1. Elliptic Flow in PHENIX
Hiroshi Masui
for the PHENIX collaboration
CIPANP 2006, Westin Rio Mar Beach
Puerto Rico
May 30 – Jun 3, 2006

2. Why Elliptic Flow ? Study the properties of matter created at RHICZ
Study the properties of matter created at RHIC
The probe for early time
The dense nuclear overlap is ellipsoid at the beginning of heavy ion collisions
Pressure gradient is largest in the shortest direction of the ellipsoid
Spatial anisotropy  Momentum anisotropy
Signal is self-quenching with time
Elliptic flow (v2) is defined by the 2nd coefficient of Fourier expansion
Reaction plane Y X Py Pz Px
Hiroshi Masui / University of Tsukuba

3. Outline direct  D Elliptic Flow, v2 , K, p Elliptic Flow, v2
partonic
hadronic
direct  D
Elliptic Flow, v2
, K, p
time
Partonic degrees of freedom, Thermalization, ..
Elliptic Flow, v2
Bulk, early probe
Charged hadrons
give us a base line
 Meson
Small interaction cross section, long lived life time (~40 fm/c)
Penetrating probes
Heavy flavor electrons (charm)
Direct 
Hiroshi Masui / University of Tsukuba

4. Baseline : charged hadrons
Charged hadrons, v2 vs pT
PHENIX : PRL 91, (2003)
v2 ~ k  
Large elliptic flow at RHIC
Consistent with hydrodynamics with rapid thermalization,  ~ 1 fm/c
v2 scales initial eccentricity () of reaction zone
Hiroshi Masui / University of Tsukuba

5. Hydro scaling Scaling holds up to KET = 1 GeV Meson/Baryon v2
M. Issah, A. Taranenko, nucl-ex/
Hydro scaling of v2
K0S,  (STAR) : PRL 92, (2004)
 (STAR) : PRL 95, (2005)
, K, p (PHENIX) : preliminary
Meson v2
Baryon v2
* KET ~ mT – m0 at y ~ 0
Scaling holds up to KET = 1 GeV
Meson/Baryon v2
Possible hint of partonic degrees of freedom
Hiroshi Masui / University of Tsukuba

6. 1.  meson v2
Hiroshi Masui / University of Tsukuba

7. Clear  signal  reconstruction via K+K- decay channel
S/N ~ 0.3
Centrality 20 – 60 %
S/N is good
Event plane resolution is good
Separation between meson/baryon v2 is good
v2 do not be varied too much.
Before subtraction
Signal + Background
Background
After subtraction
Hiroshi Masui / University of Tsukuba

8. “Meson” type flow !  v2 vs pT
Hydro. mass ordering for pT < 2 GeV/c
v2() ~ v2(), v2(K) for pT > 2 GeV/c
Mass Number of constituent quarks
Consistent with the description by quark coalescence, recombination models
Hiroshi Masui / University of Tsukuba

9. Hint of partonic d.o.f Hydro scaling of v2 Hydro + Nquark scaling
K0S,  (STAR) : PR 92, (2004)
 (STAR) : PRL 95, (2005)
 (STAR) : preliminary
, K, p, 0, d (PHENIX) : preliminary
Hydro + Nquark scaling
Works for a broad range of KET
 meson also follow the scaling
Partonic degrees of
freedom
WWND 2006, M. Issah
Hiroshi Masui / University of Tsukuba

10. 2. Heavy flavor electron v2
Hiroshi Masui / University of Tsukuba

11. Clean heavy flavor electron
Cocktail subtraction
Converter subtraction
Signal / Background ratio
Run4
Run2
Two different techniques show an excess heavy flavor electron signal
Signal/Background > 1 for pT > 1 GeV/c
Hiroshi Masui / University of Tsukuba

12. Hint of Charm flow e v2 vs pT Data indicate finite v2 of charm quark
V. Greco, C. M. Ko, R. Rapp: PL B 595, 202 (2004)
e v2 vs pT
Data indicate finite v2 of charm quark
Suggest thermalization of c quark as well as light quarks
What is the origin of high pT drop ?
Hiroshi Masui / University of Tsukuba

13. b quark contribution ?
H. Hees, V. Greco, R. Rapp: PRC 73, (2006)
e v2 vs pT
(1)
input
D or B meson v2
decrease, (2) flat at high pT
(2) v2
D  e
output
B  e (1)
Simulation
B  e (2)
SQM 2006, S. Sakai
pT (GeV/c)
High pT drop might be explained by B meson contribution
Hiroshi Masui / University of Tsukuba

