Measurement of virtual photons radiated from Au+Au collisions at E beam = 1.23 AGeV in HADES

Transcription

1 Measurement of virtual photons radiated from Au+Au collisions at E beam = 1.23 AGeV in HADES Szymon Harabasz for the HADES collaboration 1. Motivation 2. Lepton identification 3. Combinatorial background subtraction 4. Excess yield extraction 5. Comparison to models 6. Summary

2 Electromagnetic Probes of Strongly Interacting Matter g, g*: No strong final state interactions leave reaction volume undisturbed Schematic spectral distribution of lepton pairs emitted in ultra-relativistic heavy ion collisions Reflect whole history of collision: From pre-equilibrium phase From QGP and hot hadronic gas From meson decays after thermal freeze-out Figure from PHSD: Nucl. Phys. A 831 (2009) ECT* Workshop, 2017 Szymon Harabasz 2

3 Low Energy Little Excitement SPS, RHIC, LHC SIS18 < qq > T,m B < qq > T =0,m B =0 by B.J.Schaefer and J. Wambach Central cell UrQMD ECT* Workshop, 2017 Szymon Harabasz 3

4 Meet the HADES Glauber Monte Carlo Photograph by A. Rost Beams from SIS18: protons, nuclei, secondary pion beams, E kin =1-2 GeV/u Search for very rare probes -Di-lepton production governed by the factor a 2 -Branching ratio to e + e - ~ Vector meson production sub-threshold Fast detector interaction rate of 8 khz Large acceptance full azimuth, q from 18 o to 85 o Mass resolution of the order of few % Good particle identification Efficient track reconstruction Track multiplicity as large as 300 per event (incl. fakes & secondaries) combinatorial background of 40% most central Au+Au events recorded ECT* Workshop, 2017 Szymon Harabasz 4

5 Electron Identification Track quality selection Energy loss Particle velocity Electromagnetic shower Cherenkov radiation two independent analyses: Ring Finder Backtracking Correlated with momentum All combined in a multivariate analysis (neural networks) Purity of single lepton identification at least 98 % ECT* Workshop, 2017 Szymon Harabasz 5

6 Electron Identification Two Approaches to Detect Cherenkov Signal Ring finder 1. Search for rings in the photodetection plane using pattern matrix or Hough transform 2. Use angular correlations to match rings with tracks in drift chambers Side view of the RICH detector Backtracking algorithm P. Sellheim, J.Phys.Conf.Ser. 599 (2015) 1. Identify lepton candidates using velocity and energy loss 2. Check the ring hypothesis around the expected ring enter Motivation and advantages Ability to resolve overlapping rings Removal of close pairs (g-conversion, combinatorial background) ECT* Workshop, 2017 Szymon Harabasz 6

7 Signal Determination RICH Ring Finder Backtracking ECT* Workshop, 2017 Szymon Harabasz 7

8 Corrections as Function of Pair Properties Construct a cocktail, that reasonably describes raw data. All sources are simulated in Pluto as emitted from thermal fireball. Three sets of sources: p 0, h, w, thermal r 0 p 0, h, w, pairs with (M,p t,y) distribution from Coarse-Graining Pairs with (M,p t,y) distribution from HSD Variance is a contribition to systematic uncertainty ECT* Workshop, 2017 Szymon Harabasz 8

9 Invariant Mass Distribution Results obtained with the two analysis methods agree Agreement in the p 0 -Dalitz region with the prevoiusly measured 1/2(np+pp) reference confirms the constistency of the reconstruction. Above 0.15 GeV/c 2 clear enhancement compared to the reference Medium radiation Ring Finder Backtracking ECT* Workshop, 2017 Szymon Harabasz 9

10 Constraining the Cocktail by the Measurements in the Same Experiment π 0 and η reconstruction C. Behnke, J.Phys.Conf.Ser. 599 (2015) Conversion method Challenge: Low mass spectrometer! Conversion probability: 1% Radiator gas Target holder Beam pipe Segmented target Min-bias ECT* Workshop, 2017 Szymon Harabasz 10

