide rapidity gap 
297: with the central rapidity detector. 
298: 
299: \medskip
300: \noindent 4. Computing the fluctuations
301: 
302: 
303: We now compute the eccentricities entering Eqs.~(\ref{v2std}), 
304: (\ref{v24}) and (\ref{v2ZDC}), namely, $\varepsilon\{2\}$, 
305: $\varepsilon\{4\}$ and $\mean{\varepsilon_{\rm RP}}$, to 
306: leading order in the fluctuations. 
307: 
308: 
309: To that end, we write each sample average (over one event) 
310: entering Eq.~(\ref{defsigma}) as the sum of an ensemble average and 
311: a fluctuation, e.g., $\{x^2\}=\mean{x^2}+~\delta_{x^2}$, 
312: where $\delta_{x^2}$ is the fluctuation. 
313: We choose the coordinate system in Fig.~\ref{fig:ellipse}, where
314: $\mean{x}=\mean{y}=\mean{xy}=0$. 
315: Substituting in Eq.~(\ref{epspart}) and retaining terms 
316: to second order in the $\delta$'s gives
317: \begin{eqnarray}
318: \label{epspart2}
319: \varepsilon_{\rm part}=\varepsilon_{\rm s}
320: &+&\frac{\delta_{y^2}-\delta_{x^2}}{\mean{r^2}}
321: -\varepsilon_{\rm s}\frac{\delta_{y^2}+\delta_{x^2}}{\mean{r^2}}
322: -\frac{\delta_y^2-\delta_x^2}{\mean{r^2}} \nonumber  \\
323: &+&\frac{2\delta_{xy}^2}{\varepsilon_{\rm s}\mean{r^2}^2}
324: -\frac{\delta_{y^2}^2-\delta_{x^2}^2}{\mean{r^2}^2} \nonumber  \\
325: &+&\varepsilon_{\rm s}\frac{\delta_y^2+\delta_x^2}{\mean{r^2}}
326: +\varepsilon_{\rm s}\frac{(\delta_{y^2}+\delta_{x^2})^2}{\mean{r^2}^2},
327: \end{eqnarray}
328: where $\mean{r^2}=\mean{x^2+y^2}$. 
329: The reaction plane eccentricity $\varepsilon_{\rm RP}$ is given
330: by a similar expression, except for the fifth term on the right-hand
331: side (rhs) proportional to  $\delta_{xy}^2$ which does not appear in 
332: $\varepsilon_{\rm RP}$. 
333: The 2nd and 3rd terms on the rhs
334: of Eq.~(\ref{epspart2}) are linear and the remaining 5 terms are quadratic in
335: fluctuations. In an obvious notation Eq.~(\ref{epspart2}) can be rewritten as 
336: \begin{equation}
337: \label{AandB}
338: \varepsilon_{\rm part}=\varepsilon_{\rm s} +\sum_{i=1}^2 A_i\delta_i^{(1)}
339: +\sum_{i=1}^5 B_i\delta_i^{(2)}, 
340: \end{equation}
341: where $\delta_i^{(1)}$ are linear and $\delta_i^{(2)}$ are quadratic
342: in $\delta$'s. $\varepsilon_{\rm RP}$ is given by a similar
343: expression, with one less quadratic term. 
344: It is now straightforward to derive the expressions for
345: $\mean{\varepsilon_{\rm part}}, \mean{\varepsilon^2_{\rm part}}, 
346: \mean{\varepsilon^4_{\rm part}}$, etc. We get
347: \begin{eqnarray}
348: \label{epsmoments}
349: \mean{\varepsilon_{\rm part}}&=& \varepsilon_{\rm s} 
350: + \mean{B_i\delta_i^{(2)}}\nonumber \\
351: \mean{\varepsilon_{\rm part}^2}&=& \varepsilon_{\rm s}^2
352: +\mean{(A_i\delta_i^{(1)})^2}+
353: 2\varepsilon_{\rm s}\mean{B_i\delta_i^{(2)}}\nonumber \\
354: \mean{\varepsilon_{\rm part}^4}&=& \varepsilon_{\rm s}^4
355: +6 \varepsilon_{\rm s}^2 \mean{(A_i\delta_i^{(1)})^2}
356: +4 \varepsilon_{\rm s}^3 \mean{B_i\delta_i^{(2)}}, 
357: \nonumber \\
358: \end{eqnarray}
359: where we have retained terms to second order in the $\delta$'s and used
360: $\mean{\delta_i^{(1)}}=0$; summation over the repeated index is
361: understood. 
