View
217
Download
0
Category
Preview:
Citation preview
3-10 November 2005 Geant4 10th international conference
1
H4sim, a Geant4 simulation programfor the CMS ECAL supermodule
Laboratoire Leprince-Ringuet, École polytechnique, IN2P3-CNRS, Palaiseau, Francepmine@poly.in2p3.fr
Philippe Miné, on behalf of the CMS collaboration
3-10 November 2005 Geant4 10th international conference
2
The CMS electromagnetic calorimeter
Geant4 implementation
Comparison with test beam data : electrons
Comparison with test beam data : pions
Precalibration with cosmic muons
Outline
3-10 November 2005 Geant4 10th international conference
3
The CMS electromagnetic calorimeter
Compact Muon Solenoid at CERN LHC (2007)7+7 TeV p-p, heavy ions
ECAL : 75848 leadtungstate crystals
22×23×230 mm3
Density = 8.28 g/cm3
X0 = 0.89 cmRM = 2.10 cm
Fast < 15 nsRadiation hard20 interactions/25 ns
3-10 November 2005 Geant4 10th international conference
4
The CMS electromagnetic calorimeter
Non-pointing geometry 3° in η(beam axis) and in ϕ
Barrel 36 supermodules,2 end-caps
120 GeV e-
η
ϕ
3-10 November 2005 Geant4 10th international conference
5
The CMS electromagnetic calorimeter
Beam
Readout Electronics: 2 avalanche photodiodes + MGPA (multi gain pre-
amplifier) 3 gains (12,6,1), 10 samples
trigger
Scintillating
fibers
Movable table:
Used to reproduce for the impinging
particle the same almost pointing
geometry of CMS
Supermodule
4 photoelectrons / MeV 35.6 MeV / ADC count
3-10 November 2005 Geant4 10th international conference
6
The CMS electromagnetic calorimeter
Test in the H4 beam at CERN SPS since 1997 : electrons 20 to 250 GeV, pions, muons, laser light
Energy and position resolution, containment, cracks, irradiation recovery
Precalibration with electrons in the H4 beam : 2004 - 2006
Precalibration with cosmic muons : 2004 - 2006
In situ monitoring with laser
In situ calibration with electrons from Z, W, …
3-10 November 2005 Geant4 10th international conference
7
Geant4 implementation
CMS has a Geant4 simulation program, OSCAR, for the complete detector, interfaced with the reconstruction
H4sim is a standalone simulation of one supermodule in H4
Geometry description by Geant4 data cards with a text fileDDD (Detector Description Database) in XML format
Aluminiumframe
3-10 November 2005 Geant4 10th international conference
8
Geant4 implementation
No magnetic field (time consuming tracking)No matter behind the crystalsCan add specific H4 devices, e.g. hodoscopesOutput of the crystal energies in a ROOT file, interface toRRF (Raw Root File) to simulate beam data
3-10 November 2005 Geant4 10th international conference
9
.XML files
.txt or .root
files
.mac files
Geant4 implementation
XML readout
Output
3-10 November 2005 Geant4 10th international conference
10
Geant4 implementation
Production cuts :
1mm for e-,e+ and γcorrespond to :~ 1.15 MeV for electrons in PbWO4
~ 0.59 MeV in Al
Version 6.2. p 02
Execution time : 1.4 s for 50 GeV on 2 GHz
3-10 November 2005 Geant4 10th international conference
11
Geant4 implementation
- Interactiveversiongraphical interface & visualization - Batch version
Special H4sim commands
3-10 November 2005 Geant4 10th international conference
12
Geant4 implementation
Centre of crystal face Centre of crystal 120 GeV calibration
3-10 November 2005 Geant4 10th international conference
13
Geant4 implementation
100 GeV → π 500 GeV µ-
3-10 November 2005 Geant4 10th international conference
14
TEST BEAM
E25 = 97,5 %
E9 = 94,5 %
E1 = 78,9 %
MONTE CARLO
E25 = 97,5 %
E9 = 94,6 %
E1 = 78,9 %
Electron beam dataElectron beam data
