February7,2008 7:53 ProceedingsTrimSize: 9inx6in myproc 5 0 0 2 MEASUREMENT OF PROMPT PHOTON IN √s = 200GeV n a pp COLLISIONS J 5 2 K. OKADAFOR THE PHENIX COLLABORATION 1 RIKEN-BNL Research Center, v Brookhaven National Laboratory, 6 Upton, NY 11973-5000, USA 6 E-mail: [email protected] 0 1 0 5 0 Thisreportpresentsthepreliminaryresultforthepromptphotonproductioncross / section in proton-proton collisions in the mid-rapidity region using the PHENIX x detector. The NLO pQCD calculation is in good agreement with the data. The e measurement was made with and without an isolation cut. The isolation cut - p significantly improves the signal puritywithout reducing the signal yield. This is e animportantstepforthefuturespinasymmetrymeasurement. h : v i X 1. Introduction r a Prompt photon production in proton-proton collisions is a good probe of the parton structure in the proton. Its leading sub-process is gluon-quark Compton scattering. In polarized proton-protoncollisions at RHIC, it is a good tool to access the spin structure function of the gluon in the proton. Two types of processes contribute to the prompt photon production cross section: theso-called’direct’component,wherethephotonisemittedviaa pointlike(direct)couplingtoaquark,andthefragmentationcomponent,in whichthe photonoriginatesfromthe fragmentationofafinalstate parton. In the latter case, hadronicactivities areexpected to be accompaniedwith the photon. These processes are thought to be separated by applying an isolation cut on the photon. The cross section of the prompt photon at the mid-rapidity region in proton-proton collisions at √s = 200GeV has been measured1 using the PHENIX detector. In this paper, we report the measurement with increased statistics, and an isolation cut on photons is applied for the first time. 1 February7,2008 7:53 ProceedingsTrimSize: 9inx6in myproc 2 2. Experimental Setup −1 The analysis is based on the integrated luminosity of 266 nb collected by the PHENIX detector 2 at the RHIC proton-proton run in May 2003. Photons are detected by electromagnetic calorimeters(EMCal) in the cen- ◦ tral arms, each of which has an azimuthal coverage of 90 and a pseudo- rapidity (η) coverage of 0.35. In this paper, we report only on the mea- ± surement done with one arm. The EMCal has such a very fine granularity (10 10mrad2) that we can efficiently identify background photons by re- con×structing π0’s. The energy scale and resolution were tuned by checking the π0 masspeakpositionanditswidth. InadditionPHENIXhasatrack- ing system in front of the EMCal which was useful to reject background clusters produced by charged hadrons. Data were collected with a min- imum bias trigger provided by the beam-beam counters (BBC) placed in 3.0< η <3.9regionandwithanEMCaltrigger. Byrequiringahighenergy | | cluster with the EMCal trigger, event condensation by a factor of 120 was achieved. Fig. 1showsthatthetriggerefficiencyattainsaplateauat100% for single photons above E 3GeV for the active channels of EMCal. ∼ EMCal Trigger Efficiency 1 0.8 0.6 0.4 0.2 00 0.5 1 1.5 2 2.5 3 3.5 4 4.5 E[GeV] Figure1. Turn-onbehavioroftheEMCaltrigger. Thelineshowstheexpectedbehavior. 3. Analysis Procedure The prompt photon signal is obtained by subtracting known backgrounds from all EMCal clusters. The contribution of hadronic interactions is re- jectedbyaphotonshowershapecutandbyrequiringthatnochargedtrack is associated with the cluster. The dominant source of the background is two photon decays of π0’s. This background is estimated by reconstruct- ing π0’s from two photons. A photons is rejected as a π0 decay photon when the invariant mass of the photon and another photon in the same 0 event is consistent with mass of π . The accidental coincidence is taken February7,2008 7:53 ProceedingsTrimSize: 9inx6in myproc 3 into account using the vicinity of π0 mass window. If only one of the two decay photons from a π0 is detected, it mimics a prompt photon signal. Thisprobabilitydependsmainlyongeometricalacceptanceandkinematics of π0 decays, which can be well reproduced by a Monte-Carlo simulation. Thanks to the highly segmentated EMCal, two photon clusters from a π0 are well separated in the p region of this measurement. T Besides the simple subtraction method, an isolation cut is applied to prompt photon candidates. It requires that the energy sum (E ) in an sum angular cone around a photon is less than a certain fraction of the photon energy (Eq. 1). E (R<0.5)<E 0.1, R=p∆η2+∆φ2 (1) sum γ × The amount of photons from hadronic decays other than π0’s is esti- mated based on the total amount of π0’s. Inthesubtractionmethod,ofallphotonclusters,50%(20%)aretagged as photons from π0, and 85% (30%) are estimated as photons from all hadrons including missing π0’s at p = 5GeV/c (16GeV/c). In the isola- T tion method, ratios are improved to 25% (<5%) and 65% (<5%). The 0 uncertainties of the energy scale, of the π extraction, and of the hadron- 0 to-π production ratio are sources of systematic error. After the yield extraction, correction factors such as acceptance, effi- ciency, luminosity, and BBC trigger bias are applied to calculate the cross section. 4. Results and discussion Fig. 2 (left) showsour measurementofthe promptphoton crosssection as a function of p . Bands correspond to systematic errors which are listed T in Table 1 for our lowest and highest p bins. Curves show a NLO pQCD T calculation with 3 different scales 3. The theory agrees well with our ex- perimental measurement. In Fig. 2 (right), the two analysis methods are comparedand no significantreduction in the isolationmethod is observed. It suggests either (1) the ’direct’ component is the dominant source of the prompt photon or (2) the reduction of the ’fragment’ component by the isolation cut is small. In summary, we have measured the cross section of prompt photon in proton-proton collisions at √s = 200 GeV. The NLO pQCD calculation agrees with the measurement very well. The isolationmethod significantly enhances the signal purity. These results are important for future spin asymmetry measurements of this channel. February7,2008 7:53 ProceedingsTrimSize: 9inx6in myproc 4 -23)cV -23)cV PPHHEENNIIXX PPrreelliimmiinnaarryy ((SIsuoblatrtaiocnti)on) 3 (pbGe/dp103 PBNaLHCmnOd=ETs1 ENp /rQ2eQIppX6CrTMe ,D Ps pePrTn(D,ebt s2yFl ip sWmTys.iVtneomagaertylisca enrrgo)r. 3 (pbGe/dp103 Shaded box represents systematic errors s3d s3d E E 102 102 10 10 1 1 4 6 8 10 12 14 16 18 4 6 8 10 12 14 16 18 p(GeV/c) p(GeV/c) T T Figure2. [left]Resultofthesubtraction methodwithNLOpQCDcalculation. [right] Comparisontotheisolationmethod. Table1. Systematic errortable Subtraction Isolation Lowest Highest Lowest Highest 5−5.5 15−17 5−5.5 15−17 [GeV/c] [GeV/c] [GeV/c] [GeV/c] π0 photon estimation 30% 5 16 2 Nonπ0 photonestimation 27 6 8 1 Photonacceptance andsmearing 10 10 10 10 Photonconversioneffect 1 1 1 1 Luminositymeasurement 12 12 12 12 BBCtriggerbias 3 3 3 3 Total 43% 18 24 16 References 1. S.S. Adler et al.: Phys. Rev. D (to besubmitted) 2. K.Adcox et al.: Nucl. Inst. Meth. A499, 469-602 (2003) 3. Privatecommunication with WernerVogelsang (RBRC)