1. Quarkonia and heavy flavor – recent results from STAR Workshop: Hot and Dense Matter in the RHIC-LHC Era, Tata Institute of Fundamental Research, Mumbai, 12–14 February 2008 André Mischke for the STAR Collaboration Amsterdam

2. 2 Outline Charm production cross-section in heavy-ion collisions  Au+Au and Cu+Cu data (new) Heavy quark energy loss in the medium  Non-photonic electron spectra Experimental access to charm and bottom contributions  e-h correlations  e-D 0 correlations (new) Quarkonia production in QCD matter  J/  at high-pT (new)   in Au+Au (new)

3. 3 central R AA data K.J. Eskola et al., Nucl. Phys. A747 (2005) 511 increasing density Jet quenching of light quarks well described by energy loss models  medium is extremely opaque Surface bias effectively leads to saturation of R AA with density Limited sensitivity to the region of highest energy density Need: - insensitive trigger particles: Prompt photons - more penetrating probes: Heavy quarks ? light hadrons Limitations of R AA

4. 4 Heavy quark energy loss Due to their large mass heavy quarks are believed to be primarily produced by gluon fusion M. Gyulassy and Z. Lin, PRC 51, 2177 (1995) - production rates calculated by pQCD - sensitivity to initial state gluon distribution  Does it hold for heavy-ion collisions? Heavy quarks - are ”grey probes” - lose less energy due to suppression of small angle gluon radiation (dead-cone effect) Dokshitzer & Kharzeev, PLB 519, 199 (2001) Radiative and collisional energy loss M.G. Mustafa, PRC72, 014905 A.K. Dutt-Mazumder et al., PRD71, 094016 (2005) parton hot and dense medium Wicks et al, NPA784, 426 (2007)

5. 5 The STAR experiment Centrality & trigger - ZDC - ZDC - CTB+BEMC - CTB+BEMC Tracking and PID - Magnet - Magnet 0.5 Tesla - TPC - TPC |  | < 1.5  p/p = 2-4%  dE/dx /dEdx = 8% - ToF - ToF 60ps resolution - SVT - SVT Energy measurement - Barrel EMC - Barrel EMC fully installed and operational, |  | < 1 Pb/scintillator (21 X 0 ) dE/E ~16%/  E Shower maximum detector

6. 6 Heavy flavour measurements: hadronic decay channel D 0  K  (B.R.: 3.83%) D *  D 0  (B.R.: 61.9%) D   K  (B.R.: 9.51%)  Direct clean probe: signal in invariant mass distribution  Difficulty: large combinatoric background; especially in high multiplicity environments  Event-mixing  Vertex tracker - charm c  ~ 100-200  m - bottom c  ~ 400-500  m

7. 7 Heavy flavour measurements: semi-leptonic decay channels c  ℓ + + X (B.R.: 9.6%) ℓ = e or  D 0  ℓ + + X(B.R.: 6.87%) D   ℓ  + X(B.R.: 17.2%) b  ℓ - + X (B.R.: 10.9%) B   ℓ  + X(B.R.: 10.2%)  Single (non-photonic) electrons sensitive to charm and bottom  Robust electron trigger  Photonic electron background

8. 8 Charm production cross-section in heavy-ion collisions  Au+Au and Cu+Cu data Heavy quark energy loss in the medium  Non-photonic electron spectra Experimental access to charm and bottom contributions  e-h correlations  e-D 0 correlations Quarkonia production in QCD matter  J/  at high-pT   in Au+Au

9. 9 Electron identification - ToF electrons ToF patch (prototype) -    /30 - 0 >  > -1 Electron ID - |1/ß-1| < 0.03 - TPC dE/dx Momentum range: - p T < 4 GeV/c p K e 

10. 10 Muon identification - ToF Low-p T (< 0.25 GeV/c) muons can be measured with TPC + ToF Separate different muon contributions using MC simulations: -K and  decay -charm decay -DCA (distance of closest approach) distribution is very different   0.17 < p T < 0.21 GeV/c 0-12% Au+Au m inv 2 (GeV 2 /c 4 ) Inclusive   from charm  from  / K (simu.) Signal+bg. fit to data

11. 11 Open charm reconstruction - TPC Hadronic decay channel: D 0 → K +  (B.R. 3.8%) - PID in TPC using dE/dx - p T < ~3 GeV/c No reconstruction of displaced vertex up to now Background description using mixed event technique (details in PRC 71, 064902 (2005)) after background subtraction

