ELSEVIER Nuclear Physics B 435 ( 1995) 3-20

NUCLEAR PHYSICS B

Observation of hard processes events in 7~ interactions

H 1 Collaboration

in rapidity gap at HERA

T. Ahmedc, S. Aid”, V. AndreevX, B. Andrieuab, R.-D. Appuhnk, M. Arpagaus *, A. Babaev”, J. Baehra”, J. Banq, P Baranov”,

E. Barrelet ac, W. Bartel k, M. Barthd, U. Bassler ac, HP Beck ak, H.-J. Behrend k, A. Belousov ‘, Ch. Berger a, H. Bergstein a,

G. Bernardiac, R. Bernet G, G. Bertrand-Coremans d, M. Besaqon i, R. Beyer k, P Biddulph “, J.C. Bizot aa, V Blobel m, K. Borras h,

F. Botterweckd, V. Boudry ab, A. Braemer”, F. Brasse k, W. Braunschweiga, V. Brisson aa, D. Bruncko9, C. Brune O,

R. Buchholz k, L. Btingener “, J. Btirgerk, F.W. Biisser m, A. Buniatiank,‘, S. Burke’, G. Buschhorn”, A.J. Campbellk, T. Carli ‘,

F. Charles k, D. Clarke”, A.B. Clegg r, B. Clerbauxd, M. Colomboh, J.G. Contreras h, J.A. Coughlan e, A. Courauaa, Ch. Coutures’,

G. Cozzika i, L. Criegee k, D.G. Cussans e, J. Cvachad, S. Dagoret ac, J.B. Dainton ‘, M. Danilov w, W.D. Dau P, K. Daumah, M. David i,

E. Deffurk, B. Delcourtaa, L. Del Buonoac, A. De Roeck k, E.A. De Wolf d, P. Di Nezza af, C. Dollfus &, J.D. Dowel1 ‘, H.B. Dreis b,

V. Droutskoi w, J. Duboc ac, D. Dtillmann m, 0. Diinger m, H. Duhm!, J. Ebert ah, T.R. Ebert ‘, G. Eckerlin k, V. Efremenko w, S. Egli ak,

H. Ehrlichmann ai, S. Eichenberger ak, R. Eichler s, F. Eisele “, E. Eisenhandler t, R.J. Ellison “, E. Elsen k, M. Erdmann n,

W. Erdmannq, E. Evrardd, L. Favart d, A. Fedotov w, D. Feeken m, R. Felst k, J. Feltesse I, J. Ferencei O, F. Ferrarotto af, K. Flammk,

M. Fleischer k, M. Flieser ‘, G. Fltigge b, A. FomenkoX, B. Fominykh w, M. Forbush g, J. Formanek ae, J.M. Foster “, G. Frankek, E. Fretwurst ‘,

E. Gabathuler ‘, K. Gabathuler ag, K. Gamerdinger ‘, J. Garvey c, J. Gayler k, M. Gebauer h, A. Gellrich k, H. Genzel a, R. Gerhards k,

0550.3213195/$09,50 @ 1995 Elsevier Science B.V. All rights resrxved SSDlOSSO-3213(94)00541-9

-1 HI Collars ~tion/Nuclear Physics B 435 (1995) 3-20

U. Goerlachk, L. a; jerlich f, N. Gogitidze ‘, M. Goldberg ac, D. Goldner h, B. Gone,, lez-PineiroaC, A.M. Goodall s, I. Gorelov”,

P. Goritchev w, C. Grab ’ , H. GrBssler b, R. Grassier b, T. Greenshaw s, G. Grindhammer”, A. ~1 iruber”, C. Gruberr’, J. Haack”‘, D. Haidtk, L. Hajdukf, 0. Hamon ‘, M. Hampel”, E.M. Hanlon’, M. Hapkek,

W. J. Haynes e, . Heatherington t, G. Heinzelmann m, R.C.W. Henderson ‘, H. Henschelai, R. Herma”, I. Herynekad,

M.F. Hess ‘, W. Hildes .eime, P. Hille, K.H. Hiller ai, C.D. Hilton “, J. Hladky ad, K.C. Hoeger “, M. H6ppnerh, R. Horisberger”g, Ph. Huet d,

H. Hufnagel”, M. bbotson”, H. Itterbeck”, M.-A. Jabiol’, A. Jacholkowska”“, C. J> zobsson “, M. Jaffre =, J. Janoth’, T. Jansen k, L. Jonsson “, K. Johann: en m, D.P. Johnson d, L. Johnson ‘, H. Jung ac,

P.I.P. Kalmust, D. <ant’, R. Kaschowitz b, P. Kasselmann e, U. Kathage P, H.H. (aufmann”‘, S. Kazarian k, I.R. Kenyon c,

S. Kermichey, C. Keukc r”, C. Kiesling ‘, M. Klein”‘, C. Kleinwort”, G. Knies k, W. Ko s, T. K jhler a, H. Kolanoski h, F. Kole s, S.D. Kolya “,

V. Korbel k, M. corn h, P. Kostka”‘, S.K. Kotelnikov”, T. Krtimerkamper h, R/ .W. Krasny f,ac, H. Krehbiel k, D. Krticker b,

U. Kriigerk, U. Krii ier-Marquisk, J.P. Kubenka”, H. Kiister b, M. KuhlenZ, T. KurCa 1, J. Kurzh6ferh, B. Kuznikah, D. LacouraC,

F. Lamarche ab, R. Lane er s, M.P.J. Landon t, W. Lange ai, P. Lanius ‘, J.-F. Laporte ‘, A. Lebec :v ‘, C. Leverenz k, S. Levonian k,x, Ch. Ley b,

A. Lindnerh, G. Lint strijm’, F. Linselk, J. Lipinskim, B. Listk, P. Loch aa, H. Lohmand :r “, G.C. Lopez t, V. Lubimov w, D. Luke h,k,

N. Magnussen ah, E. Ma1 novski ‘, S. Mani g, R. Maracek q, P Marage d, J. Marks Y, R. Marshal ‘, J. Martens ah, R. Martin k, H.-U. Martyn a,

J. Martyniak f, S. Wasson b, T. Mavroidis’, S.J. MaxfieldS, S.J. McMahon s, A. 1 Jlehta “, K. Meier O, D. Mercer “, T. Merz k,

