Real and virtual photon structure Leif J

1
Real and virtual photon structureLeif
JönssonUniversity of Lundrepresentingthe H1
and ZEUS collaborations

• Outline of the talk
• Physics processes di-jet events
• Virtual photon structure
• Real photon structure
• Di-jet events with charm production
• Conclusions

2
Physics Processes

• Electron-proton scattering proceeds via the
  exchange of a virtual photon
• Photoproduction Q20 GeV2
• Deep inelastic scattering Q2gtgt0 GeV2
• Pointlike photons (direct) Q2gtkT2
• Resolved photons Q2ltkT2
• Direct processes with kT non-ordered parton
  emissions (CCFM)

CCFM evol.
DGLAP evol.
sep?eX ?dy fg/e(y,Q2) sgp Y electron momentum
fraction taken by the photon fgT/e dominating
fgL/e contributes as y gets small Direct
sgp ?i ?dxp fi/p(xp, mp) sig Resolved sgp
?ij ?dxg fj/p(xg, mg) ?dxp fi/p(xp, mp) sij xg
the fractional photon momentum entering the hard
scattering xp the fractional proton momentum
taken by the interacting parton
3
xgobs ?jetsETe-?/2yEe
A cut at xg around 0.7-0.8 gives good separation
between direct and resolved processes
4
Virtual photon structureTriple differential
cross sections

• Direct processes only describe data in the
  region Q2gt(ET)2
• In the region Q2lt(ET)2 the resolved processes
  become important

5
Virtual photon structureIncluding longitudinal
photon polarisation

• Recently QCD parametrisation of longitudinally
• polarised photons has been implemented in Herwig
• Herwig comes much closer to data

6
Virtual photon structureComparisons with Cascade
(CCFM)

• Cascade provides kT non-ordered parton showers
• Cascade with less degrees of freedom (no photon
  structure) describes data reasonably well

7
Structure of real photons ds/dxgOBS compared to
NLO calculations

• NLO calculations give reasonable description of
  data
• Only slight dependence on photon PDFs

8
Structure of real photonsds/dcos? compared to
NLO calculations
Mjjgt42 GeV
H1 xglt 0.75 xggt
0.75

• cos?tanh(h1-h2)/2?
• NLO calculations give reasonable agreement with
  data

9
Cross section ratio of resolved and direct
processes as a function of Q2

• The cross section ratio decreases with
  increasing Q2 as the contribution from resolved
  processes gets less important
• SaS1D falls below the data

10
Q2 dependence in charm productionRsgobslt0.75)/sg
obsgt0.75) vs Q2

• Data are not able to distinguish between Q2
  suppression or not

• Cascade (CCFM) gives good description of data
• Aroma (DGLAP) falls below

• Extrapolation to the full D phase space
  confirms no Q2 suppression

• Two different scales come into play

11
ds/dxgOBS vs xgOBS in charm productionPrediction
s by Cascade (CCFM)

• In a significant fraction of the events the
  gluon is the hardest parton
• Cascade on hadron level gives reasonable
  agreement with data

12
Conclusions

• Virtual photon structure
• Direct photon processes only describes data in
  the kinematic region Q2gtET2
• The inclusion of resolved photon processes
  provides better agreement in the region Q2ltET2
• Considering also longitudinally polarized photons
  improves the agreement even more
• CASCADE with kT non-ordered parton emissions
  (CCFM) gives similar agreement over the full
  kinematic range do we need resolved photons?
• Real photons
• NLO calculations reproduce data reasonably well
• The dependence on photon PDFs seems small
• The dominant error comes from NLO scale
  uncertainties
• Charm production
• No Q2 suppression observed in contrast to the
  case where no charm requirement is made
• Suppression due to charm and Q2 not independent