nucl-th0602013/3pi.tex
1: %#!latex 3pi-2PTP.tex ; dvips -h duplex.ps -o 3pi-2PTP.ps -Ppdf 3pi-2PTP
2: %%#!latex 3pi-2PTP.tex ; dvipdfm -o 3pi-2PTP.pdf -p a4 -t 3pi-2PTP
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: %%%%%%   template.tex for PTPTeX.cls <ver.0.9>  %%%%%
5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6: \documentclass[seceq]{ptptex}
7: %\documentclass[letter]{ptptex}
8: %\documentclass[seceq,supplement]{ptptex}
9: %\documentclass[seceq,addenda]{ptptex}
10: %\documentclass[seceq,errata]{ptptex}
11: %\documentclass[seceq,preprint]{ptptex}
12: 
13: \usepackage{graphicx}
14: %\usepackage{wrapft}
15: 
16: %%%%% Personal Macros %%%%%%%%%%%%%%%%%%%
17: 
18: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
19: 
20: %\pubinfo{Vol.~11X, No.~X, Mmmmm YYYY}%Editorial Office will fill in this.
21: %\setcounter{page}{}                  %Editorial Office will fill in this.
22: %\def\ptype{p}                        %Editorial Office will fill in this.
23: %\def\ptpsubject{}                    %Editorial Office will fill in this.
24: %\def\pageinfo{X-X}                   %Editorial Office will fill in this.
25: %-------------------------------------------------------------------------
26: %\nofigureboxrule                     %to eliminate the rule of \figurebox
27: \notypesetlogo                       %comment in if to eliminate PTPTeX 
28: %---- When [preprint] you can put preprint number at top right corner.
29: %\preprintnumber[3cm]{%<-- [..]: optional width of preprint # column.
30: %KUNS-1325\\PTPTeX ver.0.8\\ August, 1997}
31: %-------------------------------------------------------------------------
32: 
33: \markboth{%     %running head for odd-page (authors' name)
34: K. Morita, S. Muroya and H. Nakamura
35: }{%             %running head for even-page (`short' title)
36: Multiplicity Dependence of Partially Coherent Pion Production
37: }
38: 
39: \title{%        %You can use \\ for explicit line-break
40: Multiplicity Dependence of Partially Coherent Pion Production\\
41: in Relativistic Heavy Ion Collisions
42: }
43: 
44: %\subtitle{Subtitle}    %use this when you want a subtitle
45: 
46: \author{%       %Use \scshape  for the family name
47: Kenji
48: \textsc{Morita}$^{1,}$\footnote{E-mail:~morita@hep.phys.waseda.ac.jp},%
49: Shin
50: \textsc{Muroya}$^{2,}$\footnote{E-mail:~muroya@yukawa.kyoto-u.ac.jp,  Present Address:~Matsumoto University, Matsumoto 390-1295, Japan},
51: and Hiroki \textsc{Nakamura}$^{1,}$\footnote{E-mail:~naka@hep.phys.waseda.ac.jp}
52: }
53: 
54: \inst{%         %Affiliation, neglected when [addenda] or [errata]
55: $^1$Department of Physics, Waseda University, Tokyo 169-8555,
56: Japan\\
57: $^2$Tokuyama University, Shunan, 745-8511,
58: Japan
59: }
60: 
61: %\publishedin{%         %Write this ONLY in cases of addenda and errata
62: %Prog.~Theor.~Phys.\ \textbf{XX} (19YY), page.}
63: 
64: \recdate{February 4, 2006}%            %Editorial Office will fill in
65: %this.
66: 
67: 
68: \abst{%         %this abstract is neglected when [addenda] or [errata]
69:  We investigate two- and three-particle intensity correlation functions
70:  of pions in relativistic heavy ion collisions for different colliding
71:  energies. Based on three models of particle production, we analyze the
72:  degree to which the pion sources are chaotic in the SPS S+Pb, Pb+Pb and
73:  the RHIC Au+Au
74:  collisions. The ``chaoticity'', $\lambda$, of the two-particle correlation
75:  functions is corrected for long-lived resonance decays. The effect of the
76:  partial Coulomb correction is also examined. Although the partially
77:  coherent model gives a result which is consistent with that of STAR,
78:  the chaotic fraction does not exhibit clear multiplicity dependence if we
79:  take into account both the corrected chaoticity and the weight factor of
80:  the three-pion correlation function.
81:  The result of the partially-multicoherent model
82:  indicates an increasing number of coherent sources in higher multiplicity
83:  events.
84: }
85: 
86: \begin{document}
87: 
88: \maketitle
89: 
90: \section{Introduction}\label{Sec:intro}
91: Pion interferometry has been regarded as an indispensable tool in
92: relativistic heavy ion physics. Two-particle intensity interferometry
93: can be used to determine the sizes of the collision system. This fact is
94: known as the Hanbury Brown-Twiss (HBT) effect. For this reason, it has
95: been used for exploring the space-time
96: evolution of hot, dense matter created in heavy ion collisions
97: \cite{Tomasik_qgp3}. In particular, the most recent experiment at the
98: Relativistic Heavy Ion Collider (RHIC) obtained an interesting result
99: which is referred to the ``HBT puzzle''
100: \cite{Heinz_NPA702,Morita_PTP111}. Hydrodynamical models have failed to
101: reproduce the experimental results of two-particle correlation
102: functions. Hydrodynamical models are based on the assumption of local
103: thermal equilibrium. For the early stage of the space-time evolution,
104: the validity of this assumption is indirectly verified by the
105: observation of the large
106: elliptic flow at RHIC \cite{Kolb_PLB500}. However, equilibration in the
107: final hadronic stage is still ambiguous. Although an exponential particle
108: spectrum is expected for a thermal source, it does not require a system
109: that has reached local thermal equlibrium \cite{Rischke_NPA698}.
110: 
111: The two-particle intensity correlation function
112: $C_2(\boldsymbol{p_1},\boldsymbol{p_2})$ of identical particles provides
113: information not only
114: on source sizes but also on the state of the source through the chaoticity,
115: $\lambda=C_2(\boldsymbol{p},\boldsymbol{p})-1$. The HBT effect does not
116: exist if the source
117: is coherent. The chaoticity $\lambda$ is unity for a
118: perfectly chaotic source and 0 for a
119: coherent source, which does not induce the HBT effect. In
120: multi-particle production phenomena of relativistic heavy ion
121: collisions, the thermalized source can be chaotic and non-thermal
122: components can be coherent. For example, if disoriented chiral
123: condensate domains are created, such domains can decay into coherent
124: pions. Thus the final state pions may carry information regarding how partons
125: hadronize if the deconfined phase has been created. Therefore, the chaoticity
126: $\lambda$ is a very important
127: quantity to investigate the final state in heavy ion collisions. 
128: 
129: However, $\lambda$ cannot be regarded as the true chaoticity in real
130: (experimental) situations, because many other effects, such as long-lived
131: resonance decay contributions
132: \cite{Gyulassy_PLB217,Bolz_PRD47,Heiselberg_PLB,Csorgo_ZPHYS}, Coulomb
133: repulsions, and particle contaminations \cite{CERES_NPA714}
134: affect the chaoticity $\lambda$. As an alternative tool, the
135: three-particle correlation function has been proposed
136: \cite{Biyajima_PTP84,Heinz_PRC56,Nakamura_PRC60,Nakamura_PRC61}. 
137: Three-particle
138: correlations are more useful for this purpose, because long-lived resonances
139: do not affect the normalized three-pion correlator,
140: \begin{align}
141:  r_3&(\boldsymbol{p}_1,\boldsymbol{p}_2,\boldsymbol{p}_3)\nonumber\\
142: &=\frac{[C_3(\boldsymbol{p}_1,\boldsymbol{p}_2,\boldsymbol{p}_3)-1]
143: -[C_2(\boldsymbol{p}_1,\boldsymbol{p}_2)-1]
144:   -[C_2(\boldsymbol{p}_2,\boldsymbol{p}_3)-1]
145: -[C_2(\boldsymbol{p}_3,\boldsymbol{p}_1)-1]}
146:   {\sqrt{[C_2(\boldsymbol{p}_1,\boldsymbol{p}_2)-1]
147: [C_2(\boldsymbol{p}_2,\boldsymbol{p}_3)-1]
148: [C_2(\boldsymbol{p}_3,\boldsymbol{p}_1)-1]}}, \label{eq:r3-1}
149: \end{align}
150: with $C_3(\boldsymbol{p_1},\boldsymbol{p_2},\boldsymbol{p_3})$ being the
151: three-particle correlation function \cite{Humanic_PRC60}. The
152: correlation functions are defined as
153: \begin{equation}
154:  C_2(\boldsymbol{p_1},\boldsymbol{p_2}) = \frac{W_2(\boldsymbol{p_1},\boldsymbol{p_2})}{W_1(\boldsymbol{p_1})W_1(\boldsymbol{p_2})}
155: \end{equation}
156: and
157: \begin{equation}
158:  C_3(\boldsymbol{p_1},\boldsymbol{p_2},\boldsymbol{p_3}) =
159:   \frac{W_3(\boldsymbol{p_1},\boldsymbol{p_2},\boldsymbol{p_3})}
160:   {W_1(\boldsymbol{p_1})W_1(\boldsymbol{p_2})W_1(\boldsymbol{p_3})}
161: \end{equation}
162: with $W_n(\boldsymbol{p_1},\cdots ,\boldsymbol{p_n})$ being the $n$ particle distribution.
163: The index of the source chaoticity in the three-pion correlator is the weight
164: factor
165: \begin{equation}
166:  \omega=r_3(\boldsymbol{p},\boldsymbol{p},\boldsymbol{p})/2 \label{eq:omega-r3}
167: \end{equation}
168: which is unity for a chaotic source. Due to insufficient statistics, the
169: three-pion correlator measured in experiments to this time is slightly
170: different from Eq.~\eqref{eq:r3-1},
171: \begin{equation}
172:  r_3(Q_3)=\frac{[C_3(Q_3)-1]-[C_2(Q_{12})-1]-[C_2(Q_{23})-1]-[C_2(Q_{31})-1]}
173:   {\sqrt{[C_2(Q_{12})-1][C_2(Q_{23})-1][C_2(Q_{31})-1]}}, \label{eq:r3}
174: \end{equation}
175: where $Q_{ij}=\sqrt{-(p_i-p_j)^2}$ and $Q_3=\sqrt{Q_{12}^2+Q_{23}^2+Q_{31}^2}$.
176: The weight factor is defined as $\omega=r_3(0)/2$, so that it is
177: expected to be the same as the definition Eq.~\eqref{eq:omega-r3}. 
