Soft Electron Tagging

1
Soft Electron Tagging

• John Paul Chou
• DOE Review 2008
• Thursday, August 14, 2008

2
Overview

• Soft Electron b-tagging (SLTe)
• Algorithm
• Simulation
• Applying the Tagger
• Top Pair Cross Section
• WCharm Cross Section
• Top Charge

3
Soft Electron b-Tagging

• Look for soft electrons from heavy flavor decay
• BF(b?e?X) 10
• BF(b?c?e?X) 10
• 35 of top events have a soft electron from HF
• High PT Electrons
• Charged track
• Associated with electromagnetic shower
• Little to no hadronic energy
• No other tracks nearby (i.e. isolated)
• Question How do we ID electrons when the
  electron is embedded in a jet?
• Specifications
• b-jets from top decay are high PT and dense, but
  the tagger must work over a couple decades in
  scale
• To be useful, we need 1001 ep/k/p separation

4
Soft Electron Algorithm

• Begin by using calorimeter quantities
• E/P Electromagnetic energy on par with track P
• Had/Em Cluster dominated by EM component
• Not finely segmented very sensitive to local
  environment
• Main engine of the tagger CES
• Wire and strip chambers located at approximately
  shower maximum within the EM calorimeter
• Measures transverse EM shower profile in two
  orthogonal directions
• Advantage less environmental dependence
• Finely segmented 2-3 mm resolution
• Measure position and shape of shower (not
  amplitude)
• Disadvantage not well modeled in MC
• Combine CES elements into a likelihood and cut

5
Conversions

• Conversion electrons are also a significant
  background
• Conversions dominate signal at low PT
• 3 times as many candidate tracks in top are from
  conversion photons than from HF
• Two techniques for removal
• Locate partner electron track geometrically
• But partner track may not be reconstructed
• Low PT threshold for candidate electrons
• Asymmetric energy sharing
• Use material interaction behavior to identify
• conversions
• Extrapolate track through silicon detector
• Use double-sided silicon layers
• Reject tracks with more than three layers
  expecting
• hits on each side but having none
• 70 efficiency for low PT conversion electrons in
  jets
• 7 over-efficiency to misidentify prompts as
  conversions

6
Simulation
B enriched dijets

• Calorimeter variables are well-modeled,
• but CES variables are not
• Parameterize data and apply it to MC
• Consider the effects of
• Kinematics (PT)
• geometry (?)
• and local environment (track isolation)
• Cross check as many places as we can
• (Zs, b-jets, etc.)
• Tag Matrix predicts the tagging rate of
• electrons
• Use conversion electrons as a template
• Correct tag matrix for jet environment since
• most conversion electrons are not in jets
• Fake Matrix predicts the tagging rate of
  non-electrons (a.k.a. fakes)
• Use tracks from generic jets (20/50/70/100) as a
  template

7
Top Production Cross Section

• The cross section is a cross check of the tagger
  itself
• We extrapolated the SLTe tagger into a High PT,
  dense environment
• Second order effects could come into play should
  check that it works
• LeptonJets event selection
• 1 isolated, high PT lepton (electron or muon)
  with PT/ET gt 20 GeV
• 3 jets (Corrected ET gt 20 GeV, ? lt 2.0)
• Missing ET gt 30 GeV
• Scalar sum of transverse energy, HT gt 250 GeV
• 1 soft electron tag
• Missing ET, HT, and SLTe likelihood cut optimized
  for total uncertainty

Background method similar to secvtx xs
Luminosity 1.7 fb-1
AcceptanceEfficiency
8
Cross Section
First measurement of cross section with soft
electron tags in run II Moving towards PRD
9
Kinematics
10
WCharm

• Measure WCharm production cross section
• Important background in W1,2 jet sample (Higgs,
  etc.)
• Charges of W lepton and soft lepton are
  anti-correlated
• Count Opposite Sign (OS) minus Same Sign (SS)
  events
• Use 1, 2 jet bin of top
• cross section measurement
• as control region
• Plan precision measurement combining
• soft electron and soft muon channels
• with 3.0 fb-1

11
Top Charge (I)

• Theoretically, exotic top model with mass near
  175 GeV/c2 could have -4/3 charge
• Decays into W- and b, instead of W and b
• PRD59(091503)
• Measurement techniques
• Measure associated photon production
• cross section
• Not practical at Tevatron, but possible
• at LHC
• Use jet charge algorithm to determine
• b-jet charge
• Kinematic fitter determines which jets
• are leptonic b and hadronic b
• Assign top charge based on b-jet charge and
  lepton charge
• high efficiency, low purity
• Exotic quark model excluded at 87 confidence
  level
• Use soft lepton tagging instead of jet charge
• Want to maximize eD2 (Dilution, D2P-1)
• Lower efficiency, higher purity competitive,
  orthogonal measurement

12
Top Charge (II)

• Increase purity by selecting high PT
• soft lepton tags
• Optimizing on separate channels
• triples measurement significance
• SecVtx SLT (different jets)
• Use SLT tag in kinematic fitter to
• enhance fitter purity
• SecVtx SLT (same jet)
• Low purity high efficiency channel
• 2 SecVtx SLT (in same a SecVtx jet)
• Best eD2
• Electrons are better than muons!
• Can tune S/B with different operating points
• Less efficiency but higher fake rejection at high
  PT

80 purity at PTgt8 GeV/c
13
Summary

• We have implemented a soft electron tagger at CDF
• Tagger used to measure top cross section
• Godparent committee has been chosen
• Wcharm measurement is in the works
• Top charge measurement with soft lepton tags
  shows promise

14
Backup Slides
15
Figures of Merit
Per track tagging efficiency for conversion
electrons and tracks in generic jets

• Environment overstates HF electron tagging
  efficiency
• In top events, tagging efficiency for HF
  electrons 40 per track (L1)
• Electron contamination overstates non-electron
  tagging rate
• In top events, per track non-electron tag rate
  0.5 per track (L1)

16
Schematic
Data
Tag/Fake Matrix Depend on choice of Likelihood cut
MC
Both