CMS Ecal Laser Monitoring System

1
Wjets and Zjets studies at CMS
Analysis Overview
Candles and Ladders

• Data-driven strategy to study properties of W/Z
  jets production in final states with electrons
  and muons
• Focus on LHC startup order
• Using different jet definitions - in some cases,
  detector-wise orthogonal

Two inter-related analyses
Zjets candle analysis

• Test of Berends-Giele (BG) scaling through the
  measurement of the Zn jets / Z(n1) jets ratio
  as a function of n
• High efficiency signal selection provides
  candle dataset for
• detector and physics object commissioning
  studies
• normalization of irreducible Z(??)jets
  background in METjets new physics searches
• New physics searches with SM Z bosons in the
  final state, observed as a deviation from BG
  scaling

W/Zjets _at_ LHC Overview

• W/Zjets have a large cross section at LHC with
  final states including
• leptons
• jets
• missing transverse energy (neutrinos)
• Dominant background for SM measurements (ex.
  and Higgs production) and new physics searches
  involving jets/leptons/MET final states

W/Zjets ratio analysis

• Test of BG scaling through the measurement of the
  Wn jets / W(n1) jets ratio as a function of n,
  along with the double ratio
• Predict W gt 3,4 jet yields from lower jet
  multiplicities, revealing excesses from new
  physics processes with leptons, jets and MET

• High-efficiency event selection to increase S/B
  ratio to suitable level for ML fit to extract
  yields
• Z candle only minimal isolation and vertex
  requirements necessary due to discriminating
  power of di-lepton invariant mass
• W/Zjets ratio synchronized event selection for
  W and Z events gt maximal cancellation of
  systematic uncertainties in the double ratio.

Analysis Strategy

• Common selection requirements
• Single non-isolated HLT lepton trigger
• Electron/muon reconstruction
• Lepton identification
• Lepton isolation
• Lepton - primary vertex compatibility
• Jet clustering
• Electron(s) from W(Z) cleaning from jet
  collections
• Jet counting

Jet Clustering

• In these analyses, each yield measurement is done
  as a function of inclusive jet multiplicity
• We consider several types of jet with different
  experimental constituents (all clustered with the
  SISCone algorithm and a cone size of 0.5)

• calo-jets jets clustered from the calorimeter
  (ECALHCAL) cells re-projected w.r.t. the primary
  vertex (standard)
• track-jets jets clustered from tracks consistent
  with the event primary vertex (lower noise levels
  relative to calorimeters)
• JES corrected calo-jets synchronized with the
  above calo-jets definition (mature
  detector understanding)
• Particle Flow jets synchronized with the above
  calo-jets definition
• These types of jets
• have orthogonal detector systematics calorimeter
  vs tracker
• probe different regions of phase space different
  cuts in pT, 3.0 vs 2.4 in ?

• W specific requirements
• gt 1 lepton
• Z mass veto
• extra muon veto (e)
• MET gt 15 GeV (QCD rejection)

• Z specific requirements
• gt 2 leptons
• Z mass window

Event Reconstruction and Cut-Based Wjets, Zjets
selection
orthogonal selection
pT gt 30 GeV/c, h lt 3.0
pT gt 15 GeV/c, h lt 2.4

• Rather than a cutting-and-counting, we use
  maximum likelihood fits to extract event yields,
  as a function of the number of jets
• Additional orthogonal discriminating variable
  used to validate PDF parameterizations
• Control samples from data are used to determine
  distribution shapes and efficiencies required by
  the fits, minimizing the reliance on Monte Carlo
  simulation

pT gt 60 GeV/c, h lt 3.0
Maximum Likelihood Fits
pT gt 60 GeV/c, h lt 3.0
Data control samples
Maximum Likelihood Fits

• Maximal likelihood fits are performed for each
  jet multiplicity in order to extract signal and
  background yields.

• For Zjets events, the fit is based on the
  di-lepton invariant mass

Z candle dataset

• For Wjets events, the fit is based on the W
  transverse mass

• In the W fit, two different categories are
  defined a heavy flavor (hf) enriched (
  ) and depleted region ( ),
  dominated by signal and single top ,
  respectively.

• The yields from the maximum likelihood fits are
  used to calculate the ratios related to BG
  scaling and between the W/Zjets yields - most
  systematic uncertainties (PDFs, jet energy
  scale, lepton isolation, etc.) cancel in these
  ratios
• Additionally, sPlots statistical background
  subtraction, in conjunction with the maximal
  likelihood fit, can be used to produce a pure,
  high efficiency, Z(ll)jets sample - which can be
  used for a number of applications related to
  detector and physics object commissioning and
  background normalization

• The hf enriched region is defined by a cut on
  variables related to the impact parameters of
  tracks matched to jets in the event, Dxyevt and
  Dszevt.

BG scaling in W/Zjets events
Background control samples for m(ll) and mTW
shapes
W/Zjets ratio
BG scaling for W/Zjets

• Low jet multiplicities can be used to predict the
  high multiplicity yields.
• Comparison with measurements of the higher
  multiplicities can quantify deviations from BG
  scaling due to new physics in Z and
  METjetslepton final states

hf variable control samples for signal and
background
Deviations from BG scaling

• Zjets candle sample
• High signal efficiency with sPlots statistical
  background subtraction

MET correction for W(??)jets events deriived
from Z(??)jets events
Z(??)jets background normalization
Christopher S. Rogan, California Institute of
Technology - HCP2009 - Evian-les-Bains