Transcript Nessun titolo diapositiva - Istituto Nazionale di Fisica

Development of SiPMs
a FBK-irst
C.Piemonte
FBK – Fondazione Bruno Kessler, Trento, Italy
[email protected]
C. Piemonte
1
Outline
• Important parameters of SiPM
• Characteristics of FBK-irst SiPMs
• Application of FBK-irst SiPM
C. Piemonte
General view of the important
parameters in a SiPM
- Gain
- Noise
- Photo-detection efficiency
- Dynamic range
- Time resolution
C. Piemonte
3
Gain
number of carriers produced per photon absorbed
i
~(VBIAS-VBD)/RQ
~exp(-t/RS*CD)
exp(-t/RQ*CD)
t
charge collected per event is the area of the exponential
decay which is determined by circuital elements and bias.
Gain = IMAX*t
(VBIAS-VBD)*t
(VBIAS-VBD)*CD
__Q = ________
__Q = ____________
q
RQ
q
q
C. Piemonte
4
NOISE
1) Primary DARK COUNT
False current pulses triggered by non photogenerated carriers
Main source of carriers: thermal generation in the depleted region.
Critical points: quality of epi silicon; gettering techniques.
2) Afterpulsing:
secondary current pulse caused by a carrier released by a trap
which was filled during the primary event.
3) Optical cross-talk
Excitation of neighboring cells due to the emission of
photons during an avalanche discharge
C. Piemonte
5
Photodetection efficiency
PDE = Npulses / Nphotons = QE x P01x FF
1. QE Quantum efficiency is the probability for a photon to
generate a carrier that reaches the high-field region.
Maximization: anti-reflective coating, drift region location
2. P01. triggering probability probability for a carrier traversing the
high-field to trigger the avalanche.
Maximization: 1. high overvoltage
2. photo-generation in the p-side of the junction
(electrons travel through the high-field region)
3. FF. Fill Factor
“standard” SiPMs suffer from low FF due to the
structures present around each micro-cell
(guard ring, trench)
C. Piemonte
micro-cell
dead width
6
Time resolution
Statistical Fluctuations in the first stages of the current growth:
1. Photo-conversion depth
2. Vertical Build-up at the very beginning of the avalanche
t=0
pair generation
0<t<t1 drift to the high-field region
t>t1
avalanche multiplication
* for short wavelength light
the first contribution is negligible
t1
t’1
single carrier
current level
3. Lateral Propagation
the avalanche spreading is faster if
generation takes place in the center
C. Piemonte
7
FBK-irst SiPMs
Development of SiPMs started in 2005 in
collaboration with INFN.
• IRST:
development of the technology for the production of SiPMs
(large area devices/matrices) + functional characterization
• INFN (Pisa, Bari, Bologna, Perugia, Trento):
development of systems, with optimized read-out electronics,
based on SiPMs for applications such as:
- tracking with scintillating fibers;
- PET;
- TOF;
- calorimetry
C. Piemonte
8
[C. Piemonte
“A new Silicon Photomultiplier
structure for blue light detection”
NIMA 568 (2006) 224-232]
IRST technology
20
n+
7E+05
p
guard region
n+
p
p epi
p+ subst.
Doping conc. (10^) [1/cm^3]
19
6E+05
Doping
Field
18
5E+05
17
4E+05
16
3E+05
15
2E+05
14
1E+05
13
0E+00
0
0.2
0.4
0.6
0.8
depth (um)
High field
region
1
1.2
1.4
Drift region
1) Substrate: p-type epitaxial
2) Very thin n+ layer
3) Polysilicon quenching resistance
4) Anti-reflective coating optimized for l~420nm
C. Piemonte
9
E field (V/cm)
Shallow-Junction SiPM
Layout: from the first design…(2005)
SiPM structure:
- 25x25 cells
- microcell size: 40x40mm2
1mm
1mm
C. Piemonte
Geometry NOT optimized
for maximum PDE
(max fill factor ~ 30%)
10
… to the new devices (i) (2007)
Fill factor: 40x40mm2 => ~ 40%
50x50mm2 => ~ 50%
100x100mm2 => ~ 76%
Geometries:
1x1mm2
C. Piemonte
2x2mm2
3x3mm2 (3600 cells)
4x4mm2 (6400 cells)
11
…to the new devices (ii)
Circular: diameter 1.2mm
diameter 2.8mm
Matrices: 4x4 elements
of 1x1mm2 SiPMs
Linear arrays:
8,16,32 elements of
1x0.25mm2 SiPMs
C. Piemonte
12
Tests performed at FBK
• IV measurement
fast test to verify functionality and uniformity of the properties.
