Diffractive hard photoproduction at HERA and evidence for the gluon content of the pomeron

Full text

ELSEVIER
10 August 1995
PHYSICS LETTERS B
Physics Letters B 356 (1995) 129-146
Diffractive hard photoproduction at HERA and evidence for the
gluon content of the pomeron
ZEUS Collaboration
M. Derrick, D. Krakauer, S. Magill, D. Mikunas, B. Musgrave, J. Repond, R. Stanek,
R.L. Talaga, H. Zhang
Argonne National Laboratory, Argonne, IL, VSA48
R. Ayad *, G. Bari, M. Basile, L. Bellagamba, D. Boscherini, A. Bruni, G. Bruni, P. Bruni,
G. Cara Romeo, G. Castellini 2, M. Chiarini, L. Cifarelli 3, F. Cindolo, A. Contin,
M. Corradi, I. Gialas4, P. Giusti, G. Iacobucci, G. Laurenti, G. Levi, A. Margotti,
T. Massam, R. Nania, C. Nemoz, F. Palmonari, A. Polini, G. Sartorelli, R. Timellini,
Y. Zamora Garcia I, A. Zichichi
University and INFN Bologna, Bologna, Italy 38
A. Bargende 5, J. Crittenden, K. Desch, B. Diekmann6, T. Doeker, M. Eckert, L. Feld,
A. Frey, M. Geerts, M. Grothe, H. Hartmann, K. Heinloth, E. Hilger, H.-P Jakob,
U.F. Katz, S.M. Mari 4, S. Mengel, J. Mollen, E. Paul, M. Pfeiffer, Ch. Rembser,
D. Schramm, J. Stamm, R. Wedemeyer
Physikalisches Institut der Vniversitiit Bonn, Bonn, Germany35
S. Campbell-Robson, A. Cassidy, N. Dyce, B. Foster, S. George, R. Gilmore, G.P. Heath,
H.F. Heath, T.J. Llewellyn, C.J.S. Morgado, D.J.P. Norman, J.A. O’Mara, R.J. Tapper,
S.S. Wilson, R. Yoshida
H.H. Wills Physics Laboratory University of Bristol, Bristol, VK41
R.R. Rau
Brookhaven National Laboratory, Upton, L.I., VSA48
M. Arneodo7, M. Capua, A. Garfagnini, L. Iannotti, M. Schioppa, G. Susinno
Calabria University, Physics Dept. and INFN, Cosenza, Italy 38
A. Bernstein, A. Caldwell, N. Cartiglia, J.A. Parsons, S. Ritz *, F. Sciulli, P.B. Straub,
L. Wai, S. Yang, Q. Zhu
Columbia University, Nevis Labs., Irvington on Hudson, N.E, VSA4g
P. Borzemski, J. Chwastowski, A. Eskreys, K. Piotrzkowski, M. Zachara, L. Zawiejski
0370-2693/95/$09.50 0 1995 ELsevier Science B.V. All rights reserved
SSDZ 0370-2693(95)00803-9

130 ZEUS Collaboration/Physics Letters B 356 (1995) 129-146
Inst. of Nuclear Physics, Cracow, Poland42
L. Adamczyk, B. Bednarek, K. Jelefi, D. Kisielewska, T. Kowalski,
E. Rulikowska-Zarebska, L. Suszycki, J. Zajgc
Faculty of Physics and Nuclear Techniques, Academy of Mining and Metallurgy, Cracow, Poland”
A. Kottiski, M. Przybycieli
Jagellonian Univ., Dept. of Physics, Cracow, Poland43
L.A.T. Bauerdick, U. Behrens, H. Beier v, J.K. Bienlein, C. Coldewey, 0. Deppe, K. Desler,
G. Drews M Flasiliski lo, D.J. Gilkinson, C. Glasman, P. Giittlicher, J. GrolJe-Knetter,
B. Gutjak .
‘I’ T Haas, W. Hain, D. Hasell, H. Hefiling, Y. Iga, K. Johnson I*, P. Joos,
M. Kasemkn, R. Klanner, W. Koch, L. Kiipke 13, U. K(itz, H. Kowalski, J. Labs,
A. Ladage, B. Liihr, M. Liiwe, D. Liike, J. Mainusch, 0. Maficzak, T. Monteiro 14,
J.S.T. Ng, S. Nickel 15, D. Notz, K. Ohrenberg, M. Roco, M. Rohde, J. Rold&n,
U. Schneekloth, W. Schulz, F. Selonke, E. Stiliaris 16, B. Surrow, T. VoB, D. Westphal,
G. Wolf, C. Youngman, W. Zeuner, J.F. Zhou l7
Deutsches Elektronen-Synchrotron DESI: Hamburg, Germany
H.J. Grabosch, A. Kharchilava, A. Leich, M.C.K. Mattingly, A. Meyer, S. Schlenstedt,
N. Wulff
DESY-Zeuthen, Inst. fiir Hochenergiephysik, Zeuthen, Germany
G. Barbagli, P. Pelfer
University and INFN, Florence, Italy 3a
G. Anzivino, G. Maccarrone, S. De Pasquale, L. Votano
INFN, Laboratori Nazioaali di Frascati, Frascati, Italy 38
A. Bamberger, S. Eisenhardt, A. Freidhof, S. Sijldner-Rembold l*, J. Schroeder lg,
T. Trefzger
Fakultiit ftir Physik der Universitiit Freiburg i.Br., Freiburg i.Br., Germany35
N.H. Brook, P.J. Bussey, A.T. Doyle , . . 2o JI Fleck4, D.H. Saxon, M.L. Utley, A.S. Wilson
Dept. of Physics and Astronomy, University of Glasgow, Glasgow, UK47
A. Dannemann, U. Holm, D. Horstmann, T. Neumann, R. Sinkus, K. Wick
Hamburg University, I. Institute of Exp. Physics, Hamburg, Germany35
E. Badura21, B.D. Burow 22 L. Hagge, E. Lohrmann, J. Milewski, M. Nakahata23,
N. PavLl, G. Poelz, W. Schott, F. Zetsche
Hamburg University, II. Institute of Exp. Physics, Hamburg, Germany35
T.C. Bacon, N. Bruemmer, I. Butterworth, E. Gallo, V.L. Harris, B.Y.H. Hung, K.R. Long,
D.B. Miller, P.P.O. Morawitz, A. Prinias, J.K. Sedgbeer, A.F. Whitfield
Imperial College London, High Energy Nuclear Physics Group, London. UK41

ZEUS Collaboration/Physics Letters B 356 (1995) 129-146
U. Ma&k, E. McCliment, M.Z. Wang, S.M. Wang, J.T. Wu
University of Iowa Physics and Astronomy Dept., Iowa City, USA 48
P. Cloth, D. Filges
Forschungszentrum Jiilich, Institut fiir Kemphysik, Jiilich, Germany
131
S.H. An, S.M. Hong, S.W. Nam, S.K. Park, M.H. Suh, S.H. Yon
Korea University8 Seoul, South Koream
R. Imlay, S. Kartik, H.-J. Kim, R.R. McNeil, W. Metcalf, V.K. Nadendla
Louisiana State University, Dept. of Physics and Astronomy, Baton Rouge, LA, USA 48
F
. Barreiro 24 G. Cases, J.P Fernandez, R. Graciani, J.M. Hernandez, L. Her& 24,
L. Labarga , .
24 M Martinez, J. de1 Peso, J. Puga, J. Terron, J.F. de Troconiz
Univer. Autdnoma Madrid, Depto de Ftsica Tedrica, Madrid, Spain46
G.R. Smith
University of Manitoba, Dept. of Physics, Winnipeg, Manitoba, Canada33
F. Corriveau, D.S. Hat-ma, J. Hartmann, L.W. Hung, J.N. Lim, C.G. Matthews, P.M. Patel,
L.E. Sinclair, D.G. Stairs, M. St.Laurent, R. Ullmann, G. Zacek
McGill University, Dept. of Physics, Montreal, Quebec, Canada 33,34
V. Bashkirov, B.A. Dolgoshein, A. Stifutkin
Moscow Engineering Physics Institute, Moscow, Russia 44
G.L. Bashindzhagyan, PF. Ermolov, L.K. Gladilin, Yu.A. Golubkov, V.D. Kobrin,
I.A. Korzhavina, VA. Kuzmin, O.Yu. Lukina, A.S. Proskuryakov, A.A. Savin,
L.M. Shcheglova, A.N. Solomin, N.P. Zotov
Moscow State University, Institute of Nuclear Physics, Moscow, Russia 45
M. Botje, F. Chlebana, A. Dake, J. Engelen, M. de Kamps, P. Kooijman, A. Kruse,
H. Tiecke, W. Verkerke, M. Vreeswijk, L. Wiggers, E. de Wolf, R. van Woudenberg
NIKHEF and University of Amsterdam, Netherlands41
D. Acosta, B. Bylsma, L.S. Durkin, K. Honscheid, C. Li, T.Y. Ling, K.W. McLean 25,
W.N. Murray, I.H. Park, T.A. Romanowski 26, R. Seidlein27
Ohio State Universi& Physics Department, Columbus, Ohio, USA 48
D.S. Bailey, A. Byrne ,
28 R.J. Cashmore, A.M. Cooper-Sarkar, R.C.E. Devenish,
N. Harnew, M. Lancaster, L. Lindemann4, J.D. McFall, C. Nath, VA. Noyes, A. Quad&
J.R. Tickner, H. Uijterwaal, R. Walczak, D.S. Waters, F.F. Wilson, T. Yip
Department of Physics, University of Oxford, Oxford, VK47
G. Abbiendi, A. Bertolin, R. Brugnera, R. Carlin, F. Dal Corso, M. De Giorgi, U. Dosselli,
S. Limentani, M. Morandin, M. Posocco, L. Stance, R. Stroili, C. Voci
Dipartimento di Fisica dell’ Vniversita and INFN, Padova, Italy 38

