1. 1 PHENIX Overview: Status of QGP Terry C. Awes Oak Ridge National Laboratory IX Workshop on High Energy Physics Phenomenology Jan. 3-14, 2006 Bhubaneswar, India

2. 2 Quark Gluon Plasma Lattice QCD predicts transition to deconfined Quark Gluon Plasma phase at ~175MeV Goal of Relativistic Heavy Ion collisions - to produce and characterize QGP state. F. Karsch, Prog. Theor. Phys. Suppl. 153, 106 (2004)

3. 3 Pre-equilibrium Hadronization (Freeze-out) + Expansion ThermalizationQGP phase?Mixed phase ,  e + e -,  +   Hard processes (early stages): Real and virtual photons, high p T partons. PHENIX emphasis Soft hadrons reflect medium properties when interactions stop (chemical and thermal freeze-out).  K  p  n  d,… Central Relativistic Heavy Ion Collision

4. 4 Does the produced matter in RHI reach local equilibrium - allowing a discussion of “matter properties” ? Is deconfined “quark matter” produced? What is the transition temperature? What are the characteristics of the quark matter? Opacity Viscosity Heat capacity - Degrees of freedom - quarks and gluons or more complicated colored objects? etc Studying high density matter with Relativistic Heavy Ion Collisions What have we learned so far?

5. 5 PHENIX detector at RHIC Hadron measurement |  |<0.35 PID using TOF p/K/  separation up to 2 GeV/c (EMCAL) and 4 GeV/c (TOF) Electron measurement |  |<0.35 PID using RICH e/  separation up to p T ~ 4.8 GeV/c Photon measurement |  |<0.35 Two different calorimeters PbSc and PbGl PID using TOF, shower shape, charged veto Muon measurements 1.2 < |  | < 2.4 Two separate arms at forward and backward rapidity

6. 6 Centrality: Nucleon Collisions & Nucleon Participants Centrality selection : Sum of Beam-Beam Counter (BBC, |  |=3~4) and energy of Zero-degree calorimeter (ZDC) Extracted N coll and N part based on Glauber model. Spectator nucleons Participant nucleons Peripheral Central 0-5% 5-10% 10-15% Spectator neutrons Forward Mult.

7. 7 The Final State: Particle Yields Excellent description of relative yields of particles with only 2 parameters. Assuming Chemical Equillibrium: (Chemical Freeze-Out) Braun-Munzinger, Maegestro, and Stachel

8. 8 QGP to Hadron Phase transition? Chemical Freeze-Out Temperature (at  B ) is remarkably close to the Hadron to QGP phase boundary predicted by Lattice QCD. How can chemical equillibrium be attained so rapidly? (Hadronic rates/cross sections too small -- equilibrated in partonic phase?). Why is T Chem so high?

9. 9 PHENIX (PRC72 2005 014903) The Final State: Thermal Freeze-Out Particle Spectra ,K,p,  (“low” p T ) can be described consistently with common Temperature and radial flow velocity profile (max.  T ) Particle Spectra Central Au+Au T=109MeV  T (Max) =0.77

10. 10 The Final State: Thermal Freeze-Out Particle Spectra ( ,K,p,  ) can be described consistently with common T and radial flow  T that depends on overlap volume. Increase in centrality (volume) gives longer lifetime - more rescattering allows transfer from thermal to collective motion, thus larger  T and lower T. Results suggest significant rescattering. Pressure? Thermalization? PHENIX Preliminary PeripheralCentral Temperature  T >

11. 11 Anisotropic Flow: aka Elliptic Flow Elliptic flow = v 2 = 2 nd Fourier coefficient of azimuthal anisotropy  x y z  For non-zero impact parameter, the nuclear overlap volume is  -asymmetric. If the matter interacts strongly, pressure gradients will result and the initial spatial asymmetry will be converted to a momentum asymmetry. Study via angular (  ) correlations between particle and event plane (average  ), or between particle pairs.

12. 12 In-plane Out-of-plane Correlation Function V 2 Harmonic Jet Function Azimuthal Correlations include elliptic flow and di-jet contributions Elliptic Flow via 2-particle correlation More about jet component later …  R(  2-particle correlation with trigger particle selected according to event plane measured in forward rapidity region (BBC).

13. 13 Observe v 2 dependent on p T and particle mass Such dependences expected from hydro (flow) Final State: Elliptic Flow PHENIX Preliminary From Hydro: Observed large v 2 implies strong interactions in the produced matter.

14. 14 Au+Au 200 GeV T=106 ± 1 MeV = 0.571 ± 0.004 c = 0.540 ± 0.004 c R InPlane = 11.1 ± 0.2 fm R OutOfPlane = 12.1 ± 0.2 fm Life time (  ) = 8.4 ± 0.2 fm/c Emission duration = 1.9 ± 0.2 fm/c  2 /dof = 120 / 86 Final State: Putting it together One can obtain a consistent description of the final state -- particle spectra, yields, azimuthal asymmetries, and radii (from  HBT analyses) using a hydro inspired “Blastwave” model (F. Retiere, nucl-ex/0404024).

