Quarkonia (and heavy flavors) at RHIC Andry Rakotozafindrabe LLR – École Polytechnique QGP – France, Étretat 2006

05/07/2006Andry Rakotozafindrabe – LLR2 Physics motivation – the starting point  What are the properties of the hot and dense matter produced in relativistic heavy ion collisions ?  c and b are produced in the initial parton collisions, so they can be used to probe the created medium : open charm (or beauty) energy loss  energy density Only a few words here ( from a newbie ☺), (quarkonia) suppressed by color screening  deconfinement We will focus on quarkonia in this talk, especially on hidden charm.

Heavy quarks

05/07/2006Andry Rakotozafindrabe – LLR4 Heavy quarks dynamic  Measuring non-photonic electrons at RHIC: R AA R AA v 2 v 2

05/07/2006Andry Rakotozafindrabe – LLR5 Heavy quark (radiative?) energy loss  Agreement between both experiments  Significant reduction at high pT suggests sizeable heavy quark energy loss.  Data favors a strong transport coefficient* q_hat ~14 GeV²/fm (radiative energy loss only model)  large initial gluon density ~ 3500! Too high! Should take into account the collisionnal energy loss? (see M. Djordjevic, nucl-th/ ) *q_hat  density of scattering centers in the medium (1-3) N. Armesto et al., PRD 71, (charm contribution only) (4) M. Djordjevic et al., PRL 94, (beauty included) J. Dunlop, J. Bielcik; QM05 (3) q_hat = 14 GeV 2 /fm (1) q_hat = 0 GeV 2 /fm (4) dN g / dy = 1000 (2) q_hat = 4 GeV 2 /fm V. Greene, S. Butsyk; QM05

05/07/2006Andry Rakotozafindrabe – LLR6 Charm flow  Significant flow observed for heavy flavor electrons  Indication for reduction of v 2 at p T > 2 GeV/c. Due to beauty contribution? S.Butsyk QM05 Greco, Ko, Rapp, PLB 595 (2004) 202M. Djordjevic et al., Phys.Lett.B632 (2006) 81 c and b quark p T distributions at mid- rapidity before fragmentation : b contribution is dominant at high pT dN g /dy=100 0

Quarkonia Mostly J/  (mostly PHENIX results)

05/07/2006Andry Rakotozafindrabe – LLR8 Screening the J/  in a QGP  Sequential dissociation as the temperature (or energy density) increases :  Energy density (τ 0 = 1fm) vs the max. √s for SPS, RHIC and LHC ’’J/  χcχc Satz, hep-ph/   Production ~ 60% direct production J/  ~ 30% via χ c  J/  + x ~ 10% via  ’  J/  + x   Temperature of dissociation T d for χ c and  ’: T d ~ 1.1 T c for J/  : T d ~ 1.5 to 2 T c

05/07/2006Andry Rakotozafindrabe – LLR9 Physics motivation – a few complications  Sensitive to: Initial state Modification of the parton distribution functions (shadowing, CGC) p T broadening (Cronin effect) Parton energy loss in the initial state ? Final state “Normal” nuclear absorption Absorption by (hadronic ?) comovers ? Color screening ? In-medium formation (recombination) ? Flow ? J/  or  + feed-down b, ,  c  J/  + x  quarkonia production g+g fusion dominant at RHIC energies

05/07/2006Andry Rakotozafindrabe – LLR10 Production baseline : p+p  J/  Total cross section in p+p  = 2.61  0.20  0.26 µb in agreement with COM Phys. Rev. Lett. 96, y Cross section vs rapidity follow PYTHIA shape Cross section vs pT = 2.51  0.21 (GeV/c)² energyrapidityCross section vs :pTpT p+p  J/  measurement will be used as a reference for A+B  J/  :

05/07/2006Andry Rakotozafindrabe – LLR11  Available d+Au data : Weak shadowing (modification of gluon distribution) and weak nuclear absorption ( ) Weak shadowing (modification of gluon distribution) and weak nuclear absorption (σ abs ~ 1mb favored) gluons in Pb / gluons in p x Shadowing Anti Shadowing Nucl. Phys. A696 (2001) Cold nuclear effects : d+Au  J/  Vogt, PRC71, (2005), Kopeliovich, NP A696, 669 (2001) Phys. Rev. Lett. 96 (2006), Low x 2 ~ (shadowing region) y ~ 0 : intermediate x Au ~ xdxd x Au J/  North y > 0 y > 1.2 : low x Au ~ xdxd x Au J/  South y < 0 rapidity y y < -1.2 : large x Au ~ 0.090