132 ZEUS Collaboration/Physics Letters B 356 (19951129-146
J. Bulmahn, J.M. Butterworth, R.G. Feild, B.Y. Oh, J.J. Whitmore 29
Pennsylvania State University, Dept. of Physics, University Park, PA, USA 49
G. D' Agostini, G. Marini, A. Nigro, E. Tassi
Dipartimento di Fisica, Univ. ‘La Sapienza ’ and INFN, Rome, Italy 38
J.C. Hart, N.A. McCubbin, K. Prytz, T.P. Shah, T.L. Short
Rutherford Appleton Laboratory, Chilton, Didcot, Oxon, UK41
E. Barberis, T. Dubbs, C. Heusch, M. Van Hook, W. Lockman, J.T. Rahn,
H.F.-W. Sadrozinski, A. Seiden, D.C. Williams
University of California, Santa Cruz, CA, lJSA48
J. Biltzinger, R.J. Seifert, 0. Schwarzer, A.H. Walenta, G. Zech
Fachbereich Physik der Universitiit-Gesamthochschule Siegen, Germany35
H. Abramowicz, G. Briskin, S. Dagan 30, A. Levy3i
School of Physics,Tel-Aviv University, Tel Aviv, Israe13’
T. Hasegawa, M. Hazumi, T. Ishii, M. Kuze, S. Mine, Y. Nagasawa, M. Nakao, I. Suzuki,
K. Tokushuku, S. Yamada, Y. Yamazaki
Institute for Nuclear Study, University of Tokyo, Tokyo, Japan 3g
M. Chiba, R. Hamatsu, T. Hirose, K. Homma, S. Kitamura, Y. Nakamitsu, K. Yamauchi
Tokyo Metropolitan University, Dept. of Physics, Tokyo, Japan3’
R. Cirio, M. Costa, M.I. Ferrero, L. Lamberti, S. Maselli, C. Peroni, R. Sacchi, A. Solano,
A. Staiano
Universita di Torino, Dipartimento di Fisica Sperimentale and INFN, Torino, Italy38
M. Dardo
II Faculty of Sciences, Torino University and INFN - Alessandria, Italy 38
D.C. Bailey, D. Bandyopadhyay, F. Benard, M. Brkic, M.B. Crombie, D.M. Gingrich32,
G.F. Hartner, K.K. Joo, G.M. Levman, J.F. Martin, R.S. Or-r, S. Polenz, C.R. Sampson,
R.J. Teuscher
University of Toronto, Dept. of Physics, Toronto, Ont., Canada33
C.D. Catterall, T.W. Jones, PB. Kaziewicz, J.B. Lane, R.L. Saunders, J. Shulman
University College London, Physics and Astronomy Dept., London, UK4’
K. Blankenship, B. Lu, L.W. MO
Virginia Polytechnic Inst. and State University, Physics Dept., Blacksburg, VA, USA4’
W. Bogusz, K. Charchula, J. Ciborowski, J. Gajewski, G. Grzelak, M. Kasprzak,
M. Krzyianowski, K. Muchorowski, R.J. Nowak, J.M. Pawlak, T. Tymieniecka,
A.K. Wroblewski, J.A. Zakrzewski, A.F. Zarnecki
Warsaw University, Institute of Experimental Physics, Warsaw, Poland4=

133
ZEUS Collaboration/ Physics Letters B 3.56 (1995) 129-146
M. Adamus
Institute for Nuclear Studies, Warsaw, Poland42
Y. Eisenberg 30, U. Karshon 30, D. Reve130, D. Zer-Zion
Weizmann Institute, Nuclear Physics Dept., Rehovot, Israe136
I. Ali, W.F. Badgett, B. Behrens, S. Dasu, C. Fordham, C. Foudas, A. Goussiou,
R.J. Loveless, D.D. Reeder, S. Silverstein, W.H. Smith, A. Vaiciulis, M. Wodarczyk
University of Wisconsin, Dept. of Physics, Madison, WI, USA 48
T. Tsurugai
Me@ Gakuin University, Faculty of General Education, Yokohama, Japan
S. Bhadra, M.L. Cardy, C.-P. Fagerstroem, W.R. Frisken, K.M. Furutani, M. Khakzad,
W.B. Schmidke
York University, Dept. of Physics, North York, Ont., Canada33
Received 22 June 1995
Editor: L. Montanet
Abstract
Inclusive jet cross sections for events with a large rapidity gap with respect to the proton direction from the reaction
ep -+ jet + X with quasi-real photons have been measured with the ZEUS detector. The cross sections refer to jets with
transverse energies EF’ > 8 GeV. The data show the characteristics of a diffractive process mediated by pomeron exchange.
Assuming that the events are due to the. exchange of a pomeron with partonic structure, the quark and gluon content of
the pomeron is probed at a scale N ( EFt)‘. A comparison of the measurements with model predictions based on QCD
plus Regge phenomenology requires a contribution of partons with a hard momentum density in the pomeron. A combined
analysis of the jet cross sections and recent ZEUS measurements of the diffractive structure function in deep inelastic
scattering gives the first experimental evidence for the gluon content of the pomeron in diffractive hard scattering processes.
The data indicate that between 30% and 80% of the momentum of the pomeron carried by partons is due to hard gluons.
’ Supported by Worldlab, Lausanne, Switzerland.
’ Also at IROE Florence, Italy.
3 Now at Univ. of Salerno and INFN Napoli, Italy.
4 Supported by EU HCM contract ERB-CHRX-CT93-0376.
’ Now at Miibelhaus Kramm, Essen.
6 Now a self-employed consultant.
’ Now also at University of Torino.
8 Alfred P. Sloan Foundation Fellow.
g Presently at Columbia Univ., supported by DAAD/HSPII-
AUFB.
lo Now at Inst. of Computer Science, Jagellonian Univ., Cracow.
l1 Now at Comma-Soft, Bonn.
l2 Visitor from Florida State University.
I3 Now at Univ. of Mainz.
l4 Supported by DAAD and European Community Program
PRAXIS XXI.
l5 Now at Dr. Seidel Informationssysteme, Frankfurt/M.
I6 Now at Inst. of Accelerating Systems Applications (IASA),
Athens.
l7 Now at Mercer Management Consulting, Munich.
l8 Now with OPAL Collaboration, Faculty of Physics at Univ. of
Freiburg.
lg Now at SAS-Institut GmbH, Heidelberg.
*O Also supported by DESY.
*I Now at GSI Darmstadt.
** Also supported by NSERC.
23 Now at Institute for Cosmic Ray Research, University of Tokyo.
24 Partially supported by CAM.
25 Now at Carleton University, Ottawa, Canada.
26 Now at Department of Energy, Washington.
27 Now at HEP Div., Argonne National Lab., Argonne, IL, USA.
*’ Now at Oxford Magnet Technology, Eynsham, Oxon.
2g On leave and partially supported by DESY 1993-95.
3o Supported by a MINERVA Fellowship.
31 Partially supported by DESY.
32 Now at Centre for Subatomic Research, Univ.of Alberta, Canada

