Dileptons at RHIC

1. Dileptons at RHIC Alberica Toia & Torsten Dahms For the PHENIX Collaboration • Dilepton Measurements at RHIC • Single electron identification • Combinatorial Background • Cocktail of hadrons • Charm contribution • Results @ 200 GeV • p+p • Au+Au • In medium modifications • Chiral symmetry • Spectral function • Thermal radiation • Heavy quarks energy loss

2. e- e+  Strong enhancement of low-mass pairs persists at RHIC R. Rapp nucl-th/0204003  Open charm contribution becomes significant Dielectron Pairs at RHIC- low mass region Expected sources • Light hadron decays • Dalitz decays p0, h • Direct decays r/w and f • Hard processes • Charm (beauty) production • Important at high mass & high pT • Much larger at RHIC than at the SPS • Cocktail of known sources • Measure p0,h spectra & yields • Use known decay kinematics • Apply detector acceptance • Fold with expected resolution Possible modifications Chiral symmetry restoration continuum enhancement modification of vector mesons thermal radiation charm modification exotic bound states suppression (enhancement)

3. Dielectron Pairs at RHIC – intermediate mass Transverse momentum spectra of dielectrons at constrained transverse masses: RHIC with PHENIX acceptance, pT > 1 GeV and 2 GeV < Mee < 3 GeV, Mee hard chance to find thermal dileptons with M > 2 GeV. The double differential rate dNe+e−/dMT2 dQT2 with MT in a narrow interval and with a suitable pTmin cut on the individual leptons seems to allow for a window at large values of the pair pT where the thermal yield shines out. K. Gallmeister, B. Kämpfer and O. P. Pavlenko hep-ph/9801435 (January 1998) hep-ph/9712293 (December 1997)

4. g p DC e+ e- PC1 magnetic field & tracking detectors PC3 PHENIX Physics Capabilities designed to measure rare probes:+ high rate capability & granularity + good mass resolution and particle ID - limited acceptance Au-Au & p-p spin • 2 central arms: electrons, photons, hadrons • charmonium J/, ’ -> e+e- • vector mesonr, w,  -> e+e- • high pTpo, p+, p- • direct photons • open charm • hadron physics • 2 muon arms: muons • “onium” J/, ’,  -> m+m- • vector meson -> m+m- • open charm • combined central and muon arms: charm production DD -> em • global detectors forward energy and multiplicity • event characterization

5. production and of el.magn. shower in the Electro- Magnetic Calorimeter → measure of energy E All charged tracks RICH cut Real Net signal Background Energy-Momentum Electron Identification • Charged particle tracking (dm: 1%) Electron ID • emission and measurement of Cherenkov light in the Ring Imaging Cherenkov detector → measure of min. velocity mis-identification  ~ 20% hadron contamination in the electron sample RICH

6. Acceptance q0 • Define acceptance filter (from real data) • Correct only for efficiency IN the acceptance • “Correct” theory predictions IN the acceptance charge/pT z vertex pT f0 • Single electron pT > 200 MeV • Pair mT > 400 MeV Not an analysis cut, but a constrain from the magnetic field mass

7. Acceptance • Comparison of PHENIX and NA60 acceptance (pT vs. yCM) • Experiments measure in different regions of phase space NA60

8. The Double Challenge Need to detect a very weak source of e+e- pairshadron decays (m>200 MeV, pT>200 MeV) ~ 4x10-6 / π0 In the presence of hundreds of charged particlescentral Au+Au collision dNch / dy ≈ 700 And several pairs per event from trivial originπ0Dalitz decays ~ 10-2 / π0+ γ conversions (assume 0.5% radiation length) ~ 10-2 / π0 huge combinatorial background  (dNch / dy )2 pairing of tracks originating from unrecognized π0 Dalitz decays and γ conversions no means to reduce combinatorial backgroundbeyond reducing conversion length to 0.4% andpT cut at 200 MeV  Signal to background depending on mass up to1 : few hundred Electron pairs are emitted through the Whole history of the collision: need to disentangle the different sources. need excellent reference p+p and d+Au data. Experimental Challenge Analysis Challenge

9. Combinatorial Background Which belongs to which? Combinatorial background g e+ e- g e+ e- g e+ e- g e+ e- p0  g e+ e- p0  g e+ e- p0  g e+ e- p0  g e+ e- PHENIX 2 arm spectrometer acceptance: dNlike/dm ≠ dNunlike/dm  different shape  need event mixing like/unlike differences preserved in event mixing Produce like and unlike sign in the mixed events at the proper rate (B+- =2√B++B--) Use Like sign as a cross check for the shape and to determine normalization Use same event topology (centrality, vertex, reaction plane) Remove every unphysical correlation pT mass

