Study on Saturation of SiPM for scintillator calorimeter ... · “Study on Saturation of SiPM for scintillator calorimeter using UV laser”, CHEF2019 at Fukuoka Naoki Tsuji, The

/19

Study on Saturation of SiPM for scintillator calorimeter using UV laser

Naoki Tsuji, The University of Tokyo Linghui Liu, Wataru Ootani, Kosuke Yoshioka, Makoto Gonokami, Yusuke Morita

CHEF2019 at Fukuoka, 25-29 November 2019

1

/19

Naoki Tsuji, The University of Tokyo“Study on Saturation of SiPM for scintillator calorimeter using UV laser”, CHEF2019 at Fukuoka

International linear Collider (ILC)

Future high energy frontier machine Electron-positron collider

Lepton, elementary particle Low background → precise measurement

Ecm : 250 - 500 GeV Initial plan : 250 GeV Extendable to 1 TeV

Search for new physics Precise measurement of Higgs Precise measurement of top quark Search for dark matter

2

/19


International Large Detector (ILD)

ILD consists of vertex detector, central tracker, EM calorimeter, hadron calorimeter, and muon tracker Particle Flow Algorithm (PFA)

Each particle is detected by the best suited detectors Improved jet energy resolution

It is necessary to classify charged particle and neutral particle precisely ➡High granularity calorimeter is required.

3

/19


Sc-ECAL

Scintillator Electromagnetic CALorimeter (Sc-ECAL) Technology option of EM calorimeter for ILD

Based on scintillator strips readout by SiPM 5 × 45 × 2 mm3 scintillator strip

Virtual segmentation : 5mm × 5mm with strips in x-y configuration Timing resolution < 1 ns Low cost

Silicon PhotoMultiplier (SiPM) Made up of multiple APD pixels operated in Geiger mode Excellent photon-counting capability Small size Low bias operation

4Hamamatsu Photonics K.K., Opto-semiconductor Handbook

/19


Saturation of SiPMs

SiPM saturation can be an issue for Sc-ECAL Pile-up hits in small and dense EM shower When a large number of photons are injected to SiPM, output of SiPM can be saturated due to limited number of pixels

Saturation curve is usually measured by direct injecting fast laser pulse (~400 nm) to SiPM

Time constant of emission of scintillation light (few ns) is not negligible compared to recovery time of SiPM cell (dozens ns)

Our idea is to measure the SiPM saturation with scintillation light excited by UV pulse laser The measured saturation curve can directly be used for saturation correction in Sc-ECAL

5

Num

ber o

f excite

d pixels (N

det)

Number of incident photons (Nseed)

Hamamatsu Photonics K.K., Opto-semiconductor Handbook

/19


SiPMs for Sc-ECAL

Hamamatsu MPPC S12571 series Small pixel size : 10 / 15 μm Active area : 1 × 1 mm2 Breakdown voltage : 65 V No Trench isolation

Hamamatsu MPPC S14160 series Small pixel size : 10 / 15 μm Active area : 1.3 × 1.3 / 3 × 3 mm2 Breakdown voltage : 38 V 0.5 μm trench isolation → low crosstalk

6

Hamamatsu Photonics K. K., PD18

Trench : Separation between adjacent pixels in order to reduce crosstalk

Crosstalk : IR photons from avalanche in fired pixel can

trigger adjacent pixels S12571-015P

Overvoltage (V)

Cros

stalk prob

ability (%

)

MPPC : Product name of Hamamatsu SiPM

/19


Trench technique of new MPPC

MPPC S14160 series employ trench technique

Low crosstalk Low operation voltage No reduction of fill factor

Longer tail due to larger cell capacitance

Longer recovery time

Saturation is improved for new MPPC?