134 ZEUS Collaboration/Physics Letters B 356 (1995) 129-146
1. Introduction
Electron-proton collisions at HERA have shown ev-
idence for hard processes in diffractive reactions. Both
in deep inelastic scattering (DIS) (Q2 > 10 GeV2,
where Q2 is the virtuality of the exchanged photon)
[ l-31 and in photoproduction ( Q2 N 0) [ 4,5], events
characterized by a large rapidity gap towards the pro-
ton direction have been observed and interpreted as re-
sulting from diffractive scattering [ 1,2,4]. In the DIS
regime, hard scattering for this class of events has been
revealed through the virtuality of the probing photon
[ 1,3] and through the observation of jet structure in
the final state [ 21. In the photoproduction domain, the
hard scattering has been identified through jet produc-
tion [ 4,5] .
and TRIUMF, Vancouver, Canada.
33 Supported by the Natural Sciences and Engineering Research
Council of Canada (NSERC) .
34 Supported by the FCAR of Quebec, Canada.
35 Supported by the German Federal Ministry for Research and
Technology (BMW).
36 Supported by the MINERVA Gesellschaft ftir Forschung GmbH,
and by the Israel Academy of Science.
37 Supported by the German Israeli Foundation, and by the Israel
Academy of Science.
38 Supported by the Italian National Institute for Nuclear Physics
(INFN).
3g Supported by the Japanese Ministry of Education, Science and
Culture (the Monbusho) and its grants for Scientific Research.
4o Supported by the Korean Ministry of Education and Korea
Science and Engineering Foundation.
41 Supported by the Netherlands Foundation for Research on Mat-
ter (FGM).
42 Supported by the Polish State Committee for Scientific Research
(grant No. SPB/P3/202/93) and the Foundation for Polish- Ger-
man Collaboration (proj. No. 506/92).
43 Supported by the Polish State Committee for Scientific Research
(grant No. PB 861/2/91 and No. 2 2372 9102, grant No. PB 2
2376 9102 and No. PB 2 0092 9101).
44 Partially supported by the German Federal Ministry for Re-
search and Technology (BMFT).
45 Supported by the German Federal Ministry for Research
and Technology (BMW), the Volkswagen Foundation, and the
Deutsche Forschungsgemeinschaft.
46 Supported by the Spanish Ministry of Education and Science
through funds provided by CICYT.
47 Supported by the Particle Physics and Astronomy Research
Council.
48 Supported by the US Department of Energy.
4g Supported by the US National Science Foundation.
Diffractive processes are generally considered to
proceed through the exchange of a colourless object
with the quantum numbers of the vacuum, generi-
cally called the pomeron (P). Although the descrip-
tion of soft diffractive processes in terms of pomeron
exchange has been a phenomenological success, the
description of the pomeron in terms of a parton struc-
ture at first lacked experimental support. On the basis
of pp data [ 61 from the CERN ISR, Ingelman and
Schlein [7] suggested that the pomeron may have a
partonic structure. The observation of jet production
in pp collisions with a tagged proton (or antiproton)
made by the UA8 Collaboration [ 81 gave strong evi-
dence for such a structure. Further evidence has been
provided by the observations made at HERA [ l-51,
which in addition include the first measurements of
the diffractive structure function in DIS [ 9,101.
The description of diffractive processes in terms of
QCD has remained elusive, in part due to the lack of a
sufficiently large momentum transfer on which to base
the perturbative expansion, Cross sections for diffrac-
tive processes involving large transverse energy jets
or leptons in the final state are, however, amenable to
perturbative QCD calculations [ 7,l l-171. Their mea-
surement could answer several questions concerning
the structure of the pomeron such as: whether the
pat-ton picture is valid for the pomeron and universal
parton densities can be defined; what fraction of the
pomeron momentum is carried by gluons and what by
quarks; and whether a momentum sum rule applies to
the pomeron [ 181.
This paper presents the first measurement of inclu-
sive jet cross sections in photoproduction at centre-of-
mass energies N 200 GeV with a large rapidity gap.
This process is sensitive to both the gluon and quark
content of the pomeron. In order to examine the par-
tonic structure of the pomeron, these jet cross sections
are compared to predictions from models based on
perturbative QCD and Regge phenomenology. The jet
cross sections measured in photoproduction, combined
with the results on the diffractive structure function
in deep inelastic scattering [ lo], give the first experi-
mental evidence for the gluon content of the pomeron.
The result does not depend on the flux of pomerons
from the proton nor on the assumption that a momen-
tum sum rule can be defined for the pomeron. The
data sample used in this analysis corresponds to an
integrated luminosity of 0.55 pb-’ and was collected

ZEUS Collaboration/Physics Letters B 3.56 (1995) 129-146 135
during 1993 with the ZEUS detector at HERA.
2. Experimental setup
2.1. HERA operation
The experiment was performed at the electron-
proton collider HERA using the ZEUS detector. Dur-
ing 1993 HERA operated with electrons of energy
EC = 26.7 GeV colliding with protons of energy EP =
820 GeV. HERA is designed to run with 210 bunches
separated by 96 ns in each of the electron and pro-
ton rings. For the 1993 data-taking, 84 bunches were
filled for each beam and an additional 10 electron and
6 proton bunches were left unpaired for background
studies. The electron and proton beam currents were
typically 10 mA, with instantaneous luminosities of
approximately 6 . 1O29 cme2 s-l.
2.2. The ZEUS detector and trigger conditions
ZEUS is a multipurpose magnetic detector. The con-
figuration for the 1993 running period has been de-
scribed elsewhere [ 2,191. A brief description concen-
trating on those parts of the detector relevant to this
analysis is presented here.
Charged particles are tracked by two concentric
cylindrical drift chambers, the vertex detector (VXD)
and the central tracking detector (CID), operating
in a magnetic field of 1.43 T provided by a thin su-
perconducting coil. The coil is surrounded by a high-
resolution uranium-scintillator calorimeter (CAL)
divided into three parts, forward 5o (FCAL) covering
the pseudorapidity5’ regions2 4.3 2 vd > 1.1, bar-
rel @CAL) covering the central region 1.1 > qj 2
-0.75 and rear (RCAL) covering the backward re-
gion -0.75 2 r]d 2 -3.8. The solid angle coverage
M The ZEUS coordinate system is defined as right-handed with
the Z axis pointing in the proton beam direction, hereafter referred
to as forward, and the X axis horizontal, pointing towards the
centre of HERA.
51 The pseudorapidity is defined as - ln(tan $), where the polar
angle 0 is taken with respect to the proton beam direction, and is
denoted by qj (‘I) when the polar angle is measured with respect
to the nominal interaction point (the reconstructed vertex of the
interaction). i
52 The FCAL has the forward edge at v,dse = 4.3 with full
acceptance for vd < 3.7.
is 99.7% of 47~. The CAL parts are subdivided into
towers which in turn are subdivided longitudinally
into electromagnetic (EMC) and hadronic WAC)
sections. The sections are subdivided into cells, each
viewed by two photomultiplier tubes. The CAL is
compensating, with equal response to hadrons and
electrons. Measurements under test beam conditions
show that the energy resolution is Q/E = 0.18/a
(E in GeV) for electrons and Q/E = 0.35/G
for hadrons [20]. In the analysis presented here,
CAL cells with EMC (HAC) energy below 60 MeV
(110 MeV) are excluded to minimize the effect of
calorimeter noise. This noise is dominated by the ura-
nium activity and has an r.m.s. value below 19 MeV
for EMC cells and below 30 MeV for HAC cells. For
measuring the luminosity as well as for tagging very
small Q2 processes, two lead-scintillator calorime-
ters [ 211, located at 107 m and 35 m downstream
from the interaction point in the electron direction,
detect the bremsstrahlung photons and the scattered
electrons respectively.
Data were collected using a three-level trigger [ 191.
The first-level trigger (FLT) is built as a deadtime-free
pipeline. The FLT for the sample of events analysed
in this paper required a logical OR of different con-
ditions on sums of energy in the CAL cells. The av-
erage FLT acceptance for the events under study was
approximately 90%. The second-level trigger used in-
formation from a subset of detector components to
differentiate physics events from backgrounds con-
sisting mostly of proton beam gas interactions. The
third-level trigger (TLT) used the full event infor-
mation to apply specific physics selections. For this
analysis, the following conditions were required: a)
the event has a vertex reconstructed by the tracking
chambers (VXD+CID) with the Z value in the range
121 < 75 cm; b) E - pz 2 8 GeV, where E is the
total energy as measured by the CAL
E=CEi,
i
pz is the Z-component of the vector
p = C Eiri,
i
the sums run over all CAL cells, Ei is the energy of
the calorimeter cell i and ri is a unit vector along the

