Transcript MPPC Study for GLD Carolimeter

The MPPC Study
for the GLD Calorimeter Readout
2006/10/31
Takashi Maeda
Institute of Physics, University of Tsukuba
for KEK-DTP photon sensor group
for the GLD Calorimeter group
• Introduction
• Measurement of basic characteristics
– Gain, Noise Rate, Cross-talk
• Measurement of uniformity
with microscopic laser
• Summary and plans
GLD (Global Large Detector) Calorimeter
… a candidate detector for ILC (International Linear Collider)
• Sampling calorimeter with Pb/W - scintillator
sandwich structure with WLSF readout
• Particle Flow Algorithm (PFA) needs particle
separation in the calorimeter
EM-scintillator-layer model
absorber plate
1 cm x 5cm x 2 mm
1 cm x 5cm x 2 mm
• Fine granularity with strip/tile scintillators
• Huge number of readout channels
– ~10M (ECAL) + 4M (HCAL) !
• Used inside 3 Tesla solenoid
Need a new photon sensor
which is compact and low-cost,
can operate in a strong magnetic field
particles
readout
The Multi-Pixel Photon Counter (MPPC)
…novel photon sensor being developed by Hamamatsu Photonics (HPK)
400 pixels
~ 1 mm
20~100 mm
~ 8 mm
Depletion
region
~ 2 mm
Substrate
Si Resistor
Bias voltage (~70V)
Guard ring n+
p+ n
Al conductor
p-
substrate p+
1600 pixels
Requirements for the GLD Calorimeter
• Gain: ~ at least 105, preferably 106
• Dynamic range:
up to ~1000 p.e. (need > 2500 pixels)
– to measure EM shower maximum
• Single Photon Detection Efficiency: ~ 30 %
– to identify MIP signals
• Noise rate : < 1 MHz (threshold = 0.5 p.e.)
• Good uniformity, small cross-talk
• Timing Resolution ~ 1 nsec
• Sensor area: 1.5 x 1.5 mm2
– to place a larger number of pixels
• Should be stable against bias voltage / temperature /
time
Characteristics of the 1600-pixel MPPC
• Evaluate performance as a function of bias voltage
– Gain, Noise Rate, Cross-talk probability
– Photon Detection Efficiency, Linearity
(measurements still ongoing)
• Temperature dependence is also measured
– MPPC performance is known to be sensitive to temperature
Green LED
MPPC
Thermostatic Chamber
Gain
measurement
70V, 20℃
d
2 pix. fired
Pedestal
1 pix. fired
S d
Gain 
A e
S : ADC sensitivity
= 0.25 pC/ADCcount
A : Amp gain = 63
e : electron charge
= 1.6 x10-19 C
C
Gain  (VBias  Vo )
e
C : Pixel capacitance
V0: Geiger-mode starting voltage
・30℃
・25℃
・20℃
・15℃
・10℃
・0℃
・-20℃
C, V0 vs. Temperature
• C looks not sensitive to temperature, at least under < 20oC
• V0 is linear to temperature
V0=aT+b
a = (5.67 ± 0.03) x10-2 V/oC
b = 66.2V±
V +b
= aT
0 0.1
Noise Rate
… rate of avalanche signals
induced by thermal electrons
1MHz
・30℃
・25℃
・20℃
・15℃
・10℃
・0℃
・-20℃
Vbias – V0(T) [V]
Lower temperature a Lower noise rate
Cross-talk
The cross-talk to adjacent pixels
is caused by photons created in
an avalanche.
Cross-talk probability is
measured from dark noise rates :
Rate( 1.5 p.e.)
Pcrosstalk 
Rate( 0.5 p.e.)
・30℃
・25℃
・20℃
・15℃
・10℃
・0℃
・-20℃
Vbias – V0(T) [V]
• Cross-talk probability looks stable with
temperature in Vbias – V0 < 2.5V.
Measurement of uniformity
in the sensor
Using
a microscopic laser system
we perform
• scan within a pixel
• pixel-by-pixel scan
to see the variation of
• Gain
• Hit probability
• Cross-talk
1 pixel
1 pixel
Measurement with
Microscopic Laser System
•
•
•
•
•
•
１６００ pixel MPPC
Introduced by KEK-DTP
YAG Laser,  = 532 nm (green)
Pulse width ~ 2 nsec, rate ~ 8 kHz
Spot size ~ 1 mm
Light yield ~ 0.5 p.e. (not calibrated)
Can perform precise pinpoint scan with
the well-focused laser
~25 mm
Laser spot
Hit fraction vs. Bias Voltage
• Inject laser to center of a pixel.
1 pix. fired
2 pix. fired
(cross-talk)
Hit fraction
Pedestal
N ev ( 0.5 p.e.)
PHit 
N ev (all )
The hit fraction depends on bias voltage,
but is stabilized in Vbias > 70 V.
Hit probability
Uniformity within a Pixel
Bias voltage
・ -71.0V
・ -70.0V
・ -69.5V
・ -69.0V
1 pixel
• Fraction of sensitive region ～20%
• Variation within a sensitive region
~9.2% (RMS)
• The shape of sensitive region is not
changed with bias voltage
Gain Uniformity within a Pixel
Vbias = 70.0 V
y-point (1 mm pitch)
Edge of the sensor
Gain (x105)
x-point (1 mm pitch)
•Higher gain in central part
•Gain variation in a sensitive
region ~ 2.7% (RMS)
Cross-talk Variation within a Pixel
Sensitive region in a pixel
Pedestal
１ pix. fired
2 pix. fired
(cross-talk)
Bias
voltage
・ -71.0V
・ -70.0V
・ -69.5V
・ -69.0V
• Shape of the cross-talk probability
N ev (2 pix.)
PXtalk 
N ev (1 pix.)  N ev (2 pix.) depends on bias voltage
• Edge part shows larger cross-talk
Pixel-by-pixel Scan - Hit fraction
edge of the sensor
Variation ~3.2%
20 x 20 pixels
0.55
Sensor
0.44
Pixel-by-pixel Scan - Gain
edge of the sensor
• Edge pixels have
higher gain
• Strange
structure is seen,
reason unknown
• Variation ~2.4%
3.8 (x105)
3.2(x105)
Summary
• We are evaluating the MPPC performance from
viewpoint of the GLD calorimeter readout use
– Gain, Noise rate, Cross-talk are acceptable
• The MPPC properties are sensitive to Vbias-V0(T)
and temperature
– Lower Noise rate and Cross-talk with lower temperature
• The MPPC properties are observed to be uniform
within a sensor.
Plans
• Measure photon detection efficiency and Linearity
• Perform same measurements for new MPPC samples
and evaluate device-by-device variation
(We just have been provided new samples by HPK)