C.A. Meyer ak, H. Me ,erah, J. Meyer k, S. Mikockif, D. Milstead ‘, F. Moreau ab, J.V. MN )rris e, G. Mtiller k, K. Mtiller ak, P. Murin 9, V. Nagovizin w, R. p ahnhauer ai, B. Naroskam, Th. Naumann”‘,

P.R. Newman c, C Newton r, D. Neyret ac, H.K. Nguyen ac, F. Niebergall m, C. NielJuhr k, R. Nisius ‘, G. Nowak f, G.W. Noyese,

M. Nyberg-Werther , H. Oberlack”, U. Obrock h, J.E. Olsson k, E. Panarop, A. Panit#.:h d, C. Pascaud aa, G.D. Pate1 ‘, E. Peppel a’,

E. Perez i, J.P. Phillips’ , Ch. Pichler ‘, D. Pitzl 4, G. Popes, S. Prell k, R. Prosi k, G. Radel k, F. Raupach ‘, P. Reimerad, S. Reinshagen k,

P. Ribarics z, H. Rick h, V. Riech”, J. Riedlberger ‘j, S. Riess m,

HI Collaboration/Nuclear Physics B 435 (1995) 3-20 5

M. Rietz b, S.M. Robertson ‘, P. Robmann *, H.E. Roloff ai, R. Roosen d, K. Rosenbauer a A. Rostovtsev w, F. Rouse g, C. Royon i, K. Riiter ‘, S. Rusakov ‘, K. Rybicki f, R. Rylko ‘, N. Sahlmann b, E. Sanchez ‘,

D.P.C. Sankeye, M. Savitsky w, P. Schacht ‘, S. Schick k, P. Schleper”, W. von Schlippe’, C. Schmidtk, D. Schmidtah, G. Schmidtm, A. Sch6ningk, V. Schroderk, E. Schuhmann”, B. Schwab”,

A. Schwind”‘, U. Seehausen m, F. Sefkow k, M. Seidel”, R. Sell k, A. Semenov w, V. Shekelyan”, I. Sheviakov”, H. Shooshtari ‘,

L.N. ShtarkovX, G. Siegmon P, U. Siewertp, Y. Sirois ab, 1.0. Skillicornj, P. SmirnovX, J.R. Smithg, Y. SolovievX, H. Spitzer m, R. Starosta”,

M. Steenbock”, P. Steffenk, R. Steinbergb, B. Stellaaf, K. Stephens’, J. Stierk, J. Stiewe’, U. Stosslein”‘, J. Strachota3d, U. Straumannak,

W. Struczinskib, J.P. SuttonC, S. Tapprogge’, R.E. Taylor”‘,““, V. Tchernyshov w, C. Thiebaux ab, G. Thompson t, P Truol ak, J. Turnau f, J. Tutas “, P. Uelkes b, A. Usik ‘, S. Valkarae. A. Valkarova ae, C. Vallee Y,

P. Van Esch d, P. Van Mechelen ‘, A. Vartapetian k,l, Y. Vazdik”, M. Vecko od, P. Verrecchia’, G. Villet’, K. Wacker h, A. Wagener b,

M. Wagener ‘g, I.W. Walker r, A. Walther h, G. Weberm, M. Weber k, D. Wegener h, A. Wegner k, H.P. Wellisch ‘, L.R. West ‘, S. Willardg, M. Winde ai, G.-G. Winterk, A.E. Wright “, E. Wtinsch k, N. Wulffk.

T.P. Yiou ac, J. Zaeek ae, D. Zarbock I’, Z. Zhang “, A. Zhokin w, M. Zimmer k, W. Zimmermannk, F. Zomer aa and K. Zuber’

a 1. Physikulisches Institut der RWTH, A~zchen. Germitny’

h III. Physiknlisches lnstitut der RWTH. Aachen. Germany 2

’ School of Physics and Space Research, University of Birmingham, Birmingham, UK’

* inter-University Institute for High Energies l/LB-VUB, Brussets; Universitrtrre Instellingen Antwerpen.

Wilrijk, Belgium ’

c Rutherford Appleton Laboratory, Chilton, Didcot, l/K’

’ Institute for Nucleur Physics, Cracow: Polandi

s Physics Department nnd IIRPA. University of Ccltfornia, Davis CA, USA e

h Institut fir Physik, Universitiit Dortmund, Ljortmund, Germany2

i CEA. DSM/DAPNIA, CE-SACLAY Gtf-sur-Yvette. Frunce

I Department of Physics and A.stronontv University of Gla~~o~~. Glasgow, UK 1

’ DE% Hamburg. Germany’

’ /. hstitut fir Experimentalphysik. Unit,ersitiit Hcmtburg, Humburg. Germany2

m II. Institutftr Experimentalphysik, Universitiit Hamburg, Hamburg, Germany’

n Physikalisches Institut, Universitiit Heidelberg, Heidelberg, Germany2

” lnstitut fiir Hochenergiephsik, Universitiit Heidelberg, Heidelberg, Germany2

P Institut fur Reine und Angewnndte Kernphysik. Universitiit Kiel, Kiel, Germuny2

9 Institute of Experimental Physics, Slovak Arndettty of S.iences, Ko.Fice. Slovak Republic

r School of Physics and Materials. Untversity <f’ Lancaster Lancaster UK3

’ Department of Physics. University of Liverpool, Liverpool, UK3

’ Queen Mary and Westfield College, London, UK’

’ Physics Department, University of Lund Lund, Sweden’

rtion/Nuclear Physics B 435 (1995) 3-20

‘r Physics Depcn ent, University of Manchester, Manchester, UK3 w institute for T/n -eticnl and Experimental Physics, Moscow, Russia

y Leb rv Physical Institute, Moscow, Russia Y CPPM, Universrt I’Aix-Marseille II, IN2P3-CNRS, Marseille, France

’ Max-PI, ck-lnstitut fib Physik, Munich, Germany2 ‘ra LAL, Univ( tte de Paris-Sud, IN2P3CNRS Orsay, France

ah LPNHE, Eco folytechniyue, IN2P3-CNRS, Palaiseau, France x LPNHE, Unive ‘t&s Paris VI and VII, IN2P3CNRS Paris, France

ad Institute of Phy.sks Czech Academy of Sciences, Prague, Czech Republic 8 aG Nuclear Cm t; Charles University, Prague, Czech Republic8