Low crosstalk → saturation ↓ Longer recovery → saturation ↑

7

Hamamatsu Photonics K. K., PD18

Provided by Hamamatsu PhotonicsS14160 (w/ trench)

S12572 (w/o trench)

/19


Experimental setup

Laser 190 nm laser : Scintillation excitation, invisible to MPPC 470 nm laser : No scintillation excitation, directly detected by MPPC

Setup for Sc-ECAL 5 × 45 × 2 mm3 scintillator strip (EJ-212) with center dimple MPPC w/o trench : S12571-015P

Active area : 1.0 × 1.0 mm2 Pixel pitch : 15 μm 4489 pixels

MPPC w/ trench : S14160-1315PS Active area : 1.3 × 1.3 mm2 Pixel pitch : 15 μm 7296 pixels

8

Injection hole

MPPC

Scintillator stripSetup

(wrapped strip + PCB)

Cavity

/19


Experimental setup

Excite scintillation light with fast fsec UV pulse laser Laser light is split using half mirror (to scintillator, to photodiode) Incident light intensity is monitored with photodiode MPPC S12571-015P with over voltage of +4 V (Recommended voltage by Hamamatsu) MPPC S14160-1315PS with over voltage of +4 V ( 〃 ) Signal attenuations (10 - 40 dB) used to avoid saturation of electronics

9

Laser

Waveform digitizer

Photodiode

HV

Attenuator

Amp

Variable ND filters

Prism

Al mirror

Half mirror

Scintillator

MPPC

DelayTrigger signal

(S12689-02)

/19


Experiment setup

10

UV laser

Waveform digitizer

Photodiode

HV

Attenuator

Amp

Variable ND filters

Prism

Al mirror

Half mirror

Scintillator

MPPC

DelayTrigger signal

Cut off contamination of other wavelength light

Incident light intensity controlled with three ND filtersPhotodiode

Half mirror

Scintillator & PCB

Laser Light is split using half mirror

/19


Signal waveform

Waveform is compared among 190 nm laser, 470 nm laser, and Sr90 190 nm laser, Sr90 → MPPC detects scintillation light 470 nm laser → MPPC detects laser light directly

Almost the same waveform b/w 190 nm and Sr90 Faster signal for direct injection of 470 nm laser ➡Suggesting that injected UV laser really excites scintillation light !

11

190 nm laser470 nm laser

Sr90

190 nm laser470 nm laser

Sr90

S12571-015P (MPPC w/o trench) S14160-1315PS (MPPC w/ trench)

/19


Analysis

Digitized waveform is integrated to estimate charge. The charge is then converted into number of photoelectrons being divided by single photoelectron charge.

Single photoelectron charge is obtained from dark noise signal found in off-time region

12

h2Entries 42862Mean 0.0002862Std Dev 0.003065

0.02− 0.01− 0 0.01 0.02 0.03 0.04Single charge [10^10e]

1

10

210

310

410

Even

t num

ber h2

Entries 42862Mean 0.0002862Std Dev 0.003065

MPPC chargeIntegration range (100 ns)

Integration range (100 ns)

S12571-015P(MPPC w/o trench)

S14160-1315PS (MPPC w/ trench) Pedestal

Single p.e.

Signal of scintillation light

Signal of scintillation light

chargeHEntries 44629Mean 0.2499Std Dev 0.04832

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5Single charge [10^10e]

0

500

1000

1500

2000

2500

3000

3500

Even

t num

ber chargeH

Entries 44629Mean 0.2499Std Dev 0.04832

MPPC charge

Charge value

Gain value

/19


0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 29−10×

current[A]0

20

40

60

80

100

120

140

N_d

etec

ted[

p.e.

]

UV scan without attenuatorUV scan without attenuator

0 0.2 0.4 0.6 0.8 1 1.2 1.49−10×

current[A]0

20

40

60

80

100

120

140

160

180

N_d

etec

ted[

p.e.

]

UV scan without attenuatorUV scan without attenuator

Laser intensity

Incident light intensity is monitored with photodiode Laser light is split using half mirror

Photodiode current is converted to Nseed using calibration constant obtained at low light intensity where no saturation is anticipated

(Nseed : number of photoelectrons when assuming no saturation) Effect of crosstalk and after-pulse is not corrected yet

13

Linear region of S12571-015P (MPPC w/o trench) Linear region of S14160-1315PS (MPPC w/ trench)

/19


0 5000 10000 15000 20000 25000 30000N_seed[p.e.]

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

N_d

etec

ted[

p.e.

]

UV scan of S12571-015PUV scan of S12571-015P

0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000N_seed[p.e.]

