Collective Flow vs. Hard Processes in Relativistic Heavy Ion Collisions

1
Single Particle Spectra from AuAu Collisions at

J. Burward-Hoy Lawrence Livermore National Labo
ratory
for the PHENIX Collaboration
Motivation Identifying ??, K?, p? in PHENIX Res
ults and discussion
Conclusion Outlook
2
Single particle spectra sensitive to dynamics
In the initial collisions between participating
nucleons
If sufficient secondary scatterings
between produced hadrons

• jets form and fragment into high pT hadrons
• high pT hadrons sensitive to hot, dense nuclear
  medium
  
• if partons experience energy loss, jet formation
  decreases, hadron yield at high pT decreases (Jet
  Quenching)
  
• use perturbative QCD parton models tuned to
  reproduced pp and pA data (X.N. Wang et al).

local thermal equilibrium collective motion low
pT hadrons sensitive to collective behavior due
to secondary scatterings the low pt particles bo
osted to higher momenta spectra of heavier mass p
articles broader than for lighter mass particles
assume/use hydrodynamics
3
The PHENIX Detector
-0.35 Event and Trigger Selection BBC and ZDC both
fire
zvtx
4
Event centrality how head-on is the collision?
The fraction of energy deposited in
the Zero-degree Calorimeters.
A. Denisov
The fraction of counts in the Beam-beam
Counters.
For each collision class, a Glauber model
calculation is used to map centrality to the
number of participants and the number of
collisions.
5
Detecting ??, K?, p? in PHENIX
DC main bend plane
DC resolution ?p/p 0.6 ? 3.6 p
TOF resolution 115 ps
pt p sin(?0)
PC1 and Event vertex polar angle
6
Corrections to the Raw Spectra
Used MC single particles and track embedding to
correct for Tracking inefficiencies and momentu
m resolution Geometrical acceptance Decays i
n flight (?s and Ks)
Correction is p and PID dependent
Particle Acceptance
0.2 GeV/c
Note absolute momentum scale is known better
than 2.
-0.35 7
p/? Spectral Crossing
5 central
Nucleons dominate mesons at 1.5-2 GeV/c Pt r
egion where this occurs centrality dependent?
Will confirm with more statistics from Year-2 run
most peripheral (60-92)
8
?0 measurement

• Two types of electromagnetic calorimeter
  detectors are used PbGl , PbSc
  
• Measure decay channel
• ?0 ? ? ?
• Measure pairs of ? showers
• binned in pair pT, m??
• Spectra corrected for
• Energy resolution
• Shower cluster overlap
• Analysis cuts
• Dead detector areas
• Acceptance
• Energy scale known

At each pT, invariant mass distributions deter
mined
Combinatorial background is subtracted.
Yield is determined after fitting Gaussian and
integrating
Please see F. Messers talk (Jan. 25, 2002)
9
10 Central PHENIX Spectra
10
Mean transverse momentum
20/- 5 increase
20/- 5 increase
Open symbols pp collisions

• Mean pt ? with Npart , m0 ? radial flow
• Relative increase from peripheral to central same
  for ?, K, (anti)p
  
• (Anti)proton significant ? from pp collisions

11
Modeling the Source

• Interaction region
• Assembly of classical boson emitting sources in
  space-time region
  
• The source S(x,p) is the probability boson with p
  is emitted from x
  
• Determines single-particle momentum spectrum
• E d3N/dp3 ? d4x S(x,p)
• Determines the HBT two-particle correlation
  function C(K,q)
  
• C(K,q) 1 ? d4x S(x,K) exp(iqx) 2/ ? d4x
  S(x,K) 2
  
• where K ½(p1 p2) (KT, KL), q p1 p2
• The LCMS frame is used (KL 0)
• In the hydrodynamics-based parameterizations
  assume something about the source S(x,p)
  
• Gaussian particle density distribution
• Linear flow (rapidity or velocity) profile
• Instantaneous freeze-out at constant proper time
  (sharp)
  

12
Important Assumptions Used. . .

• Assume Boltzmann statistics
• Integration over ? is done exactly.
• Boost invariance (vL z/t). Space-time rapidity
  equals flow rapidity
  
• Infinitely long in y.
• In longitudinal, co-moving system (LCMS), y and
  ?L 0.
  
• Integrate using the modified K1 Bessel function
  
  
• Assume azimuthal symmetry in ? and integrate
  using the modified I0 Bessel function
  
• What is left? Integration over the radial
  coordinates. . .