‘I” LNFN Roma and Dipcl imento di Fisicn, Universita “La Sapienza “, Rome, Italy ng PUI Scherrer Institu!, Villigen, Switzerland

il” Frtchbereich Physik, Bergisch Universitiit Gesumthochschule Wuppertul, Wuppertal, Germany2 ili DESK 1ns1 rt .ftir Hochenergiephysik, Zeuthen, Germany2

ai lnstitut f Teilchenphysik, ETH, Zurich, Switzerland” ti Phvsik-Ins/

at Stllnford

tt der Universitiit Zurich, Zurich, Switzerland” inectr Accelerutor Center, Stanford CA, USA

Received 4 November 195 : revised 18 November 1994; accepted 22 November 1994

Abstract

Events with no hadronic energy

are observed in photon-proton it

GeV. These events are interpretec

in photon diffractive dissociation

a function of transverse energy, i by a Monte Carlo calculation inc

low in a large interval of pseudo-rapidity in the proton direction

xactions at an average centre of mass energy (6) of 200 1s photon diffractive dissociation. Evidence for hard scattering

demonstrated using inclusive single particle spectra, thrust as

d the observation of jet production. The data can be described ding hard photon-pomeron scattering.

1. Introduction

The electron-proton collide HERA has turned out to be an important research facil-

ity for the understanding of 1 gh energy photon-hadron interactions [l-7]. Quasi-real photons are produced in ep lllisions by electrons which are scattered through small

angles. Photon-proton centre If mass system (CMS) energies of up to 300 GeV are

’ Visitor from Yerevan Phys. Inst., rmenia.

* Supported by the Bundesministerir n ftlr Forschung und Technologie, FRG under contract numbers 6AC17P

6AC47P, 6D0571, 6HH17P. 6HH271 SHD171, 6HD271, 6KI17P 6MP171, and 6WT87P 3 Supported by the UK Particle Ph! its and Astronomy Research Council, and formerly by the UK Science

and Engineering Research Council.

4 Supported by FNRS-NFWO, IISP IIKW.

5 Supported by the Polish State Co ,mittee for Scientific Research, grant No. 204209101.

’ Supported in part by USDOE gra t DE F603 9IER40674.

’ Supported by the Swedish Nature Science Research Council.

’ Supported by GA CR, grant no. : 12/93/2423 and by GA AV CR, grant no. 19095. ‘) Supported by the Swiss National cience Foundation.

HI Colilbor-ation/Nuclear- Physics B 43:; (I 995) .f-20 7

possible. This is one order of magnitude larger than Ihe CMS energies achieved so far in fixed target experiments. At these energies high mass diffractive dissociation processes

in photoproduction can be studied. The hadronic interaction of quasi-real photons with matter is fairly well described by

a model where the photon converts into a virtual hadronic state, mainly the p”( 770)

meson, and subsequently interacts with a target hadron. In recent yp studies it was shown

that, similar to hadron-hadron scattering, photoproduction exhibits both the production of transverse jets and a substantial rate of events which contain particles with transverse momenta pr values larger than a few GeV/c [ l-61. These phenomena can be readily

interpreted in terms of the interaction of a parton in the incident hadron and a parton in

the photon, and is termed a resolved photon process. In addition to the resolved process

the photon can couple directly to quarks in the proton. This is termed a direct photon

process. Models including leading order (LO) QCD diagrams for resolved and direct

processes indeed describe the gross features of these data. As a result of the similarity with hadron-hadron collisions, one expects a diffrac-

tive scattering component in the yp cross section. Diffractive scattering involves the

exchange of energy-momentum between the incident hadrons, but no exchange of quan-

tum numbers. The diffractive cross section is expected to be large, namely about 30% of the total cross section.

In diffractive processes both incident particles can keep their original identity (elastic scattering). or one or both of them can dissociate i.e. break up into a system of generally

low invariant mass and low multiplicity (diffractive dissociation). For single diffractive

dissociation only one of the incident particles dissociates after the interaction, while for double diffractive dissociation both incident particles dissociate. At HERA it is

possible to study the single diffractive dissociation channels Vp --f Xp (meson diffractive

dissociation) and Vp - KY (proton diffractive dissociation), where V stands for a vector meson ( p”( 770), w( 782), c$( 1020). etc.). A salient feature of single diffractive

dissociation at high centre of mass energies is a gap in rapidity between the non- dissociated hadron and the particles of the dissociated system. Due to the asymmetry

between the incident proton (820 GeV) and electron (26.7 GeV) energies, we can

identify meson diffractive interactions by requiring the absence of energy in the region

of the HERA detectors around the proton direction. The diffractive interaction events, which will be isolated by the rapidity gap cut detailed below, are the sum of meson single diffractive dissociation and a contamination of double diffractive dissociation, and are termed photon diffractive dissociation in the following. The contamination of elastic events Vp ---) Vp and proton single diffractive dissociation events Vp + VX seen in the detector is negligible (< l%), as well as that of events resulting from other Reggeon exchanges.

Phenomenologically, the observed properties of the diffractive cross section are de- scribed by triple-Regge theory where this process is viewed as the exchange of a pomeron [8]. This interpretation however gives no information on the details of the hadronic final states produced in diffractive dissociation. Traditionally the final state in diffractive dissociation is assumed to be described by a multiperipheral [9] type

of model in which particles are distributed throughout the final state phase space with limited transverse momentum. This approach has been used successfully so far for

x HI Coll~hor rtion/Nuclear Physics B 435 (1995) 3-20

comparisons with the available measurements of multiplicity and rapidity distributions of charged particles from the diffractive system. On the other hand, in modern QCD

language it is tempting to consider the pomeron as a partonic system [lo] which can be probed in a hard scattering process. Models based on this idea assume that the

pomeron behaves as a hadron aiid the concept of a pomeron structure function is intro-

duced [ I I - 131. In contrast to t re approach of assuming limited pr phase space, these models predict that, similar to h gh energy hadron-hadron scattering, high mass diffrac- tive dissociation exhibits the production of jets and a large m tail in the differential

transverse momentum distribution. Thus hard hadron-pomeron scattering events should

be observed in diffractive hadrctnic collisions at high energies. The UA8 collaboration

recently has shown evidence for jet production in diffractive pp events [ 141, inter-

preted as resulting from the collisions of partons from the proton with partons from the pomeron. Furthermore, within rhis partonic picture, these data have shown sensitivity to the parton distribution in the pomeron. New data to study the partonic structure of

the pomeron are essential to cheek this picture. In particular the pr spectra of particles,

the transverse thrust distributior, and the jet production spectra are expected to provide

important information on the ur derlying dynamics of the diffractive process. This analysis presents the stt.dy of events with a rapidity gap in photoproduction at

