Transcript photon detector upgrade

Compton Photon Calorimeter
Gregg Franklin, B. Quinn
Carnegie Mellon
Design Considerations
• Light Yield and Photoelectrons
• Detector Geometry, EGS Simulations, Linearity
• Decay time
• Crystal Properties
• Light yield and Photoelectrons
Calculate contribution of finite photoelectrons per MeV energy deposited
First, write mean total photoelectrons as:
n pe 

N ,i i Ei
photon
energy
bin
where
Ei  energy of bin i
 i  mean photoelectrons per photon of energy Ei
N ,i =mean number of photons of energy Ei
(integrated flux) x (Compton cross section d/dE) x (bin size)
Probability of getting npe photoelectrons from Compton Photons of energy Ei
 photons
giving npe photoelectrons
 photons
Prob(n pe , E i )   e
-(Ni   ) 2 / 2 Ni
e
-( i Ei   n pe ) 2 / 2 i Ei 

 e
-(Ni   ) 2 / 2 Ni
e
-(  n pe /  i Ei ) 2 /(2 /  i Ei Ni )
(using 

Ni )
Convolution of two gaussians gives variance for npe,i:
 n2,i =i 2 Ei2 Ni (1+ 1 E )
i
If  energy independent, error on summed energy is:
 Esum
Esum

 n , sum
n pe


dN 
1
1



dE   E 
E max
dN
dE
E
0
dE
E max
0
Finite photoelectron
term small if
Emax large

dE E 2
Photoelectrons not a big issue for integrated energy
BUT: Electron tagged data may be easier to analyze with more photoelectrons
+Other calibration issues?
1MeV
5 MeV
Measured energy deposited for
1 Mev, 5 MeV, and 20 MeV energy
deposions
20
MeV
Simulation includes only
photoelectron statistics and
PMT gain variance
Measured Energy Deposited (MeV)
• Detector Geometry, EGS Simulations, Linearity
EGS simulation by Brian Quinn
12.75 MeV photons
ISaint-Gobain
“BrilLanCe 380”
LaBr3(Cd)
1 inch diam.
4 inch thick
(~ 5.3 rad lengths)
511 keV
escape peaks
Density: 5.29 g/cm3
Energy Deposited
Infinite slab still looses
energy due to backscattering
Finite slab energy loss goes
up with photon energy
Linearity improves with thickness,
but is it important?
4 inches
3.0%
Analyzing Power of summed Deposited
Energy as function of Deposited Energy
Threshold
1.5%
EDep Thresh.
25 MeV
% change in Analyzing Power
1% change in
analyzing power
1 MeV
EDep Thresh.
5 MeV
• Decay Time Consideration
Why not use BGO (decay time ~300 nS)?
• Bremstrahlung
• If ~10 kHz and “deadtime” 3* 300 ns, get 1% deadtime
• Other
• Coincidence and singles data
• Electronics set up for ~100 nS gate
• Larger background from tails
Prefer faster decay time (50 ns?)
• Crystal Properties
PbWO4
BGO
GSO
CeF3
BriLanCe
380
PreLude
420
Density
(6/cm3)
8.30
7.13
6.70
6.16
5.29
7.1
Rad Length
(cm)
0.90
1.12
1.39
1.68
~1.9
1.2
Moliere Radius
(cm)
2.0
2.3
2.4
2.6
?
?
Decay time
(ns)
50
300
56:600
30
16
41
Light output
(% NaI)
0.4%
9%
45%
6.6%
165%
84%
8
170
850
125
3150
1600
photoelectrons
(# / MeV)
$$$
4 in max
Natural
decay
This summer
Need to settle on crystal (at least for test)
Test FADC algorithm at CMU this summer
• Gated and integrating modes (simulate summing algorithm)
• Does ADC sum represent #photoelectrons?
• Test resolution on sources
• Need to slow down signal?
• Possibly clip large pulses?
Better linearity simulations
• GEANT4 (Optimization by Guido, some work at CMU)