Marcella Capua

1
Inclusive Diffraction at ZEUS

• Marcella Capua
• Calabria University and INFN Cosenza (Italy)

on behalf of

• Diffractive structure function and cross section
  results
• Hep-ex/0501060 (accepted by Nucl. Phys. B)
• Eur. Phys. J. C. 38,43 (2004) and hep-ex/0408009
• Comparison with models
• NLO QCD fits of diffractive data

2
Diffraction at HERA
Standard Deep Inelastic Scattering
x fraction of protons momentum carried by
struck quark ? Q2/W2 W photon-proton centre of
mass energy
DIS probes the partonic structure of the proton
Diffraction exchange of color singlet producing
a rapidity GAP in the particle flow
Diff DIS probes the partonic structure of colour
singlet exchange
3
Inclusive diffraction ?p ? Xp

• diffractive ?p cross section

• diffractive structure function

fraction of the p momentum carried by the IP
fraction of the IP momentum carried by the struck
quark
? No activity in the forward direction ? Proton
intact after the collision
4
Selection of events ?p ? Xp with LPS method
Leading Proton Spectrometer Used till the and of
the HERA I data taking
Diffractive peak

• Direct t measurement
• Free of p-diss background
• Low acceptance ? low statistics

5
Selection of events ?p ? Xp with Mx method

• Forward Plug Calorimeter (FPC)
• CAL acceptance extended by 1 unit in
  pseudorapidity from ?4 to ?5
• higher Mx (a factor 1.7) and lower W
• if MN gt 2.3 GeV deposits EFPCgt 1GeV
• recognized and rejected

• High acceptance
• t measurement not possible
• systematics from p-diss (ep?eXN)

D, c, b from a fit to data
- exponentially falling for decreasing Mx for
non-diffractive events - flat vs ln Mx2for
diffractive events
6
Kinematic domains
Large overlapping but also complementary regions
Mx method allow higher ß bins and lower xIP , LPS
method allow higher xIP
97 data 0.03ltQ2lt0.60 GeV2 (3.6pb-1) 2ltQ2lt100GeV
2 (12.8pb-1) 25ltWlt280GeV 1.5ltMxlt70GeV
tlt1GeV2
98-99 data (4.2pb-1) 2.2ltQ2lt80GeV2 37ltWlt245GeV 0
.28ltMxlt35GeV hep-ex/0501060 (accepted by Nucl.
Phys. B)
Eur. Phys. J. C38, 43 (2004) and hep-ex/0408009
7
t dependence of cross section (LPS method)
Exponential fit to t distribution

• agrees/improves previous results
• no Q2 dependences

Regge phenomenology predicts shrinkage of the
diffractive peak (b rise as xIP ? 0)
8
xIP dependence of F2D(3) (LPS method)
97 LPS sample Common xIP dependence in all bins
consistent with the assumption of Regge
factorization
with
by fitting xIP dep. at fixed ß,Q2
9
Comparison between methods
Good agreement between LPS and MX method (x0.7
for MNlt 2.3GeV)
xIP dep. of F2D(3) equivalent to W dep. of ds/dMx
10
Cross section W dependence (MX method)
98-99 FPC sample (Higher ß region)

• p-dissociation events with MNlt2.3 GeV included
  (30)

• MXgt 2 GeV
• d?/dMX rises rapidly with W

Used for a power-like fit
11
aIP(0) from diffractive and total ?p scattering
(Mx method)
fit to diffractive cross section data
from LPS data

• ?IPdiff(0) higher than soft Pomeron

• Evidence of a rise of ?IPdiff with Q2 ? mild
  Regge factorisation violation

12
sdiff/ stot W and Q2 dependence (Mx method)

• In DIS ?diff/?tot independent of W

• low MX strong decrease of diff/?tot with
  increasing Q2
• high MX no Q2 dependence

Same results from ZEUS-LPS
highest W bin Diffraction is a substantial
part of the total cross section
?diff(MXlt35 GeV)/?tot 16 Q2 4 GeV2 ? 10
Q2 27 GeV2
13
The colour dipole picture
Virtual photon fluctuates to
states (colour dipoles)
Provide a good description of the DIS diffractive
data BEKW (Bartels, Ellis, Kowalski and
Wüsthoff) BGK (Bartels, Golec-Biernat,
Kowalski) FS04(Forshaw Shaw) CGC(Colour Glass
Condensate)
14
BEKW model

• For b lt 0.7 and xb xIP lt 0.002
• xIP2FD(3) increases with increasing Q2
• Positive scaling violations
• For fixed b, Q2 dependence of
• xIPF2D(3) changes with xIP

• Transition to a constant
• cross section as Q2?0
• (similar to total cross section )

• Main features of the data described by BEKW
  parametrization (xIPlt0.01)

(Bartels, Ellis, Kowalski and Wüsthoff)
energy dep. for both (from fits to the data)
FTqq ß(1- ß), weak Q2 dep. FTqqg (1- ß)? , ln
(1Q2/Q20), Q201GeV2 FLqq only at high ß
medium ß small ß
qqg fluctuations dominant at low Q2
15
F2D(3) Q2 dependence and BGK model (LPS method)
xIP lt 0.01 (diff peak region) Will show QCD fits
for this region only
xIP gt 0.01 (exchange of subleading trajectories
important)

• Data well described by BGK (Bartels,
  Golec-Biernat, Kowalski) saturation model
  (xIPlt0.01)
• Positive scaling violation at all values of xIP
  lots of gluons!

