Presentation is loading. Please wait.

Presentation is loading. Please wait.

PFA Development – Definitions and Preparation 0) Generate some events w/G4 in proper format 1)Check Sampling Fractions ECAL, HCAL separately How? Photons,

Similar presentations


Presentation on theme: "PFA Development – Definitions and Preparation 0) Generate some events w/G4 in proper format 1)Check Sampling Fractions ECAL, HCAL separately How? Photons,"— Presentation transcript:

1 PFA Development – Definitions and Preparation 0) Generate some events w/G4 in proper format 1)Check Sampling Fractions ECAL, HCAL separately How? Photons, electrons in ECAL Neutral hadrons in HCAL (no ECAL int.) Charged Pions in HCAL (don’t forget ECAL mips) Detector – SDFeb05 Sci HCAL Photons in ECAL -> sf = 0.012 K L 0 in HCAL -> sf = 0.06

2 Pions KL0KL0 KL0KL0 N - ? Hadron Comparisons G4 Physics List? – under investigation

3 PFA Development – Definitions and Preparation 2a) Single Particle Response -> Analytic Perfect PFA Expected values for E resolution? Why not!? -> G4 problem? Go Back To 0) 2b) Analog/Digital Readout!? 2a) Calibration How? With/without threshold cut? Realistic methods? 2b) Choice of threshold cut Necessary? Realistic?

4 Analytic Perfect PFA – SDFeb05 Detector Model 57 GeV 22 GeV 11 GeV Photons Hadrons (Pions) No neutral E! Photon resolution =  22.5 x.199 = 0.94 GeV Neutral H resolution =  10.7 x.48 = 1.57 GeV -> PPFA = 19%/  E

5 PFA Development – Definitions and Preparation 3) Perfect PFA with Detector Effects Equal to 2a)? Better than 30%/√E? 4) Now ready for PFA development 1.31 GeV 1.95 GeV -> PPFA = 28%/  E

6 PFA Development – Definitions and Preparation 4) Document and archive all of the above for each Detector Model Web site for archived plots and detector documentation Also needs to include special cuts, etc. 5) Now ready for PFA development Examples of PFA use in detector optimization/evaluation ->

7 P (e,  ) = 1 – C e,  e -x/X0 P (h) = 1 – C h e -l/ I C e,  = (1, 7/9 ) C h = 1 1)PFA optimization - beginning of hadron showers separated (longitudinally) from beginning of EM showers... So, in first layers of calorimeter, want P (e,  ) >> P (h) -> x/X 0 >> l/ I -> I /X 0 should be as large as possible Material I (cm) X 0 (cm) I /X 0 W9.590.3527.40 Au9.740.3428.65 Pt8.840.30528.98 Pb17.090.5630.52 U10.500.3232.81 Calorimeter Absorber Optimization – PFA Application Material I (cm) X 0 (cm) I /X 0 Fe (SS)16.761.769.52 Cu15.061.4310.53 Dense, Non-magneticLess Dense, Non-magnetic... Use these for ECAL * * Note ~X2 difference in I for W/Pb – important for HCAL later

8 P (  ) - P (h) X0X0 X0X0 P (interaction) P ()P () P (h) Shower Probabilities in ECAL (25 X 0 ) W Absorber W Pb SS Cu P (  ) reaches ~100% while P (h) still <20% -> W,Pb probability differences >> SS,Cu -> better shower separation in dense material

9 2) Once P (e,  ) -> 1 and  ’s are fully contained (end of ECAL), want P (h) -> 1 as fast as possible...... W performs better than SS and Pb for HCAL W Pb SS W Pb SS P (h) Length (cm) I d P (h)/d I I 1 1 2 3 2 4 5 3

10 Z jets in SS/W HCAL – Absorber Comparison ~1 m ~0.9 m SD SS HCAL 34 layers – 2 cm SS (1 X 0 ) 1 cm Scintillator 4 I SD W HCAL 55 layers – 0.7 cm W (2 X 0 ) 1 cm Scintillator 4 I Same event - different shower shape in W compared to SS?

