aa r X i v : . [ nu c l - e x ] M a y Direct Photons and Photon-Hadron Correlations at PHENIX
B. Sahlmueller a Department of Physics and Astronomy, Stony Brook University,Stony Brook, NY 11790, USA
Direct photons are a powerful tool to study the hot and dense matter created in heavy-ion collisions at RHIC, since they are created in the different stages of the collision. Sincethey do not interact via the strong force, they can travel through the hot and dense mattermostly unaffected. The PHENIX experiment has measured direct photons using differentmethods, over a broad range of transverse momentum ( p T ), in different collision systems,and at different energies. These measurements help interpreting the measurement of hadronsas well as understanding the temperature of the created quark-gluon plasma (QGP). Theazimuthal anisotropy of direct photons may shed light on the thermalization time of themedium. Using direct photons to tag jets is a crucial tool to understand the energy loss ofscattered partons in the medium. The experimental program at the Relativistic Heavy-Ion Collider has found evidence forthe creation of a quark-gluon plasma in collisions of Au nuclei at center-of-mass energies of200 GeV per nucleon. 1 One of the crucial signatures are direct photons that are produced insuch collisions and can traverse the created QGP mostly unaffected.Direct photons are defined as photons that are not from decays of hadrons, such as π or η . These photons are produced in different stages of a heavy-ion collision over a broadrange of transverse momentum. At large and intermediate transverse momentum ( p T ) they areproduced mainly from initial hard scattering processes of the colliding quarks or gluons suchas q + g → q + γ or q + ¯ q → g + γ , as bremsstrahlung emitted by a scattered parton, fromthe fragmentation of such quarks and gluons, or from the interaction of a scattered parton withthe strongly interacting medium created in such collisions 2. In the hard scattering processes, aparton is emitted opposite to the photon, that will subsequently fragment into a hadronic jet.Hence, the energy of the jet is balanced with the energy of the direct photon on the oppositeside. At low p T , the medium can emit thermal direct photons directly. Their p T distributiondepends on the average temperature of the medium. 3 To account for nuclear effects, the directphoton yield in Au+Au collisions is compared to the cross section in p + p collisions with thehelp of the nuclear modification factor which is defined as R AA = d N/dp T dy | AuAu h T AA i d σ pp /dp T dy , (1)where h T AA i is the nuclear overlap function.The azimuthal anisotropy of direct photons is sensitive to the different production processes,it is measured in terms of the anisotropy parameter v which is the second harmonic of the Fourier a for the PHENIX collaborationigure 1: Average nuclear modification factor < R AA > for direct photons in Au+Au and Cu+Cu collisions at200 GeV, plotted versus N part . expansion of the azimuthal distribution. The elliptic flow of thermal direct photons is sensitiveto the thermalization time τ of the QGP, small τ would lead to small v . 4Direct photons can be measured using different subsystems of the central arm of the PHENIXdetector, with different methods of measurement. The PHENIX detector is described else-where. 5 Photons can be measured directly using the Electromagnetic Calorimeters, the decayphotons from π and other mesons are subtracted statistically from the measured inclusive pho-ton sample, charged hadrons and electrons are rejected with the help of the central arm trackingdetectors. This method is described in more detail in earlier publications 6. It is most feasibleat high p T , at low p T the signal to background ratio gets too small to make a significant mea-surement. Therefore, two other methods have been developed, using the electron ID capabilitiesof the PHENIX detector and measuring direct photons indirectly through conversions.The so-called internal conversion method uses the idea that virtual direct photons are pro-duced in conjunction with real direct photons, and convert into low mass e + e − pairs, in a processsimilar to the π Dalitz decay. The method benefits from the limited phase space of such pairsfrom the π Dalitz decay, hence the signal to background ratio is improved compared to thedirect calorimeter measurement at low p T . The method is based on the measurement of e + e − pairs, the background is removed using like-sign pairs, the resulting distribution is comparedto a cocktail of dilepton pairs that includes all expected hadronic sources. An excess over thatcocktail at low invariant mass is then interpreted as a virtual direct photon signal. A moredetailed description of this method is given in 7.A new method has been developed to measure direct photons through external conversionsin the detector material. The back plane of the PHENIX Hadron-Blind Detector (HBD), whichwas installed during the 2007 RHIC run for commissioning, offers a well-defined conversionpoint for photons, about 60 cm away from the interaction point, with a radiation length ofabout 4% X . To account for the wrongly reconstructed opening angle of such conversion pairs,that leads to an apparent invariant mass, an alternate track model was developed that assumesthe origin of the pairs at the HBD back plane. Using this model moves the peak of the invariantmass from about 12 MeV/ c to 0, it also helps separating the conversion pair from pairs fromDalitz decays. Using this method, the misidentification rate of conversion photons is found tobe less than 3%. This method is described in more detail in 8.The nuclear modification factor R AA has been measured in 200 GeV Au+Au collisions andwas found to be consistent with unity for most p T . 9 This result is a confirmation that h T AA i scaling works, since at a first approximation, photons are not affected by the QGP. However,there appears to be a possible suppression of photons at p T >