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

N_d

etec

ted[

p.e.

]

UV scan of S14160-1315PSUV scan of S14160-1315PS

Saturation curve with 190 nm laser

Over-saturation is observed for both MPPCs

N.B. still some uncertainties in estimation for signal attenuation factors We observed that the attenuation factor depends on light intensity Probably due to change in signal shape caused by MPPC saturation

14

0dB10dB20dB

40dB30dB

Npixel = 4489

Npixel = 7296

0dB10dB20dB

40dB30dB

PreliminaryPreliminary

UV scan of S12571-015P (MPPC w/o trench) UV scan of S14160-1315PS (MPPC w/ trench)

/19


0 5000 10000 15000 20000 25000 30000N_seed[p.e.]

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

N_d

etec

ted[

p.e.

]

470 nm laser scan470 nm laser scan

0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000N_seed[p.e.]

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

N_d

etec

ted[

p.e.

]

470 nm laser scan470 nm laser scan

Saturation curve with 470 nm laser

Saturation curve using 470 nm laser is measured to compare with 190 nm laser Stronger saturation compared to 190 nm laser is observed Over-saturation is still observed for both MPPCs

Due to after pulse and delayed crosstalk?

N.B. still some uncertainties in estimation for signal attenuation factors

15

0dB10dB20dB

40dB30dB

Npixel = 4489

Npixel = 7296

0dB10dB20dB

40dB30dB


470 nm scan of S12571-015P (MPPC w/o trench) 470 nm scan of S14160-1315PS (MPPC w/ trench)

/19


0 5000 10000 15000 20000 25000 30000N_seed[p.e.]

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

N_d

etec

ted[

p.e.

]

Comparison with 190nm and 470nm laserComparison with 190nm and 470nm laser

0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000N_seed[p.e.]

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

N_d

etec

ted[

p.e.

]

Comparison with 190nm and 470nm laserComparison with 190nm and 470nm laser

Comparison of 190 nm and 470 nm laser

Significant difference between saturation curves with 190 nm and 470 nm laser The effect of time constant of scintillation light emission is observed

Can be a big impact on saturation correction

16


Comparison of S12571-015P (MPPC w/o trench) Comparison of S14160-1315PS (MPPC w/ trench)

190 nm laser470 nm laser 190 nm laser

470 nm laser

Npixel = 4489

Npixel = 7296

/19


0 10 20 30 40 50 60 70 80 90 100310×

N_seed[p.e.]/area/PDE/(1+CT)0

5000

10000

15000

20000

25000

30000

N_d

etec

ted[

p.e.

]/are

a/(1

+CT)

Comparison of MPPCs with 190 nm laserComparison of MPPCs with 190 nm laser

Comparison of two MPPCs with 190 nm laser

Normalized by sensor area, PDE and crosstalk probability to compare saturation curves between two MPPCs MPPC w/ trench is less saturated compared to MPPC w/o trench

Few plots above the linear function (because of wrong attenuation factor?) [For S14160-1315PS (w/ trench)] (Lower crosstalk → saturation ↓) ＞ (Longer recovery time → saturation ↑)

Effect of longer recovery time is small because of scintillation emission time

17

S14160-1315PS (w/ trench)

S12571-015P (w/o trench)

Preliminary

Number of photons [p.e./mm2]

Num

ber o

f detec

ted ph

oton

s [p.e./m

m2 ] Comparison with two types of MPPCs (S14160-1315PS and 12571‒015P)

/19


0 10 20 30 40 50 60 70 80 90 100310×


5000

10000

15000

20000

25000

30000

N_d

etec

ted[

p.e.

]/are

a/(1

+CT)

Comparison of MPPCs with 470 nm laserComparison of MPPCs with 470 nm laser

Comparison of two MPPCs with 470 nm laser

S12571-015P (w/o trench) is less saturated compared to S14160-1315PS (w/ trench) [For S14160-1315PS (w/ trench)] (Lower crosstalk → saturation ↓) ＜ (Longer recovery time → saturation ↑)

Effect of longer recovery time is large because of short duration of laser pulse

18Nphoton[p.e.] = Nseed/area/PDE/(1+PCT)

S14160-1315PS (w/ trench)

S12571-015P (w/o trench)

Nphoton[p.e.] = Nseed/area/PDE/(1+PCT)

Nno

r_de

t = N

det/a

rea/PD

E/(1+P

CT) [p.e.]

