Ilya Selyuzhenkov October 27, 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.

Shower shapes scaling options in data-driven maker:

1. scale = E_smd^geant / E_smd^library (default)
   E_smd^geant is SMD energy associated with given photon
   integrated over +/- 12 strips from raw Monte-Carlo,
   and E_smd^library is SMD energy from +/- 12 strips for the library photon.
2. scale = E_Geant / E_library (used before in all posts)
   E_Geant is thrown photon energy from Geant record,
   and E_library is stand for energy of the library photon.

In all figures below (exept for pp2006 data and raw Monte-Carlo)
the SMD based shower shape scaling has been used.

Figure 1: SMD shower shapes: data, raw, and data-driven MC (40 strips).
Vertical axis shows average energy per strip (no overall shower shapes scaling)

Figure 2: Shower shapes: data, raw, and data-driven MC (12 strips)

Figure 3: Shower shapes: data, raw, and data-driven MC (5 strips)

Figure 4: 25 strips SMD cluster energy for data-driven Monte-Carlo
(SMD based shower shape scaling)

Figure 5: 25 strips SMD cluster energy for raw Monte-Carlo
Note, the difference between results in Fig. 4 and 5. for MC gamma-jets (shown in red)
at low energy (Esmd < 0.04) for pre1=0 pre2=0 case.
This effect is due to the "Number of strips fired in 5-strips cluster > 3" cut.
In data-driven Monte-Carlo we may have shower shapes
with small number of strips fired (rejected in raw Monte-Carlo)
to be replaced by library shape with different (bigger) number of strips fired.
This mostly affects photons which starts to shower
later in the detector and only fires few strips (pre1=0 pre2=0 case)