HERA. Based on the compariscn with Monte Carlo calculations these events are found

to be well compatible with the hypothesis of photon diffractive dissociation. Evidence

for hard scattering properties of these events is presented. The present analysis is based

on data collected in 1993 with the Hl detector at HERA, for collisions of e and p

beams with energies of 26.7 GcV and 820 GeV respectively.

2. Experimental set-up

The Hl detector is described in detail elsewhere [ 151. Here we describe briefly the components of the detector relevant for this analysis.

Measurements of charged particle tracks and the interaction vertex are provided by a central and forward tracking system, consisting of drift and multiwire proportional chambers. The central and forward tracking chambers cover the complete azimuth and

the range from about -2.0 to 3.0 in pseudo-rapidity 77 = - ln(tan(8/2)). Here B is

a polar angle with respect to the proton beam direction (z-axis), termed the forward region. Tracks found in the central and forward tracker are used to define the event vertex of the interaction. In this analysis we use for the inclusive single particle analysis

tracks fitted to the event vertex, with a minimum m of 150 MeVlc and / Ztrack - zvertex I< 12 cm, with &a& being the z-value of the track at the distance of closest approach. The central chamber is interspersed by an inner and an outer double layer of cylindrical multiwire proportional chambers (MWPC). Together with MWPCs from the forward tracking system, these chambers were used in the trigger to select events with charged tracks pointing to the interactil,n region. The MWPC trigger covers the rapidity range from -1.5 to 2.8.

The tracking region is surrounded by a fine grained liquid argon (LAr) calorime- ter [ 161 consisting of an electromagnetic section with lead absorber and a hadronic

HI C~)lluboration/Nuclea~ Physics B 43.i (1995) %20 9

section with steel absorber. The energy resolutions achieved in test beams were a/E =

12%/a for electrons and z 50%/a for pions, with E in GeV [ 15,171. The

LAr calorimeter covers the complete azimuth and the range from - 1.5 to 3.65 in pseudo-rapidity r]. The backward region (-3.3 < 7 < -1.5) is covered by a lead-

scintillator electromagnetic calorimeter (BEMC). For the measurement of the energy flow we use the cells of the LAr calorimeter and of the BEMC. The reconstruction of

calorimetric energies is described in more detail in Refs. [ 1.5,17 1, For the transverse thrust analysis tracks are also included. The calorimeters and the tracking system are

placed inside a superconducting solenoid which, together with the surrounding octagonal iron yoke, maintains a uniform magnetic field of 1.15 T along z in the tracking region.

An electron detector placed at a distance of z = -33 m allows tagging of electrons

scattered at small angles 6’ < 5mrad (t9’ = r - 8) from photoproduction processes. Together with a photon detector at z = - 103 m the electron tagger is used to mea-

sure Bethe-Heitler ep ---f epy events for luminosity determination. Both detectors are

TlCl/TlBr crystal Cerenkov calorimeters with an energy resolution of lo%/&.

3. Event selection

This analysis is based on a sample of tagged events in which the energy of the scattered

electron is measured in the HI small angle electron tagger. This limits the acceptance for

the virtuality of the incident photons to the range 3 x IO-’ GeV* < Q* < lo-* GeV’

where Q2 is given by Q* = 4E,E: cos2(6,/2). Here E, and EL are the energies of the incoming and scattered electron respectively and 8, is the angle of the scattered electron

with respect to the proton direction. The fractional energy of the photon as measured by

the small angle electron detector is required to be in the interval 0.25 < y < 0.7, where

v = 1 - E:/E,. This range in y corresponds to the energy interval of the yp system

(W) f rom 150 GeV to 250 GeV, with an average of about 200 GeV. The condition on y removes events from the tails of the electron energy distribution where the acceptance

of the electron tagger is small. It also removes the elastic Vp + Vp and single proton

diffraction dissociation Vp + Vx component where, due to the event kinematics, the vector meson escapes detection.

We use two data samples in this analysis. A first data sample, henceforth called minimum bias sample, has been collected with a loose, minimum bias trigger, designed for collecting events for general multiparticle studies of ‘yp collisions. This trigger was only active for an equivalent integrated luminosity of 117 nbb’. A second data sample,

henceforth called jet sample, was collected with a more selective trigger but permits the use of a data sample corresponding to an integrated luminosity of 289 nb-‘.

A coincidence of the small angle electron detector signal (Ei > 4 GeV ) with at least one track pointing to the vertex region was used to trigger on events from interactions of protons with quasi-real photons. For the minimum bias sample, the track condition

is derived from a very rough measurement from the MWPC trigger and requires a track pr 2 150 MeV/c. The acceptance of this trigger was studied in Ref. [7]. For the jet sample, a more restrictive track trigger condition was used which was based on fast signals of the central drift chamber trigger, It required a well defined track with

1 (’ c“,;’ 1ts ki:pl ” 1.1. (

It11 owing ,clcction 17 CI

cm Iyf ccl0ditions g!‘, .: ,\

in he lrz ~i~verse ( \ I’ 1 I’ w t 11 ver .*x must lil: it a’ is :‘ov~ern.J by the total pr 11.~

the l:ventr; .o contain II Inil, 1x1 the criggcr used for ,hi,; i,;- 11 wiih jets. Throughout IIris 1 !;I’ wx.

llvents containing COilll I,

of patterns recognized ir ! ht.

teractions of the proton h :a

random coincidence w ItI7 L cuts on low y/1 in coin~cide ICI Here JJ/~ := C(E-p:.)/:2

determined by summing 1; momentum component.

ln total about 125 000 e\ et sample and 174 000 events ft.

lh the tollowing the unc~ n-1

which were simulated tbro .tl; detailed s,imuiation of the c et

wi1.h stat :stical errors only rl

expected to be small (;< 2 been considered in this analy

4. Monte Carlo simulation

Soft hadronic events, i.e. el

[ 181 Monte Carlo program section. The differential cros: from hadronic diffractive d ISI

be’haviour. Here t is the 1’01 Mx is the: invariant mass 0 MS: and 1 has been con&n: diffractive system fragmerns only fragnnentation pr for tl properties of a longitudinal diffractive model.