11 Constraining the Cocktail by the Measurements in the Same Experiment Measurement of π + and π - yields, thus giving π 0 In addition: Extraction of slope parameters of π +/- Centrality A Part M (p + + p - )/2 0-10% % % % % ECT* Workshop, 2017 Szymon Harabasz 11

12 Constraining the Cocktail by the Measurements in the Same Experiment [arxiv: ] M(f) = (0.99±0.26) 10-4 per evt. M(w) = per evt. (from m t scaling) ECT* Workshop, 2017 Szymon Harabasz 12

13 Constraining the Cocktail by the Measurements in the Same Experiment HADES Preliminary p 0 from charged pions multiplicity, cross-checked with the conversion method h from the g conversion f from the K + K - channel w from the m t scaling ECT* Workshop, 2017 Szymon Harabasz 13

14 Isolating the Medium Radiation o Subtracting the first chance (NN reference) and freeze-out emission (h) o Correcting for acceptance ECT* Workshop, 2017 Szymon Harabasz 14

15 Measuring the Fireball Temperature with LMR Dileptons ECT* Workshop, 2017 Szymon Harabasz 15

16 NA60: H.J.Specht, AIP Conf.Proc (2010) Model: Rapp/Wambach/Hees r - /N * couplings play substantial role in r melting observed in UrHIC At low energies the same couplings govern off-shell r production by resonance decay HADES Au+Au with r spectral function within the coarse graining framework ECT* Workshop, 2017 Szymon Harabasz 16

17 Vector Meson Dominance 0.9 < M miss ee < 1.03 GeV/c 2 Total π 0 e + e - γ η e + e - γ N(1520) ne + e ECT* Workshop, 2017 Szymon Harabasz 17

18 Data vs. Calculations Inside Hades acceptance o Coarse Graining (CG); medium contribution calculated through r in-medium spectral function with thermodynamic parameters obtained from UrQMD ambient o The calculations provide spectra consistent with each other and in good agreement with measured data CG FRA: Phys. Rev. C 92, (2015) CG GSI-Texas A&M:: Eur. Phys. J. A, 52 5 (2016) 131 HSD: Phys. Rev. C 87, (2013) ECT* Workshop, 2017 Szymon Harabasz 18

19 Centrality Dependence 0.3 < M ee < 0.7 GeV/c 2 The nature of the excess radiation Regeneration of baryonic resonances Contribution in the spectral function ECT* Workshop, 2017 Szymon Harabasz 19

20 Everything flows, nothing stays still Heraclitus Extraction of Fourier coefficients for signal is ongoing ECT* Workshop, 2017 Szymon Harabasz 20

21 Summary and Perspectives Conclusions HADES explores baryon rich matter at SIS 18 Heavy-ion and elementary collisions are measured simultaneously Properly extracted dilepton excess yield agrees well with theory predictions Coming soon from dilepton analysis Completion of the e + e - excitation function Extraction of a fireball lifetime centrality dependence of the excess yield Outlook Now: HADES Upgrade FAIR phase 0: Ag+Ag at 1.65 GeV/u and pion induced reactions FAIR phase 1: SIS 100 Ag+Ag at 3.5 GeV/u Stat. errors correspond to 4 weeks of the beam time ECT* Workshop, 2017 Szymon Harabasz 21

22 Thank You for Your Attention ECT* Workshop, 2017 Szymon Harabasz 22

23 BACKUP SLIDES ECT* Workshop, 2017 Szymon Harabasz 23

24 Combinatorial background estimation Single leptons are produced independently, with Poisson statistics. [PHENIX analysis note] Suitable for: semileptonic decays Compton scattering... Two formalisms Leptons are always produced in pairs, no assumption about the statistics. [Adare et al. (STAR) PRC ] Suitable for: virtual photons real photon conversion Reality is somewhere between the two sets of assumptions Fortunately, "factors with epsilons" can be replaced with ratios of spectra from event mixing: ECT* Workshop, 2017 Szymon Harabasz 24