13
Calculating the Single Particle Spectra
1/mt dN/dmt A ? f(?) ? d? mT K1( mT /Tfo cosh
? ) I0( pT /Tfo sinh ? )
?t
Integration variable ? r/R Geometric radius R

Three parameters normalization A freeze-out t
emperature Tfo
surface velocity ?t
1
Particle density distribution f(?)
box or Gaussian Linear velocity profile ?t(?)
?t?
Boost ?(?) atanh( ?t(?) )
In order to minimize contributions from hard
processes fit mt-m0
Ref E. Schnedermann, J. Sollfrank, and U.
Heinz, Phys. Rev. C 48, 2462 (1993)
Ref S. Esumi, S. Chapman, H. van Hecke, and N.
Xu, Phys. Rev. C 55, R2163 (1997)
14
Fitting the Single Particle Spectra
Simultaneous fit (mt -m0 ) PHENIX Preliminary
PHENIX Preliminary
Exclude ? resonances by fitting pt 0.5 GeV/c
The resonance region decreases T by 20 MeV.
This is no surprise! Sollfrank and Heinz also
observed this in their study of SS collisions at
CERN energies. NA44 also had a lower pt cut-off f
or pions in PbPb collisions.
PHENIX Preliminary
15
5 Central Single Particle Spectra
T 122?4 MeV ?t 0.72 ?0.01 ?2/dof 30.0/40.0
Note For the 5-15 centrality (in S.C.
Johnsons talk),
Tfo 125 MeV and ?t 0.69
16
Suppression of hadron production high pT
nucl-ex/0109003 (See F. Messers talk Jan. 25,
2002)
Ratio of yield to binary-scaled yield from NN
collisions. Suppression of high pT ?0 is larger
than for hadrons Difference is due to the nucl
eons in the hadron spectra
J. Velkovska, DNP Fall 2001
Strong radial flow in the nucleon spectra and
suppression of pion yield at high pt may cause
the observed crossing in the spectra.
17
Particle Ratios and Npart
PHENIX Preliminary

PHENIX Preliminary
Independent of Npart Negative to positive for ?,
K, p
PHENIX Preliminary
PHENIX Preliminary
Dependent on Npart K/? and p/?
PHENIX Preliminary
18
Hadron Yields

• dN/dy scaled by Np pair rises faster for (anti)p
  than pions with Npart
  
• Note scaled for clarity

20 /- 6
65 /- 7
Similar Np dependence as observed at both the AGS
and SPS energies Larger measured range in Np Not
e scaled for clarity Plot courtesy of Vince Cian
ciolo
19
Total charged multiplicity
PHENIX Preliminary
PRL 86 (2001) 3500
PRL 87 (2001) 112303
PRL 85 (2000) 3100
Is the increase in dNch/dh / (0.5 Np) due to
hard processes ? How does the yield per participa
nt behave for different particles ?
5 GeV/fm3 (70 higher than at CERN-SPS
energies)
20
Summary and Conclusion

• In addition to charged pions and kaons, PHENIX
  also measured (anti)protons, ?0, and h up to 3.5,
  4, 5 GeV/c
  
• All measurements are internally consistent
• Protons dominate over pions at high pt in central
  collisions
  
• Mean pT increases with Npart for all particles
• dN/dy per participant pair rises faster for
  (anti)protons than for pions
  
• The data suggest radial flow at RHIC
• Hadron suppression at high pt
• Different for mesons and baryons
• Suppression of high pt pions and strong radial
  flow in the protons may explain the observed
  crossing region in the spectra
  

21
Outlook

• In July 2001, Year-2 run started in PHENIX with
  both a muon spectrometer arm and fully active
  central spectrometer arms.
  
• New detectors include
• 8 complete sectors of EMCAL
• (6 PbSc, 2 PbGl)
• Rebuilt drift chamber
• Multiplicity vertex detector (MVD)
• South Muon Spectrometer
• Upgraded trigger and data acquisition
• Both AuAu and pp collisions at

22
Minimum Bias Spectra
23
Comparing apples to apples. ..y and ?
?
24
The mass-squared width how particle identifica
tion is done.
momentum p pathlength L m p/?? where ?
L/ct time-of-flight t mmeas2 - mcent2 2?m22
Momentum resolution
Time-of-Flight resolution
H. Hamagaki
25
From HydrodynamicsRadial Flow Velocity Profiles
Velocity profile linear on freeze-out hyper
surface
? ? radius r r/R
At each snapshot in time during the expansion,
there is a distribution of velocities that va
ry with the
radial position r
Plot courtesy of P. Kolb