10. Background shape: like sign • Small signal in like sign at low mass • N++ and N–- estimated from the mixed events like sign B++ and B-- normalized at high mass (> 700 MeV) Normalization: 2√N++ N-- • Uncertainty due to statistics of N++ and N--: 0.12% • Correction for asymmetry of pair cut • Pair cut works differently in like and unlike sign pairs • K=k+-/√k++ k-- = 1.004estimated with mixed events • Systematic error (conservative): 0.2% TOTAL SYSTEMATIC ERROR = 0.25% χ2 Pearson test Kolmogorov-Smirnov test High p-valuecompatibility of the real and mixed distributions at any common significance level

11. z Dalitz decay Conversion pair z e- B B y y x e+ e- x e+ Physical background Photon conversion Semi-correlated Background • p0g g* e+e- e+e- • ge+e- at r≠0 have m≠0(artifact of PHENIX tracking) • effect low mass region • have to be removed • Conversion removed with orientation angle of the pair in the magnetic field X External conversion removed with fV cut Inclusive Removed by phiV cut After phiV cut Data: unlike Data: like Monte Carlo: Cross Like Cross Unlike

12. The Background in p+p • Near side located at small mass and high pT • Away side at low pT and large mass • In between exists a region that can be described by mixed events γ e- e+ e+ π0 e+ e- π0 π0 γ e- γ Observe difference from mixed events at near- and away-side In like and unlike sign Background normalized to yield in Δφ = (π/2± π/10) rad • Could jet correlations show up as signal? • Would produce like and unlike sign pairs • Generated p+p events with PYTHIA • compare same event spectra with mixed events

13. Correlated Signal in Data & MC • Subtract combinatorial background from data and PYTHIA • normalize EXODUS and PYTHIA to subtracted like sign spectra • EXODUS in 60–120MeV • PYTHIA in 0.8–2.0 GeV • like sign data and MC in reasonable agreement • Normalization of MC in like sign fixes contribution in unlike sign mee [GeV/c2] • No Problem in Au+Au • Jets are suppressed in Au+Au • As part of cross pair subtraction full like sign signal was subtracted, difference to unlike sign signal from EXODUS included in systematic uncertainty

14. Comparison of BG subtraction Methods Monte Carlo method Like sign method(with some variations) give consistent results over the full invariant mass range to determine syst. uncertainty: spread of two extreme cases (blue & orange): 5-10% 14

15. Ne Electron yield converter 0.8% 0.4% 1.7% With converter Photonic W/O converter Dalitz : 0.8% X0 equivalent radiation length Non-photonic 0 Material amounts: 0 Cross check Converter Method We know precise radiation length (X0) of each detector material The photonic electron yield can be measured by increase of additional material (photon converter was installed) The non-photonic electron yielddoes not increase Photonic single electron: x 2.3 Inclusive single electron :x 1.6 Combinatorial pairs :x 2.5 Photon Converter (Brass: 1.7% X0)

16. submitted to Phys. Rev. Lett arXiv:0706.3034 The raw subtracted spectrum • Same analysis on data sample with additional conversion material • Combinatorial background increased by 2.5 Good agreement within statistical error ssignal/signal = sBG/BG * BG/signal 0.25% large!!! 300,000 pairs 50,000 above p0 From the agreement converter/non-converter and the decreased S/B ratio scale error < 0.1%(well within the 0.25% error we assigned)

17. Single electron pT efficiency pT mass Efficiency Correction Correction for eID cut Mee 0.8 MeV Pair eff = single eff 2 Mee 0.4—0.8 pair eff decreases as consequence of single eff drop Mee < 0.4 MeV: pair eff increases as consequence of mT cutoff (from single pT cutoff which truncates the single distribution leading to larger pT Correction for RICH ghost cut mee (GeV/c2)

18. Cocktail Ingredients pp • Start from the π0 , assumption: π0 = (π+ + π-)/2 • parameterize PHENIX pion data: p0→ γγ(Phys. Rev. D 76, 051106 (2007)) p±(Phys. Rev. C 74, 024904)

19. p+p Cocktail Tuning (ω & φ) • ω and φ are fit with: • modified Hagedorn (as on previous slide: all parameter free) • π0 parameterization with modified Hagedorn + mT scaling (as on previous slide: A is only free parameter, pT→√(pT2+mω2-mπ2)) • exponential in mT • Fits of ω cross section • mod. Hagedorn: χ2/NDF = 21.6/18 • mT scaled π0: χ2/NDF = 34.1/22 • expo in mT: χ2/NDF = 77.0/8 • Fits of φcross section • mod. Hagedorn: χ2/NDF = 30.5/13 • mT scaled π0: χ2/NDF = 32.4/17 • expo in mT: χ2/NDF = 73.3/15 19

