Synopsis

• Jet energy resolution is discussed on this session.

Pass2

• One of the main contributions to jet energy resolution is the pT resolution of charged tracks.
• Jet selection:
  • 3x3 clusterization: cluster Et > 150MeV
  • Hybrid tracks: track pt > 150MeV
  • 100% hadronic correction
  • Anti-kt, R=0.4
  • Jets fully in EMCal acceptance
  • 0.02 < NEF < 0.98

pT resolution

• Hybrid tracks are used to get track classes with uniform phi distribution and good pT resolution. Due to the mixed quality of tracks, different bands of pT resolution show up.
  • Category 1: tracks with ITS refit and SPD hits
  • Category 2: tracks with ITS refit, but without SPD hits
  • Category 3: tracks without ITS refit. Note: some of the tracks might have SPD hits.

• The following plots show the pT resolution of tracks that contribute to the jets.
  

• Summary
  • About 90% of the tracks are the ones with best pT resolution
  • Tracks of category 1 and 2 have similar pT resolution.
  • Tracks of category 3 have significantly worse pT resolution and account for about 7% of the total track population. Interestingly, two bands are observed in this category. Might be due to the presence of SPD hits?

Relative contribution to jets

• In order to find out what are the dominate track categories for jets in different pT region, a few QA plots are produced. Note: different jet pT regions are used for MB and EMC-trigger data.

• This panel shows the distribution: (sum of pT of tracks in a particular category)/(charged pT of the jets) for the three different track categories.
  • MB:
    
  • EMC trigger:
    

• This panel shows the distribution: (pT of track)/(charged pT of the jets) for the three different track categories.
  • MB:
    
  • EMC trigger:
    

MC

• By matching particle jets and detector jets, we can estimate the jet energy resolution from simulation.
• All jets are selected to be fully in EMCal acceptance.
• A detector jet is matched to a particle jet if their distance: sqrt(dEta^2+dPhi^2)<0.25

• E_reco vs E_truth
  

• (E_reco - E_truth)/E_truth VS E_truth
  

• Distribution of (E_reco - E_truth)/E_truth in different bins of E_truth for R=0.4