Direct Photons at the PHENIX Experiment : From Large to Small Systems †

Direct photons are a unique probe to study the properties of the medium created in heavy ion collisions. Low transverse momentum (pT) direct photons are of special interest since thermal photons are supposed to be dominant, while at high pT direct photons come from initial hard scattering (pQCD). PHENIX has observed a large excess of direct photon yield as well as large azimuthal anisotropy at low pT in Au+Au collisions at the c.m.s energy per nucleon pair √ sNN = 200 GeV. The mechanism to produce a large direct photon yield with a large elliptic anisotropy (v2) is not well understood yet. PHENIX has made systematic measurements of direct photons with different collision energies and system configurations. It has been found that direct photon yield dNγ/dη is proportional to charge particle multiplicity (dNch/dη). This behavior holds for beam energies measured both at RHIC and at the LHC in large systems. This scaling suggests that there is a transition from p+p to A+A system which could be understood with the analysis of smaller systems like p+Au and d+Au.


Introduction
Due to the fact that photons do not interact strongly in heavy ion collisions, they can carry primordial information about the collision at the time of their production, namely temperature evolution and collective motion of the matter.Photons are very abundant in heavy ion collisions, however, most of them come from hadronic decays, mainly π 0 and η to γγ.The so called direct photons are those which do not come from hadronic decays; they can be determined experimentally by subtracting from the total yield the expected fraction coming from hadron decays.In large systems and at low p T , it is expected that the dominant source of direct photons is emission from the QGP and from the hadronic gas phase; whereas at high p T , their origin is connected to initial state interactions produced in the hard scattering.

Challenges
The PHENIX experiment measured not only the direct photon yield but also its azimuthal anisotropy in Au+Au collisions at √ s NN = 200 GeV.A large yield and a large v 2 were found, which is challenging to theoretical models because a large yield implies early emission when the temperature is higher, but a large v 2 implies late emission when the flow is fully developed.Details can be found in [1].
In order to understand this observation, PHENIX has performed similar measurements on different datasets.In large systems, PHENIX has measured: Au+Au collisions at √ s NN = 200, 62, 39 GeV and Cu+Cu at √ s NN = 200 GeV.In small systems, PHENIX has measured: p+p, p+Au, d+Au (MB) at √ s NN = 200 GeV.We will discuss these measurements in the next sections.

Yield
The measurement of direct photon yield in Au+Au collisions at √ s NN = 200 GeV is well established [2].The yield follows the p+p signal scaled to the number of binary collision (N coll ) above 4 GeV/c in p T , but there is a significant excess of the direct photon yield at p T below 3 GeV/c.The excess has a nearly exponential shape.Recently, the direct photon yield at low p T was measured in Au+Au collisions at 62.4 GeV and 39 GeV via the external conversion method [3].A clear direct photon signal excess was found at low p T compared to the pQCD prediction scaled by the N coll .The results can be seen in Figure 1.Cu+Cu collisions at 200 GeV were also explored [4] via the internal conversion method.The direct photon yield for Cu+Cu minimum bias (MB) can be seen in the right panel of Figure 1.It shows that the Cu+Cu signal has a clear direct photon yield excess over the binary-scaled p+p baseline.Similar behavior is expected in other large systems, and it was recently reported in Pb+Pb 2760 GeV by the ALICE experiment [5].

Scaling
In order to compare the different collision systems and energies, and gain insight in the mechanism of the direct photon production, the yield was integrated.It can be represented as a function of the number of participants (N part ), but N part has limitation because it saturates at the same value for similar system size at different beam energies.So instead of N part , the number of charged particles (N ch ) was used.Figure 2 shows the integrated direct photon yield for p T > 1 GeV/c (left) and p T > 5 GeV/c (right) versus the number of charge particles, dN ch /dη.In the plot there is data from Au+Au collisions at 200 GeV, 62 GeV, 39 GeV, Cu+Cu collisions at 200 GeV and Pb+Pb collisions at 2760 GeV from ALICE [5].A universal scaling behavior of dN γ /dη with (dN ch /dη) 1.25 is observed, independent of the system, energy or centrality for p T > 1 GeV/c.The p+p fit scaled by the number of collisions is significantly lower.This scaling behavior can indicate that the low p T direct photons are produced around the critical temperature when the transition from QGP to hadron gas happens.The integrated yield at hight p T , p T > 5 GeV/c, follows the p+p fit scaling by the number of collision for Au+Au data as expected, where the dominant photons are the ones that come from hard scattering.

Small Systems
PHENIX has recently measured low momentum direct photon production in p+Au collisions at 200 GeV.While the p+Au MB direct photon signal is consistent with N coll p+p scale, the most central p+Au show a hint of excess over the binary-scaled baseline, as seen in Figure 3.A previous measurement in small systems, d+Au [6] did not show a significant modification of the direct photon p T distribution.These last data are consistent with the results from MB in p+Au collisions, Figure 3 left.Figure 4 left shows the ratio of the direct photon yield in p+Au to p+p scaled by the number of binary collisions as a funtion of p T .The ratio shows a hint for a significant excess, which is compatible with theoretical calculations of thermal emission [7].

Conclusions
In order to understand better the low p T direct photon signal, PHENIX has performed the measurement at different energies and systems: Au+Au collisions at 39 and 62.4 GeV, Cu+Cu collisions at 200 GeV.A unique scaling behavior of the integrated direct photon yield with respect to the dN ch /dη was observed with collision energies and different systems and centralities.This scaling may be an indication that direct photons are produced during the transition from QGP to the hadron gas.The hint of an excess in p+Au can indicate a possible formation of QGP droplets in the small systems.The data suggest the transition between p+p to A+A like yields in a narrow multiplicity range.

Figure 2 .
Figure 2. Integrated direct photon yield vs. dN ch /dη for different collision systems and energies, integrated in p T > 1 GeV/c (left) and p T > 5 GeV/c (right).The plots are from [3].

Figure 3 .Figure 4 .
Figure 3. Direct photon yield at low p T in d+Au and p+Au MB (left), and p+Au central collision at 200 GeV (right).Going back to the integrated yields, the data from p+Au and d+Au are added to what was shown in Figure 2 left in Figure 4 right.The data from small systems indicate a transition from p+p like yields to large systems like yields in a narrow multiplicity range.