20. p+p Cocktail Tuning (J/ψ) • Fits of J/ψcross section • mod. Hagedorn: χ2/NDF = 9.86/12 • mT scaled π0: χ2/NDF = 12.5/16 • expo in mT: χ2/NDF = 15.0/14 Published J/ψ is fit with: • modified Hagedorn (all parameter free) • π0 parameterization with modified Hagedorn + mT scaling (one free parameter) • exponential in mT • also shown is the published fit with a power law

21. p+p Cocktail Tuning

22. (1) q_hat = 0 GeV2/fm (4) dNg / dy = 1000 (2) q_hat = 4 GeV2/fm (3) q_hat = 14 GeV2/fm Non-photonic electron • Non-photonic electron RAAstrong suppression even for heavy quark • non-photonic electron v2 • Indication for reduction of v2 at pT > 2 GeV/c ; Bottom contribution?? • Compared with quark coalescence model prediction. • with/without charm quark flow • (Greco, Ko, Rapp: PLB 595 (2004) 202) • Below 2.0 GeV/c ; • consistent with charm quark • flow model. • indicate charm quark flowcharm thermalization?consequences for dielectrons?

23. p+p Cocktail Comparison Data abs. normalized to J/ψ (acceptance correct our J/ψ yield and normalize to published yield) Cocktail tuned for p+p π0 Hagedorn parameterization η mT scaling φ and J/ψadjusted (like in Au+Au) Filtered in PHENIX acceptance cc contribution from PYTHIA(567±57±193mb) correlated signal in LMR & IMR had not been taken into account if subtracted, much better agreement with PYTHIA calculation in IMR and hadronic cocktail in LMR

24. submitted to Phys. Rev. Lett arXiv:0706.3034 Cocktail comparison • Data and cocktail absolutely normalized • Cocktail from hadronic sources • Charm from • PYTHIA • Single electron non photonic pectrum w/o angular correlations • Predictions are filtered in PHENIX acceptance & resolution • Low-Mass Continuum:enhancement 150 <mee<750 MeV • Intermediate-Mass Continuum: • Single e  pt suppression • PYTHIA softer than p+p but coincide with Au+Au • Angular correlations unknown • Room for thermal contribution?

25. p+p MULTIPLIED by Ncoll pp – AuAu comparison pp and AuAu normalized to p0 region Agreement at the resonances (w, f) Enhancement in 0.2-0.8 Agreement in intermediate mass and J/ just for ‘coincidence’(J/ happens to scale as p0 due to scaling with Ncoll + suppression) p+p NORMALIZED TO mee<100 MeV

26. individual centralities

27. Yield in different mass ranges 0-100 MeV: p0 dominated; approximately scales with Npart 150-750 MeV: continuum 1.2-2.8 GeV: charm dominated; scales with Ncoll Study yield in these mass regions as a function of centrality

28. Centrality Dependence π0 production scales with Npart Low Mass: If in-medium enhancement from ππ or qq annihilation yield should increase faster than proportional to Npart Intermediate Mass: charm follows binary scaling yield should increase proportional to Ncoll Enhancement rise does not depend on 1D vs 2D correction Enhancement present even before correction submitted to Phys. Rev. Lett arXiv:0706.3034 LOW MASS INTERMEDIATE MASS Torsten Dahms - Stony Brook University 28

29. pT dependencyin the PHENIX aperture 0<pT<0.7 GeV/c 0.7<pT<1.5 GeV/c PHENIX Preliminary 1.5<pT<8 GeV/c PHENIX Preliminary While the p+p data follow the cocktail lines, the Au+Au (min bias) are enhanced at low pT and get closer to p+p for high pT PHENIX Preliminary

30. pT dependency II PHENIX Preliminary PHENIX Preliminary While the p+p data follow the cocktail lines, the Au+Au (min bias) significantly deviate at low pT

31. pT – centrality dependency Enhancement in Au+Au observed for the most central collisions (0-20%) The increase at low pT in the enhancement region is present in central, not peripheral data PHENIX Preliminary

