1
IIII
Upgrade of the
MEG liquid xenon calorimeter
with VUV-light sensitive
large area SiPMs
Kei Ieki for the MEG-II collaboration
Contents
2


MEG and MEG II
Upgrade of liquid Xe detector
 VUV-sensitive
MPPC
 Signal transmission


MPPC performance tests
Simulated dector performance
MEG: μeγ search
3
LFV decay μeγ is a
good probe for new physics!
γ
52.8
MeV
History of
μeγ upper limits
μ+
52.8
MeV
γ
e+
𝜇
μ+
𝑒
e+
BSM (e.g. SUSY)
Upper limit (MEG):
5.7×10-13
New physics
MEG II goal:
prediction
4×10-14
Signal and background
4
Main BG
Signal
γ
μ+
e+
Same energy (52.8MeV)
Same timing
Back to back
from μeννγ
or e+e-γγ
Accidental
coincidence
γ
+
e+from μeνν
Energy, position and timing resolution are important!
The MEG experiment
5
Liquid Xe
γ-ray detector
γ
e+
e+ drift chamber
& timing counter
Gradient magnetic field
μ+ beam
The MEG experiment
6
Liquid Xe
γ-ray detector
900l liquid Xe
• large & homogeneous
• High light yield (~75% of NaI)
• Fast signal (τdecay~45ns)
• Short radiation length (2.8cm)
γ
e+
e+ drift chamber
& timing counter
μ+
beam
Successfully operated in 2009~2013
Upgrade of the Xe detector
7
Resolution was limited by the non-uniformity of
the photon collection efficiency.
 Replace the PMTs at the γ incident face to
12x12 mm2 MPPCs.
Upgrade!
2’’ PMT
× 216
12x12mm2
MPPC × 4096
(CG)
Upgrade of the Xe detector
8
PMT
MPPC+PMT
Imaging power will be significantly improved.
 Better energy & position resolution
The layout of the PMTs will also change.
 less energy leakage, better uniformity
VUV-sensitive MPPC
9
We have successfully developed VUV-MPPC in collaboration with
Hamamatsu Photonics. K.K.
● Large area (12x12 mm2)
● Sensitive to VUV-light
 signal tail become long due to
 Protection coating is removed,
large capacitance.
VUV-transparent quartz window
Reduce capacitance by
is used for protection.
connecting 4 chips in series.
- 50μm pitch pixel
- 4 independent chips
- metal quench resister
or
4-segments
2.5mm
Hamamatsu S10943-3186(X)
2-segments
Signal transmission
10
feedthrough
PCB
inside
chamber
DAQ
(WaveDream)
outside
chamber
• MPPC chips are connected in series on PCB.
• Co-axial like structure PCB and feed-through.
 good shielding, high bandwidth, small crosstalk (<0.3%)
“co-axial like structure”
MPPC performance tests
11