12. 12 Charm cross section in Au+Au ±± e±e± D0D0 Combined fit covers ~95% of cross section  NN cc = 1.40 ± 0.11 ± 0.39 mb in 12% most central Au+Au  NN cc agrees with NLO calculation progress in determining theoretical uncertainties, R. Vogt, hep/ph: 0709.2531 More data points needed to constraint models STAR p+p and d+Au data, PRL 94, 062301 (2005) STAR preliminary

13. 13 d  NN cc /dy follows binary collision scaling → charm is exclusively produced in initial state; as expected → no room for thermal production STAR preliminary Open charm in Cu+Cu Cu+Cu minbias Run 5

14. 14 Charm production cross-section in heavy-ion collisions  Au+Au and Cu+Cu data Heavy quark energy loss in the medium  Non-photonic electron spectra Experimental access to charm and bottom contributions  e-h correlations  e-D 0 correlations Quarkonia production in QCD matter  J/  at high-pT   in Au+Au

15. hadrons electrons 1.TPC: dE/dx for p T > 1.5 GeV/c only primary tracks (reduces effective radiation length) electrons can be discriminated well from hadrons up to 8 GeV/c allows to determine the remaining hadron contamination after EMC 2.EMC: a)tower E & p/E b)shower Max Detector (SMD) hadrons/Electron shower develop different shape use # hits in SMD to discriminate 85-90% purity of electrons (p T dependent) hadron discrimination power ~10 3 -10 4 electrons  Kp d hadronselectrons Electron identification - EMC Advantage: triggering to enrich high- p T particle sample

16. 16 mass (GeV/c 2 ) Photonic electron background – unlike-sign pairs – like-sign pairs Most of the electrons in the final state are originating from other sources than heavy-flavor decays Dominant photonic contribution - gamma conversions -  0 and  Dalitz decays Exclude electrons with low invariant mass m inv < 150 MeV/c 2  Non-photonic electron excess at high-p T  70% photonic background rejection efficiency e-e-  e+e+ e-e- (assigned as primary track) (global track) (primary track) dca

17. 17 R AA for non-photonic electrons Surprise in central Au+Au: Electrons from D and B decays are strongly suppressed; not expected due to dead-cone effect Smoothly decreases from peripheral to central Au+Au Models implying D and B energy loss are inconclusive yet Phys. Rev. Lett. 98, 192301 (2007)

18. 18 D/B crossing point M. Cacciari et al., PRL95, 122001 (2005) Large uncertainty in D/B crossing point: p T = 3-10 GeV/c Goal - access to underlying production mechanisms - separate D and B contribution experimentally Approach: Two heavy-flavor particle correlations D/B crossing point in FONLL

19. 19 Charm production cross-section in heavy-ion collisions  Au+Au and Cu+Cu data Heavy quark energy loss in the medium  Non-photonic electron spectra Experimental access to charm and bottom contributions  e-h correlations  e-D 0 correlations Quarkonia production in QCD matter  J/  at high-pT   in Au+Au

20. 20 Electron-hadron correlations Azimuthal e-h correlations - Exploit different fragmentation of associated jets e-h correlations - B much heavier than D  subleading electrons get larger kick from B (decay kinematics)  near-side e-h correlation is broadened Extract relative B contribution using PYTHIA simulations B  e (PYTHIA) D   e (PYTHIA) D+B fit to data p+p 200 GeV Run 5 p+p, L  3 pb -1

21. 21 Access to underlying production mechanism Identification and separation of charm and bottom production events using their decay topology and azimuthal angular correlation of their decay products - electrons from D/B decays are used to trigger on charm or bottom quark pairs - associate D 0 mesons are reconstructed via their hadronic decay channel (probe) Electron-D 0 correlations charm production trigger side probe side 3.83% 54% ~10% c cc g g flavor creation (LO) gluon splitting (NLO) g g g g c cc   0

22. 22 LO PYTHIA simulations Near-side - B decays (dominant) Away-side - charm flavor creation (dominant) - small B contribution Away-side - B decays (dominant) - small charm contribution like-sign e-K pairs unlike-sign e-K pairs LO PYTHIA 3<p T trg <7 GeV/c

23. 23 K  invariant mass distribution w/o electron trigger combinatorial background is evaluated using like-sign pairs STAR preliminary dn/dm D 0 reconstruction: dE/dx cut (±3  ) around Kaon band Significant suppression of combinatorial background: factor 200 S/B = 14% and signal significance = 3.7  D 0 +D 0 w/ non-photonic electron trigger position: 1892 ± 0.005 GeV/c 2 width: 16.2 ± 4.7 MeV/c 2

