CMS 2013 Public Electron Performance Results

Table of contents:

Cut Based Identification Efficiencies

pdf, png	pdf, png	Top: Electron selection efficiency for the medium working point (WP) on data and on a Drell-Yan Monte Carlo simulation sample as a function of the electron pT, only statistical errors are shown. Bottom: data/simulation scale factor. Both statistical and systematic errors are included. Left: Electrons in 0 <abs(η)<0.8. Right: Electrons in 0.8 <abs(η)<1.4442
pdf, png	pdf, png	Top: Electron selection efficiency for the medium working point (WP) on data and on a Drell-Yan Monte Carlo simulation sample as a function of the electron pT, only statistical errors are shown. Bottom: data/simulation scale factor. Both statistical and systematic errors are included. Left: Electrons in 1.556 <abs(η)<2.0. Right: Electrons in 2.0 <abs(η)<2.5
pdf, png	pdf, png	Top: Electron selection efficiency for the medium WP on data and on a Drell-Yan Monte Carlo simulation sample as a function of the number of reconstructed vertices. Bottom: data/simulation scale factor. Both statistical and systematic errors are included Left: Electrons barrel. Right: Electrons in endcap.

High Energy Electrons (HEEP)

pdf, png	pdf, png	Top: HEEP selection efficiency on data and on a Drell-Yan Monte Carlo simulation as a function of the number of vertices. Only the statistical error is shown. Bottom: data/MC scale factor. Both statistical and systematic errors are included. Left: Barrel. Right: Endcap
pdf, png	pdf, png	Distance in η between the energy-weighted supercluster position in the electromagnetic calorimeter (ECAL) and the track direction at the innermost tracker position. Probe electrons with ET > 100 GeV reconstructed in data and compared with signal and background predictions estimated from simulation except the multi-jet background which is taken from data. Left: Barrel. Right: Endcap.
pdf, png	pdf, png	Distance in φ between the energy-weighted supercluster position in the electromagnetic calorimeter (ECAL) and the track direction at the innermost tracker position. Probe electrons with ET > 100 GeV reconstructed in data and compared with signal and background predictions estimated from simulation except the multi-jet background which is taken from data. Left: Barrel. Right: Endcap.
pdf, png	pdf, png	Ratio between the energy in the hadronic calorimeter (HCAL) computed in a cone ΔR=0.15 behind the electron seed and the supercluster ECAL energy. Probe electrons with ET > 100 GeV reconstructed in data and compared with signal and background predictions estimated from simulation except the multi-jet background which is taken from data. Left: Barrel. Right: Endcap.

Data To Simulation Comparisons for ID Variables

pdf, png	pdf, png	Distance in η between the track direction at the outer tracker position and the electron cluster in the ECAL. Electron candidates with pT > 20 GeV are shown. Left: barrel. Right: endcap.
pdf, png	pdf, png	Ratio between the electron cluster energy and the track momentum at the outer tracker layer. Electron candidates with pT > 20 GeV are shown. Left: barrel. Right: endcap.
pdf, png	pdf, png	Cluster covariance matrix in the η direction. Electron candidates with pT > 20 GeV are shown. Left: barrel. Right: endcap

