Net charge fluctuations in Au + Au interactions at sqrt[s(NN)]=130 GeV.

Abstract

Data from Au + Au interactions at sqrt[s(NN)]=130 GeV, obtained with the PHENIX detector at the Relativistic Heavy-Ion Collider, are used to investigate local net charge fluctuations among particles produced near midrapidity. According to recent suggestions, such fluctuations may carry information from the quark-gluon plasma. This analysis shows that the fluctuations are dominated by a stochastic distribution of particles, but are also sensitive to other effects, like global charge conservation and resonance decays.

Full text

Page 1

arXiv:nucl-ex/0203014v1 21 Mar 2002
Net Charge Fluctuations in Au + Au Interactions
at√sNN = 130 GeV
K. Adcox,40S.S. Adler,3N.N. Ajitanand,27Y. Akiba,14J. Alexander,27L. Aphecetche,34Y. Arai,14S.H. Aronson,3
R. Averbeck,28T.C. Awes,29K.N. Barish,5P.D. Barnes,19J. Barrette,21B. Bassalleck,25S. Bathe,22V. Baublis,30
A. Bazilevsky,12,32S. Belikov,12,13F.G. Bellaiche,29S.T. Belyaev,16M.J. Bennett,19Y. Berdnikov,35S. Botelho,33
M.L. Brooks,19D.S. Brown,26N. Bruner,25D. Bucher,22H. Buesching,22V. Bumazhnov,12G. Bunce,3,32
J. Burward-Hoy,28S. Butsyk,28,30T.A. Carey,19P. Chand,2J. Chang,5W.C. Chang,1L.L. Chavez,25
S. Chernichenko,12C.Y. Chi,8J. Chiba,14M. Chiu,8R.K. Choudhury,2T. Christ,28T. Chujo,3,39M.S. Chung,15,19
P. Chung,27V. Cianciolo,29B.A. Cole,8D.G. D’Enterria,34G. David,3H. Delagrange,34A. Denisov,12
A. Deshpande,32E.J. Desmond,3O. Dietzsch,33B.V. Dinesh,2A. Drees,28A. Durum,12D. Dutta,2K. Ebisu,24
Y.V. Efremenko,29K. El Chenawi,40H. En’yo,17,31S. Esumi,39L. Ewell,3T. Ferdousi,5D.E. Fields,25S.L. Fokin,16
Z. Fraenkel,42A. Franz,3A.D. Frawley,9S.-Y. Fung,5S. Garpman,20,∗T.K. Ghosh,40A. Glenn,36A.L. Godoi,33
Y. Goto,32S.V. Greene,40M. Grosse Perdekamp,32S.K. Gupta,2W. Guryn,3H.-˚ A. Gustafsson,20J.S. Haggerty,3
H. Hamagaki,7A.G. Hansen,19H. Hara,24E.P. Hartouni,18R. Hayano,38N. Hayashi,31X. He,10T.K. Hemmick,28
J.M. Heuser,28M. Hibino,41J.C. Hill,13D.S. Ho,43K. Homma,11B. Hong,15A. Hoover,26T. Ichihara,31,32
K. Imai,17,31M.S. Ippolitov,16M. Ishihara,31,32B.V. Jacak,28,32W.Y. Jang,15J. Jia,28B.M. Johnson,3
S.C. Johnson,18,28K.S. Joo,23S. Kametani,41J.H. Kang,43M. Kann,30S.S. Kapoor,2S. Kelly,8B. Khachaturov,42
A. Khanzadeev,30J. Kikuchi,41D.J. Kim,43H.J. Kim,43S.Y. Kim,43Y.G. Kim,43W.W. Kinnison,19
E. Kistenev,3A. Kiyomichi,39C. Klein-Boesing,22S. Klinksiek,25L. Kochenda,30V. Kochetkov,12D. Koehler,25
T. Kohama,11D. Kotchetkov,5A. Kozlov,42P.J. Kroon,3K. Kurita,31,32M.J. Kweon,15Y. Kwon,43G.S. Kyle,26
R. Lacey,27J.G. Lajoie,13J. Lauret,27A. Lebedev,13,16D.M. Lee,19M.J. Leitch,19X.H. Li,5Z. Li,6,31D.J. Lim,43
M.X. Liu,19X. Liu,6Z. Liu,6C.F. Maguire,40J. Mahon,3Y.I. Makdisi,3V.I. Manko,16Y. Mao,6,31S.K. Mark,21
S. Markacs,8G. Martinez,34M.D. Marx,28A. Masaike,17F. Matathias,28T. Matsumoto,7,41P.L. McGaughey,19
