Particle production from ee to AA and search for the QCD Critical Point with PHENIX

Transcription

1 Particle production from ee to AA and search for the QCD Critical Point with PHENIX Máté Csanád (Eötvös University, Budapest) for the PHENIX Collaboration and for Roy L. et al (the authors of arxiv: ) 54th International School of Subnuclear Physics, Erice, Sicily, 016 June 19, 016 Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

2 Particle production from ee to AA Particle production mechanisms from ee to AA A+A(B) collisions: frequently described with thermo/hydrodynamics Model ingredients: macroscopic variables (temperature, entropy) see e.g. W. Kittel and E. A. DeWolf, Soft Multihadron Dynamics, (World Scientic, 005) or recent PHENIX, PHOBOS, STAR, ALICE, CMS and ATLAS papers Microscopic phenomenology used in e + e +, e ± + p, p( p)+p or p+a Perturbative gluon exchange, gauge fields, strings, partons see e.g. Kharzeev et al., NPA747; Armesto et al., PRL94, Dusling et al., PRD87 and other references in arxiv: Similarity in particle production available E eff for part. production Feinberg Phys Rept. 5, Albini et al. Nuovo Cim A3, Basile et al. PLB9&Nuovo Cim A67, Nyiri, IJMP A18 Sarkisyan and Sakharov, hep-ph/ κ 1 see κ spp κ 3 sep with κ 1 1 (1) Role of quark participant pairs in pp, pa and AA describing initial stage What do the measurements tell us? Máté Csanád (Eötvös University) 54th ISSP, Erice, 016 / 15

3 Particle production from ee to AA N ch vs. s scaling in e + e +, p( p)+p and e ± + p Entropy ansatz: S (TR) 3, dn ch /dη and N ch S Initial stage variable N pp number of participant pairs (N pp = 1 here) Global assumption: Npp 1/3 R Simple result: [ N ch /N pp ] 1/3 T p T Scaling versus κ n s Fit result: N ch = [ b Nch + m Nch log(κ n s) ] 3 b Nch = 1. ± 0.01 m Nch = ± See details and references in Lacey et al., arxiv: Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

4 Particle production from ee to AA Quark participant scaling from pp through pa to AA Initial stage size measure: N pp Use quark participant pairs N qpp Recall [(dn ch /dη)/n qpp ] 1/3 T Flat size (N qpp ) dependence Strikingly similar s NN trends for p+p and A+A(B) collisions Common production mechanism? Small deviation for TeV Smooth trend over full range Details in Lacey et al., arxiv: When does QGP production start? Phases of QCD? Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

5 Emergent QCD phenomena Exploring the QCD phase diagram QCD: fundamental theory mostly understood Emergent phenomena (phases, nucleons, etc): hard to handle Important part of the phase diagram: RHIC energies! Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

6 Emergent QCD phenomena The PHENIX Experiment at RHIC PHENIX Detector PbSc PC3 PC PbSc DC Central Magnet PC3 TEC 010 PbSc PbSc DC RICH PbSc RICH PbGl BB HBD MPC RxNP PC1 PbSc PC1 7.9 m = 6 ft TOF-W PbGl Aerogel TOF-E West Beam View East RHIC collisions: versatile in energy, GeV/nucleon Versatile in colliding nuclei: p, d, Cu, Au, Al, He, U Tracking via Drift Chambers & Pad Chambers PID via time of flight from TOF & EMCal Momentum resolution: δp/p 1.3% 1.% p GeV/c Charged pion ID from p 0.18 to GeV/c Ma te Csana d (Eo tvo s University) 54th ISSP, Erice, / 15

7 Emergent QCD phenomena The RHIC Beam Energy Scan Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

8 Bose-Einstein correlations Introduction to Bose-Einstein correlations In case of free bosons, momentum correlation function: S(q) C (q) 1 +, S(q) = S(x)e iqx d 4 x, q = p 1 p () S(0) Final state interactions distort the simple Bose-Einstein picture Coulomb interaction important, handled via two-particle wave function Resonance pions: Halo around primordial Core Halo part unresolvable Bolz et al, Phys.Rev. D47 (1993) Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15 J.

