W/Z + jets CSC analyses
Joey Huston, Bruce Mellado, Ulla Blumenschein, Ester Segura, Maria Fiascaris, Alessandro Tricoli, Monika Wielers, Ellie Dobson, Guillaume Kirsch, Sophia Chouridou, Jason Nielson, Ada Farilla, Monica Verducci
Outline
●
Introduction: Motivations and Goals
●
Lepton Trigger and Reconstruction
●
Signal Selection & Backgrounds
●
Systematic Uncertainties: JES and PDF
●
Z+jets cross-section: comparison data-theory
W/Z + jets Motivation: Testing ground of perturbative QCD Measure PDFs Background for other SM and Beyond SM processes
Goal: Measure inclusive and differential W/Z+jets cross sections (jet multiplicity, jet PT..) in ATLAS -> compare with theory -> test performance of generators
Preparation: Optimize signal selection and background suppression in complex multi-jet environment Prepare necessary techniques (data-driven extraction of trigger efficiencies and backgrounds, unfolding to hadron level...) Extract the expected precision of the measurement
W/
Lepton triggering and reconstruction in the presence of jets
Lepton reconstruction in the presence of jets
DR between leptons and between leptons and jets becomes smaller with increasing jet multiplicity -> bias in trigger and reco efficiency ?
Electron trigger & reconstruction: tag-and probe study Trigger efficiency (e25i ) per electron drops with increasing hadronic activity Reco efficiency stable with increasing hadronic activity
Trigger (e25i)
Reconstruction + Medium IsEM
Good agreement between tag-and-probe and Monte Carlo results
Signal Selection & Backgrounds
Jet Multiplicity Z->ee
W->e
Z->
W->
With increasing jet multiplicity, Top replaces QCD multi-jets as most important background
Leading Jet Pt Z->ee
W->e
QCD important at low jet pt, Top at high jet Pt Z+jets: Top distributions different from signal distributions
Z->
W->
Second Leading Jet Pt Z->ee,
Top distributions different from signal distributions: Important to understand Top background with first data
Z + jets : Signal and Backgrounds
Invariant Mass Cut to reduce background: 81 < Mll < 101 GeV
W+ jets : Signal and Backgrounds W->e
Plan to extract QCD from loose selection: - no cut in Missing ET - Z->ee removed by di-electron mass cut Missing ET cut applied after QCD removal
Z->ll + bjets Z-> +jets selection B-tagging applied on jets: b-tagging weight >5 (SV1,IP3D)
w/o b-tagging
with b-tagging
• B-jets fraction dominant after b-tagging applied • C+light jet background: order 30%
Systematic Uncertainties
JES uncertainty on Jet Multiplicity Z->
● ●
W->e
W->
Uncertainty increases with jet multiplicity With an error of 5% (10%) on the JES scale we expect an uncertainty of 7-10% (15-25%) on the measured cross section for 1-3 jets
JES uncertainty on Leading Jet Pt Z->
W->e
• JES error increasing for large jet PT
W->ev + jets: Impact of PDF uncertainties on MC predictions Reweight from CTEQ6LL to the central value of CTEQ6M Reweight to CTEQ6M errors sets -> uncertainty on the jet kinematics
PT leading jet
• PDF uncertainty at the order of 5% • larger at low jet-ET
W->ev + jets: Impact of PDF uncertainties on MC predictions Eta leading jet
• PDF uncertainty at the order of 5% • larger at large jet Eta
Z+jets: Comparison Data-Theory
Z + jets: unfolding of detector effects Correct for electron reconstruction and trigger efficiency Correct for Non-linearity of the jet calibration, jet reconstruction efficiency, jet resolution effects Electron channel
Hadron-level distributions recovered after corrections
Z+jets: correction from parton to hadron level Z-> Z->µµ standard / Z->µµ w/o fragmentation
Z->µµ standard / Z->µµ w/o fragmentation w/o underlying event
-> Corrections decrease with increasing jet PT -> Residual corrections cancel for jet PT > 40 GeV up to percent level
Z + jets: Comparison of Pythia and Alpgen and MCFM Z-> Compare Generators at hadron level with MCFM corrected to hadron level Global normalization of Pythia and Alpgen data to the NLO inclusive cross section Pythia/Alpgen: statistical errors MCFM: PDF, residual fragmentation
• Both Pythia and Alpgen predict lower jet multiplicities than MCFM • MCFM NLO/LO differ by 20-30% • Typical differences MCFM-PythiaAlpgen at the level of 10-60%
Z + jets: comparison Pythia, Alpgen, MCFM Z->
•Pythia Parton Shower jets are softer than Alpgen Matrix Element jets, as expected
Z + jets: Analysis simulation Z->ee
Use fully-simulated Alpgen data to simulate al Analysis steps Compare data to MCFM predictions
Systematic errors from unfolding and background estimate (assume 20% on QCD multi-jet background) Statistic errors: scale data (and errors) to 1fb-1 Assume JES uncertainty of 3% (more later)
Z + jets: uncertainty on the ratio data/theory Add all uncertainties from different sources
• With an error of 5% (10%) on the JES scale we expect an uncertainty of 10-20% (20-30%) on the ratio data/theory -> PDF uncertainties smaller (if JES unc. ≥ 5%) -> uncertainty with 10% larger than difference LO-NLO
• PDF uncertainty larger for low jet ET • Statistical error becomes important for large jet PT (>>200GeV) or large number of jets
References
The contribution on W/Z + jets to the Standard Model CSC chapter has been pared down to 11 pages and concentrates only on Z + jets final states; it can be found at www.pa.msu.edu/~huston/atlas/csc/chapter_draft/june9 A longer (preliminary) version of the note that includes discussion on W + jets and Z + (b) jets can be found at www.pa.msu.edu/~huston/atlas/csc/chapter_draft/APR27_forTL Ellie Dobson has given a recent talk (DIS08) on W/Z + jets in ATLAS that can be found at http://indico.cern.ch/materialDisplay.py?contribId=53&sessionId=27 &materialId=slides&confId=24657 A webpage will be set up containing ATLAS plots useful for presentations