Comparison of data and simulation of Argonne Beam Test

1
Comparison of data and simulation of Argonne Beam
Test

• July 10, 2004
• Tsunefumi Mizuno
• mizuno_at_SLAC.Stanford.EDU

2
Experimental Setup
Objective to examine the performance of PoGO and
verify the Geant4-based Monte-Carlo Simulator
with a PoGO prototype
PoGO 200-400 units of fast/slow plastic
scintillators and BGOs Prototype 7 units of
fast scintillators (detection part)

• Polarization vector was along the horizontal axis
  (x-axis)
• Irradiated polarized synchrotron beam of 60 keV,
  73 keV and 83 keV at the center of the central
  scintillator. The same beam was used to calibrate
  the energy response.
• Rotated the detector in x-y plane in 15 degree
  steps.

y
beam direction
x
3
Definitions

• We defined that xy plane is normal to the
  scintillator principle axis.
• Scintillators are numbered from 1 to 7. Central
  scintillator is number 4.
• Beam goes from z to -z. Polarization vector is
  along x-axis.
• We rotated the detector. When we rotated it by 30
  degree, scintillator number 2 was along the y
  axis.

y
30degree
polarization vector
1
2
3
4
5
x
Beam Direction
6
7
4
Run Summary

• Coincidence Trigger (in 15 degree steps)
• run048-073 83.5 keV run
• run106-149 60.2 keV run
• run164-190 73.2 keV run
• Ch4 Trigger (in 30 degree steps. They dont cover
  the whole azimuth angle)
• run191-197 73.2 keV run
• run210-212 83.5 keV run
• Calibration Run
• run088,090-096 83.5 keV run
• run090-105,150-156 60.2 keV run
• run157-163 73.2 keV run
• In this report, we use only ch4 trigger run data.
  (Coincidence run data gave similar modulation
  factor.)

5
Calibration Constants

• We had been using calibration constants
  (channel/energy conversion factors) derived from
  the calibration run, but they gave 5 lower
  energy deposition. For reference, please look at
  http//www.slac.stanford.edu/mizuno/PoGO/Argonne/
  report/ArgonneResult_2004-02-07.ppt. I suppose
  that shaper output was slightly affected by the
  trigger condition (due to the change of input
  impedance?), and determined the conversion
  factors so that total energy deposition is equal
  to the beam energy.
• Below we list channel/energy conversion factors
  of 73.2 keV beam.
• ch1 f4.20 (it was 4.46 when we used the
  calibration run data)
• ch2 f4.32 (it was 4.53)
• ch3 f4.35 (it was 4.58)
• ch4 f4.39 (it was 4.34)
• ch5 f4.39 (it was 4.79)
• ch6 f4.17 (it was 4.45)
• ch7 f4.85 (it was 5.15)

6
Geant4 vs. EGS4
We have compared the scattering process of
polarized photons between Geant4 and EGS4, and
found that G4 gave less asymmetry. There are two
reasons for this 1) In Geant4, polarization
vector after the Compton scattering is
excessively randomized. 2) Rayleigh scattering in
Geant4 does not take into account the
polarization. We have fixed them and obtained a
good agreement (within a few ) between two
simulation toolkits.
Simulated modulation curves observed by PoGO
(geometry is simplified) for Crab spectrum in
25-200 keV.
0.2184-0.0022(EGS4) 0.2223-0.0023(Geant4)
7
73.2 keV Ch4 Trigger Run (1)

• Event selection criteria
• Detection threshold was 2 keV
• 2 scintillators detected a hit (one was the
  central scintillator)
• Deposit energy of the central scintillator was
  below 40 keV and less than half of the total
  deposit energy.
• total deposit energy 73.2-25keV

Total deposit energy (keV)
deposit energy in the central scinti. (keV)
Total deposit energy (keV)
8
73.2 keV Ch4 Trigger Run (2)
ch3
ch7
ch2
ch6
ch1
ch5
We fitted data with a sinusoidal curve.
Modulation Factor (an average of 6 channels) is
0.430-0.006. On the other hand, MF predicted by
G4 is 0.500-0.005. Difference between data and
simulation is 15.
9
Effects that affect the MF

• Passive materials
• If we put sensor mount and table (see page 2), MF
  predicted by G4 decreases from 0.5 to 0.48 (i.e.,
  MF decrease by 4). This is mostly due to the
  splash component from the table.
• Accuracy of the simulation
• If we regard EGS4 simulation to be correct, G4
  overpredicts the MF by 2 (see page 6).
• Polarization degree of the beam
• According to Doug Robinson, the calculated source
  polarization was 98-99. This can reduce the MF
  by 1-2.
• Alignment of sensors
• If we shorten the distance between sensors by 2mm
  (from 2.22cm to 2.02cm), predicted MF decreases
  by 1. If we use an extended beam (7mm diameter)
  instead of a pencil beam, MF predicted decreases
  by 1. Considering them, we expect that
  misalignment of sensors could explain 2-3
  difference of MF.
• There still be 5 difference. What else do we
  have?

10
Comparison between data and simulation
Below we compare between data and simulation and
can see the differences. One difference which
could affects the MF is that, data seems to be
contaminated by background. To examine how this
influences the MF, we change the selection
criteria and revaluate the MF (next page).
73.2keV run, 0degree, data
73.2keV run, 0degree, simulation
Total deposit energy (keV)
deposit energy in the central scinti. (keV)
deposit energy in the central scinti. (keV)
11
Ratio of MF factor

• We selected the event based on criteria given in
  page 7, but changed a range of total deposit
  energy. Narrower range (high S/N ratio) gives
  closer values of Modulation Factor between data
  and simulation. Background could explain the
  difference of 3-5.

Ratio of the MF (data to simulation), as a
function of the range of total deposit energy.
Passive materials (sensor mount and table) are
already taken into account in simulation.
totE73-25keV, -20keV, -15keV, -10keV,
-5keV
12
Summary

• We have fixed the Geant4 physics processes
  (polarized Compton and Rayleigh scattering), and
  obtained a good agreement with EGS4 prediction
  (page6).
• Modulation Factor observed in Argonne beam test
  data was 0.43, whereas that predicted by Geant4
  was 0.50. There are 15 difference (page 8).
• Passive materials, accuracy of the simulation
  toolkit, depolarization of the beam and
  instrumental effect (misalignment of sensors)
  could explain 10 difference (page9).
• There seems to be background component of the
  beam (continuum under the 73.2keV line?). If this
  explains the 5 difference, we obtain a good
  agreement with data (pages 10 and 11).