05/07/2006Andry Rakotozafindrabe – LLR12 σ abs =1 mb RHIC : beyond cold nuclear effects ?  Au+Au data : even compared to the « worst » σ abs ~ 3mb case Factor 2 of suppression Au+Au bin Factor 2 of suppression beyond cold effects in the most central Au+Au bin σ abs =3 mb Suppression Factor ~ 2 Number of participants Cold nuclear matter predictions from Vogt, nucl- th/ (shadowing + σ abs = 1, 3mb)

05/07/2006Andry Rakotozafindrabe – LLR13 RHIC vs SPS (I) : raw comparison  SPS : √s ~ 17 GeV i.e. a factor 10 below RHIC √s ~ 17 GeV i.e. a factor 10 below RHIC Cold effect = σ abs = 4.18 ± 0.35 mb Cold effect = normal nuclear absorption σ abs = 4.18 ± 0.35 mb Maximum ε ~ 3 GeV/fm 3 (τ 0 = 1) Maximum ε ~ 3 GeV/fm 3 (τ 0 = 1)  Compare to RHIC : Cold effect = σ abs ~ 1mb (Vogt, nucl- th/ ) Cold effect = shadowing + nuclear absorption σ abs ~ 1mb (Vogt, nucl- th/ ) Maximum ε ~ 5 GeV/fm 3 (τ 0 = 1), higher than at SPS, but still, the ! Maximum ε ~ 5 GeV/fm 3 (τ 0 = 1), higher than at SPS, but still, the same pattern of J/  suppression ! SPS normalized to NA51 p+p value (NA60 preliminary points from Arnaldi, QM05). NA50 Pb+Pb NA60 In+In PHENIX Au+Au PHENIX Cu+Cu PHENIX cold effect NA50 cold effect PHENIX : |y|~1.7

05/07/2006Andry Rakotozafindrabe – LLR14 RHIC vs SPS (II) : extrapolating suppression models  Suppression models in agreement with NA50 data overestimate the suppression when extrapolated at RHIC energies : quite striking for mid and most central Au+Au bins quite striking for mid and most central Au+Au bins already the case for Cu+Cu most central bins ? already the case for Cu+Cu most central bins ? (Hadronic?) co-mover scattering Direct suppression in a hot medium : Cu+CuAu+Au

05/07/2006Andry Rakotozafindrabe – LLR15 Some recombination effects ?  Adding some regeneration that partially compensates the suppression : there is a better agreement between the model and the data. Direct suppression in a hot medium : Cu+CuAu+Au Regeneration : Cu+CuAu+Au Total : Cu+CuAu+Au Grandchamp et al. hep-ph/

05/07/2006Andry Rakotozafindrabe – LLR16 Recombination predictions for vs N coll  Recombination predicts a narrower p T distribution with an increasing centrality, thus leading to a lower  Within the large error bars : seems to be consistent with a flat dependence seems to be consistent with a flat dependence falls data falls between the two hypothesis  partial recombination ? p+p, d+Au, Cu+Cu, Au+Au No recombination With recombination No recombination With recombination Open markers : |y|<0.35 Solid markers : |y|~1.7 Thews & Mangano, PRC73 (2006) c

05/07/2006Andry Rakotozafindrabe – LLR17 Predictions for vs N coll : Cronin effect ?  Random walk of the initial gluons in the transverse plane : AA = pp +  0  g-N  p T 2 L AA  Use this linear L dependence to fit the brodening seen in dimuon data from p+p to d+Au at RHIC  Using L↔N coll, plot the result vs N coll Lower energy survey pp, pA and AA RHIC (y~1.7) VN Tram, Moriond 2006 & PhD thesis p+p d+Au Au+Au = *L Open symbol: y ~ 0 Full & curves: y ~ 2 Cu+Cu