Comparison with two types of MPPCs (S14160-1315PS and 12571‒015P)

Preliminary

/19


Summary & To do

Saturation curves for MPPCs are measured using scintillation light excited by fast UV-laser Saturation recovery with scintillation light is observed.

It should be taken into account for saturation correction The measured saturation curves can directly be used for correction

Saturation curves are measured for two types of MPPCs for Sc-ECAL (w/ and w/o trench) The effect of longer recovery time of S14160 (MPPC w/ trench) is found small for scintillation light

Longer-tail effect is observed when the measurement of fast pulse

To do Investigate light intensity dependence of attenuation factor Compare with theoretical model of saturation

19

/10

Backup

20

/19


Attenuation factor

Dynamic range of DAQ is not sufficient to cover whole light intensity range

Need signal attenuation to avoid saturation in electronics 10-40 dB attenuator used

There is a frequency dependence of attenuator Attenuation rate depends on input pulse shape

21

Waveform comparison with 0dB vs 10dB (S12571-015P) Waveform comparison with 0dB vs 10dB (S14160-1315PS)

0 dB

10 dB

0 dB

10 dB

MPPC

Waveform digitizer

Attenuator

Amp

HV

/19


Variation of waveform

Variation of waveform depending on light intensity was observed Effect of constant background light at low light intensity

Background light comes from optical setup and laser reflection Faster than scintillation light because directly injected to MPPC

Significant contribution from fast component of background at low light intensity

22

Waveform comparison with 0dB attenuator (S12571-015P) Waveform comparison with 0dB attenuator (S14160-1315PS)

Background

Scintillation

Background

Scintillation

At low light intensity, background is dominant

At higher light intensity, background is negligible

BackgroundScintillation

Background

Scintillation

/19


Variation of waveform

Effect of SiPM saturation at high light intensity Saturation deforms waveform at very high light intensity

These effects change the attenuation rate depending on light intensity Not yet calibrated

23

Waveform comparison with 40dB attenuator (S12571-015P) Waveform comparison with 40dB attenuator (S14160-1315PS)

Not saturated

Saturated

Not saturated

Saturated

/19


4000 6000 8000 10000 12000 14000 16000 18000 20000N_incident[p.e.]

1000

2000

3000

4000

5000

6000

7000

8000

N_d

etec

ted[

p.e.

]

UV scan with 30dB and 40dB attenuatorUV scan with 30dB and 40dB attenuator

0 20 40 60 80 100 120 140 160 180N_incident[p.e.]

0

20

40

60

80

100

120

140

160

N_d

etec

ted[

p.e.

]

UV scan without attenuator and 10dB attenuatorUV scan without attenuator and 10dB attenuator

Attenuation Calibration

Attenuation rate is optimized with factor and offset : Nopt = A * Npe + B Fit with linear function and calculate coefficients to match the lines

Specious factor and offset are used at high light intensity Saturation and waveform variation change the attenuation rate depending on light intensity Fit with polynomial function

24

Comparison with 0dB vs. 10dB (S12571-015P) Comparison with 30dB vs. 40dB (S12571-015P)

0 dB

10 dB

30 dB

40 dB(Attenuation rate) ×3.11

×3.39

×3.45

×3.63

/19


4000 6000 8000 10000 12000 14000 16000 18000 20000N_incident[p.e.]

4000

5000

6000

7000

8000

N_d

etec

ted[

p.e.

]

UV scan with 30dB and 40dB attenuatorUV scan with 30dB and 40dB attenuator

0 20 40 60 80 100 120 140 160 180N_incident[p.e.]

0

20

40

60

80

100

120

140

160

N_d

etec

ted[

p.e.