24. 24 like-sign e-K pairs Relative B  e contribution Model uncertainties not included yet Good agreement between different analyses Data consistent with FONLL within errors Small gluon splitting contribution Comparable D and B contributions  Au+Au: suggests significant suppression of non-photonic electrons from bottom in medium essentially from B decays only  75% from charm  25% from beauty

25. 25 Charm production cross-section in heavy-ion collisions  Au+Au and Cu+Cu data Heavy quark energy loss in the medium  Non-photonic electron spectra Experimental access to charm and bottom contributions  e-h correlations  e-D 0 correlations Quarkonia production in QCD matter  J/  at high-pT   in Au+Au

26. 26 Quarkonia production in QCD matter H. Satz, HP2006 “Melting” of Quarkonia states in QGP - color screening of static potential between heavy quarks: J/  suppression: Matsui and Satz, Phys. Lett. B 178 (1986) 416 - suppression states is determined by T C and their binding energy  QCD thermometer Sequential disappearance of states - color screening  deconfinement - QCD thermometer  access to properties of QGP Lattice QCD: Evaluation of spectral functions RHIC Charmonia: J/ ,  ’,  c Bottomonia:  (1S),  (2S),  (3S)‏

27. 27 J/  production at high-pT J/  /c H. Liu, K. Rajagopal and U.A. Wiedemann PRL 98, 182301(2007) and hep-ph/0607062 Data consistent with no suppression at high p T : R AA (p T > 5 GeV/c) = 0.9 ± 0.2 While at low-p T R AA = 0.5-0.6 (PHENIX) Indicates R AA increase from low p T to high p T But, most models expect a decrease R AA at high p T - two component approach by X. Zhao and R. Rapp, hep-ph/07122407 - AdS/CFT by H. Liu, K. Rajagopal and U.A. Wiedemann, PRL 98, 182301(2007) and hep-ph/0607062

28. 28 STAR detector can resolve individual  states Low statistics currently  yield is extracted from combined  +  +  states FWHM ≈ 400 MeV/c 2  mass resolution in STAR   e + e - RHIC II with  20 nb -1 will definitely allow separation of 1S, 2S and 3S states Measurement will greatly benefit from the future Muon Telescope Detector upgrade simulation with inner tracker (X/X 0  6%)

29. 29  (1s+2s+3s) signal in p+p unlike-sign pairs — like-sign pairs background subtracted  (1s+2s+3s) d  /dy at y=0 Peak width consistent with expected mass resolution preliminary Run 6 p+p, L  9 pb -1

30. 30 Mid-rapidity  (1s+2s+3s) cross-section STAR preliminary p+p 200 GeV y d  /dy (nb) Counts Integrated yield at mid-rapidity: |y|<0.5  (1s+2s+3s)  e + e - : BR ee  d  /dy = 91 ± 28(stat.) ± 22(sys.) pb Consistent with NLO pQCD calculations and world data trend

31. 31 Sample  - triggered event  e + e - m ee = 9.5 GeV/c 2 (offline mass)‏ cosθ = -0.77 (offline opening angle)‏ E 1 = 6.6 GeV (online cluster energy hardware trigger)‏ E 2 = 3.2 GeV (online cluster energy software trigger)‏ charged tracks electron momentum > 3 GeV/c  in Au+Au (Run 7)

32. 32  signal in Au+Au Run 7 Au+Au, L  300  b -1 First  measurement in nucleus-nucleus collisions 4  signal These data should provide a first measurement of R AA after corrections for efficiency are completed

33. 33 Charm production cross-section - follows binary collision scaling  no room for thermal production - agrees with FONLL within errors Non-photonic electron spectra - energy loss in heavy-ion collision is much larger than expected - models invoking radiative + collisional energy loss get closest to data Electron-D 0 azimuthal correlations - first heavy-flavor particle correlation measurement at RHIC - B and D contributions similar at p T >  5 GeV/c; consistent with FONLL - STAR data + MC@NLO simulations: Small gluon-splitting contribution Quarkonia -  cross-section in pp consistent with pQCD and world data trend - first  signal in Au+Au - Next: absolute cross-section in p+p, d+Au (Run 8) and Au+Au and R AA Summary

34. 34 More exciting results are about to come with the STAR detector upgrades - Heavy Flavor Tracker - Full barrel Time-of-Flight Outlook SSD HFT  e  improved D → K  reconstruction excellent electron ID (J/  range 0.2<p<4 GeV/c) J/  trigger in Au+Au low-p T muon ID (0.2 GeV/c) active pixel sensor technology 2 layers CMOS Active Pixel sensors pixel dimension: 30  m  30  m resolution: ~10  m ToF

