Triplet polarimeter
M. Dugger, March 2015
1
Triplet production
• Pair production off a nucleon: γ nucleon → nucleon e+ e-.
• For polarized photons σ = σ0[1 + PΣ cos(2φ)], where σ0 is the
unpolarized cross section, P is the photon beam polarization
and Σ is the beam asymmetry
• Triplet production off an electron: γ e- → eR- e+ e-, where eR
represents the recoil electron
• Any residual momentum in the azimuthal direction of the e- e+
pair is compensated for by the slow moving recoil electron.
This means that the recoil electron is moving perpendicular to
the plane containing the produced pair and can attain large polar
angles.
M. Dugger, February 2012
2
Event generator
• Diagrams used
• Screening correction
• Most important diagrams
3
Triplet production (pair like)
3
4
5
4
q
Q
1
2
3
5
2
4
5
q
1
5
q
Q
1
3
time
3
4
q
Q
2
1
Q
2
• Here we have two electrons in final state and must include
diagrams that have 4 ↔ 5 interchange
4
Triplet production (Compton like)
5
5
4
3
3
4
Q
time
Q
q
2
1
4
3
1
2
4
5
Q
5
q
1
3
Q
2
• Includes 4 ↔ 5 exchange
1
2
5
Screening correction
M. Dugger, February 2014
6
Screening correction
• Leonard Maximon informed me that the screening correction for
triplet production should not be important for the beam asymmetry
but was very important for the cross section
• Used the screening correction provided in the paper by Maximon
and Gimm [1] to compare cross section results of the event
generator to values from the NIST
[1] L. C. Maximon, H. A. Gimm Phys. Rev. A. 23, 1, (1981).
7
Comparison plot of total cross section for
triplet production
• Black line: Total cross
section from NIST
• Blue points: Total cross
section from event
generator without
screening correction
• Red points: Total cross
section from event
generator with screening
corrections included
• Very nice agreement once screening is included 
8
Most important diagrams
• The Mork paper [2] says that the Compton-like diagrams and the
switched electron leg diagrams should be negligible at high
photon energy
[2] K. J. Mork Phys. Rev. 160, 5, (1967).
9
Calculations
3
4
5
4
q
1
3
5
time
q
Q
1
2
Q
2
• Ran 100,000 events at Eγ = 9.0 GeV using:
• Full calculation: σ = 10.856 mb
• Reduced calculation: σ = 10.855 mb
• Reduced calculation neglects Compton-like and crossed electronline diagrams
• Reduced calculations are well within 0.1% of full calculations
10
Comparison of GEANT4 study of triplet
polarimeter to previous study
• The SAL detector
• δ-rays
• Pair to triplet ratio
• ASU simulation of SAL
• Could SAL have been modified to work in the Hall-D environment?
M. Dugger, February 2012
11
The GW SAL detector
• Eγ = 220 to 330 MeV
• 2 mm scintillator converter
• Polarimeter located ~39 cm downstream of converter
• Recoil θ = 15 to 35 degrees
• Recoil φ = 0, 90, 180, 270 degrees with Δφ = 44 degrees
• Analyzing power at the event generator level = 12%
• Analyzing power from simulation 3-4% (post experiment)
• Measured analyzing power = 2.7%
• Quick check - Can ASU reproduce the GW results?
M. Dugger, February 2012
12
δ-ray comparison with Iwata 1993 simulation
• Eδ is δ-ray kinetic energy
after traveling through 1 mm of
scintillator using Iwata’s polar
geometry and scintillator
widths
• BLUE: Current ASU
GEANT4 results
• RED: Iwata GEANT3 (scaled
to GEANT4 results by ratio of
signal integration)
• Shapes of the distributions
look similar
• ASU simulation Eγ = 300 MeV
Eδ (MeV)
• Iwata simulation Eγ = 250, 365, 450 MeV
Note: ASU simulation did not wrap scintillators
M. Dugger, February 2012
13
NIST cross sections for triplet and pair
production off carbon
• σpair/σtriplet:
5.75 @ 300 MeV
5.16 @ 9.0 GeV
• Ratio does not vary much over large energy range
M. Dugger, February 2012
14
Comparison of ASU MC of SAL detector to GW results
Note: ASU results are for Eγ = 300 MeV and GW is of Eγ = 220 to 330 MeV
Analyzing power:
• At event generator level: 12.6 ± 0.1 % ASU; ~12% GW (no error reported)
• ASU simulation: 2.65 ± 0.05%
• ASU simulation (30 μm Al wrapped scintillators): 2.8 ± 0.1 %
• GW experiment: 2.7% (no error reported)
• GW simulation: 3-4% (range given with no error reported)
• ASU results are in agreement with the GW results for the SAL detector ☺
M. Dugger, February 2012
15
How could the SAL detector be modified to
work in Hall-D environment?
