Comparison of Experimental Electron Signals with GEANT3 and...

Rachid MaziniUniversite de Montreal

Peter LochUniversity of Arizona

Comparison of Experimental Electron

Signals with GEANT3 and GEANT4

Simulations for the ATLAS Forward

Calorimeter Prototype

• Overview of the Forward Calorimeter (FCal) in ATLAS

• FCal Module 0’s

• Monte Carlo set up with Geant4

• Results

• conclusions and Outlook

Rachid Mazini, UdM CALOR2000, October 12, 2000

ATLAS Calorimetry

ATLAS Calorimetry (Geant)

Calorimeters

Calorimeters

Calorimeters

Calorimeters

Hadronic Tile

EM Accordion

Forward LAr

Hadronic LAr End Cap

η ≅ 1.3

η ≅ 4.9

η ≅ 3.2

η ≅ 1.8

Hadronic Tile Calorimeter

Hadronic LAr EndCap Calorimeters

Forward LAr Calorimeters

Electromagnetic LAr Accordion Calorimeters

-6-4

-20

24

6

0

50

100

150

200

250

300

350

0

5

10

15

20

run 17 event 6

pseudo-rapidity η

azimuth φ [deg]

tran

sver

se e

nerg

y E

t [G

eV]

-6-4

-20

24

6

0

50

100

150

200

250

300

350

0

5

10

15

20

run 17 event 6

pseudo-rapidity η

azimuth φ [deg]

tran

sver

se e

nerg

y E

t [G

eV]


ATLAS FCal

• one electromagnetic liquid argon/copper (FCal1)and two hadronic liquid argon/tungsten sections(FCal2/3) integrated into a common cryostat withthe electromagnetic endcap calorimeters;

• cylindrical section 45 cm deep with 90 cm diameter;

Depth[X0] Depth[λ] Weight[t]FCAL1 27.6 2.7 2.1FCAL2 91.2 3.7 3.9FCAL3 89.5 3.6 3.8

• thin gaps of liquid argon are required to avoid positivecharge build-up;

• Electrode Design:

– tube/rod electrodes assembly with cylindricalshell gap of 250/375/500 µm and tube spacing is7.5/8.18/9.0mm centre-to-centre in FCal1/2/3

– Tube material is copper, rods are copper inFCal1 and pure tungsten in FCal2/3. About12330/10320/8120 electrodes in FCal1/2/3

– liquid argon gap is maintained by peek fibre spac-ers


ATLAS FCal

Hexagonal tube/rod pattern Tungsten ”slugs” in FCal2/3

• Two stage summing of electrodes (4/6/9 −→16/24/36) leads to “tiles” readout geometry;

• typical tile size is 0.2×0.2 in ∆η × ∆φ


FCal Module 0

• prototypes for the electromagnetic (Fcal1) and onehadronic module (Fcal2);

• 1/4 ring modules at full depth sufficient for lateralelectromagnetic and hadronic shower containment;6.4 λ total hadronic depth −→ longitudinal accep-tance limitations at higher energies;

η = 3.7

50% lateral containment(pions, 100 GeV)

FCal1 Module 0(front)

80% containment

90% containment

r = 15 cm

# tiles: 192# channels: 256

bi−gain: 64

beam spot ~5 cm OFCal2 Module 0(back)

η = 3.7

r = 15 cm

# tiles: 128# channels: 160

bi−gain: 32

• Testbeam setup at CERN: H6(North area beam line;10-200 GeV/c pions, electrons and muons:)

Beam Particle Trigger(3 Scintillator Counters)

S1 S3

S2

Multi-Wire Proportional Chambers(1 x, 1 y plane each)

Veto System(Scintillator Veto Wall and Hole Veto)

Beamline(Particle Trigger, Position and Direction Measurement)

z

y

Concrete Beam Stop

Muon Counter(Scintillator)

Tail Catcher(Iron/Scintillator Calorimeter)

Low DensityArgon Excluder

Forward CalorimeterModules

1 2

Liquid ArgonCryostat

Test Area(Calorimeter Modules, Leakage Detectors, Particle Identification)


FCal Module 0

• FCal modules in cryostat


Monte Carlo Simulation

GEANT 3.21/11 and GEANT4.2.0R2 base simulationprogram;

Geometry Description

• Beam line: MWPC’s, S1/S2/S3 scintillators, Hole veto includ-ing Pb shielding, lead and iron walls in front of cryostat, Tail-Catcher, concrete beam stop, muon counter;

• Cryostat: wall structure with superinsulation and front/backliquid argon excluders;

• FCal1 and FCal2 modules:

– Dimensions of electrodes and modules measured (FCal1) orfrom drawings (FCal2);

– Electrode positioning (x, y) read from external file describ-ing the readout;

Particle generation:

• particle vertices in (x, y) and directions from reconstructed ex-perimental data; correlation between vertex and direction (beamfocusing) is automatically included; the same data files are usedfor both Geant3 and Geant4;

• particle momentum (20-200 GeV) smeared by an estimated0.5% beam momentum spread.