14. 3. Direct  v2
Hiroshi Masui / University of Tsukuba

15. Clear Direct  signal Large direct  excess
PRL 94, (2005)
Large direct  excess
0 is suppressed but direct  not
Both results are consistent with pQCD
Hiroshi Masui / University of Tsukuba

16. Possible 3 different scenarios
 v2 vs pT
PRL 96, (2006)
Direct  v2
1. Hard scattering
v2 = 0
2. Parton fragmentation
v2 > 0
3. Bremsstrahlung
v2 < 0
Inclusive 
Consistent with expected  v2 from hadron decays
Hiroshi Masui / University of Tsukuba

17. v2direct = 0 ?  v2 vs pT Direct  v2
PRL 96, (2006)
Direct  v2
Direct  excess ratio is consistent with v2BG/v2inclusive, suggest v2direct = 0
Favors prompt photon production for dominant source of direct 
Hiroshi Masui / University of Tsukuba

18. Conclusions
Elliptic Flow is powerful tool to study hot and dense matter at RHIC
 meson  Partonic degrees of freedom
Hydro. mass ordering for pT < 2 GeV/c
For pT > 2 GeV/c, v2() prefer quark composition not mass
Hydro + Nquark scaling works for  meson as well as other hadrons
Heavy flavor electron  Thermalization
Consistent with non-zero charm flow, suggest thermalization of c quark as well as light quarks
Bottom contribution at high pT need to be studied experimentally
Direct   Coming soon …
v2direct = 0, consistent with the scenario of direct  production from initial hard scattering
Year-4 data may enable us to reduce statistical error bars, and extend pT reach, and to measure thermal  v2
Hiroshi Masui / University of Tsukuba

19. Thank you
Hiroshi Masui / University of Tsukuba

20. Back up
Hiroshi Masui / University of Tsukuba

21. Converter Subtraction methodNe
0.8%
1.1%
? %
1.7%
With converter
Photonic
electron
W/o converter
Photon Converter (Brass 1.7 % X0) around
beam pipe
Conversion from beam pipe
Dalitz :
0.8 % X0 equivalent
Non-photonic electron X0
Source of background electrons
Photon conversion 0     e+e-
Dalitz decay (0 ee,  ee, etc)
Di-electron decays of , , 
Kaon decays
Conversion of direct photons
Hiroshi Masui / University of Tsukuba

22. Inclusive electron v2 (w., w/o. converter)
Converter subtraction method
Inclusive electron v2 w. and w/o. converter.
Clear difference between them.
Electron v2 with converter include large photonic component.
Hiroshi Masui / University of Tsukuba

23. Inclusive & photonic electron v2
pT < 1 GeV/c
 Converter method
pT > 1 GeV/c
Simulation
v2(inclusive) < v2 (photonic)
 v2(non-photonic) < v2(photonic)
photonic e v2
inclusive e v2 (w/o. converter)
Inclusive electron v2
Hiroshi Masui / University of Tsukuba

24. Centrality dependence
Non zero heavy flavor electron v2 !
Hiroshi Masui / University of Tsukuba

25. Run4 inclusive  v2 v2 pT 0 inclusive 
Hiroshi Masui / University of Tsukuba

26. Independent on system size
Eccentricity scaling
Eccentricity scaling
A wide range of centrality
Independent of system size
Integrated v2  eccentricity
Reduce systematic error from eccentricity calculation
Cancel systematic error by the ratio of v2(pT) / (integrated v2)
Hiroshi Masui / University of Tsukuba

27. Hadron identification
Time-of-flight
|| < /4, || < 0.35
Timing resolution ~ 120 ps
/K separation ~ 2 GeV/c
K/p separation ~ 4 GeV/c
EM Calorimeter
|| < /2, || < 0.35
Timing resolution ~ 400 ps
/K separation ~ 1 GeV/c
K/p separation ~ 2 GeV/c
Hiroshi Masui / University of Tsukuba

28. The PHENIX experiment Acceptance Centrality Event plane Tracking
Central arm: || < , || < 0.35
Centrality
Beam-Beam Counter, Zero Degree Calorimeter.
Event plane
BBC
Tracking
Drift chamber, Pad chambers.
Particle identification
Electron
Electro-Magnetic Calorimeter
Hadron
Time-of-Flight, Aerogel Cherenkov Counter, EMCal
Hiroshi Masui / University of Tsukuba