E. Melnikov,12M. Merschmeyer,22F. Messer,28M. Messer,3Y. Miake,39T.E. Miller,40A. Milov,42
S. Mioduszewski,3,36R.E. Mischke,19G.C. Mishra,10J.T. Mitchell,3A.K. Mohanty,2D.P. Morrison,3
J.M. Moss,19F. M¨ uhlbacher,28M. Muniruzzaman,5J. Murata,31S. Nagamiya,14Y. Nagasaka,24J.L. Nagle,8
Y. Nakada,17B.K. Nandi,5J. Newby,36L. Nikkinen,21P. Nilsson,20S. Nishimura,7A.S. Nyanin,16J. Nystrand,20
E. O’Brien,3C.A. Ogilvie,13H. Ohnishi,3,11I.D. Ojha,4,40M. Ono,39V. Onuchin,12A. Oskarsson,20L.¨Osterman,20
I. Otterlund,20K. Oyama,7,38L. Paffrath,3,∗A.P.T. Palounek,19V.S. Pantuev,28V. Papavassiliou,26S.F. Pate,26
T. Peitzmann,22A.N. Petridis,13C. Pinkenburg,3,27R.P. Pisani,3P. Pitukhin,12F. Plasil,29M. Pollack,28,36
K. Pope,36M.L. Purschke,3I. Ravinovich,42K.F. Read,29,36K. Reygers,22V. Riabov,30,35Y. Riabov,30
M. Rosati,13A.A. Rose,40S.S. Ryu,43N. Saito,31,32A. Sakaguchi,11T. Sakaguchi,7,41H. Sako,39T. Sakuma,31,37
V. Samsonov,30T.C. Sangster,18R. Santo,22H.D. Sato,17,31S. Sato,39S. Sawada,14B.R. Schlei,19Y. Schutz,34
V. Semenov,12R. Seto,5T.K. Shea,3I. Shein,12T.-A. Shibata,31,37K. Shigaki,14T. Shiina,19Y.H. Shin,43
I.G. Sibiriak,16D. Silvermyr,20K.S. Sim,15J. Simon-Gillo,19C.P. Singh,4V. Singh,4M. Sivertz,3A. Soldatov,12
R.A. Soltz,18S. Sorensen,29,36P.W. Stankus,29N. Starinsky,21P. Steinberg,8E. Stenlund,20A. Ster,44S.P. Stoll,3
M. Sugioka,31,37T. Sugitate,11J.P. Sullivan,19Y. Sumi,11Z. Sun,6M. Suzuki,39E.M. Takagui,33A. Taketani,31
M. Tamai,41K.H. Tanaka,14Y. Tanaka,24E. Taniguchi,31,37M.J. Tannenbaum,3J. Thomas,28J.H. Thomas,18
T.L. Thomas,25W. Tian,6,36J. Tojo,17,31H. Torii,17,31R.S. Towell,19I. Tserruya,42H. Tsuruoka,39
A.A. Tsvetkov,16S.K. Tuli,4H. Tydesj¨ o,20N. Tyurin,12T. Ushiroda,24H.W. van Hecke,19C. Velissaris,26
J. Velkovska,28M. Velkovsky,28A.A. Vinogradov,16M.A. Volkov,16A. Vorobyov,30E. Vznuzdaev,30H. Wang,5
Y. Watanabe,31,32S.N. White,3C. Witzig,3F.K. Wohn,13C.L. Woody,3W. Xie,5,42K. Yagi,39S. Yokkaichi,31
G.R. Young,29I.E. Yushmanov,16W.A. Zajc,8Z. Zhang,28and S. Zhou6
(PHENIX Collaboration)
1Institute of Physics, Academia Sinica, Taipei 11529, Taiwan
2Bhabha Atomic Research Centre, Bombay 400 085, India
3Brookhaven National Laboratory, Upton, NY 11973-5000, USA
4Department of Physics, Banaras Hindu University, Varanasi 221005, India
5University of California - Riverside, Riverside, CA 92521, USA
6China Institute of Atomic Energy (CIAE), Beijing, People’s Republic of China
7Center for Nuclear Study, Graduate School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo, Tokyo 113-0033, Japan
8Columbia University, New York, NY 10027 and Nevis Laboratories, Irvington, NY 10533, USA
9Florida State University, Tallahassee, FL 32306, USA
10Georgia State University, Atlanta, GA 30303, USA
11Hiroshima University, Kagamiyama, Higashi-Hiroshima 739-8526, Japan
1

