Direct-Photon Physics with ALICE

Transcription

1 Direct-Photon Physics with ALICE New ersectives on Photons and Diletons in Ultrarelativistic Heavy-Ion Collisions at RHIC and LHC, EC*, rento, 30 Nov - 11 Dec, 2015 Klaus Reygers Physikalisches Institut University of Heidelberg

2 he Role of Direct Photons in Heavy-Ion Physics τ = 1.2 fm/c τ = 7.0 fm/c τ = 4.1 fm/c τ =.1 fm/c Escae medium unscathed Produced over the entire duration of the collision est of sace-time evolution, in articular of the hydro aradigm Exerimental access to initial QGP temerature (?) QGP hoton rate rγ (lowest order): E dr d 3 / s 2 e E / log E k 2 c otal emission rate: r / 4 Klaus Reygers, Direct Photons with ALICE rento 8 December

3 A Comlication for the emerature Measurement: Blueshift due to Radial Flow C. Shen, U. Heinz, J.-F. Paquet, C. Gale, PRC 89 (2014) 4, 0449 E d3 N d 3 / e E / e s 1+ flow e = 1 flow {z } 2for flow=0.6 Large blueshift at late times when MeV Extraction of initial temerature from data requires comarison to (hydro) model Klaus Reygers, Direct Photons with ALICE rento 8 December

4 ALICE ZDC PMD V0 and 0 (A side) EMCal HMPID IS PC IS SSD IS SDD IS SPD FMD V0 and 0 (C side) diole magnet OF RD ZDC L3 magnet PHOS muon tracking (2.5 < η < 4) Photons/π 0 s: Conversions (PC [+ IS]) PHOS EMCal used in arxiv: (focus of this talk) 4

5 Direct Photons in Pb-Pb at snn = 2.76 ev: PHOS One PHOS module 4.6 m from interaction oint 260 < φ < 320 η < modules Each module lead tungstate crystals Crystal size: cm3 Avalanche hotodiode readout Energy resolution 1.3% = E E /GeV E E E 3.3% E /GeV 1.12% = 3.7% for E = 1 GeV Klaus Reygers, Direct Photons with ALICE rento 8 December

6 Direct Photons in Pb-Pb at snn = 2.76 ev: Photon Conversion Method (PCM) e + e + e - e - Photon candidates from airs of secondary, oositely charged tracks with reconstructed conversion oint in 5 < r < 180 cm (PC [+ IS]) Large accetance: η < 0.9, 0 φ 360 Electron ID with PC de/dx (+ OF, if available) γ 1 γ 2 hoton conversion Purity of the hoton candidate samle for < 3 GeV/c 98% in eriheral Pb-Pb 90% in central collisions Very good hoton energy resolution σe/e 1.6% for 1 E 2 GeV Klaus Reygers, Direct Photons with ALICE rento 8 December

7 Reconstructed hoton conversion oints ALICE, arxiv:

8 Material Budget Reconstructed hoton conversion oints Precise knowledge of material key for conversion method 4.5% uncertainty ( η < 0.9, r < 180 cm) X/X0 = ± Corresonds to conversion robability of 8.5% Klaus Reygers, Direct Photons with ALICE rento 8 December

9 he Statistical Subtraction Method Idea: Cancellation of uncertainties common to hoton and π 0 measurement direct = incl decay =(1 1 R ) incl R = incl decay incl 0 aram decay 0 aram measured decay hoton calculation Which uncertainties cancel (artially)? Calorimeter: global energy scale, energy non-linearity Photon conversions: conversion robability, hoton selection Klaus Reygers, Direct Photons with ALICE rento 8 December

10 No Significant Direct Photon Signal ev (Photon Conversion Method) /π 0 ) decay /π 0 )/(γ , s = 7 ev QCD consistent with data inc (γ Direct hoton double ratio NLO rediction: 1 + (γ for µ = 0.5,1.0,2.0 /γ ) direct,nlo decay ALI PREL Klaus Reygers, Direct Photons with ALICE rento 8 December 2015

