validation of geant4 by the cms experiment

GEANT4 workshop ’03 GEANT4 Validation by CMS Pedro Arce(CIEMAT/CERN) 1

Validation of GEANT4 by the CMS Experiment

Pedro Arce (CIEMAT/CERN)

On behalf of the CMS Simulation Group

GEANT4 Workshop ’03, TRIUMF, 2nd August 2003


Index

Status of CMS GEANT4 simulation (OSCAR)

Silicon tracker studies

EM calorimeter studies

Muon physics studies

Hadronic calorimeter studies

Conclusions


OSCAR status

OSCAR (Object Oriented Simulation for Cms Analysis and Object Oriented Simulation for Cms Analysis and Reconstruction)Reconstruction) is (almost) ready to substitute GEANT3-based simulation in the official CMS production

Extensive physics checking has been carried to test GEANT4 physics vs. GEANT3 and vs. TestBeam data where available

1.2 M single particle events + 300 k full events have been produced

CMS DC04 challenge (50 M events) have started this summer with GEANT3 simulation, we hope to make a major part of this production with GEANT4


OSCAR status (II)

CPU is acceptable: 1.5 - 2 X GEANT3

Memory is acceptable: 220 Mb OSCAR vs. 100 Mb GEANT3 simulation

Crashing rate is acceptable: 1 / 20000 events

Becoming smaller at each GEANT4 release

Still some problems in tracking and hadronic physics


View of CMS muon system

Sliced view of CMS barrel detectors

View of CMS pixel detector


dE/dx in Silicon Detector

• dE/dx has been studied for:– Particle Types: Muon, Pion, Kaon and Electron– Momenta: 2, 3, 4, 5 and 10 GeV

• It is compared with:– GEANT3– ALEPH data– PHOBOS data

OSCAR (GEANT4) CMSIM (GEANT3)


dE/dx in Silicon Detector (II)

RELATIVISTIC RISE

• We do not have that much data available ;)

– PHOBOS: relativistic rise for pions between 0.5 – 8 GeV is 4% (with pads!)

– ALEPH: relativistic rise for pions between 1 – 18 GeV is 7% (but only 2 layers!)

– We lack any data about electrons

– GEANT4/GEANT3 with the proposed settings: we “measure” a 4% rise between 1 and 10

GeV.


Resolutions in Silicon Detector

Pt resolution:

Lower resolution probably due to smaller spread from multiple scattering (~< 10 %)

PT = 10 GeV

PT = 100 GeV


EM calorimeter

• SETUP: • CMS electromagnetic calorimeter: lead tungstate crystals pointing to interaction point• Photon of E = 30 GeV enters front face of crystal• No magnetic field

• What was compared:• E in hottest crystal, in 3x3, in 5x5 matrix• E total

• dNhits/dt, for 1 ns time slices

• GEANT3.21 vs. GEANT4.5.1


EM calorimeter

Shower shape: From GEANT4 gallery:Electron 5 Gev in PbWO4

(more detailed)

• GEANT4 electromagnetic shower looks thinner than GEANT3• Test beam data seems to agree with GEANT4


Energy resolution in EM calorimeter

GEANT4

GEANT3

GEANT4

GEANT3


Muon Physics

Single block of iron

• To concentrate on physics differences• Similar results for other materials

Single muons of 10, 100, 1000 GeV • Energy loss• Delta rays• Bremsstrahlung• Pair production• Muon nuclear interaction (negligible)• Multiple scattering

Dependency with production cuts (plots shown with 0.1 mm)

Check new ´integral´ processes in GEANT4 (GEANT4.5.0.ref02)

– No difference found


Total Energy Lost

10 GeV

100 GeV

1000 GeV


Bremsstrahlung

10 GeV

100 GeV

1000 GeV


Pair Production

10 GeV

100 GeV

1000 GeV


Multiple scattering

10 GeV

100 GeV

1000 GeV


Multiple scattering

LEP I Z events in L3 detector (45 GeV)

50 GeV GEANT3.21 vs

GEANT4.5.2.ref01GEANT3 GEANT4.5.2.ref01


Muon Physics

PT Resolution:

GEANT3

GEANT4


EM calor. + Hadr. calor. testbeam

2002 Testbeam Set-up for CMS HCAL• A module of copper absorber plates with scintillator tile sampling

• 28 scintillator plates with absorber of varying thickness in between (only 3 samplings)

• A prototype lead tungstate crystal electromagnetic calorimeter• Energies: 225 GeV (for calibration), 10-300 GeV , 10-300 GeV e-• Magnetic Field: 0, 0.75, 1.5, 3 tesla (direction parallel to the face of the scintillator)• Configuration: Only HCAL or ECAL + HCAL

• 5000 events were taken for each set-up• Response of the Calorimeter was studied as a function of:

• Magnetic field: effect on scintillator• Absorber thickness: optimisation of resolution versus containment• Absorber depth: energy containment • Electromagnetic Calorimeter contribution: e/ effects

• GEANT4.4.1 (physics list 1.8)


Resolutions: M.C. & data agree very well

Source of discrepancy could be e/h (e.m. & nuclear x-sec), punch through (shower length) ?

Resolution and Energy Response

• All known contributions to ECal and HB resolutions added• Ecal resolution “matched” to data measured e- resolution


Hadr. calor. GEANT4 vs GEANT3

- 50 GeV: Energy deposited in each layer of Hadr. Calor. shows good agreement GEANT3-GEANT4


Hadr. calor. GEANT4 vs GEANT3

- 20, 30, 50, 100, 300 GeV: Energy deposited in EM Calor., Hadr. Calor. and outer layers of Hadr. Calor. is different

GEANT4 shower is shortero Test beam data seems to agree better with GEANT3, new data being analysed now

EM Calor. Hadr. Calor. Outer layers Hadr. Calor.


Summary

CMS GEANT4-based simulation is almost ready to substitute GEANT3 one

Big effort has been made during recent months to compare GEANT4 with GEANT3 and with Test Beam data where available

Small differences observed sometimes

Precision of test beam data can hardly favour GEANT3 or GEANT4 in most cases

All physics groups agree in switching to GEANT4 as official CMS simulation tool

validation of geant4 by the cms experiment

Documents

geant4 geant4

geant4 physics

cms hcal

cms analysis

gev50 gev geant3

gev pt

geant4 gallery

geant4 release