136 ZEUS Collaboration/Physics Letters B 3.56 (1995) 129-146
line joining the reconstructed vertex and the geometric
centre of the cell i; c) pz/E 5 0.94 to reject beam-
gas interactions; and d) the total transverse energy as
measured by the CAL, excluding the cells whose polar
angles are below lo”, exceeds 12 GeV.
3. Diffractive hard photoproduction
Diffractive hard photoproduction processes in ep
collisions are characterized by Q2 M 0 and by a final
state consisting of a hadronic system X containing one
or more jets, the scattered electron and the scattered
proton
e+pi--te+X+pf+e+(jet+X,)+pf (1)
where pi (pf) denotes the initial (final) state proton
and X consists of at least one jet plus the remaining
hadronic system (X,.) .
The kinematics of this process are described in
terms of four variables. Two of them describe the
electron-photon vertex and can be taken to be the
virtuality of the exchanged photon (Q’) and the
inelasticity variable y defined by
y_,_+l-~B:
e
where EL denotes the scattered electron energy and 9:
is the electron scattering angle. The other two vari-
ables describe the proton vertex: the fraction of the
momentum of the initial proton carried by the scat-
tered proton (xf), and the square of the momentum
transfer (t) between the initial and final state proton.
In terms of these variables and at low values of Q2
and t, the square of the mass of the hadronic system
X is given by
M$ M (1 -“f)ys (2)
where s is the square of the ep centre-of-mass energy.
Diffractive processes in which the photon disso-
ciates give rise to a large rapidity gap between the
hadronic system X and the scattered proton:
AYGAP = YpI - YIk$ (3)
where yPf is the rapidity of the scattered proton and
Y
z& is the rapidity of the most forward going hadron
belonging to the system X. The same signature is ex-
pected for double dissociation where the scattered pro-
ton is replaced by a low mass baryonic system (N).
In this paper, the outgoing proton (or system N) was
not observed, and instead of y$& the pseudorapidity
(r) of the most forward-going hadron in the de-
tector was used.
Two cross sections are presented in this paper. First,
the cross section for inclusive jet production is mea-
sured as a function of the pseudorapidity of the jet
( viet) (for the definition of the jet variables see Sec-
tion 4) in reaction (1) with the most-forward going
hadron at 72; < 1.8. This corresponds to a rapidity
gap of at least 2.5 units measured from the edge of
the CAL (Avo~p = r]&e - qz). This cross section
is denoted by
(4)
and is measured in the det range between -1 and 1.
Second, the integrated cross section for inclusive jet
production is determined as a function of vzax
and is measured in the range of T&,~ between 1 and
2.4. Both measurements include contributions from
double dissociation where the large-rapidity-gap re-
quirement is satisfied.
The jet cross sections refer to jets at the hadron
level with a cone radius, R = dw, of one
unit in pseudorapidity (7) - azimuth (4) space, and
integrated over the transverse energy of the jet EFt >
8 GeV. They are given in the kinematic region Q2 <
4 GeV2 and 0.2 < y < 0.85. This region corresponds
to photoproduction interactions at centre-of-mass en-
ergies in the range 130-270 GeV with a median Q2 M
10P3 GeV2.
3.1. Models
The description of diffractive hard processes in
terms of QCD is still in an early stage. Two main
theoretical approaches have been considered. Both
assume that a pomeron (P) is emitted by the proton.
The variable xp f 1 - of is then the fraction of the

ZEUS Collaboration/ Physics Letters B 356 (1995) 129-146 137
initial proton’s momentum carried by the pomeron
and M$ M xpys is the square of the yP centre-
of-mass energy. The two approaches differ in the
modelling of jet production in YIP collisions. One of
them [ 7,11- 13 ] assumes factorisation ( factorisable
models) while the other one does not [ 14,161 (non-
factorisable models). The latter are, however, not
considered in what follows due to the lack of a Monte
Carlo generator with an appropriate description of the
event jet structure.
Calculations based on factorisable models involve
three basic ingredients: the flux of pomerons from the
proton as a function of xp and t, the parton densities
in the pomeron and the matrix elements for jet pro-
duction. The pomeron is assumed to be a source of
partons which interact either with the photon (direct
component) or with a partonic constituent of the pho-
ton (resolved component). As an example, the con-
tribution of the direct component to the cross section
for reaction (1) is given by
x .fi/P(P, P2> (6)
where fyi, is the flux of photons from the electron 53
and fpjI, is the flux of pomerons from the proton. The
sum in i runs over all possible types of partons present
in the pomeron, and fi/p( /?, p2) is the density of par-
tons of type i carrying a fraction p of the pomeron
momentum at a scale p2 and is assumed to be inde-
pendent of t. The sum in j and k runs over all possible
A
types Of final State pEU%OIlS and ui+y+j+k is the cross
section for the two-body collision i+y -+ j+k and de-
pends on the square of the centre-of-mass energy ( ji),
the transverse momentum of the two outgoing partons
(fir) and the momentum scale (,u) at which the strong
coupling constant (cz, ( p2) ) is evaluated, One possi-
ble choice is p 2 - A2 In these models, the pomeron
- pT.
flux factor is extracted from hadron-hadron collisions
using Regge theory, and the matrix elements are com-
puted in perturbative QCD. However, the parton den-
sities in the pomeron have to be extracted from exper-
53 The Q2 dependence has been integrated out using the
Weizslcker-Williams approximation.
iment. While the recent measurements of the diffrac-
tive structure function in DIS at HERA [9,10] give
information on the quark densities in the pomeron, the
gluon content has so far not been determined.
Two forms of the pomeron flux factor are commonly
used. The Ingelman-Schlein form (IS) [ 221 uses a
parameterisation of UA4 data [ 231:
frP&w) =
2 (3.19e” + 0.212e3f) (7)
where CO = & GeVF2 and t is in GeV2. The
Donnachie-Landshoff form (DL) [ 121 is calculated
in Regge theory, with parameters determined by fits
to hadron-hadron data:
i
fP/p(XP, t> = SF1 ( t)2X;-2a(t) (8)
using the elastic form factor Fl (t) of the proton, the
pomeron-quark coupling ba N 1.8 GeV-’ and the
pomeron trajectory a(t) = 1.085 + 0.25t with t in
GeV2.
Various parameterisation of the parton densities
in the pomeron have been suggested on theoretical
grounds [ 7,l l-131. The following represent extreme
possibilities for the shape of the quark and gluon
momentum densities:
- hard gluon density pfglp(/?, p2> = 6p( 1 - p);
- soft gluon density /3fg/p(/?,p2) = 6( 1 - /3)5;
- hard quark density (for two flavours) /3fqfp(p, ,x2)
= @(l-p).
The first two assume a pomeron made entirely of glu-
ons and the last one a pomeron made of uii and dd
pairs. In all cases a possible p2 dependence of the par-
ton densities is neglected 54 and the densities are nor-
malised such that all of the pomeron’s momentum is
carried by the partons under consideration, Zp ( ,u2) 3
JtdPCi Pfi/P(P.P') = 1.
However, since the pomeron is not a particle, it
is unclear whether or not the normalisations of the
pomeron flux factor and the momentum sum of the
pomeron can be defined independently. Nevertheless,
for an assumed normalisation of the flux factor, the
momentum sum Xp(p2) can be measured. The defi-
nition used for the DL form of the pomeron flux factor
54 The p2 dependence of the parton densities in the pomeron is
expected to be smaller than the differences between the various
parton densities considered [ 71.

138 ZEUS Collaboration/Physics Letters B 356 (1995) 129-146
is the appropriate one if the pomeron were an ordinary
hadron and, hence, the one in which the momentum
sum rule might be fulfilled [ 241.
Factorisable models presently account only for
diffractive hard processes in which the proton re-
mains intact. Since the measurements are based on
the requirement of a large rapidity gap in the central
detector, the contribution to the measured cross sec-
tions from double dissociation has to be taken into
account when comparing with model predictions.
4. Data selection and jet search
Events from quasi-real photon proton collisions
were selected using the same criteria as reported ear-
lier [ 251. The main steps are briefly discussed here.
Events satisfying the TLT selection described in
Section 2.2 are first selected. A cone algorithm in q-4
space with a cone radius of 1 unit [ 26,271 is then used
to reconstruct jets, for both data and simulated events
(see next section) from the energy deposits in the CAL
cells (Cal jets), and for simulated events also from the
final state hadrons (had jets). The axis of the jet is de-
fined according to the Snowmass convention [ 271, qjet
(@-‘) is the transverse energy weighted mean pseu-
dorapidity (azimuth) of all the objects (CAL cells or
final state hadrons) belonging to that jet. This proce-
dure is explained in detail elsewhere [ 251. The vari-
ables associated with the cal jets are denoted by E$,,,
dzl. and #ii, while the ones for the had jets by EFt,
?1
jet, and eet.
A search for jet structure using the CAL cells is
performed in the data. Events with at least one jet
fulfilling the conditions E$,, > 6 GeV and -1 <
$A < 2 are retained. Beam-gas interactions, cosmic-
ray showers, halo muons and DIS neutral current
events are removed from the sample as described
previously [ 251. The sample thus consists of events
from ep interactions with Q2 < 4 GeV* and a median
Q2 M lop3 GeV*. The yp centre-of-mass energy
(W) is calculated using the expression W = @i.
The event sample is restricted to the kinematic range
0.2 < y < 0.85 using the following procedure. The
method of Jacquet-Blonde1 [28] is used to estimate
y from the energies measured in the CAL cells (see
Section 2.2)
E-PZ
YJB = 2E,’
As can be verified using photoproduction events
with an electron detected in the luminosity monitor
(tagged events), yJB systematically underestimates y
by approximately 20%, which is adequately repro-
duced in the Monte Carlo simulation of the detector.
To allow for this effect, the event selection required
0.16 < yJB < 0.7. The sample thus obtained consists
of 19,485 events containing 24,504 jets. The only sig-
nificant background, which is from misidentified DIS
neutral current interactions with Q2 > 4 GeV’, is es-
timated to be below 2%. The photoproduction origin
of the sample is verified by the expected contribution
(26%) of tagged events.
5. Monte Carlo simulation
Events from diffractive hard photoproduction pro-
cesses were simulated using the program POMPYT
[ 221. These events were used to determine both the
response of the detector to the hadronic final state and
the correction factors for the cross sections for jet pro-
duction with a large rapidity gap.
The POMPYT generator is a Monte Carlo im-
plementation of the model proposed in [7]. The
generator makes use of the program PYTHIA [29]
to simulate electron-pomeron interactions via re-
solved and direct photon processes. In PYTHIA, the
lepton-photon vertex is modelled according to the
Weizsacker-Williams approximation and the effects
of initial state bremsstrahlung from the electron are
simulated by using the next-to-leading order electron
structure function [ 301. Radiative corrections in our
kinematic region, where W is larger than 100 GeV,
are expected to be negligible [ 3 11. For the resolved
processes, the parton densities of the photon were pa-
rameterised according to GS-HO [ 321 and evaluated
at the momentum scale set by the transverse momen-
tum of the two outgoing partons, ,u~ = &. The parton
densities in the pomeron were parameterised accord-
ing to the forms described in Section 3.1 and were
taken to be independent of any scale. In PYTHTA, the
pat-tonic processes are simulated using leading order
matrix elements, with the inclusion of initial and final
state parton showers. Fragmentation into hadrons was
performed using the Lund string model [ 331 as im-