11 Decay Photon Calculation (Pb-Pb, snn = 2.76 ev) decay /γ source γ π γγ (e e γ) η' ργ (ωγ, γγ) η γγ (π π γ,e + e γ,π 0 γγ) φ ηγ (π γ, ωγ) ω π 0 γ (ηγ) ρ 0 π + π γ (π γ, ηγ ) Inut: π 0 sectrum η, ω: modeling m scaling alternatively (for η): shae from Ks 0 All other hadrons negligible % Pb-Pb = 2.76 ev Monte Carlo simulation s NN π 0 : 82-88% η: -15% ω: 2-3% 1 ALICE, arxiv: Klaus Reygers, Direct Photons with ALICE rento 8 December

12 η/π 0 Ratio in Pb-Pb (1): m Scaling vs. Ks 0 Shae 0 η/π ALICE, s = 7 ev Phys.Lett. B717 (2012) Cocktail Calculations η from s = 7 ev as inut η from m scaled π % Pb-Pb s NN = 2.76 ev, η from K scaled s ALI DER At = 2.5 GeV/c: η/π 0 m scaling = 0.4, η/π 0 Ks0 shae = 0.5 η/π 0 in the decay hoton calculation: Average of the two aroaches Klaus Reygers, Direct Photons with ALICE rento 8 December

13 η/π 0 Ratio in Pb-Pb (2): ALICE Data Particle ratio Pb-Pb, 0 s NN = 2.76 ev η /π, 0-% K ± /π ±, 0-% 0 η/π, 20-50% PLB 736 (2014) 196 η/π 0 ratio from 2011 Pb-Pb data set Current measurement not recise enough for use in decay hoton calculation ALICE Preliminary ALI PREL Klaus Reygers, Direct Photons with ALICE rento 8 December

14 Comarison btw. PHOS and PCM Inclusive Photon Sectra (Data / Combined) inc γ % Pb-Pb s NN = 2.76 ev PCM ALICE PHOS % Pb-Pb s NN = 2.76 ev For both methods: Stat. Syst. A 1.2 Syst. B Syst. C % Pb-Pb s NN = 2.76 ev Systematic uncertainties classified as Point-by-oint uncorrelated (tye A) Correlated and deendent (tye B) Normalization uncertainty (tye C) Level of agreement btw. PCM and PHOS quantified with MC seudo exeriments PCM and PHOS inclusive hoton sectra agree on the level of 1.2σ 0.8 ALI PUB ALICE, arxiv: Klaus Reygers, Direct Photons with ALICE rento 8 December

15 Combined Inclusive Photon Sectra N γ 2 d c 2 ) -2 (GeV inc 1 dy d N ev. 2π Pb-Pb s NN = 2.76 ev 0-20% ALICE 20-40% ALICE 40-80% ALICE 2 x 1 x 0 x Analysis strategy Combined PHOS+PCM direct hoton sectrum from combined inclusive hoton sectrum and combined double ratio Rγ Uses information from both measurements even if measured Rγ from one system fluctuates below unity reach 0.9 < < 14 GeV/c 7 ALI PUB Klaus Reygers, Direct Photons with ALICE rento 8 December

16 Double Ratio (PCM, PHOS) % Pb-Pb = 2.76 ev ALICE PCM PHOS 20-40% Pb-Pb s NN s NN = 2.76 ev Rγ,X with π 0 sectrum from X (X = PCM, PHOS) Cancellation of systematic uncertainties Level of agreement: 1.4σ 1.5 R γ % Pb-Pb s NN = 2.76 ev ALI PUB π 0 s from PCM and PHOS ublished in arxiv: (Eur. Phys. J. C (2014) 74:38) Klaus Reygers, Direct Photons with ALICE rento 8 December

17 Systematic Uncertainties: PHOS cancel in double ratio Centrality 0 20% 2 g incl yield Efficiency (B) Contamination (B) Conversion (C) Accetance (C) Global E scale (B) Non-linearity (B) yield Yield extraction (A) Efficiency (B) Accetance (C) Pileu (C) Feed-down (B) g decay / 0 0 sectrum (B) h contribution (B) otal R g otal g incl systematic uncertainties of the PHOS ana Energy scale uncertainties cancel in the double ratio Many source with similar contributions γ and π 0 reconstruction efficiency π 0 yield extraction Contamination of the γ samle Feed-down correction of the π 0 sectrum (mostly from Ks 0 ) Klaus Reygers, Direct Photons with ALICE rento 8 December