362: 
363: The averages on the rhs of Eqs.~(\ref{epsmoments}) 
364: are easily computed by using the following identity, which holds
365: for $N$ independent participant nucleons 
366: \begin{equation}
367: \mean{\delta_f\delta_g} =
368: \frac{\mean{fg}-\mean{f}\mean{g}}{N},
369: \end{equation}
370: where $f$ and $g$ are any functions of $x$ and $y$. We get 
371: \begin{eqnarray}
372: \label{epscumul2}
373: \varepsilon\{2\}^2
374: &=&\varepsilon_{\rm s}^2+\cr
375: &&\frac{\mean{r^4}}{N\mean{r^2}^2}
376: \left(1+3 \varepsilon_{\rm s}^2+
377: 4\varepsilon_{\rm s}\frac{\mean{r^4 \cos 2\phi}}{\mean{r^4}}\right),
378: \end{eqnarray}
379: \begin{equation}
380: \label{epscumul4}
381: \varepsilon\{4\}^4
382: =\varepsilon_{\rm s}^4+
383: \frac{2\mean{r^4}}{N\mean{r^2}^2}
384: \left(\varepsilon_{\rm s}^4-
385: \varepsilon_{\rm s}^2\frac{\mean{r^4 \cos  4\phi}}{\mean{r^4}}\right),
386: \end{equation}
387: and 
388: \begin{equation}
389: \label{epsRP}
390: \mean{\varepsilon_{\rm RP}}=
391: \varepsilon_{\rm s}
392: +\frac{\mean{r^4}}{N\mean{r^2}^2}\left(\varepsilon_{\rm s}+
393: \frac{\mean{r^4 \cos 2\phi}}{\mean{r^4}}\right),
394: \end{equation}
395: where $(r,\phi)$ are the polar coordinates in the $(x,y)$ plane.
396: 
397: 
398: \begin{figure}
399: \includegraphics*[width=\linewidth]{Au_Npart.eps}
400: \includegraphics*[width=\linewidth]{Cu_Npart.eps}
401: \caption{Standard eccentricity $\varepsilon_{\rm s}$,
402: participant eccentricities 
403: $\varepsilon\{2\}$ and $\varepsilon\{4\}$, and reaction-plane 
404: eccentricity $\mean{\varepsilon_{\rm RP}}$, 
405: vs the number of participant nucleons for a Au-Au collision (top)
406: and a Cu-Cu collision (bottom).}
407: \label{fig:Au_Npart}
408: \end{figure}
409: 
410: 
411: The number of participants, $N$, and the moments of their distribution, 
412: $\mean{r^4}$, $\mean{r^4\cos2\phi}$, \dots, 
413: are easily computed numerically in the Glauber model with Woods-Saxon
414: distribution of participants. 
415: The parameters of our Glauber calculation 
416: are the same as in Ref.~\cite{Kharzeev:2000ph}, with one minor 
417: difference: 
418: we neglect fluctuations in the number of participants, i.e., we assume
419: a one-to-one relation between impact parameter and number of
420: participants. 
421: Orders of magnitude of the various coefficients involved are 
422: $\mean{r^4}/\mean{r^2}^2\sim 1.4-2$, 
423: $\mean{r^4\cos 2\phi}/\mean{r^4}\sim-\varepsilon_{\rm s}$, 
424: $\mean{r^4\cos 4\phi}/\mean{r^4}\sim \varepsilon_{\rm s}^2$. 
425: Numerical results for the various eccentricities as a function of 
426: the number of participants are given in Fig.~\ref{fig:Au_Npart} 
427: for Au-Au and Cu-Cu collisions.
428: 
429: The standard eccentricity $\varepsilon_{\rm s}$ vanishes for the most
430: central as well as for the most peripheral collisions as expected. 
431: On the other hand, $\varepsilon\{2\}$ rises steeply as 
432: $N$ decreases; this is due to the $1/N$ contribution of the 
433: fluctuations, see Eq.~(\ref{epscumul2}). 
434: This contribution originates from the 
435: $\delta_{y^2}-\delta_{x^2}$ and $\delta_{xy}^2$ terms in 
436: Eq.~(\ref{epspart2}).
437: The number of participants $N$ being smaller for the Cu-Cu collision than
438: for the Au-Au collision, the magnitude of the fluctuations is
439: relatively larger in the former case.