ZEUS Collaboration/Physics L.etfers B 356 (19951 129-146 139
plemented in JETSET [ 341. Samples of events were
generated with different values of the minimum cut-
off for the transverse momentum of the two outgoing
partons, starting at pren = 3 GeV.
The program PYTHIA was also used to simu-
late standard (non-diffractive) hard photoproduction
events via resolved and direct photon processes.
Events were generated using the leading order pre-
dictions of GRV [ 3.51 for the photon parton densities
and MRSD_ [ 361 for the proton parton densities.
All generated events were passed through the ZEUS
detector and trigger simulation programs. They were
reconstructed using the same standard ZEUS off-line
programs as for the data.
6. Event characteristics
The event variable vmax, as in previous studies by
ZEUS [ 1,2,4,10], was used to select events with a
large rapidity gap. For the data, this variable is defined
as the pseudorapidity ($&) of the most forward con-
densate with an energy above 400 MeV. A condensate
is a contiguous energy deposit above 100 MeV for
pure EMC and 200 MeV for HAC or mixed energy de-
posits in CAL. In the samples of simulated events, the
qmaX variable is defined at both the hadron and CAL
levels. At the hadron level, all particles with lifetimes
larger than lo-l3 s, energies in excess of 400 MeV
and pseudorapidities below 4.5 are considered as can-
didates for the most forward final state particle, and
qhad defines the pseudorapidity of the most forward
max
particle. The CAL level uses the same definition as for
the data.
The mass of the hadronic system (Mx) of each
event is reconstructed using the CAL cells, MC,a’ =
dw [ 41. The correlation between MT’ and r&&
for the sample of events with at least one cal jet ful-
filling the conditions Eg& >6GeVand-l<r$z<
1 is displayed in Fig. la. As shown in our previous
publication [ 41, there exists a distinct class of events
with low r]& values. The large-rapidity-gap events
( %%aX
< 1.8) are found to populate the region of low
My’ values, in contrast to the bulk of the data which
have large My’ values. These features of the data are
reproduced by the Monte Carlo simulations: the events
from a simulation of standard hard photoproduction
processes using PYTHIA populate the region of large
v;k and large MC,a’
values; the events from a sim-
ulation of diffractive hard processes using POMPYT
extend into the region of low r&& and My1 values.
A study of the region of low My in the data sam-
ple reveals the following features. The r$& distri-
bution for events with My’ < 30 GeV is shown in
Fig. lb along with the predictions of PYTHIA and
of POMPYT with a pomeron made of hard gluons
(normalised to the number of data events above and
below v,L& = 2.5, respectively). The simulation of
non-diffractive processes by PYTHIA cannot repro-
duce the shape of the measured ~g& distribution. On
the other hand, the predictions of POMPYT describe
well the shape of the data below 72; N 3.
The MC,a’ distribution for the sample of events with
rl
$& < 1.8 is shown in Fig. lc. This sample consists
of 49 events containing 68 jets. The data exhibit an en-
hancement at low masses, 15 GeV 5 Mp 5 30 GeV,
which is reproduced by the simulation of POMPYT
with a hard gluon density (normalised to the number
of data events). The W of each event is reconstructed
using YJB, W”’ = ,/jj$jY. The distribution of WC for
events with 7&$X < 1.8 is shown in Fig. Id along
with the expectations of POMPYT with a hard gluon
density (normalised to the number of data events).
The WCs’ dependence exhibited by the data sample is
well reproduced by the simulation of POMPYT. The
expectations of POMPYT using a hard quark density
(not shown) also give a good description of the distri-
butions of the data. Note that POMPYT assumes the
cross section to be independent of W as expected for
diffractive processes mediated by pomeron exchange.
The good agreement with the data gives evidence for
the diffractive nature of the large-rapidity-gap events.
In summary, the data exhibit a different behaviour
in the region of low masses of the hadronic system
compared to that of high masses. At low masses, the
shape of the r$& distributionin the data sample cannot
be accounted for by the simulation of non-diffractive
processes as in PYTHIA. On the other hand, the fea-
tures of the data are described by the predictions of
diffractive processes mediated by pomeron exchange
as in POMPYT. These facts support the interpretation
of these large-rapidity-gap events as being produced
by diffractive processes via pomeron exchange. There-
fore, the measurements of jet cross sections presented
in the next section are compared to the predictions of

140 ZEUS Collaboration/Physics Letters B 356 (1995) 129-146
ZEU% 1993
F 208
~
s? Cd 2
@ZEUS Bdo i. :,
3, I50
I
100 -
0
10 15 20 25 30 35
lv$~VSeKl
8 ZEUS Dam
_- POMPM
/JII/IIIII//I
150 200 250
PbGeVl
Fig. 1. (a) The scatter plot of My’ versus n,‘& for the sample of events with at least one cal jet fulfilling the conditions I$& > 6 GeV
and -1 < I& < 1; (b) the distribution of vrnax cal for the events with My1 < 30 GeV along with the predictions of PYTHIA (shaded area)
and POMPYT with a hard gluon density in the pomeron (solid line). The predictions are normalised to the number of data events above
and below ~$4~ = 2.5, respectively; (c) the distribution in kf’$” for the events with qm,
cal < 1.8 together with the prediction of PGMPYT
with a hard gluon density in the pomeron (solid line) normalised to the number of data events; (d) the distribution in wCa’ for the events
with q$ < 1.8 and the prediction of POMPYT as in (c).
models based on pomeron exchange. However, with-
out a detected fast proton in the forward direction, the
jet cross sections refer to events with a large rapidity
gap. These include events with a diffractively scattered
proton as well as those with a diffractively dissoci-
ated proton with mass less than approximately 4 GeV
[ 101. In this way, the measurements are presented in a
model independent form suitable for comparison with
calculations other than those presented here.
6.1. Energy and acceptance corrections
The method to correct the transverse energy of a
jet as reconstructed using the CAL cells has been
discussed elsewhere [ 251. For samples of simulated
events, the transverse energy of a jet as measured by
the CAL (E&) was compared to that reconstructed
using the final state hadrons (E$?) . The corrections
to the jet transverse energy were constructed as multi-
plicative factors, C (E$&,, T,$:), which, when applied
to the ET of the cal jets, give the corrected transverse
energies of the jets: EFt = C ( .!$Lalr v’,“,:) x EcLal. The
function C corrects for energy losses, and for values