C. Piemonte et al.
“Characterization of the first prototypes
of SiPM fabricated at ITC-irst”
IEEE TNS, February 2007
• Functional characterization in dark
for a complete characterization of the output signal and
noise properties (signal shape, gain, dark count, optical cross-talk, after-pulse)
• Photo-detection efficiency
C. Piemonte
13
Static characteristic (IV)
Matrix 4x4 1-9IV
Reverse
SiPM4 - W12
1.E-05
Breakdown current:
determined by dark events
1.E-06
1.E-07
I [A]
Breakdown voltage
1.E-08
1.E-09
Leakage current: mainly due to
surface generation at the
micro-diode periphery
1.E-10
1.E-11
0
5
10
15
20
Vrev [V]
25
30
35
Very useful fast test. Gives info about:
- Device functionality
- Breakdown voltage
- (Dark rate)x(Gain) uniformity
- Quenching resistance (from forward IV)
C. Piemonte
Performed on
several thousands of
devices at wafer level
14
Signal properties – NO amplifier
Dark signals are exactly equal to photogenerated signals
functional measurements in dark give a complete
picture of the SiPM functioning
Thanks to the large gain it is possible to connect the SiPM
directly to the scope
7.E-03
VBIAS
SiPM: 1x1mm2
Cell: 50x50mm2
SiPM
Digital
Scope
50W
Amplitude (V)
6.E-03
5.E-03
4.E-03
3.E-03
2.E-03
1.E-03
0.E+00
0.0E+00
1.0E-07
2.0E-07
3.0E-07
4.0E-07
Time (s)
C. Piemonte
15
Signal properties – NO amplifier
pedestal.
1p.e. 2
800
700
~ns
Counts
Pulse gen.
Excellent
cell
uniformity
3
600
500
4
400
300
200
100
Laser
0
0
20
40
SiPM
60
80
Charge (a.u.)
100
120
3.5E+06
3.0E+06
Pulse area
= charge
Gain
histogram
collection
2.5E+06
2.0E+06
1.5E+06
Linear
gain
1.0E+06
5.0E+05
0.0E+00
31
C. Piemonte
32
33
34
Bias voltage (V)
35
36
16
Signal properties – with amplifier
A voltage amplifier allows an easier characterization,
but attention must be paid when determining the gain
Pulses at the scope.
VBIAS
SiPM
Av
100x
50W
C. Piemonte
Digital
Scope
s = single
d = double pulses
a = after-pulse
17
Let’s look at the electro-optical characteristics of
these devices:
1x1mm2 (400 cells)
4x4mm2 (6400 cells)
Micro-cell size: 50x50mm2
C. Piemonte
18
1x1mm2 SiPM - 50x50mm2 cell
Set up: SiPM current signal converted into voltage on a 50W
resistor and amplified with a wide-band voltage amplifier.
1.2
1.00
Signal shape
1.0
Amplitude(a.u.)
(a.u.)
Amplitude
Fast transient:
avalanche current
through parasitic
capacitance in
parallel with
quenching res.
0.8
0.6
0.10
0.4
`
= 25C
T =T25C
T = 15C
T = 5C
T = -5C
T = -15C
T = -25C
0.2
0.0
0.01
-1.00E-08
-1.0E-08
4.00E-08
4.0E-08
9.00E-08
9.0E-08
Time
Time (s)
(s)
Slow transient:
Exponential recharge
of the diode capac.
through the
quenching resistor
1.40E-07
1.4E-07
Important to note:
The value of the quenching resistor increases with decreasing
temperature and so the time constant follows the same trend
C. Piemonte
19
1x1mm2 SiPM – 50x50mm2 cell
Dark count
Growing
threshold
1.0E+06
1.0E+03
DC 28
DC 28.5
DC 29
DC 29.5
DC 30
DC 30.5
DC 31
DC 32
DC 33
1.0E+02
-0.70
-0.60
C. Piemonte
Counts
1.0E+05
From this plot we get
idea of dark rate and
optical cross-talk
probability
T = -30C VBD = 27.2V
1.0E+04
-0.50
-0.40
-0.30
Threshold (V)
-0.20
-0.10
20
0.00
1x1mm2 SiPM - 50x50mm2 cell
5.0E+06
3.0E+06
31.5
Breakdown voltage (V)
4.0E+06
Gain
Gain
T = 25C
T = 15C
T = 5C
T = -5C
T = -15C
T = -25C
2.0E+06
1.0E+06
0.0E+00
31
30.5
30
29.5
29
y = 0.0674x + 29.2
28.5
28
27.5
27
27
28
29
30
31
32
33
34
35
-30
-10
10
Temperature (C)
Voltage (V)
25.00
15.00
5.00
-5.00
-15.00
-25.00
1.0E+06
1.E+07
Dark count
Dark count (Hz)
Dark count (Hz)
1.0E+07
30
2V
• 2V DC
overvoltage
3V
• 3V DC
overvoltage
4V
• 4V DC
overvoltage
1.E+06
y = 1E+14e-5.213x
1.E+05
3.2
1.0E+05
27
C. Piemonte
28
29
30
31
32
Voltage (V)
33
34
35
3.4
3.6
3.8
1000/T (1/K)
4.0
21
4.2
4x4mm2 SiPM - 50x50mm2 cell
4x4mm2
3.E+06
Gain
-25C
-15C
Gain
2.E+06
1.E+06
0.E+00
28
29
30
31
32
33
Voltage (V)
Signal shape
1mm2
T = -15C
T = -25C
1.00
0.10
1mm2
0.01
-1.0E-08
Dark count
7.E+06
SiPM
16 x Dark Count
of 1mm2 SiPM
6.E+06
Dark count (Hz)
Amplitude (a.u.)
10.00
5.E+06
4.E+06
-15C
3.E+06
-25C
2.E+06
1.E+06
5.0E-08
1.1E-07 1.7E-07
Time (s)
2.3E-07
0.E+00
28
29
30
31
32
33
Voltage (V)
C. Piemonte
22
4x4mm2 SiPM - 50x50mm2 cell
T=-25C Vbd=27.6V
2
1
3
28.6V
4
5
4
5
6
29.2V
3
2
Charge spectra when
illuminating the device
with short light pulses
Same conclusions as for
the previous device:
1
• Excellent cell response uniformity
4
5
6
7
8
3
29.6V
over the entire device (6400 cells)
Width of peaks dominated by
electronic noise
2
1
-5.E-10
2.E-09
C. Piemonte
4.E-09
Charge (V ns)
6.E-09
8.E-09
23
Photo-detection efficiency (1)
C. Piemonte
24
Photo-detection efficiency (2)
DC curr
with light
DC curr.
wo light
dark pulses
C. Piemonte
light pulses
25
…what is the PDE of these devices?
Measured on 1x1mm2 SiPM using photon counting technique
50x50mm cell - ~50% fill factor
0.35
L=400nm
L=425nm
L=450nm
L=475nm
L=500nm
L=550nm
0.3
PDE
0.25
0.2
0.15
500nm
450nm
425nm
400nm
0.1
0.05
0
30
31
32
33
34
35
Bias voltage (V)
Broad peak between 450 and 600nm
C. Piemonte
26
Time resolution (1)
• Laser: - wavelength: 400 or 800nm
- pulse width: ~60fs
- pulse period: 12.34ns with time jitter <100fs
• Filters: to have less than 1 photodetection/laser pulse
• SiPMs: 3 devices from 2 different batches measured
G. Collazuol, NIMA, 581, 461-464, 2007.
C. Piemonte
27
Time resolution (2)
Distribution of the
time difference
Timing performance (s)
as a function of the
over-voltage
C. Piemonte
28
Microcell functionality measurement
(measurement at RWTH Aachen)
C. Piemonte
29
Microcell functionality measurement
Measurement of microcell
eficiency with a 5 um
LED spot diameter
C. Piemonte
Microcell functionality measurement
C. Piemonte
Some applications and
projects in which we are
involved
C. Piemonte
32
SiPM matrix – for PET (1)
1mm
First, small monolithic matrix of SiPM:
Element 1x1mm2
Micro-cell size: 40x40mm2
IV
curves
of 9 matrices
of one wafer
SiPM4
- W12
Matrix 4x4 1-9
1.E-05
9x16 IV curves
Non working SiPM
1.E-06
I [A]
1.E-07
1.E-08
• Uniform BD voltage
• Uniform dark rate
1.E-09
1.E-10
1.E-11
0
C. Piemonte
5
10
15
20
Vrev [V]
25
30
35
33
SiPM matrix – for PET (2)
Tests are ongoing in Pisa
(DASiPM project, A. Del Guerra)
on these devices coupled with pixellated
and slab LSO scintillators
Na22
spectrum with LSO
on a single SiPM
Res = 18%
(1x1mm2, 40x40mm2 cell)
C. Piemonte
NEXT STEP: Larger monolithic matrices
34
Circular SiPM - 50x50mm2 cell
SiPM for CMS- Outer Hadron calorimeter
for CMS – Outer Hadron Calorimeter
Muon
in YB2
using SiPMs
SiPMs
Muonresponse
response
using
package designed by Kyocera
6mm2 area SiPM
Muon Efficiency in YB2
Baseline
HPD response
at 8 kV
Muon
response
using HPD
at 8kV
Muon Efficiency for 1 kHz noise
100%
module with
95%
18 SiPMs 90%
85%
80%
75%
70%
65%
60%
55%
50%
Each SiPM
reads a bundle
of 5 fibers
0
2.8 mm round IRST
1
2
3
4
SiPM dark current (microAmps/mm^2)
C. Piemonte
35
Array of SiPM for Fiber Tracking
INFN PG (R. Battiston) + Uni Aachen
32x array connected to ASIC designed for strip detectors
=> capacitive divider at the input to reduce signal
Response uniformity under LED illumination
C. Piemonte
36
HYPERimage project
Seventh Framework programme, FP7-HEALTH-2007-A
coordinator
C. Piemonte
37
HYPERimage project
Development of hybrid TOF-PET/MR test systems
with dramatically improved effective sensitivity
First clinical whole body PET/MR investigations of
breast cancer
TOF-PET building blocks
C. Piemonte
38
Conclusion
• The SiPM is going to play a major role as a
detector for low intensity light, because of:
- comparable/better proprieties than PMT;
- the inherent characteristics of a solid-state det..
• IRST has been working on SiPMs (GM-APDs) for about
3 years obtaining very good results in:
- performance;
- reproducibility;
- yield;
- understanding of the device.
C. Piemonte
39