Dileptons : Getting to the Heart of the Matter

1. Dileptons: Getting to the Heart of the Matter Thomas K Hemmick Stony Brook University

2. COURAGE INTENTION

3. For the Students! • This talk is not targeted at the experts. • Students should EXPECT to understand. • Whenever the speaker fails to meet this expectation: INTERRUPT! Thomas K Hemmick

4. What Physics do You See? Thomas K Hemmick

5. The Focus of the School will be the QCD Phase Structure • The water droplets on the window demonstrate a principle. • Truly beautiful physics is expressed in systems whose underlying physics is QED. • Does QCD exhibit equally beautiful properties as a bulk medium. • ANSWER: YES! Thomas K Hemmick

6. Too hot for quarks to bind!!! Standard Model (N/P) Physics • Collisions of “Large” nuclei convert beam energy to temperatures above 200 MeV or 1,500,000,000,000 K • ~100,000 times higher temperature than the center of our sun. • “Large” as compared to mean-free path of produced particles. Too hot for nuclei to bind Nuclear/Particle (N/P) Physics HadronGas Nucleosynthesis builds nuclei up to He Nuclear Force…Nuclear Physics E/M Plasma Universe too hot for electrons to bind E-M…Atomic (Plasma) Physics SolidLiquidGas Today’s Cold Universe Gravity…Newtonian/General Relativity Stars convert gravitational energy to temperature. They “replay” and finish nucleosynthesis ~15,000,000 K in the center of our sun. Reheating Matter Evolution of the Universe Quark-GluonPlasma??

7. Paradigm #1: Hard Probes • We accelerate nuclei to high energies with the hope and intent of utilizing the beam energy to drive a phase transition to QGP. • The created system lasts for only ~10 fm/c • The collision must not only utilize the energy effectively, but generate the signatures of the new phase for us. • I will make an artificial distinction as follows: • Medium: The bulk of the particles; dominantly soft production and possibly exhibiting some phase. • Probe: Particles whose production is calculable, measurable, and thermally incompatible with (distinct from) the medium.

8. Nuclear Collision Terminology Peripheral Collision Semi-Central Collision Central Collision • Centrality and Reaction Plane determined on an Event-by-Event basis. • Npart= # of Participants • 2  394 • Nbinary=# of Collisions 100% Centrality 0% f Reaction Plane • Fourier decompose azimuthal yield:

9. schematic view of jet production hadrons leading particle q q hadrons leading particle q/g jets as probe of hot medium Jets from hard scattered quarks observed via fast leading particles or azimuthal correlations between the leading particles However, before they create jets, the scattered quarks radiate energy (~ GeV/fm) in the colored medium Jet Quenching

10. AA AA If no “effects”: RAA < 1 in regime of soft physics RAA = 1 at high-pT where hard scattering dominates Suppression: RAA < 1 at high-pT AA RAA Normalization 1. Compare Au+Au to nucleon-nucleon cross sections 2. Compare Au+Au central/peripheral Nuclear Modification Factor: nucleon-nucleon cross section <Nbinary>/sinelp+p

11. Thermally-shaped Soft Production Hard Scattering Calibrating the Probe(s) • Measurement from elementary collisions. • “The tail that wags the dog” (M. Gyulassy) p+p->p0 + X

12. Suppression Discovered in Year One QM2001 • Quark-containing particles suppressed. • Photons Escape! • Gluon Density = dNg/dy ~ 1100 QM2001

13. Escaping Jet “Near Side” Out-plane Lost Jet “Far Side” In-plane Jet Tomography • Jets are produced as back-to-back pairs. • If one jet escapes, is the other shadowed? • Map the dynamics of Near-Side and Away-Side jets. • Vary the reaction plane vs. jet orientation. • Study the composition of the jets • Reconstruct the WHOLE jet • Find “suppressed” momentum & energy. X-ray pictures areshadows of bones Can Jet Absorption be Used to“Take an X-ray” of our Medium?

14. Peripheral Au + Au STAR Central Au + Au Out-plane In-plane Back-to-back jets • Given one “jet” particle, where are it’s friends: • Members of the “same jet” are in nearly the same direction. • Members of the “partner jet” are off by 180o • Away-side jet “gone”

15. y py px x y z x Paradigm 2: Collective Flow Almond shape overlap region in coordinate space Origin: spatial anisotropy of the system when created, followed by multiple scattering of particles in the evolving system spatial anisotropy  momentum anisotropy v2: 2nd harmonic Fourier coefficient in azimuthal distribution of particles with respect to the reaction plane

16. Anisotropic Flow • Process is SELF-LIMITING • Sensitive to the initial time • Delays in the initiation of anisotropic flow not only change the magnitude of the flow but also the centrality dependence increasing the sensitivity of the results to the initial time. Liquid Li Explodes into Vacuum Position Space anisotropy (eccentricity) is transferred to a momentum space anisotropy visible to experiment • Gases explode into vacuum uniformly in all directions. • Liquids flow violently along the short axis and gently along the long axis. • We can observe the RHIC medium and decide if it is more liquid-like or gas-like