ZEUS Collaboration/Physics Letters B 356 (1995) 129-146 141
ET& > 10 GeV is approximately flat as a function
of ‘.!&t and varies between 1.08 and 1.18 depending
on r&,. For E$i, near threshold _E@ M 6 GeV this
correction procedure can give vahte?& large as ; .40.
No correction is needed for 7jet (9 M $i>. The
procedure was validated by comparing the momenta
of the tracks in the cal jet in data and in Monte Carlo
simulations. From this comparison it was concluded
that the energy scale of the jets is corrected to within
f 5% [ 251. The correction procedure was applied to
the data sample of jets with Egial > 6 GeV to select
for further study those jets with corrected transverse
energies of EFt > 8 GeV and with the jet pseudora-
pidity in the range between -1 and 1.
The events generated by POMPYT were used to
compute the acceptance correction for the inclusive jet
distributions. This correction function takes into ac-
count the efficiency of the trigger, the selection criteria
and the purity and efficiency of the jet and &$ selec-
tion. It also corrects for the migrations in the variable
$& and yields cross sections for the true rapidity gap
defined by 1722 and 7 = 4.5. After applying the jet
transverse energy corrections, the purity was N 40%
and the efficiency was N 50%. Cross sections were
then obtained by applying bin-by-bin corrections to the
inclusive jet distributions of the data sample in the vari-
ables ,ljet and qmax.
had The acceptance correction factors
for the inclusive jet cross section du/d@(qz$ <
1.8) (4~:; < v&x)) were found to vary between
0.63 and 0.93 (0.59 and 0.84). The dependence of
these correction factors on the choice of parameterisa-
tion of the parton densities in the pomeron were found
to be below N 20%, and are taken into account in the
systematic uncertainty assigned to the measurements
reported in the next section.
7. Results
In this section, first the measured jet cross sections
are presented and the uncertainties of the measure-
ments discussed. These results are model-independent.
Second, the expectations from non-diffractive pro-
cesses are found not to account for the measurements.
Third, the predictions from diffractive models are
compared to the data and estimates of the momentum
sum of the pomeron and of the relative contribution of
ZEUS 1993
::
IO -
-1 -0.5 0 0.5 1
17
id
Fig. 2. Measured differential ep cross section &r/d$“(r&$ <
1.8) for inclusive jet production for .I$ > 8 GeV in the kinematic
region Q2 5 4 GeVZ and 0.2 < y < 0.85 (dots). The measure-
ments are not corrected for the contributions from non-diffractive
processes and double dissociation. The inner error bars represent
the statistical errors of the data, and the total error bars show the
statistical and systematic errors - not associated with the energy
scale of the jets - added in quadrature. The shaded band displays
the uncertainty due to the energy scale of the jets. For compar-
ison, POMPYT predictions for single diffractive jet production
(e+p --f e+p+jet+Xr) using the DL flux factor for direct plus
resolved processes for various parameterisation of the pomeron
parton densities (hard gluon, upper solid line; hard quark, mid-
dle solid line; soft gluon, lower solid line) are also shown. The
GS-HO photon parton densities have been used in POMPYT. The
contribution from non-diffractive processes is exemplified by the
PYTHIA predictions using MRSD_ (GRV-HO) for the proton
(photon) parton densities (dashed line).
quarks and gluons in the pomeron are extracted using
solely the diffractive jet measurements. Fourth, the
jet cross sections in photoproduction are combined
with the measurements of the DIS diffractive struc-
ture function to constrain further the parton content
of the pomeron.
7.1. Jet cross sections
The results for du/dq-@(7$$ < 1.8) and
+j+$: < v:,> are presented in Figs. 2 and 3 [ 371.
The differential cross section is flat as a function
of vjet. Since the measured jet cross sections refer
to events with a large rapidity gap they include a

ZEUS Collaboraiion/Physics Letters B 356 (199.7) 129-146
ZEUS 1993
soft glum
I I I ,
1 1.5 2 2.5
0
?7mra
fig. 3. Measured integrated ep cross section r( ~$2 < $,,) for
inclusive jet production for -1 < 77jet < 1 and I$ > 8 GeV in
the kinematic region Q* 5 4 GeV’ and 0.2 < y < 0.85 (dots).
The measurements are not corrected for the contributions from
non-diffractive processes and double dissociation. The inner error
bars represent the statistical errors of the data, and the total error
bars show the statistical and systematic errors - not associated with
the energy scale of the jets - added in quadrature. The shaded band
displays the uncertainty due to the energy scale of the jets. For
comparison, PYTHIA and POMPYT calculations (for the same
conditions as in Fig. 2) are included.
contribution from double dissociation. The statistical
errors of the measurements are indicated as the in-
ner error bars in Figs. 2 and 3. They are N 30% for
du/d@ ( r$i < 1.8) and constitute the dominant
source of uncertainty. For a( ~2, < v&,) the statisti-
cal error increases from 8% to 20% as qiW decreases.
A detailed study of the systematic uncertainties of
the measurements has been carried out [ 25,381. The
sources of uncertainty include the dependence on the
choice of the parton densities in the pomeron, the sim-
ulation of the trigger, the cuts used to select the data,
and the absolute energy scale of the cal jets [25].
The following systematic uncertainties related to the
qmax-cut were studied: the ~zk variable in the data and
the simulated events was recomputed after removing
the CAL cells with 7 > 3.25 in order to check the de-
pendence on the detailed simulation of the forward re-
gion of the detector, resulting in changes up to 13% for
fl(11E < ?1:ax ) and up to 18% for da/dvjet (72: <
1.8) (except at the most forward data point, where the
statistics are small and the change amounts to 37%);
the energy threshold in the computation of v;& for
data and simulated events was decreased to 300 MeV,
yielding changes up to 11% for da/d@ (r]Fz < 1.8)
and up to 14% for U( 172; < vz,) (except at the most
backward data point, where the statistics are small and
the change amounts to 27%).
The dominant source of systematic error is the ab-
solute energy scale of the cal jets, known to within
5%, which results in a 20% error. The systematic un-
certainties not associated with the energy scale of the
jets were added in quadrature to the statistical errors
and are shown as the total error bars. The additional
uncertainty due to the energy scale of the jets is shown
as a shaded band. The systematic uncertainties have
large bin to bin correlations. They are to be understood
as a conservative estimate of the error associated with
each data point. An additional overall normalisation
uncertainty of 3.3% from the luminosity determina-
tion is not included.
7.2. Comparison to non-diffractive model predictions
The contribution to the measured cross sections
from non-diffractive processes was estimated using
PYTHIA including resolved and direct processes. Had
jets were selected in the generated55 events using
the same jet algorithm as for the data and calculating
&g as explained in Section 6. In PYTHIA the occur-
rence of a rapidity gap is exponentially suppressed and
arises from a fluctuation in the pseudorapidity distribu-
tion of the final state hadrons. The calculations using
MRSD- [ 361 for the proton and GRV-HO [ 351 for
the photon parton densities are compared to the mea-
surements in Figs. 2 and 3. The non-diffractive contri-
bution does not reproduce the measurements. For the
measured dv/d@( &$, < 1.8) the non-diffractive
contribution is close to the data only at the most for-
ward measured point. For the remaining gjet range, the
data are a factor between 2 and 7 above the expecta-
tions from non-diffractive processes. In the measured
&E& < $&, the non-diffractive contribution as
predicted by PYTHIA is smaller than the data by fac-
tors between 3 and 9. These comparisons, together
with the features of the data shown in the previous
section, demonstrate that the measured jet cross sec-
55 These generated events were analysed at the hadron level.