Small sample test in liquid Xe
Mass test in room temperature
Mass test in liquid Xe
MPPC small sample test
12
Several performance tests have been done in
2 litter Xe chamber.
Measurement of MPPC properties with
• LED  Gain, crosstalk, afterpulse etc.
• α source  PDE, energy resolution
300
In liquid Xe
0p.e.
1p.e.
250
example of charge
distribution with LED
200
MPPC
LED
150
100
50
0
α source (241Am)
on wire
-0.01
0
0.01
0.02
0.03
0.04
0.05
example of LED signal
100ns
0.06
0.07
0.08
Charge
MPPC small sample test
13
Basic performance of final model MPPC @ Vover~7V (4 segment series connection):
Gain: 8x105, Crosstalk+Afterpulse probability: 25%, Signal decay time: 30ns
PDE: 16~25% (large uncertainty due to geometry of the setup)
PDE vs Over Voltage
Energy Resolution vs Photon Statistics
PDE vs. over voltage
Energy resolution vs. Np.e.
4´10-1
3´10-1
Energy Resolution
PDE
0.35
0.3
2´10-1
0.25
0.2
0.15
10-1
4´10-2
3´10-2
0.1
I2
L1
K1
chip 488 for LProto
August, 2MPPC(Even-odd)
August, 4MPPC(Even-odd)
Micro Prototype, 2MPPC(Even-odd)
Micro Prototype, 4MPPC(Even-odd)
2´10-2
0.05
0
0
10-2
1
2
3
4
5
6
7
Over Voltage (V)
30
102 2´102
VUV sensitivity and short decay time are confirmed.
10 2´10
3
3
104
# of p.e.
Mass test in room temperature
14
LED
Temperature
controlled chamber
• 600 prototype MPPCs have been tested
in room temperature.
• Basic properties at T=20 deg are
measured for each chip
• No dead channel found.
• Same setup is being used for the test of
final model MPPCs (4000 pcs).
Readout PCB
relays to
change readout ch
MPPC
insertion/removal
tools
Mass test in liquid Xe
15
• 600 prototype MPPCs are tested in liquid Xe
• Goal: Test all the readout system (PCB etc.)
feed-through
in LXe
α sources
241Am
MPPC & PCB
LED
PCBs
(12+12)
Mass test in liquid Xe
16
• 600 prototype MPPCs are tested in liquid Xe
• Goal: Test all the readout system (PCB etc.)
feed-through
in LXe
MPPC & PCB
Mass test in liquid Xe
17
• 600 prototype MPPCs are tested in liquid Xe
• Goal: Test all the readout system (PCB etc.)
feed-through
in LXe
MPPC & PCB
Mass test in liquid Xe
18
Example of α event for one PCB
α
Example of LED
charge distribution
• Response to the LED light and α scintillation signal are
measured for all PCBs.
• We confirmed that MPPC, PCB and feed-through
work OK in liquid Xe.
• There were ~5% of bad channels found. Most of those were caused by bad
connections of MPPCs and cables at PCB and feed-through.
 Assembly procedure and the design of the connectors and
boards will improve.
reduce density
of connectors
direct soldering
of cables
Expected detector performance
19
• Simulation based on measured properties
of MPPCs.
• Waveform of 1p.e. signal
• PDE, gain
• Crosstalk & afterpulse
• Reconstruction algorithm has been optimized
to exploit the advantages of MPPC.
w
u
v
Expected detector performance
20
t(rec) - t(MC)
Energy resolution
Timing resolution
depth > 2cm
depth < 2cm
1.7%
2.4%
Entries
Mean
RMS
Constant
Mean
Sigma
250
200
150
1.0%
1.1%
h0
7277
1.126e-11
7.372e-11
224.8
3.513e-12
5.965e-11
σ=60ps
100
50
48
50
52
54
56 58
E [MeV]
Position resolution
48
50
52
54
56 58
E [MeV]
-9
0
-0.5
-0.4
-0.3
-0.2
- 0.1
-300 -200 -100 0 100 200 300
0
0.1
0.2
0.3
0.4
´10
0.5
time(s)
trec – tMC [ps]
Resolution
MEGI
MEGII
u (mm)
5
2.4
v (mm)
5
2.2
w (mm)
6
3.1
Eγ (w<2cm)
2.4%
1.1%
Eγ (w>2cm)
1.7%
1.0%
tγ (ps)
67
60
Position & energy
resolution improve
by a factor of 2!
Also, γ det. efficiency
63%69%
thanks to less material
Summary & prospect
21




MEG liquid Xe calorimeter will be upgraded by replacing
some PMTs to VUV-sensitive MPPCs.
MPPC, PCB and feed-through have been successfully
developed and tested in liquid Xe.
Thanks to the high granularity, the position and the energy
resolution will improve by a factor of 2.
Mass test of final model MPPCs is ongoing. Construction of the
detector will start this autumn.
22
Backup slides
Mass test results
23
1p.e. charge @ Vover=3.0V
1p.e. pulse height @ Vover=3.0V
Breakdown voltage
Noise rate @ Vover=3.0V
CTAP @ Vover=3.0V
Bad channels
24
~1% of the channels had problems in th PCB and cable connecors.
Bad soldering of chips
Connector too hard
Imperfect isolation of pin
Series connections
25
100ns
parallel
Fall Time
No Segment
135ns
×2
Segmented
49ns
×4
Segmented
25ns
2 segmented
4 segmented