2008.11.06 Gamma-jet reconstruction with the Endcap EMC (Analysis status update)

Ilya Selyuzhenkov November 06, 2008

Gamma-jet reconstruction with the Endcap EMC (Analysis status update for Spin PWG)

• latest version (pdf)
• Version 3: pdf or odp (for spin PWG)
• Version 2: pdf or odp (after phana comments)
• Version 1: pdf or odp (draft for phana meeting)

 

2008.11.11 Yields vs. analysis cuts

Ilya Selyuzhenkov November 11, 2008

Data sets:

• pp2006 - STAR 2006 pp longitudinal data (~ 3.164 pb^1)
  after applying gamma-jet isolation cuts (R_cluster > 0.9 is used below).
• 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.

Figure 1: Reconstructed gamma pt for di-jet events and
Geant cuts: pt_gamma[Geant] > 7GeV and 1.05 < eta_gamma[Geant] < 2.0
Total integral for the histogram is: N_total = 5284
(after weighting different partonic pt bins and scaled to 3.164pb^-1).
Compare with number from Jim Sowinski study for
Endcap East+West gamma-jet and pt>7 GeV: N_Jim = 5472
( Jim's numbers are scaled to 3.164pb^-1 : [2539+5936]*3.164/4.9)

Figure 2: Reconstructed jet pt for di-jet events and the same cuts as in Fig. 1

Yield vs. various analysis cuts

List of cuts (sorted by bin number in Figs. 2 and 3):

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} > 0.9 : Energy in 3x3 cluster of EEMC tower to the total jet energy
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 3: 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: cuts applied sequentially
lower left: ratio of pp2006 data vs. MC sum of gamma-jet and QCD-jets events (cuts applied independently)
lower right:ratio of pp2006 data vs. MC sum of gamma-jet and QCD jets events (cuts applied sequentially)

Figure 4: Number of accepted events vs. various analysis cuts
Data from Fig. 3 (upper plots) scaled to the initial number of events from first bin:
left: cuts applied independently
right: cuts applied sequentially

2008.11.18 Cluster isolation cuts: 2x1 vs. 2x2 vs. 3x3

Ilya Selyuzhenkov November 18, 2008

Data sets:

• pp2006 - STAR 2006 pp longitudinal data (~ 3.164 pb^1)
  after applying gamma-jet isolation cuts (no R_cluster cut applied for this study).
• 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.

2x1, 2x2, and 3x3 clusters definition:

• 3x3 cluster: tower energy sum for 3x3 patch around highest tower
• 2x2 cluster: tower energy sum for 2x2 patch
  which are closest to 3x3 tower patch centroid.
  3x3 tower patch centroid is defined based
  on tower energies weighted wrt tower centers:
  centroid = sum{E_tow * r_tow} / sum{E_tow}.
  Here r_tow=(x_tow, y_tow) denotes tower center.
• 2x1 cluster: tower energy sum for high tower plus second highest tower in 3x3 patch
• r=0.7 energy is calculated based on towers
  within a radius of 0.7 (in delta phi and eta) from high tower

Cuts applied

all gamma-jet candidate selection cuts except 3x3/r=0.7 energy isolation cut

 

Results for 2x1, 2x2, and 3x3 clusters

1. Energy fraction in NxN cluster in r=0.7 radius
   2x1, 2x2, 3x3 patch to jet radius of 0.7 energy ratios
2. Yield vs. NxN cluster energy fraction in r=0.7
   For a given cluster energy fraction yield is defined as an integral on the right
3. Efficiency vs. NxN cluster energy fraction in r=0.7
   For a given cluster energy fraction
   efficiency is defined as the yield (on the right)
   normalized by the total integral (total yield)

 

2008.11.21 Energy fraction from 2x1 vs. 2x2 vs. 3x3 or 0.7 radius: rapidity dependence

Ilya Selyuzhenkov November 21, 2008

Data sets:

• pp2006 - STAR 2006 pp longitudinal data (~ 3.164 pb^1)
  after applying gamma-jet isolation cuts (no R_cluster cut applied for this study).
• 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.

2x1, 2x2, and 3x3 clusters definition:

• 3x3 cluster: tower energy sum for 3x3 patch around highest tower
• 2x2 cluster: tower energy sum for 2x2 patch
  which are closest to 3x3 tower patch centroid.
  3x3 tower patch centroid is defined based
  on tower energies weighted wrt tower centers:
  centroid = sum{E_tow * r_tow} / sum{E_tow}.
  Here r_tow=(x_tow, y_tow) denotes tower center.
• 2x1 cluster: tower energy sum for high tower plus second highest tower in 3x3 patch
• r=0.7 energy is calculated based on towers
  within a radius of 0.7 (in delta phi and eta) from high tower

Cuts applied

all gamma-jet candidate selection cuts except 3x3/r=0.7 energy isolation cut

Results

There are two sets of figures in links below:

• Number of counts for a given energy fraction
• Yield above given energy fraction
  [figures with right integral in the caption]
  

  Yield is defined as the integral above given energy fraction
  up to the maximum value of 1

Gamma candidate detector eta < 1.5
(eta region where we do have most of the TPC tracking):

1. Cluster energy fraction in 0.7 radius
2. 2x1 and 2x2 cluster energy fraction in 3x3 patch

Gamma candidate detector eta > 1.5:
(smaller tower size)

1. Cluster energy fraction in 0.7 radius
2. 2x1 and 2x2 cluster energy fraction in 3x3 patch

Some observation

• For pre1>0 condition (contains most of events)
  yield in Monte-Carlo for eta > 1.5 case
  is about factor of two different than that from pp2006 data,
  while for eta < 1.5 Monte-Carlo yield agrees with data within 10-15%.
  This could be due to trigger effect?
• For pre1=0 case yiled for both eta > 1.5 and eta < 1.5 are different in data and MC
  This could be due to migration of counts from pre1=0 to pre1>0
  in pp2006 data due to more material budget than it is Monte-Carlo
• For pre1=0 condition pp2006 data shapes are not reproduced by gamma-jet Monte-Carlo.
  With a larger cluster size (2x1 -> 3x3) the pp2006 and MC gamma-jet shapes
  are getting closer to each other.
• For pre1>0 condition (with statistics available),
  pp2006 data shapes are consistent with QCD Monte-Carlo.

 

2008.11.25 Yiled vs. analysis cuts: eta dependence

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