PHENIX measurements of low momentum direct photon radiation from large and small systems in (ultra)relativistic heavy ion collisions: direct photon scaling
PPHENIX measurements of low momentum direct photonradiation from large and small systems in (ultra)relativisticheavy ion collisions: direct photon scaling ∗ Vladimir Khachatryan for the PHENIX Collaboration
Department of Physics and Astronomy, Stony Brook University, Stony Brook,New York 11794-3800, USAThe PHENIX collaboration has measured low momentum direct photonradiation in Au+Au collisions at 200 GeV, 62.4 GeV and 39 GeV, in Cu+Cuat 200 GeV as well as in p+p, p+Au and d+Au at √ s NN =200 GeV. Inthese measurements PHENIX has discovered a large excess over the scaledp+p yield of direct photons in A+A collisions, and a non-zero excess, ob-served within systematic uncertainties, over the scaled p+p yield in centralp+A collisions. Another finding is that at low- p T the integrated yield ofdirect photons, dN γ /dy , from large systems shows a behavior of universalscaling as a function of the charged-particle multiplicity, ( dN ch /dη ) α , with α = 1 .
25, which means that the photon production yield increases fasterthan the charged-particle multiplicity.PACS numbers: 25.75.-q, 25.75.Nq, 24.85.+p
1. Introduction
Direct photons determine the excess yield, which one obtains by sub-tracting the hadronic decay photon yield (mostly from π and η decays)from the total observed photon yield. By measuring these photons, we canstudy the strongly interacting medium produced in (ultra)relativistic heavyion collisions, and gain information on the properties and dynamics of theproduced matter integrated over space and time. The direct photons possi-bly originate from the hot fireball of the Quark-Gluon Plasma (QGP), latehadronic phase as well as from initial hard scattering processes like QCDCompton scattering among the incoming and outgoing partons.After PHENIX measured large invariant yield and large anisotropy oflow momentum direct photons in Au+Au collisions at √ s NN = 200 GeV [1, ∗ Presented at the XIIIth Workshop on Particle Correlations and Femtoscopy(WPCF 2018), Krakow, Poland, May 22-26, 2018. (1) a r X i v : . [ nu c l - e x ] D ec acta-template printed on December 27, 2018
2, 3], a challenging problem arose, commonly referred to as “thermal photonpuzzle”, where different theoretical models encounter difficulties when theyare used to describe these two quantities simultaneously (though there isalso some progress in recent years [4, 5, 6] on this matter). For resolvingthis puzzle, PHENIX continues measuring low momentum direct photonsin large and small collision systems. These past and new measurementsrevealed very interesting findings, which we report in this proceedings.
2. Some results on direct photon elliptic flow from large systemsand on direct photon p T spectra from large/small systems In this section we represent some of the PHENIX recent and previouslypublished results on direct photon measurements. Fig. 1 shows two plotson direct photon elliptic flow ( v ) in two centrality bins, from an ongoinganalysis in Au+Au at √ s NN = 200 GeV obtained based upon an externalconversion method. Fig. 2 shows p T spectra, obtained with another externalconversion method, for minimum bias data samples in Au+Au at √ s NN =62.4 GeV and 39 GeV [7], including also the previously published result inAu+Au at 200 GeV [2]. In the external conversion method the photons aremeasured through their conversions to e + e − pairs at the VTX or HBD in thePHENIX detector system, and the fraction of direct photons is determinedafter tagging photons from neutral pion decays. Comparing these data to N coll scaled p+p fit (or pQCD calculations) one finds a significant excessover the scaled p+p yield of low- p T direct photons in all three systems. [GeV/c] T p d i r g v - PRC94, 064901PRL109, 122302 (BBC)2014 data = 200GeV NN sAu+Au, |<0.35, 0-20% h , | dir g PH ENIX preliminary [GeV/c] T p d i r g v - PRC94, 064901PRL109, 122302 (BBC)2014 data = 200GeV NN sAu+Au, |<0.35, 20-40% h , | dir g PH ENIX preliminary