Page 2

12Institute for High Energy Physics (IHEP), Protvino, Russia
13Iowa State University, Ames, IA 50011, USA
14KEK, High Energy Accelerator Research Organization, Tsukuba-shi, Ibaraki-ken 305-0801, Japan
15Korea University, Seoul, 136-701, Korea
16Russian Research Center ”Kurchatov Institute”, Moscow, Russia
17Kyoto University, Kyoto 606, Japan
18Lawrence Livermore National Laboratory, Livermore, CA 94550, USA
19Los Alamos National Laboratory, Los Alamos, NM 87545, USA
20Department of Physics, Lund University, Box 118, SE-221 00 Lund, Sweden
21McGill University, Montreal, Quebec H3A 2T8, Canada
22Institut f¨ ur Kernphysik, University of M¨ unster, D-48149 M¨ unster, Germany
23Myongji University, Yongin, Kyonggido 449-728, Korea
24Nagasaki Institute of Applied Science, Nagasaki-shi, Nagasaki 851-0193, Japan
25University of New Mexico, Albuquerque, NM 87131, USA
26New Mexico State University, Las Cruces, NM 88003, USA
27Chemistry Department, State University of New York - Stony Brook, Stony Brook, NY 11794, USA
28Department of Physics and Astronomy, State University of New York - Stony Brook, Stony Brook, NY 11794, USA
29Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA
30PNPI, Petersburg Nuclear Physics Institute, Gatchina, Russia
31RIKEN (The Institute of Physical and Chemical Research), Wako, Saitama 351-0198, JAPAN
32RIKEN BNL Research Center, Brookhaven National Laboratory, Upton, NY 11973-5000, USA
33Universidade de S˜ ao Paulo, Instituto de F´isica, Caixa Postal 66318, S˜ ao Paulo CEP05315-970, Brazil
34SUBATECH (Ecole des Mines de Nantes, IN2P3/CNRS, Universite de Nantes) BP 20722 - 44307, Nantes-cedex 3, France
35St. Petersburg State Technical University, St. Petersburg, Russia
36University of Tennessee, Knoxville, TN 37996, USA
37Department of Physics, Tokyo Institute of Technology, Tokyo, 152-8551, Japan
38University of Tokyo, Tokyo, Japan
39Institute of Physics, University of Tsukuba, Tsukuba, Ibaraki 305, Japan
40Vanderbilt University, Nashville, TN 37235, USA
41Waseda University, Advanced Research Institute for Science and Engineering, 17 Kikui-cho, Shinjuku-ku, Tokyo 162-0044,
Japan
42Weizmann Institute, Rehovot 76100, Israel
43Yonsei University, IPAP, Seoul 120-749, Korea
44KFKI Research Institute for Particle and Nuclear Physics (RMKI), Budapest, Hungary†
(February 8, 2008)
Data from Au + Au interactions at√sNN = 130 GeV, obtained with the PHENIX detector
at RHIC, are used to investigate local net charge fluctuations among particles produced near mid-
rapidity. According to recent suggestions, such fluctuations may carry information from the Quark
Gluon Plasma. This analysis shows that the fluctuations are dominated by a stochastic distribution
of particles, but are also sensitive to other effects, like global charge conservation and resonance
decays.
PACS numbers: 25.75.Dw
The PHENIX detector [1] at the Relativistic Heavy-Ion
Collider (RHIC) is a versatile detector designed to study
the properties of nuclear matter at extreme temperatures
and energy densities, obtained in central heavy-ion col-
lisions at ultra-relativistic energies.
these studies is to collect evidence for the existence of the
Quark-Gluon Plasma (QGP) characterized by deconfined
quarks and gluons.
A central goal of
There are several proposed ways to experimentally ver-
ify the existence of a QGP [2]. A general problem is that
many of these signals also can be produced in a hadronic
scenario, albeit special conditions of highly compressed
matter have to prevail. Furthermore, it is not straight-
forward to determine how the various plasma signals are
distorted when the deconfined matter transforms back to
hadronic matter. Recent theoretical investigations [3–5]
predict a drastic decrease of the event-by-event fluctu-
ations of the net charge in local phase-space regions as
a signature of the plasma state. These fluctuations are
not related to the transition itself, but rather with the
charge distribution in the primordial plasma state. The
basic idea is that each of the charge carriers in the plasma
carries less charge than the charge carriers in ordinary
hadronic matter. The charge will thus be more evenly
distributed in a plasma. The main concern of the theo-
retical discussions is how and why the original distribu-
tion survives the transition back to ordinary matter [6,7].
Predictions, for a rapidity coverage ∆y ≥ 1, range up to
2