For further study we use. a POMPYTl .O [ 201. This mc the proton vertex. The resul

~tion/Nuclear Physics B 435 (1995) 3-20

gular interval of 25” < 8 < 155”. per conditions were reconstructed and subjected to the

:nts were accepted if, apart from the electron tagger tt least one track originated from the interaction region

rnd thus an interaction vertex was reconstructed. This

n the region (-35 < z < 25 cm), the length of which

unch length. In addition, for the jet sample we required transverse energy, ET, of 5 GeV. With this requirement

was found to have an unbiased acceptance for events the transverse energy is defined with respect to the yp

showers and beam halo muons were rejected by means :ntral tracking system and in the LAr calorimeter. In- with the residual gas in the fiducial vertex region in

ral in the electron tagger are removed by appropriate ith large c pz / cp values (see Ref. [ 71 for details).

) is measured in the calorimeter where C( E - pz ) is _ cos0) for all cells, and pz denotes the longitudinal

satisfy the final selection criteria for the minimum bias

he jet sample.

ed data will be compared with Monte Carlo predictions he detector and reconstructed as for the real data. The 3r parts is described in Ref. [ 151. All figures are shown

QED radiative corrections to the jet cross section are

for the present experimental conditions and have not

Igrams

ts with no hard scale, are generated with the PYTHIA5.6

ich includes the diffractive components of the yp cross ction for diffractive events follows the properties known iation, namely an exponential t dependence and a 1 /Mi

nomentum transfer between the incident particles and ie diffractive system. Experimentally this behaviour in n a fixed target photoproduction experiment [ 191. The 1 hadrons and produces a final state involving essentially jroduced hadrons. This model reproduces the kinematic [se space model. Henceforth it is referred to as the soft

)del which explicitly includes diffractive hard scattering:

1 assumes the emission of a (space-like) pomeron at g photon-pomeron interaction is simulated as the hard

Hl Collaboration/Nuclear Physics B 435 (1995) 3-20 I I

scattering of the photon (direct process) or partons in the photon (resolved process)

with partons in the pomeron according to LO QCD calculations for the hard scattering processes. These collisions give rise to the production both of particles with a large pr and of jets.

To compare with non-diffractive hard scattering in photon-hadron interactions, the

Monte Carlo program PYTHIA5.6 was used in its high pr option for photoproduction.

Here yp interactions are simulated as the hard scattering of the partons in the photon and in the proton, according to the LO QCD calculations, summing the contributions

from direct and resolved photon interactions. The effects of initial and final state QCD

radiation are described by leading logarithm type parton showers. The possibility of multiple interactions between the partons from the photon and proton is included. This model is referred to as the non-diffructive model in the following. It describes well the basic features of the high pr inclusive yp data [ 31.

Both the high pr PYTHIA and POMPYT models are based on LO QCD calculations.

Since a QCD calculation is divergent for a --i 0, where ,& is the transverse momentum

of the out-going partons in the hard scattering process, a minimum fi cut-off value

byi” is applied. For this analysis the cut-off has been chosen to be @Tin = 2.0 GeVlc. Due to the divergence, no absolute normalization of the Monte Carlo predictions will be used in the following. Instead comparisons are made with the Monte Carlo calculations normalized to the number of events in the data in the regions where these models

are expected to be applicable. For the parton distributions we use GRV leading order parametrizations both for the proton [ 211 and the photon [ 221. In this analysis the pomeron was assumed to consist of gluons and their distribution function within the pomeron was taken to be either “hard”, z,g( Z) - x ( 1 - z ), hereafter labelled “GO”, or “soft”, zg( z) N ( 1 - z)~, hereafter labelled “G.5”. The variable z = xg/p is the fraction

of the pomeron momentum carried by the struck gluon involved in the interaction.

The results from high pr jet production in diffractive proton anti-proton interactions

mentioned above favour the hard GO distribution. Unless otherwise specified we use for comparison with the data the GO densities for the pomeron. For hadronic fragmentation the Lund model [23] is used in all Monte Carlo programs.

5. Rapidity gap events

The energy flow for the selected events was investigated using the variable vmax, the pseudo-rapidity either of the most forward calorimetric energy deposit of 400 MeV or the most forward detected track with a transverse momentum pr > 150 MeV/c. In Fig. 1

the Tmax spectrum is shown for the minimum bias sample of photoproduction events. The largest pseudo-rapidity which can be observed in the HI liquid argon calorimeter, TLAr, is about 3.65 and in the forward tracker, vm, about 3. These values vary slightly due to the position of the ep event vertex. Fig. 1 shows that for most events there is energy close to r]LAr. However there is also a class of events which have a small vmax value, i.e. a large empty region in the calorimeter in the proton direction.

The Tmax spectrum can be qualitatively understood by comparing it with model pre- dictions for diffractive and non-diffractive processes. The comparison with the PYTHIA

iii ! r~licd)rsi don/Nuclear Physics B 435 (1995) 3-20

I I

--~- PY rHlA sd ~ PY rHlA nd+sd

Fig. I. Maximum pseudo-rapidity vmax distribution in yp events compared to a diffractive (dashed line) and

a non-diffractive (shaded area) Monte &lo model, and their sum (full line).