3-10 November 2005 Geant4 10th international conference
15
y (mm)-10 -5 0 5 10
E1
0.5
0.6
0.7
0.8
0.9
1
1.1
E1 vs Y (hodoscopes) normalized
testbeam
simulation
x (mm)-10 -5 0 5 10
E1
0.5
0.6
0.7
0.8
0.9
1
1.1
E1 vs X (hodoscopes) normalizedtestbeam
simulation
y (mm)-10 -5 0 5 10
E1
0.5
0.6
0.7
0.8
0.9
1
1.1
E1 vs Y (hodoscopes) normalized
testbeam
simulation
x (mm)-10 -5 0 5 10
E1
0.5
0.6
0.7
0.8
0.9
1
1.1
E1 vs X (hodoscopes) normalizedtestbeam
simulation
Test beam
Simulation
X (mm) Y (mm)
Position scan at centre of crystal 204
Electron beam data
Y goodX ~ 2 %
3-10 November 2005 Geant4 10th international conference
16
Phi EtaData
Simulation
log(E2/E1)-3 -2 -1 0 1 2 3
SU
M 9
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
1.005
1.01
Test-Beam
Monte-Carlo
comparaison Test-Beam / Monte-Carlo @120 GeV Xtal 1104ηS9 vs log(E1/E2)
log(E2/E1)-3 -2 -1 0 1 2 3
SU
M 9
0.96
0.965
0.97
0.975
0.98
0.985
0.99
0.995
1
1.005
1.01
Test-Beam
Monte-Carlo
comparaison Test-Beam / Monte-Carlo @120 GeV Xtal 1104φS9 vs log(E1/E2)
Electron beam data
3x3 crystals impact position measured by log(E2/E1) function
3-10 November 2005 Geant4 10th international conference
17
Electron beam data
3X3 5X5
S = stochastic
N = noise
C = constant
2
222
CE
N
E
S
E+
+
=
σ
3-10 November 2005 Geant4 10th international conference
18
Electron beam data
ECAL used in trigger
Simulation of the TPG linearity (Trigger Primitive Generator)
Data (green)Simulation (black)
3-10 November 2005 Geant4 10th international conference
19
Pion beam data
GeV
# Entries5x5 energy
for 50, 120 &
180 GeV
pions
3-10 November 2005 Geant4 10th international conference
20
Pion beam data
Compare the mean valuesµ ± σof Landau-fit for 4 crystals@ 3 energies
for 4 crystals @ 3 energies
204 704 1104 1404
204 1104 1404
204 704 1104 1404
3-10 November 2005 Geant4 10th international conference
21
Pion beam data
Minimum ionizing particle (mip) are visible
50 GeV
λ ~ 23 cm
E = 270 MeVGeV
# E
ntr
ies
Pedestal
mip
3-10 November 2005 Geant4 10th international conference
22
Less good agreement for crystals 1104 & 1404
Pion beam data
Almost constant, as it should be for mips
3 energies for 4 crystals
204 704 1104 1404
3-10 November 2005 Geant4 10th international conference
23
Precalibration by cosmic muons
Raise the APD gain by a factor 4
Select the muons parallel to the crystal axisVeto neightbouring crystals
3-10 November 2005 Geant4 10th international conference
24
MC
Precalibration by cosmic muons
data
Module 1
Event selection: E1 > 10 ADC counts & E2 < 3 ADC counts
3-10 November 2005 Geant4 10th international conference
25
More difficult for module 4 (eta angle)⇒ incline the supermodule by10 degrees
Precalibration by cosmic muons
Intercalibration precision of 2-3% can be reached with 1 week of continuous cosmic ray data taking
3-10 November 2005 Geant4 10th international conference
26
Peak after final selection:
good agreement between data (after calibration) and MC
Precalibration by cosmic muons
Data: <evts/xtal/day> = 55±2 MC: <evts/xtal/day> = 61±2
Disagreement ~ 10% due to trigger inefficiency and uncertainty on the overall normalization
3-10 November 2005 Geant4 10th international conference
27
Conclusion
Geant4 based H4sim is widely used by the ECAL CMS communityInterpretation of test beam dataTest of algorithms and cosmic precalibration
Acknowledgments to :T. Frisson, J. Gilly, P. Mora de Freitas, S. Bimbot, P. Paganini, P. Meridiani, R. Salerno, P. Jarry, J. Descamps, A. Ghezzi, S. Vigano, C. Marchica, A. Holzner, P. Bloch, A.Zabi and the ECAL group
Recommended