Page 3

an 80% reduction in the magnitude of the fluctuations,
as measured by the variance of the net charge.
Decays of hadronic resonances influence the net charge
fluctuations, whether or not deconfinement is reached.
In the absence of a QGP, deviations from statistical be-
haviour can be used to determine the abundance of e.g.
ρ and ω mesons [8]. In a hadron gas resonances are ex-
pected to decrease the fluctuations by about 25%. Glob-
ally, fluctuations will be further reduced, since charge is
a conserved quantity. Deviations from a statistical be-
haviour in the net charge are also of relevance when the
phase transition is treated as a semiclassical decay of a
Polyakov loop condensate [9]. Although multiplicity fluc-
tuations have been studied extensively in both hadronic
and nuclear processes [10], net charge fluctuations have
not been addressed experimentally.
In this Letter we report results from an analysis of
net charge fluctuations for particles produced in Au+Au
interactions at√sNN = 130 GeV. The fluctuations are
studied in the variables R = n+/n−, the ratio between
positive and negative particles, and Q = n+−n−, the net
charge [3]. The advantages and disadvantages of these
variables will also be discussed.
Information from one of the PHENIX central track-
ing arms (west) is used in this analysis, where events are
required to have a well-defined vertex close to the cen-
ter of the apparatus (|Z| < 17 cm), as defined by the
two beam-beam counters (BBC). These are Cherenkov-
counters surrounding the beam, placed on both sides 1.44
m from the interaction region, covering the pseudorapid-
ity region 3.0 < |η| < 3.9. Together with the informa-
tion from the two zero-degree calorimeters (ZDC), placed
further away (18 m), the BBC information is used for
off-line centrality selection [11]. A total of about 5×105
minimum bias events has been analyzed. The PHENIX
west arm spectrometer has an acceptance of 0.7 units of
pseudorapidity (-0.35 < η < 0.35) and π/2 radians in
azimuth ϕ. Charged-particle trajectories are recorded in
a multiwire focusing drift chamber (DC) [12]. The com-
bination of reconstructed DC tracks [13] with matching
hits in the innermost pad-chamber plane (PC1) defines
the sign of the charge of the particle and also provides
a high resolution measurement of the transverse momen-
tum pT of tracks originating from the collision vertex.
Tracks with a reconstructed pT less than 0.2 GeV/c have
been excluded from the analysis due to a low reconstruc-
tion efficiency and large contributions from background
sources, as revealed by simulations.
The tracking efficiency and the charge assignment have
been studied using GEANT [14] simulations. Of partic-
ular importance in this context is a realistic description
of the drift chamber response. The drift distances have
been calculated based on a geometric model of the indi-
vidual drift cells. Additional parameters describing the
response of the chamber, i.e. single-wire efficiency, pulse
width, single-hit resolution, and space drift-time relation,
have been extracted from measured data, parameterized,
and applied empirically in the simulation.
RQMD [15] simulations are used to study the detection
efficiency, and the fraction of reconstructed particles that
preserve their charge, as well as to evaluate the results of
the analysis. The charge fluctuations in RQMD are con-
sistent with calculations based on other hadronic mod-
els like UrQMD and HIJING [3]. The overall efficiency
for detecting a charged particle within the acceptance
is found to be around 80% for both positive and nega-
tive particles. Depending on pT, between 70% and 85%
of the reconstructed tracks are in one-to-one correspon-
dence with a primarily produced particle. The remaining
tracks come from secondary interactions in the detector
material and from decays, where the original charge in-
formation is lost.
In each event the numbers of positively charged parti-
cles n+, negatively charged particles n−, and their sum
nchare recorded. In a stochastic scenario, with a fixed
number of charged particles within the acceptance, where
each particle is assigned a random charge (+1 or −1 with
the same probability), the variance of the net charge, Q,
is
V (Q) ≡ ?Q2? − ?Q?2= nch.
The normalized variance in Q is
(0.1)
v(Q) ≡V (Q)
nch
= 1.(0.2)
For the charge ratio, in the stochastic scenario, V (R) ≡
?R2?−?R?2will approach the value 4/nchas nchincreases
and v(R) ≡ nch· V (R) asymptotically approaches 4. A
small asymmetry between positive and negative parti-
cles affects v(R) drastically, whereas the effect on v(Q)
is negligible. If we write the probability p+, that a given
particle has positive charge, in the form p+=
and subsequently p−=
1/2(1+ε),
1/2(1 − ε), we find
v(Q) = 1 − ε2,(0.3)
while the asymptotic value of v(R) is 4 + 16ε + O(ε2).
Detector or reconstruction inefficiencies do not influence
those results in the stochastic scenario. To calculate v(Q)
as a function of multiplicity, Eq. 0.3 can be used and v(R)
can be calculated from
?R? =1
A
nch−1
?
i=1
nch− i
i
?nch
i
?
pnch−i
+
pi
−, (0.4)
and
?R2? =1
A
nch−1
?
i=1
?nch− i
i
?2?nch
i
?
pnch−i
+
pi
−, (0.5)
where A = 1− pnch
to discarding events with n+ or n−equal to zero. The
+
−pnch
−
is the new normalization due
3