15 GeV/ c , which is not fullyunderstood. Such a suppression could be an initial state effect, for example an effect of thedifferent isospin composition of the proton and of the gold nuclei, and would thus be visible in igure 2: Fraction of direct photons over inclusive photons, for different collision systems, at center-of-mass energyof 200 GeV: p + p , d +Au. Cu+Cu, Au+Au (from left to right). ξ -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 ξ d N / d t r i g / N -4 -3 -2 -1 <7 GeV/c T,h <15 GeV/c x 0.5
T,trig
PHENIX Au+Au 0-20% 5
T,h <15 GeV/c x 1
T,trig
PHENIX p+p 5
8% p+p Global Scale Uncertainty ± correlated systematic error Figure 3: ξ distribution for PHENIX Au+Au (black cir-cles) and p + p (open blue circles), compared to a MLLAprediction (red line) and TASSO data (green triangles). ξ ( P H E N I X A u + A u ) / ( . * T A SS O ) (PHENIX Au+Au 0-20%) / (0.1*TASSO)correlated systematic error 8.8% Global Scale Uncertainty ± fit to flat line AA I Figure 4: Ratio of Au+Au to TASSO data scaled by afactor of 10. The black line is a fit to the ratio, the blueline is a fit to the I AA using PHENIX p + p as reference. d+Au collisions at the same energy. The isospin effect would also be visible in Au+Au collisionsat 62.4 GeV, at lower p T , since it scales with x T = 2 p T / √ s NN . PHENIX has done bothmeasurements, with the 2003 and 2004 datasets, respectively, and they are both statistically toolimited to draw any conclusion. Direct photons have also been measured in Cu+Cu collisions at200 GeV, the averaged nuclear modification factor is compared with Au+Au collisions in Fig. 1and found to be consistent for similar numbers of participants.The virtual direct photon measurement has been done for four different collision systems at200 GeV collision energies, a p + p measurement works as a baseline to understand the othermeasurements, a d +Au measurement is used to look for effects of cold nuclear matter, and themeasurements in Au+Au and Cu+Cu are used to study properties of the QGP, also with respectto different system sizes. The ratio of direct photons and inclusive photons is shown in Fig. 2for all four collision systems. While the p + p measurement shows that pQCD agrees well withthe data, there is a clear excess in the Au+Au data. The excess is smaller in Cu+Cu and doesnot appear in the d +Au measurement, which gives evidence that it is indeed a final state effect.When fitting the excess over scaled pQCD of direct photon yield in Au+Au with an exponential,the average temperature of the medium can be extracted as the inverse slope of the function. Itis found to be 221 ± stat ) ± sys ) MeV.The external conversion measurement is still on its way, its final goal is to measure theelliptical flow of thermal photons. The method has been found to produce reliable results forthe inclusive photon v , this measurement agrees well with an earlier measurement using theEMCal. 8nother way of using direct photons to study the QGP is using direct photons as trigger totag jets, for photons from hard scattering processes balance the energy from jets on the oppositeside. To measure the modification of the jet, PHENIX uses correlations of direct photons andhadrons. Since direct photons cannot be measured on an event-by-event basis, first the yieldper trigger ( Y inc ) of inclusive photons and hadrons is measured as well as the yield per triggerof the π and hadrons. The yield per trigger for decay photons ( Y dec ) is calculated based on the π -hadron correlation, and finally, using the ratio of inclusive and decay photons, R γ , the yieldper event for direct photons is calculated as Y dir = R γ Y inc − Y decR γ − p + p and Au+Au data, it isplotted as a function of the fragmentation variable z = p hT /p triggerT to show the fragmentationfunction. An alternative way of plotting is showing the distribution as a function of ξ = − ln( x E ),where x E = p hT cos(∆ φ ) /p triggerT , this plot is shown in Fig. 3 for both p + p and Au+Au, in-cluding also e + e − data from TASSO 11 and a theoretical prediction from the Modified LeadingLogarithmic Approximation (MLLA) in the medium 12, the latter two curves are scaled downarbitrarily by a factor of 10 to account for the limited PHENIX acceptance.Dividing the yield in Au+Au by the yield in p + p , I AA can be calculated, a variable to quantifyeffects of the medium in Au+Au. Since the scaled TASSO agrees with the PHENIX p + p data,and since it extends to higher ξ , it can be used as a baseline instead of the p + p . The I AA likeratio calculated with the TASSO data is shown in Fig. 4. The shape is consistent with a flat lineand a suppression of the Au+Au yield below ξ = 1 .
8, but the points above indicate a change inshape and suggest an enhancement at highest ξ values, which can be interpreted as the responseof the medium to the lost energy.In summary, direct photons are a powerful tool to study the QGP created in ultrarelativisticheavy-ion collisions. The measurement of direct photons shows that binary scaling works whencomparing heavy-ion collisions to baseline p + p or d +Au collisions. Photons are also emittedfrom the medium directly or through interaction of partons with the medium. An excess ofdirect photons at low p T can be interpreted as a thermal signal from the QGP, and an averagemedium temperature of 221 ± stat ) ± sys ) MeV could be extracted. The measurement offlow of direct photons could further disentangle different photon production mechanisms. Directphotons are also crucial for probing the matter with direct photon-hadron correlations wherethe photon balances the jet energy. This measurement showed suppression of the away side anda shape suppression of the fragmentation function at high ξ which can be related to the mediumresponse to energy loss.1. K. Adcox et al., Nucl. Phys. A , 184 (2005)2. C. Gale, arXiv:0904.2184 (hep-ph) (2009)3. P. Stankus, Ann.Rev.Nucl.Part.Sci. , 517 (2005)4. R. Chatterjee, D. Srivastava, Phys. Rev. C , 021901 (2009)5. K. Adcox et al., Nucl. Instrum. Methods A , 469 (2003)6. S. S. Adler et al., Phys. Rev. Lett. , 232301 (2005)7. A. Adare et al., Phys. Rev. Lett. , 132301 (2010)8. R. Petti, conference proceedings, WWND 2011, to be published (2011)9. T. Isobe, J.Phys.G , 1015 (2007)10. M. Connors, Nucl. Phys. A , 335 (2011)11. W. Braunschweig et al., Z.Phys.C47