17. Large v2 • Hydrodynamic limit exhausted at RHIC for low pT particles. • Can microscopic models work as well? • Flow is sensitive to thermalization time since expanding system loses spatial asymmetry over time. • Hydro models require thermalization in less than t=1 fm/c Adler et al., nucl-ex/0206006

18. v2 Scales with valence quarks BRAHMS • A flowing system would be expected to have mass-dependent v2 values STAR & PHENIX • Valence quark scaling is superb indicating that the final state hadrons may have come from recombination

19. High Order Moments vn • Event Plane method yields <vn> (vodd=0). • 2-particle yields SQRT(<vn2>) (vodd>0). • How to deal: • PHENIX = EP method + factorization. • ATLAS = Rapidity OUTSIDE other Jet. • Everyone else = Factorization.

20. v2{2}, v3{3}, v4{4} at 200GeV Au+Au arXiv:1105.3928 charged particle vn : ||<0.35 reaction plane n : ||=1.0~2.8 v3 is comparable to v2 at 0~10% weak centrality dependence on v3 (3) v4{4} ~ 2 x v4{2} All of these are consistent with initial fluctuation.

21. Paradigm 3: Hadrochemistry • Hadronization by random choice or recombination will follow simple statistical distributions:

22. Limitations of these Paradigms • Hard or Jet Probes provide useful information BECAUSE their initial production is well known. • Flow is driven by “pre-collision” spatial anisotropy. • Hadro-chemistry (and HBT) probe the final state at de-coupling time. PENETRATING (color-less) Probes are Transparent to the QGP medium and directly probe the initial state e+e- IS THE MOST DIFFICULT MEASUREMENT IN HEAVY ION PHYSICS! History of Success: CERES, PHENIX, STAR, ALICE??? Thomas K Hemmick

23. Chiral symmetry restoration continuum enhancement modification of vector mesons thermal radiation & modified heavy flavor suppression (enhancement) Lepton-Pair Continuum Physics Modifications due to QCD phase transition known sources of lepton pairs at s = 200 GeV • Sources “long” after collision: • p0, h, w Dalitz decays • (r), w, f, J/y, y‘ decays • Early in collision (hard probes): • Heavy flavor production • Drell Yan, direct radiation • Baseline from p-p • Thermal (blackbody) radiation • in dileptons and photons • temperature evolution • Medium modifications of meson • pp r l+l- • chiral symmetry restoration • Medium effects on hard probes • Heavy flavor energy loss 23

24. Problem #1: Electrons are Rare • Particles are produced in accordance with the strength of the coupling. • Colorless particles are rare compared to those produced via strong interactions. • Requires a superb detector to distinguish electrons from other species. • Which property of the electron is KEY to distinguishing electrons in your detector? • charge, mass, spin, flavor, colorless, weak interaction, Thomas K Hemmick

25. q e - How do particles lose energy in matter? “kinematic term” “minimum ionizing particles”   3-4 “relativistic rise” density effect Bethe-Bloch Formula ionization constant Thomas K Hemmick

26. Pb-Scint Calorimeter Bremsstrahlung • Electromagnetic Calorimeter modules use the “Shish-Kebab” geometry and achieve excellent energy resolution (7%/sqrt(E)) and timing (< 200 psec). Thomas K Hemmick

27. PID via Cherenkov No Ring Unless 1/b > n High Index distinguisheshadrons Ring Radius determines qc

28. 2560 PMTs, 107 gain. • Ethane or CO2 radiator. • 104 charge pion rejection. • 1 degree ring res. • N0= 118. • 12-18 p.e./ring • C-fiber/Rohacell mirror 0.3% X0 Thomas K Hemmick

29. Optics of the RICH • Two circular mirrors each of which reflects light onto arrays of PMT’s. • PMT’s are located so that the central magnet poles shield the array. • Offset focal plane leads to ellipse rather than circle images. Thomas K Hemmick

30. p K p STAR e Particle Identification by dE/dx • dE/dx: • The 1/ b2 survives integration over impact parameters • Measure average energy loss to find b • Looks tough for e-sep. • How possible??? Thomas K Hemmick

31. Transition Radiation Small angles • Charged Particles may radiate X-ray photons upon “transition”. • LARGE ionization at small angle & short distance. ALICE Thomas K Hemmick

32. γ e- e+ π0 e+ π0 e+ X π0 e- e- γ γ Challenge: Pair Background • No background rejection  Signal/Background  1/100 in Au-Au • Unphysical correlated background • Track overlaps in detectors • Not reproducible by mixed events: removed from event sample (pair cut) • Combinatorial background: e+and e-from different uncorrelated source • Need event mixing because of acceptance differences for e+ and e- • Use like sign pairs to check event mixing • Correlated background: e+ and e- from same source but not “signal” • “Cross” pairs  “jet” pairs • Use Monte Carlo simulation and like sign data to estimate and subtract background