Page 4

variance of R, even for a purely stochastic charge distri-
bution, depends on multiplicity and on the fractions of
positive and negative particles.
The data show a small excess of positive particles,
growing proportionally with nch, in qualitative agree-
ment with calculations using RQMD and GEANT. A
part of this excess comes from the intrinsic isospin asym-
metry and a part from secondary interactions in the de-
tector and surrounding materials.
In Fig. 1a, v(R) and v(Q) are displayed as functions
of nch. v(Q) is multiplied by a factor of 4 to compensate
for the asymptotical difference between v(R) and v(Q).
Both v(Q) and v(R) are well described by the results ob-
tained from the stochastic scenario, including the positive
excess, as given by the curves.
Since v(Q) is independent of nchone expects v(Q) to
be close to unity also in representations where other cen-
trality measures are used. On the other hand, since v(R)
depends on multiplicity, it will have a complicated be-
haviour as a function of centrality, making it difficult to
draw any further conclusions. We will thus focus on v(Q)
for the rest of this analysis.
In Fig. 1b, v(Q) is displayed as a function of central-
ity based on the ZDC/BBC information. The full event
sample, corresponding to 92% of the inelastic cross sec-
tion [11], is divided into 20 centrality classes, where each
class corresponds to 5% of the events. Class 20 repre-
sents the most central events. With the increased reso-
lution on the y-axis in Fig. 1b, it is evident that v(Q)
is consistently below unity, and deviates from stochastic
behaviour. The value is, however, far above the most
optimistic QGP predictions v(Q) ∼ 0.2 [3], although one
should keep in mind that our coverage in rapidity is on
the limit for these predictions and that we have only par-
tial coverage in azimuth.
There may be other explanations for deviations from
stochastic behaviour than the one offered by the quark-
gluon plasma. These include global charge conservation
and neutral resonances decaying into correlated pairs of
one positive and one negative particle. Both of these
effects will decrease the fluctuations, and the decrease
will grow in proportion to the experimental acceptance.
In a stochastic scenario, taking global charge conserva-
tion into account, the normalized variance v(Q) becomes
(1−p), where p is the fraction of observed charged parti-
cles among all charged particles in the event. Eventually,
if all charged particles are detected, v(Q) will become 0.
Experimentally we can change the fraction p of parti-
cles within the acceptance by using different regions of
the PHENIX west arm. In Fig. 2, v(Q), for the 10 %
most central events, is displayed as a function of ∆ϕd, i.e.
the chosen azimuthal interval of the spectrometer. For
comparison, the results from RQMD processed through
GEANT are shown. The data and the simulation show a
similar trend. Note that the errors given are correlated,
