Transcript Multi-Pixel Photon Counter (MPPC)

Integration of SiPM in a high-pressure noble gas
scintillation detector for homeland security
Romualdo Santoro
Università dell’Insubria
M. Caccia, V. Chmill, S. Martemiyanov – Insubria
R. Chandra, G. Davatz, U. Gendotti – Arktis
MODES_SNM

Approved by the European Commission within
the Framework Program 7

Modular Detector System for Special Nuclear Material
The Main Goal is the development of a system
with detection capabilities of “difficult to detect
radioactive sources and special nuclear materials”



Neutron detection with high γ rejection power
γ-rays spectrometry
Other requirements



Mobile system
Scalability and flexibility to match a specific
monitoring scenario
Remote control, to be used in covert operations
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
2
Baseline technology


The Arktis technologies is based on the use of 4He
for the neutrons detection
The main key features of 4He




Reasonably high cross section for n elastic scattering
Good scintillating properties
Two component decays, with τ at the ns and μs levels
Cheaper and easier to be procured wrt 3He

44 cm diameter x 47 cm sensitive length

180 bar 4He sealed system maintaining
gas purity
Topical Seminar on Innovative Particle and Radiation Detectors
R. Santoro et al., 2012 JINST13th
R. Chandra
7 C03035
(IPRD13) 7-10 Oct. 2013, Siena, Italy
3
MODES_SNM System overview
With γ-ray
spectroscopy
capability
Modular system optimized for:
 Fast neutron (4He)
 Thermal neutron (4He with Li
converter)
 Gamma (Xe)
All components are being integrated,
we are approaching the
commissioning and qualification phase
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
4
MODES_SNM R&D


The project baseline is based on PMTs coupled with scintillating material
R&D activities was planned since the beginning to investigate the possibility of
using SiPM as light detector

Why SiPM is so appealing?




high sensitivity (single photon discrimination)
compactness, robustness,
low operating voltage and power consumption
low cost
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
5
SiPM






SiPM is a High density (~103/mm2 ) matrix of diodes with a common
output, working in Geiger-Müller regime
Common bias is applied to all cells (few % over breakdown voltage)
Each cell has its own quenching resistor (from 100kΩto several MΩ)
When a cell is fired an avalanche starts with a multiplicative factor of
about 105-106
The output is a fast signal (Trise~ ns; Tfall ~ 50 ns) sum of signals
produced by individual cells
SiPM works as an analog photon detector
Signals from SiPM
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
6
SiPM






SiPM is a High density (~103/mm2 ) matrix of diodes with a common
output, working in Geiger-Müller regime
Common bias is applied to all cells (few % over breakdown voltage)
Each cell has its own quenching resistor (from 100kΩto several MΩ)
When a cell is fired an avalanche starts with a multiplicative factor of
about 105-106
The output is a fast signal (Trise~ ns; Tfall ~ 50 ns) sum of signals
produced by individual cells
SiPM works as an analog photon detector
The selected device is a large area (13.6 x 14.3
mm2) monolithic array of SiPM units produced
by Hamamatsu: S11829-3344M
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
7
Lab charcterization to fullfill the simulation hints


Minimum detectable light
Detector sensitivity (i.e. S/N or capability to
discriminate an “event” against noise )
Model developed by
Arktis
We expect 255 photons / matrix for 100 keV
deposited energy assuming the 95% of reflectivity
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
8
Lab charcterization to fullfill the simulation hints

Firing the matrix with a calibrated photon flux, we measured the minimum
detectable light a two different temperatures (different performances due to
a combined effect of increased noise and gain drift)
≈ 250 ph @ 25°C
≈ 60 ph @ 21.6°C
We expect 255 photons / matrix for 100 keV
deposited energy assuming the 95% of reflectivity
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
9
Experimental set-up for proof of principle

A short tube (19 cm) used for the proof of
principle

Filled with 4He at 140 bar, an integrated
wavelength shifter and two SiPMs mounted
along the wall (by ARKTIS)

Two SIPMs read-out through the Hamamatsu
electronic board (C11206-0404FB)

2-channels 3-stage amplification with leading
edge discrimination (SP5600A – CAEN)

Digitizer with a sampling rate of 250 Ms/s 12
bit digitization (V720 – CAEN)
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
10
Counting measurements
Test performed measuring:
 Background, n and γ counting rate using 252Cf and 60Co
source in contact
Two triggering scheme:
 Trailing edge discrimination in coincidence
 Trailing edge and delayed gate of each single SiPM in
coincidence

1st Trigger Scheme
Few parameters to be optimized:



Leading and trailing threshold
Delay time (ΔT)
Gate aperture
typical γ event
R. Santoro
2nd Trigger Scheme
typical n event
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
11
Counting measurements
Result for the different trigger scheme @ 28°C
An amazing result, corresponding to a γ rejection power at the 106 level
[ 10 counts in 1000s, for a number of γ given by acceptance*activity*time
= 1/3 * 3 * 104 * 103 ~ 107 ]
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
12
Off-line data analysis

Data recorded with a minimum bias trigger




For each triggered events we digitize signal of both SiPMs with sample rate
of 125 MS/s and a total duration of 4μs.
Three data set:
1.
2.
3.

Low threshold on the pulse height discrimination
No coincidence between the two SiPMs
400 events without radioactive sources
6000 events with 60Co source in contact
10000 events with 252Cf source in contact
Analysis strategies:





Identify an observable allowing to measure the ratio between noise & particle
induced triggers in samples 2 & 3
Filter noise from particle induced events through a multivariate analysis
Identify the ratio between γ and n events in sample 3
Filter γ from n through a multivariate analysis
Measure the rejection power of interacting γ and the selection efficiency of
interacting neutrons
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
13
Sample composition: (% of Background and signal)


Each signal is baseline subtracted
in the integrated time windows
FAST and SLOW component is
calculated as the integral of the
signal to the left / right side of the
peak
Definition of Fast and
Slow Component


The areas underneath the fits are
used to measure sample
composition
These numbers are used to
estimate the selection efficiency
and bkg rejection power
R. Santoro
BKG
Co60
Cf252
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
14
A multivariate Bayesian analysis
The strategy:

Select 4 no-correlated variables where bkg, γ and n appear to be “reasonably” different

TOT_Diff, Charge Diff, Charge Skewness, Full_charge

Bkg data-set is used to build the experimental probability density functions (p.d.f.)

The corresponding cumulative distributions function (c.d.f.) Ii is then constructed:
Ii (~
xi ) 


~
xi

h i ( x i ) dx i
h i ( x i ) : signal over noise distributions
The four Ii’s are combined to get the final distribution:
3
P  

i 0
(  log  )
i!
i
,   I1  I 2  I 3  I 4
Two step procedure:

1st step: the selection criteria based on P is used to remove the background from n and
γ induced events

2nd

step: the procedure is reiterated to define the n/γ selection criteria (60Co sample
used to build the p.d.f.)
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
15
1st Step
c.d.f. obtained for one of the
selected variables (total charge),
based on the bkg data-set
bkg data-set
As expected we have a random quantity
with a flat distribution
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
16
1st Step
c.d.f. obtained for one of the
selected variables (total charge),
based on the bkg data-set
Moving towards a different p.d.f (60Co
or 252Cf), we have an accumulation of
events on the right part of the histogram
which allows as to separate the signal
form bkg
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
17
1st step: bkg rejection

Repeat the exercise for all the quantities and combine the c.d.f. as follow
3
P  

(  log  )
i 0

i!
i
,   I1  I 2  I 3  I 4
These two plots show that almost all the signal is on the right part of the
histograms (peak) while the bkg is a flat component on the left

The bin at P>0.995 contains
~78% of the γ events
The bin at P>0.995 contains
~99% of the γ +n events for Cf
Selection for
the 2nd step
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
Selection for
the 2nd step
18
2nd step

after the Bkg has been filtered out, the ratio between γ
and n events in 252Cf can be measured:
γ- events in the
60Co data-set
n and γ
composition in the
252Cf data-set
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
19
2nd step: γ rejection

The procedure is reiterated using the 60Co data-set to build the cumulative
distributions to identify γ over n in the 252Cf data-set
Results for a P cut of 0.995
Results are well beyond the expectation
which are pushing us to continue with
 Further test to measure the neutron
detection efficiency and to qualify the
γ/n separation using TOF technique
 Tube layout optimization
 Improved electronics (see next slide)
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
20
Customized electronics


Adjustable bias voltage for each of 16 channels
SiPM temperature readout and gain compensation




The thermo-chip placed onto the SiPM generates a
digital pickup noise which cannot be removed
The frequency of the temperature readout is settable
The board include a lemo connector that can be used
for veto trigger
Improved minimum detectable light and dynamic
range
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
21
Customized electronics


Adjustable bias voltage for each of 16 channels
SiPM temperature readout and gain compensation




The thermo-chip placed onto the SiPM generates a
digital pickup noise which cannot be removed
The frequency of the temperature readout is settable
The board include a lemo connector that can be used
for veto trigger
Improved minimum detectable light and dynamic
range
≈ 30 ph @ 19.6°C
Pedestal
R. Santoro
≈ 60 ph @ 19.6°C
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
22
Conclusion




The Proof of Concept of SiPM arrays in the tube has been
successfully completed
The tube design optimization is certainly required
A new front-end electronics (designed at Uni. Insubria)
has been characterized in the lab and fulfils the
requirements
New test campaign is on the way to optimize the on-line
and off-line analysis
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
23
Spares
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
24
Typical events with PMTs


Neutron/γ discrimination is based on the difference between the fast/slow
component of the scintillation light
Typical neutron and gamma events
Typical plot from the 4He detectors
showing the discrimination between
neutrons (Am-Be) and gamma (60Co)
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
25
Bkg_rejection
Bkg flat distribution
(1st selection)
Co60 flat distribution
(2nd selection)
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
26
2nd step
Distributions of the four discriminant variables used in the second step of the
procedure after having filtered the noise induced events
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
27
2nd step: γ rejection
 n _ select 
N cf
_ sig _ sel
  bkg _ rej * N cf
N cf
_ total _ bkg
 bkg _ rej 
_ total _ signal
N co 60 _ sel
N co 60 _ total _ signal
measured with a data-set
with pure bkg
 rej _ power  1   bkg _ rej



P_value=0.995
Neutron eff = 94%
γ rejection power= 92%
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
28
Few spectra used in the linearity plot
R. Santoro
13th Topical Seminar on Innovative Particle and Radiation Detectors
(IPRD13) 7-10 Oct. 2013, Siena, Italy
29