18 Systematic Uncertainties: PCM Centrality 0 20% g incl yield rack quality (A) Electron PID (A,B) Photon selection (A,B) Material (C) g incl / 0 rack quality (A) Electron PID (A,B) Photon selection (A,B) yield (A) Material (C) g decay / 0 0 sectrum (B) h yield (C) h shae (B) otal R g otal g incl Material budget uncertainty dominant 4.5% for γincl/π 0 (9% for π 0 sectrum) Followed by hoton selection uncertainties 3.2% (from cut variations) Klaus Reygers, Direct Photons with ALICE rento 8 December

19 Combined Double Ratio % Pb-Pb = 2.76 ev ALICE s NN 20-40% Pb-Pb = 2.76 ev ALICE s NN Significance of the excess for 0.9 < < 2.1 GeV/c 2.6σ for 0-20% Calculated from seudo exeriments taking into account correlated uncertainties R γ % Pb-Pb = 2.76 ev ALICE s NN ALI PUB Klaus Reygers, Direct Photons with ALICE rento 8 December

20 Counts esting the Null Hyothesis Rγ R0 = est statistics ALI PUB t data 0-20% Pb-Pb s NN = 2.76 ev ALICE seudo data -value = σ R mod,i = R 0 (1 + " B B,i,rel )(1 + " C C,rel ) t = ALICE-PUBLIC n data X oints i=1 Rd,i R 0 0,i 2, 0,i = R mod i,stat+a,rel t Model of the measurement based on tye A, B, C uncertainties Deviations from central values described by Gaussian-distributed nuisance arameters εb, εc Distribution of test statistic t for H0(R0=1) from seudo exeriments -value: fraction of seudo exeriments above tdata ranslated into units of Gaussian standard deviations with a two-sided test Klaus Reygers, Direct Photons with ALICE rento 8 December

21 Double Ratio: Comarison with QCD R γ % Pb-Pb s NN = 2.76 ev ALICE NLO QCD PDF: CEQ6M5 FF: GRV JEPHOX PDF: C, FF: BFG2 JEPHOX npdf: EPS09, FF: BFG2 (all scaled by N coll ) 20-40% Pb-Pb s NN = 2.76 ev ALICE NLO QCD PDF: CEQ6M5 FF: GRV JEPHOX PDF: C, FF: BFG2 JEPHOX npdf: EPS09, FF: BFG2 (all scaled by N coll ) 40-80% Pb-Pb s NN = 2.76 ev ALICE NLO QCD PDF: CEQ6M5 FF: GRV JEPHOX PDF: C, FF: BFG2 JEPHOX npdf: EPS09, FF: BFG2 (all scaled by N coll ) QCD curves: R QCD =1+N coll QCD decay calculated based measured π 0 sectrum QCD agrees with data for 5 GeV/c Evidence for an additional hoton source at lower 1.0 ALI PUB Klaus Reygers, Direct Photons with ALICE rento 8 December

22 Combined Direct Photon Sectra N γ 2 d c 2 ) -2 (GeV dir dy d 3 2 Pb-Pb s NN = 2.76 ev 0-20% ALICE 20-40% ALICE 40-80% ALICE Aroximately exonential sectrum for 3 GeV/c N ev. 2 π x 4 1 x x 7 8 ALI PUB Klaus Reygers, Direct Photons with ALICE rento 8 December

23 Combined Direct Photon Sectra N γ 2 d c 2 ) -2 (GeV dir dy d N ev. 2 π Pb-Pb s NN = 2.76 ev 0-20% ALICE PDF: CEQ6M5, FF: GRV 20-40% ALICE (n)pdf: CEQ6.1M/EPS09, 40-80% ALICE FF: BFG2 JEPHOX PDF: C, FF: BFG2 npdf: EPS09, FF: BFG2 (all scaled by N coll ) Aroximately exonential sectrum for 3 GeV/c QCD uncertainties growing towards lower 20-30% for 5 GeV/c 50-60% at 1 GeV/c 3 2 x x 6 0 x 7 8 ALI PUB Klaus Reygers, Direct Photons with ALICE rento 8 December