Multivariate Electron Identification

pdf, png	pdf, png	Distance in η between the track direction at the outer tracker position and the electron cluster in the ECAL. Electron candidates with 7 GeV <pT< 35 GeV are shown. Signal: Drell-Yan Monte Carlo sample. Background: jets faking electrons in a data sample dominated by Z+jets events. Left: barrel. Right: endcap
pdf, png	pdf, png	Ratio between the electron cluster energy and the track momentum at the outer tracker layer. Electron candidates with 7 GeV <pT< 35 GeV are shown. Signal: Drell-Yan Monte Carlo sample. Background: jets faking electrons in a data sample dominated by Z+jets events. Left: barrel. Right: endcap
pdf, png	pdf, png	Cluster covariance matrix in the η direction. Electron candidates with 7 GeV <pT< 35 GeV are shown. Signal: Drell-Yan Monte Carlo sample. Background: jets faking electrons in a data sample dominated by Z+jets events. Left: barrel. Right: endcap
pdf, png	pdf, png	Distance in η between the track direction at the outer tracker position and the electron cluster in the ECAL. Jets faking electrons with 7 GeV < pT < 35 GeV are compared in data samples dominated by W+jets (training sample) and Z+jets (testing sample) events. Left: barrel. Right: endcap
pdf, png	pdf, png	Ratio between the electron cluster energy and the track momentum at the outer tracker layer. Jets faking electrons with 7 GeV < pT < 35 GeV are compared in data samples dominated by W+jets (training sample) and Z+jets (testing sample) events. Left: barrel. Right: endcap
pdf, png	pdf, png	Cluster covariance matrix in the η direction. Electrons with 7 GeV < pT <35 GeV are shown. Jets faking electrons with 7 GeV < pT < 35 GeV are compared in data samples dominated by W+jets (training sample) and Z+jets (testing sample) events. Left: barrel. Right: endcap
pdf, png	pdf, png	Multivariate electron identification output (Boosted Decision Trees). Signal: Drell-Yan Monte Carlo simulation sample. Background: Jets faking electrons are compared in data samples dominated by W+jets (training sample) and Z+jets (testing sample) events. Electrons candidates with 7 GeV <pT< 35 GeV are shown. Left: barrel. Right: endcap.
pdf, png	pdf, png	ROC curves for the electron multivariate identification (Boosted Decision Trees) compared with the cut-based selection working points. Signal from Drell-Yan Monte Carlo simulation sample. Background from jets faking electrons in a data sample dominated by Z+jets. Electron candidates with pT > 20 GeV are shown. Left: Barrel. Right: Endcap

Efficiency for Multivariate Identification

pdf, png	pdf, png	Efficiency on data and on a Drell-Yan Monte Carlo sample for the isolation plus multivariate electron selection as a function of the electron pT. Both statistical and systematic errors are included. Left: barrel. Right: endcap
pdf, png	pdf, png	Efficiency on data and on a Drell-Yan Monte Carlo sample for the isolation plus the multivariate electron selection as a function of the number of vertices. Both statistical and systematic errors are included. Left: barrel. Right: endcap
pdf, png	pdf, png	Data/simulation scale factor for the isolation plus the multivariate electron selection for two separate pseudorapidity regions in the ECAL barrel. Both statistical and systematic errors are included. Left: barrel. Right: endcap

Isolation Pileup Dependency and Correction

pdf, png	pdf, png	Led: average density energy of the event (ρ) and particle-isolation components as a function of the number of reconstructed vertices. The charged particles are associated to the primary vertex. Right: effect of the ρ-based correction on the total particle-isolation. Electrons are selected with pT > 20 GeV and in a data sample dominated by Z->ee events. Electrons in barrel.
pdf, png	pdf, png	Led: average density energy of the event (ρ) and particle-isolation components as a function of the number of reconstructed vertices. The charged particles are associated to the primary vertex. Right: effect of the ρ-based correction on the total particle-isolation. Electrons are selected with pT > 20 GeV and in a data sample dominated by Z->ee events. Electrons in endcap.

Electron Resolution

pdf, png	Expected resolution for prompt and isolated electrons in the ECAL barrel as a function of the initial electron energy from the ECAL, the tracker and the combined estimates. The resolution is evaluated as half the minimal width that contains 68.3% of the reconstructed energy or momentum distribution (effective resolution) and using a Gaussian fit of the core of the momentum distribution.
pdf, png	Instrumental di-electron mass resolution as measured from Z → e+e− events and compared to simulation. Events are categorized according to the electron class and pseudorapidity region of each leg (G1: electron is golden or bigbrem, G2: electron is showering or crack or bad-track, EB: electron is in ECAL barrel, EE: electron is in ECAL endcaps).

Electron Scale

pdf, png	pdf, png	Relative difference between the di-electron mass scales in data and simulation extracted from J/Ψ, Upsilon and Z decays, as function of the average electron pT and for different pseudorapidity regions. Left: 8 TeV data. Right: 7 TeV data
pdf, png		Relative difference between the di-electron mass scale in data and in simulation for Z→e+e- events as a function of the number of the number of vertices.

J/Ψ → ee And Υ → ee

pdf, png	J/Ψ → e+e- reconstructed in the barrel with one electron with 7 <pT< 10 GeV.
pdf, png	Υ (1S) and not resolved Υ (2S+3S) reconstructed in the barrel with one electron with 10 <pT< 15 GeV.