since the data in one bin are a subset of the data in the
next. The solid line corresponds to the (1−p) dependence
discussed above. The linear relationship between p and
∆ϕd is estimated from the phase-space distribution of
particles in RQMD, including effects from reconstruction
efficiency and background tracks. For larger angles, both
data and the RQMD results lie consistently below the
line, which indicates that effects from resonance decays
are important.
Due to the influence of the magnetic field the positive
and negative particles will have different azimuthal ac-
ceptance. The ∆ϕd study in Fig. 2 thus selects partly
non-overlapping regions of phase space for positive and
negative particles. A remedy for this is to use the recon-
structed ϕ-angle for each particle ϕr, i.e. the azimuthal
direction of the particle at the primary vertex, before it
is deflected by the magnetic field. Figure 3a shows the
acceptance in transverse momentum and ϕrfor positive
and negative particles. By choosing the azimuthal inter-
val ∆ϕr symmetrically around the center of the accep-
tance, a better phase space overlap is achieved for small
azimuthal intervals. In Fig. 3b, v(Q), for the 10 % most
central events, is displayed as a function of ∆ϕr. The
(1−p) dependence, which is no longer linear, is given by
the solid curve. Again data and the RQMD results show
a similar trend, but the deviations from the curve are
larger in this representation, indicating that an overlap
in phase space is of importance. For large values of ∆ϕ,
the acceptance approaches the limits determined by the
boundaries of the tracking arm, and the two representa-
tions are essentially the same.
The effects of the detector inefficiency and background
tracks not assigned the correct charge have been investi-
gated in a Monte Carlo simulation. It is assumed that the
inefficiency independently removes positive and negative
particles with the same probability, and that the back-
ground consists of uncorrelated positive and negative par-
ticles. The reconstruction efficiency and the amount of
background have been determined from the RQMD and
GEANT simulations discussed earlier. Both the ineffi-
ciency and the background contribution have the effect of
diluting the signal and pushing the value of v(Q) closer
to 1. The dilution due to these effects can be treated
as an experimental systematic error, estimated from the
simulations, setting a lower limit on v(Q). For the net
charge fluctuations in the region -0.35 < η < 0.35, pT >
0.2 GeV/c, ∆ϕ = π/2,
v(Q) = 0.965± 0.007(stat.) − 0.019(syst.)(0.6)
is obtained for the 10 % most central events. A linear ex-
trapolation of this value to full azimuthal coverage gives
a value of v(Q) in the range 0.78 - 0.86, in qualitative
agreement with calculations from a hadronic gas. For
comparison, it would be desirable to have a Monte Carlo
model for the QGP which exhibits the predicted reduc-
tion in the charge fluctuations and which could be used
4