ZEUS Collaboration/Physics L,etters B 356 (1995) 129-146 143
tions with a large rapidity gap cannot be accounted for
by non-diffractive processes. However, in the discus-
sion below, this non-diffractive contribution will be
subtracted from the data.
7.3. Comparison to diffractive model predictions
The measured cross sections are compared to the
predictions of the models for diffractive hard scatter-
ing mediated by pomeron exchange, as implemented
in the POMPYT generator. The predictions have been
obtained by selecting had jets in the generated events
using the same jet algorithm as for the data and cal-
culating vg& as explained in Section 6.
In a first step, the predictions of POMPYT, using
the DL flux factor and the parameterisation of the
pomeron parton densities suggested theoretically (see
Section 3.1)) and assuming 2~ = 1, are compared
to the measured cross sections in Figs. 2 and 3. For
this initial comparison, the contributions from non-
diffractive and double dissociation processes have not
been taken into account. As mentioned earlier (see
Section 3.1) , the ,u2 dependence of the parton densi-
ties has been neglected and, hence, the argument of
Zp( p2) is omitted. The scale relevant for the mea-
sured jet cross sections is p2 N (EFt)2. We start by
discussing the results for da/d$“‘( v:A < 1.8). The
shape of the predictions of POMPYT using a hard par-
ton density compares well with the measured shape of
the cross section. The shape predicted by POMPYT
using a soft gluon density does not describe the data.
The calculations based on a soft gluon density are
smaller than the measurements by factors between 20
and 50. This type of parameterisation was already dis-
favoured by previous studies [ 51. The predictions us-
ing a hard quark density are too small by factors be-
tween 3 and 10, but those using a hard gluon den-
sity reproduce the measurements well. The predictions
based on the IS flux factor lead to similar conclusions.
For the integrated cross section (~(7% < vi,), the
measured shape is in each case described by the ex-
pectations from POMPYT, although the normalisation
is incorrect by a factor depending on the model. A soft
gluon (hard quark) density yields a prediction which
is too small by factors between 30 and 60 (5 and 10).
A hard gluon density for the pomeron gives a good de-
scription of the data. Based on the samples of events
of POMPYT, the measurements are sensitive to p val-
ues above approximately 0.3. Therefore, the data is
not sensitive to a possible additional contribution due
to a soft parton component in the pomeron. The data
do not rule out a possible contribution from a super-
hard pat-ton component in the pomeron [ 15-171.
In principle, other processes could contribute to jet
production with a large rapidity gap. For example, the
proton may emit a m+ (instead of a pomeron), pi +
nfn-+, and a partonic constituent of the & undergoes
a hard interaction with the photon or its constituents.
The contribution from this reaction to the data is ex-
pected to be small due to the power law decrease,
N We4, for pion exchange. Monte Carlo calculations
using POMPYT confirm these expectations.
In a second step, the data were compared to the pre-
dictions of POMPYT based on a pomeron consisting
of both quarks and gluons but without assuming Xp =
1. In addition, the contribution from non-diffractive
processes and from double dissociation to the mea-
sured cross sections were taken into account. The non-
diffractive contribution as predicted by PYTHIA56
was subtracted bin by bin from the data. The contri-
bution from double dissociation for large-rapidity-gap
events was estimated to be ( 15110) % [ lo]. This con-
tribution was assumed to be independent of p and
was also subtracted from the data. After the above sub-
tractions, the data were compared with the predictions
of POMPYT using the DL flux factor and allowing for
a mixture of the hard gluon (6p( 1 - p) ) and the hard
quark ($/3( 1 - ,@) densities in the pomeron: a frac-
tion es for hard gluons and cs = 1 -cg for hard quarks.
The overall normalisation of the POMPYT prediction
was left as a free parameter: 2~. For this study, the
contribution to the cross sections and to Za from pos-
sible soft gluon and soft quark components has been
neglected. For each value of es, a one-parameter (&>
X2-fit to the measured du/drZjet(qr$ < 1.8) was
performed. The results are presented in Fig. 4. The
thick solid line represents the value of Zcp for the min-
imum of the X2-fit for each value of cg and the shaded
band represents the la range around those minima.
For cs = 1 (gluons only) the fit yields Zp = 0.5 f 0.2
with ~2, = 2.3 for three degrees of freedom, while
for cg = 0 (quarks only) the fit yields Zp = 2.5 f 0.9
56 These calculations give a good description of the inclusive jet
differential cross sections (without the large-rapidity-gap require-
ment) in the range -1 < qjet < 1 [25].

144 ZEUS Collaboration/Physics Letters B 356 (1995) 129-146
ZEUS 1993
w’ 4
3.5
3
2.5
2
1.5
0.5
0 1 / I I I
0 0.2 0.4 0.6 0.6 1
Fig. 4. The plane of the variables ZZ, (momentum sum) and cs
(relative contribution of hard gluons in the pomeron). The thick
solid line displays the minimum for each value of cs obtained
from the x2 fit (the shaded area represents the 1 (T band around
these minima) to the measured dtr/d$‘($$& < 1.8) using the
predictions of POMPYT. The constraint imposed in the I;p - cs
plane by the measurement of the diffractive structure function in
DIS (FfC3) ) [lo] for two choices of the number of flavours
(upper dot-dashed line for Xpq = 0.40 and lower dot-dashed line
for Y.Q = 0.32) is also shown. The horizontal dashed line displays
the relation Cp = 1.
with x&n = 2.8. The momentum sum rule (BP = 1)
is approximately satisfied for 0.2 < cg < 0.6 (statis-
tical errors only). Note that for this estimate the DL
form for the pomeron flux factor was assumed.
This comparison of cross sections for jet produc-
tion with a large rapidity gap between data and model
predictions is subject to the following uncertainties:
- The jet cross sections obtained from the Monte
Carlo calculations presented here are leading order
calculations. In these calculations, (u,(~~) and the
parton densities in the proton and the photon are
evaluated at p2 = &. These computations may be
affected by higher order QCD corrections, which are
expected to change mainly the normalisation (K-
factor). The agreement between the PYTHIA cal-
culations of the inclusive jet differential cross sec-
tions and the measurements [25] indicate that in
the case of the non-diffractive contribution the K-
factor is close to 1, within an uncertainty of - 20%.
The K-factor in the case of POMPYT is expected to
be similar (with a similar uncertainty), as the same
hard subprocesses are involved in the calculation of
jet cross sections.
- The amount of the non-diffractive contribution to
the measured cross section was modelled using
PYTHIA with some choices for the proton and
photon parton densities. This contribution is more
sensitive to the choice of photon parton densities.
- The uncertainty in the estimation of the contribution
from double dissociation.
- The POMPYT model for diffractive hard scatter-
ing assumes factorisation of the hard process with
respect to the soft diffractive reaction. The extent
to which this assumption is valid has to be deter-
mined experimentally through a detailed compar-
ison of measurements for different reactions (see
next section).
- The pomeron flux factors adopted in the various
models are based on different assumptions for the t
and xp dependencies which are obtained from data
on soft diffractive hadronic processes. The uncer-
tainty in the procedure used to extract the flux is
about 30%.
The differences between the results obtained in each
of the studies listed above and the central values were
combined in quadrature to yield the theoretical sys-
tematic uncertainties (not shown in Fig. 4) of the fit-
ted values of Zp. These uncertainties were then added
in quadrature with the statistical and systematic un-
certainties of the measurements resulting in the fol-
lowing ranges at the la level: 1.4 < BP < 3.8 for
cg = 0 and 0.3 < 2~ < 0.9 for cg = 1. The range in
Zp assumes the DL convention for the pomeron flux
factor. This normalisation has recently been discussed
by Landshoff [ 391 who concludes that the normalisa-
tion is arbitrary up to a multiplicative factor A. If the
normalisation is changed by a factor A, the range of
the momentum sum is given by 1.4/A < Zp < 3.8/A
for cg = 0 and by 0.3/A < Xp < 0.9/A for cg = 1.
In summary, the comparison of model predictions
with the jet cross section measurements favours those
models where the partonic content of the pomeron has
a hard contribution. Given the uncertainties mentioned
above and the DL convention for the normalisation of
the pomeron flux factor, the data can be reproduced
by a pomeron whose partonic content varies between
a pure hard quark density with momentum sum given
by 1.4 < 2~ < 3.8 and a pure hard gluon density with

ZEUS Collaboration/Physics Letters B 356 (1995) 129-146 145
0.3 < zp < 0.9.
7.4. The gluon content of the pomeron
The HERA experiments have recently presented the
first measurements of the diffractive structure func-
tion in DIS [9,10]. The results show that the quark
densities in the pomeron have a hard and a soft contri-
bution. Assuming the DL form for the pomeron flux
factor, the DIS data do not favour a pomeron structure
function which simultaneously fulfills the momentum
sum rule and consists exclusively of quarks.
If the pomeron parton densities are universal and
describe both DIS and photoproductionprocesses, the
DIS results together with the photoproduction data
further constrain the partonic content of the pomeron.
The measured diffractive structure function in DIS
(Ft’3’(p, Q2,xp)) [9,10] can be used to extract
the contribution of the quarks to the momentum sum
(XpJQ2)). The integral of Ffc3) over XP and p is
proportional to X:pq ( Q2 ) :
Qmax 1
dp F~O(~)(P~ ~2, xd
.m”i” 0
= kf . &,(Q2) . hu (9)
where ZsuX is the integral of the pomeron flux factor
over t and over the same region in xp, and kf is a
number which depends on the number of flavours
assumed (5/18 for two flavours and 2/9 for three
flavours). For the left-hand side of Eq. (9), the
parameterisation of FfC3’(/3, Q2, XP) obtained in
[ lo] was used. The integral was performed over the
range 6.3 . 10e4 < XP < 10m2 of the ZEUS DIS
measurements. The DL form for the pomeron flux
factor was used to compute Ifi,, for the right hand
side of Eq. (9). This procedure yields an estimate of
Xpg( Q2) : 0.32 f 0.05 (0.40 f 0.07) for two (three)
flavours. These estimates are based on a parameteri-
sation of FfC3) (j3, Q2, xp) which was determined in
the large p region (0.1 < p < 0.8) and is assumed
to be valid for the entire region 0 < /3 < 1. In the
range of Q2 where the DIS measurements were done,
8 GeV* < Q2 < 100 GeV’, the pomeron structure
function is approximately independent of Q2 and,
thus, the estimated ZP~(&~> does not depend upon
Q2. It should be noted that the scales at which the
parton densities in the pomeron are probed in DIS,
Q2, and in photoproduction, p2, are comparable. The
estimate from DIS imposes a constraint on the Zp - cg
plane which, combined with the estimates obtained in
the preceding section, restricts the allowed ranges for
Xp and the relative contributions of quarks and glu-
ons (cg) . The DIS constraint, which can be written
as Zp . ( 1 - cs) = 0.32 (0.40) for the two choices of
the number of flavours, is included in Fig. 4 (the dark
shaded area represents the uncertainty in this con-
straint) . Combining the estimates from photoproduc-
tion (thick solid line) and DIS yields 0.5 < &U < 1.1
and 0.35 < cg < 0.7 (statistical errors only).
These results are subject to the uncertainties listed at
the end of Section 7.3. The allowed range for Cp which
results from the combination of the DIS and photopro-
duction measurements was evaluated for each source
of systematic uncertainty. Taking into account all the
uncertainties mentioned, the comparison between the
DIS and photoproduction measurements gives 0.4 <
Bp < 1.6 for the momentum sum of the pomeron
assuming the DL convention for the flux. If the nor-
malisation of the pomeron flux factor is changed by a
multiplicative factor A, the allowed range of the mo-
mentum sum is given by 0.4/A < Xp < 1.6/A.
It should be noted that the evaluation of the ce
range allowed by the DIS and photoproduction mea-
surements is not affected by the normalisation of the
pomeron flux factor or the uncertainty on the double
dissociation contribution since they cancel out in the
Y
comparison . Taking into account the remaining un-
certainties the combination of the DIS and photopro-
duction data gives 0.3 < cg < 0.8. This rest.& does
not depend on the validity of the momentum sum rule
for the pomeron.
8. Summary and conclusions
Measurements of ep cross sections for inclusive jet
photoproduction with a large rapidity gap in ep col-
lisions at J3; = 296 GeV using data collected by the
ZEUS experiment in 1993 have been presented. The
measured jet cross sections are compared to pertur-
bative QCD calculations of diffractive hard processes
57 This cancellation occurs as long as the same pomeron flux
factor is used in both DIS and photoproduction.