440: 
441: The 4-cumulant $\varepsilon\{4\}$ is only slightly larger than 
442: $\varepsilon_{\rm s}$, down to low values of $N$.
443: This is because the $1/N$ term in Eq.~(\ref{epscumul4}) 
444: is multiplied by a term of order $\varepsilon_{\rm s}^4$,
445: As a result, the {\it relative\/} difference between 
446: $\varepsilon\{4\}$ and $\varepsilon_{\rm s}$ is only of order
447: $1/N$. Similarly, the reaction plane eccentricity, 
448: $\mean{\varepsilon_{\rm RP}}$, is almost equal to 
449: $\varepsilon_{\rm s}$. 
450: 
451: We have considered so far the fluctuations in the eccentricity of the
452: overlap zone. We now briefly discuss fluctuations in the size or the
453: transverse area of the overlap zone, $S\sim 2\pi\sigma_x\sigma_y$.
454: They can be calculated exactly in the same way as fluctuations in 
455: the eccentricity. The result is 
456: \begin{equation}
457: \frac{\mean{S^2}-\mean{S}^2}{\mean{S}^2}=\frac{1}{4N}
458: \left(\frac{\mean{y^4}}{\mean{y^2}^2}+
459: \frac{\mean{x^4}}{\mean{x^2}^2}+
460: \frac{2\mean{x^2y^2}}{\mean{x^2}\mean{y^2}}-4\right).
461: \end{equation}
462: Numerically, these fluctuations are found to be practically
463: negligible. The reason why eccentricity fluctuations are important 
464: is that the $1/N$ term in Eq.~(\ref{epscumul2}) must be compared with 
465: $\varepsilon_{\rm s}^2$, which is itself a small number. 
466: 
467: 
468: 
469: \medskip
470: \noindent 5. Discussion
471: 
472: We have shown that the values of $v_2$ analyzed with different methods 
473: should be scaled by different eccentricities:
474: $v_2\{ {\rm EP2}\}$ (standard $v_2$), 
475: $v_2\{4\}$ and $v_2\{ {\rm ZDC}\}$ 
476: should be scaled respectively by $\varepsilon\{2\}$,
477: $\varepsilon\{4\}$ and $\mean{\varepsilon_{\rm RP}}$, 
478: defined in Eqs.~(\ref{v2std}), (\ref{v24}) and (\ref{v2ZDC}). 
479: 
480: Our most important result is that $\varepsilon\{4\}$ and 
481: $\mean{\varepsilon_{\rm RP}}$ are almost equal to the 
482: standard eccentricity, while $\varepsilon\{2\}$ is strongly
483: affected by fluctuations for small systems and/or peripheral
484: collisions. 
485: An important contribution to the fluctuations comes from 
486: the angle between the $x'$-axis and the 
487: $x$-axis in Fig.~\ref{fig:ellipse}, i.e., from the angle of tilt 
488: of the participant ellipse relative to the reaction plane. 
489: This effect was neglected in Ref.~\cite{Miller:2003kd} and 
490: first taken into account in Ref.~\cite{Manly:2005zy}. 
491: The ZDC analysis eliminates the effect of the tilt angle 
492: by measuring the eccentricity in the (x,y) axes; in the case of
493: the 4-cumulant analysis, most of the fluctuations happen to 
494: cancel in the subtraction of Eq.~(\ref{v24}).
495: 
496: Our results show that higher-order estimates 
497: of elliptic flow, $v_2\{4\}$ and $v_2\{\rm ZDC\}$, are not only 
498: insensitive to nonflow effects, but also, to a large extent, 
499: to fluctuations in the participant eccentricity.
500: This is confirmed by transport calculations~\cite{Zhu:2005qa}, and 
501: explains the observed agreement between $v_2\{4\}$ and 
502: $v_2\{\rm ZDC\}$ in Au-Au collisions~\cite{Wang:2005ab}. 
503: 
504: Fluctuations in the participant eccentricity, on the other hand,  
505: tend to increase the value of the standard, event-plane $v_2$. 
506: We have estimated this increase quantitatively, assuming 
507: independent nucleons. 
508: Within this simple model, fluctuations  account for less than 
509: one half of the observed difference between $v_2\{2\}$ 
510: and $v_2\{4\}$ in Au-Au collisions~\cite{Wang:2005ab}. 
511: The remaining difference can be (at least partly) 
512: ascribed to nonflow effects. Nonflow effects are 
513: clearly seen in the different $p_T$-dependences of $v_2\{2\}$ and 
514: $v_2\{4\}$, which cannot be explained by fluctuations. 
