Ilya Selyuzhenkov May 27, 2008

Shower shapes and triple Gaussian fits for gammas from eta-meson

Figure 1: Shower shapes and triple Gaussian fits for photons from eta-meson
sorted by different conditions of EEMC 1st and 2nd pre-shower layers.
Note: All shapes have been normalized at peak to unity

Triple Gaussian fit parameters:
Pre1=0 Pre2=0
0.669864*exp(-0.5*sq((x-0.46016)/0.574864))+0.272997*exp(-0.5*sq((x-0.46016)/-1.84608))+0.0585682*exp(-0.5*sq((x-0.46016)/5.49802))
Pre1=0 Pre2>0
0.0694729*exp(-0.5*sq((x-0.493468)/5.65413))+0.615724*exp(-0.5*sq((x-0.493468)/0.590723))+0.314777*exp(-0.5*sq((x-0.493468)/2.00192))
Pre1>0 Pre2>0
0.0955638*exp(-0.5*sq((x-0.481197)/5.59675))+0.558661*exp(-0.5*sq((x-0.481197)/0.567596))+0.345896*exp(-0.5*sq((x-0.481197)/1.9914))

Shower shapes: pp2006, MC gamma-jet and QCD jets, gammas from eta

Shower shapes comparison between different data sets:

• gammas from eta-meson decay. Obtained from Will's eta-meson analysis
• pp2006 - STAR 2006 pp longitudinal data (~ 3.164 pb^1) after applying gamma-jet isolation cuts.
• gamma-jet - data-driven Pythia gamma-jet sample (~170K events). Partonic pt range 5-35 GeV.
• QCD jets - data-driven Pythia QCD jets sample (~4M events). Partonic pt range 3-65 GeV.

Some observations:

• Shapes for gammas from eta-meson decay
  are in a good agreement with those from MC gamma-jet sample
  (compare red squares with blue triangle in Fig. 2 and 3).

  MC gamma-jet shapes obtained by running a full gamma-jet reconstruction algorithm,
  and this agreement indicates that we are able to reconstruct gamma shapes
  which we put in with data-driven shower shape library.

• MC gamma-jet shapes match pp2006 data shapes
  for pre1=0 condition, where we expect to be very efficient in background rejection
  (compare red squares with black circles in upper plots of Fig. 2 and 3).

  This indicates that we are able to reproduce EEMC SMD of direct photons with data-driven Monte-Carlo.

• There is no match between Monte-Carlo QCD background jets and pp2006 data
  for the case when both pre-shower layer fired (pre1>0 and pre2>0).
  (compare green triangles with black circes in bottom right plots of Fig.2 and 3).
  This is the region where we know background dominates our gamma-jet candidates.

  This shows that we still do not reproduce SMD response for our background events
  in our data-driven Monte-Carlo simulations
  (note, that in Monte-Carlo we replace SMD response with real shapes for all background photons
  the same way we do it for direct gammas).

Figure 2: Shower shapes comparison between different data sets.
Shapes for gamma-jet candidates obtained with the same gamma-jet reconstruction algorithm
for three different data samples (pp2006, gamma-jet and QCD jets MC).
pt cuts of 7GeV for the gamma and of 5 GeV for the away side jet have been applied.

Figure 3:Same as Fig. 2, but with no cuts on gamma and jet pt.
All shapes are similar to those in Fig. 2 with an additional pt cuts.
Note, that blue triangles are the same as in Fig. 2.