178: 
179: As shown in a previous work \cite{Nakamura_PRC66}, more precise
180: information concerning the source can be extracted using model analyses
181: combining
182: two- and three-particle correlations. In this paper, we analyze the
183: two-pion and three-pion correlations and investigate the chaoticity 
184: of the pion sources. Firstly, assuming that the main background
185: contribution is a long-lived
186: resonance decay and other effects are successfully removed in the
187: experimental data, we make a correction to account for the long-lived
188: resonance decays to the chaoticity of the two-pion correlation
189: functions, with the help of a statistical model. Secondly, we extract the
190: weight factor $\omega$ through the  simultaneous construction of $C_2$
191: and $C_3$.
192: Finally, we carry out model analyses using the ``true'' chaoticity
193: $\lambda^{\text{true}}$ after applying the resonance correction and the
194: weight factor
195: $\omega$. 
196: In addition to the analysis of Au+Au collisions at the RHIC given in a
197: previous paper \cite{Morita_3pirhic}, we give analyses on lower
198: colliding energy collisions at the CERN Super Proton Synchrotron (SPS)
199: based on the same method. We treat data from 200$A$ GeV (laboratory
200: system) S+Pb
201: collisions measured by the NA44 collaboration
202: \cite{NA44_2pi_PLB302,NA44_2pi_ZPHYS,NA44_3pi_SPb} and data from 158$A$ GeV
203: (lab. sys) Pb+Pb collisions measured by the NA44 collaboration
204: \cite{NA44_2pi_PbPb,NA44_3pi_pbpb} and the WA98 collaboration
205: \cite{WA98_3pi_prl,WA98_pbpb_prc}. On the basis of the results obtained
206: from these
207: experimental data, we investigate the multiplicity dependence of the
208: extracted model parameters. 
209: 
210:  In the next section, we explain the three models used
211: in this paper. The correction to account for the long-lived resonance
212: decays is discussed \S3. In \S4, we will give a combined analysis of the
213: 2$\pi$ and 3$\pi$ correlation functions, with the goal of extracting the weight
214: factor $\omega$. The results of the model analyses are given in \S5.
215: The paper is summarized in \S6.
216: 
217: \section{Model description}\label{Sec:model}
218: 
219: We use three kinds of models in this analysis. In all the models,
220: $\lambda^{\text{true}}$ and $\omega$ are used as inputs to fix the model
221: parameters. The output
222: quantities are the chaotic fraction and the
223: mean number of coherent sources. Here we do not discuss the origin of
224: the coherences in these models. Figure \ref{fig:models} presents a
225: schematic depiction of the models. 
226: 
227: \begin{figure}[!b]
228:  \begin{center}
229:   \includegraphics{models.eps}
230:  \end{center}
231:  \caption{\label{fig:models}Schematic depiction of Models I--III. The
232:  shaded area
233:  denote the chaotic sources. The unshaded area and the small circles
234:  represent the coherent sources. (a) Model I, (b) Model II, (c) Model
235:  III.}
236: \end{figure}
237: 
238: One is a partially coherent model \cite{Heinz_PRC56} whose only
239: parameter is the chaotic fraction, $\varepsilon_{\text{I}}$, defined as the
240: ratio of the number of particles coming from the chaotic source to the total
241: number of particles. In this model, pions are emitted
242: from a mixture of a chaotic and a coherent source. In general, 
243: we need to fix the source function in a manner that depends on
244: the momentum. In the present
245: analysis, fortunately, the true chaoticity $\lambda^{\text{true}}$ and the
246: weight factor $\omega$ are given by the correlation functions at
247: vanishing relative momenta, and hence they can be expressed in terms of
248: the chaotic fraction.
249: The relations among
250: $\varepsilon_{\text{I}}$, $\lambda^{\text{true}}$ and $\omega$ are
251: \begin{equation}
252:  \lambda^{\text{true}} = \varepsilon_{\text{I}}(2-\varepsilon_{\text{I}}), 
253: \quad
254:  \omega = \sqrt{\varepsilon_{\text{I}}}
255:  \frac{3-2\varepsilon_{\text{I}}}{(2-\varepsilon_{\text{I}})^{3/2}}.
256:  \label{eq:modelI}
257: \end{equation}
258: We refer to this model as Model I.
259: 
260: The second model, referred to as Model II, is a multicoherent source model
261: \cite{Nakamura_PRC61}. The parameter in this model is the mean
262: number of coherent sources, $\alpha_{\text{II}}$, which obeys the Poisson
263: distribution. Because small coherent sources are independent, a
264: chaotic source is realized as a cluster of an infinite number of coherent
265: sources. The parameter $\alpha_{\text{II}}$ is thus related to
266: $\lambda^{\text{true}}$
267: and $\omega$ as
268: \begin{equation}
269:  \lambda^{\text{true}} = \frac{\alpha_{\text{II}}}{\alpha_{\text{II}}+1},
270:   \quad
271:  \omega = \frac{1}{2}\frac{2\alpha_{\text{II}}^2+2\alpha_{\text{II}}+3}
272:  {\alpha_{\text{II}}^2+3\alpha_{\text{II}}+1}
273:  \sqrt{\frac{\alpha_{\text{II}}+1}{\alpha_{\text{II}}}} \label{eq:modelII}.
274: \end{equation}
275: As shown in Fig.~2 of Ref.~\citen{Nakamura_PRC61}, $\omega$ as a function
276: of $\alpha_{\text{II}}$ has a minimum value $\omega_{\text{min}}\simeq
277: 0.82$. Hence, there is no
278: corresponding $\alpha_{\text{II}}$ for $\omega$ smaller than
279: $\omega_{\text{min}}$.
280: 
281: As each of these models possesses only a single parameter, the
282: parameters calculated from $\lambda^{\text{true}}$ and $\omega$ should
283: give the same value if all corrections are correctly made. 
284: 
285: Model III is a ``partially multicoherent'' source model
286: \cite{Nakamura_PRC61}. This model is a mixture of Model I and Model II,
287: and its parameters are the chaotic fraction $\varepsilon_{\text{III}}$
288: and the mean number
289: of coherent sources $\alpha_{\text{III}}$, which are related to
290: $\lambda^{\text{true}}$ and $\omega$ as
291: \begin{align}
292:  \lambda^{\text{true}} &=
293:  \frac{\alpha_{\text{III}}}{\alpha_{\text{III}}
294:  +(1-\varepsilon_{\text{III}})^2},\label{l-pm}
295:  \\
296:  \omega &= \frac{2\alpha_{\text{III}}^2 + 2\alpha_{\text{III}}
297:  (1-\varepsilon_{\text{III}})^2 + 3(1-\varepsilon_{\text{III}})^3
298:  (1-2\varepsilon_{\text{III}})}
299:  {2[\alpha_{\text{III}}^2  + 3\alpha_{\text{III}}
300:  (1-\varepsilon_{\text{III}})^2  +(1-\varepsilon_{\text{III}})^3]} 
301: %\nonumber\\
302: %&\times
303:  \sqrt{\frac{\alpha_{\text{III}}+(1-\varepsilon_{\text{III}})^2}
304:  {\alpha_{\text{III}}}}\label{ome-pm}.
305: \end{align}
306: In this model, there are two parameters that correspond to two
307: experimentally measurable quantities, $\lambda^{\text{true}}$ and
308: $\omega$. Below, we investigate the allowed parameter regions for given sets
309: of $\lambda^{\text{true}}$ and $\omega$.
310: 
311: \section{Extracting $\lambda^{\text{true}}$ from two-pion correlations}\label{sec:lambda}
312: 
313: In relativistic heavy ion collisions, a non-negligible fraction of pions
314: comes from the decay of long-lived resonances. Recent analyses based on
315: statistical models show that hadrons are chemically frozen near the
316: critical temperature $T_c$ \cite{Braun-Munziger_PLB518}. Short-lived
317: resonances, such as $\rho$ and $\Delta$, decay before hadrons reach
318: the kinetic freeze-out or soon after the freeze-out, but long-lived
319: resonances, such as hyperons, can decay long after the kinetic freeze-out of
320: pions. In the two-pion correlation function, these long-lived resonances
321: appear as a spike near $q\sim 0$, whose width is too small to be resolved
322: with the current experimental resolutions, $\Delta q \simeq 5-10$ MeV. Thus,
323: chaoticities $\lambda^{\text{exp}}$ measured experimentally 
324: are smaller than the true chaoticities due to long-lived resonance decays
325: \cite{Bolz_PRD47,Csorgo_ZPHYS,Wiedemann_PRC56}. Following
326: Ref.~\citen{Wiedemann_PRC56}, we take into account resonances up to
327: $\Sigma^*(1385)$. We treat resonances whose widths are less
328: than 5 MeV as long-lived ones, i.e., 
329: $K_s^0, \eta, \eta', \phi,\Lambda, \Sigma$ and $\Xi$ are considered
330: long-lived resonances in the
331: calculation. Though $\omega$ mesons have an
332: intermediate width, it is known that they distort the shape of the
333: correlation function at low $q$ but do not reduce the
334: chaoticity. \cite{Bolz_PRD47}. 
335: 
336: \begin{table}[t!]