Monte Carlo predictions in Fig 1 shows that the events with small 7,rrnax are consistent

with photon diffractive dissocia ion. The non-diffractive prediction (PYTHIA nd; shaded area) exhibits a sharp fall off \ ith increasing gap size, i.e. decreasing vmax, and clearly

does not account for the vrnax tail for vrnax < 2. The photon diffractive dissociation component (PYTHIA sd; dash1 d line), calculated with the soft diffractive model, gives

a good description of the spec trum for vmax < 2. The shape of the vmax distribution

for the Monte Carlo events w th a rapidity gap results from the l/M: ansatz for the differential diffractive cross se :tion. The distribution for the soft diffractive model is normalised to the region vmax : 2, while the distribution for the non-diffractive model

is normalised to qmax > 3. In all, the sum of the soft diffractive and non-diffractive

yp Monte Carlo calculations ( ull line) accounts reasonably well for the observed vmax spectrum. This observation doe ; not change by varying the minimum cluster energy and

the track PT requirements by 3 1%. Other production mechanisr IS for producing rapidity gap events can be envisaged.

Other Reggeon exchanges (e. :. rr, f~( 1270) exchange) can give rise to rapidity

gaps as well. However due tc the nature of the qmax cut applied here, we implicitly

select a region of M",/s FZ 2 p,t, < IO-*, where pomeron exchange is expected to

dominate over Reggeon exch: nge [ 241. Here xn/,, is the momentum fraction of the

proton carried by the pomeron

Hl Colhboration/Nuclear Physics B 435 (199.5) 3-20 13

10 -’

I,, / 13

+ Hi data ---‘-- soft diffraction 1 ..---- hard diffraction -__ soft+hard j

3

1 I

i-‘-.- - -.’ I I

E ; C I / I, I I,,,, I ,,I,,/,,,, I I /

0 1 2 3 4 5

PT [GeVlcl

Fig. 2. Transverse momentum distribution of charged particles for events with large pseudo-rapidity gap

( vrnax < 1.5) compared to Monte Carlo predictions, explained in the text.

To study the properties of the rapidity gap events, we select events with vmax < 1.5.

We additionally require for this and all further analyses in this paper that the calorimetric energy with 7 > 1.5 be less than 1 GeV and that the forward rapidity gap be the largest

in the event. These additional requirements remove only a few % of the events with

a forward rapidity gap, but enhance the purity of the sample. In total 7249 events survive this cut for the minimum bias sample. A consequence of this qmax selection

is that diffractive events with a dissociative system of mass Mx less than about 20

GeV are selected. The background in the rapidity gap data sample, due to accidental coincidence of a proton beam gas interaction with an electron scattered at small angle in the same event, was estimated to be four events using data taken with a non-colliding

proton bunch (pilot bunches) and using the rate of the small angle electron detector

alone. This background is neglected in the following. The remaining background from electron gas events was found to be more important. From electron pilot bunch studies we derived a contamination of 6% in the minimum bias sample. This background is subtracted statistically from the data presented in this section.

If diffractive dissociation involves a hard process then the corresponding underlying

parton interaction should be detectable in the event shape at the hadronic level. Possible signatures should be a large pi tail in the inclusive single particle distribution, and an azimuthal back to back correlation of transverse energy flow, growing with ET of the event. Such a correlation can be quantified by studying the thrust in the transverse momentum plane.

In Fig. 2 the pr spectrum for charged tracks in the range - 1.5 < q < 1.5 is shown, where pr is measured with respect to the beam axis. The shape of the distribution shows an exponential fall off at small pi values with a large tail extending to pr N 5

HI Cdlahon tion/Nuclear Physics B 435 (1995) 3-20

0.76 -

0.74 i

31 data

I J 10

ET [GeV]

Fig. 3. Average observed transverse thust as function of event ET. Data (points) and expectation for az-

imuthally isotropic events (line) with tile same average multiplicity as the data points for a given ET,

GeV/c. The same observation was made for the pi spectrum measured in the total

inclusive photoproduction samp e [ 31, where the events in the tail were identified with

hard scattering in ‘yp interactions. The shape of the pa spectrum is compared with the soft diffractive model predictior from PYTHIA (dashed-dotted line) and the diffractive

hard scattering model prediction from POMPYT (dashed line). For purpose of display the POMPYT prediction is nonnalised to the region pr > 1.5 GeV/c, and the sum of

POMPYT and PYTHIA is norrralised to the total spectrum. The soft diffractive model

describes well the exponential fall of the data at small m which represents the bulk of the data, but clearly cannot account for the pi tail. POMPYT gives a satisfactory description of the large Pr region. The overall Pr distribution is well described by the

sum of a soft and hard diffractive component (full line). Events from hard scattering of the partonic content of the photon with the proton, i.e. the remaining non-diffractive

events in the rapidity gap sample, are predicted to give a small contribution to the pi spectrum, namely less than 20 events. Similar results were obtained from studies of

the transverse energy spectrum of the rapidity gap events (not shown). The agreement with the diffractive hard scattering model may thus be taken as an indication for hard scattering at the parton level in photon diffraction.

A variable sensitive to the e\,ent shape is the transverse thrust

calculated from charged tracks and calorimetric clusters, without double counting. Here pr is the transverse momentum vector for each final state object, and In] = 1. The transverse thrust axis is given by the n vector for which the maximum of TL is obtained. For a two-body decay the reconstructed value is 7’_~(2) = 1, and only in the limit of

HI Collabaration/Nucleur Physics B 455 (I 995) 3-20 1.5

infinite multiplicity does the reconstructed thrust adopt the “theoretical” value TL (co) = 2/lr N” 0.64 for an isotropic production of particles in the transverse plane. Fig. 3 shows the average transverse thrust, (T_L), as a function of the ET of the event. At ET = 5 GeV. (T_L) falls with increasing ET, but changes behaviour l’or ET larger than 9 GeV. On the

other hand the average multiplicity of particles in our data is found to increase with

increasing ET. In the absence of a preferred transverse direction, i.e. for an isotropic particle distribution in the transverse plant (7’1) is expected to continuously decrease with growing ET, as a result of the increased average multiplicity. This is demonstrated by the curve in Fig. 3 which shows the expectation for thrust in isotropic final states

with the same average multiplicity as in the data. The data however does not continue

to fall at fir > 9 GeV, contrary to the implications of growing multiplicity. This is

an unambiguous signature for an underlying two-body structure in the transverse plane such as a hard scattering process. This evidence is independent of any specific model

for soft and hard scattering.

6. Jets in rapidity gap events

To substantiate the above evidence for hard scattering, a search was made for jet structures in the data sample of photoproduction events with a total transverse energy larger than 5 GeV (jet sample), which used an integrated luminosity of 289 nbb’.