515: 
516: However, one should be aware that estimates of the 
517: participant eccentricity are model dependent:
518: in particular, Monte-Carlo Glauber 
519: calculations~\cite{Manly:2005zy,Voloshin:2006gz} generally 
520: yield higher fluctuations. 
521: The difference is the following: in a Monte-Carlo Glauber, 
522: each nucleon is modelled as black disk of transverse area $\sigma$, 
523: and a nucleon from nucleus A is a participant if it overlaps 
524: with at least a nucleon from nucleus B, and vice-versa.
525: Participant nucleons are therefore correlated, and the black-disk
526: approximation maximizes these correlations. 
527: Due to such correlations, which are neglected in our
528: calculation, our estimate of the participant eccentricity can be 
529: considered a lower bound. 
530: Quantitatively, correlations can increase the effect of fluctuations
531: by a factor up to 2: if the participant nucleons can only be found 
532: in pairs of overlapping disks, this amounts to replacing $N$ by 
533: $N/2$ (the number of pairs) in Eqs.~(\ref{epscumul2}-\ref{epsRP}). 
534: 
535: Finally, let us compare our results with the recent Monte-Carlo 
536: Glauber calculations of Ref.~\cite{Voloshin:2006gz}. 
537: The results are in qualitative agreement with ours: 
538: $\varepsilon\{4\}$ is much closer to $\varepsilon_{\rm s}$ 
539: than to $\varepsilon\{2\}$ for moderate centralities. 
540: For large impact parameters, however, $\varepsilon\{4\}$ is almost 
541: equal to $\varepsilon\{2\}$. This means that the participant 
542: eccentricity, although much larger than the standard eccentricity, 
543: fluctuates little from one event to the other. This intriguing
544: behaviour could be a consequence of the strong correlations 
545: mentioned above, and deserves further investigation. 
546: 
547: To summarize, the elliptic flow scaled by the eccentricity of the
548: overlap zone, $v_2/\varepsilon$, is an important observable at RHIC as
549: well as LHC, because if it is found to be independent of the system
550: size, one has a strong pointer toward thermalization. We have 
551: discussed various definitions of $v_2$ and $\varepsilon$ and studied how
552: they are affected by fluctuations in nucleon positions.
553: We have shown that when $v_2$ is analyzed using 4-particle cumulants or 
554: the event-plane from directed flow in a ZDC calorimeter, the resulting 
555: estimate essentially scales with the standard eccentricity, and is 
556: insensitive to the fluctuations in the participant eccentricity 
557: considered by PHOBOS~\cite{Manly:2005zy}. 
558: 
559: 
560: \section*{Acknowledgments}
561: 
562: We acknowledge the financial support from CEFIPRA, New Delhi, under
563: its project no. 3104-3. JYO thanks G. Wang, G. Roland, M. Nardi, 
564: S. Manly and A. Poskanzer for discussions.
565: 
566: \begin{thebibliography}{99}
567: 
568: 
569: %\cite{Ackermann:2000tr}
570: \bibitem{Ackermann:2000tr}
571:   K.~H.~Ackermann {\it et al.}  [STAR Collaboration],
572:   %``Elliptic flow in Au + Au collisions at s(N N)**(1/2) = 130-GeV,''
573:   Phys.\ Rev.\ Lett.\  {\bf 86}, 402 (2001).
574:   %%CITATION = NUCL-EX 0009011;%%
575: 
576: 
577: %\cite{Ollitrault:1992bk}
578: \bibitem{Ollitrault:1992bk}
579:   J.~Y.~Ollitrault,
580:   %``Anisotropy as a signature of transverse collective flow,''
581:   Phys.\ Rev.\ D {\bf 46}, 229 (1992).
582:   %%CITATION = PHRVA,D46,229;%%
583: 
584: %\cite{Masui:2005aa}
585: \bibitem{Masui:2005aa}
586:   H.~Masui  [PHENIX Collaboration],
587:   %``Anisotropic flow in s(NN)**(1/2) = 200-GeV Cu + Cu and Au + Au collisions
588:   %at PHENIX,''
589:   nucl-ex/0510018.
590:   %%CITATION = NUCL-EX 0510018;%%
591: 
592: %\cite{Manly:2005zy}
593: \bibitem{Manly:2005zy}
594:   S.~Manly {\it et al.}  [PHOBOS Collaboration],
595:   %``System size, energy and pseudorapidity dependence of directed and elliptic
596:   %flow at RHIC,''
597:   nucl-ex/0510031.