15. 15 QGP to Hadron Phase transition? Particle Yields: Chemical Freeze-Out at T~175MeV and  B ~ 30 MeV. ??? Consistent description of final state indicates that system lives ~10fm/c, with pion emission occurring in a final burst of ~2fm/c duration at T~100MeV Did system initially enter QGP phase? How far - what T?

16. 16 Probing the Early Phase: Theory PHENIX Huovinen et al Ideal Hydrodynamics (1+1Dim) can describe the particle spectra and v 2 if Equation of State includes QGP phase. EOS without QGP too hard. Parton Cascade (Boltz.Eq.) requires unphysically large cross sections (~45mb). Why? Suggests initial matter of Quarks and Gluons is strongly interacting (sQGP) and non- viscous (Ideal Hydro). “Perfect Liquid” Large initial energy density:  ~ 15-25 GeV/fm 3 (  çrit ~1GeV/ fm 3 )

17. 17

18. 18 Perfect Liquid:  /s =1/4   /s =1/4 

19. 19 PHENIX white paper, NPA757,184(2005) Press release based on RHIC “White Papers” elliptic flow p T spectra p  PHENIX ( Nucl. Phys. A757, 2005 I&II ): Model comparisons show Hydro+ChemEq doesn’t work, Hydro+HadronCascade is better.

20. 20 It ’ s because  /s is small that Ideal Hydro works so well. Hirano & Gyulassy, nucl-th/0506049 Absolute value of viscosity Its ratio to entropy density  : shear viscosity, s : entropy density State of the Art: CGC+3D hydro+hadron cascade (Hirano et al) Reproduces all quite well including rapidity dependence of v 2 for non-peripheral collisions.  /s =1/4 

21. 21 If initial phase is thermalized it should radiate photons. Measure the initial temperature via the spectrum of thermal photon radiation. If you measure T 0 much greater than T C one can be sure to have started in QG phase. Study production of hard probes produced early in the collision to deduce properties of the produced medium that they must traverse Jets, i.e. hard scattered partons. More particularly high p T particles from jet fragmentation. Charm production J/  production Probing the Early Phase: Back to experiment…

22. 22 Photons: Continum Spectrum with Many Sources Turbide, Rapp, Gale EE Rate Hadron Gas Thermal T f QGP Thermal T i Pre-Equilibrium Jet Re-interaction pQCD Prompt Final-state photons are the sum of emissions from the entire history of a nuclear collision. +Weak+EM decay  ’s (    = Bkgd

23. 23   and  - On the way to Measuring Direct  in  s=200 GeV/c Au+Au collisions Au-Au PRL 91 072301 PHENIX Preliminary  Spectra Measure  0 and  distributions- Input to MC to calculate decay   Compare measured  to decay  to extract direct  yield Centrality

24. 24 High-P T   spectra in p+p collisions at 200 GeV/c p-p PRL 91(2003) 241803 Spectra for  0 out to 12 GeV/c compared to NLO pQCD predictions (by W.Vogelsang) pQCD works very well! Calculations with different (gluon) FF’s (Regions indicate scale uncertainty) Large contribution from gluon fragmentation.

25. 25 High-p T  in p+p (d+Au) Collisions at 200 GeV/c As observed for   production, the direct photon measurement in p+p agrees with NLO pQCD calculations. The preliminary d+Au  yield also agrees with - scaled NLO pQCD calculation. Baseline for comparison with Au+Au  results.

26. 26 First RHIC Au+Au Direct Photon Results Direct  excess consistent with NLO pQCD p+p predictions, scaled by the number of binary collisions. Fragmentation?, Bremsstrahlung?, Thermal? PHENIX PRL 94, 232301 (2005)

27. 27 Centrality Dependence of Direct Photons Within errors scaled NLO pQCD describes  yield even to low p T ! Need to improve errors on p+p and Au+Au measurements to search for deviations from pQCD as evidence for other contributions, e.g. thermal  PHENIX PRL 94, 232301 (2005)

28. 28 Thermal Photon Expectations? Hydrodynamical predictions for thermal  (HRG + QGP) plus prompt NLO pQCD prediction yields. Consistent with thermal with QGP with T 0 of 590MeV. Measured  yield is consistent with NLO pQCD prediction with or without thermal contribution. NLO pQCD works too well!? Fragmentation  contributions are large (~50% at 3 GeV/c, 35% at 10 GeV/c). Why not modified? d’Enterria and Peressounko nucl-th/0503054 Central Au+Au

29. 29 Preliminary result from higher statistics Run4 data set. Different method. Can the errors on the data, pQCD and thermal model calculations be reduced sufficiently to deduce initial temperature? Probably not... Can deduce that the photon yield is consistent with various predictions with T 0 max ~ 500-600 MeV T 0 ave ~ 300-400 MeV Thermal Photons: Initial Temperature PHENIX preliminary