A jet finding algorithm was applied to search for jets. The definition of jets was

based on transverse energy in the calorimeter contained within cones of radius R =

dw = 1 .O in the space of pseudo-rapidity 7 and azimuth 9 (in radians). For the present study, calorimeter cells from the region -2 < vcetl < 2.5 were considered

in the jet search. Within this region, we applied a ‘sliding window’ with radius R = 1 in order to find the cone with the highest transverse energy in the event. The transverse

energy ET within the cone was calculated as a scalar sum of the transverse energy of individual calorimeter cells inside it, while the cone axis is given as the vector pointing

from the event vertex to the transverse energy centroid of all cells within the cone [ 251. Cones with ET > 4 GeV and pseudo-rapidity in the range -1 < vjet < 1.5 were accepted as jets in this analysis.

The rlmax distribution for all yp events of the jet sample containing at least 1 jet is shown in Fig. 4. The distribution shows the same characteristics as Fig. 1, namely most events have an vmax close to VLAr, but a clear signal of events with vmax < 2 is observed.

The reduction in the proportion of events in the vrnas < 1 region compared to all data (cf. Fig. 1) is a consequence of the reduced phase space available for jet production.

The rllwX cut is strongly correlated with M x, the hadronic mass of the diffractively produced system. Requiring vrnax to be small preferably selects small Mx events, which have less phase space for the production of jets. The data have been compared with non-diffractive hard scattering yp (PYTHIA; shaded area) and y-pomeron (POMPYT,

dashed and full lines) predictions. The non-diffractive yp prediction, normalized to the total number of events, cannot account for the vmax distribution at smaller values. The POMPYT model, normalized to the number of events with vmax < 2, accounts better for the shape of the vmax spectrum, with a slight preference for the configuration including

Hi (‘ollabor- ~tion/Nuclear Physics B 435 (1995) 3-20

lo4

lo3

lo2

10’

loo

-1 0 1 2 3

rl max

f ;: !. Maximum pseudo-rapidity Q,,~~ ‘ktribution in tagged yp events containing jets with ET ,c, > 4 GeV in

t I 1 A rval - 1 .O < qjet < 1.5, compared to Monte Carlo expectations from a non-diffractive process (hatched

a ‘1 : I Ind a diffractive process assuming a pomeron with hard (GO; full line) and soft (G5; dashed line)

E I 11 11 nomentum distribution.

: 1 w 1 gluon distribution (GO) for the pomeron.

A+ isolate a diffractive sampl: by selecting events with r),,, < 1.5, as before. Without 1 IC 1.1 :quirement for a jet to be present 1632 events are selected. In this rapidity gap

: E I .:I) e we find in total 116 events which contain one jet and 19 events which contain

1 i’ : jets. From Monte Carlo studies we expect a contamination of four non-diffractive “7 i’lteractions with at least clne jet. The profiles of jets with 4 GeV < Erjet < 6

f ;f ‘bi and (vjet\ < 0.5 are show? in Fig. 5a and 5b and compared with those of jets in 1 1‘1 -li(rproduction events selected as above but with the requirement that there be no large (II ‘JQ trd rapidity gap, namely l;lmax > 1.5. The profiles are observed to be similar with t t, : xception of the large 11~ region (1 < 87 < 2 in Fig. 5b). This corresponds to III: r :gion required to be devoid of energy in the rapidity gap selection. The diffractive

I 3 .ci scattering Monte Carlo calculation as given by POMPYT describes fairly well the ill tii:ity gap event jet profiles. An example of an event with two jets, with an ETjet of ’ :;.EV and 4 GeV for the jets, is shown in Fig. 5c. The back-to-back structure of the 1~ 11 lets is clearly observed in the adjacent transverse view of the detector and in the !’ ~;t 9) distribution.

‘l’lle transverse energy, ETj,t. spectrum of all jets in the rapidity gap events is shown 11 1:; g. 6 as well as the distribution of the jet pseudo-rapidities, vjet, and the azimuthal 3r 131: between the jets, Ap(jet 1 -jet 2), for the two jet events. For two jet events, the 18: ; are clearly observed to be back-to-back in azimuth, a characteristic feature for a

: c rrf scattering process. The ETjet and the qj,t spectra are compared with the shape of t: :: POMPYT expectation, no,-malized to the number of events with qmax < 1.5, for ti‘ :: lard and soft choices of the pomeron parton distribution functions. The observed j : :s behave as expected from parton-parton scattering kinematics and are well described

- l Hl data q,.,<1.5 b) A Hl data ~,_>1.5

-- POMPYT

I

’ I

Fig. 5. (ah) Transverse energy flow around the jet axis for jets with 4 GeV < Evict < 6 GeV and Ivler( <

0.5: for data with vrnax <I..5 (circles), for data with 7 ,W > 1.5 (tl‘iangles), and for Monte Carlo events with %W < I.5 (full line). (c) Two-jet event with a large rapidity gap in the HI detector.

HI ~‘oll~~honrtion/Nuclear Physics B 435 (1995) 3. -20

10

\

, __- _--

--lb-

I I

30

20

10

0 J 4 -1 0 1 ~0 2 4 t?

0 Ii1 data Cl - POMPYT (GO)

ET jet 77 jet &J two ,ets

Fig. 6. (a,b) Inclusive jet distributions for large pseudo-rapidity gap events (7 mnx < 1.5) : transverse energy ET@ and pseudo-rapidity vjct. (c) Distribution of the azimuthal angle Ap between the jets for two jet events.

The data are compared with Monte Carlo predictions assuming a pomeron with hard (GO; full line) and soft

(GS; dashed line) gluon momentum distribution.

by the POMPYT Monte Carlo predictions. With the present statistics there is no clear preference for one or the other parton distribution, GO or G5, in the pomeron.