33. Normalization Mathematics • Combinatorial Mathematics • Leptons are ALWAYS produced as pairs. • Within a collection of events, any single event produced N pairs governed by some probability distribution P(N). • For a single event with N primary lepton pairs, there is a BINOMIAL distribution for the number of these pairs that are fully reconstructed: • Additionally, this same event will contain partially reconstructed events (+ only; - only; neither) that are governed by a MULTINOMIAL • We will 1count the number of pairs (true and combinatorial) for a single np, 2sum over all np, 3sum over all N, and 4demonstrate the relationship for combinatorial rates: Thomas K Hemmick

34. Useful Relations • In General: • Binomial: • Poisson: • Multinomial: Thomas K Hemmick

35. For a given np, unlike sign pairs are: • Three combinations: fully-fully, fully-partial, partial-partial: <n>=eN Cov(M) algebra algebra Thomas K Hemmick

36. Sum these over np: Substitute <n+_> • Total Reconstructed Pairs <N+->: <n2>=<n>2+s2 algebra algebra This result contains epN true pairs and combinatorial. Thomas K Hemmick

37. Same thing for like-sign: algebra • Like pairs as 2n++ = n+(n+-1): • Similarly for -- pairs <n2>=<n>2+s2 algebra Thomas K Hemmick

38. Rate of like-sign pairs <n2>=<n>2+s2; s2(B) =eN(1-e) • Summing over np: algebra algebra Thomas K Hemmick

39. Lastly, sum over all N • Unlike sign • Like sign • By inspection: Thomas K Hemmick

40. Side Notes • For independent production: • However, experimental determination of <FG+> is limited by the in ability to select collisions of identical centrality. • Such effects must be modeled using a Glauber Monte Carlo of your experiment’s centrality measure and generates some systematic error in the determination of the absolute combinatorial background. • This technique, however, has proved viable in the analysis of background in 2-particle correlations and obviates the need for a “ZYAM” normalization. • Muons are not as well behaved as electrons: • Many muons come from hadron decays that don’tproduce pairs. • These are “in-between” statistics and require • When using an absorber muons from K+ decays are enhanced as compared to other sources. Thomas K Hemmick

41. Mathematical Summary: • Proper Statistics depend upon fundamental production stats: • If particles are always produced in pairs: • Nearly correct for e+e- (violated in double-semi-leptonic decays). • Normalization is absolutely independent of the singles efficiency. • NOT (necessarily) the same as: N++ + N— • Independent of “pooling” • Normalization IS sensitive to correlated particle losses: • Fv, Double-Track Resolution , Ring Sharing, EMC Sharing. • Handling the intrinsic pair cuts is the toughest part. • If particles are produced independently • Nearly correct for hadron correlations (Jets). • Sensitive to width of pools: requires x correction. • If you use the “wrong” statistics: • z correction is necessary:

42. p+ e- Unphysical Background Rejection: “Pair Cut” Effectively promotes pions into electrons, RUINS statistics • Pions identified as electrons in presents of electron • RICH measures angle only and not position!!! • Pion can be misidentified as electron • Leads to correlated but unphysical pairs • Not reproduced by mixed events • Different probability and kinematics for like and unlike sign pairs pT UNLIKE mass Remove by rejecting events with parallel tracks in RICH Thomas K Hemmick

43. Mixing Without Pair Cut Large unphysical background! Thomas K Hemmick

44. Combinatorial Background: Like Sign Pairs • Shape from mixed events • Excellent agreements for like sign pairs • Normalization of mixed pairs • Small correlated background at low masses from double conversion or Dalitz+conversion • normalize B++ and B- - to N++ and N- - for m > 0.7 GeV • Normalize mixed + - pairs to • Subtract correlated BG • Systematic uncertainties • statistics of N++ and N--: 0.12 % • kappa: 0.2 % --- Foreground: same evt N++ --- Background: mixed evt B++ Au-Au TOTAL SYSTEMATIC ERROR = 0.25% Thomas K Hemmick

45. yield in4p yield in acceptance Signals in the like sign! -- “Cross” Pairs • p0g g* unlike cross like cross unlike 4-body e+ e- X e+ e- Unlike: data - mixed Like: data - mixed Monte Carlo: Cross Like Cross Unlike Include also h decay Thomas K Hemmick

46. Background Description of Function of pT Good agreement Thomas K Hemmick

47. Continuum in p+p and AuAu Phys. Lett. B 670, 313 (2009) arXiv:0912.0244 • Data and Cocktail of known sources • Striking Enhancement at and below the w mass. • Data and Cocktail of known sources • Excellent Agreement 47

48. STAR has Excellent Results: Sees enhancement, but smaller than PHENIX. PHENIX aperture with minimum acc at pT=m, is a significant factor. Need detailed calculations to show whether data agree.

49. STAR Enhancement in Central collisions The fit of charm cocktail to the data gives: 0.62 ± 0.14 mb PYTHIA setting for charm: V6.416 MSEL=1, PARP(91)=1.0 (kt), PARP(67)=1 (parton shower level)

50. Centrality Dependence • Enhancement in low mass region is a strong function of centrality. • Statistics are also sufficient to analyze pT dependence. • Need methodical approach to the spectra. Thomas K Hemmick