9 Bose-Einstein correlations Lévy shaped source and the Critical Point Lévy distributed source and resulting correlation function: S(r) α,r = 1 (π) 3 d 3 qe iqr e 1 qr α C (q) = 1 + λ e (R q )α (3) Parameters: λ, R, α, may depend on average pair momentum Related to anomalous diffusion or generalized random walk Metzler, Klafter, Physics Reports 339 (000) 1-77 α = : Gauss, α = 1: Cauchy; power-law tail for α < Levy index α identified with critical exponent η Spatial correlation function r (d +η) critical η exponent Symmetric stable distributions (Levy) spatial corr. r 1 α QCD universality class 3D Ising Halasz et al., Phys.Rev.D58 (1998) , Stephanov et al., Phys.Rev.Lett.81 (1998) D Ising: η = (3), Random field 3D Ising: η = 0.50 ± 0.05 Rieger, Phys.Rev.B5 (1995) 6659, El-Showk et al., J.Stat.Phys.157 (4-5): 869 Change in α proximity of CEP Motivation for precise Levy HBT Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

10 PHENIX Levy HBT analysis results Physical parameters: R, λ, α; measured versus pair m T Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15 Example correlation function measurement result Measured in 31 m T = m + p T bins for π+ π + and π π pairs

11 PHENIX Levy HBT analysis results Physical parameters: R, λ, α; measured versus pair m T Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15 Example correlation function measurement result Measured in 31 m T = m + p T bins for π+ π + and π π pairs

12 PHENIX Levy HBT analysis results Levy scale parameter R versus pair m T R [fm] 10 9 MinBias s NN = 00 GeV PH ENIX preliminary π π π + π + α free m T [GeV/c ] Similar decreasing trend as usual Gaussian HBT radii Hydro behaviour: 1/R m T, not incompatible Interesting, since simple hydro would mean α = Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

13 PHENIX Levy HBT analysis results Correlation strength λ versus pair m T λ MinBias s NN = 00 GeV - - π π π + π + α free PH ENIX preliminary m T [GeV/c ] From the Core-Halo model: λ = (N C /(N C + N H )) Large and correlated systematic uncertainties Observed decrease at small m T increase of halo fraction May be interesting phyics in λ, strange speculations: Resonance effects, chiral symmetry restoration? Partially coherent pion production? Aharanov-Bohm effect? Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

14 PHENIX Levy HBT analysis results Levy exponent α versus pair m T α MinBias s NN = 00 GeV - - π π π + π + α 0 = ± χ /ndf = 159/61 PH ENIX preliminary m T [GeV/c ] Far from Gaussian (α = ) and Cauchy/exponential (α = 1) Also far from 3D Ising value at CEP (α 0.5) More or less constant (at least within systematic errors) Although the constant fit is statistically not acceptable Motivation to do fits with fixed α = Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

15 PHENIX Levy HBT analysis results Newly discovered scaling parameter R 1/R [1/fm] MinBias - - π π π + π + α free R = R λ (1+α) = 00 GeV s NN PH ENIX preliminary [GeV/c ] Empirically found scaling parameter Linear in m T Physical interpretation open question Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15 m T

16 Summary Scaling of N ch from ee to AA data from GeV to TeV s regions Effective energy notation successful Quark participants emerging Special signatures at special energies? Seach for the QCD critical point at RHIC New measurement method of Levy sources developed with PHENIX Levy exponent α at s NN = 00 GeV far from Gaussian or critical values Correlation strength λ hints at interesting physics New empirically found scaling parameter R Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

17 Thank you for your attention! Let me invite You to the 16th Zimanyi Winter School on Heavy Ion Physics Budapest, Hungary, Dec. 5 9, Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