-4
Air → Vacuum Converter rad length → 10 E → 9 GeV ΔE pair < 1.5 GeV Analyzing power
2.8(1) %
X
6.9(1) %
X
X
12.5(2) %
X
X
X
10.6(2) %
X
X
X
X
17.5(3) %
M. Dugger, February 2012
16
Dependence of analyzing power on
ΔEpair for 16 sector detector
• Analyzing power fairly constant
for ΔEpair < 1750 MeV
• The 16 sector design increases
the analyzing power to 19.1 ± 0.7 %
from 17.5 ± 0.3 % of the 4 paddle
design (ΔEpair < 1500 MeV)
Analyzing power (%)
• Same parameters as previous best configuration but now with 16
sectors instead of 4 paddle SAL design
M. Dugger, February 2012
ΔEpair (MeV)
17
Recap of modifications that would make
SAL type detector work in Hall-D
-4
Air → Vacuum Converter rad length → 10 E → 9 GeV ΔE pair < 1.5 GeV 16 Sector Analyzing power
X
X
X
X
X
X
X
X
X
X
X
X
M. Dugger, February 2012
X
X
X
2.8(1) %
6.9(1) %
12.5(2) %
10.6(2) %
17.5(3) %
19.1(7) %
18
Detector
• Decided to use
Micron S3 design
instead of the S2
70 mm
• The S3 has 32
azimuthal sectors
instead of 16 for the
S2
22 mm
1 mm thick
19
Preliminary design (slide 1)
• Micron S3
• Converter
• 200 events thrown
20
Preliminary design (slide 2)
• Micron S3
• Converter
• Mounting plate and brackets
• Having a removable plate will allow for
easy modification of how the detector is
mounted without having to modify the
chamber
21
Preliminary design (slide 3)
• Micron S3
• Converter
• Mounting plate and
brackets
• Chamber with
electrical feedthrough flange, and
blank flange
• Inner dimensions:
11in by 9in by 9in
Actual design: 12 in by 12 in by 12 in
22
Preliminary design (slide 4)
• Includes ribbon cable from
detector to electrical feedthrough
23
Preliminary design (slide 5)
• With front door in
see through mode
• 200 events thrown
24
• Generated pairs and
triplets with Eγ= 9 GeV
• Required energy of e+
e- pair to be within 1.75
GeV of each other
Energy deposited (MeV)
Simulation
• Required energy
deposition in detector
to be greater than 200
keV
φ (degrees)
• Converter: 35 μm beryllium
• Used full calculation (all diagrams included)
25
• Assumed collimated photon
rate in coherent peak : 99 MHz
• Δt = 4 hours
• Analyzing power: 0.186
• Polarization uncertainty: 0.01
Assumed
P
P

1
2
1
2 2
N  P
cross section weighted counts
Simulation results
φ (degrees)
where N ≡ Rate surviving cuts * 4 hour,
P ≡ Polarization = 0.4, and
α ≡ analyzing power
Rates:
• Total rate on device = 955 Hz
• Rate surviving software and trigger cuts = 64 Hz
26
Future home
Polarimeter location (red box)
• Sources of magnetic fields in red circles
27
B-field study
• Study performed in April 2012
• Applied magnetic field in vertical direction
28
y (cm)
Effect of B-field on δ-rays
No field →
y (cm)
x (cm)
350 gauss field →
M. Dugger, April 2012
x (cm)
29
Azimuthal distribution with applied B-field
• 350 gauss field applied
A[1+Bcos(2φ)]
A[1+Bcos(2φ)+Ccos(φ)]
φ
30
Analyzing power (B)
Analyzing power vs. B-field
Small field → Small systematic effect
Field strength (gauss)
31
Polarimeter stand in collimator cave
32
Upstream of polarimeter stand
• Secondary
collimator
• Secondary
sweep magnet
• Polarimeter
stand
33
Further upstream of polarimeter stand
• Primary sweep
magnet
• Secondary
collimator
34
Construction
35
Vacuum system
• When the polarimeter is installed in the collimator cave, the
vacuum will come from the beam line
• In the test bench, the vacuum has to be provided by a
temporary system
• Using a rotary vane pump for the vacuum system of the test
bench
• Rotary vane pumps will back-stream oil and this issue must be
addressed
36
VisiTrap
• VisiTrap will
catch any backstreaming before
it hits the
vacuum hose
37
Molecular Sieve and Stinger
• Molecular sieve
will catch stray
contaminates
• Loaded sieve
with zeolite and
heated for two
hours
38
