sPHENIX HCal discussion

[Kistenev <kistenev AT bnl.gov>]

Hi, Liang,

better late then never. I think everything is as expected. Doing a bit of correction on the inner detector will help too (every bit counts). Most important - we now killed the high tail. Low tail will be there forever - calorimeter is just not deep enough but it will work. What is left is to try and figure leakage correction  algorithm which should work for events in the low tail (nothing can be done for events in the “peak” except … if energy in the first two sections  is consistent with MIP - flag the guy as “unreliable”). For the rest - LCG computed as weighted “section number” should still produce better results for lerakage (computed as Eleak=f(LCG, Etot measured)) compared to  Eleak = f(Etotal measured).

Edward

On May 26, 2015, at 11:28 AM, lxue <lxue AT rcf.rhic.bnl.gov> wrote:

John, Edward, and All,

Following Atlanta workshop last week, please find some slides in following link for the tile grading response check in GEANT4.

Instead of using prototype, I generated 50k 30 GeV pion in HCal reference design, and use new sPHENIX software world for this

study. No towering clustering is used, analysis code and ntuple file can be found :

/direct/phenix+sim01/phnxreco/users/lxue/G4Sim_RefDesignLightYield/analysis

https://www.phenix.bnl.gov/WWW/p/draft/lxue/sPHENIX/Calorimeters/Talks/sPHENIX_HCal_LightYieldCorr.pdf

The sampling fraction radial / longitudinal CG dependence can be remove by assuming Eff(outer R)=1.0 and Eff(inner R)=0.4 for

outer HCal. For reconstructed energy spectra, the spike at 0 due to channeling/punch through is <3%, and lower side tail due to

energy leakage of deep shower is <5%. The double gaussian structure of reconstructed energy that existed before correction is

eliminated with correction, but need to confirm with more simulations.

Pleas look at the slides for more details. If time allowable I will present this study in sPHENIX simulation meeting this afternoon.

Thanks

Liang