high mass drell yan differential cross-section (atlas)
DESCRIPTION
High Mass Drell Yan Differential Cross-section (ATLAS). Paul Bell University of Geneva on behalf of the high mass DY analysis team 18 / 2 / 2013. The measurement. Measurement of DY differential cross section in di -electron channel As function of di -electron invariant mass - PowerPoint PPT PresentationTRANSCRIPT
High Mass Drell Yan Differential Cross-section(ATLAS)
Paul BellUniversity of Geneva
on behalf of the high mass DY analysis team
18/2/2013
The measurement Measurement of DY differential cross section in di-electron channel As function of di-electron invariant mass In “high mass” region = 116 – 1500 GeV Within “extended fiducial” region of electron pT > 25 GeV and | eta | < 2.5
CONF note: https://cds.cern.ch/record/1493623
(paper in approval)
Main backgrounds: W + jets & multijets where jet fakes the electron ID => data driven, 4 methods Other backgrounds from MC
2
“Unfolding” method
Fiducial cross section within extended fiducial region given by
where CDY is the correction factor, given by
where NMC,reco = sum of MC signal events after simulation and selection in measured fiducial region (excludes the “crack” region)
NMC,gen-fid = sum of MC signal events at generator level within extended fiducial region
i.e. CDY includes the small extrapolation over the crack region.
Cross-section measured at different levels of QED FSR correction corresponding to different truth level definitions of NMC,gen-fid:
• Born level: Using electrons before QED FSR, i.e. full QED correction
• dressed level: Where 4-vectors of the bare lepton and all QED FSRphotons within a cone of ∆R < 0.1 are resummed
to perform a partial QED FSR correction3
Results and systematics
4
Statistical uncertainty dominates for mee > 500 GeV
Reaches 50% in highest mee bin
Total systematic uncertainty within 10%
Dominated at low mee by electron efficiency and at high mee by background estimate
Results: generator comparison
MC comparison at dressed level
MC predictions agree well in shape only, after scaling globally
5
Results: FEWZ 3.1 comparison
FEWZ 3.1 comparisonat Born level with recent NNLO PDFs
FEWZ allows (N)NLO QCD corrections to be combined with the NLO electroweak corrections
Choice of EW scheme, Gµ or α(MZ), at most 0.4% effect, after applying NLO electroweak corrections
Contribution from photon-induced (PI) background, evaluated at LO using MRST2004qed
Correction from real W/Z radiation evaluated using MADGRAPH 5, at most 2% at highest mee (not shown) 6