Halina Bilokon, Vitaliano Chiarella, Simonetta Gentile, Giovanni Nicoletti

1
Data based method for background subtraction Z
?µµ-
Halina Bilokon, Vitaliano Chiarella, Simonetta
Gentile, Giovanni Nicoletti Laboratori Nazionali
di Frascati, Università di Roma La Sapienza,
INFN
2
Outline

• Motivation.
• Example of application bb Z ? m m-
• Method proposed
• Software Generators
• Simulation, Reconstruction, Analysis.
• Sample
• Conclusion and future development

3
MonteCarlo
Isajet
Ariadne
Sherpa
Cojest

• Les Houches Guidebook hep-ph/040345

Jetset
Pythia
Herwig
4
Shower MonteCarlo
5
Motivation

• It is then extremely important to rely
• on alternative method to MonteCarlo, as much we
  can
• We propose a method based on a different
• data sample of experimental data
• Only hypothesis lepton universality.
• Example of application bb Z ? m m-
• (most copious background of bb h0/A ? m m-
  studies (14/12/2005) S.G. presentation,
  ATLAS-PHYS-2003-015, June 2003.)

6
Signal background

• 0.2 - 0.01 pb
• (even lower)

s 22.8 pb

• h0 bb? m- mbb

• Zbb ? m-m bb

(Pythia 6.226)AcerMC (v.2.3) interfaced with
Pythia 6.2 (hep/ph0405247).
7
Background Subtraction Method
Precise Knowledge of background is crucial

• Experimental method
• proposed based on
• Z? m m- and Z? e e-
• Relying on experimental data
• Br( h0? e e-) neglegible
• Different Inner Bremsstrahlung

8
Z production
9
Software Release Sample

• Generation Athena 9.0.4 - Pythia (v.6.226) - Acer
  (v. 2.3)
• Simulation, Digitalization, Reconstruction Athena
  10.0.1
• Analysis based on CBNT (root ntuples).
• Final study in scenario
• 600000 events
• In this presentation Lint 20 fb-1
• 400000 events

10
Study performed

• Different Inner Bremsstrahlung
• Zbb ? m-m bb (g) Ng 0.086 Nev464522
• Zbb ? e-e bb (g) Ng 0.157 Nev442828
• Different Detector Response
• The most significative variable of
• bb h0/A ? m m- analysis has be chosen Minv
  as test .
• NB. The combination of a same flavor lepton
  originating from b and one of Z are included
  inevitably in the plots

11
Bremsstrahlung Photons
Zbb ? m-m bb
Zbb ? e-e bb
Ng MC 0.086
Ng MC 0.157
e
12
Bremsstrahlung Photons
Zbb ? m-m bb
Zbb ? e-e bb
13
Generator level Zbb ? m-m bb
Minv(mm-)g
Minv(mm-)
14
Generator level Zbb ? m-m bb
Minv(mm-)
Minv(mm-)g
Only events with a g of Bremsstrahlung
15
Generator level Zbb ? m-m bb
16
Generator level Zbb ? e-e bb
Minv(ee-)g
Minv(ee-)
17
Generator level Zbb ? e-e bb
Minv(ee-)g
Minv(ee-)
Only events with a g of Bremsstrahlung
18
Generator level Zbb ? e-e bb
19
Generator level
Bremsstrahlung events
20
Generator level
21
Generator level
22
Leptons
muons
electrons
23
Photons Zbb ? m-m bb
Bremsstrahlung photons
All photons
watch the different scale! Subsample(25000 ev)
PT vs. Distance between g and m
24
Photons Zbb ? m-m bb
All photons
Bremsstrahlung photons
DR 0.15
Distance between g and m
watch the different scale!
25
Photons Zbb ? e-e bb
Bremsstrahlung photons
All photons
5x5 cluster in ATLAS
watch the different scale! Subsample(25000 ev)
PT vs. Distance between g and e
26
Photons Zbb ? e-e bb
All photons
Bremsstrahlung photons
DR 0.25
watch the different scale!
Distance between g and e
27
Photons

• If the photon is emitted close to electron,
• only one cluster is reconstructed and the
  combined
• 4-momentum is measured. For
  
• the g is effectively absorbed in the same
  cluster of the electron.
• 
  corresponds to a 5x5 cluster in ATLAS. The energy
  of electron is measured in e.m. calorimeter, q
  and f are measured in tracker.
• Photons, inside cone DR, are taken in account
  summing their 4-momentum to the 4-momentum of the
  nearest lepton.
• Photons reconstructed are e.m. clusters without
  track associated (Dh 0.025,Df 0.05 ) and PT
  gt5GeV, h gt2.5.

28
All Photons Zbb ? m-m bb
Zbb ? e-e bb
Zbb ? m-m bb
PT ggt 5 GeV h gt2.7 recontructed g
29
Reconstruction

• Muons combined reconstruction (inner detector
  spectrometer), PT gt10 GeV, h gt2.5.
• Electrons energy measured in calorimeter,(q,f)
  measured from tracker, PT gt10 GeV , h gt2.5.
• Photons e.m. cluster without track associated
• Dh 0.025, Df 0.05, PT gt 5 GeV, h gt2.5.
• Brem photons gt2 leptons reconstructed
• g reconstructed in a cone around lepton
  direction lt DR.
• Electrons sample only one cluster is
  reconstructed.
• Muon sample m track near photon cluster,
  misidentification.
• The number events reconstructed with g brem
  identified is at percent level.

30
Reconstruction Zbb ? m-m bb
Minv(mm-)g
Minv(mm-)
31
Reconstruction Zbb ? m-m bb
Minv(mm-)
Minv(mm-)g
Only events with a g of Bremsstrahlung
32
Reconstruction Zbb ? e-e bb
Minv(ee-)g
Minv(ee-)
33
Reconstruction Zbb ? e-e bb
Minv(ee-)g
Minv(ee-)
Only events with a g of Bremsstrahlung
34
Reconstruction
Only events with a g of Bremsstrahlung
35
Reconstruction
The number g bremsstrahlung reconstructed
separatly in muons events is higher than
electron events, where are mostly absorbed in
the electron cluster
36
Generation-reconstruction
37
Generation-reconstruction
38
Conclusions

• A study for an experimental method of background
  subtraction is developed based on fully
  simulated samples, Lint 20 fb-1 of
• bb Z ? m m- and bb Z ? e e-
• comparing
• the Bremsstrahlung of the two samples.
• the ATLAS detector response.
• The study is preliminary and needs careful
  tuning on data collected.
• DR is an extremely important parameter, that
  should be carefully tuned.

39
Conclusions

• In the region Minv 98 -135 GeV the ratio is
  reasonably
• stable and the effect of Bremsstrahlung is
  negligible with respect to detector response
• The same method can be applied to H? tt and in
  general to all decays where the electron decay
  is negligible in the signal.
• The method relies on lepton universality and
  needs a good knowledge of detector response for
  different leptons
• The method doesnt need an implementation on MC
  of higher order corrections

Acknowledgements Borut Kersevan, Dirk
Zerwas,Donatella Cavalli
40
Riserve

• Riserve

41

• Matching track narrow
• Dh 0.025 Df 0.05
• Isolation broad
• h 0.05 Df 0.10

DR 0.07 size of cluster 5x 5 Dh 0.025 Df 0.05
42
(No Transcript)
43
(No Transcript)