February7,2008 8:40 WSPC/TrimSize: 10inx7inforProceedings ichep06˙writeup PROMPT PHOTONS AND PARTICLE MOMENTUM DISTRIBUTIONS AT HERA N.GOGITIDZE 7 0 DESY, Notkestrasse 85, 22607 Hamburg, Germany 0 E-mail: [email protected] 2 n Recentresults,obtainedbytheH1andZEUScollaborations,arepresentedondifferentialcrosssec- a tions,forinclusivepromptphotonproductioninDISandforphotoproductionofpromptphotons J accompaniedbyahadronicjet. Alsopresentedarecrosssectionsofnormalisedscaledmomentum 6 distributionofchargedfinalstatehadrons,measuredbyH1inDISepcollisionsathighQ2 inthe 1 Breitframeofreference. 1 v Keywords:Promptphotons; Chargedparticlemultiplicity;HERA 3 3 1. Prompt photon production in photoproduction2. 0 1 1.1. Introduction 0 1.2. Prompt Photon identification 7 Isolated high transverse energy photons in 0 the final state are a powerful tool for de- Themainexperimentaldifficultyisthesepa- / x tailed studies of the Quantum Chromody- ration of the prompt photons from hadronic e namics (QCD) in hard interaction processes background,inparticularfromsignalsdueto - p andofthehadronicstructureoftheincoming π0 mesons. e particles. In the H1 analysis photons are iden- h : The photons are called “prompt” if tified in the liquid argon calorimeter by v they are directly coupled to the interact- a compact electromagnetic cluster with no i X ing quarks, instead of being produced as track pointing to it. The photon transverse r hadronic decay products. energy Eγ and pseudorapidity ηγ are re- a T In contrast to jet measurements, where stricted to 3 GeV < Eγ < 10 GeV and T the partonic structure is obscured by −1.2 < ηγ < 1.8. In order to com- the non-perturbative hadronisation process, pare with perturbative QCD (pQCD) calcu- prompt photons at large transverse energy lations, the photon isolation requirement is Eγ can be directly related to the partonic defined in an infrared-safe way, using: z = T event structure. Furthermore, the experi- Eγ/Ephotonjet > 0.9, i.e. z is the ratio of mental uncertainties connected with the en- the photon energy to the energy of the jet, ergy determination of an electromagnetic which contains the photon. The photon sig- showerinitiatedbyaphotonaresmallercom- nal is extracted by a shower shape analysis pared to the measurement of a hadron jet. which uses six discriminating shower shape However, the cross section for prompt pho- functions in a likelihood method. tonproductionissmallandtheidentification In the ZEUS analysis photons are iden- of photons in the detector is not trivial. tified using the barrel preshower detector Preliminary results for two analyses (BPRE). The BPRE prompt photon signal are presented here: An H1 study of in- is determined using the conversionprobabil- clusive prompt photons in deep inelastic ity in the detector, known from a study of scattering (DIS)1, and a ZEUS study of DVCS data3. prompt photons with an accompanying jet Thephotonkinematicrangeisrestricted 1 February7,2008 8:40 WSPC/TrimSize: 10inx7inforProceedings ichep06˙writeup 2 wtoh5er<e pEoTγsit<ive16ηGγ ecVorarensdp−on0d.7s t<o tηhγe p<ro1t.o1n, eV]eV] 4400 -1.2<h <1.8 GG 3355 H1 Data (prel.) beam direction. The photon isolation crite- b/b/ pp 3300 LO(a 3) Sum ria are similar to the ones used in the H1 [ [TT 2255 LO(a 3) LL analysis. Hadronic jets were selected in the dEdE LO(a 3) QQ kinematic range 6 < Ejet < 17 GeV and // 2200 −1.6<ηjet <2.4. T ssdd 1155 1100 55 1.3. Results 00 44 66 88 1100 Differential cross sections for the production EE [[GGeeVV]] TT of isolated photons in DIS measured by H1 areshowninFig. 