Abstract:  We investigate losses in a single-photon MC event sample due to cuts applied in the conversion method.  We find that most of the (potentially recoverable) losses are due to the "hadronic veto" cut using the postshower detector.

 

Method: We generate a MC event sample consisting of 1 gamma / event which may be w/in the EEmc acceptance, and 4 muons per event which are guarenteed not to be w/in the EEmc acceptance.  Next we apply the cuts used in the conversion method in the same order, and determine the fraction of gammas which survived after each cut.

 

Figure 1 -- Generated pT spectrum for single gammas.  Eta range is generated wider than the EEmc acceptance, z vertex has a 30cm sigma.

 

 

Figure 2 -- Fraction of gamma candidates vs pT, which show up with E_T > 5 GeV in the endcap.  About 45% of the initial event sample satisfies acceptance cuts.

Figure 3 -- Fraction of events vs pT which survive the hadronic veto cut.  We immediately lose 1/2 - 3/4 of the events depending upon pT.  This cut needs to be relaxed and/or eliminated from the analysis.

Figure 4 -- Fraction of events which survive the charged-particle veto cut, plotted vs pT.  These losses are an unavoidable part of the gamma identification method.  It appears we lose a factor of ~3 of our photons here.

 

Figure 5 -- Fraction of events vs pT which satisfy the isolation cut.  This shows essentially no losses.

 

Conclusions:

 

A cut on the postshower response shows a somewhat surprising result... we cut out over half of the number of photons incident on the endcap, and in a pT dependent way.  I thought that the endcap was designed such that a ET=60 GeV gamma or electron would be 90-95% contained w/in the calorimeter stack, and so my assumption was that a piddly little 6-10 GeV photon would be fully contained.  But this shows otherwise.  So we should relax the hadronic veto.  Next posting will investigate this.