Fig. 1. (Left) Direct photon v vs. p T in 0-20% compared with the published resultsfrom [3]. (Right) Direct photon v vs. p T in 20-40% compared with the publishedresults from [3]. The new results are based upon external conversions with VTX. With the external conversion method PHENIX recently also measuredlow momentum direct photons in p+p and p+A collisions (shown in Fig. 3).Within systematic uncertainties, the observed a non-zero excess yield ( ∼ one sigma) in central p+Au collisions above the scaled p+p fit may comefrom the possible production of QGP droplets in small central systems. cta-template printed on December 27, 2018 [GeV/c] T p ] - d y [ ( G e v / c ) T N / dp d T p p / -5 -4 -3 -2 -1 T = 0.5 p m -scaled pQCD AA T Extrapolated to 1 GeV/c T = 1.0 p m T = 2.0 p m ] eff /T T Exp[-p (cid:181)
Fit 0.045 GeV – – = 0.214 eff T + X, |y|<0.35 g fi Au+Au = 62.4 GeV NN s0-86%, p T [ GeV / c ] − − − − − − − − − π p T d N d p T d y [ ( G e V / c ) − ] √ s NN = pp data PRL104,132301PRL98,012002PRD86,072008 pp fit Au+Au data
PRL104,132301PRL109,152302Presentdata N coll -scaled pp fit [GeV/c] T p ] - d y [ ( G e v / c ) T N / dp d T p p / -5 -4 -3 -2 -1 T = 0.5 p m -scaled pQCD AA T Extrapolated to 1 GeV/c T = 1.0 p m T = 2.0 p m ] eff /T T Exp[-p (cid:181)
Fit 0.070 GeV – – = 0.176 eff T + X, |y|<0.35 g fi Au+Au = 39 GeV NN s0-86%, Fig. 2. Direct photon p T spectra in Au+Au at √ s NN = 200 GeV [2] (central panel),and in Au+Au at √ s NN = 62.4 GeV and 39 GeV [7] (left and right panels). Alldata are in minimum bias and from external conversions with HBD. [GeV/c] T p c - m b G e V / dp s d (cid:215) E - - - - + X g fi p + p = 200 GeVs Int. conv.Ext. conv. n0 p T2 p1+)=A T f(p PH ENIX preliminary [GeV/c] T p c - m b G e V / dp s d (cid:215) E - - - - -
10 1 + X g fi p + Au = 200 GeV: NN s 0-5%, Ext. conv. n0 p T2 p1+ A coll )=N T f(p Fig. 3. The left and right panels show direct photon p T spectra in p+p and in thecentral 0-5% p+Au collisions at 200 GeV (external conversions with VTX). ThePHENIX previously measured p+p and minimum bias d+Au cross sections froman internal conversion method [8] are included in this figure and shown in the leftpanel (with p+p), and in the central panel (with p+p and d+Au).
3. Scaling properties of direct photons
For a given center-of-mass energy one can compare data from differentcentrality classes (or system size) using number of participants, N part , orthe number of binary collisions, N coll . But this way is not useful when wecompare data at different energies. Instead, we use charged-particle multi-plicity, dN ch /dη , which itself has an interesting scaling behavior with N coll shown in Fig. 4. Here N coll scales like ( dN ch /dη ) α for all center-of-mass en-ergies with a logarithmically slowly increasing function called specific yield,SY. Four datasets are simultaneously fitted by a power-law, with verticaland horizontal uncertainties of N coll and dN ch /dη | η =0 , respectively [9, 10]. acta-template printed on December 27, 2018 Then we get α = 1 . ± .