146 ZEUS Collaboration/Physics Letters B 356 (1995) 129-146
and allow a model dependent determination of the
parton content of the pomeron. The measurements re-
quire a contribution from a hard momentum density
of the partons in the pomeron. This result is consis-
tent with the observations of the UA8 Collaboration
made in pp collisions. When the measured jet cross
sections are combined with the results on the diffrac-
tive structure function in deep inelastic scattering at
HERA, first experimental evidence for the gluon con-
tent of the pomeron is found. This evidence is inde-
pendent of the normalisation of the flux of pomerons
from the proton and does not rely on assumptions on
the momentum sum of the pomeron. The data indicate
that between 30% and 80% of the momentum of the
pomeron carried by partons is due to hard gluons.
Acknowledgements
We thank the DESY Directorate for their strong
support and encouragement. The remarkable achieve-
ments of the HERA machine group were essential for
the successful completion of this work and are greatly
appreciated. We would like to thank J. Collins, G. In-
gelman and G. Kramer for valuable discussions.
References
[l] ZEUS Collab., M. Derrick et al., Phys. Lett. B 315 (1993)
[21
131
[41
I51
161
[71
[81
[91
t101
[111
r121
[I31
r141
481.
ZEUS Collab., M. Derrick et al., Phys. L&t. B 332 (1994)
228.
HI Collab., T. Ahmed et al., Nucl. Phys. B 429 (1994) 477.
ZEUS Collab., M. Derrick et al., Phys. Lett. B 346 (1995)
399.
Hl Collab., T. Ahmed et al., Nucl. Phys. B 435 (1995) 3.
R608 Collab., A.M. Smith et al., Phys. L&t. B 163 (1985)
267, B 167 (1986) 248;
T. Henkes et al., Phys. I.&t. B 283 (1992) 15.5.
G. Ingelman and P.E. Schlein, Phys. Lett. B 152 (1985) 256.
UA8 Collab., A. Brandt et al., Phys. I&t. B 211 (1988)
239, B 297 (1992) 417.
Hl Collab., T. Ahmed et al., Phys. Lett. B 348 (1995) 681.
ZEUS Collab., M. Derrick et al., DESY 95-093.
E.L. Berger et al., Nucl. Phys. B 286 (1987) 704.
A. Donnachie and P.V. Landshoff, Nucl. Phys. B 303 (1988)
634, Phys. Lett. B 285 ( 1992) 172.
K.H. Streng, Proc. of the HERA Workshop, edited by R.D.
Peccei, Hamburg, Germany (DESY, 1987) p. 365.
N.N. Nikolaev and B.G. Zakharov, Z. Phys. C 53 (1992)
331;
M. Genovese, N.N. Nikolaev and B.G. Zakharov, KFA-
IKP(Th)-1994-37 and CERN-TH.13/95.
r151
[I61
1171
[I81
1191
1201
I211
t221
1231
~241
[251
[261
~71
[281
[291
[301
[311
[321
r331
t341
[351
J.C. Collins et al., Phys. Rev. D 51 (1995) 3182.
J.C. Collins, L. Frankfurt and M. St&man, Phys. Lett. B
307 (1993) 161;
A. Berera and D.E. Soper, Phys. Rev. D 50 (1994) 4328.
B.A. Kniehl, H.-G. Kohrs and G. Kramer, Z. Phys. C 65
(1995) 657.
A. Capella et al., Phys. Lett. B 343 (1995) 403.
The ZEUS Detector, Status Report (1993), DESY 1993.
A. Andresen et al., Nucl. Inst. Meth. A 309 (1991) 101;
A. Caldwell et al., Nucl. Inst. Meth. A 321 (1992) 356;
A. Bernstein et al., Nucl. Inst. Meth. A 336 (1993) 23.
J. Andruszkdw et al., DESY 92-066 (1992).
P. Bmni and G. Ingelman, Proc. of the International
Europhysics Conference, edited by J. Carr and M. Perrotet,
Marseille, France, July 1993 (Ed. Frontieres, Gif-sur-Yvette,
1994) p. 595.
UA4 Collab., M. Bozzo et al., Phys. I&t. B 136 (1984) 217.
A. Donnachie and PV. Landshoff, Proc. of the HERA
Workshop, edited by R.D. Peccei, Hamburg, Germany
(DESY, 1987) p. 351.
ZEUS Collab., M. Derrick et al., Phys. Lett. B 342 (1995)
417.
UAl Collab., G. Amison et al., Phys. I_&. B 123 (1983)
115.
J. Huth et al., Proc. of the 1990 DPF Summer Study on
High Energy Physics, Snowmass, Colorado, edited by E.L.
Berger (World Scientific, Singapore, 1990) p. 134.
Method proposed by E Jacquet and A. Blonde1 in: Proc. of
the Study for an ep Facility for Europe, U. Amaldi et al.,
DESY 79/48 (1979) 377;
see also G. D’Agostini and D. Monaldi, Z. Phys. C 48
(1990) 467;
K.J. Abraham et al., Proc. of ECFA Workshop on LHC,
Aachen (1990) 899;
R. van Woudenberg et al., Proc. of the Workshop on Physics
at HERA, edited by W. Buchmtlller and G. Ingelman,
Hamburg, Germany (DESY, 1991) p. 739.
H.-U. Bengtsson and T. Sjiistrand, Comp. Phys. Comm. 46
(1987) 43 and T. Sjiistrand, CERN-TH.6488/92.
R. Kleiss et al., in: Z Physics at LEP 1, eds. G. Altarelli,
R. Kleiss and C. Verzegnassi, CERN-89-08, Vol. 3, p. 1.
ZEUS Collab., M. Derrick et al., Phys. Lett. B 293 (1992)
465.
L.E. Gordon and J.K. Storrow, Z. Phys. C 56 (1992) 307.
B. Andersson et al., Phys. Rep. 97 (1983) 31.
T. Sjostrand, Comp. Phys. Comm. 39 (1986) 347;
T. Sjostrand and M. Bengtsson, Comp. Phys. Comm. 43
(1987) 367.
M. Gliick, E. Reya and A. Vogt, Phys. Rev. D 46 (1992)
1973.
[36] A.D. Martin, W.J. Stirling and R.G. Roberts, Phys. Rev. D
47 (1993) 867.
[37] Tables of cross sections are available: ZEUS Collab., M.
Derrick et al., DESY 95-115.
[38] K. Desch, Ph.D. Thesis, Physikalisches Institut der
Universitat Bonn ( 1994).
[39] P.V. Landshoff, HEP-PH-9505254, talk given at the
Workshop on Deep Inelastic Scattering and QCD (DIS 95))
Paris, France, 24-28 April 1995 and at the 10th Workshop
on Photon-Photon Collisions (PHOTON ‘95), Sheffield,
England, 8-13 April 1995.