13th Conference on Innovative Particle
and Radiation Detectors
Siena – 7-10 October 2013
Performances of SiPMT array readout for
Fast Time-of-Flight Detectors
M. Bonesini1, R. Bertoni1, A. De Bari2, R. Nardo’2, M. Prata2, M. Rossella2
Presented by M. Bonesini
INFN, Sezione di Milano Bicocca - Dipartimento di Fisica G. Occhialini1
INFN, Sezione di Pavia - Dipartimento di Fisica Nucleare e Teorica2
Outline
 Introduction
 Scintillator based TOF detectors
 PMTs vs SiPMT arrays
 Test setup
 Results
 Conclusions
M. Bonesini - 13th IPRD conference
2
PID methods
• Particle identification
(PID) is crucial in most
experiments (from /K
identification in B physics
to e/ separation at 10-2
level for p< 1GeV)
• At low momenta TOF
methods are used (p 3-4
GeV/c)
M. Bonesini - 13th IPRD conference
3
Particle ID with TOF
 TOF based on measure of t
over a fixed length L

p
c 2t 2
m p
1
2
L
K
d
c 2t 2
m p
1
2
L
 Mass resolution dominated
by t (not measure of L, p)
Beam particle separation in
HARP beam Tof , for a 3 GeV
beam
 Separation power in
standard deviation
L ( m1  m2 )

2 p 2 c t
2
n t ,1 2
M. Bonesini - 13th IPRD conference
2
4
Examples of TOF detectors
 Based on scintillator counters:
simple to made, sensitive to B,
read at both ends by PMTs, good
resolutions -> 50-100 ps
(depends mainly on L,Npe)
 Based on PPC or spark
Rate
chambers: some care in
problems
production, not sensitive
to B, very good resolutions ->
30-50 ps
 Based on RPC’s: cheap (suitable
for large areas), not sensitive to
B, R&D in development, very
good resolutions -> 50 ps
M. Bonesini - 13th IPRD conference
5
Basics of double-sided scintillator counters
Pmt
Scintillator +
lightguide
t  t1  t 2  l v
light
0
2

t t
 x0  1 2  vlight

2
t
t
t
 t0   1    2   1
2
 2   2 
2
 x  vlight
2
t
t
t
  1    2   vlight  1
2
 2   2 
2
2
Precise TOF and Hit position
Tof resolution can be expressed as:
 2 scint   2 PMT   2 pl
t 
  2 elec
N pe
Some points to look have
M. Bonesini - 13th IPRD conference
high resolution tof’s
6
Problems for high resolution scintillator
based TOF (t < 100 ps)
•
pl dominated by geometrical dimensions (L/Npe)
•
scint  5060ps (mainly connected with produced
number of ’s fast and scintillator characteristics, such
as risetime)
•
PMT PMT TTS (typically 70-150 ps)
•
Additional problems in harsh environments:
1. B field (-> fine mesh PMTs/shielding of conventional
PMTs)
2. High particle rates (tuning of operation HV for PMTs)
M. Bonesini - 13th IPRD conference
7
SiPMT arrays vs PMTs
• PMTs
1. Large active area > 0.5 – 1 inches
2. Gain G depends on external magnetic fields B (needs shielding,
aside fine-mesh PMTs)
3. Good TTS: typical values in the range 150-400 ps
4. Fast PMTs are quite expensive: 1000-1500 E
5. Needs HV: typically 1000-2000 V
6. Low noise rate ~1KHZ
•
1.
2.
3.
4.
5.
6.
SiPMT arrays
Active area up to 1x1 cm2 typically
Gain G insensitive to external magnetic fields, but depend on
temperature T (needs feedback)
Good STPR response for single SiPMT ~140-300 ps
Quite cheap
Needs low voltages: ~30 V for SenSL, Advansid , ~70 V for
Hamamatsu
High noise rate up to MHZ
M. Bonesini - 13th IPRD conference
8
Timing resolution of photo detectors
Resolution improves:
•
•
From K. Arikasa NIM A422 (2000)
By decreasing active
area
As
From G. Colazzuol
(LIGHT11 2011)
M. Bonesini - 13th IPRD conference
9
Conventional fast TOF with PMT readout
have found application in many experiments
Tof detectors
MICE at RAL
(Hamamatsu R4998
PMTS)
Liq. Xe Scintillation
Detector
Liq. Xe Scintillation
Detector
Thin Superconducting Coil

Stopping Target
Muon Beam
+
e

e+
Timing Counter
Drift Chamber
Drift Chamber
AMS-02 exp
(Hamamatsu fine mesh
PMTs)
M. Bonesini - 13th IPRD conference
1m
MEG at PSI
(Hamamatsu fine mesh
PMTs)
10
Studies of SPTR (timing for single
photoelectrons)
• Timing studies are usually
done for single SiPMT (not
arrays) with single p.e.
• For scintillation counters
needs to study multiphoton
response
from V.Puill et al NIM A695 (2012) 354
M. Bonesini - 13th IPRD conference
11
A small remark
• Timing studies are usually done with fast lasers (eg 30-50 ps
FWHM ): good for single p.e. studies, but scintillator have
typically 200-300 p.e. signals and scintillator risetime are in
the 1-2 ns range …
• Needs laser signals that resemble more physical
scintillation light
M. Bonesini - 13th IPRD conference
12
Experimental lab setup
Light
MM fiber
Laser head
BS
x/y/z flexure (fiber
launch system)
Laser driver
Fast photodiode
Sync out
Fast Amplifier
Prism injection system
Scintillation counter
PMTL
PMTR
t0
VME
tR
M. Bonesini - 13th IPRD conference
13
Test setup: home-made laser system
1.
2.
3.
4.
5.
Fast Avtech AVO pulser + Nichia violet laser
diode (l ~408 nm)
Laser pulses width selectable between 120 ps
and 3 ns length, with a ~200 ps risetime
(simulate scintillator response)
Laser pulse height selectable to give scintillator
response between a fraction of MIP and 10-50
MIPS
Laser repetition rate selectable between ~100
Hz and 1 MHz
The laser beam is splitted by a 50%
beamsplitter to give a reference t0 on a fast
photodiode (Thorlabs DET10A risetime ~1 ns)
amplified via a CAEN A1423 wideband inverting
fast amplifier (up to 51 dB, ~1.5 GHz
bandwidth)
M. Bonesini - 13th IPRD conference
14
Some details
Laser injection system:
• Newport 20X microscope
objective
• x/y/z Thorlabs micrometric
flexure system
•
•
•
•
•
Acquisition system:
VME based (CAEN V2718 interface)
VME CAEN TDC V1290A (25 ps res)
VME CAEN QADC V792
VME CAEN V895 L.E. discriminator
(typ discr values -50 -100 mV)
home-written acquisition software
M. Bonesini - 13th IPRD conference
15
Tuning of laser setup
From R. Bertoni et al., NIM A615 (2010) 14
• Tune laser settings to
reproduce testbeam
results (t )with a single
counter equipped with
R4998 PMTs and MIP
response
• Study single counter
response substituting
PMTS readout with SiPMT
arrays readout
• Advantage as respect to
cosmics is the possibility
to collect a high statistics
in a short time, with
different exp conditions
(amplifier tuning, …)
M. Bonesini - 13th IPRD conference
16
BC 404 scintillation counter (60
cm long, 6 cm wide) equipped
with Hamamatsu R4998 PMTs
(as in MICE expt)
PMT signals with laser
PMT signals with cosmics
M. Bonesini - 13th IPRD conference
17
Readout chain for SiPMT arrays
• SiPMT array custom mount
• 16 macrocells signals are
summed up in the basette
and then amplified
Schematic of one
``basette’’
M. Bonesini - 13th IPRD conference
18
Custom amplifier
Amplifier:
• Custom made (INFN Pv)
• 1 or 2 channels
• Gain up to 100X (30X with pole
zero suppression)
• Input dynamic range: 0-70 mV
• Bandwith : 600 MHz
This may limit timing response,
tests will be redone soon with a 50x
PLS 774 amplifier (bandwith ~1.50
GHZ)
M. Bonesini - 13th IPRD conference
19
Amplifier performances
Vout
Vin
100X amplifier
Amplifier linearity
M. Bonesini - 13th IPRD conference
20
SiPMT arrays under test
Available SiPMT arrays use 3x3 mm2 or 4x4 mm2 macro-cells
arranged in 4x4 (or more) arrays.
• SENSL ArraySL-4-30035-CER arrays with 3x3 mm2
macrocells, Vop ~29.5 V
• Hamamatsu S11828-3344 , S12642 arrays with 3x3 mm2
macrocells, Vop~72.5 V
• Advansid FBK/IRST ASD-SiPM3S-4x4T (RGB) arrays with
3x3 mm2 macrocells, Vbkw ~28.5 V
• Advansid FBK/IRST ASD-SiPM4S-4x4T (RGB) arrays with
4x4 mm2 macrocells, Vbkw ~29.2 V
• We plan to extend study to new NUV types, better matched
to scintillator light emission
M. Bonesini - 13th IPRD conference
21
Results with conventional PMTs (as
benchmark)
Very low laser light intensity
(1 MIP or less)
Standard laser light intensity
(2-3 MIP)
Vop = V0+DV
M. Bonesini - 13th IPRD conference
22
SenSL ArraySL-4-30035-CER arrays
SiPMT I-V characterization
(our measure)
• Risetime of SenSL arrays much bigger than one of
Hamamatsu or FBK/IRST
• Preliminary results quite bad , we need further studies with
new blue extended arrays from SenSL to get conclusions
M. Bonesini - 13th IPRD conference
23
Results with FBK/IRST arrays
SiPMT I-V characteristics
(manufacture specs)
M. Bonesini - 13th IPRD conference
24
Results with FBK/IRST SiPMT arrays
Typical difference DTDC
(converted in ps) : t=DTDC/2
Standard light intensity
Vop
M. Bonesini - 13th IPRD conference
t~60 ps at best, but
RGB array !!
25
Results with Hamamatsu S11828-3344
Arrays
SiPMT I-V characterisation
(our characterisation)
M. Bonesini - 13th IPRD conference
26
Results with Hamamatsu S11828 Arrays
Standard light intensity
We foresee soon tests with Hamamatsu
S12642 arrays, TSV package , where
better results may be expected
M. Bonesini - 13th IPRD conference
27
Foreseen improvements
• Higher bandwith amplifiers (bandwidth > 1 GHz)
• Extension of measurements to NUV extended SiPMT arrays
(better matched to scintillator emission lmax ~400 nm)
• Use of better signal cables (eg RG213 instead of RG58)
• See effect of CF discriminators vs LE discriminators (even if
we expect no big change: laser light signal/cosmics signal at
counter center + equalized response of photodetectors)
• Test of fully equipped detectors at BTF
M. Bonesini - 13th IPRD conference
28
Conclusions
SiPMT arrays may be a good repacement for fast
PMTs in scintillator time-of-flight system
Preliminary conclusions show a “comparable” timing
resolution with fast PMTs
Clearly results must be validated by testbeam (one
at BTF is foreseen)
Some optimization may be needed: use of fast (> 1
GHz) amplifiers, NUV SiPMT arrays (instead of RGB
ones) to better match scintillator emission
Acknowledgements: many thanks Mr. F. Chignoli, G.
Stringhini and O. Barnaba for skilful help and work in the setup
installation and laboratory measurements and Dr. Bombonati of
Hamamatsu Italia
M. Bonesini - 13th IPRD conference
29
Backup slides
M. Bonesini - 13th IPRD conference
30
Estimation of Npe
• If QADC P.H. distribution is fully described by
photoelectron statistics : Npe=(<R>/R)2 where peak
value R and sigma R are obtained with a gaussian fit.
• From published data (R. Bertoni et al NIM …) we expect
200-300 pe per MIP; from QADC fit we obtain that our
standard light intensity is equivalent to 2-3 MIP
M. Bonesini - 13th IPRD conference
31
Timing resolution as function of
discriminator threshold
M. Bonesini - 13th IPRD conference
32
Fine Mesh Photomultiplier Tubes
•
•
•
•
•
Secondary electrons accelerated parallel to the B-field.
Gain with no field: 5 x 10 5 – 10 7
With B=1.0 Tesla: 2 x 104 - 2.5 x 10 5
Prompt risetime and good TTS
Manufactured by Hamamatsu Photonics
Measures at INFN LASA laboratory to study behaviour
in B field (up to 1.2 T ) as respect to gain, rate capability,
timing
R5505
R7761
R5924
Tube diameter
1”
1.5”
2“
No. Of stages
15
19
19
Q.E.at peak
.23
.23
.22
Gain (B=0 T) typ
5.0 x 10
5
1.0 x 10
7
1.0 x 10
7
Gain (B=1 T) typ
1.8 x 10 4
1.5 x 10
5
2.0 x 10
5
Risetime (ns)
1.5
2.1
2.5
TTS (ns)
0.35
0.35
0.44
M. Bonesini - 13th IPRD conference
33