]


Attenuation Calibration

Attenuation rate is optimized with factor and offset : Nopt = A * Npe + B Fit with linear function and calculate coefficients to match the lines

Specious factor and offset are used at high light intensity Saturation and waveform variation change the attenuation rate depending on light intensity Fit with polynomial function

25

Calibrated results with 0dB vs. 10dB (S12571-015P) Calibrated results with 30dB vs. 40dB (S12571-015P)

0 dB

10 dB

30 dB

40 dB

/19


Comparison of two types of MPPCs

These plots contain the difference of active area, crosstalk probability and PDE between two types of MPPCs Nseed and Ndet are divided by each active area, crosstalk probability, PDE for comparison purpose

Nseed is converted into number of inserted photon (Nphoton) Ndet is converted into normalized number of detected photons (Nnor_det)

Then, we can see directly saturation tendency of each MPPC

26

S12571-015P

Overvoltage (V)

Overvoltage (V)

PDE (%

)Cr

osstalk prob

ability (%

)

/19


270 20 40 60 80 100 120 140 160

N_incident[p.e.]

0

20

40

60

80

100

120

140

160

N_d

etec

ted[

p.e.

]


Waveforms with 0dB attenuator

1 1.2 1.4 1.6 1.8 2 2.2 2.49−10×

current

0.2

0.4

0.6

0.8

1peak

attenuation ratioattenuation ratio

x3.38772

x3.15034

x3.1103

Attenuation factor 0dB vs. 10dB

/19


28

Waveforms with 10dB attenuator Waveforms with 30dB attenuator

Waveforms with 40dB attenuator Waveform transition with 0 - 40dB attenuator

/19


29

1 1.2 1.4 1.6 1.8 2 2.2 2.49−10×

current

0.2

0.4

0.6

0.8

1peak


3 3.5 4 4.5 5 5.5 6 6.5 79−10×

current

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1peak


10 12 14 16 18 20 22 24 26 289−10×

current

0.2

0.4

0.6

0.8

1pea

k


0.05 0.1 0.15 0.2 0.256−10×

current

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1peak


x3.38772

x3.15034

x3.1103

x3.49093

x3.33339

x3.21422

x3.60312

x3.5701

x3.49469

x3.55659

x3.63195

x3.45095

Attenuation factor 0dB vs. 10dB Attenuation factor 10dB vs. 20dB


/19


300 20 40 60 80 100 120 140 160 180

N_incident[p.e.]

0

20

40

60

80

100

120

140

160

180

N_d

etec

ted[

p.e.

]


Waveforms with 0dB attenuator

1 1.2 1.4 1.6 1.8 29−10×

current

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6peak


x3.38914

x3.23162

x3.08691

Attenuation factor 0dB vs. 10dB

/19


31

Waveforms with 10dB attenuator Waveforms with 30dB attenuator

Waveforms with 40dB attenuator Waveform transition with 0 - 40dB attenuator

/19


3 3.5 4 4.5 5 5.5 69−10×

current

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8peak


20 30 40 50 60 70 80 909−10×

current

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8peak


10 12 14 16 18 209−10×

current

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

2.2

peak


1 1.2 1.4 1.6 1.8 29−10×

current

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6peak


32

x3.38914

x3.23162

x3.08691

x3.46571

x3.27096

x3.35106

x3.80956

x3.36394

x3.51893

x3.7942

x3.67675

x3.53735



/19


0 20 40 60 80 100310×


2000

4000

6000

8000

10000

N_d

etec

ted[

p.e.

]/are

a/(1

+CT)

UV comparison of MPPCsUV comparison of MPPCs

Comparison with two types of MPPCs

The difference of two plots should be consistent with the difference of PDE However, S14160-1315PS is less saturated compared with S12571-015P even after correction

Need further verification We plan to do the same measurement with fast 400 nm laser

33

S14160-1315PS (New MPPC w/ trench)

S12571-015P (Standard MPPC w/o trench)

Preliminary

Nphoton[p.e.] = Nseed/area/PDE/(1+PCT)

Nno

r_de

t = N

det/a

rea/PD

E/(1+P

CT) [p.e.]

Documents

Study on Saturation of SiPM for scintillator calorimeter ... · “Study on Saturation of SiPM for scintillator calorimeter using UV laser”, CHEF2019 at Fukuoka Naoki Tsuji, The