32. Conclusions & Outlook • First dielectron continuum measurement at RHIC • Despite of low signal/BG • Thanks to high statistics • Very good understanding of background normalization p+p • Very precise measurement: will fix cocktail components and extract the charm cross section Au+Au LOW MASS: • Enhancement above the cocktail expectations: 3.4±0.2(stat.) ±1.3(syst.)±0.7(model) • Centrality dependency: increase faster than Npart • pT dependency: enhancement concentrated at low pT with a slope softer than the hadronic cocktail INTERMEDIATE MASS: • Coindidence agreement with PYTHIA • Room for thermal radiation? “POINTLESS QUEST’” A needle in the haystack may be difficult to find. Your chance of ever finding one is small. Especially with haystacks of the ordinary kind which don´t have any needles in at all. (Piet Hein - Grooks IV )

33. signal electron Cherenkov blobs e- partner positron needed for rejection e+ qpair opening angle ~ 1 m A Hadron Blind Detector (HBD) for PHENIX • Dalitz & Conversion rejection via opening angle • Identify electrons in field free region • Veto signal electrons with partner • HBD concept: • windowless CF4 Cherenkov detector • 50 cm radiator length • CsI reflective photocathode • Triple GEM with pad readout • Reverse bias (to get rid of ionization electrons in the radiator gas) • Status • installed and taking data in RUN7 • x3 more statistics

34. Background studies • HBD simulation • Change acceptance filter in simulation • Assume HBD rejects hadrons 1:10 • Assume single electron rejection 1:10 Expect S/B 1:10 at m~ 500 MeV • Not included • Realistic hadron pT distribution • Mass resolution • Additional conversion background reject hadrons total reject electron charm

35. Projections N. events 1x109 10x109

36. RHIC at low energy • Npart = 109.1 (MinBias Au+Au) • Nch/Npart and <pT> at mid-rapidity at sqrt(s_NN)=17.2 from SPS • pT distributions and particle ratios according to CERES paramterization in EXODUS pt*(a1*exp(-1.0*mt/T1)+a2*exp(-1.0*mt/T2)+a3*exp(-1.0*mt/T3)) • other hadrons: • T = 0.175+0.115*mass .; • pt*exp(-1.0*mt/T); • PHENIX acceptance filter for +- magnetic field configuration • The expected ZDC rates at beam energy of ~8.5GeV is 100Hz. * 75% to convert into a BBC rate* 50% for the lifetime event rate of 37.5Hz Expected whole vertex minbias event rate [Hz] Which scaling do we use?

37. Projections 50M events(=2 weeks runtime = 1.2*106 s) 500M events(+ electron cooling)

38. Brazil University of São Paulo, São Paulo China Academia Sinica, Taipei, Taiwan China Institute of Atomic Energy, Beijing Peking University, Beijing France LPC, University de Clermont-Ferrand, Clermont-Ferrand Dapnia, CEA Saclay, Gif-sur-Yvette IPN-Orsay, Universite Paris Sud, CNRS-IN2P3, Orsay LLR, Ecòle Polytechnique, CNRS-IN2P3, Palaiseau SUBATECH, Ecòle des Mines at Nantes, Nantes Germany University of Münster, Münster Hungary Central Research Institute for Physics (KFKI), Budapest Debrecen University, Debrecen Eötvös Loránd University (ELTE), Budapest India Banaras Hindu University, Banaras Bhabha Atomic Research Centre, Bombay Israel Weizmann Institute, Rehovot Japan Center for Nuclear Study, University of Tokyo, Tokyo Hiroshima University, Higashi-Hiroshima KEK, Institute for High Energy Physics, Tsukuba Kyoto University, Kyoto Nagasaki Institute of Applied Science, Nagasaki RIKEN, Institute for Physical and Chemical Research, Wako RIKEN-BNL Research Center, Upton, NY Rikkyo University, Tokyo Tokyo Institute of Technology, Tokyo University of Tsukuba, Tsukuba Waseda University, Tokyo S. Korea Cyclotron Application Laboratory, KAERI, Seoul Kangnung National University, Kangnung Korea University, Seoul Myong Ji University, Yongin City System Electronics Laboratory, Seoul Nat. University, Seoul Yonsei University, Seoul Russia Institute of High Energy Physics, Protovino Joint Institute for Nuclear Research, Dubna Kurchatov Institute, Moscow PNPI, St. Petersburg Nuclear Physics Institute, St. Petersburg St. Petersburg State Technical University, St. Petersburg Sweden Lund University, Lund 12 Countries; 58 Institutions; 480 Participants* USA Abilene Christian University, Abilene, TX Brookhaven National Laboratory, Upton, NY University of California - Riverside, Riverside, CA University of Colorado, Boulder, CO Columbia University, Nevis Laboratories, Irvington, NY Florida State University, Tallahassee, FL Florida Technical University, Melbourne, FL Georgia State University, Atlanta, GA University of Illinois Urbana Champaign, Urbana-Champaign, IL Iowa State University and Ames Laboratory, Ames, IA Los Alamos National Laboratory, Los Alamos, NM Lawrence Livermore National Laboratory, Livermore, Ca University of New Mexico, Albuquerque, NM New Mexico State University, Las Cruces, NM Dept. of Chemistry, Stony Brook Univ., Stony Brook, NY Dept. Phys. and Astronomy, Stony Brook Univ., Stony Brook, NY Oak Ridge National Laboratory, Oak Ridge, TN University of Tennessee, Knoxville, TN Vanderbilt University, Nashville, TN