35. 35 The STAR collaboration 52 institutes from 12 countries, 582 collaborators

36. 36 Backup

37. 37 Compilation: Charm cross section at RHIC Yifei Zhang, QM2008

38. 38 Extraction of charm cross-section Inelastic cross section in p+p (UA5) Conversion to full rapidity Ratio obtained from e + e - collisions Number of binary collisions (Glauber)

39. 39 Obtaining the Charm Cross-Section   cc From D 0 mesons alone:  N D0 /N cc ~ 0.54  0.05  Fit function from exponential fit to m T spectra Combined fit:  Assume D 0 spectrum follows a power law function  Generate electron spectrum using particle composition from PDG  Decay via routines from PYTHIA  Assume: dN/dp T (D 0, D*, D , …) have same shape only normalization In both cases for d+Au  p+p:   pp inel = 41.8  0.6 mb  N bin = 7.5  0.4 (Glauber)  |y|<0.5 to 4  : f = 4.7  0.7 (PYTHIA)  R dAu = 1.3  0.3  0.3 PRL 94, 062301 (2005)

40. 40 Non-photonic electron spectra FONLL calculation factor of about 5 lower than p+p data Spectra shape of p+p spectrum is well described Large uncertainty over where B  e dominates over D  e. p+p at  s NN = 200 GeV Phys. Rev. Lett. 98, 192301 (2007)

41. 41 Latest efforts to describe the data Radiative energy loss with typical gluon densities is not enough Djordjevic et al., PLB 632 (2006) 81 Models involving a very opaque medium agree better Armesto et al., PLB 637 (2006) 362 Collisional energy loss / resonant elastic scattering Wicks et al., nucl-th/0512076 and van Hees and Rapp, PRC 73 (2006) 034913 Heavy quark fragmentation and dissociation in the medium → strong suppression for c and b Adil and Vitev, hep-ph/0611109 Comparisons to models Theoretical explanations are inconclusive Is it possible to determine B contribution experimentally?

42. 42 K-K- bb b B-B- D *0 D0D0 ++ ee e-e- B+B+ -- K+K+ D0D0 Electron tagged correlations: bottom production Near- and away-side correlation peak expected for B production like-sign pairs  near-side correlation unlike-sign pairs  away-side correlation Charge-sign requirement on trigger electron and decay Kaon gives additional constraint on production process

43. 43

44. 44 NLO Process: Gluon splitting FONLL/NLO calculations only give single inclusive distribution  no correlations PYTHIA PYTHIA is not really adequate for NLO predictions STAR measurement of D* in jet  access to charm content in jets Gluon splitting rate consistent with pQCD calculation  Small contribution from gluon splitting at top RHIC energy PYTHIA PYTHIA charm production E. Norrbin & T. Sjostrand, Eur. Phys. J. C17, 137 (2000) STAR Preliminary Mueller & Nason PLB 157 (1985) 226 P29: X. Dong, Charm Content in Jets

45. 45 Δ  (c,c) Δ  (e,D 0 ) NLO QCD computations with a realistic parton shower model  Remarkable agreement of the away-side peak shape between LO PYTHIA and MC@NLO  Gluon-splitting contribution is small - S. Frixione, B.R. Webber, JHEP 0206 (2002) 029 - S. Frixione, P. Nason, and B.R. Webber, JHEP 0308 (2003) 007 - private code version for charm production MC@NLO LO PYTHIA like-sign e-K pairs 3<p T <7 GeV/c MC@NLO predictions for charm production

46. 46  trigger in STAR Real Data, p+p Run V Large acceptance Fast L0 trigger (hardware) - Select events with at least one high energy EMC cell (E~4 GeV) L2 trigger (software) -EMC cell clustering -Using topology (cosθ) and mass reconstruction in real time Very clean to trigger up to central Au+Au Trigger efficiency ~80% Offline: Match TPC tracks to triggered EMC cells

47. 47   =  geo ×  L0 ×  L2 ×  2 (e)×  mass  cross section and uncertainties geometrical acceptance  geo 0.263±0.019 efficiency L0 trigger  L0 0.928±0.049 efficiency L2 trigger  L2 0.855±0.048 efficiency of e reconstr.  2 (e) 0.47±0.07 efficiency of mass cut  mass 0.96±0.04  0.094±0.018  L dt = (5.6±0.8) pb -1 N  = 48±15(stat.)