Page 5

to study the sensitivity of the method for limited accep-
tance.
To summarize, we have shown that the data behave in
an almost stochastic manner. There are also clear indica-
tions that effects from hadronic decays are seen; the data
are in good agreement with RQMD calculations, which
includes the effects of global charge conservations as well
as neutral hadronic resonance decays. Furthermore, the
data show no centrality dependence, which is in contra-
diction to the expectations from a Quark-Gluon Plasma
scenario. We have clearly demonstrated that the fluc-
tuations of the charge ratio v(R) and of the net charge
v(Q) are well understood in a stochastic model.
however, advise against the usage of the proposed R vari-
able [3], since it unnecessarily complicates the evaluation
of the fluctuations, and the intrinsic decrease of v(R), as
a function of centrality, can be mistaken as a ’plasma fin-
gerprint’. The measured value of v(Q) = 0.965 is far from
the value predicted for a plasma. Even extrapolating the
linear trend seen in the data in Fig. 2 to full azimuthal
coverage, renders values of the fluctuations, which are far
above the proposed values. With the caveat of our lim-
ited acceptance in rapidity, these results clearly indicate
either the absence of a plasma or that the proposed signal
does not survive the transition back to hadronic matter.
We,
We thank the staff of the Collider-Accelerator and
Physics Departments at BNL for their vital contri-
butions. We acknowledge support from the Depart-
ment of Energy and NSF (U.S.A.), Monbu-sho and
STA (Japan), RAS, RMAE, and RMS (Russia), BMBF,
DAAD, and AvH (Germany), VR and KAW (Swe-
den), MIST and NSERC (Canada), CNPq and FAPESP
(Brazil), IN2P3/CNRS (France), DAE and DST (India),
KRF and CHEP (Korea), the U.S. CRDF for the FSU,
and the US-Israel BSF.
∗
Deceased
Not a participating Institution.
[1] PHENIX Collaboration, D. P. Morrison, et al., Nucl.
Phys. A638, 565c (1998).
[2] see e.g. J.-P. Blaizot, Nucl. Phys. A661, 3c (1999).
[3] S. Jeon and V. Koch, Phys. Rev. Lett. 85, 2076 (2000);
M. Bleicher, S. Jeon and V. Koch, Phys. Rev. C62,
061902(R) (2000).
[4] M. Asakawa, U. Heinz and B M¨ uller, Phys. Rev. Lett.
85, 2072 (2000).
[5] H. Heiselberg and A. D. Jackson, Phys. Rev. C63, 064904
(2001).
[6] E. V. Shuryak and M. A. Stephanov, Phys.Rev. C63,
064903 (2001).
[7] K. Fialkowski and R. Wit, Europhys. Lett. 55 (2), 184
(2001).
†
[8] S. Jeon and V. Koch, Phys. Rev. Lett. 83, 5435 (1999).
[9] A. Dumitru and R. D. Pisarski, Phys. Lett. B504, 282
(2001).
[10] see e.g. E.A. De Wolf, I.M. Dremin and W. Kittel, Phys.
Rep. 270, 48 (1996).
[11] PHENIX Collaboration, K. Adcox, et al., Phys. Rev.
Lett. 86, 3500 (2001).
[12] V. G. Riabov, Nucl. Instr. and Methods A 419, 363
(1998).
[13] J. T. Mitchell et al., nucl-ex/0201013, accepted for pub-
lication in Nuclear Instr. and Methods A.
[14] GEANT 3.2.1, CERN program library.
[15] H. Sorge, Phys. Rev. C 52 3291 (1995).
ch
n
020406080100120140
v(R) , 4 v(Q)
0
2
4
6
8
10
12
14
16
18
v(R)
4 v(Q)
a)
Centrality Class
246810 12 14 16 18 20
v(Q)
0.9
0.95
1
b)
FIG. 1.
as functions of nch, together with curves for stochastic be-
haviour. b) The normalized variance v(Q) for different cen-
trality classes, as described in the text.
a) The normalized variances v(Q) and v(R)
5

Page 6

d
ϕ∆
0 10 2030 40506070 80 90
v(Q)
0.94
0.95
0.96
0.97
0.98
0.99
1
1.01
1.02
Data
RQMD
FIG. 2.
imuthal coverage of the detector for data and events simulated
with the RQMD model. The solid line shows the expected re-
duction in v(Q) in the stochastic scenario when global charge
conservation is taken into account. (Angles in degrees.)
v(Q), for central events, as a function of the az-
r
ϕ
-150 -100-50050100 150
(GeV/c)
T
p
0
0.5
1
1.5
2
r
ϕ∆
a)
r ϕ∆
02040 60 80100
v(Q)
0.94
0.95
0.96
0.97
0.98
0.99
1
1.01
1.02
Data
RQMD
b)
FIG. 3.
mentum pT and azimuthal angle ϕr. The solid curves indicate
the acceptance bands for positive and negative particles, re-
spectively. b) The effect on v(Q) of varying the acceptance in
ϕr. The solid curve shows the expectation from global charge
conservation. The 10 % most central events are used for data
and RQMD. (Angles in degrees.)
a) The acceptance in reconstructed transverse mo-
6