24 Direct Photon RAA ALICE-PUBLIC R AA % Pb-Pb s NN = 2.76 ev ALICE γ dir reference: (n)pdf: CEQ6.1M/EPS09, FF: BFG2 Rel. error JEPHOX npdf: EPS09, FF: BFG2 R AA % Pb-Pb s NN = 2.76 ev ALICE γ dir reference: (n)pdf: CEQ6.1M/EPS09, FF: BFG2 Rel. error JEPHOX npdf: EPS09, FF: BFG ALI PUB ALI PUB Calculated using QCD reference (McGill calculation) Confirmation of Ncoll scaling of QCD hotons for 5 GeV/c Klaus Reygers, Direct Photons with ALICE rento 8 December

25 Comarison with Au-Au at 200 GeV (PHENIX) N γ 2 d c 2 ) -2 (GeV dir dy d N ev. 2 π ALICE 0-20% Pb-Pb A ex(- / eff ) eff = 304 ± PHENIX s NN stat % Au-Au A ex(- / eff ) = 239 ± 25 eff s NN stat = 2.76 ev ± sys 40 MeV = 0.2 ev ± 7 sys MeV Inverse sloe arameter eff for 0-20% Pb-Pb@2.76 ev without QCD subtraction: eff = 304 ± 11 stat ± 40 sys MeV with QCD subtraction: eff = 297 ± 12 stat ± 41 sys MeV 0-20% Au-Au@0.2 ev eff = 239 ± 25 stat ± 7 sys MeV ( arameterization subtracted) Indication for larger inverse sloe arameter in Pb-Pb@2.76 ev 5 ALI PUB Klaus Reygers, Direct Photons with ALICE rento 8 December

26 Do we Understand the snn Deendence of the Direct Photon Signal? Hydro exectation for snn deendence, figure courtesy of Jean-François Paquet, based on arxiv: Low hoton excess (0-20%) 0.2 ev Au-Au: Rγ ev Pb-Pb: Rγ 1.12 Observed increase of thermal yield (LHC/RHIC, from revious slide) factor ~2.5 at = 1 GeV/c factor ~5 at = 2 GeV/c Hydro exectation Similar or smaller Rγ at the LHC than at RHIC exected Observed increase of thermal yield somewhat smaller than in hydro, but aroximately consistent Klaus Reygers, Direct Photons with ALICE rento 8 December

27 Comarison with State-of-the-Art Models N γ 2 d c 2 ) -2 (GeV dir dy d N ev. 2 π ALI PUB Pb-Pb Paquet et al. arxiv: Linnyk et al. arxiv: s NN = 2.76 ev 0-20% ALICE 20-40% ALICE 40-80% ALICE A ex(- / eff ) 0-20% 20-40% 2 x 1 x x Chatterjee et al. v. Hees et al. PRC 85(2012) 0649 NPA 933(2015) JHEP 1305(2013) Models (all with QGP formation) v. Hees et al.: Ideal hydro with initial flow, τ = 0.2 fm/c, 0-20% = 682 MeV Chatterjee et al.: 2+1 hydro, fluctuating initial conditions, τ = 0.14 fm/c, 0-20% 740 MeV Paquet et al.: 2+1 viscous hydro with IP-Glasma initial conditions, τ = 0.4 fm/c, 0-20% = 385 MeV Linnyk et al.: Off-shell transort, microscoic descrition of evolution Currently not ossible to rule out one or more of these models Klaus Reygers, Direct Photons with ALICE rento 8 December

28 Direct Photon Azimuthal Asymmetries hermal hoton v2 (QGP, HG) v2(hg) v2(π) Collective flow velocity takes time to build u in hydro models: v2(qgp) v2(hg) arxiv:nucl-th/05179v2 Exect v2(γdir) v2(π) if QGP hotons dominate at low v2(γdir) tests hydro models and hoton roduction rates (QGP, HG) Non-hydro mechanisms roducing v2(γdir) conceivable E.g. hoton roduction related to large initial B field in non-central A-A Can be tested with v3 Klaus Reygers, Direct Photons with ALICE rento 8 December