05/07/2006Andry Rakotozafindrabe – LLR18 Recombination predictions vs rapidity  Recombination predicts a narrower rapidity distribution with an increasing N part.  Going from Cu+Cu to the most central Au+Au : no significant change seen in the shape of the rapidity distribution. No recombination All recombination Thews & Mangano, PRC73 (2006) c Blue bands: cold nuclear matter prediction from Vogt, nucl-th/ (shadowing + σ abs = 0, 3mb)

05/07/2006Andry Rakotozafindrabe – LLR19 Ending where it began: revisiting the sequential dissociation  Data driven parametrization of cold nuclear effect (expected)  Sequential melting  overall J/  survival probability (measured/expected): S = 0.6 S direct J/  S J/  ←  ’, χc  Excited states melting from ψ ’ suppression SPS  Recent lattice QCD results : direct J/  melting at GeV/fm 3  SPS and RHIC data seems to be consistent with the sequential melting. Real Au+Au systematic errors i.e. pt-to-pt and global scale added (small systematic errors associated with NA50 published data) A. Bickley, Hard Probes 06 Karsch, Kharzeev & Satz, PLB 637 (2006) 75

05/07/2006Andry Rakotozafindrabe – LLR20 Summary (I) PHENIX preliminary results on J/  dileptons at forward and mid-rapidity in Cu+Cu and Au+Au :  Suppression pattern Beyond cold nuclear effects, at least factor 2 of suppression in most central Au+Au events Beyond cold nuclear effects, at least factor 2 of suppression in most central Au+Au events Similar to SPS suppression? despite a higher energy density reached Similar to SPS suppression? despite a higher energy density reached Overestimated by models in agreement with NA50 data and extrapolated at RHIC energy Overestimated by models in agreement with NA50 data and extrapolated at RHIC energy  Understandable as recombinations that partially compensate the J/  suppression ? Still open question (test vs dependance and rapidity distribution) Still open question (test vs dependance and rapidity distribution)

05/07/2006Andry Rakotozafindrabe – LLR21 Summary (II)  Alternate explanations ? Direct J/  is not melting at present energy densities ? Only the higher mass resonances  ’ and χ c ? (recent lattice QCD results) Direct J/  is not melting at present energy densities ? Only the higher mass resonances  ’ and χ c ? (recent lattice QCD results) J/  transport (with high p T J/  escaping QGP region) + QGP suppression ? ( J/  transport (with high p T J/  escaping QGP region) + QGP suppression ? (Zhu, Zhuang, Xu, PLB607 (2005) 107)  Need to improve knowledge on cold nuclear effects at RHIC

05/07/2006Andry Rakotozafindrabe – LLR22 Hint of things to come  Improved reference p+p : x10 higher statistics from run 5 x10 higher statistics from run 5 x30 higher statistics from run 6 x30 higher statistics from run 6  Future measurements in  ’ ?  Future measurements in χ c  Planning d+Au (28 nb -1 vs 2.7 nb -1 in run 3) and Au+Au (1 nb -1 vs 0.24 nb -1 in run 4 ) with high luminosity Run 6 200GeV p+p Invariant Mass (GeV/c 2 ) A. Bickley, Hard Probes 06 (  c - J/  ) Mass (GeV/c 2 ) PHENIX Run 5 200GeV p+p (  c - J/  ) Mass (GeV/c 2 ) Work under progress

05/07/2006Andry Rakotozafindrabe – LLR23 First upsilon measurement 1st Upsilons at RHIC from ~3pb -1 collected during the 2005 run. Dimuon mass spectrum for the two muon arms added together. Hie Wei, QM05

05/07/2006Andry Rakotozafindrabe – LLR24 STAR results and near future STAR – J/  Run5 pp STAR – J/  Run4 AuAu STAR Preliminary  Dataset GeV : No trigger due to high background No trigger due to high background Just a faint signal Just a faint signal For efficient J/  trigger, full barrel ToF is needed (just patch in Run5) For efficient J/  trigger, full barrel ToF is needed (just patch in Run5)  (Run5): trigger commissioning (~1.7M events) trigger commissioning (~1.7M events)  Run 6 (this year): expect (work in progress) M. Cosentino, QWG06