337:  \caption{\label{tbl:t-mu_SPS}Thermodynamic parameters obtained from the
338:  particle ratio in nucleus-nucleus collisions at SPS. Here, $h^-$ represents
339:  the negatively charged hadrons.}
340:  \begin{center}
341:   \begin{tabular}{cccccc}\hline
342:    System &  Ratio & Data [Reference] & $T$ [MeV]& $\mu_{\text{B}}$ [MeV] & $\chi^2/N_{\text{dof}}$\\ \hline
343:    SPS S+Pb & $p/\pi^+$ & 0.18$\pm$0.03
344:    \cite{Murray_NA44_NPA566,Gazdzicki_NA35_NPA590} & 173 $\pm$ 2 & 196
345:    $\pm$ 2 &
346:    36/7 \\
347:    & $\bar{p}/p$ & 0.12$\pm$0.02 \cite{Jacak_HDNM} & & & \\ 
348:    & $\bar{p}/\pi^-$ & 0.024$\pm$0.009
349:    \cite{Murray_NA44_NPA566,Gazdzicki_NA35_NPA590} & & & \\
350:    & $(K^-+K^+)/2K_s^0$ & 1.07$\pm$0.03 \cite{DiBari_NPA590}& & & \\
351:    & $K^+/K^-$ & 1.67$\pm$0.15 \cite{DiBari_NPA590} & & & \\
352:    & $K_s^0/\Lambda$ & 1.4$\pm$0.1 \cite{DiBari_NPA590} & & & \\
353:    & $K_s^0/\bar{\Lambda}$ & 6.4$\pm$0.4 \cite{DiBari_NPA590}& & & \\ 
354:    & $\bar{\Lambda}/\Lambda$ & 0.20$\pm$0.01 \cite{Abatzis_NPA566}& & & \\
355:    & $\Xi^+/\bar{\Lambda}$& 0.21$\pm$0.02 \cite{Abatzis_NPA566}& & & \\
356: 
357:    \hline
358:    SPS Pb+Pb & $\bar{p}/p$ & 0.085$\pm$0.009 \cite{Jones_NPA610} & 161
359:    $\pm$ 4 & 223 $\pm$ 7 & 44/9\\
360:    & $K_s^0/\pi^-$ & 0.125$\pm$0.019 \cite{Rohrich_Hirschegg97}& & & \\
361:    & $K_s^0/h^-$ & 0.123$\pm$0.020 \cite{WA97_PLB499}& & & \\
362:    & $\Lambda/h^-$ & 0.077$\pm$0.011 \cite{WA97_PLB499}& & & \\
363:    & $\Lambda/K_s^0$ & 0.63$\pm$0.08 \cite{WA97_PLB499}& & & \\
364:    & $\bar{\Lambda}/\Lambda$ & 0.131$\pm$0.017 \cite{WA97_PLB499}& & & \\
365:    & $\Xi^-/\Lambda$ & 0.110$\pm$0.010 \cite{WA97_PLB499}& & & \\
366:    & $\Xi/\bar{\Lambda}$ & 0.188$\pm$0.039 \cite{NA49_JPhys25}& & & \\
367:    & $(\Xi+\bar{\Xi})/(\Lambda+\bar{\Lambda})$ & 0.13$\pm$0.03
368:    \cite{NA49_JPhys25_189}& & & \\
369:    & $\Xi^+/\Xi^-$ & 0.232$\pm$0.033 \cite{NA49_JPhys25}& & & \\
370:    & $K^+/K^-$ & 1.85$\pm$0.09 \cite{Kaneta_NPA638}& & & \\\hline
371:   \end{tabular}
372:  \end{center}
373: \end{table}
374: 
375: For a chaotic source, the reduced chaoticity $\lambda^{\text{eff}}$ is
376: given in terms of the
377: ratio of the number of pions from the long-lived resonances to the total
378: number of pions as
379: \begin{equation}
380:  \sqrt{\lambda^{\text{eff}}} = 1 - \frac{N_\pi^{\text{r}}}{N_\pi},
381:   \label{eq:lambdaeff}
382: \end{equation}
383: where $N_\pi$ is the total number of emitted pions, and $N_\pi^{\text{r}}$ is
384: the number of pions from the decay of long-lived resonances
385: \cite{Csorgo_ZPHYS}.
386: We calculate this ratio with the help of the statistical model. For
387: midrapidity, the particle ratio can be written as the ratio of the number
388: densities, i.e.,
389: \begin{equation}
390:  \frac{N_i}{N_j} = \frac{n_i^0}{n_j^0}, \label{eq:ratio}
391: \end{equation}
392: where 
393: \begin{equation}
394:  n_i^0 = \frac{g_i}{2\pi^2}\int_{0}^{\infty}dp \, p^2 
395:   f(E,T,\mu_{\text{B}},\mu_{\text{S}},\mu_{\text{I$_3$}}),
396:   \label{eq:density}
397: \end{equation}
398: with $f(E,T,\mu_j,\cdots)$ being the equilibrium distribution function and
399: $g_i$ being the number of the degree of freedom of particle $i$
400: \cite{Cleymans_PRC60}.
401: For simplicity, we fix the chemical potential of the third component
402: of the isospin as $\mu_{I_3} = 0$. Then, the thermodynamic
403: parameters to be determined are the temperature $T$ and the baryon number
404: chemical potential $\mu_{\text{B}}$, because the strangeness chemical potential
405: $\mu_{S}$ is determined from the strangeness neutrality
406: condition. Results for several collision systems obtained from the
407: $\chi^2$ fit are shown
408: in Tables \ref{tbl:t-mu_SPS} and \ref{tbl:t-mu_RHIC}.\footnote{
409: 	 Though experimental data can contain contributions from
410: 	 coherent sources, we assume that particle ratios are not
411: 	 affected by the existence of coherent sources.}
412: 
413: \begin{table}[t!]
414:  \caption{\label{tbl:t-mu_RHIC}Thermodynamic parameters obtained from
415:  the particle ratio in Au+Au collisions at RHIC.}
416:  \begin{center}
417:   \begin{tabular}{cccccc}\hline
418:    System &  Ratio & Data & $T$ [MeV]& $\mu_{\text{B}}$ [MeV] & $\chi^2/N_{\text{dof}}$\\ \hline
419:    RHIC Au+Au & $\bar{p}/p$ & 0.71$\pm$0.05
420:    \cite{STAR_PbarPratio130}& 158 $\pm$ 9 & 36 $\pm$
421:    6 & 2.4/5\\
422:    & $\bar{p}/{\pi^-}$ & 0.072$\pm$0.014 \cite{STAR_pbar,STAR_PLB567}& &
423:    & \\
424:    & $K^-/\pi^-$ & 0.146$\pm$0.02 \cite{STAR_PLB595}& & & \\
425:    & $K^{*0}/h^-$ & 0.042$\pm$0.014 \cite{STAR_PRC66}& & & \\
426:    & $\bar{K}^{*0}/K^{*0}$ & 0.92$\pm$0.14 \cite{STAR_PRC66}& & & \\
427:    & $\bar{\Lambda}/\Lambda$ & 0.71$\pm$0.05 \cite{STAR_PLB567}& & & \\
428:    & $\bar{\Xi}/\Xi$ & 0.83$\pm$0.09 \cite{STAR_PRC66}& & & \\\hline
429:   \end{tabular}
430:  \end{center}
431: \end{table}
432: 
433: From Tables \ref{tbl:t-mu_SPS} and \ref{tbl:t-mu_RHIC}, we find
434: that the quality of the fit becomes better as the collsion
435: energy increases. It is reasonable to assume that the
436: system has reached thermal equilibrium and that the pion source becomes
437: chaotic as the collision energy increases. Below we see
438: whether this naive assumption is valid.
439: 
440: Since this particle ratio is obtained from the particle numbers
441: integrated with the particle momenta, $\lambda^{\text{eff}}$ is
442: calculated from Eqs.~\eqref{eq:lambdaeff}--\eqref{eq:density} using the
443: integrated particle numbers.
444: The measured chaoticity $\lambda^{\text{exp}}$, however, depends on
445: the average momentum of pion pairs. For example, STAR has measured
446: $\lambda^{\text{exp}}$ for three bins of the transverse momentum ($k_t$)
447: \cite{STAR_PRL87}. In this paper, we assume, for simplicity, that
448: the true chaoticity $\lambda^{\text{true}}$ does not depend on the particle
449: momentum.\footnote{Note that this is also assumed in Model I--III.} 
450: Therefore, we need to average $\lambda^{\text{exp}}$ over the momenta in order
451: to evaluate $\lambda^{\text{true}}$. Assuming the $k_t$ dependence of
452: the measured
453: $\lambda^{\text{exp}}$ is dominated by long-lived resonances
454: [i.e., Eq.~\eqref{eq:lambdaeff}], we obtain the averaged
455: chaoticity $\overline{\lambda}^{\text{exp}}$ as
456: \begin{equation}
457:   \overline{\lambda}^{\text{exp}} 
458:    = \frac{\displaystyle \sum_{i=1}^{n} \lambda^{\text{exp}}_i
459:   \int_{i\text{-th bin}} k_t dk_t \left(\frac{dN}{k_t dk_t}\right)^2 }
460:   {\displaystyle \sum_{i=1}^{n} \int_{i\text{-th bin}} 
461:   k_t dk_t \left(\frac{dN}{k_t dk_t}\right)^2},
462:   \label{eq:lambdaav}
463: \end{equation}
464: with $\lambda_i^{\text{exp}}$ being the measured chaoticity in the $i$-th
465: $k_t$ bin. [See the appendix of Ref.~\citen{Morita_3pirhic} for the derivation
466: of Eq.~\eqref{eq:lambdaav}.] The transverse momentum distribution $dN/k_t
467: dk_t$, in the above equation is taken from experimental results. For S+Pb
468: collisions, in
469: which three-particle correlation data are available only for minimum-bias
470: data, we calculate $\overline{\lambda}^{\text{exp}}$ by simply averaging
471: the 
472: $\lambda^{\text{exp}}$ with different multiplicities. This should not affect
473: our conclusion, because the data exhibit little multiplicity dependence
474: \cite{NA44_2pi_ZPHYS}.
475: For $\lambda^{\text{exp}}$, it is known
476: that the value of $C_2(\boldsymbol{q})-1$ at $\boldsymbol{q}=0$ depends on the
477: dimension of the projection onto
478: $\boldsymbol{q}$-space. In general, the value of $\lambda^{\text{exp}}$
479: obtained from the
480: 1-dimensional Gaussian fitting differs from that obtained from the
481: 3-dimensional Gaussian fitting because of the projection,
482: experimental resolution, and other effects, though these should be the
483: same for an ideal measurement. In this paper, we use the value extracted
484: from the 3-dimensional Gaussian fitting,
485: \begin{equation}
486:  C_2^{\text{fit}}(\boldsymbol{q}) = 1 + \lambda^{\text{exp}}
487:   \exp\left( -R_{\text{side}}^2 q_{\text{side}}^2
488:       -R_{\text{out}}^2 q_{\text{out}}^2
489:       -R_{\text{long}}^2 q_{\text{long}}^2 \right).
490:   \label{eq:c2fit}
491: \end{equation}
492: as $\lambda^{\text{exp}}$.
493: 
494: The true chaoticity
495: is then given by $\lambda^{\text{true}} =
496: \overline{\lambda}^{\text{exp}}/\lambda^{\text{eff}}$. 
497: Results for $\lambda^{\text{true}}$ obtained from various systems are
498: summarized in Table~\ref{tbl:lambdatrue}. The error on
499: $\lambda^{\text{true}}$ is the sum of the experimental one on
500: $\overline{\lambda}^{\text{exp}}$, calculated from the errors on
501: $\lambda_i^{\text{exp}}$ and $dN/k_tdk_t$, and the errors
502: propagated from the fit of the thermodynamic quatities at the 1-$\sigma$ level,
503: shown in Tables \ref{tbl:t-mu_SPS} and \ref{tbl:t-mu_RHIC}.
504: 
505: \begin{table}[t!]
