For Publisher’s use DEEPLY VIRTUAL COMPTON SCATTERING AT HERA LAURENT FAVART On behalf of the H1 and ZEUS Collaborations Universit´e Libre de Bruxelles, I.I.H.E., Belgium 1 E-mail: [email protected] 0 0 2 ResultsonDeeplyVirtualComptonScatteringatHERAmeasuredbytheH1andZEUSCollaborations n arepresented. Thecrosssection, measuredforthefirsttime,isreportedforQ2>2GeV2. a J 5 1 Introduction e γ e 2 e Along these last years, the exclusive vector e e γ 1 meson production (as ρ and J/Ψ) has been γ* v 6 studied extensively at HERA and has pro- γ* 4 vided very interesting results, in particular, p p p p 0 testing for the domain of applicability and 1 the relevance of perturbative QCD in the Figure2. Bethe–Heitler process. 0 1 field of diffraction (see e.g.1,2). Here, we re- 0 port the first analyses of a similar process, / Bethe–Heitler process (Fig. 2). x theDeeplyVirtualComptonScattering,con- e sisting in the hard diffractive scattering of - Sincethevirtualphotonisscatteredonto p a virtual photon off a proton (Fig. 1). The e DVCSprocessoffersanewandcomparatively mass shell in the final state it is necessary h to transfer longitudinal momentum from the clean way to study diffraction at HERA. In : v photon to the proton, i.e. forcing a nonfor- comparison to vector meson production it i ward kinematic situation3,4. In the picture X avoids large uncertainties on the theoretical of the two gluonexchange the fractionalmo- r predictions due to the meson wave-function. a mentumofthegluonsarethereforenotequal, The largest interest comes from the access it which implies the involvement of the skewed gives to the skewed parton distributions of the proton3. (or nonforward) parton distribution. InpresenceofahardscaletheDVCSpro- e cess can be completely calculated in pertur- e bative QCD. In the present case, the photon γ* γ virtuality Q2, above a few GeV2, insures the presence of a hard scale. LO QCD calcula- tions exist, based on the two gluon exchange model5. The factorization theorem in per- p p turbativeQCDhavingbeendemonstrated4,6, the DVCS process provides a unique way to Figure1. DVCSprocess. extractthesedistributionsfromexperimental data through the measurement of the asym- TheDVCSprocesscontributesto the re- metryofthephotonazimuthalangledistribu- actione+p→e+γp,whosetotalcrosssection tion due to the interference with the Bethe– is dominated by the purely electromagnetic Heitler process7. For Publisher’s use is observed. The LO calculation including From an experimental point of view, the DVCS and the Bethe–Heitler processes HERA kinematic enables the DVCS process achives a good description of the experimen- tobestudiedinalargerangeinQ2,W andt taldata. AclearDVCSsignalisstillseenaf- andoffersthepossibilitytostudythisdiffrac- ter the photon energy cut is increased. Also tive mechanism in detail. a shower shape analysis of the calorimetric clusters was performed that shows that the 2 Analysis strategy signal originates from photons and not from π0 background. Around the interaction region both exper- ZEUS 1996/97 Preliminary iments, H1 and ZEUS, are equipped with γN tracking devices which are surrounded by 140 γ candidates calorimeters. Since the proton escapes the 120 main detector through the beam pipe only 100 the scattered electron and photon are mea- sured. Therefore the event selection is based 80 on demanding two electromagnetic clusters, 60 one in the backward and one in the central 40 or forwardpart of the detector (θ<∼140o - the 20 backward direction (θ = 0) is defined as the directionoftheincomingelectron). Ifatrack 00 0.25 0.5 0.75 1 1.25 1.5 1.7θ5 (r2adia2.n25s) can be reconstructed