1,asfunctionofETγ andηγ. pb]pb] 5500 3<ET<10 GeV A new LO(α3) calculation4 gives a good de- [ [ H1 Data (prel.) hhdd 4400 scription, although it lies slightly below the // LO(a 3) Sum ssdd LO(a 3) LL data. Atlargepseudorapiditiesthedominant 3300 LO(a 3) QQ contribution is radiation off the quark line (QQ),whereasinthebackwardregionthera- 2200 diation off the electron line (LL) dominates 1100 the cross section. A comparison with predictions of the 00 --11 00 11 PYTHIA and HERWIG generators (radia- hh tion off the quark) plus photon radiation off the electron is also made1. Both generators Fig. 1. Prompt photon differential cross sections describe the shapeinETγ well,butare lower dσ/dETγ for −1.2 < ηγ < 1.8 and dσ/dηγ for inthe absolutescale(factor2.3forPYTHIA 3GeV<Eγ <10GeV,forphotonvirtualitiesQ2> T and 2.6 for HERWIG). The ηγ distribution 4GeV2 andye>0.05. Curves showanLOcalcula- tionwithLLandQQgivingthecontributionofradi- is better described by PYTHIA. ationofftheelectronandthequarklinerespectively. The ZEUS differential cross sections as Astheinterferenceisverysmallitisnotshown,but functions of Eγ and ηγ for the prompt pho- includedinthesum. T tons are shown in Fig. 2. Two next-to- or- der (NLO) pQCD predictions are compared A comparison with the prediction of Lipa- tothedata. Inbothcalculationsseveralpho- tov and Zotov7 (LZ), which is based on k - T tonandprotonpdf’s,andseveralfragmenta- factorisation, corrected for hadronisation ef- tion functions are used. The FGH (Fontan- fects, is also shown. The LZ prediction gives naz, Guillet and Heinrich)5 calculation con- the best description of the Eγ and ηγ cross T tains additional higher order corrections to sections. In particular, it describes the low- the resolved photon process. Like the KZ est Eγ region better than the KZ and FGH T (Krawczyk and Zembrzuski)6 prediction, it NLO predictions. PYTHIA and HERWIG describesthedataratherwell. However,they do not rise as steeply at low Eγ as do the T both underestimate the observed cross sec- data and underestimate the measured cross tion at low Eγ and in the backward region. section. T The difference between the data and the Since the largest difference between the NLOQCDcalculationsismainlyobservedin NLO calculations and the data is observed thexobs < 0.75region(notshown),whichis in the regionof lowEγ andlowηγ, the level γ T sensitivetotheresolvedphotoncontribution. of agreement with NLO QCD was verified February7,2008 8:40 WSPC/TrimSize: 10inx7inforProceedings ichep06˙writeup 3 ZEUS ZEUS pb /GeV) 10 ep→ gprompt + jet + X (a) gh/d (pb) 112050 67e p<<→ EE TTgjepg tr<<om 11p67t + GG jeeetVV + X (a) g (T sd 5 sd/E 1 ZNELUOS+h (a7d7. p(Kb-Z1)) 0.5< mR < 2 0 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 NLO+had. (FGH) mR=1 h g -1 kT -fact.+had. (LZ) 0.5< mR < 2 10 PHYETRHWIAIG 6 6.3.5 eV) (b) 6 8 10 12 14 16 G b) 40 ETg (GeV) jetE (pb / T101-1 ZNNkPTEYLL -UOOTfaHS++chh It(A.aa7+ dd76h.. .a p3((dKFb.-GZ 1())LH 0Z) .) 5 m <0R .=m51R<< m 2R < 2 p / HERWIG 6.5 ( (b) sd 6 8 10 12 14 16 ghd Eje t(GeV) / T sd 20 jet (pb) 1680 (c) hd / 4 sd 2 0 0 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -1.5 -1 -0.5 0 0.5 1 1.5 2 h g h jet Fig.2. Thedifferentialcrosssectionfortheprompt Fig. 3. The differential cross section for the γ + photoneventswithanaccompanyingjetasfunctions jet events as function of: a) ηγ, b) Ejet and c) ηjet T of Eγ and ηγ compared to theoretical QCD calcu- compared to QCD calculations (with hadronisation T lations (including hadronisation corrections). The corrections) and Monte Carlo models. The cut on shaded bands correspond to the uncertainty in the Eγ isincreasedto7GeV. T renormalisation scale which was changed by factors 0.5and2. fragmentation universality. by increasing the minimum transverse en- The energy scale for the current region, ergy of prompt photons from 5 to 7 GeV. set by the virtual photon, is given by Q/2 Inthis casehadronisationcorrectionsareex- and, for purpose of comparison, is taken to pected to be smaller. As shown in Fig. 3 be equivalent to one half of the e+e− c.m. with Eγ > 7 GeV the NLO QCD and the energy E∗/2. T LZ predictions all agree well with the data. IntheBreitframethescaledmomentum ± The PYTHIA model then also agrees well, variable x is defined to be 2p /Q, where p h ± while HERWIG is still below the data. p is the momentum of a charged particle. h In e+e− annihilation events the equivalent variable is 2p±/E∗. 