02. For other details see the caption of Fig. 4. h /d ch dN c o ll N [GeV] NN s SY )-1.83 NN s0.98log( a ) h /d ch (dN) NN sSY(1Fit: – =1.25 a = 39 GeV NN sAu+Au, = 62.4 GeV NN sAu+Au, = 200 GeV NN sAu+Au, = 2.76 TeV NN sPb+Pb, PHENIX
Fig. 4. N coll vs. dN ch /dη | η =0 forgiven four beam energies. The smallbox on the bottom right shows datademonstrating a scaling between N coll and dN ch /dη of the form: N coll = SY ( √ s NN ) (cid:16) dN ch dη (cid:17) α , wherethe specific yield is a function loga-rithmically increasing with √ s NN :SY( √ s NN ) = 0 . · log( √ s NN ) − . . ................................................................................ Thereby, one can scale the direct photon yield by ( dN ch /dη ) α , which fora specific √ s NN is equivalent to N coll . Let us take, e.g., the photon spectra inminimum bias Au+Au collisions at 62.4 and 39 GeV with pQCD curves fromFig. 2, and normalize them by ( dN ch /dη ) α . It results in the data falling ontop of each other at low- p T as shown in the panel (a) of Fig. 5. As expectedat high- p T the p+p data coincide with the pQCD calculations within thequoted uncertainties. In the panel (b) all Au+Au data at 200 GeV are on topof each other at high- and low- p T , and at low- p T they are distinctly abovethe p+p data, fit and pQCD. In (c) the data are compared for different √ s NN from 62.4 GeV to 2760 GeV. Again all the data coincide at low- p T ,while at high- p T we see the expected difference with √ s NN and N coll scaling.In Fig. 5 all error bars are the quadratic sum of the systematic andstatistical uncertainties. Uncertainties on dN ch /dη are not included. Allnormalized data, p+p fit, pQCD curves are from [7], and they are obtainedwith Au+Au data from [1, 2, 11], Pb+Pb data from [12], p+p data at200 GeV from [8], at 62.4 GeV from [13], at 63 GeV [14, 15], the empiricalfit to the p+p data at 200 GeV from [7], the pQCD calculations at differentbeam energies from [5, 16], and the data on dN ch /dη from [9, 10].Now in order to quantify the direct photon spectra, we first integrate the p T spectra above p T = 1 GeV/c and obtain the left plot of Fig. 6. This plotis another representation of the direct photon scaling, where the integratedyield from the large systems scales with dN ch /dη by the same power α =1 .
25, i.e., dN γ /dy grows faster than dN ch /dη . Also, we show the integratedyield of extrapolations (extrapolated down to p T = 1 GeV/c) of the fit top+p data and of the three different pQCD calculations scaled by N coll . It isquite interesting that the prompt photons (the purple band) and integratedpQCD curves have nearly the same slopes as that of the large systems. cta-template printed on December 27, 2018 [GeV/c] T p . ) h / d c h d y / ( d N T p N / d d - - - - - - [GeV/c] T p . ) h / d c h d y / ( d N T p N / d d - - - - - - (a) = 62.4 GeV, 0-86% NN sAu+Au, = 39 GeV, 0-86% NN sAu+Au, = 62.4 GeVsp+p, = 63 GeVsp+p, = 62.4 GeVs pQCD, = 39 GeVs pQCD, [GeV/c] T p . ) h / d c h d y / ( d N T p N / d d - - - - - - = 200 GeV: NN s Au+Au, 0-20%Au+Au, 20-40%Au+Au, 40-60% = 200 GeVsp+p, = 200 GeVsp+p fit, = 200 GeVs pQCD, (b) [GeV/c] T p . ) h / d c h d y / ( d N T p N / d d - - - - - - = 2760 GeV, 0-20% NN sPb+Pb, =200 GeV, 0-20% NN sAu+Au, = 62.4 GeV, 0-20% NN sAu+Au, = 200 GeV, 0-40% NN sCu+Cu, = 2760 GeVs pQCD, = 200 GeVs pQCD, PHENIX (c)
Fig. 5. This three-panel plot is from [7] showing the direct photon spectra nor-malized by ( dN ch /dη ) . . The comparison is shown for Au+Au data in minimumbias collisions at 62.4 GeV and 39 GeV in the panel (a); for Au+Au data in threecentrality bins at 200 GeV in the panel (b); and for different A+A systems at fourbeam energies in the panel (c). The panels (a) and (b) also show p+p data, andall the panels show perturbative QCD calculations at respective energies. »h | h /d ch dN > . G e V / c ) T / d y ( p g d N - - - -
10 110 + X dir g fi p(d,A) + p(A) = 2760 GeV NN sPb+Pb, = 200 GeV NN sAu+Au, = 62.4 GeV NN sAu+Au, = 39 GeV NN sAu+Au, = 200 GeV NN sCu+Cu, = 200 GeV NN sd+Au, = 200 GeV NN sp+Au, = 200 GeVsp+p, scaled prompt photons coll N = 200 GeVsp+p fit, = 2760 GeVspQCD, = 200 GeVspQCD, = 62 GeVspQCD, = 1.25 a PH ENIX preliminary »h | h /d ch dN > . G e V / c ) T / d y ( p g d N - - - - - - - + X dir g fi A+A/p+p = 2760 GeV NN sPb+Pb, = 200 GeV NN sAu+Au, = 200 GeVsp+p, = 62.4 GeVsp+p, scaled prompt photons coll N = 200 GeVsp+p fit, = 2760 GeVspQCD, = 200 GeVspQCD, = 62 GeVspQCD, = 1.25 a PHENIX
Fig. 6. The left plot shows the direct photon yield, integrated > . p T , vs. dN ch /dη , for five A+A datasets at different collision energies and for onep+Au, one d+Au and one p+p datasets at 200 GeV (some of the unintegrated dataare shown in Fig. 2 and Fig. 3). The right plot shows the yield integrated > . p T , for two A+A data sets and two p+p datasets. In both plots theintegration is carried out for the data, p+p fit and pQCD curves from Fig. 5. By integrating above p T = 5 GeV/c, we get the right plot of Fig. 6. Herethe observed scaling behavior is expected (since R AA = 1 [11]), thoughwe see that the slopes are almost the same as those in the left plot. Theintegrated pQCD yields scaled by N coll are also shown. The black dashedlines are the power-law fits over the A+A data with a fixed α = 1 .