29 How to Measure the Direct-Photon v2? Reaction lane (RP) from charged articles in forward direction Inclusive hotons (mid-raidity): 2 = hcos(2(' RP 2 ))i C v,incl 2.8 < η < < η < -1.7, C = resolution correction Decay hoton v2 from cocktail calculation based on measured ion v2 (+ higher mass hadrons, based on KE scaling where KE = m - m) Inclusive hoton v2 is weighted average of decay hoton and direct hoton v2. hus one can calculate the direct-hoton v2 as v,direct 2 = R v,incl 2 v,decay 2 R 1 with R = incl decay =1+ direct decay Klaus Reygers, Direct Photons with ALICE rento 8 December

30 Measured Inclusive Photon and Calculated Decay Photon v2 γ 2 v 0.20 ALICE reliminary γ,decay v % Pb-Pb, V0 event lane s NN = 2.76 ev γ,incl v 2 v2(incl) < v2(decay) for > 3 GeV/c Exected from v2 = 0 of romt hotons decay + NLO 0.04 Phys.Rev. D50 (1994) decay + NLO + thermal (Shen et al.) arxiv: decay + NLO + thermal (Holoainen et al.) Phys.Rev. C84 (2011) ALI PREL v2(incl) v2(decay) for < 3 GeV/c (v2(incl) maybe slightly smaller) If there is a large direct hoton comonent its v2 must be very similar to the decay hoton v2 v2(incl) described by models with small Rγ ( 1.05) redicted by the same models Klaus Reygers, Direct Photons with ALICE rento 8 December

31 Direct Photon Ellitic Flow Aears to be Larger than Exected in Hydro Models γ,dir 2 v ALI PREL arxiv: ALICE reliminary 0-40% Pb-Pb, V0 event lane s NN = 2.76 ev ALICE reliminary NLO (Vogelsang) + thermal (Shen et al.) NLO (Vogelsang) + thermal (Holoainen et al.) PHSD(O. Linnyk et al.) PCM, using the reliminary Rγ v2(γdir) only slightly lower than v2(incl) Many direct hotons from late stage with c MeV? hen large inverse sloe arameter due to Doler blueshift with tyical hadronic flow velocity βflow 0.6 c? However, current systematic uncertainties are sizable so that there is no big uzzle looking at the ALICE data alone Klaus Reygers, Direct Photons with ALICE rento 8 December

32 Inclusive Photon v3 γ 3 v ALICE reliminary 0-40% Pb-Pb, s NN = 2.76 ev V0 event lane γ,decay v 3 γ,incl v 3 decay + NLO 0.02 Phys.Rev. D50 (1994) decay + NLO + thermal (Shen et al.) arxiv: decay + NLO + thermal (Chatterjee et al.) arxiv: ALI PREL v3(incl) v3(decay) for < 3 GeV/c: If there is a large direct hoton comonent its v3 must be very similar to the decay hoton v3 v3(incl) described by models with small Rγ ( 1.05) redicted by the same models Uncertainty σ(rγ) is key for concluding whether v2(γdir) and v3(γdir) agree or disagree with models Error roagation non-trivial (no v3(γdir) from ALICE so far) Klaus Reygers, Direct Photons with ALICE rento 8 December

33 he Direct Photon Puzzle Plots: Paquet et al., arxiv: /(2 ) dn /d (GeV -2 ) 1/(2 ) dn /d (GeV -2 ) Au-Au 0-20% s=200 GeV Pb-Pb 0-20% s=2.76 ev PHENIX Direct hermal Promt Non-cocktail (GeV) ALICE (2015) Direct hermal Promt Non-Cocktail (GeV) v 2 {SP} v 2 {SP} Pb-Pb 0-40% s=2.76 ev PHENIX (2015) Direct hermal (GeV) ALICE (relim) Direct hermal (GeV) Au-Au 20-40% s=200 GeV Au-Au at RHIC Models fail to describe direct hoton data Puzzle has two arts yields v2 Pb-Pb at the LHC Similar trends However, no uzzle with current uncertainties Klaus Reygers, Direct Photons with ALICE rento 8 December