Vacuum system attached to chamber (view 1)
• Cleaned chamber
with:
• Acetone
• Methonal
• DI water
• Attached the
vacuum system
39
Leakage and outgassing tests
Procedure:
• Pump down chamber
• Close butterfly valve between vacuum system and chamber
• Record the pressure as a function of time
Slope = 0.0266 mTorr/min
cycle 1
40
Vacuum leakage test results (empty chamber)
• For area and volume calculations of the chamber I included all of
the flanges
• Volume ~ 29.5 liters. Surface area ~ 6500 cm2
• Outgassing rate = (Volume/Area)*dP/dt
• Tim Whitlatch said steel will outgas at a rate of
2E-09 Torr*l/(cc*s)
Cycle
1
2
3
4
dP/dt (mTorr/min) dP/dt (Torr/s) V/A*dP/dt (Torr*l/cc*s)
0.0266
4.4 * 10^-7
2.00E-09
0.0216
3.6 * 10^-7
1.60E-09
0.0169
2.8 * 10^-7
1.30E-09
0.0091
1.5 * 10^-7
6.90E-10
• It looks like the test is consistent with the dP/dt of the chamber
being from the outgassing of steel: No big obvious leaks 
41
Vacuum test results (detector in chamber)
• Five hour pump down to ~20 mTorr
• Ql = V*dp/dt found to be 1.80x10-4 Torr*liter/s
• A turbo pump on the chamber with a flow rate of 100 liters/s with
a working pressure of 2x10-5 Torr will have a Qw = 2x10-3
Torr*liter/s
• Pfeiffer vacuum says that a system is adequately tight if Qw >
10*Ql
• Putting a turbo pump with flow rate 100 liters/s on the base of the
chamber should be sufficient to maintain a vacuum of 2x10-5 Torr
42
Detector upstream view with source stand and
Po210 source
• Teflon fasteners
connect detector
to supports
43
Detector downstream view
44
Preamps
• Decided to have a parallel development of the preamps:
• Glasgow is building a pre-amplification system based off
of the Rutherford Appleton Laboratory RAL-108 pramps
and custom motherboards
• ASU is using a pre-amplification system from Swan
research (“Box16 preamps”)
45
Swan preamps
The STARS detector uses Micron S2
with swan research preamps
Stars detector
Preamp hybrid
Input view
Output view
16 channel box
46
Preamp to feedthrough cable assembly
•
The Micron S3 detector
uses special ribbon
cable connectors
•
Could not find suitable
ribbon cables. Instead
used Kapton wires that
were individually
placed in the cable
connector
•
40 Wires attached to
connector in picture
shown
47
Detector, cable and source
48
Ring side cable
• Only enough
preamps to
instrument the
sectors but
made the ring
side cables
first
49
Preamps wired up and ground connections
• Sector side
cables
• Ring side set
to ground
50
Distribution box connected to preamp
enclosure
• Original
distribution
box
51
New distribution box (view 1)
• While Kei was at
ASU getting trained
to be a polarimeter
expert he was able to
help assemble to new
distribution box 
Signal plate
52
New distribution box (view 2)
Power plate
53
Copper preamp-box grounding (slide 1)
• Preamp supports made
out of anodized
aluminum
• Decided to help ground
the preamp boxes by
using copper foil on the
preamp supports
54
Copper preamp-box grounding (slide 2)
• Lined three sides of the
preamp enclosure with
the copper foil
• View: looking into the
preamp enclosure
through the opening for
the vacuum chamber
feedthrough flange
• Ground connector to ring side of detector
55
Copper preamp-box grounding (slide 3)
• View: looking into the
preamp enclosure from
the top
• Can see the EMshielding copper mesh
for the fan inlet/outlet
grounded to the copper
foil
56
Signal plate grounding
• Cutting copper foil
for the signal plate
grounding
• Also grounded to
the input voltages
(power plate)
57
Signal plate and power plate grounding
58
Fan leads
• Routed the fan leads
through the preamp
enclosure towards the
distribution box
59
View of polarimeter with original distribution
box completely removed
60
Noise reduction
• Wrapping signal wire around
toroidal core reduces noise