39. Backup

40. Centrality dependence of background

41. Photon conversion cut No cut M<30 MeV & fV<0.25 & M<600 MeV & fV<0.04 M<600 MeV & fV<0.06 M<600 MeV & fV<0.08 M<600 MeV & fV<0.10 M<600 MeV & fV<0.12 M<600 MeV & fV<0.14 M<600 MeV & fV<0.20 M<600 MeV & fV<0.40

42. Hadrons suppressed, photons not • Suppression of 0 and  reduced background, and highlighted direct photons • Ratio increases as centrality increases • Direct photon yield for pT>6GeV/c is well described by NLO pQCD calculation • NO direct photon suppression (initial state), and large 0 suppression (final state) measured / background Nuclear Modification factor Direct photons 0, h [g/p0]measured / [g/p0]background= gmeasured/gbackground S.S.Adler, et. al. (PHENIX Collaboration), PRL 94, 232301(2005)

43. Decay photons(p0→g+g, h→g+g, …) hard: thermal: Thermal photons? • Direct measurement of photons (EMCAL) in this energy region impaired by: • Neutral hadron contamination • Energy resolution in π0 reconstruction  Try to use dielectrons! No significant excess at low pT

44. The idea e+ Compton e- g* q g q Compton g q g q • Start from Dalitz decay • Calculate inv. mass distribution of Dalitz pairs‘ invariant mass of Dalitz pair invariant mass of Dalitz pair invariant mass of virtual photon invariant mass of virtual photon form factor form factor phase space factor phase space factor • Now direct photons • Any source of real g produces virtual gwith very low mass • Rate and mass distribution given by same formula • No phase space factor formee<< pT photon

45. e+ Low mass dielectron pairs e- g* g q q g g q q 0-30 90-140 140-200 200-300 MeV Rdata ÷ ÷ ÷ • Invariant mass distribution of Dalitz pairs fully calculable with QED • Different contributions are well measured • Any source of real g (on-shell) produces virtual gwith very low mass (~ mass shell)formee<< pT photon:Rate and mass distribution same shape(no parent  no phase space factor  no cutoff) Experimental technique • Measure yield in different mass windows (different rate of different contributions) • Calculate ratios of various Minv bins to lowest one: Rdata • If no direct photons: ratios correspond to Dalitz decays • If excess:direct photons

46. Comparison to conventional result Measured ratio From conventional measurement

47. The spectrum • Compare to NLO pQCD • L.E. Gordon and W. Vogelsang • Phys. Rev. D48, 3136 (1993) • Above (questionable) pQCD • Compare to thermal model • D. d’Enterria, D. Peressounko • nucl-th/0503054 • Data above thermal at high pT • Data consistent with thermal+pQCD • Needs confirmation from p+p measurement 2+1 hydro T0ave=360 MeV (T0max=570 MeV) t0=0.15 fm/c

48. Single electron cuts • Event cut: • zvertex <= 25 • Single electron cuts: • Pt: = 150 MeV – 20 GeV • Ecore >= 150 MeV • Match PC3 & EMC • PC3 (Phi+z) < 3 sigma • EMC (Phi+z) < 3 sigma • Dispmax < 5 (ring displacement) • N0min >= 3 tubes • dep >= -2 sigma (overlapping showers NOT removed) • chi2/npe0 < 10 • Quality = 63, 51, 31

49. Efficiency Efficiency Corrections applied as a function of (m, pT) 5-15% 17%

50. Non-photonic electron v2 • non-photonic electron v2 • Indication for reduction of v2 at pT > 2 GeV/c ; Bottom contribution?? • Compared with quark coalescence model prediction. • with/without charm quark flow • (Greco, Ko, Rapp: PLB 595 (2004) 202) • Below 2.0 GeV/c ; • consistent with charm quark • flow model. • indicate charm quark flow • Challenge: fit pT suppression AND v2 simultaneously 1 2 3 0