it has to be associated 2 tooneoftheclustersanddeterminestheelec- Figure3. Distribution(uncorrected)ofthepolaran- gle of the photon candidate with an energy above tron candidate. To enhance the DVCS con- 2GeV. Datacorrespondtothefullcircles. Thepre- tributionin comparisonto the Bethe–Heitler dictionfor the Bethe–Heitler process isindicated by process the phase space has to be restricted the open triangles. The prediction of Frankfurt et al. basedoncalculationsincludingtheBethe–Heitler by demanding the photon candidate in the andtheDVCSprocessisshownbytheopencircles. forward part of the detector. The H1 analysis selects more specifically the elastic component by using, in addition, 3.2 H1 detectors which are placed close to the beam In the H1 analysis, the DVCS cross sec- pipe and which are used to identify parti- tion is measured in the kinematic region: clesoriginatingfromprotondissociationpro- 2 < Q2 < 20GeV2, |t| < 1GeV2 and cesses. 30 < W < 120GeV. The proton disso- ciation background has been estimated at 3 Results around 10% and subtracted statistically as- suming the same W and Q2 dependence as 3.1 ZEUS for the elastic component. The acceptance, The first observation of the DVCS process initialstateradiationofrealphotonsandde- was reported by the ZEUS collaboration in tector effects have been estimated by MC to 19998. In the analysis a photon virtuality extract the elastic cross section. Q2 > 6GeV2 is demanded. In Fig. 3 the polarangulardistributionofthe photoncan- InFig.4the differentialcrosssections as didates is shown. A clear signal above the a function of Q2 and of W are shown. The expectations for the Bethe–Heitler process data are compared with the Bethe–Heitler For Publisher’s use predictionaloneandwiththe fullcalculation 4 Conclusion including Bethe–Heitler and DVCS. The de- The DVCS process has been observedby the scription of the data by the calculations is H1 and ZEUS Collaborations. Cross section good, in shape and in absolute normaliza- tion when a t slope is chosen between 7 and measurements have been presented for the 10 GeV−2. first time. The experimental results are well described by the calculations of Frankfurt et al. Itisimportanttonoticethat,attheLO, the interference term cancels when integrat- ingovertheazimuthalangleofthefinalstate References photon(asinthedifferentialcrosssectionsin Q2 and of W). 1. P.Schleper, these proceedings. 2. P.Marage, “Hard Diffraction in Vector Meson Production at HERA”, Review 2 10 presented at the XXVIII International ] 2 V H1 Preliminary Symposium on Multiparticle Dynamics, e G BH + DVCS (FFS) hep-ph/9904255. b/ BBHH 3. A.V.Radyushkin Phys. Rev. D56, p 10 [ 30 < W < 120 GeV 5524 (1997). 2 | t | < 1 GeV2 Q 4. J.C.Collins, A.Freund, Phys. Rev. d / D59, 074009 (1999). σ d 1 5. L.L.Frankfurt, A.Freund, M.Strikman Phys. Rev. D58, 114001 (1998) and Phys. Rev. D59, 119901E(1999). -1 10 6. X.-D. Ji and J.Osborne, Phys. Rev. 5 10 15 20 D58, 094018 (1998). Q2 [ GeV2 ] 7. A.Freund Phys. Lett. B472, 412 (2000). ] V H1 Preliminary 8. P.R.B.Saull, Proceedings of the Int. e 3 G BH + DVCS (FFS) Europhysics Conf. on High En- / b BBHH ergy Physics, Tampere, Finland, 15- p [ 2 < Q2 < 20 GeV2 21 July 1999, edited by K. Huitu, W 2 | t | < 1 GeV2 H. Kurki-Suonio and J.Maalampi (hep- d / ex/0003030). σ d 1 0 40 60 80 100 120 W [ GeV ] Figure 4. The measured cross sections of the reac- tion e+p → e+γp as a function of Q2 and W are shownandcomparedtotheoretical predictions. The uncertaintyinthetheoretical prediction,shownhere asashadedbandisdominatedbytheunknownslope of the t-dependence of the DVCS part of the cross section,assuming7<b<10GeV−2