34 owards Higher Precision (1): Virtual Photon Method direct hoton ratio arxiv: ALI PREL ALICE reliminary, s=7 ev virtual hotons (1+r ) conversions (γ /π 0 )/(γ /π 0 ) inc decay wo-comonent fit to mee distribution Hadronic cocktail Virtual direct hoton mee shae (Kroll-Wada formula) Extraolated to m = 0 (real hotons) More model assumtions than in direct measurement Reasonably small uncertainties in Klaus Reygers, Direct Photons with ALICE rento 8 December

35 owards Higher Precision (1): Virtual Photon Method c 3 ) -2 (b GeV 3 σ/d 3 d E ALICE reliminary, s=7 ev 95 % C.L. arxiv: γ direct γ incl = r (PCM) γ direct (W.V.) NLO µ=0.5 µ=1.0 µ=2.0 γ incl Virtual hoton method ev Agreement with QCD Virtual Photon Method in Pb-Pb? Method is statistics hungry Hadronic cocktail needs to be understood very well In Pb-Pb likely only doable with LHC Run 3 data ( ) 5 ALI PREL Klaus Reygers, Direct Photons with ALICE rento 12 December

36 owards Higher Precision (2): Standard Statistical Subtraction Method More statistics hels π 0 statistical uncertainty traded for systematic uncertainty by using a arameterization of the π 0 sectrum in the double ratio Statistics for 0-20% Pb-Pb - Current analysis (20 data): 4 6 events : 17 6 events : 30 6 events (exected) Photon conversion method: Reduce material budget uncertainty No direct-hoton signal in for < 2 GeV/c Estimate material budget (and other) uncertainties from double ratio? /π 0 ) decay /π 0 )/(γ inc (γ ALI PREL 27932, s = 7 ev Direct hoton double ratio NLO rediction: 1 + (γ for µ = 0.5,1.0,2.0 /γ ) direct,nlo decay contains information about material budget uncertainty Klaus Reygers, Direct Photons with ALICE rento 8 December

37 owards Higher Precision (3) agging Method, External Converter Use π 0 tagging with EMCal, PHOS (like PHENIX) Conversion robability dros out At the cost of adding, e.g., the energy scale uncertainty of the calorimeter (via the matching of the lower energy threshold for the π 0 tagging in data and MC) Will be tried, not clear if errors will be smaller For LHC Run 3 ( ) the use of an external converter of known X0 is an otion R = incl h"f i decay = N N, 0 decay 0 incl tag calc data N incl incl = conv " ee a ee N, 0 tag = h"f i conv " ee a 0 ee Klaus Reygers, Direct Photons with ALICE rento 12 December

38 Photon conversion and virtual hoton method No significant direct hoton excess for < 2.5 GeV/c Consistent with QCD ev (20 data set) Electromagnetic calorimeter (PHOS) and hoton conversion (PCM) Consistent results Direct hoton excess in 0-20%: Rγ 1.12 for 0.9 < < 2.1 GeV/c, 2.6σ significance Excess w.r.t. QCD consistent with thermal hoton source Direct hoton uzzle Mostly at RHIC Similar trend at the LHC, but no big uzzle with current uncertainties Klaus Reygers, Direct Photons with ALICE rento 8 December

39 Extra slides 39

40 Omega meson / π 0 in at 7 ev 0 ω/π 1.4 ALICE reliminary, s = 7 ev Cocktail Calculations ω from m scaled π ALI PREL Klaus Reygers, Direct Photons with ALICE rento 8 December

41 Model Predictions for Rγ R γ 1.8 ALICE reliminary 0-40% Pb-Pb, s NN = 2.76 ev Direct hoton double ratio NLO rediction: 1 + (N γ /γ ) coll direct,,nlo decay thermal (Shen et al., arxiv: ) thermal (Holoainen et al., Phys.Rev. C84 (2011) ) ALI PREL Klaus Reygers, Direct Photons with ALICE rento 8 December

42 Virtual Photon Method (ALICE,, 7 ev) (mb/ GeV/c 2 ) d σ dm ee d ALICE reliminary s=7 ev, e >0.2 GeV/c η e < < ee < 3.2 GeV/c π0 η φ data f γ,dir (1-r)*f cocktail + cocktail sum η ω c r*f γ,dir -3-4 ALI PREL (GeV/c 2 ) m ee Klaus Reygers, Direct Photons with ALICE rento 8 December