25 Meaning of the k-factor Not only geometrical acceptance but also recontruction efficiency! ECT* Workshop, 2017 Szymon Harabasz 25

26 Efficiency and Acceptance Correction in HADES Simulate e +, e -, homogoneous in p, cos(q), f Geant Acceptance: acc 1 (p, q, f), acc 2 (p, q, f) Digitization Embedding to real events Particle reconstruction and identification Efficiency: eff 1 (p, q, f), eff 2 (p, q, f) Cocktail of pair sources in 4p Accepting with the probability given by acc 1 and acc 2 Cocktail spectra in HADES (M ee or m t,ee or y ee or helicity ) Weighting with product of eff 1 * eff 2 Cocktail spectra equivalent to raw data ECT* Workshop, 2017 Szymon Harabasz 26

27 Efficiency and Purity Ring Finder Case Original distribution Purity is estimated based on measured data by reproducing random RICH ring-mdc track matches (matching tracks and rings from different HADES sectors) Random matches Signal Efficiency of different methods can be confronted by comparing total yields of identified leptons Most points lie in the same place this indicates the stability of the "training" procedure (for illustration of the method all preselected lepton candidates, without strict PID cuts, are shown) ECT* Workshop, 2017 Szymon Harabasz 27

28 Centrality dependence 0.3 < M ee < 0.7 GeV/c 2 HADES Preliminary o Two observables indicate the formation of longer-lived and hotter medium in the most central collisions ECT* Workshop, 2017 Szymon Harabasz 28

29 ECT* Workshop, 2017 Szymon Harabasz 29

30 Event Selection DST gen 8 New standard event flags File lists of good sectors Track selection and sorting also standard Lepton identification: Neural networks trained on data with two sets of input variables Neural network trained on SIM like on data Neural network trained on Geant PID Hard cuts All based on RICH ring finder, no usage of backtracking Bin Centrality N META hit, cut > N META hit,cut 0 Multiplicity overflow % % % % Multiplicity underflow ECT* Workshop, 2017 Szymon Harabasz 30

31 Low-mass Dileptons at 1 2A GeV HADES Resonance clock ½[pp+ pn]=c+c x x 8-10! w e + e - Phys.Lett. B 690 (2010) 118 Phys.Rev.C 84 (2011) C+C: After h subtraction, coincides with (pp+np) Ar+KCl: First evidence for radiation from the medium in this energy regime! Rapid increase of relative yield reflects the number of s/ N* s regenerated in fireball ECT* Workshop, 2017 Szymon Harabasz 31 S. Harabasz, P. Sellheim (PhD)

32 HSD input file 197, MASSTA: target mass 79, MSTAPR: protons in target 197, MASSPR: projectile mass 79, MSPRPR: protons in projectile 1.25, ELAB: (= Lab energy per nucleon LHC),=21300 RHIC,= (5 TeV) 8.0, BMIN: minimal impact parameter in fm 9.0, BMAX: maximal impact parameter in fm 0.5, DBIMP: impact parameter step in fm 200, NUM: number of parallel events(=20 for RHIC,=150 for SPS and below) 10, ISUBS: number of subsequent runs 113, ISEED: ANY INTEGER number 0, ICHARM: perturbative charm degrees of freedom = 0: no; 1: yes 1, Idilept: =0 no dileptons; =1 electron pair; =2 muon pair 3, ICQ = 0, free rho's, =1 dropping mass, =2 broadening, =3 drop.+broad. 0, IGLUE: 1 with partons; 0 w/o partons 50., FINALT: final time of calculation in fm/c 0, IHARD: =1 compute hard collisions+ direct charm; =0 no 10, ILOW: output level ECT* Workshop, 2017 Szymon Harabasz 32