05/07/2006Andry Rakotozafindrabe – LLR25 Back-up

05/07/2006Andry Rakotozafindrabe – LLR26 Heavy flavour energy loss? Energy loss for heavy flavours is expected to be reduced: i)Casimir factor light hadrons originate predominantly from gluon jets, heavy flavoured hadrons originate from heavy quark jets light hadrons originate predominantly from gluon jets, heavy flavoured hadrons originate from heavy quark jets C R is 4/3 for quarks, 3 for gluons C R is 4/3 for quarks, 3 for gluons ii)dead-cone effect gluon radiation expected to be suppressed for  < M Q /E Q gluon radiation expected to be suppressed for  < M Q /E Q [Dokshitzer & Karzeev, Phys. Lett. B519 (2001) 199] [Armesto et al., Phys. Rev. D69 (2004) ] Casimir coupling factor transport coefficient of the medium average energy loss distance travelled in the medium  R.Baier et al., Nucl. Phys. B483 (1997) 291 (“BDMPS”)

05/07/2006Andry Rakotozafindrabe – LLR27 Charm flow  Disagreement between STAR and PHENIX v 2

05/07/2006Andry Rakotozafindrabe – LLR28 Alternate model : Hydro + J/  transport One detailed QGP hydro + J/ψ transport (Zhu et al)  g + J/ψ  c + c First published without cold nuclear effects, but here : + Nuclear absorption (1 or 3 mb) + Cronin effect from dAu ok (as on previous slide) ok (as on previous slide)  Model should be valid for y=0 But match y=1.7 But match y=1.7 (and central y=0) (and central y=0) Zhu, Zhuang, Xu, PLB607 (2005) private communication Predicted R AA for y=0

05/07/2006Andry Rakotozafindrabe – LLR29 Recombination predictions vs rapidity  Recombination ( Thews et al., nucl-th/ ) predicts a narrower rapidity distribution with an increasing N part.  Going from p+p to the most central Au+Au : no significant change seen in the shape of the rapidity distribution. No recombination All recombination

05/07/2006Andry Rakotozafindrabe – LLR30  Small centrality dependence  Model with absorption + shadowing ( black lines* ): shadowing EKS98 shadowing EKS98 σ abs = 0 to 3 mb σ abs = 0 to 3 mb  σ abs = 1 mb good agreement  σ abs = 3 mb is an upper limit  weak shadowing and weak nuclear absorption J/  production in d+Au vs centrality High x 2 ~ 0.09 Low x 2 ~ *Colored lines: FGS shadowing for 3 mb

05/07/2006Andry Rakotozafindrabe – LLR31 RHIC vs SPS  Plotted « à la SPS » way i.e. normalize the J/  production with the cold nuclear effects : nuclear absorption with σ abs = 4.18 ± 0.35 mb at SPS nuclear absorption with σ abs = 4.18 ± 0.35 mb at SPS Shadowing + nuclear absorption with σ abs ~ 1 mb at RHIC (Vogt, nucl- th/ ) Shadowing + nuclear absorption with σ abs ~ 1 mb at RHIC (Vogt, nucl- th/ ) (NA60 preliminary points from Arnaldi, QM05). NA50 Pb+Pb NA60 In+In PHENIX Au+Au PHENIX Cu+Cu PHENIX : |y|~1.7

05/07/2006Andry Rakotozafindrabe – LLR32 SPS vs RHIC  RHIC : √s = 200 GeV √s = 200 GeV R AA i.e. (J/  in A+A) / (Ncoll * J/  in p+p) R AA i.e. (J/  in A+A) / (Ncoll * J/  in p+p) « expected » = nuclear absorption (σ ~ 1 à 3 mb ) + shadowing « expected » = nuclear absorption (σ ~ 1 à 3 mb ) + shadowing |y| = [0,0.35] or [1.2,2.2] |y| = [0,0.35] or [1.2,2.2]  SPS : √s ~ 17 GeV √s ~ 17 GeV Measured/expected Measured/expected measured = J/  / D.Y measured = J/  / D.Y expected = normal nuclear absorption expected = normal nuclear absorption σ = 4.18 ± 0.35 mb NA50: |y*| = [0,1] NA50: |y*| = [0,1] (Arnaldi, QM05) |y|<0.35 |y|~1.7 + private communications d+Au Cu+Cu Au+Au