18 Similarities from p+p through p+a to A+A(B) Similar charged particle multiplicities (N ch ) Similar pseudorapidity densities (dn ch /dη) Azimuthal long range ( η 4) angular correlations, ridge Collective anisotropic flow in A+A collisions Also in p+p, p+pb, d+au and He+Au ALICE PLB719, ATLAS PRL110, CMS PLB718, PHENIX PRL114, PHENIX PRL115 Qualitative consistency achieved with hydro See e.g. Bozek, PRC85, the Buda-Lund model from Csörgő et al., NPA661, JPhysG30, EPJA38,... Common underlying particle production mechanism dominating? Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

19 Our framework to capture underlying physics Macroscopic entropy (S) ansatz S (TR) 3 const. (4) dn ch /dη and N ch S (5) Initial stage variable N pp number of participant pairs N pp = 1 for e + e +, e ± + p and p( p)+p Nucleon or quark participant pairs (N npp, N qpp ) in p+a, A+A(B) Further assumption: N 1/3 pp R [(dn ch /dη)/n pp ] 1/3 T p T Monte Carlo Glauber calculations performed to obtain N npp and N qpp. Lacey et al. PRC83, Eremin et al. PRC67, Bialas et al. PLB649, Nouicer EPJC49, PHENIX PRC89 Subset of initial particles become participants by an initial inelastic N+N or q+q interaction. N np = N npp or N qp = N qpp N+N (q+q) cross sections taken from literature Fagunders et al, J. Phys. G40 Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

20 Nucleon participant scaling in A+A(B) collisions All systems: [ dnch /dη η =0.5 N npp ] 1/3 T (a): log(dn ch /dη) log S, (b): N 1/3 npp R Logarithmic S-dependence Linear size dependence (at a given s NN ) p T increases with snn and log(dn ch /dη) Pseudorapidity density factorizes into contributions depending on s NN and N 1/3 npp Slope increaseses with beam energy Lack of sensitivity to system type (Cu+Cu, Cu+Au, Au+Au, U+U), for fixed s NN. Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

21 PHENIX Levy HBT analysis Dataset: s NN = 00 GeV Au+Au, min. bias, 7 billion events Goal: detailed shape analysis of two-pion correlation functions Correlation functions measured in fine pair m T = m + p T bins Levy source instead of Gaussian better agreement with data Precision measurement of source parameters versus pair m T Result: λ Levy (m T ), α Levy (m T ), R Levy (m T ) Tracking: DCH & PC1, σ matching cuts in TOF & EMCal Particle identification: time-of-flight data from EMCal & TOF, momentum, flight length σ cuts on m distribution Pair-cuts: Exclude particles hitting the same tower/slat/strip Customary shaped cuts in ϕ z plane for TOF, EMCal, DCH Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

22 Example correlation functions C 1.6 MinBias s NN - - = 00 GeV, π π, p = GeV/c T (data-fit)/error λ = 1.14 ± 0.06 R = 6.90 fm ± 0.8 fm α = 1.09 ± 0.03 ε = ± N = ± χ /NDF = 95/51 conf. level = Raw corr. function Raw corr. Coulomb factor Coulomb factor C C Coul Levy 0 Levy N(1+ε k ) (λ,r,α; k ) N(1+ε k ) (λ,r,α; k ) N(1+ε k ) 0 α C =(1+λ Levy exp(-( R k ) )) N(1+ε k ) PH ENIX preliminary k [GeV/c] Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

23 Example correlation functions C 1.6 MinBias s = 00 GeV, π + π + NN, p = GeV/c T (data-fit)/error λ = 1.05 ± 0.09 R = 4.84 fm ± 0.30 fm α = 1.0 ± 0.05 ε = ± N = ± χ /NDF = 413/380 conf. level = Raw corr. function Raw corr. Coulomb factor Coulomb factor C C Coul Levy 0 Levy N(1+ε k ) (λ,r,α; k ) N(1+ε k ) (λ,r,α; k ) N(1+ε k ) 0 α C =(1+λ Levy exp(-( R k ) )) N(1+ε k ) PH ENIX preliminary k [GeV/c] Máté Csanád (Eötvös University) 54th ISSP, Erice, 016 / 15