CUTS:10 KeV (Geant3) and 0.5, 1, and 2mm range cut (Geant4):

G3 G4(0.5mm) G4(1mm) G4(2mm)ELAr(KeV); γ 10 4.4 6.18 8.67ECu(KeV); γ 10 17.1 24.6 35.7CPU (s/GeV) 3.8 0.75 0.68 0.61


Test Beam set up in Geant4

Geant4 set up:


FCal Modules in Geant4

FCal1 FCal2

Reconstruction of Monte Carlo Data

• visible energy in individual electrodes collected intotiles using the experimental cabling/readout descrip-tion database;

• experimental noise from randomly triggered “empty”events is added cell by cell, using the experimentaland Monte Carlo electron calibration constants cexp

(in GeV/ADC) and cMC (inverse sampling fraction):

Erec = cMCEvis + cexpA (GeV )

in any given tile; A is the noise signal in ADC counts.

• cexp and cMC are both calculated from the average60 GeV signal.

Sampling fraction:

G3(10KeV) G4(0.5mm) G4(1mm) G4(2mm)

c−1MC(%) 1.44 1.42 1.41 1.36


Results

40 GeV Electron Signal Distributions

0

0.01

0.02

0.03

0.04

0.05

0.06

-20 0 20 40 60 80Reconstructed Energy [GeV]

rela

tive

entr

ies/

1 G

eV

Experiment

G3 + Noise

Reconstructed Energy [GeV]

Experiment

G4(0.5mm) + Noise

0 20 40 60 80


Results

200 GeV Electron Signal Distributions

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

50 100 150 200 250 300Reconstructed Energy [GeV]

rela

tive

entr

ies/

4 G

eV ExperimentG3 + Noise

Reconstructed Energy [GeV]

ExperimentG4(0.5mm) + Noise

100 150 200 250 300


Results

FCal1 Electron Energy Resolution

Beam Energy [GeV]

Rel

ativ

e E

nerg

y R

esol

utio

n σ/

E [%

]

Erec G3 + noise

Erec G4(0.5mm) + noise

Erec Experiment

0

5

10

15

20

25

30

35

40

45

0 25 50 75 100 125 150 175 200 225

Exp. G3(10KeV) G4(0.5mm) G4(1mm) G4(2mm)

A 39.2±10.9 38.4±9.1 30.5±13.1 30.8±14.8 45.9±29.9

B 7.87±0.21 7.64±0.17 7.54±0.16 7.73±0.17 7.55±0.62

c 4.1±0.5 4.5±0.4 5.2±0.6 5.6±0.3 4.9±1.3

A[%GeV−1/2], B[GeV] and C[%] are the fitted resolution parameters


Results

FCal1 Electron Signal Linearty

-3

-2

-1

0

1

2

3

4

-4

-3

-2

-1

0

1

2

3

4

0 25 50 75 100 125 150 175 200 225

Geant3

Geant4 (1.0 mm range cut)

Rel

ativ

e D

iffe

renc

e M

C-D

ata

[%]

Beam Energy [GeV]


Geant4 Simulation

FCal1 Electron Energy Resolution for

Various Noise Cuts

0

5

10

15

20

25

30

35

40

45

0

5

10

15

20

25

30

35

40

0 25 50 75 100 125 150 175 200 225

no noise cut

1σ noise cut

2σ noise cut

3σ noise cut

no noise cut

1σ noise cut

2σ noise cut

3σ noise cut

Geant3

Geant4 (0.5 mm range cut)

Rel

ativ

e E

nerg

y R

esol

utio

n σ/

E [%

]

Beam Energy [GeV]


Results

Detailed Look at Signal Composition

Cell Signal Significance [σnoise]

Rel

. Ent

ries

10-6

10-5

10-4

10-3

10-2

10-1

0 50 100 150 200 250 300

More highly significant signals in experiment than in both

Geant3 and Geant4!


Conclusion

Conclusion and Outlook

• comparison of Geant3 and Geant4 for FCAL1 Module0 shows a good agreement for both signal and energyresolution;

• adding noise to simulated data at cell level producesa rather good agreement between data and MonteCarlo for global; energy sums and resolution, at thelevel of (few) percent;

• Geant4 compares well to the experimental data. How-ever, more statistics is going to be produced for moredetailed comparisons of shower development and sig-nal fluctuations


Comparison of Experimental Electron Signals with GEANT3 and...

Documents

Transcript of Comparison of Experimental Electron Signals with GEANT3 and...