05/07/2006Andry Rakotozafindrabe – LLR33 PHENIX detector J/  µ + µ – 1.2< |y| < 2.2 P µ > 2 GeV/c  = 2  Tracking, momentum measurement with cathode strip chambers µ ID with penetration depth / momentum match J/   e + e – |y| < 0.35 P e > 0.2 GeV/c  =  Tracking, momentum measurement with drift chambers, pixel pad chambers e ID with EmCAL + RICH Centrality measurement, vertex position Beam-beam counters (charged particle production) Zero-degree calorimeters (spectator neutrons)

05/07/2006Andry Rakotozafindrabe – LLR34 Invariant yield vs p T at forward rapidities  we fit the p T spectrum using to extract Cu+Cu (|y|  [1.2,2.2])Au+Au (|y|  [1.2,2.2])

05/07/2006Andry Rakotozafindrabe – LLR35 Invariant yield vs p T at mid-rapidity  we fit the p T spectrum using to extract Cu+Cu (|y|~0.35)Au+Au (|y|~0.35)

05/07/2006Andry Rakotozafindrabe – LLR36 Computing the J/  yield : number of ‘s reconstructed : probability for a thrown and embeded into real data to be found (considering reconstruction and trigger efficiency) : total number of events : BBC trigger efficiency for events with a : BBC trigger efficiency for minimum bias events Invariant yield : For Au+Au or Cu+Cu collision : ~ i : i-th bin (centrality for e.g.)

05/07/2006Andry Rakotozafindrabe – LLR37 –1.95 < y < –1.70 Signal extraction in Cu+Cu  Cuts : Dimuons cuts 2.6 < mass < 3.6 GeV/c² 2.6 < mass < 3.6 GeV/c² 1.2 < |rapidity| < < |rapidity| < 2.2 Track quality cuts …  Combinatoric background from uncorrelated dimuons : N bgd = 2√(N ++. N – – ) N bgd = 2√(N ++. N – – )  Signal = number of counts within the J/  invariant mass region  (2.6 – 3.6 GeV/c²) after subtracting N bgd to the distribution of the opposite sign dimuons.  Systematic errors : ~10% from varying fits of the background subtracted signal. Also account for the physical background that can be included into the previous counting.

05/07/2006Andry Rakotozafindrabe – LLR38 Getting acc*eff correction factors in Cu+Cu  Using Monte Carlo J/  generated by PYTHIA over 4π  embed the J/  within muon arm acceptance into real minimum bias Cu+Cu data  Apply to them the same triggers and signal extraction method as the ones applied to the data  Acc.eff(i) is the probability that a J/  thrown by PYTHIA in a given bin i to survive the whole process followed by the data Acc*eff vs rapidity (statistical errors only) Systematic errors : 5% from track/pair cuts and uncertainities in p T, y and z-vertex input distribution 8% from run to run variation (mainly due to the varying number of dead channels in MuTr). Acc*eff vs p T (statistical errors only)Acc*eff vs centrality (statistical errors only)

05/07/2006Andry Rakotozafindrabe – LLR39 Collision geometry and centrality (eg : Cu+Cu) For a given b, Glauber model (Woods-Saxon function) predicts: ● N part (No. participants) ● N coll (No. binary collisions) 10 fm b 0 fm 0 N part N coll 198 Spectators (neutrons) E ZDC Participants (charged particles) Q BBC BBC charge N+S distribution for min.bias Cu+Cu √s = 200 GeV data 80-88% 70-80% etc Monte-Carlo Glauber model Probability for a given Npart Each participant contributes to a Negative Binomial distribution of hits Fit BBC charge distribution