2. Charged Particle Momentum h The use of much higher statistics now distributions available at high Q as compared to previous InanH1analysis8theprocessofpartonfrag- studies9,10, as well as an improved under- mentationandhadronisationisstudiedusing standing of the H1 detector and associated inclusive charged particle spectra in DIS. In systematics, provide a much improved mea- the current region of the Breit frame a com- surement of the scaled momentum spectra. parisonwithonehemisphereofe+e− annihi- Results are nowavailableup to <Q>∼ 100 lationoffersa directpossibility to testquark GeV, close to the LEP-1c.m. energy, andin February7,2008 8:40 WSPC/TrimSize: 10inx7inforProceedings ichep06˙writeup 4 the full range of x (0 <x < 1). p p InFig. 4theinclusive,eventnormalised, chargedparticlescaledmomentumspectrum xp d H1 Preliminary is shown as a function of Q for nine differ- –/dn104 e+e- xp range ent bins of x . Also shown is a comparison N Rapgap 0.0 - 0.02 (x30) p 1/ to results from e+e− annihilation data (see 103 references in ref. 8). As seen, the ep and 0.02 - 0.05 (x5) e+e− data are in excellent agreement,which 102 supportsthe conceptofquarkfragmentation 0.05 - 0.1 (x2) universality. 10 0.1 - 0.2 Moving from low to high Q the x spec- p 0.2 - 0.3 tra become softer, i.e. there is a dramatic 0.3 - 0.4 increase in the number of hadrons with a 1 0.4 - 0.5 small share of the initial parton’s momen- 0.5 - 0.7 tum and a decrease of those hadrons with a 10-1 large share. These scaling violations (parton 0.7 - 1.0 splittinginQCD)arecompatibletothescal- 10-2 ing violations observedfor the DIS structure 10 102 Q, E* (GeV) functions. In the RAPGAP simulation11, also Fig. 4. H1 data for the event normalised inclusive showninFig. 4,the PartonShowermodelis scaled momentum spectrum as a function of Q for implemented. Itdescribesthe fragmentation nvainrieoudsiffee+reen−t xepxpreergiimonesn.tsA(lstoaksihnogwQn a=reEda∗t)a. frTohme processasthesplittingofparentpartonsinto DIS data are compared with the RAPGAP genera- two daughters(e.g. q →qg,q →qq,g →qq), tor. the splitting continues with daughters go- ing on to form parents. The evolution of 6. M. Krawczyk and A. Zembrzuski, Phys. Rev the partonshoweris basedonleading logQ2 D64, 14017 (2001); A. Zembrzuski and M. Krawczyk, hep-ph/0309308, 2003. DGLAP splitting functions. RAPGAP gives 7. A.V. Lipatov and N.P. Zotov, Phys. Rev. a very good description of the ep scaled mo- D72, 054002 (2005). mentum spectra over the whole range of xp. 8. H1 Coll., A. Aktas et al., Contributed paper tothe33rdInternationalConferenceonHigh Energy Physics, Moscow (2006). 9. H1 Coll., S. Aid et al., Nucl. Phys. B445, 3 References (1995); H1Coll., S.Adloffet al.,Nucl. Phys. 1. H1 Coll., A. Aktas et al., Contributed paper B504, 3 (1997); H1 Coll., S. Adloff et al., tothe33rdInternationalConferenceonHigh ICHEP98: 29th Int. Conf. on High Energy Energy Physics, Moscow (2006). Physics, Vancouver,531 (1998) 2. ZEUS Coll., S. Chekanov et al.,Contributed 10. ZEUSColl.,M.Derricketal.,Z.Phys.C67, paper to the 33rd International Conference 93 (1995); ZEUS Coll., M. Derrick et al., on High Energy Physics, Moscow (2006). Phys. Lett. B414, 428 (1997); ZEUS Coll., 3. ZEUS Coll., S. Chekanov et al., Phys. Lett. J. Breitweg et al., Eur. Phys. J. C11, 251 B573, 46 (2003). (1999). 4. A. Gehrmann-De Ridder, T. Gehrmann and 11. H. Jung, Computer Phys. Comm. 86, 147 E.Poulsen,hep-ph/0601073,2006; ibid.hep- (1995). ph/0604030, 2006. 5. M. Fontannaz, J.P. Guillet and G. Heinrich, Eur. Phys. J. C21, 303 (2001); M. Fontan- nazandG.Heinrich,Eur. Phys. J.C34,191 (2004).