598:   %%CITATION = NUCL-EX 0510031;%%
599: 
600: %\cite{Wang:2005ab}
601: \bibitem{Wang:2005ab}
602:   G.~Wang  [STAR Collaboration],
603:   %``Anisotropic flow in Au Au and Cu Cu at 62-GeV and 200-GeV,''
604:   nucl-ex/0510034.
605:   %%CITATION = NUCL-EX 0510034;%%
606: 
607: %\cite{Voloshin:2006gz}
608: \bibitem{Voloshin:2006gz}
609:   S.~A.~Voloshin,
610:   %``Toward the energy and the system size dependece of elliptic flow: working
611:   %on flow fluctuations,''
612:   nucl-th/0606022.
613:   %%CITATION = NUCL-TH 0606022;%%
614: 
615: %\cite{Bhalerao:2005mm}
616: \bibitem{Bhalerao:2005mm}
617:   R.~S.~Bhalerao, J.~P.~Blaizot, N.~Borghini and J.~Y.~Ollitrault,
618:   %``Elliptic flow and incomplete equilibration at RHIC,''
619:   Phys.\ Lett.\ B {\bf 627}, 49 (2005).
620:   %%CITATION = NUCL-TH 0508009;%%
621: 
622: %\cite{Sorge:1998mk}
623: \bibitem{Sorge:1998mk}
624:   H.~Sorge,
625: %``Highly sensitive centrality dependence of elliptic flow: A novel  signature
626:   %of the phase transition in QCD,''
627:   Phys.\ Rev.\ Lett.\  {\bf 82}, 2048 (1999).
628:   %%CITATION = NUCL-TH 9812057;%%
629: 
630: %\cite{Socolowski:2004hw}
631: \bibitem{Socolowski:2004hw}
632:   O.~J.~Socolowski, F.~Grassi, Y.~Hama and T.~Kodama,
633:   % ``Fluctuations of the initial conditions and the continuous emission in
634:   %hydro description of two-pion interferometry,''
635:   Phys.\ Rev.\ Lett.\  {\bf 93}, 182301 (2004).
636:   %%CITATION = HEP-PH 0405181;%%
637: 
638: %\cite{Kolb:2003dz}
639: \bibitem{Kolb:2003dz}
640:   P.~F.~Kolb and U.~W.~Heinz,
641:   %``Hydrodynamic description of ultrarelativistic heavy-ion collisions,''
642:   nucl-th/0305084.
643:   %%CITATION = NUCL-TH 0305084;%%
644: 
645: %\cite{Kharzeev:2000ph}
646: \bibitem{Kharzeev:2000ph}
647:   D.~Kharzeev and M.~Nardi,
648:   %``Hadron production in nuclear collisions at RHIC and high density QCD,''
649:   Phys.\ Lett.\ B {\bf 507}, 121 (2001).
650:   %%CITATION = NUCL-TH 0012025;%%
651: 
652: %\cite{Drescher:2006pi}
653: \bibitem{Drescher:2006pi}
654:   T.~Hirano, U.~W.~Heinz, D.~Kharzeev, R.~Lacey and Y.~Nara,
655:   % ``Hadronic dissipative effects on elliptic flow in ultrarelativistic
656:   %heavy-ion collisions,''
657:   Phys.\ Lett.\ B {\bf 636}, 299 (2006);
658:   %%CITATION = NUCL-TH 0511046;%%
659:   A.~Kuhlman, U.~W.~Heinz and Y.~V.~Kovchegov,
660:   %``Gluon saturation effects in relativistic U + U collisions,''
661:   Phys.\ Lett.\ B {\bf 638}, 171 (2006);
662:   %%CITATION = NUCL-TH 0604038;%%
663:   H.~J.~Drescher, A.~Dumitru, A.~Hayashigaki and Y.~Nara, 
664:   %``The eccentricity in heavy-ion collisions from color glass condensate
665:   %initial conditions,''
666:   nucl-th/0605012.
667:   %%CITATION = NUCL-TH 0605012;%%
668: 
669: %\cite{Miller:2003kd}
670: \bibitem{Miller:2003kd}
671:   M.~Miller and R.~Snellings,
672:   %``Eccentricity fluctuations and its possible effect on elliptic flow
673:   %measurements,''
674:   nucl-ex/0312008.
