Transcript １６００pixel MPPC 特性の温度依存性

Performance of 1600-pixel MPPC
for the GLD Calorimeter Readout
Jan. 30(Tue.) Korea-Japan Joint Meeting
@ Shinshu Univ.
Takashi Maeda（Univ. of Tsukuba)
for the GLD Calorimeter Group
The Multi-Pixel Photon Counter (MPPC)
…Novel photon sensor that used for GLD calorimeter readout
1600-pixel MPPC
Si Resistor
Guard ring
The Multi-Pixel Photon Counter (MPPC)


20 improved samples in last October
and
400 samples in last December.
Very compact plastic package
for Beam Test @ DESY
4 mm


Old Sample
Can package
---Improved Point--•Higher Gain
•Lower Noise rate
•Package becomes compact
•etc…
3 mm
1.3 mm
1 x 1 mm
Pulse Shape
1 p.e.
2 p.e.
Measurement of Basic Characteristics

Evaluate 1600-pixel MPPC performance as a
function of Bias Voltage and Temperature
– Gain, Noise rate, Cross-talk, P.D.E.
Set up
Blue LED
MPPC
Thermostatic chamber
Gain
•30oC
•25oC
•20oC
•15oC
•10oC
•0oC
•-20oC
Pedestal peak
d
1 p.e. peak
2 p.e. peak
S : ADC Sensitivity
= 0.25 pC/ADCcount
A : Amp gain = 63
S d
e : electron charge Gain 
= 1.6 x10-19 C
A e
C
Gain  (VBias  Vo )
e
C : Pixel Capacitance
V0: Breakdown voltage
V0 with Temperature Variation

 DV = VBias – V0(T) is
sensitive to temperature
– Most of MPPC
performances are
affected by temperature
change
DV0/DT = (56.0 ± 0.1) mV/oC
Gain 
C
(VBias  Vo )
e
C : Pixel Capacitance
V0: Breakdown voltage
V0 is linear to
temperature

Must be improved
Noise Rate

Dark noise :
Avalanche amplification
by thermal electron
0.5 p.e.
Threshold
•
•
•
•
•
•
•
30 oC
25 oC
20 oC
15 oC
10 oC
0 oC
-20 oC
Threshold
0.5 p.e.
1.5 p.e.
Threshold
Noise rate is lower in lower DV( = Vbias – V0) and temperature
Cross-talk Probability


Cross-talk :
The cross-talk to adjacent
pixels is caused by photons
created in an avalanche.
2 pixels fired signals in dark
noises are
caused by cross-talk
•
•
•
•
•
•
•

30 oC
25 oC
20 oC
15 oC
10 oC
0 oC
-20 oC
Cross-talk probability
is not sensitive to
temperature change
Photon Detection Efficiency (P.D.E)
~ Detection probability for single photon injection
e MPPC  Q.E.  e Geiger  e geom
: e- h+ pair production probability for single photon
injection ( Quantum Efficiency )
 eGeiger : Avalanche amplification probability from single p.e.
 egeom : Fraction of sensitive region in a sensor
( Geometrical Efficiency)
Measurement method
 Compare # of p.e. of MPPC with # of p.e. of PMT (Reference)

Q.E.
e MPPC 
N
MPPC
p .e .
PMT
PMT
p .e. ~ 16 %
N
e
MPPC
PMT
WLSF LED
0.5 mm f Pin-hole
P.D.E. Result
~ 6.3 %
P.D.E. of PMT
6.3 % uncertainty comes from estimation of PMT’s P.D.E
Summary
 We are evaluating 1600-pixel MPPC characteristic for the
GLD calorimeter readout
 Gain, Noise rate are sufficient for our requirement
 Breakdown voltage is sensitive to temperature change
– Have to monitor the temperature
Photon Detection Efficiency is higher than PMT
Plans
 Response curve (Input light-yield vs. Output signal)
 Evaluate Uniformity in the sensor
 Measure long-term stability
 Figure out radiation damage effect and magnetic field
stability

Back up…
Old sample results - Gain
•30oC
•25oC
•20oC
•15oC
•10oC
•0oC
•-20oC
V0=aT+b
a = (5.67 ± 0.03) x10-2 V/oC
b = 66.2 ± 0.1 V
Old sample results – Noise rate
•30oC
•25oC
•20oC
•15oC
•10oC
•0oC
•-20oC
Vbias – V0(T) [V]
Old sample results – Cross-talk
The cross-talk to adjacent pixels
is caused by photons created in
an avalanche.
・30℃
・25℃
・20℃
・15℃
・10℃
・0℃
・-20℃
Cross-talk probability is
measured from dark noise rates :
Rate( 1.5 p.e.)
Pcrosstalk 
Rate( 0.5 p.e.)
Vbias – V0(T) [V]
Cross-talk
probability looks stable
with temperature in
Vbias – V0 < 2.5V.
Set up
PC
Gate
Clock
Generator
Generator
Gate
Signal
Delay
input
AMP *63
Green
MPPC
WLSF
Voltage
Voltage
source
Stage
PMT
HV
LED
Thermo-static chamber
source
Light yeild measurement
( with noise and cross-talk subtraction)
Measure light yeild of
LED light pulse
 Fit ADC distributon
 Supposed signals are
dominated by Poisson
statistics
 Count number of events
below 0.5 p.e. threshold
( both LED on and off )

0.5 p.e.
pedestal threshold
Events
0.5 p.e.
threshold
pedestal
Events
Calculation of Np.e.
• f(n,μ) is Poisson distribution function
• μ is Expectated number of Np.e.
• f(0,μLED on), f(0,μLED off ) are probability of 0 p.e.
off / NLED off = e-μnoise
f(0,μnoise) = NLED
all
pedestal
on / NLED on =e-μLED on
f(0,μLED on) = NLED
pedestal
all
 f(0,μLED on) = f(0, μ+μnoise)
= f(0,μ) × f(0,μnoise)
f(0,μ) = f(0,μLED on) / f(0,μnoise) = e-μ
μ = -ln( f(0,μ) )
NpeMPPC / NpePMT (Npe ratio)
WLS Fiber Y-11
Reference : JLCストリップ・ファイバー電磁カロリメータの性能研究
Katsumi Sekiguchi March. 2003
QE of H1161GS
Mean of QE for 1 p.e.
QE × relative light yeild on each wave length
∑ （ QE × relative light yeild on each wave length ）
= Mean of QE for 1 p.e. ~16.7 %
Response / Correction curves ( with small cross-talk )
R (p;Npe)
R-1(p;Nfired)
p=0.1
p=0.1
p=0
(no cross-talk)
Response curve
Correction curve