• Putting AC Power Entry Module
(with inline filter and earth-line
choke) into LV supply also helped
with the noise
61
The silicon detector
• The detector is very much like a diode operated in reverse bias
mode
• As the voltage is increased across the detector, the depletion
region gets larger
• The larger the depletion region, the smaller the capacitance of
the detector
• For each 3.6 eV of energy deposited in the depletion region there
is one electron-hole pair that is created and then swept out of the
detector
62
Noise of preamps versus input capacitance
• Preamp noise has a linear
relationship with the
detector capacitance
Typical noise versus
detector capacitance plot
63
High voltage
• For the test bench we are using a
temporary power supply that is
rather old
Tennelec TC 952
• The permanent power supply will
be provided by JLab and will be of
higher quality
• The temporary power supply has a
ripple of about +/- 5 mV at 60 Hz
64
Ripple and other noise as function of HV (slide 1)
10 mV/div
10 ms/div
HV = 0V
HV = 40V
Using
Po210
source
HV = 20V
HV = 60V
65
Ripple and other noise as function of HV (slide 2)
10 mV/div
10 ms/div
HV = 80V
HV = 100V
HV = 120V
HV = 140V
66
Ripple and other noise as function of HV (slide 3)
10 mV/div
10 ms/div
HV = 160V
HV = 180V
α
HV = 200V
HV = 200V
200 mV/div & 1 µs/div
67
Alpha o-scope picture
• Polonium 210
source
• Alpha energy =
5.3 MeV
• Signal about 500
mV
68
Electron o-scope picture
• Cesium 137
source
• Signal about 25
mV for this shot
• Finer time scale
for this shot (50
ns/div)
69
Data acquisition system at ASU
• Using a Tektronix logging
oscilloscope as a slow ADC
• Acquisition rate ~ 1 Hz 
LabView signal express GUI
70
Fit to signal
• Assume voltage has same form
V = [Γr Vm/(Γr - Γf)][exp(Γrt) – exp(Γrt)]
Voltage
Po210 signal
time (μs)
71
• Po210 alpha source
Counts
Calibration (sector 3)
Center = 505.5 mV
• Ek = 5304.33 keV
• One hour of data
mV
• Calculated sensitivity = 95mV/MeV
72
Alpha-test widths
• Looked at all 32 sectors
• Looks good except for a
single channel (sector 32,
lower preamp-box channel
10)
73
Voltage
Typical fit to signal for Ba133 source
time (μs)
74
Ba133 MC compared to data (sector 3)
MC
•
MC
• Generated photon energies: 223, 276, 302, 356, 383 keV
•
•
•
Data
Smeared energy deposited by standard deviation of 12 keV
Fit:
• Centers locked to same photon energies that were generated in MC
• Standard deviation was the same for each Gaussian and allowed to vary
• Standard deviation from fit found to be 11+/- 1 keV
Therefore, resolution of detector plus electronics is about 12 keV for this sector
75
Pile-up study
• Threw 10 million photon events
• Only 469 events seen on detector
• Assume
• 108 Hz in photon range
between 8.4 and 9 GeV
• Timing window of preamp
pulse to be 18 μs
• For a single sector we expect 0.7%
of events to have more than one
signal in the timing window
• Pile-up should not be much of an
issue
2.1% of total
76
Work still to be done at ASU
• Need to complete the positioning system (should be able to finish
this week)
77
Convertor tray
Top of converter tray
Bottom of converter tray
78
Positioning system
• Still need to clean
parts and install
limit switches
• I expect the
positioning system
to be ready to ship
by the end of this
week
79
Chamber crated up
80
Initial work to be done at JLab
• Set up the chamber with vacuum system attached and do leakage
and outgassing tests
• Attach the preamps and make sure that the signals look as they did
at ASU
• Send the signals through the fADC and take data
• Attach the position control system and test
New stuff
81
Chamber at JLab
Chamber
• JLab started
receiving the
polarimeter parts
last week 
• Initial test bench
in the
Experimental
Equipment
Laboratory (EEL)
at Jefferson Lab
82