25 slope. acta-template printed on December 27, 2018
4. Concluding remarks and summary
The PHENIX collaboration has measured low momentum direct photonsin Cu+Cu ([17]) and Au+Au at √ s NN = 200 GeV, in Au+ Au at √ s NN =62.4 and √ s NN = 39 GeV as well as in p+p, p+Au and d+Au at √ s NN =200 GeV. Considering all the available data on small and large systems atvarious energies, we observe a surprising scaling behavior of direct photonsin large systems, namely: at a given center-of-mass energy the low- andhigh- p T direct photon invariant yields from A+A collisions scale with N coll ;then N coll is proportional to dN ch /dη across different energies; meanwhile,for all energies the low- p T yield scales like ( dN ch /dη ) α . PHENIX has alsodiscovered direct photon excess yield (within systematic uncertainties) atlow- p T in central p+Au collisions above N coll scaled p+p fit, which mayoriginate from possibly existing QGP droplets in small central systems. Inthe low dN ch /dη region of the left plot of Fig. 6 we see a gradually increasingtrend of the integrated yield from small systems, which seems to intersectwith a trend from large systems. Both observed trends suggest the existenceof a “thermal transition region or point” between small and large systems.REFERENCES [1] A. Adare et al. (PHENIX Collaboration), Phys. Rev. Lett. , 132301 (2010).[2] A. Adare et al. (PHENIX Collaboration), Phys.Rev. C , 064904 (2015).[3] A. Adare et al. (PHENIX Collaboration), Phys.Rev. C , 064901 (2016).[4] H. van Hees, M. He and R. Rapp, Nucl. Phys. A , 256 (2015); R. Rapp,H. van Hees, M. He, Nucl. Phys. A , 696 (2014).[5] J. F. Paquet et al. , Phys. Rev. C , no. 4, 044906 (2016); C. Shen, U. Heinz,J.-F. Paquet, and C. Gale, Phys. Rev. C , 044910 (2014).[6] Y. M. Kim, C. H. Lee, D. Teaney and I. Zahed, Phys. Rev. C , no. 1, 015201(2017); C.-H. Lee and I. Zahed, Phys. Rev. C , 025204 (2014).[7] A. Adare et al. (PHENIX Collaboration), [arXiv:1805.04084 [nucl-ex]].[8] A. Adare et al. (PHENIX Collaboration), Phys. Rev. C , 054907 (2013).[9] A. Adare et al. (PHENIX Collaboration), Phys. Rev. C , 024901 (2016).[10] K. Aamodt et al. (ALICE Collaboration), Phys. Rev. Lett. , 032301 (2011).[11] S. Afanasiev et al. (PHENIX Collabor.), Phys. Rev. Lett. , 152302 (2012).[12] J. Adam et al. (ALICE Collaboration), Phys. Lett. B , 235 (2016).[13] A. L. S. Angelis et al. (CCOR Collaboration), Phys. Lett. B , 106 (1980).[14] A. S. Angelis et al. (CMOR Collaboration), Nucl. Phys. B , 541 (1989).[15] T. Akesson et al. (AFS Collaboration), Sov. J. Nucl. Phys. , 836 (1990).[16] J.-F. Paquet, Private communication, (2017).[17] A. Adare et al.et al.