675:   %%CITATION = NUCL-EX 0312008;%%
676: 
677: %\cite{Adler:2002pu}
678: \bibitem{Adler:2002pu}
679:   C.~Adler {\it et al.}  [STAR Collaboration],
680:   %``Elliptic flow from two- and four-particle correlations in Au + Au
681:   %collisions at s(NN)**(1/2) = 130-GeV,''
682:   Phys.\ Rev.\ C {\bf 66}, 034904 (2002).
683:   %%CITATION = NUCL-EX 0206001;%%
684: 
685: %\cite{Poskanzer:1998yz}
686: \bibitem{Poskanzer:1998yz}
687:   A.~M.~Poskanzer and S.~A.~Voloshin,
688:   %``Methods for analyzing anisotropic flow in relativistic nuclear
689:   %collisions,''
690:   Phys.\ Rev.\ C {\bf 58}, 1671 (1998);
691:   %%CITATION = NUCL-EX 9805001;%%
692:   J.~Y.~Ollitrault,
693:   %``On the measurement of azimuthal anisotropies in nucleus--nucleus
694:   %collisions,''
695: nucl-ex/9711003.
696:   %%CITATION = NUCL-EX 9711003;%%
697: 
698: %\cite{Back:2002gz}
699: \bibitem{Back:2002gz}
700:   B.~B.~Back {\it et al.}  [PHOBOS Collaboration],
701:   %``Pseudorapidity and centrality dependence of the collective flow of  charged
702:   %particles in Au + Au collisions at s(NN)**(1/2) = 130-GeV,''
703:   Phys.\ Rev.\ Lett.\  {\bf 89}, 222301 (2002).
704:   %%CITATION = NUCL-EX 0205021;%%
705: 
706: %\cite{Adler:2003kt}
707: \bibitem{Adler:2003kt}
708:   S.~S.~Adler {\it et al.}  [PHENIX Collaboration],
709:   %``Elliptic flow of identified hadrons in Au + Au collisions at  s(NN)**(1/2)
710:   %= 200-GeV,''
711:   Phys.\ Rev.\ Lett.\  {\bf 91}, 182301 (2003).
712:   %%CITATION = NUCL-EX 0305013;%%
713: 
714: %\cite{Adcox:2002ms}
715: \bibitem{Adcox:2002ms}
716:   K.~Adcox {\it et al.}  [PHENIX Collaboration],
717:   %``Flow measurements via two-particle azimuthal correlations in Au + Au
718:   %collisions at s(NN)**(1/2) = 130-GeV,''
719:   Phys.\ Rev.\ Lett.\  {\bf 89}, 212301 (2002).
720:   %%CITATION = NUCL-EX 0204005;%%
721: 
722: %\cite{Borghini:2000sa}
723: \bibitem{Borghini:2000sa}
724:   N.~Borghini, P.~M.~Dinh and J.~Y.~Ollitrault,
725:   %``A new method for measuring azimuthal distributions in nucleus nucleus
726:   %collisions,''
727:   Phys.\ Rev.\ C {\bf 63}, 054906 (2001).
728:   %%CITATION = NUCL-TH 0007063;%%
729: 
730: %\cite{Borghini:2001vi}
731: \bibitem{Borghini:2001vi}
732:   N.~Borghini, P.~M.~Dinh and J.~Y.~Ollitrault,
733:   %``Flow analysis from multiparticle azimuthal correlations,''
734:   Phys.\ Rev.\ C {\bf 64}, 054901 (2001).
735:   %%CITATION = NUCL-TH 0105040;%%
736: 
737: %\cite{Adams:2003zg}
738: \bibitem{Adams:2003zg}
739:   J.~Adams {\it et al.}  [STAR Collaboration],
740:   %``Azimuthal anisotropy at RHIC: The first and fourth harmonics,''
741:   Phys.\ Rev.\ Lett.\  {\bf 92}, 062301 (2004).
742:   %%CITATION = NUCL-EX 0310029;%%
743: 
744: %\cite{Zhu:2005qa}
745: \bibitem{Zhu:2005qa}
746:   X.~l.~Zhu, M.~Bleicher and H.~Stoecker,
747:   %``Elliptic flow analysis at RHIC: Fluctuations vs. non-flow effects,''
748:   Phys.\ Rev.\ C {\bf 72}, 064911 (2005).
749:   %%CITATION = NUCL-TH 0509081;%%
750: 
751: \end{thebibliography}
752: 
753: 
754: \end{document} 
755: 
756: 
757: