1. md-NUV PET project meeting
Influence of SiPM Single Photon Timing Resolution on Coincidence Timing Resolution with Fast Scintillator .
1E. POPOVA, 1,2S. VINOGRADOV, 1D. PHILIPPOV,
1P. BUZHAN, 1A. STIFUTKIN,
1National Research Nuclear University «MEPhI»
2Lebedev Physical Institute RAS
18 April 2018
AQUA EPFL Microcity, Neuchatel

2. Estimate a coincidence time resolution CTR on the basis of known
For the TOF PET ( and another timing applications) we are interested to
Estimate a coincidence time resolution CTR on the basis of known
photodetector and scintillator parameters.
Choosing of the best photodetector
Choosing of the best scintillator
Chosing of the best photodetector and scintillator
* Photodetector – analogue SiPM
The goal of our studt – to derive CTR as an analytical function of SiPM and scintillator parameters:
single photon time resolution SPTR, pulse shape (single electron response, SER), PDE - SiPM
trise and tdecay of scintillator, photon numbers - scintillator 2

3. Before we will talk about analytical model for the CTR,
we have to define SPTR itself
Uniform light distribution over the sensitive SiPM area, selection of single photoelectron events
Draft Timing group, S.Gundaker
We assume that SPTRdetector is important parameter that affects CTR, however
we often measure SPTR and CTR for the different electronics conditions
(different noise and SER)
How we can estimate electronics noise and BW influence on SPTR value in order to perform transition between the different setups?
Or how we can estimate for the used electronics the SPTRdetector ?

4. Experimental study of SPTR and MPTR and analytical model for MPTR
The main idea of experiment:
Instead of measuring of CTR with annihilation gammas we study MPTR by using light source with variable pulse shape and intensity 4

5. Timing measurements with KETEK SiPM+amplifiers assembly
Experimental setup:
picosecond laser (405 nm, FWHM ≈ 40 ps)
advanced timing optimized 3x3 mm2 KETEK SiPM chip and specially designed (by S. Ageev) and produced monolithic trans-impedance amplifier(s) (BW 1.5GHz) on PCB assembly
External KETEK evaluation kit amplifier
thermal chamber with light protection T=-30° C
digital oscilloscope LeCroy WaveRunner 620Zi (2GHz, 20GS/s )
PMT-monitor for calibration light intensity into Npe
New timing optimized SiPM
SiPM + Amplifiers PCB 5

6. Digital signal processing
Linear fit of baseline
Time stamp is the moment of crossing threshold with linear fit of rising edge of pulse
Charge selection: only 1phe pulses (for SPTR data)
LED threshold = 15 mV
FWHM ≈ 1 ns
DSP: time stamps (Uov = 7.5V)
1_phe pulse shape, Uov = 4.5 V
1_phe pulse shape, Uov = 9.5 V 6 6

7. SPTR measurements LED 15 mV Temperature = -30 С° 3x3 mm2 SiPM,
SPTR = 112 ps
Temperature = -30 С°
SPTR histogram, Uov = 9.5 V
LED 15 mV 7

8. MPTR measurements
In order to study MPTR dependence on SPTR we carried out two experiments (T = -30°C, Uov = 4.5V, SPTR = 147 ps):
1) Light source – laser, FWHM = 40 ps, MPTR measurements
2) Light source –
laser + WLS-fiber,
Tr ≈ 80 ps, Td ≈ 1.8ns,
scintillator-simulated experiment 8

9. Timing resolution - analytical model (S.Vinogradov)
Filtered marked point process
Analytical model “Amplitude noise” for timing resolution
Number of photoelectrons
Excess noise factor of SiPM (include DCR, XT, AP)
Probability density function of light
Probability density function of SiPM SPTR
Single-electron response function (SER)
Constant threshold at the first photon- no CT, no AP, no dark rate
ENF≈1 9

10. Analytical model (laser light)
Gaussian shape of laser pulse and SPTR allows to get CTR dependence on SPTR:
in case if SER is a Heaviside step response it has an analytical form:
in case if SER is a bi-exponential with rise Tr and fall Tf times it has an analytical form:
For typical SiPM pulses (Tr = ns, Tf = 1…100 ns) dependence of CTR on Tr and Tf is rather weak, so it can be approximated as:

11. MPTR experiment – short laser pulse
Analytical model:
Tr = 0.5 ns, Tf = 1 ns
Experimental Fit
Uov = 4.5 V, T = -30 °C
SPTR
Laser trigger electronic jitter?
(not include in model) 11

12. Experiment MPTR with laser+WLS-fiber
MPTR histograms (Tr ≈ 80ps, Td ≈ 1.9ns) :
top – Npe ≈ bottom – Npe ≈ 52.3
Experimental Fit
MPTR FWHM, ps (CTR with scintillator simulation) vs Light intensity
Uov = 4.5 V, T = -30 12

13. Analytical model calculations: MPTR as function of SPTR
for scintillator-simulated pulse
Npe = 100
Npe = 10
Npe = 1
Npe = 100, Td = 1.8 ns
Npe = 100, Td = 18 ns
MPTR has regions with different dependence on SPTR
Kind of plateau for smaller SPTR value is connected with WLS rise time (80 ps) 13

14. Summary
PCB assembly of 3x3 mm2 KETEK SiPM with amplifier was measured, SPTR of 112 ps was achieved (9.5V OV, -30 °C).
The multi-photon timing measurements with different pulse shapes were carried out to show how coincidence timing resolution depends on SPTR.
Analytical model of “Amplitude noise” has a good agreement with experiment results for light intensity Npe > 1.
MPTR for short light pulse may allow to extract true SPTRdetector (not affected by noise) – should be checked
Analytical model shows how MPTR depends on SPTR for long scintillator-like pulses, but it should be checked with more experimental data. 14

15. BACKUP 15

16. Timing measurements with new PCB – multi-photon TR results
N, pixel
SPTR, ps
0.168
163
0.288
185.5
0.549
204
1.072
217
1.932
144
3.585
111
7.353 85
12.6 66 39
36.6 59
32.6 95
29,8
112
28,8
140
27,9
212
26,1
Timing resolution vs Light intensity (in fired pixels), Uov = 4.5 V 16

17. Timing measurements with new PCB – CTR simulation experiment – results
N, pixel
CTR, ps
0.18
1200
0.46
1501
0.86
1202
1.75
850
3.71
499
6.94
376
17.5
240
26.8
188
41.5
151
52.3
131
Simulated coincidence timing resolution vs Light intensity (in fired pixels), Uov = 4.5 V 17

18. Light intensity calibration: laser & laser + WLS (CTR simulation)18

19. Analytical model: CTR as function of SPTR and other parameters
Modern analytical approaches:
Monte Carlo simulations,
Detection event statistics,
Order statistics of photoelectron detection time,
Cramer-Rao lower bound estimation. 19

20. MPTR (CTR with scintillator simulation) vs Light intensity
0.1 1 10
100 3
Experiment
Model
Experimental fit
Number of photoelectrons per pulse
Time resolution (FWHM), ps
Uov = 4.5 V, T = -30 ℃
Analytical model:
Tr = 0.5 ns, Tf = 1 ns
Experimental Fit 20

21. Uov = 4.5 V, T = -30 ℃
Experimental Fit 21