16
Comparison with Colour Dipole Model - I

• Forshaw Shaw (FS04) model ? Refer to
  hep-ph/0411337 Regge dipole model with/without
  saturation
• Iancu, Itakura, Munier Colour Glass Condensate
  (CGC) model ? Refer to hep/0310338

• ZEUS LPS97

FS04(nosat)
Q22.4 GeV2
F2
Fit F2 and then predict xIPF2D(3)
CGC
FS04(sat)
F2
Q239 GeV2
Q239 GeV2
xIP
b0.007
b0.03
b0.13
b0.48
x
(Heuijin Lim DIS05)
17
Comparison with Colour Dipole Model - II
Low Q2 from ZEUS MX 98-99
(x 1./0.7)
MX30 GeV
MX20 GeV
MX11 GeV
MX6 GeV
MX3 GeV
MX1.2 GeV
xIP
? Predictions of model are corrected by 1/0.7 for
the MNlt2.3 GeV of ZEUS MX method.
(Heuijin Lim DIS05)
18
Comparison with Colour Dipole Model - III
High Q2 from ZEUS MX 98-99
(x 1./0.7)
MX30 GeV
MX20 GeV
MX11 GeV
MX6 GeV
MX3 GeV
MX1.2 GeV
xIP

• CGC and FS04(sat) are able simultaneously to
  describe F2 and xIPF2D(3).
• Forshaw Shaw have not been able to find a good
  fit which does not invoke saturation.

(Heuijin Lim DIS05)
19
(Diffractive) hard scattering factorisation
Diffractive DIS, like inclusive DIS, is
factorisable into a hard part and a soft part
QCD Hard Scattering factorizationTrentadue,
Veneziano Berera, Soper Collins
pq/p(xIP,t,x,Q2) probability to find, with probe
of resolution Q2, in a proton, parton q with
momentum fraction x, under the condition that
proton remains intact and emerges with small
energy loss, xIP, and momentum transfer t

Additional assumption REGGE FACTORISATION
Regge motivated IP flux
Shape of diffractive PDFs, independent on xIP and
t
universal hard scattering cross section (same as
in inclusive DIS)
Diffractive Deep Inelastic Scattering probes the
diffractive PDFs of the proton relevant when the
vacuum quantum numbers are exchanged
20
NLO QCD fit on LPScharm data

• NLO DGLAP fit, QCDNUM
• MXgt2 GeV, xIPlt0.01, Q2gt2 GeV2
• Regge factorisation assumption
• DL flux
• F2D(3) (xIP,ß,Q2)FIP(xIP) F2IP(ß,Q2)
• initial scale Q022 GeV2
• Quark flavour singlet and gluon
• parameterized according to
• zf(z)(a1a2za3z2)(1-z)a4

QCD fit describes data fractional gluon momentum
is
Similar to H1 75 15 (exptheor) Higher than Mx
method data 55 (see A. Levy DIS05)
F2D (3)cc from DESY-03-094
21
Summary

• Recent data from ZEUS with improved precision and
  extended kinematic range
• Slope of ds/dt is measured it is indipendent
  of Q2 but decreases with xIP
• Indication that aIP increases with Q2
• Similar W and Q2 dep. of diffractive and total
  cross section
• Diffraction shows evidence for pQCD evolution
  with Q2
• Data described by dipole models (BEKW, GBW,
  FS04, CGC)
• Diffractive PDFs extracted from DGLAP fits
  dominated by gluons
• LPScharm data suggest momentum fraction carried
  by gluons 82 (consistent with H1 results)

22
Reserve
23
(No Transcript)
24
(No Transcript)
25
F2D(3) /F2 Q2 dependence (LPS method)
The ratio is largely Q2-indipendent
26
Comparison with H1

• H1 Large rapidity gap selection
• MYlt1.6 GeV and tlt1 GeV2
• LPS proton selection MY mp
• extrapolated to tlt1 GeV2
• constrain IR contribution at high xIP
• H1 LRG / LPS ratio
• p-dissociation contribution 10
• ?Good agreement between two methods and two
  experiments
• ZEUS Mx / LPS ratio
• ?p-dissociation contribution 30

27
F2D(3) ß dependence

• Different ß dep. at
• low and high xIP

• Data well described by
• BGK saturation model
• (xIPlt0.01)

28
NLO QCD fits to the ZEUS data

• Donnachie and Landshoff pomeron flux
• QCDNUM (Sasha) CTEQ
  (Tel Aviv)
• udsubardbarsbar (Sasha) ssbar0 (Tel
  Aviv)
• Fits to F2D (Sasha)
  reduced cross section, ie including FL(Tel A.)
• Only statistical errors (Sasha) statsyst
  in quadrature (Tel Aviv)
• Initial scale2 GeV2 (Sasha) 3 GeV2
  (Tel Aviv)
• None of these differences significant. Given same
  input, same fits are obtained.
• PARAMETRISATIONS of PDFs at initial scale
• A) zf(z) AzB (1-z)C
• B) zf(z) H1 parametrisationSCjPj(2z-1)2expa/
  (z-1) ,
• PjChebychev polynomials
  a0.01 3 terms kept in sum
• C) zf(z) (a1a2za3z2) (1-z)a4
• D) zf(z) (a1a2za3z2) (1-z)

29
LPS vs H1 PDFs
Results similar to H1s
S. Proskuryakov