05/07/2006Andry Rakotozafindrabe – LLR40 Run 1 to Run 5 capsule history and J/  in PHENIX YearIons  s NN LuminosityStatus J/  (ee + μμ) 2000Au-Au130 GeV 1  b -1 Central (electrons) Au-Au200 GeV 24  b -1 Central [1] 2002 p-p200 GeV0.15 pb muon arm [2] 2002 d-Au200 GeV2.74 nb -1 Central [3] 2003 p-p200 GeV0.35 pb muon arms [3] Au-Au200 GeV 240  b -1 preliminary~ [4] 2004Au-Au63 GeV 9.1  b -1 analysis~ 13 p-p200 GeV324 nb -1 Cu-Cu200 GeV4.8 nb -1 preliminary~ [4] 2005Cu-Cu63 GeV190 mb -1 analysis~ p-p200 GeV3.8 pb -1 ~ p-p200 GeV~10 pb -1 Just done…~ [1] PRL92 (2004) PRL92 (2004) [2] PRC69 (2004) PRC69 (2004) [3] PRL96 (2006) PRL96 (2006) [4] QM05, nucl-ex/ nucl-ex/

05/07/2006Andry Rakotozafindrabe – LLR41 Cu+Cu 200 GeV data taking: triggers and level2 filtering Volumeof data 50 Tb 2.5 Tb 173 Tb 500 Mb Minimum Bias trigger: BBC Level1 trigger: MuID Deep-Deep Reconstruction Level2 filtering: MuTr + MuID New ! fast online reconstruction of 1D road in MuID di-road accepted if last MuID gap reached (with a minimum number of hit gaps) reconstruction of 2D road in MuID used as seed for MuTr track finding rough track momentum estimate  invariant mass cut for tracks pairs at 2 GeV/c²

05/07/2006Andry Rakotozafindrabe – LLR42 J/  as a probe of the produced medium (I)  Hard probe Large charme quark mass (m J/  ~3.1 GeV/c²)  J/  produced at early stages of the collision Large charme quark mass (m J/  ~3.1 GeV/c²)  J/  produced at early stages of the collision Size r J/  ~0.2 fm < typical hadronique size (~1 fm) Size r J/  ~0.2 fm < typical hadronique size (~1 fm) Recent lattice QCD result : melting temperature in a deconfined medium is Recent lattice QCD result : melting temperature in a deconfined medium is T ~ 1.5 à 2 T C  Production g+g fusion g+g fusion p+p  reference for p+A or A+A r ratios (p+A)/(p+p) or (A+A)/(p+p)  Suppression or enhancement of the J/  yield : Due to nuclear matter or to deconfined medium Due to nuclear matter or to deconfined medium ? PDF ~ ~60% direct production J/  ~30% via χ c  J/  + x ~30% via χ c  J/  + x ~10% via  ’  J/  + x ~10% via  ’  J/  + x

05/07/2006Andry Rakotozafindrabe – LLR43 J/  as a probe of the produced medium (II)  Initial state effect CGC, shadowing CGC, shadowing Cronin effect : multiple elastic scattering  p T broadening Cronin effect : multiple elastic scattering  p T broadening Evaluated via p+A ou d+A   Final state effect Nuclear ( hadronic ) absorption QGP ? suppression : « colour screening » or enhancement : recombinaison From 10 to 20 cc in central Au+Au at RHIC Accessible via A+A PDF J/J/ L Projectile Cible V(r) r ’’J/  χcχc

05/07/2006Andry Rakotozafindrabe – LLR44 Background sources  Physical background: correleted dimuons Drell-Yan: Drell-Yan: Open charm: Open charm: D, D  µ ± + …  Combinatoric background: uncorrelated dimuons  , K   µ  + … (decay before the absorber)  , K   µ  + … (decay before the absorber)

05/07/2006Andry Rakotozafindrabe – LLR45 Energy density  Longitudinally expanding plasma : dE T /dη measurement at mid-rapidity by PHENIX EMCal dE T /dη measurement at mid-rapidity by PHENIX EMCal Which τ 0 ? Which τ 0 ? R2R2 00 ~ 1 fm/c  cross ~ 0.13 fm/c  secondaries formation ~ 0.35 fm/c  therm ~ fm/c

05/07/2006Andry Rakotozafindrabe – LLR46 Commonly used variables  Transverse : perpendicular to the beam direction  Transverse momentum : p T = sqrt( p 2 x + p 2 y )  Rapidity : y = 1/2 ln (E+p z )/(E-p z )  Pseudorapidity : η = 1/2 ln (p+p z )/(p-p z )  Invariant mass of a pair : M 2 inv = (E 1 +E 2 ) 2 - (p 1 +p 2 ) 2