Transcript Document

•
D.Chokheli, V.Grebenyuk, A.Kalinin
Study of SiPM with help of transient processes
induced by light, dark-noise and current pulses.
A SiPM is a solid-state photodetector that is consist of a matrix
of independent micro avalanche photodiodes ( micro-pixels),
that are operated in limited Geiger mode. The pixels are all
connected onto a common substrate.The sizes are about 1*1 mm
square.The thin depletion region is in the upper part of device.
The film resistor connects the micropixel to a conduction grid
on the SiPM top surface.
Spectrum of Light Diod
SiPM
2000
S60
E= 60V
430nm
S= 6,6V
1800
1600
D19600
1400
1200
N
1000
800
600
400
200
0
-200
0
500
Channels
1000
1500
2000
N
Interest to this devices is explained by the its unique properties: an
ability to detect single photons, a high photon detection efficincy,
low operating voltage and others. The signal is created, when the
hitting photon is converted to electron and hole and the multiplied
charge is accumulated on the C. The gain of SiPM G=10^6 as in
PMT, but high single photon resolution in vacuum PMT is not
possible.
The equivalent circuit of SiPM is shown in Fig 1(upper), when the
photodiode is not exited. There are resistor Rq in series with diode
capacitance in every pixel, but load resistor R is common.
.
•
The equivalent circuit of SiPM of course based on the operating
principle of one pixel that was been worked up in the end of the last
century. It’s essential for us now that the quenching resistor do not be
low resistive. It’s about 1 Mom. As a result of this we have to take into
account the capacitance Cq. It’s verysimilary because Rq disposes in
the upper part of SiPM in the vicinity with ears electrodes.
•
•
Down Fig1 shows the case in which only one micropixel at a
time fires i.e. when a single dark-noise pulse is considered. As it seen
the signal in this case is the fired cell and the rest of cells together
with resistor R are the load, but now the existing voltage is dE= EEbr, which has the form of a step. The rise time of the step is small. It
is determined by t=RsCd, where Rs- internal resist of itself avalanch.
Its value t<<1 ns, and we may it neglect in our consideration.
•
Rq
Rq
Rq
E
.
R
Q R (1+P CqRq)
Uout(p) =
---------------------------------------------------------------------------------------
P2 RRq(CdC+CqC+CqCd) +P[R(Cd+C)+Rq(Cd+Cq)]+1
Z= -1/CqRq
P1= - 1/[R(Cd*+C)+Rq*(Cd*+Cq*)]
P2= - 1/R[C+(Cd*Cq*)/(Cd*+Cq*)]
The pulse form of SiPM
Compemsation of the zero by the pole.
• At low R the output signal is the sum of two components A1 and A2.
The former is very short (about T=RCq) and big, the last is long and
small, but with increasing R it quickly rises. It is may to show when
RqCq=RC the output pulse is getting one-exponent form. It is result
of the compensation of the zero by the pole. The rise time in this
case is nil as before.( we suppose, that the avalanche rise time is
equal zero).
At the further increasing of the load resistor R the signal form gets
• two components again, but the rise time =RC and the pulse
duration is increased.
The SiPM has the unique resolution of single electron (or the
same, of single pixel). It’s result of the outstanding uniformity of the
pixels. What possibilities offer this property? First of all, it allows
the simple calibration and determination of the photon quantity,
hitting on the detector. Besides, the distance between the peaks D
(in channels) is proportional to the SiPM gain G. So G= Q/q=D* Q
1ch/ q= 10^4- 10^6.
The more, it seems, by operative measuring of the distance
between the peaks give an opportunity to restore the gain after any
changing of temperature or others conditions, if only this changing
is in the limited range. Also, by this method the digital gain system
stabilization may be created.
Light yield of SiPM and PMT
10
N pe
on wavelength 375 nm
PMT
8
photoeletrons
6
MP3
4
S60
2
0
3,10
3,15
3,20
3,25
3,30
3,35
3,40
Voltage on the Lightdiode V
3,45
N44&4131
It is the picture from the only publication (NIM A, 572(2007) where the
dark pulse has the clear fast exponent, received with help of fast
amplifier (1,8GHz) and LeQroy oscillograph.
VA-characteristics of SiPM
60
50
nA
MP3
40
Current
30
20
S 60
10
0
0
10
20
30
40
Voltage
50
V
60
Conclusion
•
•
•
•
1.The SiPM equivalent circuit is proposed and the analitical formulas
for the SiPM output signal shape are deduced at different load R.
2.At low R the signal is the sum of two falling exponents, one very
short and high and the other long and small. This signal is preferable
for time measurements.
3.The signal amplitude increases in proportion to increasing R and
when RC=RqCq, the compensation of the zero by the pole take plase
and the pulse gets a one-exponent shape. In this case the pulse
duration may be minimized by introducing the second clipping
compensation in the amplifier. It may be useful for minimizing darknoise, because the gate of Q-COD may be the smallest.
4 As the resistance R goes on increasing, the signal grows long and
the rise time becomes equal to RC. If R is more than 50k, it is possible
to use the charge-sensitive preamplifier.In this case the low thermal
noise allows one to work with small overvoltage.
The noise and the interpixel cross-talk.
The noise of SiPM is result of accident excitation of p-n
junction by thermal electrons and is determined by reverse
current of diode. So, it’s increases with temperature and with
voltage. In reality, the resolution worsening from noise is
dependent from technology. Usually no more, then 15% at the
optimum voltage increasing. The noise pulse count consists
about 10-100 kHz/s.
The specific effect of the interpixel cross-talk is exciting of the
neighbour pixels. It displays the first peaks of spectrum, but at
moderate voltage it’s not marked.
Light yield on green wavelength
4,0
3,5
Npe
MP3
(E=43V)
MP3
(E=42V)
3,0
photoelectrons
2,5
PMT
S60
2,0
1,5
1,0
0,5
0,0
3,20
3,25
3,30
3,35
3,40
Voltage on the light diode
3,45
V
25
PMT H5773
%
20
MPD S60
Q.E.
15
MPD MP3
10
5
0
300
350
400
450
500
550
Wavelenght nm
Quantum Efficiency for PMT and MPDs
600