Ilya Selyuzhenkov November 25, 2008

Data sets:

• pp2006 - STAR 2006 pp longitudinal data (~ 3.164 pb^1)
• gamma-jet[gamma-filtered] - data-driven Prompt Photon [p6410EemcGammaFilter] events.
  Partonic pt range 2-25 GeV.
• QCD jets[gamma-filtered] - data-driven QCD [p6410EemcGammaFilter] events.
  Partonic pt range 2-25 GeV.

Some observation

• Fig. 1 [upper&lower left, 3rd bin] indicates that
  cluster energy isolation is the most important cut
  for signal/background separation
• Fig.1 [lower right, 3rd bin] shows that
  R_cluster cut is independent from (or orthogonal to) other cuts
• Fig.1 [upper&lower left 4th bin] shows that
  cut on neutral energy fraction for the away side jet
  rejects more signal that background events
  We probably need to reconsider that cut
• Fig.2 [lower left, 5th bin] shows that
  charge particle veto significantly improves
  signal to background ratio

Fig.2 [lower right, 5th bin] shows that
charge particle veto also independent from other cuts

• Fig.3 [lower left, 5th bin] shows that
  in the region were we do not have TPC tracking (photon eta > 1.5)
  charge particle veto is not efficient,
  although there is still some improvement from this cut.
  This probably due to tracks with eta <1.5
  which fall into large isolation radius r=0.7.

Yield vs. various analysis cuts

List of cuts (sorted according to bin number in Figs. 1-3. [No SMD sided residual cuts]):

1. N_events : total number of di-jet events found by the jet-finder
2. cos(phi_gamma - phi_jet) < -0.8 : gamma-jet opposite in phi
3. R_{3x3cluster}: Energy in 3x3 cluster of EEMC tower to the total jet energy
   R_{3x3cluster}>0.9 for Fig. 1, and it is disabled in Fig. 2 and 3
4. R_EM^jet < 0.9 : neutral energy fraction cut for on away side jet
5. N_ch=0 : no charge tracks associated with a gamma candidate
6. N_bTow = 0 : no barrel towers associated with a gamma candidate (gamma in the endcap)
7. N_(5-strip cluster)^u > 2 : minimum number of strips in EEMC SMD u-plane cluster around peak
8. N_(5-strip cluster)^v > 2 : minimum number of strips in EEMC SMD v-plane cluster around peak
9. gamma-algo fail : my algorithm failed to match tower with SMD uv-intersection, etc...
10. Tow:SMD match : SMD uv-intersection has a tower which is not in a 3x3 cluster

Figure 1: Number of accepted events vs. various analysis cuts
The starting number of events (shown in first bin of the plots) is
the number of di-jets with reconstructed gamma_pt>7 GeV and jet_pt>5 GeV
upper left: cuts applied independently
upper right: expept this cut fired
(event passed all other cuts and being rejected by this cut)
lower left: "cuts applied independently" normalized by the total number of events
lower right: "expept this cut fired" normalized by the total number of events

Figure 2: Same as Fig.1 except: no R_cluster cut and photon detector eta < 1.5
(eta region where we do have most of the TPC tracking)

Figure 3: Same as Fig.1 except: no R_cluster cut and photon detector eta > 1.5