Run6T FPD Energy Scale Correction Update

As discussed last week, I am proposing a scheme to correct for the energy in the FPD. The basic idea is the following.

a. If photon is in the central 4X4, we use cluster energy.

b. If photon is between central 4X4 and 5X5, we use lightly corrected cluster energy with maximum correction of ~2%.

c. Between 5X5 and 6.5X6.5, we use reconstructed energy that is normalized at the 5X5 boundary, and with both distance to the edge and distance to the other photon dependent corrections.

d. FIducial cut is at 0.25cell distance from the edge.

1. Cluster Energy Normalization

We have found in the simulation that the cluster energy away from the edge has very little dependence on anything. It is a very reliable measure of the thrown energy, albeit overall normalization factor. In order to calculate this factor, we look at a kinematic region where we believe the separation measurement is very reliable.

Fig. 1. Reconstructed/Generated Separation for the selected cuts, full simulation

We apply the same cuts to the data, and use the reconstructed pi0 mass centroid to get the normalization.

2. Get distance-to-the-edge dependent corrrection factors from simulation

We use the full simulation to look at the ratio between cluster energy(or reconstructed energy) and thrown energy as a function of the distance to the edge. Turns out these ratios depend very little on the overall energy scale, and we use the fits at E=40~45GeV to get the factors.

Fig. 2.  Cluster Energy Corrections for 2.0 < (3.5 minus distance to the edge) < 2.5

Cluster energy starts to fall off at ~1.5 cell away from the edge, and we approximate the dependence as a linear function between 2. and 2.5

Fig.3. Reconstructed Energy Corrections for 2.5 < (3.5 minus distance to the edge) < 3.25

For Fig. 3, only photons that came from 1-photon clusters and that didn't have a neighboring photon in 1.8cell width radius were chosen. 

3. Get the distance-to-the-neighbor correction for the edge photons using reconstructed energy

Apart from overall normalization, we use the following graph to apply rough corrections for edge photons (from 1-photon cluster) that have a neighbor closeby.

Fig. 4. Reconstructed/Generated energy vs. Generated separation

4. Edge photons from 2-photon cluster --> Deal with it later, since it does not really affect the center cut Pi0s or Etas

The first pass using these new energy corrections were made for the data, but for simulation it is still in the process. Let's just take a look at the Eta mass distribution using the reconstructed energy as before (in red) and newly corrected energy. (in black)

Fig. 5. Mass Distribution in 5GeV energy bin from 50~70GeV, Data