Steps toward ttH Hbb in the Atlas PhysicsAnalysis Environment'

1
Steps toward ttH (H-gtbb) in the Atlas
PhysicsAnalysis Environment.

• Chris Collins-Tooth
• Christian Shaw

2
Outline

• Begin with environment description
• from digi files to ESD, AOD and histograms
• getting full (G4) and Atlfast output
• DC2 MC generation Signal and BG
• generators used, processes used, forced decays,
  status.
• DC2 data currently used (Wjets)
• for practice in new PhysicsAnalysis env.
• Future getting the ttH signal

3
Background

• Reminder In DC2, there is no CBNT.
• Instead, in athena 8.7.0 we have ESD AOD
• Atlfast can now run on GEN,SIM,DIGI and ESD
  i.e. we can produce Combined AOD.

GEN
AOD (G4Fast) (10k/event)
SIM
DIGI
ESD (100k/event)
MC
MC
MC
MC
Reconstruction
4
Environment I.

• athena 8.7.0 with latest Atlfast releases
• AtlfastUtils-00-01-33 AtlfastAlgs-00-01-60
• RecExCommon-00-02-59
• ParticleEventAthenaPool-00-00-57
• UserAnalysis-00-01-04.
• source share/RecExCommon_links.sh
• Pitfall! - gets the distribution copy of
  RecExCommon_topOptions.py so re-copy it from your
  (checked out) RecExCommon-00-02-59/share area!
• run athena several times! (ESD, AOD, analysis)
• athena optRecExToESD.py RecExCommon_topOptions.py
• Including CaloTowerContainer in JO makes ESD
  Extremely large!
• athena optESDtoCombAOD.py RecExCommon_topOptions.p
  y
• produces COMBINED G4 Fast output !
• athena AnalysisSkeleton_jobOptions.py
• nb set the OutputLevel to DEBUG to get ANY msgs
  !

5
Environment - II.

• Many gen, sim or dig files can be processed
  into the same ESD file.
• PoolRDOInput rfio/castor/cern.ch/atlas/projec
  t/dc2/..., rfio/castor...
• Reco on LSF batch can take hours just to get from
  PEND to RUN.
• Can UK group do better? UK sites?
• Currently a 100 event input file produces 300Mb
  ESD !
• We assume multiple ESDs can then be processed
  into one AOD file.
• Back navigation enabled in 3rd run of athena
  (AnalysisSkeleton).
• We assume also that AOD -gt ESD backnavigation
  works under a many-to-many scenario (not tested
  yet!!).

6
Output histograms

• As yet, no ttH DC2 samples to work with (they are
  coming!)
• Analysis written in C using UserAnalysis and
  AnalysisTools packages.
• Simple output histograms to test AOD and
  backnavigation to ESD.

FAST
G4
TRUTH
7
MC in DC2 the story so far...

• Signal ttH (H-gtbb) 20000 events each at
  MH100,110,120,130.
• Generator filter requires 1 lepton Pt gt 10Gev.
• efficiency 65
• Generated with Pythia msub processes
• qq -gt QQH
• gg -gt QQH
• Forced decays
• W leptons- Isolated high Pt lepton is the
  trigger.
• W- quarks.
• Background MC 100k ttbb, W4jets also being
  generated as part of DC2.

8
Plans for ttH
Reproduce previous analysis in paedm (Gen files
will be ready before full simulation) Would like
to look at Kt algorithm again (Previous results
suggest may not be optimal for this channel) Some
interesting ideas from top studies (Continuous
jets, constraint fit for hadronic W) Much
improvement in b-tqgging recently. Should be able
to exceed TDR levels (5 increase in b
efficiency would give 50 increase in 4 b cut)