506:  \caption{\label{tbl:lambdatrue}Summary of $\lambda^{\text{true}}$}
507:  \begin{center}
508:   \begin{tabular}{ccccc}\hline
509:     System & Experiment [Reference] & $\overline{\lambda}^{\text{exp}}$
510:    & $\lambda^{\text{true}}$ & $\lambda^{\text{true}}_{\text{pc}}$\\ \hline
511:    SPS S+Pb & NA44, min.bias \cite{NA44_2pi_ZPHYS} & 0.585 $\pm$ 0.06
512:    &  0.94 $\pm$ 0.06 & 0.7$\lambda^{\text{true}}$\\
513:    SPS Pb+Pb & NA44 Central \cite{NA44_2pi_PbPb} & 0.55 $\pm$ 0.03 &
514:    0.98 $\pm$ 0.03 & 0.8$\lambda^{\text{true}}$\\
515:    SPS Pb+Pb & WA98 Central \cite{WA98_pbpb_prc} & 0.58 $\pm$ 0.04 & 
516:    1.03 $\pm$ 0.04 & 0.8$\lambda^{\text{true}}$\\
517:    RHIC Au+Au & STAR Central \cite{STAR_PRL87} & 0.57 $\pm$ 0.06 &
518:    0.93 $\pm$ 0.08 & 0.8$\lambda^{\text{true}}$ \\\hline
519:   \end{tabular}
520:  \end{center}
521: \end{table}
522: 
523: From Table \ref{tbl:lambdatrue}, we see that the value of
524: $\lambda^{\text{true}}$
525: are not so different and become close to unity in all systems. However, it
526: should be noted that there may be an overestimation in $\lambda^{\text{exp}}$
527: due to an overcorrection for the Coulomb interaction. 
528: Because the two-pion correlation function is
529: affected by the Coulomb interaction between the two detected pions,
530: there are many issues concerning how the Coulomb interaction can be subtracted
531: from the observed correlation function
532: \cite{Pratt_PRD33,Bowler_PLB270,Biyajima_PLB353,Sinyukov_PLB432}.
533: Recently, it has been shown that
534: a new procedure called the \textit{partial} Coulomb correction, leads to
535: significant correction to the source sizes
536: \cite{PHENIX_HBT200PRL,CERES_NPA714}. Such corrections are known to
537: partially resolve the ``HBT puzzle'', that $R_{\text{out}}$ becomes smaller
538: than $R_{\text{side}}$, though this correction does not completely resolve this
539: puzzle. In this paper, we stress that the partial Coulomb
540: correction affects not only the source size but also the observed chaoticity,
541: $\lambda^{\text{exp}}$. For example, CERES reports that the $k_t$ dependent
542: correction to $\lambda^{\text{exp}}$ reaches 15--40 \%
543: \cite{CERES_NPA714}. This is a
544: significant correction. Unfortunately, partially corrected data for the
545: collisions that we treat in this paper (S+Pb, Pb+Pb at the SPS and Au+Au at
546: 130$A$ GeV)
547: are not available. Ideally, these data should be treated within a
548: theoretical approach with a Coulomb correction \cite{Biyajima_PLB601},
549: but this is beyond the scope of this paper. 
550: In the further analyses given in \S\ref{sec:results},
551: we also present the results obtained using ``partially corrected''
552: $\lambda^{\text{exp}}$ in order to see how the results change when
553: $\lambda^{\text{exp}}$ is reduced. Following the report from CERES
554: \cite{CERES_NPA714}, according to which the correction is larger at smaller
555: $k_t$, and 
556: considering that the two-particle data used here are those of the lowest
557: momentum bin, we simply multiplied a correction factor of 0.8
558: except in the S+Pb case, where the correction factor is set to 0.7, because
559: only S+Pb data are corrected by the Gamow factor. We denote this
560: ``partially Coulomb corrected''$\lambda^{\text{true}}$ by
561: $\lambda^{\text{true}}_{\text{pc}}$.
562: 
563: \section{Extraction of the weight factor $\omega$}\label{sec:omega}
564: 
565: \subsection{Procedure}
566: In the previous section, we extracted $\lambda^{\text{true}}$ from the
567: two-particle correlation data. Next, we must determine the weight factor
568: $\omega$ in order to analyze the degree to which the pion sources are
569: chaotic by using the
570: models, especially Model III [Eq.~\eqref{l-pm}]. To obtain
571: $\omega=r_3(0)/2$, we must extrapolate $C_2$ and $C_3$ to
572: $\boldsymbol{q}=0$. The method of extrapolation is described in a
573: previous paper \cite{Morita_3pirhic} in detail. Here we briefly review
574: the procedure.
575: 
576: In experimental papers, $\omega$ is extracted by
577: simply averaging $r_3(Q_3)$ (NA44 \cite{NA44_3pi_SPb, NA44_3pi_pbpb}, WA98
578: \cite{WA98_pbpb_prc}), in which there is little dependence on $Q_3$, or
579: using quadratic and quartic fits (STAR \cite{STAR_3pi}). 
580: A shortcoming of these methods is that the $Q_3$ dependence of $r_3$ may be
581: more complex than quadratic or quartic \cite{Nakamura_PRC60}. Such
582: $Q_3$ dependences result from both an asymmetry of the source and
583: coherence.\footnote{In Ref.~\citen{Heinz_PRC70}, it is pointed out that
584: there is a possibility that the standard projection method adopted here
585: may lead to an artificial momentum dependence of the projected
586: correlators.} 
587: In this paper, we assume observed $Q_3$ dependences are due to
588: coherence. This is an assumption, but it is plausible, because the
589: asymmetry of the source causes a $Q_3$ dependence of
590: $r_3$ that is somewhat different from that in the observed $r_3$ data
591: \cite{Nakamura_PRC60}. We reproduce
592: $C_2(Q_{\text{inv}})$ and $C_3(Q_3)$ using a common source function with
593: a set of parameters which is determined by minimizing $\chi^2$ with
594: respect to the experimental data. Then, we evaluate $C_2(0)$ and
595: $C_3(0)$. The quantity $r_3(Q_3)$ is also calculated for a
596: consistency check. For simplicity, we use a spherically symmetric
597: Fourier-transformed source
598: function with simulateneous emission, 
599: $F_{ij} = f_{ij}(|\boldsymbol{q}_{ij}|) e^{i(E_i-E_j)t_0}$, 
600: in which the
601: exponential term corresponds to emission at a constant time $t_0$.
602: The assumption of simulateneous emission should be a good
603: approximation, because the experimental data suggest emissions of short 
604: duration throughout the broad range of colliding energies \cite{STAR_PRL87}.
605: Since the finite emission time duration is not related to
606: $\lambda^{\text{true}}$ but, rather, to the width of the outward correlation
607: functions, it does not affect our results below.
608: For the spatial part, $f_{ij}(|\boldsymbol{q}_{ij}|)$, we try the
609: following three
610: kinds of source functions:
611: 
612: \begin{align}
613:  f_{1,ij}(|\boldsymbol{q}_{ij}|) &= 
614:  \exp\left(-R^2 |\boldsymbol{q}_{ij}|^2 /2 \right),\label{eq:fgauss} \\
615:  f_{2,ij}(|\boldsymbol{q}_{ij}|) &= 
616:  \exp\left(-R|\boldsymbol{q}_{ij}|/2 \right), \label{eq:fexp}\\
617:  f_{3,ij}(|\boldsymbol{q}_{ij}|) &= 
618:  \frac{1}{\sqrt{\cosh(R|\boldsymbol{q}_{ij}|)}}\label{eq:fcosh}.
619: \end{align}
620: Here, $\boldsymbol{q}_{ij} = \boldsymbol{p}_i - \boldsymbol{p}_j$
621: and $R$ is a size parameter which is to be determined by
622: the $\chi^2$ fitting.
623: The third function, Eq.~\eqref{eq:fcosh}, is chosen so as to be quadratic at
624: small $|\boldsymbol{q}|$ and exponential at large
625: $|\boldsymbol{q}|$. 
626: 
627: The two- and three-particle correlation functions are then calculated
628: using the relations
629: 
630: \begin{align}
631:  C_2(\boldsymbol{p_1,p_2})&=1+\lambda_{\text{inv}}\frac{f_{12}^2}{f_{11}f_{22}} 
632:   \label{eq:c2c},\\
633:  C_3(\boldsymbol{p_1,p_2,p_3}) &= 
634:  1+\nu\left(\sum_{(i,j)}\frac{f_{ij}^2}{f_{ii}f_{jj}}
635:  +2\nu_3 \frac{f_{12}f_{23}f_{31}}{f_{11}f_{22}f_{33}}\right) \label{eq:c3c}.
636: \end{align}
637: 
638: where $\lambda_{\text{inv}}$, $\nu$ and $\nu_3$ are phenomenological adjustable
639: parameters accounting for the non-trivial coherence effect. 
640: The summation $\Sigma_{(i,j)}$ here runs over $(i,j)=(1,2),(2,3),(3,1)$.
641: The quantities $\lambda_{\text{inv}}$, $\nu$ and $\nu_3$ are unity in
642: the case of a
643: fully chaotic source. We can set $\nu_3=1$ for a
644: description of $C_3(Q_3)$ at small $Q_3$ \cite{Morita_3pirhic}.
645: Then, $\lambda_{\text{inv}}$ and $\nu$ are determined by the
646: $\chi^2$ fitting, like $R$.
647: We stress that the $\chi^2$ fit is carried out \textit{simultaneously} for the
648: two- and three-particle correlation data.  
649: 
650: The results of the $\chi^2$ fittings to the experimental data and the resultant
651: $\omega$ are listed in Table \ref{tbl:omegafit}.
652: 
653: \begin{table}[t!]
654:  \caption{\label{tbl:omegafit}Results of the $\chi^2$ fitting to $C_2$
655:  and $C_3$}
656:  \begin{center}
657:  \begin{tabular}[t]{ccccccc}\hline
658:  system & $f(|\boldsymbol{q}|)$ & $R$ [fm] &  $\lambda_{\text{inv}}$ &
659:   $\nu $ &   $\chi^2 / \text{dof}$ & $\omega$\\  \hline
660: 
661:   SPS S+Pb & $f_1$ & 4.85$\pm$0.31 & 0.49$\pm$0.04 &
662:   0.34$\pm$0.04 & 2.3/7 & 0.33$\pm$0.38 \\
663: 
664:   (NA44) \cite{NA44_2pi_PLB302,NA44_3pi_SPb} & $f_2$ & 7.55$\pm$0.84 &
665:   0.79$\pm$0.10 & 0.61$\pm$0.09 & 3.9/7 & 0.48$\pm$0.45 \\
666: 
667:   & $f_3$ & 8.99$\pm$0.80 & 0.55$\pm$0.06 & 0.40$\pm$0.05 & 1.9/7 &
668:   0.40$\pm$0.44 \\ \hline
669: 
670:   SPS Pb+Pb & $f_1$ & 7.37$\pm$0.61 & 0.54$\pm$0.05 &
671:   0.49$\pm$0.07 & 1.2/6 & 1.06$\pm$0.59 \\
672: 
673:   (NA44) \cite{NA44_2pi_PbPb,NA44_3pi_pbpb}  & $f_2$ & 11.4$\pm$1.5 &
674:   0.83$\pm$0.11 & 0.85$\pm$0.16 & 2.7/6 & 1.16$\pm$0.69 \\
675:   
676:   & $f_3$ & 13.8$\pm$1.5 & 0.61$\pm$0.07 & 0.58$\pm$0.10 & 0.9/6 &
677:   1.15$\pm$0.67 \\ \hline
678: 
679:   SPS Pb+Pb & $f_1$ & 7.73$\pm$0.10 & 0.36$\pm$0.01 &
680:   0.29$\pm$0.01 & 113/17 & 0.81$\pm$0.12 \\
681: 
682:  (WA98) \cite{WA98_pbpb_prc} & $f_2$ & 14.2$\pm$0.29 & 0.75$\pm$0.02 &
683:   0.62$\pm$0.02 & 14/17 & 0.68$\pm$0.12 \\
684: 
685:   & $f_3$ & 15.8$\pm$0.28 & 0.47$\pm$0.01 & 0.39$\pm$0.01 & 17/17 &
686:   0.78$\pm$0.14 \\ \hline
687: 
688:   RHIC Au+Au & $f_1 $& 7.0$\pm$0.07 &
689:   0.54$\pm$0.01  &
690:   0.48$\pm$0.01 & 110/30 &0.958$\pm$0.09 \\
691: 
692:  (STAR) \cite{STAR_PRL87,STAR_3pi} & $f_2$ & 14.4$\pm$0.2  &
693:   1.18$\pm$0.03 & 1.08$\pm$0.03 & 79.7/30 & 0.736$\pm$0.09 \\
694: 
695:   & $f_3$ & 15.2$\pm$0.2 & 0.71$\pm$0.01 & 0.64$\pm$0.02 & 15.8/30 &
696:   0.872$\pm$0.097 \\\hline
697:  \end{tabular}
698:  \end{center}
699: \end{table}
700: 
701: \subsection{SPS S+Pb}
702: 
703: 
704: \begin{figure}[ht]
705:  \begin{center}
706:   \includegraphics[width=\textwidth]{fig_C2-Qinv.eps}
707:  \end{center}
708:  \caption{\label{fig:c2}Two-pion correlation function
709:  $C_2(Q_{\text{inv}})$ in the (a) S+Pb collisions at the SPS, (b) Pb+Pb
710:  collisions at the SPS, measured by NA44, (c) Pb+Pb collisions at the
711:  SPS, measured by WA98, and (d) Au+Au collisions at the RHIC. The lines
712:  represent our results for the fits of each source function (see
713:  text). Filled circles
714:  represent the experimental results (SPS S+Pb by NA44\cite{NA44_2pi_PLB302},
715:  SPS Pb+Pb by NA44\cite{NA44_2pi_PbPb}and WA98\cite{WA98_pbpb_prc}, and RHIC
716:  Au+Au by STAR\cite{STAR_PRL87}).}
717: \end{figure}
718: 
719: \begin{figure}[ht]
720:  \begin{center}
721:   \includegraphics[width=\textwidth]{fig_C3-Q3.eps}
722:  \end{center}
723:  \caption{\label{fig:c3}Three-pion correlation function
724:  $C_3(Q_3)$ in the various collisions. The identification of the symbols
725:  is the same as in
726:  Fig.~\ref{fig:c2}. The experimental results presented in (a)--(d) are
727:  taken from Refs.~\citen{NA44_3pi_SPb}, \citen{NA44_3pi_pbpb},
728:  \citen{WA98_pbpb_prc}, and \citen{STAR_3pi}, respectively.}
729: \end{figure}
730: 
731: \begin{figure}[ht]
732:  \begin{center}
733:  \includegraphics[width=\textwidth]{fig_r3.eps}
734:  \end{center}
735:  \caption{\label{fig:r3}The normalized three-pion correlator
736:  $r_3(Q_3)/2$.  The identification of the symbols is the same as in
737:  Fig.~\ref{fig:c2}. The experimental results presented in (a)--(d) are
738:  taken from
739:  Refs.~\citen{NA44_3pi_SPb},\citen{NA44_3pi_pbpb}, \citen{WA98_pbpb_prc},
740:  and \citen{STAR_3pi}, respectively. The dot-dashed line in (a) represents
741:  the uncertainty propagated from the fitting parameters for the ``cosh'' case
742:  (see text).}
743: \end{figure}
744: 
745: Comparisons of the fitting results and the experimental data are displayed
746: in Figs.~\ref{fig:c2}(a), \ref{fig:c3}(a) and \ref{fig:r3}(a). The curves in
747: Figs.~\ref{fig:c2}(a) and \ref{fig:c3}(a) are fitted to the
748: data using three kind of source functions: Gaussian
749: [Eq.~\eqref{eq:fgauss}], exponential [Eq.~\eqref{eq:fexp}] and 
750: cosh [Eq.~\eqref{eq:fcosh}]. The fitting range is
751: $0 < Q_{ij}, Q_3 < 60$ MeV, which is adjusted to the available $Q_3$
752: range for $r_3(Q_3)$. In the case of $C_2$, we see
753: that the experimental data are fit well by a quardratic function at small
754: $Q_{\text{inv}}$ and an exponential function at large
755: $Q_{\text{inv}}$. For this reason, the $\chi^2$ values in the
756: exponential case and the
757: Gaussian case are larger than those in the cosh case. A similar tendency is
758: also seen in the case of $C_3$, though the exponential case is not
759: excluded by the
760: data point at the smallest $Q_{\text{inv}}$. The value of $r_3$ obtained from
761: the fitted $C_2$ and $C_3$ do not
762: differ greatly. With a simple naked-eye extrapolation in
763: Fig.~\ref{fig:r3}(a), it appears that
764: $\omega=r_3(0)/2$ becomes $\sim$0.4 for all source functions. Taking the
765: smallest $\chi^2$ value, we adopt the cosh case and obtain
766: $\omega=0.40\pm 0.44$. Note that, by the definition of $r_3(Q_3)$, where
767: $C_2(Q_{ij})-1$ is in the denominator, there exists a large uncertainty in
768: the calculated $r_3(Q_3)$ value, because of the errors on
769: $\lambda_{\text{inv}}$. In Fig.~\ref{fig:r3}(a), we plot the error band
770: for the cosh case. The uncertainty 0.44 associated with $\omega$ is
771: also obtained by extrapolation of the error band to $Q_3=0$.
772: 
773: \subsection{SPS Pb+Pb}
774: 
775: For Pb+Pb collisions at 158$A$ GeV, we compared our results to data
776: measured by both NA44 and WA98. Figures
777: \ref{fig:c2}(b), \ref{fig:c3}(b) and \ref{fig:r3}(b) plot the results
778: of the $\chi^2$ fit to the NA44 data and
779: Figs.~\ref{fig:c2}(c), \ref{fig:c3}(c) and \ref{fig:r3}(c) plot those for the
780: WA98 data.
781: The fitting ranges are adjusted to the available $r_3(Q_3)$ data range, as
782: in the S+Pb case, and we use the data for $Q_{ij},Q_3 < 60$ MeV.
783: 
784: Comparing our fit to the NA44 data, we find that all source functions seem to
785: give nice
786: descriptions of the data [see Figs.~\ref{fig:c2}(b) and
787: \ref{fig:c3}(b)]. The statistics are still insufficient, especially
788: at low $Q_3$, in the three-particle correlation function to discriminate
789: the best source function. The weight factor $\omega$ is larger than
790: unity at the best fit value. This is associated with the large
791: errorbars and consistent with the WA98 case within the errorbars.
792: For further analysis (given below), we adopt the cosh case because it gives the
793: best $\chi^2$ value. On the other hand, the WA98 data exhibit better
794: statistics [see Figs.~\ref{fig:c2}(c) and \ref{fig:c3}(c)]. This
795: excludes the Gaussian case in the fit to
796: $C_2$ and $C_3$. For the $\chi^2$ values, the exponential case has the best
797: value, $\chi^2/\text{dof}=14/17$. This tendency seems to be different
798: from that in the NA44 case. In Ref.~\citen{WA98_pbpb_prc}, the three-particle
799: correlation function is fit well by a double-exponential correlation
800: function. Although the exponential case seems to be the best of the
801: three, $r_3(Q_3)$ from the exponential source function deviates from the
802: experimental result [Fig.~\ref{fig:r3}(c)]. 
803: This should not be regarded as a serious problem, however, because errors on
804: $\lambda_{\text{inv}}$ and $\nu$ lead to an uncertainty on $r_3(Q_3)$, as
805: shown in the SPS S+Pb case [Fig.~\ref{fig:r3}(c)]. A likely reason
806: for this deviation is that $C_3(Q_3)$ is smaller than that obtained from
807: the experimental data in
808: the exponential case at low $Q_3$. Naive extrapolation by naked eye in
809: Fig.~\ref{fig:r3}(c)
810: again suggests that the cosh or
811: Gaussian case is better, but this may be a coincidence in the Gaussian case,
812: because both
813: $C_2$ and $C_3$ deviate from the experimental data at low relative
814: momenta. Hence, we adopt the cosh case as the result for $\omega$. 
815: Despite the different behavior of $C_3(Q_3)$, the result for
816: $\omega$ obtained from the WA98 data is consistent with the that
817: obtained from the NA44 data.
818: 
819: \subsection{RHIC Au+Au}
820: 
821: Results presented in this subsection are the same as those presented in
822: Ref.~\citen{Morita_3pirhic}. 
823: Figures \ref{fig:c2}(d), \ref{fig:c3}(d) and \ref{fig:r3}(d) display
824: the correlation functions and the three-pion correlator for the Au+Au
825: collisions at the RHIC energy. The fitting range is $0 < Q_{\text{inv}} <
826: 90$ MeV for $C_2$ and $0 < Q_3 < 100$ MeV for $C_3$. The high
827: statistics of the data allows us to discriminate the source function.
828: In the cosh case, the value of $\chi^2$ is much smaller than other two
829: cases. Finally, we obtain $\omega=0.872 \pm 0.097$.
830: 
831: \section{Results and discussions}\label{sec:results}
832: 
833: In \S\ref{sec:lambda}, we extracted $\lambda^{\text{true}}$ from 
834: the experimental data, $\lambda^{\text{exp}}$, with the help of a
835: statistical model. In \S\ref{sec:omega}, we extracted the weight factor
836: $\omega$ from the experimental data. Now we have two input
837: parameters for the analysis employing Models
838: I--III. [see Eqs.~\eqref{eq:modelI}--\eqref{ome-pm}].
839: In Models I and II, there is one model parameter
840: ($\varepsilon_{\text{I}}$ for Model I and $\alpha_{\text{II}}$ for Model
841: II) corresponding to the two input quantities, $\lambda^{\text{true}}$
842: and $\omega$. We determined the model parameters by minimizing
843: 
844: \begin{equation}
845:  \chi^2 \equiv
846:   \frac{[\lambda^{\text{true}}_{\text{exp}}
847:   -\lambda^{\text{true}}_{\text{cal}}]^2}
848:   {(\delta\lambda^{\text{true}}_{\text{exp}})^2}
849:   +
850:   \frac{[\omega_{\text{exp}}-\omega_{\text{cal}}]^2}
851:   {(\delta\omega_{\text{exp}})^2},\label{eq:chi2}
852: \end{equation}
853: where $\lambda^{\text{true}}_{\text{exp}}$ and $\omega_{\text{exp}}$
854: were extracted from the experimental data in \S\ref{sec:lambda} and
855: \ref{sec:omega}. The quantities $\delta\lambda^{\text{true}}_{\text{exp}}$ and
856: $\delta\omega_{\text{exp}}$ are the errors on
857: $\lambda^{\text{true}}_{\text{exp}}$ and $\omega_{\text{exp}}$, which are
858: given in Tables~\ref{tbl:lambdatrue} and \ref{tbl:omegafit}, respectively.
859: In Models I and II, $\lambda^{\text{true}}_{\text{cal}}$ and
860: $\omega_{\text{cal}}$ are functions of $\varepsilon_{\text{I}}$ and
861: $\alpha_{\text{II}}$, respectively, calculated using Eqs.~\eqref{eq:modelI} and
862: \eqref{eq:modelII}. In Model III, we solve
863: Eqs.~\eqref{l-pm} and \eqref{ome-pm} for the given sets of
864: $\lambda^{\text{true}}_{\text{exp}}$ and $\omega_{\text{exp}}$ to obtain
865: $\varepsilon_{\text{III}}$ and $\alpha_{\text{III}}$. However, solutions
866: of these equation may exist in unphysical parameter regions, such as
867: $\varepsilon_{\text{III}} < 0$. In such cases, we determine a ``Best
868: fit'' solution by minimizing the above $\chi^2$ within the physical
869: model parameter region, $0 \leq \varepsilon_{\text{III}} \leq 1$ and 
870: $0 \leq \alpha_{\text{III}} < \infty$.
871: 
872: \subsection{Partially coherent model (Model I)}
873: 
874: 
875: We plot $\varepsilon_{\text{I}}$ determined from Eq.~\eqref{eq:chi2}
876: as a function of the multiplicities of various collision systems in
877: Fig.~\ref{fig:partial}. We consider three cases: The open squares represent the
878: results obtained from Eq.~\eqref{eq:chi2} using
879: $(\lambda^{\text{true}},\omega)$, the solid triangles represent those
880: obtained using
881: $(\lambda^{\text{true}}_{\text{pc}},\omega)$,  and the closed circles represent
882: those calculated from $\omega$ only, using the
883: second equation in Eq.~\eqref{eq:modelI}.
884: 
885: While the result denoted ``$\omega$ only'' exhibits clear increase of
886: $\varepsilon_{\text{I}}$ with multiplicity, which is consistent with the
887: result obtained by the STAR\cite{STAR_3pi}, neither
888: $(\lambda^{\text{true}},\omega)$
889: nor $(\lambda^{\text{true}}_{\text{pc}},\omega)$ exhibit such a clear
890: dependence. This is because $\lambda^{\text{true}}$ in
891: Table~\ref{tbl:lambdatrue} does not display a clear multiplicity dependence
892: and 
893: $\delta\lambda_{\text{exp}}^{\text{true}} \ll \delta\omega_{\text{exp}}$. 
894: The value of $\varepsilon_{\text{I}}$ which minimizes $\chi^2$ is
895: mostly determined by $\lambda^{\text{true}}$ and
896: $\lambda^{\text{true}}_{\text{pc}}$ only, i.e., the values of
897: $\omega$ are not clearly reflected in the minimization of $\chi^2$ due
898: to the fact that they possess 
899: larger errors than $\lambda^{\text{true}}$.
900: If this model is good enough,
901: and if the experimental background
902: for the correlation, such as the Coulomb correction is successfully removed,
903: the result of $(\lambda^{\text{true}}_{\text{pc}},\omega)$ should agree
904: well with
905: the ``$\omega$ only'' result. In Fig.~\ref{fig:partial}, these results seem
906: to be quantitatively consistent. If the experimental
907: accuracy of the three-particle correlation measures improves, these
908: results will be more conclusive.
909: 
910: \begin{figure}[t!]
911:  \begin{center}
912:   \includegraphics[width=3.875in]{partial.eps}
913:  \end{center} 
914: \caption{\label{fig:partial}$\varepsilon_{\text{I}}$
915:  as a function of the multiplicity. For the Pb+Pb collisions at the SPS,
916:  we plot both the
917:  NA44 and WA98 data with a slightly shifted horizontal axis for a clear
918:  comparison of the results.}
919: \end{figure}
920: 
921: \subsection{Multicoherent model (Model II)}
922: 
923: \begin{table}[ht!]
924:  \caption{\label{tbl:multic}The parameter $\alpha_{\text{II}}$ in the
925:  multicoherent model (Model II)}
926:  \begin{center}
927:  \begin{tabular}[t]{cccc}\hline
928:   System & From ($\lambda^{\text{true}},\omega$) & From
929:   ($\lambda^{\text{true}}_{\text{pc}},\omega$) & From $\omega$ only\\  \hline
930:   SPS S+Pb & 12.97$^{+57.85}_{-6.16}$ & 1.93$^{+0.41}_{-0.31}$ & --- \\
931:   SPS Pb+Pb (NA44) & 40.12$^{+\infty}_{-23.84}$ & 3.55$^{+0.67}_{-0.51}$ &
932:   0.72$^{+0.81}_{-0.38}$ \\
933:   SPS Pb+Pb (WA98) & 201.93$^{+\infty}_{-179.2}$ & 4.59$^{+1.34}_{-0.92}$
934:   & --- \\ 
935:   RHIC Au+Au & 10.54$^{+21.03}_{-4.88}$ &  2.86$^{+1.25}_{-0.76}$ &
936:   7.59$^{+9.66}_{-6.53}$\\
937:   \hline
938:  \end{tabular}
939:  \end{center}
940: \end{table}
941: 
942: The results for $\alpha_{\text{II}}$ in the multicoherent model
943: [Eq.~\eqref{eq:modelII}] are displayed in Table~\ref{tbl:multic}.
944: As in the case of Model I, three cases are shown. The results for
945: $(\lambda^{\text{true}},\omega)$ and
946: $(\lambda^{\text{true}}_{\text{pc}},\omega)$ cases 
947: mainly reflect their $\lambda^{\text{true}}$ and
948: $\lambda^{\text{true}}_{\text{pc}}$ values due to the fact that
949: $\delta\lambda^{\text{true}}$ is much smaller
950: than $\delta\omega$.
951: In the $(\lambda^{\text{true}},\omega)$ case, $\alpha_{\text{II}}$ takes
952: a very large value, coming from $\lambda^{\text{true}}\simeq 1$.
953: The value of $\alpha_{\text{II}}$ in the
954: $(\lambda^{\text{true}}_{\text{pc}},\omega)$ case is significantly
955: smaller, but there is no clear multiplicity
956: dependence in this case as expected from the fact that
957: $\lambda^{\text{true}}_{\text{pc}}$ does not possess a clear multiplicity
958: dependence. In the ``$\omega$ only'' case, there are no solutions for
959: the SPS S+Pb and Pb+Pb (WA98) data (the blank entries in
960: Table~\ref{tbl:multic}), because this model has no
961: corresponding value of $\alpha_{\text{II}}$ below $\omega\simeq 0.82$
962: [see Fig.~2 of
963: Ref.~\citen{Nakamura_PRC61}].
964: This implies that
965: this model is not suitable for studying multiplicity dependence;
966: i.e., models with a chaotic background give a better
967: description of the data.
968: 
969: \subsection{Partially multicoherent model (Model III)}
970: 
971: In the analysis using Model III, there are two output parameters
972: ($\varepsilon_{\text{III}}$ and $\alpha_{\text{III}}$) corresponding to
973: the two inputs,
974: $\lambda^{\text{true}}$ and $\omega$. In the following, we display the
975: allowed regions of $\varepsilon_{\text{III}}$ and $\alpha_{\text{III}}$
976: which correspond to the
977: sets of $\lambda^{\text{true}}$ and $\omega$ in Fig.~\ref{fig:e-a} and
978: $\lambda^{\text{true}}_{\text{pc}}$ and $\omega$ in Fig.~\ref{fig:e-a_lpc}.
979: 
980: \begin{figure}[b!]
981:  \begin{center}
982:   \includegraphics[width=\textwidth]{fig_e-a.eps}
983:  \end{center} 
984: \caption{\label{fig:e-a}Allowed regions for
985:  $\lambda^{\text{true}}$ in the various systems. (See text for details.)}
986: \end{figure}
987: 
988: \begin{figure}[ht]
989:  \begin{center}
990:   \includegraphics[width=\textwidth]{fig_e-a_reducelam.eps}
991:  \end{center} 
992: \caption{\label{fig:e-a_lpc}Allowed region for
993:  $\lambda^{\text{true}}_{\text{pc}}$ in the various systems. (See text
994:  for details.)}
995: \end{figure}
996: 
997: In each figure, the lightly shaded area labeled ``A'' is the allowed
998: parameter
999: region corresponding to the value of $\lambda^{\text{true}}$ or
1000: $\lambda^{\text{true}}_{\text{pc}}$, whose
1001: boundary is
1002: indicated by the solid line. Area B, bounded by the dashed (upper
1003: limit of $\omega$) and dotted (lower limit of $\omega$) curves, is the
1004: allowed parameter region corresponding to $\omega$. The darkest area,
1005: which is the region in which Areas A and B overlap, is Area C, which is the
1006: allowed parameter region for $\alpha_{\text{III}}$ and
1007: $\varepsilon_{\text{III}}$. The best fit
1008: values calculated from the values in
1009: Tables \ref{tbl:lambdatrue} and \ref{tbl:omegafit} are indicated by the
1010: squares. 
1011: 
1012: Figure \ref{fig:e-a}(a) plots the result for the S+Pb collisions using
1013: $\lambda^{\text{true}}$. It is seen that Area C is narrow, and the chaotic
1014: fraction has a lower bound near $\varepsilon_{\text{III}} =0.6$. The best
1015: fit value is $\varepsilon_{\text{III}}\simeq 1$ which reflects that
1016: $\lambda^{\text{true}}$ is near unity and a lower value of $\omega$. If
1017: we adopt a partially Coulomb
1018: corrected
1019: $\lambda^{\text{true}}_{\text{pc}}$, the situation changes. Due to
1020: the smaller
1021: $\lambda^{\text{true}}_{\text{pc}}$[Fig.~\ref{fig:e-a_lpc}(a)],
1022: $\varepsilon_{\text{III}}$ is
1023: allowed for a wider
1024: region and it has a
1025: maximum. The best fit value also shifts to $\varepsilon_{\text{III}} =
1026: 0.67$ and
1027: $\alpha_{\text{III}}= 0.21$. Note that, as we can see from
1028: Fig.~\ref{fig:e-a_lpc}, the upper bound of $\varepsilon_{\text{III}}$ is
1029: mainly dominated by the lower bound of
1030: $\omega$.
1031: 
1032: Figure \ref{fig:e-a}(b) displays the allowed region from
1033: $\lambda^{\text{true}}$ and $\omega$ for the NA44 Pb+Pb collision
1034: dataset. Because the lower bound of $\lambda^\text{true}$ has a large value,
1035: 0.95, Area C allows both a mostly chaotic source with a
1036: small number of coherent sources and a number of the coherent sources
1037: with a small chaotic background. In this case, a solution of
1038: Eqs.~\eqref{l-pm}--\eqref{ome-pm} exists in the unphysical region,
1039: $\varepsilon_{\text{III}}=1.01$ and $\alpha_{\text{III}}=0.003$.
1040: Hence, we determine the ``Best fit'' point by minimizing $\chi^2$ in
1041: Eq.~\eqref{eq:chi2}. The result of the minimization gives the ``Best
1042: fit'' at $\varepsilon_{\text{III}}=0$ and $\alpha_{\text{III}}=40.1$.
1043: This result does not always imply that the multicoherent picture is good;
1044: the difference between the value of
1045: $\chi^2$ for this minimum and in another case, for example,
1046: $\epsilon_{\text{III}}=0.99$ and $\alpha=0.01$, is much smaller than
1047: unity. 
1048: 
1049: If we adopt the partially Coulomb
1050: corrected chaoticity $\lambda^{\text{true}}_{\text{pc}}$, Area C in
1051: Fig.~\ref{fig:e-a_lpc}(b) becomes narrow, as in the S+Pb
1052: case. Because the lower bound of $\omega$ is larger than that in the S+Pb
1053: case, the upper limit of $\varepsilon_{\text{III}}$ becomes smaller.
1054: Similarly, this case does not
1055: have a solution of $\varepsilon_{\text{III}}$ and $\alpha_{\text{III}}$
1056: within the physical region ($\varepsilon_{\text{III}}=1.56$,
1057: $\alpha_{\text{III}}=1.12$), and therefore the ``Best fit'' point is
1058: determined by
1059: minimizing $\chi^2$ in Eq.~\eqref{eq:chi2}. Though the location of the
1060: ``Best fit''
1061: point corresponds to the multicoherent picture
1062: ($\varepsilon_{\text{III}}=0$), the opposite case
1063: ($\varepsilon_{\text{III}}\sim 1$) is
1064: statistically allowed for the same reason as in the previous case.
1065: 
1066: Similarly, Figs.~\ref{fig:e-a}(c) and \ref{fig:e-a_lpc}(c) display the
1067: results for the WA98 dataset. Since the central value of
1068: $\lambda^{\text{true}}$ exceeds unity, only a very small region near
1069: $\varepsilon_{\text{III}}=1$ is allowed in Fig.~\ref{fig:e-a}(c). This
1070: situation changes drastically if one adopts the
1071: $\lambda^{\text{true}}_{\text{pc}}$.
1072: In Fig.~\ref{fig:e-a_lpc}(c), the allowed region has a shape similar
1073: to that in the NA44 case [Fig.~\ref{fig:e-a_lpc}(b)]. The best fit value is
1074: located at $\varepsilon_{\text{III}} = 0.47$ and $\alpha_{\text{III}}=1.29$. 
1075: 
1076: Finally, we display the results for Au+Au collisions at RHIC in
1077: Figs.~\ref{fig:e-a}(d) and \ref{fig:e-a_lpc}(d). From
1078: $\lambda^{\text{true}}$ and $\omega$, it is difficult to distinguish the
1079: structure of the source; both a large chaotic fraction with a small number
1080: of coherent sources ($\varepsilon_{\text{III}}\sim 1$ and
1081: $\alpha_{\text{III}}< 1$) and small chaotic fraction with a large number of
1082: coherent sources ($\varepsilon_{\text{III}}\sim 0$ and
1083: $\alpha_{\text{III}}> 1$) can reproduce the experimental data. The best
1084: fit value
1085: is located at $\varepsilon_{\text{III}}=0.75$ and
1086: $\alpha_{\text{III}}=0.77$. However,
1087: if we use the partially Coulomb corrected two-particle correlation
1088: data, the allowed region is strongly restricted. Although the solution
1089: of Eqs.~\eqref{l-pm}--\eqref{ome-pm} exists in the unphysical region,
1090: $\varepsilon_{\text{III}}=-0.54$ and $\alpha_{\text{III}}=6.83$, the
1091: ``Best fit'' point is located at $\epsilon_{\text{III}}=0$ and
1092: $\alpha_{\text{III}}=2.86$. The maximum value allowed
1093: for $\varepsilon_{\text{III}}$ is 0.4. This means that the strong chaotic
1094: behavior observed at the RHIC is due to the production of a cluster of coherent
1095: sources.
1096: 
1097: \begin{figure}[b!]
1098:  \begin{center}
1099:   \includegraphics[width=3.875in]{partialmulti.eps}
1100:  \end{center} 
1101: \caption{\label{fig:partialmulti}The minimum (solid squares) and
1102:  maximum (solid circles) value of $\alpha_{\text{III}}$ in Model III
1103:  for various
1104:  collisions as a function of the multiplicity. For the Pb+Pb collisions at
1105:  the SPS, we plot both the NA44 and WA98 data with a slightly shifted
1106:  horizontal axis for a clear comparison of the results.}
1107: \end{figure}
1108: 
1109: As shown in the
1110: Fig.~\ref{fig:e-a_lpc}, the maximum bounds on
1111: $\varepsilon_{\text{III}}$ are mainly
1112: determined by the lower bounds on $\omega$, which become larger as the
1113: energy and the multiplicity increase. The fact that
1114: $\alpha_{\text{III}}$ becomes larger as the multiplicity increases is
1115: consistent with the result of
1116: a previous analysis by one of the authors (H.~N.), in which the chaoticity
1117: and the weight factor were experimentally
1118: determined.\cite{Nakamura_PRC66} In Fig.~\ref{fig:partialmulti}, we plot
1119: the
1120: maximum and minimum values of $\alpha_{\text{III}}$ for the sets of
1121: $\lambda^{\text{true}}_{\text{pc}}$ and $\omega$ extracted from
1122: Fig.~\ref{fig:e-a_lpc}. We can see that minimum value of
1123: $\alpha_{\text{III}}$
1124: increases with the multiplicity. This suggests that, as the collision energy
1125: and the multiplicity increase, the number of coherent sources
1126: increases. Note that the maximum values are the same as
1127: those obtained using Model II for
1128: $\lambda^{\text{true}}_{\text{pc}}$ (see Table~\ref{tbl:multic}), because
1129: $\alpha_{\text{III}}$ takes its maximum value when
1130: $\varepsilon_{\text{III}}=0$. 
1131: Furthermore, this tendency seems
1132: to be correlated with the plausibility of the statistical model
1133: ($\chi^2/N_{\text{dof}}$), given in
1134: Tables~\ref{tbl:t-mu_SPS} and \ref{tbl:t-mu_RHIC}. Though we do not know any
1135: explicit relation between Models
1136: I--III and the hadronization mechanism, this result may reflect a possible
1137: hadronization mechanism from the quark-gluon plasma phase created in the
1138: collisions. 
1139: 
1140: \section{Summary}
1141: In summary, we have investigated the degree to which the pion sources
1142: are chaotic in
1143: various heavy ion collisions by analyzing the two- and 
1144: three-particle correlation data with three kinds of particle
1145: production models. For the two-particle correlation data, we have
1146: extracted the ``true'' chaoticity by considering long-lived resonance
1147: contributions to the pion multiplicity, with the help of a statistical
1148: model. Using simple source functions, we simultaneously investigated
1149: the two- and three-particle correlation functions to extract the
1150: weight factor $\omega$ of the three-particle correlator. Incorporating the
1151: chaoticity and the weight factor into the models, we have studied
1152: the chaotic fraction and mean number of the coherent sources.
1153: The results for $\varepsilon_{\text{I}}$ obtained from $\omega$
1154: indicates that the system becomes chaotic as the multiplicity
1155: increases. This result is consistent with Ref.~\citen{STAR_3pi}. 
1156: From a multicoherent source point of view, it is concluded that pions at higher
1157: collision energies may be emitted from a cluster of coherent sources
1158: and the number of sources increases as the collision energy and
1159: the multiplicity increase (Fig.~\ref{fig:partialmulti}). 
1160: 
1161: 
1162: 
1163: 
1164: \section*{Acknowledgements}
1165:  The authors would like to acknowledge Professors I.~Ohba and H.~Nakazato for
1166:  their insightful comments.
1167:  This work was partially supported by the
1168:  Ministry of Education, Science and Culture, of Japan (Grant No.13135221
1169:  and Grant No.18540294)
1170:  , Waseda University Grant for Special Research Projects No.~2003A-095
1171:  and 2003A-591, and a Grant for the 21st Century COE Program at Waseda
1172:  University from Ministry of Education, Science and Culture, of Japan.
1173:  One of the author (S.~M.) would like to thank the YITP computer room.
1174: %We would like to thank ...........
1175: 
1176: 
1177: %\appendix
1178: %\section{First Appendix} %Empty argument \section{} yields `Appendix'. 
1179: %
1180: %\section{Second Appendix}
1181: 
1182: 
1183: \begin{thebibliography}{99}
1184: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1185: % Some macros are available for the bibliography:
1186: %  o for general use
1187: %    \JL : general journals                 \andvol : Vol (Year) Page
1188: %  o for individual journal 
1189: %    \AJ   : Astrophys. J.           \NC         : Nuovo Cim.
1190: %    \ANN  : Ann. of Phys.           \NPA, \NPB  : Nucl. Phys. [A,B]
1191: %    \CMP  : Commun. Math. Phys.     \PLA, \PLB  : Phys. Lett. [A,B]
1192: %    \IJMP : Int. J. Mod. Phys.      \PRA - \PRE : Phys. Rev. [A-E]     
1193: %    \JHEP : J. High Energy Phys.    \PRL        : Phys. Rev. Lett.
1194: %    \JMP  : J. Math. Phys.          \PRP        : Phys. Rep.
1195: %    \JP   : J. of Phys.             \PTP        : Prog. Theor. Phys.     
1196: %    \JPSJ : J. Phys. Soc. Jpn.      \PTPS       : Prog. Theor. Phys. Suppl.
1197: % Usage:
1198: %  \PRD{45,1990,345}          ==> Phys.~Rev.\ \textbf{D45} (1990), 345
1199: %  \JL{Nature,418,2002,123}   ==> Nature \textbf{418} (2002), 123
1200: %  \andvol{B123,1995,1020}    ==> \textbf{B123} (1995), 1020
1201:  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1202:  
1203:  \bibitem{Tomasik_qgp3}
1204: 	 B.~Tom\'{a}\v{s}ik and U.~A. Wiedemann, in
1205: 	 \textit{Quark-Gluon Plasma 3}, ed. R.~Hwa and X.~N. Wang (World
1206: 	 Scientific, 2003), p.~715; hep-ph/0210250.
1207: 	 
1208:  \bibitem{Heinz_NPA702}
1209: 	 U.~Heinz and P.~Kolb, \NPA{702,2002,269}.
1210: 	 
1211:  \bibitem{Morita_PTP111}
1212: 	 K.~Morita and S.~Muroya, \PTP{111,2004,93},
1213: 	 and references therein.
1214: 	 
1215:  \bibitem{Kolb_PLB500}
1216: 	 P.~F. Kolb, P.~Huovinen, U.~Heinz and H.~Heiselberg,
1217: 	 \PLB{500,2001,232}.
1218: 	 
1219:  \bibitem{Rischke_NPA698}
1220: 	 D.~H. Rischke, \NPA{698,2002,153c}.
1221: 	 
1222:  \bibitem{Gyulassy_PLB217}
1223: 	 M.~Gyulassy and S.~S. Padula, \PLB{217,1988,181}.
1224: 	 
1225:  \bibitem{Bolz_PRD47}
1226: 	 J.~Bolz, U.~Ornik, M.~Pl\"{u}mer, B.~R. Schlei and R.~M.Weiner,
1227: 	 \PRD{47,1993,3860}.
1228: 	 
1229:  \bibitem{Heiselberg_PLB}
1230: 	 H.~Heiselberg, \PLB{379,1996,27}.
1231: 	 
1232:  \bibitem{Csorgo_ZPHYS}
1233: 	 T.~Cs\"{o}rg\H{o}, B.~L\"{o}rstad and J.~Zim\'{a}nyi,
1234: 	 Z. Phys. C\textbf{71} (1996) 491.
1235: 	 
1236:  \bibitem{CERES_NPA714}
1237: 	 D.~Adamov\'{a} et~al. (CERES Collaboration),
1238: 	 \NPA{714,2003,124}.
1239: 	 
1240:  \bibitem{Biyajima_PTP84}
1241: 	 M.~Biyajima, A.~Bartl, T.~Mizoguchi, N.~Suzuki and O.~Terazawa,
1242: 	 \PTP{84,1990,931}.
1243: 	 
1244:  \bibitem{Heinz_PRC56}
1245: 	 U.~Heinz and Q.~H. Zhang, \PRC{56,1997,426}.
1246: 	 
1247:  \bibitem{Nakamura_PRC60}
1248: 	 H.~Nakamura and R.~Seki, \PRC{60,1999,064904}.
1249: 	 
1250:  \bibitem{Nakamura_PRC61}
1251: 	 H.~Nakamura and R.~Seki, \PRC{61,2000,054905}.
1252: 	 
1253:  \bibitem{Humanic_PRC60}
1254: 	 T.~J. Humanic, \PRC{60,1999,014901}.
1255: 	 
1256:  \bibitem{Nakamura_PRC66}
1257: 	 H.~Nakamura and R.~Seki, \PRC{66,2002,027901}.
1258: 
1259:  \bibitem{Morita_3pirhic}
1260: 	 K.~Morita, S.~Muroya and H.~Nakamura, \PTP{114,2005,583}.
1261: 
1262:  \bibitem{NA44_2pi_PLB302}
1263: 	 H.~B{\o}ggild et~al. (NA44 Collaboration),
1264: 	 \PLB{302,1993,510}.
1265: 
1266:  \bibitem{NA44_2pi_ZPHYS}
1267: 	 K.~Kaimi et~al. (NA44 Collaboration), 
1268: 	 Z. Phys. C \textbf{75} (1997), 619.
1269: 	 
1270:  \bibitem{NA44_3pi_SPb}
1271: 	 H.~B{\o}gglid et~al. (NA44 Collaboration),
1272: 	 \PLB{455,1999,77}.
1273: 	 
1274:  \bibitem{NA44_2pi_PbPb}
1275: 	 I.~G. Bearden et~al. (NA44 Collaboration),
1276: 	 \PRC{58,1998,1656}.
1277: 
1278:  \bibitem{NA44_3pi_pbpb}
1279: 	 I.~G. Bearden et~al. (NA44 Collaboration),
1280: 	 \PLB{517,2001,25}.
1281:  \bibitem{WA98_3pi_prl}
1282: 	 M.~M. Aggarwal et~al. (WA98 Collaboration),
1283: 	 \PRL{85,2000,2895}.
1284: 	 
1285:  \bibitem{WA98_pbpb_prc}
1286: 	 M.~M. Aggarwal et~al. (WA98 Collaboration),
1287: 	 \PRC{67,2003,014906}.
1288: 	 
1289:  \bibitem{Braun-Munziger_PLB518}
1290: 	 P.~Braun-Munzinger, D.~Magestro, K.~Redlich and J.~Stachel,
1291: 	 \PLB{518,2001,41}.
1292: 	 
1293:  \bibitem{Wiedemann_PRC56}
1294: 	 U.~A. Wiedemann and U.~Heinz,
1295: 	 \PRC{56,1997,3265}.
1296: 
1297:  \bibitem{Cleymans_PRC60}
1298: 	 J.~Cleymans and K.~Redlich,
1299: 	 \PRC{60,1999,054908}.
1300: 
1301:  \bibitem{Murray_NA44_NPA566}
1302: 	 M.~Murray (NA44 Collaboration),
1303: 	 \NPA{566,1994,515c}.
1304: 	 
1305:  \bibitem{Gazdzicki_NA35_NPA590}
1306: 	 M.~Ga\'{z}dzicki (NA35 Collaboration),
1307: 	 \NPA{590,1995,197c}.
1308: 
1309:  \bibitem{Jacak_HDNM}
1310: 	 B.~Jacak (NA44 Collaboration), in \textit{Hot and Dense Nuclear
1311: 	 Matter},
1312: 	 ed. W.~Greiner, H.~St\"{o}cker and A.~Gallmann
1313: 	 (Plenum, New York, 1994), p.~607.
1314: 
1315:  \bibitem{DiBari_NPA590}
1316: 	 D.~D. Bari (WA85 Collaboration),
1317: 	 \NPA{590,1995,307c}.
1318: 
1319:  \bibitem{Abatzis_NPA566}
1320: 	 S.~Abatzis et~al. (WA85 Collaboration),
1321: 	 \NPA{566,1994,225c}.
1322: 
1323:  \bibitem{Jones_NPA610}
1324: 	 P.~G. Jones (NA49 Collaboration),
1325: 	 \NPA{610,1996,188c}.
1326: 
1327:  \bibitem{Rohrich_Hirschegg97}
1328: 	 D.~R\"{o}hrich (NA49 Collaboration), in \textit{Proceedings of
1329: 	 International Workshop XXV on Gross Properties of Nuclei and
1330: 	 Nuclear Excitations, Hirschegg 1997} (1997), p.~299.
1331: 
1332:  \bibitem{WA97_PLB499}
1333: 	 E.~Andersen et~al. (WA97 Collaboration),
1334: 	 \PLB{449,1999,401}.
1335: 
1336:  \bibitem{NA49_JPhys25}
1337: 	 F.~Gabler et~al. (NA49 Collaboration), 
1338: 	 J. of Phys. \textbf{G25} (1999), 199.
1339: 
1340:  \bibitem{NA49_JPhys25_189}
1341: 	 S.~Margetis et~al. (NA49 Collaboration),
1342: 	 J. of Phys. \textbf{G25} (1999), 189.
1343: 
1344:  \bibitem{Kaneta_NPA638}
1345: 	 M.~Kaneta et~al. (NA44 Collaboration),
1346: 	 \NPA{638,1998,419c}.
1347: 	 
1348:  \bibitem{STAR_PbarPratio130}
1349: 	 C.~Adler et~al. (STAR Collaboration),
1350: 	 \PRL{86,2001,4778}. 
1351: 	 [Errata; \andvol{90,2003,119903}].
1352: 
1353:  \bibitem{STAR_pbar}
1354: 	 C.~Adler et~al. (STAR Collaboration),
1355: 	 \PRL{87,2001,262302}.
1356: 
1357:  \bibitem{STAR_PLB567}
1358: 	 J.~Adams et~al. (STAR Collaboration),
1359: 	 \PRL{567,2003,167}.
1360: 
1361:  \bibitem{STAR_PLB595}
1362: 	 C.~Adler et~al. (STAR Collaboration),
1363: 	 \PLB{595,2004,143}.
1364: 
1365:  \bibitem{STAR_PRC66}
1366: 	 C.~Adler et~al. (STAR Collaboration), 
1367: 	 \PRC{66,2002,061901(R)}.
1368: 
1369:  \bibitem{STAR_PRL87}
1370: 	 C.~Adler et~al. (STAR Collaboration),
1371: 	 \PRL{87,2001,112301}.
1372: 
1373:  \bibitem{Pratt_PRD33}
1374: 	 S.~Pratt, \PRC{33,1986,1314}.
1375: 
1376:  \bibitem{Bowler_PLB270}
1377: 	 M.~G. Bowler, \PLB{270,1991,69}.
1378: 	 
1379:  \bibitem{Biyajima_PLB353}
1380: 	 M.~Biyajima, T.~Mizoguchi, T.~Osada and G.~Wilk,
1381: 	 \PLB{353,1995,340}.
1382: 	 
1383:  \bibitem{Sinyukov_PLB432}
1384: 	 Y.~M. Sinyukov, R.~Lednicky, S.~V. Akkelin, J.~Pluta and
1385: 	 B.~Erazmus,
1386: 	 \PLB{432,1998,248}.
1387: 
1388:  \bibitem{PHENIX_HBT200PRL}
1389: 	 S.~S. Adler et~al. (PHENIX Collaboration),
1390: 	 \PRL{93,2004,152302}.
1391: 
1392:  \bibitem{Biyajima_PLB601}
1393: 	 M.~Biyajima, M.~Kaneyama and T.~Mizoguchi,
1394: 	 \PLB{601,2004,41}.
1395: 
1396:  \bibitem{STAR_3pi}
1397: 	 J.~Adams et~al. (STAR Collaboration), 
1398: 	 \PRL{91,2003,262301}.
1399: 
1400:  \bibitem{Heinz_PRC70}
1401: 	 U.~Heinz and A.~Sugarbaker, \PRC{70,2004,054908}.
1402: \end{thebibliography}
1403: \end{document}
1404: