This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
ARTICLE IN PRESS
Nuclear Instruments and Methods in Physics Research A 602 (2009) 391–395
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in
Physics Research A
journal homepage: www.elsevier.com/locate/nima
Active control of the gain of a 3 mm 3 mm Silicon PhotoMultiplier
P.S. Marrocchesi a,, M.G. Bagliesi a, K. Batkov a, G. Bigongiari a, M.Y. Kim b, T. Lomtadze b, P. Maestro a,
F. Morsani b, R. Zei a
a
b
Department of Physics, University of Siena and INFN, V. Roma 56, 53100 Siena, Italy
INFN sez. di Pisa, Largo B. Pontecorvo 3, 56127 Pisa, Italy
a r t i c l e in fo
abstract
Article history:
Received 15 December 2008
Accepted 29 December 2008
Available online 10 January 2009
Solid-state photodetectors working in the Geiger mode—known as Silicon PhotoMultipliers (SiPM) or
MultiPixel Photon Counters (MPPC)—are currently being developed with increasingly large active areas.
Potential applications for low-light-level detection were investigated with a 9 mm2 MPPC, recently
made available by Hamamatsu. The device was optically coupled with a scintillator and its
performances, in terms of single photoelectron discrimination, were studied in a series of
measurements under different operating conditions. An active control of the gain was implemented
by a linear feedback on the operating voltage. The results of the tests are discussed.
& 2009 Elsevier B.V. All rights reserved.
The Silicon PhotoMultiplier (SiPM) [1,2] is a novel solid-state
photon counting device. It consists of a matrix of Avalanche
PhotoDiodes operating in the Geiger regime with resistive
quenching and connected in parallel into a single readout
element. Along with a number of appealing features (insensitivity
to magnetic fields, low-voltage bias, ruggedness, small material
budget), this device suffers from a high dark-count rate and a
dependence of the gain on the main operating conditions (bias
voltage and temperature). Nevertheless, its excellent singlephoton detection capabilities [3,4] makes the SiPM a natural
candidate to replace conventional photomultipliers (PMTs) in
many applications including high energy physics and nuclear
physics instrumentation as well as in other disciplines.
In this paper, we describe the measurements carried out in our
laboratory with a Hamamatsu [5] MultiPixel Photon Counter
MPPC-S10362-33-050C. This new photosensor has a 3 mm 3 mm active area covered by 3600 square pixels of 50 mm side.
It provides approximately a 9 times larger detection area with
respect to the 1 mm2 SiPM devices that we have used in our
previous measurements [6].
with a wavelength shifted emission peak in the blue. As the
sensitive surface of the SiPM is protected by a transparent film, an
optical grease was used to ensure a good coupling between the
SiPM window and a 6 mm diameter cylindrical scintillator (T2) of
30 mm length. One side of the scintillator was optically connected
to the SiPM, while the opposite end was coupled to the
photocathode of a conventional PMT via a 3 mm diameter clear
optical fiber.
The signal from the photosensor was matched via a 50 O
impedance to a 3 stage Gali-5 wideband monolithic amplifier. The
output signal was digitized by a CAEN V792 12-bit ADC with a
gate width of 120 ns and a gain of 104 fC per ADC count. The
readout of the module was carried out by a VME controller
interfaced to a PC via an optical-fiber link.
Front-end Board
Ru-106 Source
1mm pin-hole
T1: Thin Scintillator + PMT-1
Clear optical fiber (φ = 3mm)
2. Test setup
The spectral sensitivity of the device (Fig. 1) allows for a direct
optical coupling with plastic scintillators and Cherenkov radiators
PMT-2
T2: Scintillator
(φ = 6mm; Length = 30mm)
Corresponding author. Tel.: +39 0502214363; fax: +39 0502214317.
E-mail address: marrocchesi@pi.infn.it (P.S. Marrocchesi).
0168-9002/$ - see front matter & 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.nima.2008.12.199
Si-PM
3mm x 3mm
Fig. 1. Laboratory tests with a radioactive source: experimental layout.
Fan
1. Introduction
Copper for thermal contact
Peltier Cooling plate
Keywords:
Photodetector
Silicon PhotoMultiplier
MPPC
Gain control
Author's personal copy
ARTICLE IN PRESS
392
P.S. Marrocchesi et al. / Nuclear Instruments and Methods in Physics Research A 602 (2009) 391–395
The scintillator was illuminated, from a direction orthogonal to
the cylinder axis, either with UV light from an LED or with a
radioactive b-source. In this way, we could monitor the intensity
of the scintillating light, using the PMT as a reference, and
measure the relative response and efficiency of the SiPM. The
experimental layout is shown in Fig. 1.
A thin (1.5 mm thick) scintillator T1, readout by a conventional
PMT, was placed between the source collimator and the
scintillator under test. The coincidence between the T1 signal
and the signal of the reference PMT connected to T2 was used as a
trigger for the measurements performed with the electron source.
The latter was a 106 Ru b-emitter with an energy spectrum endpoint at approximately 3.5 MeV.
Both the SiPM and the Gali-5 amplifier were in good thermal
contact with a copper cold mass connected to a Peltier cooler. The
temperature was constantly monitored using a thermistor. During
all tests, the bias voltage and the SiPM dark current were
constantly monitored using a Keithley 487 picoammeter. Temperature and current were readout via a GPIB interface and then
recorded and displayed using LabView. The SiPM digitized data
were readout and displayed by custom data acquisition and
monitoring programs, running under Linux and using the CERN
ROOT package (Fig. 2).
Fig. 3. Measured I2V curves as a function of the temperature T.
3. Static measurements
The MPPC-S10362-33-050C was first characterized by measuring the I2V curve. The measurements were carried out for
different values of the temperature T. During each measurement
the temperature was monitored and kept stable using the Peltier
cooler. The family of curves, shown in Fig. 3, refer to a temperature
range between 8.1 and 20.4 1C.
4. Measurements with a scintillator
In a first series of measurements with the radioactive source,
we were able to observe photopeaks in the pulse height
distribution of the SiPM signals (Fig. 2) as a result of the detection,
by individual cells of the photodetector, of the light generated by
the incident electrons in the cylindrical scintillator.
Data were collected with the source trigger for different values
of the bias voltage Vbias , while the temperature was kept at a
constant value. The pedestal distribution was measured with
random triggers and its mean value was subtracted from the data.
4.1. Measurement of the gain
Using the data collected with the scintillator, the distance of
the first photopeak from the pedestal and the relative distances
among the first three consecutive photopeaks were fitted to
Gaussians. By averaging these values, the gain G of the SiPM was
determined and plotted as a function of the reverse bias voltage
V bias in Fig. 4, where each curve refers to a given value of the
temperature.
By linear extrapolation of each curve of Fig. 4 to zero gain, the
reverse bias breakdown voltage Vbd was determined and plotted
in Fig. 5 as a function of the temperature T.
The linear increase of V bd with the temperature was fitted as
V bd ðTÞ ¼ a0 T þ V 0 with a slope a0 ¼ 50:2 0:1 mV= C and a
constant term V 0 ¼ 66:07 0:02 V.
For a given value of the reverse bias, the gain decreases
approximately by a factor of 2 for an increase of 10 C. The actual
dependence was fitted from the data of Fig. 4 to the expression:
GðTÞ ¼ G0 ðV bias V bd ðTÞ), where G0 ¼ 4:57 105 is the photodetector gain corresponding to an overvoltage DV ¼ V bias V bd
of 1 V.
4.2. Measurement of the dark count rate
Fig. 2. Photopeaks observed with the scintillator: (a) at 10 1C and (b) at 3 1C. Lower
panel: pedestal and first two photopeaks at 3 1C.
Data taken with a random trigger showed a pedestal width
s0 20 ADU. This translates into an rms noise of 8 fC, for an
Author's personal copy
ARTICLE IN PRESS
P.S. Marrocchesi et al. / Nuclear Instruments and Methods in Physics Research A 602 (2009) 391–395
393
amplifier gain close to 250. At a typical SiPM gain of G ¼ 4:0 105 ,
it corresponds to about 0.12 photoelectrons (pe).
For each run, taken at fixed temperature, the values of a lower
threshold (0.5 pe) and of an upper threshold (1.5 pe), were
Fig. 7. Cross-talk probability among adjacent SiPM cells as a function of the
temperature for different values of the gain.
Fig. 4. SiPM gain vs. bias voltage for different temperature values.
determined. The measured dark count rates, corresponding to
either threshold setting, are shown in Fig. 6 as a function of the
temperature, each curve referring to a fixed value of the gain
ranging from 2.5 to 4:0 105 . The dark count rate increases at
larger gain and approximately doubles for a temperature increase
of one decade around room temperature. A typical (lower
threshold) dark count rate of about 2 MHz was found at 23 1C
and G ¼ 4:0 105 .
4.3. Cross-talk probability
The dark count rates corresponding to the 1.5 pe threshold are
about one order of magnitude lower than the rates measured with
the lower threshold. Their ratio gives an estimate of the cross-talk
probability in adjacent photocells of the photosensor. The plot
shown in Fig. 7 shows an increase of the cross-talk at larger gains
and higher temperatures. A typical value between 15% and 20%
was found at 23 1C and G ¼ 4:0 105 .
Fig. 5. Reverse bias breakdown voltage Vbd as a function of the temperature.
5. Operational stability
The temperature dependence of the gain is quite large and can
have a negative impact on pulse height measurements and
seriously impair the photon counting capability of the photodetector, as shown in Fig. 8. In this example, the temperature of
the device was allowed to vary in time, within 5 C, following a
sine wave with a 8 h period. This was achieved by remotely
controlling the Peltier cooling system.
The result, shown for three different values of the temperature,
is that adjacent photopeaks broaden and merge together with a
sizeable loss of information. For stable operations, the gain should
be kept constant and this can be done either by stabilizing the
temperature or by feedback control on the reverse bias voltage.
6. Active control of the gain
Fig. 6. Dark count rates vs. temperature for different values of the gain. The upper
(lower) set of curves refer to a threshold of 0.5 (1.5) photoelectrons.
As an alternative to cool the SiPM and maintain it at a very
stable temperature, we implemented an active control of the gain
by varying the SiPM bias as a function of the temperature. We
could remotely control the bias voltage of the SiPM using a
Keithley power source connected to a PC via a GPIB interface,
Author's personal copy
ARTICLE IN PRESS
394
P.S. Marrocchesi et al. / Nuclear Instruments and Methods in Physics Research A 602 (2009) 391–395
while at the same time monitoring the current. The temperature
sensor was in thermal contact with the SiPM.
We implemented a feedback loop where the measured values
of the temperature were used as input to calculate the appropriate
value of the bias voltage for a given gain, as shown in Fig. 9.
The active control loop was embedded into a LabView virtual
instrument. Almost two complete temperature cycles were
performed during a long stability test run (more than 15 h) with
a maximum temperature span of about 9 1C. The measured ADC
values of the pedestal and the first two photopeaks are shown, as
a function of time, in Fig. 10.
The effect of the active gain control can be seen by noting that
the distance between the curves (i.e.: the gain) is approximately
constant.
The amount of stability achieved with the active control is
shown in Fig. 11, where the measured gain is plotted as a function
of time during the stability test run. Despite the large variation of
the temperature, the average gain was kept stable within 1% at
140
120
100
80
60
40
20
0
5
G = 4.15 x 10
T = 13.5 °C
0
200 400 600 800
220
200
180
160
140
120
100
80
60
40
20
0
G = 3.18 x 105 450
T = 19.0 °C 400
350
300
250
200
150
100
50
0
0 200 400 600 800
5
G = 2.18 x 10
T = 23.0 °C
0 200 400 600 800
700
Fig. 10. Active control of the gain. Upper panel: ADC values for pedestal, first and
second photopeaks as a function of time. Lower panel: temperature vs. time.
600
ADC
500
400
300
200
100
0
0
100
200
300
Time (Minutes)
400
Fig. 8. Effect of gain variation during a 5 C sinusoidal temperature cycle with a
period of 8 h. The upper histograms are pulse height distributions taken at three
different times.
Fig. 11. Active control of the gain: measured gain as a function of time during two
temperature cycles with a maximum temperature span of about 9 1C.
the value G ¼ 3:6 105 . The gain variation during the long run,
plotted as a function of the temperature in Fig. 12, is consistent
with a stability of DG=G within 1%.
7. Temperature dependence of the current at fixed gain
Fig. 9. Bias voltage vs. temperature for given values of the gain.
The temperature dependence of the current was measured, in a
series of short runs, as shown in Fig. 13, where each curve is
relative to a given gain. The curves show a non-linear dependence
on the temperature that is directly related to the increase of the
dark count rate (Fig. 6).
During the stability test run, when the gain was kept fixed
at a value G ¼ 3:6 105 via the active control, the measured
temperature dependence of the photodetector current (Fig. 14)
Author's personal copy
ARTICLE IN PRESS
P.S. Marrocchesi et al. / Nuclear Instruments and Methods in Physics Research A 602 (2009) 391–395
395
showed a non-linear behavior, consistent with the curves of
Fig. 13.
8. Conclusions
We tested the performance of the Hamamatsu MPPC-S1036233-050C photosensor, coupled with a thin plastic scintillator,
under different conditions of gain and temperature. The photon
counting capability of the device, its dark count rate and
operational stability were measured.
The problem of the temperature dependence of the gain—a key
point for stable operations—was addressed by implementing an
active control loop via a temperature dependent (linear) feedback
on the bias voltage. Once kept at a stable gain, the device was
found to provide reproducible measurements over extended
periods of time.
Acknowledgment
Fig. 12. Active control of the gain: measured fractional gain variation as a function
of the temperature during the stability test run (more than 15 h).
This work is part of the R&D program SPIDER funded by the
Istituto Nazionale di Fisica Nucleare (INFN).
References
[1]
[2]
[3]
[4]
[5]
[6]
Fig. 13. Measured current as a function of the temperature for different values of
the gain.
Fig. 14. Active gain control: measured current as a function of the temperature
when the gain was stabilized at the value G ¼ 3:6 105 .
V. Saveliev, V. Golovin, et al., Nucl. Instr. and Meth. A 442 (2000) 223.
P. Buzhan, B. Dolgoshein, et al., Nucl. Instr. and Meth. A 504 (2003) 48.
V. Golovin, et al., Nucl. Instr. and Meth. A 518 (2004) 560.
N. Dinu, et al., Nucl. Instr. and Meth. A 572 (2007) 422.
K. Yamamoto, et al., in: IEEE Nuclear Science Symposium, S. Diego, CA, 2006.
P.S. Marrocchesi et al., in: Proceedings of the 30th Cosmic Ray Conference,
Merida, 2007.