EMID Certification Status

1
EMID Certification Status

• EM selection
• Description
• Data-MC comparison
• Background
• Efficiencies
• EM reconstruction
• Geometrical corrections
• EM-scale corrections
• Electronic calibration
• Energy resolution
• Some preliminary physics studies
• Z mass
• Low energy ee- resonances

Sabine Crépé-Renaudin ISN Grenoble For the EMID
group
2
Documentation

• Certification documentation available on the EMID
  web page
• www-d0.fnal.gov/phys_id/emid/d0_private/
• certification/welcome.html
• Results presented here are based on version 1.2
• Corrections are available in D0 em_util package

3
Selection description
High pT and isolated electrons

• Emreco gt p10.08.00, CellNN and Scone algorithm
• Abs(id)10 abs(id)11
• 0.95 lt EM fraction lt 1.05
• -0.05 lt isolation lt 0.1
• HMx9 lt 100 (depends on purity/efficiency
  needed)
• HMx9 8x8 matrix without CPS energy (Michel
  Jaffré)
• available from p10.11
• Note will be called HMx8 for version gt
  p10.14.00
• Note Possibility to compute this new Hmatrix
  for p10.08 to p10.10 version (thanks to Scott
  Snyder)

4
Data-MC comparison PT gt 20 GeV

• Data EM_HI trigger, EM objects of pTgt 20
• pre-shutdown data
• MC QCD plate MC

Marc Hohlfeld
5
Data-MC comparison e from Z
Data di-EM sample around Z mass MC Zee plate
MC
Sergey Burdin
6
Background studies single EM rate

• Data pre-shutdown data, EM_HI trigger, 0.5
  pb-1,
• pT gt 20 GeV, emfrac gt 0.95
• MC QCD MC

Marc Hohlfeld
Note MC has been rescaled (x2)
7
Fake rate measurement

• Post-shutdown data, Rate vs pT
• Jet triggers, 2nd unbias jet, Standard EM
  selection

Shaohua Fu
Blue HMx9 lt 100 Red HMx9 lt 50
CC
EC
Integrated from pTmin to 200 GeV
EC
CC
Rate0.6/-0.1
8
Efficiencies measurement L1
Sarosh N. Fatakia / Ulrich Heintz

• Post shutdown data (240 k events)
• Compute turn-on curves for single EM trigger wrt
  a trigger with lower threshold (max ET trigger
  tower used)
• Translate the curves to efficiencies for EM
  object (matching ET tower with)
• EM frac gt 0.95, isolation lt 0.1,
• HM9 lt 50
• Needs to remove hot cells carefully

9
Efficiencies measurement L3
Ia Iashvili

• Last runs before shutdown
• Use L3 mark and pass info
• Select good EM objects
• EM frac gt 0.9, isolation lt 0.2,
• HM41 lt 200, h lt 0.8
• EM_LOW rejection 15
• EM_HIGH rejection 5.5
• (rejection comes mostly from differences in
  energy scale between L1 and precision readouts)

10
Efficiencies measurement offline
Leo Chan

• Offline efficiency
• Preselection efficiency (acceptance)
• ID efficiency
• Preselection efficiency measured by MC
  (p10.06.00)
• Indeed good data-MC agreement for eta
  distribution (see Sergey plot below)

11
Efficiencies measurement offline

• ID efficiencies
• Select di-EM objects
• Required one is good quality EM-object, if the
  Mee reconstructed is in Z mass window assume the
  second is good also. Apply ID cut to this second
  object. Compute efficiency after bgd subtraction
• Efficiency 0.99 - 0.53 for standard cuts in CC
• Require SMT track matching for one EM object
  allow to reduce significantly the background
• Efficiency HMx9 lt 1000.940.060.05
• HMx9 lt 50 0.940.060.05
• HMx9 lt 25 0.820.090.05

Leo Chan
Yuri Gershtein
12
Reconstruction

• Aim improve electron reconstruction and compute
  EM-scale
• Geometrical correction (from MC)
• Cuts for phi cracks
• Correct for energy variation with h due mainly to
  different amount of material present in front of
  the calorimeter seen by electron at different
  physical eta
• EM-scale (data / MC comparison)
• Rescale data electron energy to obtain right Z
  mass
• All correction functions available in em_util
  package

13
Geometrical corrections
Sabine Crépé-Renaudin

• Obtained on single e- from MC (p10.08.00)
• Available for both CellNN and Scone algorithm in
  CC
• Fit energy variation with eta and then compute
  this variation as a function of energy (5 to 250
  GeV)

Results
E loss vs eta
before
after
Scone
CNN
14
EM-scale correction
Marie-Claude Cousinou / Alexis Cothenet

• Compare data and MC Mee distributions
• Pre and post-shutdown data
• Standard selection
• Apply geometrical correction for CC
• Etrue E (1 e )
• Fit Mee and find e which maximize a likelihood

Mee91.5 GeV (89 GeV without corrections)

• Results
• e 5.6 0.7 (CC)
• e 0.95 2. (EC zlt-150)
• e 11.2 2.5 (EC zgt 150)

15
Energy resolution
Sabine Crépé-Renaudin Yannick Arnoud

• MC

• Scone
• s (E)/E(0.230.10)/?E ?
• (0.2020.006)/E ?
• (0.0040.002)
• CellNN
• s (E)/E(0.250.10)/?E ?
• (0.2020.006)/E ?
• (0.0030.002)

Data
Fit Zee mass and reproduce it using MC and
varying energy resolution parameters a, b and c.
Not enough statistics to get precise result
16
Electronic calibration
Robert Zitoun

• Takes into account
• Cell by cell gain (ADC/DAC) dispersion (5 to 10)
• ADC to GeV non linearity (300 MeV
  underestimation per cell)
• Improves both Zmass mean and resolution
• -gt EM-scale correction will need to be recomputed

Di-EM mass, after geometrical corrections
Di-EM mass, after geom. corr. elec. calib.
17
Results Z mass

• Z mass from standard EM selection and ET gt 20 GeV
• Pre and post-shutdown data

Marie-Claude Cousinou / Alexis Cothenet
Final plot after all corrections
18
Low mass resonances
Certification results are used for physics !
Oleg Kouznetsov
Sergey Burdin

• J/y

U ???
Very preliminary !!
19
Conclusion

• Stable results for Moriond analysis
• Updates with post-shutdown data
• Results become available for both CellNN and
  Scone clustering algorithms
• Energy corrections will be updated in near future
  for EC and EM-scale recomputed after R. Zitoun
  electronic calibration
• Use of combined EM/tracks information has started