11 conemean (GeV)rms  /mean 22.0251.921.44.789.36.052.941.39.414.29.0753.591.28.312.42.104.011.23.252.35.254.641.30.232.70.504.771.29.232.50.754.791.28.232.41 1.004.801.28.232.40 conemean (GeV)rms  /mean 22.0252.071.62.7910.61.052.961.66.514.51.0753.631.56.382.74.104.081.48.312.56.254.761.44.252.49.504.851.43.252.42.754.861.42.252.25 1.004.871.42.252.45 SSW rms cone Energy in fixed cone size : -> means ~same for SS/W -> rms ~10% smaller in W Tighter showers in W 3) And, hadron showers should be as compact as possible...... W looks like the best choice for HCAL Single 5 GeV 

12 SS W Energy measurement in calorimeter – Analog ECAL, Digital HCAL ->  /mean smaller in W HCAL -> same behavior for analog HCAL 4) Energy resolution comparisons for SS, W...  E /mean ~ 24%  E /mean ~ 20% Single 5 GeV Pion W – 2 X 0 sampling SS – 1 X 0 sampling

13 SS W Single 5 GeV Pion – Number of hits (1/3 mip thresh) More hits in W HCAL than in SS -> 30% more hits in the HCAL for W -> better digital resolution for W! W – 2 X 0 sampling SS – 1 X 0 sampling

14 SS W Single 5 GeV Pion – Visible Energy in HCAL More visible energy in W HCAL -> better analog resolution in W W – 2 X 0 sampling SS – 1 X 0 sampling

15 SSW e+e- -> Z (jets) – PFA performance Fits Better PFA performance with the W HCAL for conical showers... however, simple iterative cone reconstructs smaller fraction of events* True PFA -> SS 33%/  E True PFA -> W 28%/  E W – 2 X 0 sampling SS – 1 X 0 sampling

16 ... W looks like the best choice for HCAL -> hadron E resolution depends on I, not X 0 Single particle, PFA resolution comparison results...  E /mean ↓, X 0 ↑ ! Coarser X 0 sampling gives better  E !?  E /mean ↓, I ↓ Finer I sampling gives better  E  E /mean W – 2 X 0 sampling SS – 1 X 0 sampling WWSS

17 Dense HCALs (W absorber) - 4 I in ~82.5 cm IR -> OR SDFeb05 SCI HCAL 55 layers of 0.7 cm W/0.8 cm Scin. Sampling fraction ~6% SDFeb05 RPC HCAL 55 layers of 0.7 cm W/0.8 cm RPC 1.2 mm gas gap Sampling Fraction ~0.0025%!!! HCAL Readout Optimization – PFA Application

18 First – Calorimeter Performances Scin. – Analog ReadoutRPC – Digital Readout Hard to compete with no visible energy? Not a great start, but lets continue anyway 

19 Track/CAL Cell Association Algorithm Scin. – Analog ReadoutRPC – Digital Readout Resolution still better in scintillator, but algorithm reproduces perfect ID in both cases

20 Neutral Finding Algorithm Scin. – Analog ReadoutRPC – Digital Readout Once again, very similar performance

21 PFA Results Scin. – Analog ReadoutRPC – Digital Readout PFA performance is very similar (with same cuts) but reflects underlying CAL resolution

22 Confusion – Leftover Hits! Scin. – Analog ReadoutRPC – Digital Readout Better use of hits in RPC? – good since aren’t that many

23 Summary For LC Detector, HCAL should be as dense as possible -> hadron showers more compact in W – smaller HCAL volume -> more I per cm – smaller Solenoid B-field volume -> more layers for fixed total I HCAL – better resolution since more sampling -> more hits - better digital resolution -> more visible E – better analog resolution PFA (incomplete) used to optimize HCAL absorber First look at comparison of analog (scintillator) and digital (RPC) readout modes for HCAL -> very little visible E, number of hits in W/RPC showers – try finer I sampling? -> compared analog and digital modes with same analysis program Once again, PFA used to evaluate detector performance

24 Particle Flow Algorithm – Horse or Cart? PFA used for : 1)Detector Optimization – absorber type/thickness, longitudinal segmentation and transverse granularity, B-field, tracking volume (radius), etc. -> Detector Model(s) 2)Detector Model evaluation – comparisons, tradeoff evaluations, etc. -> PFA is the Horse! Physicists still have to do work!


Download ppt "PFA Development – Definitions and Preparation 0) Generate some events w/G4 in proper format 1)Check Sampling Fractions ECAL, HCAL separately How? Photons,"

Similar presentations


Ads by Google