LCG Meeting, May 14th 2003 V. Daniel Elvira1 G4 (OSCAR_1_4_0) Validation of CMS HCal V. Daniel Elvira Fermilab

LCG Meeting, May 14th 2003 V. Daniel Elvira2 General Description Simulate HCal test beam 2002 setup using OSCAR_1_4_0, physics list 1.8 Includes beam line (trigger scint. tiles, wire chambers), ECal box, HB, HO, aluminium slab (representing magnet), iron (muon) layer Mimic the data analysis, including calibration from electrons (ECal) and pions (HB): study resolution, response (linearity), transverse shower profiles Compare with data measurements for G4 validation

LCG Meeting, May 14th 2003 V. Daniel Elvira3 ECal 7x7 Crystal Matrix (movable) 2 HCal HB Modules HO Panels Aluminum Slab HCal 2002 Test Beam beam Moving table Small scale experiment to demonstrate that HCal works: 49 ECal crystals, 144 HB channels, 16 HO channels. Over 100 Million Events!  - : 225 GeV e - : 20,30,50,100 GeV  - : 20,30,50,100, 300 GeV

LCG Meeting, May 14th 2003 V. Daniel Elvira4     Based on 1-5,000   events onto the (  )=(6,2) tower of the HB and crystal 25 of the ECal matrix TB02 Simulation ( OSCAR/G4 ) Crystal 25

LCG Meeting, May 14th 2003 V. Daniel Elvira5 Angle view of the full TB02 detector 10 GeV electron HB ECal HO Beam Line HO HB ECal Beam Line TB02 Simulation ( OSCAR/G4 ) 100 GeV pion

LCG Meeting, May 14th 2003 V. Daniel Elvira6 E  corr = 122*E HB 5x *E ECal 3x3 E HB scint E HB scint + 0.1* E HB scint MeV * Rand E HB tower E HB tower MeV * Rand Long. Non-uniformity Electronic Noise Calibration factors (already there in previous presentation) E ECal tower E ECal tower +  match * Rand (To match the measured electron resolution-much worse than M.C.) Add more energy points at 10, 25, 200, 250, 300 GeV (in addition to 20, 30, 50, 100, 150 GeV) 5 times more statistics GeV Improved M.C. Simulation

LCG Meeting, May 14th 2003 V. Daniel Elvira7  /E=30.9% (similar to the fit to ¾ of second peak by Jordan) Data analysis: sources systematic errors: muon (pion decay) & electrons (from scrapping?) backgrounds calibration Data Analysis

LCG Meeting, May 14th 2003 V. Daniel Elvira8 From J.Damgov Electron background Muon background Background subtracted muon peak It should be possible to estimate background subtraction error by defining a band limited by an “aggressive” (bckgnd rejection) and a “conservative cut” (signal efficiency) The resolution comes down from 31% to 26.7% ! Pion Signal TB02:   E /E and e/ 

LCG Meeting, May 14th 2003 V. Daniel Elvira9 Pion peak Muon background Background subtracted pion peak  E /E = 20.3% TB02:   E /E and e/  50 GeV pions Pions showering in the crystals Data energy resolution for 50 GeV pions (almost negligible background close to the pion peak)

LCG Meeting, May 14th 2003 V. Daniel Elvira10 TB02:   E /E and e/  The agreement is excellent in all the energy range Shape difference: e/h (e.m. & nuclear x-sec), leakage……? Data systematic error analysis in progress Validate GEANT4 physics models Resolution Linearity

LCG Meeting, May 14th 2003 V. Daniel Elvira11 Transverse Shower Profile Future: remove ECal, add integrated profiles, longitud. profile studies, HO fraction & compare with OSCAR CMS & CMSIM (Salavat), and TB data (Jordan) G4/OSCAR TB02 crystals HB towers Fill ( D, E ) With D: distance of crystal/tower to the pion direction E: energy deposition

LCG Meeting, May 14th 2003 V. Daniel Elvira12 TB02 OSCAR_1_4_0 fully functional. Validation studies in progress, for week data uncertainties are essential. Need to port the simulation to OSCAR_2, hopefully for next CMS week in June fighting technical problems related to translation to XML, among other things Document on TB02 M.C. simulation by July Summary & Plans

LCG Meeting, May 14th 2003 V. Daniel Elvira13 Data & MC Electron Resolution 2.6% at 100 GeV for 3x3 tower matrix 2.8% at 100 GeV for 3x3 tower matrix E ECal tower E ECal tower MeV * Rand Electronic noise M.C resol much lower than measurement

LCG Meeting, May 14th 2003 V. Daniel Elvira14 E ECal tower E ECal tower MeV * Rand (noise) 2.3%/ sqrt(E ele ) (photo-statistics) 0.3% (longitudinal non-uniformity) 0.4% (calibration) Other Contributions No significant change E ECal tower E ECal tower +  match * Rand (To match the measured electron resolution) Temperature  R esp =2%/ o C (crystals) 0.5%/ o C (tubes) Longitudinal crystal uniformity Radiation damage Causes (?):