Transcript Notte_gapd_gsi

SiPM from ST-Microelectronics
Nepomuk Otte & Hector Romo
Santa Cruz Institute for Particle Physics
University of California, Santa Cruz
[email protected]
Main Characteristics (Module H)
•
•
•
3 samples packaged in TO-39 cans
The devices belong to the lot Y745439, Wafer 3
The SiPM has the following characteristics:
 1 mm2 total area (excluding metal pads)
 n-on-p device
 289 (17 x 17) pixels ( 60 µm pitch)
 40 µm single pixel active area side (45% geometrical fill factor)
 Single pixel quenching resistor value about 1.3 MΩ
 Optical insulation to avoid optical cross-talk effects between adjacent pixels
Info from ST Microelectronics
Thanks to Massimo Mazzillo for samples
Layout Mod H
Array Area: 1 x1
mm2 (excluding
metal pads)
Chip size: 4.37 x
4.37 mm2
Anode
Cathode
Measurements with Noise: Gain
Procedure:
Gain
derived from average
single cell amplitude
Temperature: 0°C
ΔQ
assumption of triangular
pulse shape
ΔU
single cell signals:
• 4 ns full width
• symmetric
effective capacitance of single cell:
Ceff=ΔQ/ΔU
~ 15fF
breakdown voltage:
extrapolation to zero gain
Ubreak=29.3 V
Operational range: 30V-40V
>30% above breakdown
Gain/Breakdown/Capacitance vs.
Temperature
@ 35V parameters from linear fit
of gain vs. bias measurements
eff. cell capacitance
uncertainty ~ 5%
breakdown voltage
change of gain 0.5% per 1°C !
0.1% per 1°C
Dark Rates
discriminator set to < single cell signal
Sensor area: 1mm²
Dark rate: 100kHz-1MHz
@ 0°C:
rate doubles every 2 Volts
or
rate doubles if gain
increases by 2·105
Dark Rates vs. Temperature
Gain ~ 500,000
factor 2 change per 5°C
factor 2 change per 12°C
Measurements with Noise:
Optical Crosstalk
direct and indirect (delayed by max 20 ns)
+ extra dark counts
optical crosstalk
1 phe
2phe
naively expect change equal to relative change of gain
but optical crosstalk increases faster
 probably due to increased breakdown probability
 nice task to simulate with SiSi
Optical Crosstalk vs. Temperature
3% optical crosstalk
rise above 20°C can be
explained by
additional darkcounts
e.g.:
2 MHz @ 25°C & 20ns gate
4% probability for
additional dark count
groves !
0.8 µm wide, 8 µm long
IEEE PTL, VOL. 18, NO. 15, 2006
Photon Detection Efficiency
Yeah, if only I would know the PDE.
packaging complicates mounting into our
setup but we will fix this next
probably green sensitive (n-on-p)
IEEE TED, VOL. 55, NO. 10, 2008
Otherwise:
•
•
•
•
good gain vs. temperature dependence (0.5% / °C)
large bias range (30V-40V)
good dark rate behaviour
fast signals
3x3 mm² devices in the pipeline