Parton density functions and jet crosssections

1
Parton density functions and jet cross-sections

• Christmas meeting progress update, 12/12/05

Oliver Harris, UCL/ATLAS
2
Introduction

• Why study PDFs?
• allows predictions to be made for hadron-collider
  processes
• Use data from finished/existing experiments as
  input
• HERA experiments are ideal PDF probes
• Knowledge of PDFs important for predicting both
  signal and background event rates

3
Example of a proton PDF
xf(x) expectation value of partons momentum
fraction x (for parton of flavour f)
behaviour of PDF depends strongly on energy
scale, Q
contribution from u valence quarks
contribution from d valence quarks
contributions from sea quarks and gluons (scaled
down)
partons momentum/protons momentum
4
My work

• Use a next-to-leading-order program to generate
  dijet and trijet events at LHC energies
• LO example process (2-gt2) NLO example
  process (2-gt3)
• Use PDFs based on data taken at ZEUS experiment
  as input to program
• Analyse contributions to x-section uncertainty
  from various sources

5
Example results

• Began by generating my own PDF sets based on ZEUS
  data, as published sets didnt go to high enough
  Q2

Q2 10 GeV2
Q2 10,000 GeV2
xf(x)
xg/20
xuv
xdv
x
x
6
Example results

• New format for PDF data LHAPDF convention
  included ZEUS-2005
• includes jet data which helps reduce uncertainty
  on gluon w.r.t older ZEUS fits
• is available with Q2 values of up to 2 x 108 GeV2
• comes in 23 member sets a central value and
  up/down excursions along 11 eigenvectors of an
  uncertainty matrix

total fractional error on ds/dpT
error contribution due to uncertainty on gluon PDF
7
Now working on ds/dpT spectra in different ?
regions sensitive to different regimes of x and
Q2
8
Verification of results

• Two others working on similar topics as me
  Claire Gwenlan, using JETRAD, and Mark Sutton,
  using a C wrapped version of the FORTRAN code
  Im using

9
Further plans

• Investigate ? dependence of parton uncertainties
  by flavour and x-value
• Make statistical comparison of inclusive jet
  cross-section to key signal processes at the LHC
• Begin Atlfast analysis using alternative event
  generator, most likely MC_at_NLO