With the given selections the relative fractions of one- and two-jet events with respect to the total number of events with ET > 5 GeV are 7.1% and 1.2% respectively. To

reduce the sensitivity to the #” cut used in POMPYT and to restrict to a region where the data show dominantly hard scattering features (set Fig. 3)) a comparison is made of our data with this mode1 by increasing the minimum E.r of the event to 9 GeV. The one- and two-jet fractions are then 38.7% and 13.4% respectively, and the ratio (2 jets) /( 1 jet) is 0.35 X!Z 0.09. These results are compared with POMPYT predictions in Table 1. The (2 jets) /( 1 jet) ratio, which is only weakly sensitive to the remaining soft diffractive contribution, compares favourably with the prediction of a hard pomeron parton distribution, but it sholuld be noted that this ratio depends somewhat on the choice of the 6,” value-changing fir”‘” by 500 MeV/(, leads to a change in the ratio (2 jets)/( 1 jet) of 15%.

Thus, the ansatz of hard scattering between partons in the photon and partons in the pomeron is compatible with the results of the jet analysis of the data.

Table 1

HI Collaboraiion/~‘uclens Physics B 435 (1995) 3-20 19

Jet rates: data compared to POMPYT Monte Carlo calculations for yp events with E-r > 9 GeV and

vmnx < 1.5, and for jets with ETj,t > 4 GeV and -1 < 7,1et < 1.5.

Sample 1 jet events(%) 2 jet (%) 2 jets/l jet

Data (142 events) 38.7 13.4 0.35 & 0.09

POMPYT GO (@,“I” = 2 GeV) 46.4 10.1 0.22 & 0.05

POMPYT G.5 (@‘” = 2 GeV) 27.3 < 0.1

7. Conclusions

Events with a large rapidity gap with respect to the proton direction are observed in

yp interactions in Hl at HERA. These events are interpreted as diffractive dissociation of the photon. When a sample of such diffractive events is selected by means of

a cut vmax < 1.5, features attributable to the presence of hard partonic scattering are

observed. The yield of charged tracks from these events, as a function of their transverse momentum, pr, extends to values which cannot be accounted for in a model of pi limited

phase space, characteristic of our present knowledge of soft diffractive processes. A model which includes hard scattering between partons in the proton and in the pomeron reproduces well the charged particle yield m in the data. The transverse thrust does

not decrease with ET for large transverse energies indicating an underlying two-body structure in the transverse plane, giving a model-independent signature for a hard process.

Jets are found in the diffractive data sample. In two jet events, the jets are back-to-back in azimuth, substantiating the evidence for hard partonic scattering in diffractive collisions.

Acknowledgements

We are very grateful to the HERA machine group whose outstanding efforts made

this experiment possible. We acknowledge the support of the DESY technical staff. We

appreciate the big effort of the engineers and technicians who constructed and maintained the detector. We thank the funding agencies for financial support of this experiment. The non-DESY members of the collaboration also want to thank the DESY directorate for the hospitality extended to them. Finally we would like to thank G. Ingelman for useful discussions.

References

( I I HI Collab., T. Ahmed et al., Phys. Lett. B 297 ( 1992) 20.5.

121 HI Collab., I. Abt et al., Phys. Lett. B 314 (1993) 436.

131 Hl Collab., I. Abt et al., Phys. Lett. B 328 (1994) 176.

141 ZEUS Collab., M. Derrick et al., Phys. Lett. B 297 (1992) 404.

(5 1 ZEUS Collab., M. Derrick et al., Phys. Lett. B 322 ( 1994) 287.

16 I ZEUS Collab., M. Derrick et al., Inclusive jet differential cross sections in photoproduction at HERA,

DESY preprint DESY 94-176 (1994).

( HI C‘ol/nhorcirion/Nuclear Physics B 435 (I 995) 3-20

17 / HI Collab., T. Ahmed et al., Phys. Lett. B 299 ( 1993) 374;

M. Erdmann et al., New results from the Hl Experiment at HERA on photoproduction, deep inelastic

scattering and searches for new patticles, DESY preprint DESY 93-077 ( 1993)

S. Levonian, Proc. 28th Rencontre De Motiond, Lcs Arcs, France, ( 1993) 529.

j 81 K. Goulianos, Phys. Rep. 101 (1983) 169.

19 1 D. Amati et al., Nuovo Cimento 26 ( 1962) 896.

I IO 1 F. E. Low, Phys. Rev. D12 ( 1975) 163; S. Nussinov, Phys. Rev. Lett. 34 ( 1975) 1286.

1 I I 1 G. ingelman and P. Schlein, Phys. Lett. B IS2 (1985) 256.

/ I2 1 A. Donnachie and P. V. Landshoff, Phys. Lett. B I91 (1987) 309.

I 13 I N. N. Nikolaev and B. G. Zakharov, Z. Phys. C53 ( 1992) 331.

I 141 UA8 Collab., A. Brandt et al., Phyc. Lett. B 297 (1992) 417;

UA8 Collab., R. Bonino et al., Phys. Lctt. B 211 (1988) 239.

( IS I HI Collab., The H 1 detector at HERA, DESY preprint 93-l 03 ( 1993). to be published in Nucl. lnstrum.

Methods.

[ I61 HI Calorimeter Group, B. Andrieu et al., Nucl. Instrum. Methods A 336 ( 1993) 460.

1 I7 ] Hl Calorimeter Group, B. Andrieu et al., Nucl. Instrum. Methods A 336 (1993) 499.

[ I8 1 M. Bengtsson and T. Sjostrand, Comput. Phys. Commun. 43 (1987) 367;

H.-U. Bengtsson and T. Sjiistrand, Comput. Phys. Commun. 46 (1987) 43;

T. Sjostrand, CERN-TH-6488-92 ( 1992).

[ I91 E612 Collab., T. J. Chapin et al.. Phys. Rev. D 31 (1985) 17.

1201 P. Bruni and G. Ingelman. Proc. of the Europhysics Conference, Marseilles, France, July 1993, p. 595;

DESY preprint DESY 93. I87 ( 1993).

1211 M. Cluck, E. Reya and A. Vogt, Z. Phys. CS3 (1992) 127.

[ 221 M. Gliick. E. Reya and A. Vogt, Phys. Rev. D46 ( 1992) 1973.

[ 23 ] B. Andersson, G. Gustafson and T. Sjostrand, Phys. Lett. B 94 ( 1980) 211;

B. Andersson et al., Phys. Rep. 97 ( 1983) 31.

[ 24 1 G. Alberi and G. Goggi, Phys. Rep. 74 ( 1981) I.

125 1 J. E. Huth et al., Towards a standardization of jet definitions, Fermilab Conf. 901249-E ( 1990).