24 Levy scale parameter R versus pair m T R [fm] MinBias s NN = 00 GeV PH ENIX preliminary ] [1/fm 1/R MinBias s NN = 00 GeV - - π π π + π + α free PH ENIX preliminary π π 0.0 π + π α free m T [GeV/c ] m T [GeV/c ] Similar decreasing trend as usual Gaussian HBT radii Hydro behaviour: 1/R m T, not incompatible Hard to say whether the 1/R scaling is linear or not Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

25 Correlation strength λ versus pair m T λ MinBias s NN = 00 GeV - - π π π + π + α free PH ENIX preliminary m T [GeV/c ] From the Core-Halo model: λ = (N C /(N C + N H )) Large and correlated systematic uncertainties Observed decrease at small m T increase of halo fraction Different effects can cause change in λ, strange speculations Resonance effects, partial chiral symmetry restoration? Partially coherent pion production? Aharanov-Bohm effect? Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

26 Levy exponent α versus pair m T α MinBias s NN = 00 GeV - - π π π + π + α 0 = ± χ /ndf = 159/61 PH ENIX preliminary m T [GeV/c ] Far from Gaussian (α = ) and Cauchy/exponential (α = 1) Also far from 3D Ising value at CEP (α 0.5) More or less constant (at least within systematic errors) Although the constant fit is statistically not acceptable Motivation to do fits with fixed α = Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

27 Levy scale parameter R with fixed α = R [fm] MinBias s NN = 00 GeV PH ENIX preliminary ] [1/fm 1/R MinBias s NN = 00 GeV - - π π π + π + α = PH ENIX preliminary π π 0.0 π + π α = m T [GeV/c ] More smooth trend Hydro behaviour seems to be more valid The linearity of 1/R holds m T [GeV/c ] Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

28 Correlation strength λ with fixed α = λ MinBias s NN = 00 GeV - - π π π + π + α = PH ENIX preliminary m T [GeV/c ] More smooth trend Smaller systematic errors Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

29 Newly discovered scaling parameter R 1/R [1/fm] MinBias - - π π π + π + α = R = R λ (1+α) = 00 GeV s NN PH ENIX preliminary α = fixed [GeV/c ] Empirically found scaling parameter m T Linear in m T Physical interpretation open question Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

30 Beam energy & system size dependence of HBT radii Corr. func. in 3D, Gaussian fit (details: arxiv: ) π + π +, π π data combined Linear 1/ m T scaling of HBT radii Interpolation to common m T, PHENIX and STAR consistent Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

31 Beam energy & system size dependence of HBT radii quantities related to emission duration and expansion velocity non-monotonic patterns indication of CEP? More precise mapping and further detailed studies required Is there any other way to find the critical point? Maybe Levy exponent α! Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

32 Generalized random walks Random walk, step size with infinite variance Random Walk Generalized random walk Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

33 (Speculative) connection between λ(m T ) and U A (1) Partial chiral U A (1) restoration reduction of the η mass Reduced η mass larger abundance of η Decay: η ηπ + π, η π + π π 0 (some probability) In summary: U A (1) restoration larger number of decay pions Larger number of decay pions larger halo, i.e. reduced λ These decay pions have small momentum, as η has no available energy to give them momentum λ reduces only at small momentum! Hole in λ(m t ) distribution! Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15

34 (Speculative) connection: λ(m T ) and Aharanov-Bohm Mixing of a A, b B and a B, b A Correlation strength: Ψ 1, with Ψ = 1 + cos( x k) and x = x 1 x, k = k 1 k Average on local production probability and random phases at emission (c.f. incoherent production); if no random phases, no correlation! Observation: closed loop in the above figure Particles go through an expanding, fluctuating medium Pairs picking up additional random phases φ, Ψ = 1 + cos( x k + φ) Assume φ distribution of exp[ φ /(σ ), average on possible φ values Ψ 1 φ = exp[ σ /] cos( x k) Correlation strength λ reduced by exp[ σ /] Suppose σ 1/m T hole in λ(m T ) Idea from Antal Jakovác (Eötvös University) Máté Csanád (Eötvös University) 54th ISSP, Erice, / 15