30. 30 Hard Probes: Nuclear Effects? Compare A+A to p+p cross section Nuclear Modification Factor: AA “Nominal effects”: R < 1 in regime of soft physics R = 1 at high-p T where hard scattering dominates AA Spectators Participants AA Suppression: R < 1 at high-p T k T broadening (Cronin): R > 1 AA

31. 31 Centrality Dependence R AA for  0 and charged hadrons PHENIX AuAu 200 GeV  0 data: PRL 91 (2003) 072301. charged hadron: PRC 69 (2004) 034909. Large suppression (factor of 5 - huge “nuclear effect”!) implies large energy loss, implies high initial densities… Strong Suppression! Suppression increases with increasing nuclear overlap volume. Increasing density and pathlength. (Difference between   and charged hadrons due to contributions from protons - more later)

32. 32 Quenching of Hard Scattered Partons Hard parton scatterings in nucleon collisions produce jets of particles. hadrons q q leading particle leading particle schematic view of jet production In the presence of a dense strongly interacting medium, the scattered partons will suffer soft interactions losing energy (dE/dx~GeV/fm). Softer fragmentation spectrum: “Jet Quenching” Alternatively, reduced hard scattering rate due to initial state PDF modification? “Gluon Saturation”

33. 33 Theoretical Interpretation of High-p T π 0 Suppression Large suppression implies large energy loss. Model calculations indicate high gluon densities dN g /dy ~ 1100 Implies large energy density (as do also E T measurements)  > 10 GeV/fm 3 well above critical energy density  crit ~ 1 GeV/fm 3 Strong Suppression!

34. 34 A Closer Look at p T Dependence of Direct Photons and   Production for Central Au+Au High p T  yield consistent with binary scaled pQCD in contrast to factor of 5 suppression of   &  yields. Direct  are not suppressed - strong evidence that hadron suppression is due to final state, i.e. parton energy loss. PHENIX PRL 94, 232301 (2005)

35. 35 Baryon “Anomaly”  While   show strong high p T suppression, high p T protons seem not to be suppressed. Surprising result if p and pbar produced from fragmentation.  shows suppression similar to pions. Not a “mass effect”. Can be explained as a quark recombination effect (thermal+fragmentation quarks) - strong evidence that quark matter has been formed.

36. 36 Quark Scaling of Elliptic Flow (v 2 )  Scale baryon/meson v 2 and p T by number of quarks (n q = 3, 2). Observe near universal scaling (better if account for decay contribution to pions). Strongly suggests that collective flow develops during the quark phase.

37. 37 Even heavy quarks flow… “ Measure ” Charm via single electrons after subtracting photon conversion contribution. Recombination model indicates that the charm quark itself flows at low p T. Charm flow supports high parton density and strong coupling in the matter. It is not a weakly coupled gas. Drop of v 2 at high p T perhaps due loss of collectivity or to b-quark contribution. Greco,Ko,Rapp: PLB595(2004)202

38. 38 (3) q_hat = 14 GeV 2 /fm (2) q_hat = 4 GeV 2 /fm (1) q_hat = 0 GeV 2 /fm (4) dN g / dy = 1000 Heavy Quarks also lose energy “ Measure ” Charm via single electrons after subtracting photon conversion contribution. Even heavy quark (charm) suffers substantial energy loss in the matter. The data suggest large c- quark-medium cross section; evidence for strongly coupled QGP? The data provide a strong constraint on energy loss models. Theory curves: (1-3) from N. Armesto, et al., hep-ph/0501225 (4) from M. Djordjevic, M. Gyulassy, S.Wicks, PRL. 94, 112301

39. 39 J/  Suppression: System Size Dependence Models that were successful to describe SPS data assuming disociation in QGP or by comovers fail to describe data at RHIC. Predict too much suppression!

40. 40 The preliminary data are in better agreement with models with the predicted suppression + re-generation (quark recombination) at the energy density of RHIC collisions. Can be tested by measurement of v 2 (J/  )? J/  Suppression: System Size Dependence

41. 41 Di-jet analyses: just a taste… Di-jet tomography is a powerful tool to probe the matter - study yields, widths, p T, and centrality dependence. The shapes of jets are modified by the matter. Mach cone? Cerenkov? Flow? Can the properties of the matter be measured from the shape? Sound velocity Dielectric constant PHENIX preliminary

42. 42 Summary and Conclusions Although there is no “smoking gun” signature for deconfined (QGP) matter, there is now a large body of data that provides many model constraints. We’re on the way towards development of a “Standard Model” of RHI collisions (eg. CGC+3D Hydro + Hadron Cascade). Some experimental observations and inferences at RHIC: Produced matter is strongly interacting: large collective flow and parton energy loss, including charm. Locally thermalized: very likely because of above and success of Hydro interpretation. Quark Gluon phase: very likely per success of quark recombination interpretation of baryon anomaly, flow. Much more to come…