Direct Photon-Hadron Correlations Measured with PHENIX

Direct photon-hadron correlations greatly improve our ability to perform jet tomography in heavy-ion collisions because the momentum of the direct photon can be used to constrain the initial momentum of the opposing jet. By comparing the spectrum of away-side hadrons observed in heavy ion collisions to the spectrum seen in nucleon collisions we can quantify the medium modification to the fragmentation function due to energy loss of the away-side parton. High pT direct photon-hadron correlations have been measured with the PHENIX detector using a statistical subtraction method to remove the photon contribution from meson decays. The increased integrated luminosity in the most recent Au+Au RHIC run at √ sNN = 200 GeV provides substantially improved statistical precision and enhances the kinematic reach. These measurements are compared to PHENIX p+p results and several theoretical models of energy loss. In addition, we compare direct photon-hadron and π0-hadron correlations. Direct photon jet correlations are considered a “golden channel” for studying jet tomography since the transverse momentum of the jet is approximately balanced by the transverse momentum of the photon and photons do not interact with the medium created at RHIC. The lack of medium modification to the direct photons has previously been shown by the PHENIX direct photon RAA measurement [1], which shows that RAA of direct photons is consistent with 1 out to at least 14 GeV/c in pT . Direct photon hadron correlations also allow measurement of the fragmentation function of the opposing parton [2]. The fragmentation function is Dq(z) = dN(z)/(Nevtdz), where z = phadron/p jet. For γdirect − h correlations, z is approximately known, since the transverse momentum of the photon has the same magnitude as the opposing jet momentum to leading order. At NLO, however, the kT effect is important [3]. The measurement of these correlations, however, is complicated by the need to separate correlations involving direct photons from those involving inclusive photons. In heavy ion collisions, one of the largest sources of photons is decay photons, mostly from π0 → γγ. Direct photons are defined as all photons which do not result from a hadronic decay process since the decay photons are the only contribution subtracted from the inclusive sample. Prompt photons are the direct result of a hard scattering such as Compton scattering (qg → qγ) or quark-antiquark annihilation. Since Compton scattering is dominant, the away-side fragmentation is mostly quark jet fragmentation. Photons from jet fragmentation and medium-induced photons also contribute to the inclusive sample. To measure γdirect − h correlations, the contribution of decay photons are subtracted from inclusive γ − h correlations via a statistical subtraction method. More details on the method and results from the 2004 data have been presented elsewhere [4]. First, inclusive γ − h correlations are determined by measuring the angle between photons detected in the electromagnetic calorimeter and charged hadrons measured in the PHENIX tracking system. The Preprint submitted to Nuclear Physics A September 24, 2009 measured inclusive photon yield can be written as: Yinc = NdirYdir + NdecYdec Ninc . (1) The conditional yield here is defined as Y = Npair/Ntrig. To determine the decay contribution, first, π0−h correlations are measured. To translate π0−h correlations to γdecay−h, a Monte Carlo simulation is used to determine the probability that a pion decays into a photon in a particular pT bin. A correction is applied in Au + Au to account for the contribution from the η decays based on the p + p measurement. Higher mass decays are accounted for in the systematic error based on PYTHIA. Once the yield from γdecay − h is determined, the subtraction described in Equation 2, which is Equation 1 rearranged, is performed, where Rγ = Ninc/Ndec and has been previously measured [1]. Ydir = RγYinc − Ydec Rγ − 1 . (2) To measure the modification of the fragmentation function of the opposing parton, the awayside yield of the resulting γdirect − h jet function is measured in the head region, |∆φ − π| < π/5, and plotted as a function of zT . For correlations between trigger and associated particles, zT = 〈pT,assoc〉/〈pT,trig〉. For γdirect − h, zT = 〈pT,h〉/〈pT,γ〉 〈pT,h〉/〈pT, jet〉 and is used to approximate z. The zT distributions are shown in Figure 1. The upper set of points are for p + p collisions from the 2005 and 2006 data combined and the lower set are for Au + Au collisions from the 2007 data. The 2007 data set is a factor of four larger than the previously shown 2004 data. The improved data set allows for a measurement at higher pT,h, extending the results to higher zT . Figure 1: (Color online) The zT distribution for p + p scaled by a factor of 10 (blue) and Au + Au (red). All pT bins appear to obey approximate zT scaling. Therefore, all points for each collisional system were fit with the function, dN dzT = Ne −bzT . For p + p the slope was measured to be b = 6.89 ± 0.64. The slope in Au + Au is b = 9.49 ± 1.37 which exceeds that in p+p, as expected for modified fragmentation in the QCD medium due to constant fractional energy loss of the away-side parton. However, the statistical uncertainties in these data limit the significance of the difference to only 1.3 σ.