Characterisation of radiation-damaged SiPMs · C-V measurements and $% &,’(,$(- Upper limit $(< 5 fF (~1 fF-pulse shape) - Significant differences for SiPM samples ICASP - Schwetzingen

Characterisation of radiation -damaged SiPMs

1. Aims2. SiPMs + measurements3. Results

3.1 C-V: Electrical parameters and doping profile3.2 Idark : Vbd, DCR 3.3 PHdark : DCR3.4 PHlight : photo-detection

4. Conclusions + next steps

ICASP - Schwetzingen --- Robert Klanner --- 11-15.6.2018 1

S. Cerioli, E. Garutti, R. Klanner, D. Lomidze, S. Martens, J. Schwandt, M. Zvolsky

Hamburg University

Aims + SiPMs investigated

Aim: Develop and test methods to characterise hadron-damaged SiPMs Demonstrate them on a specific SiPM which is the best method? Which parameters change with fluence? physics of damage Improvements? Quantify optimal operating conditions + pulse shapi ng for a given task

SiPM: KETEK MP15 *) with 4384 pixels 15x15 µm 2

n-irradiated at JSI (reactor) to 10 9, 1011, 5⋅1011, 1012, 5⋅1012, 1013, 5⋅1013, 5⋅1014 cm -2


e h MP15

- Depth ampli-region ~ 1µm- Vbd 27.5 V at 20°C- Voff Vbd – 850 mV

- Rq poly-Si

*) SiPM well studied @ UHH (RK. et al., NIM-A 854 (2016) 70)

Measurements


C/G-V: V = 0…26.5V, = 100 Hz…2 MHz T = -30°C and +20 °C

Cse

rial[F

]R

seria

l[Ω

]

Idark : V = -2…+35 (40) V, T = -30°C and +20 °C

°C

12 mA current limit →

used in analysis

Measurements


Current transients: V = 27.7…30.5V, T = -30°C PHdark and PH light -spectra for 15 … 75 ns

PHdark PHlight

lightdark

Φ=1012cm -2

lightdark

Φ=0 Φ=5⋅1013cm -2npe = 1 npe ≈ 100

gate:gate:

used in analysis + used in analysis

Parameters and formulae for analysis


EQF and ENF determined from high statistics Φ=0 data at +20°C for mostcases a small correction

find: EQF ≈ ENF

0 1 2 3 4 5 6 7 81

1.05

1.1

1.15

1.2

1.25

1.3

V - Vbd [V]

EN

F

+20°C

Cpix … pixel capacitanceRq… quenching resistanceCq … quenching capacitance

Vbd … breakdown voltageVoff … Geiger turnoff voltage

DCR … dark count rate

EQF … excess charge factor

ENF … excess noise factor /!

/!

PDE … photon detection efficiencyµ … no. photons initiating Geiger discharge

μ

pulse shape (slow comp.): # $%&⋅'(

) ($%& + $()⋅( )⋅

( … fraction of signal in gate)

-$'⋅⋅)( .)

μ⋅)⋅

μ⋅)!⋅!⋅

3 ways to obtain µ +

)

∙∙0

N.B: valid if no saturation effects !

1 23mV from high statistics Φ=0 dataat +20°C assume 1

does not depend on Φ, T

Vbd versus V off

From PH-fits: )() find )() d)/d⋅( ) with d)/d=const

straight line describes )() from extrapolation to ) .( .)

for 43ns (compatible results for other values ):

5)/d 9.75⋅104 V-1 (compare to: $%&/( 17.8 fF/1.6⋅10-19C = 11.1⋅104 V-1 ≈ ok)

1 850 mV


V [V]

Idar

k[A

] Gain

24 25 26 27 28 29 30 31 32 331 10 14−

×

1 10 13−×

1 10 12−×

1 10 11−×

1 10 10−×

1 10 9−×

1 10 8−×

1 10 7−×

1 10 6−×

1 10 5−×

0

2 105×

4 105×

6 105×

8 105×

Idark-E0Idark-E9Idark-E11Gain-E0Fit

Vbd

Voff

Difference: 1 850 mV to be taken into account for analysis

for further analysis assume no dependence of 1 on 6

d)/d ≈ compatible with $%&/(

C-V measurements and $%&, '(, $(

- Upper limit $(< 5 fF (~1 fF-pulse shape)- Significant differences for SiPM samples


Cpi

x[fF

]R

q[k

Ω]

Dependence of $%& and '( on and 6 (neutron fluence):

$%& not influenced by and 6

'( increases with decreasing (poly-Si)'( appears to increase for 6 8 .. cm -2

Cq

Cpix

Rpar C/V

Rq

x Npix

SiPM model

C-V measurements: Doping and E -field

Doping profile (&)and E-field (&)assuming 1-D abrupt p +n-junction:


[k

V/c

m]

Minor decrease of for 6 = 5⋅1014 cm -2

(expected from radiation-induced donor removal)

°C, & !9V

[c

m-3

]

°C, & !9V

5⋅1014Φ=0

Breakdown voltage V bd


1 using max( :-) with

Only at +20 °C measured for all SiPMs for 6 variations observed

difference 1(6) 1(6 )significant only at 5 ⋅1014 cm -2

1 ≈ const. up to ≈5⋅1013 cm -2

Increase by ≈ 300 mV at 5 ⋅1014 cm -2

Fluence [cm -2]

1 108× 1 109

× 1 1010× 1 1011

× 1 1012× 1 1013

× 1 1014× 1 1015

×

26.3

26.4

26.5

26.6

26.7

0

≈

Vbd

[V]

°C

Dark current vs. fluence and voltage

Does ( 1) scale with 6? ' (6)/ (...)vs. 1


(6)/ (...) increases with 1: higher 6 faster increase

; 1 ' const ; for 8 1increase by factor up to 2 with

Model: -$'⋅($%&⋅( ))⋅ x2 by and gain implausible

assume change due to DCR “trap assisted high-field generation”?

E11

5E14

3− 2− 1− 0 1 2 3 4 5 6 7 8 90.9

1.1

1.3

1.5

1.7

1.9

2.15E11/E115E12/E11E13/E115E13/E115E14/E11

V - Vbd [V] C

urre

nt r

atio

norm. to 1

5E11

5E14

E13

5E115E13

currentlimit

-$' vs. fluence and voltageModel: -$'⋅($%&⋅( ))⋅ assume $%& .2 fF(1 23 mV and 0from 6 data)


(6, <) allows estimation of -$'(6,<)

-dependence of -$' very different after irradiation! Exponential increase for 1 8 !V + saturation at high 6

Comment: Estimate fraction of busy pixels (occupancy) (-$'⋅#)/%& 0.5 for -$' 1011 s-1, # 20 ns, %& 4384

PHdark – irradiated SiPM () vs. Φ at -30°C:


1 pulse (=/#)/# per sec in random gate :

>. – #(. =/#) Solid curves, which describe data ( # 20 ns)

) 1 no CN (corr. noise) : >.⋅-$'

) @1 no CN: )!⋅>.⋅-$'

+ CN: Var ⋅!⋅)!⋅>.⋅ABC

(N.B. only valid in absence of saturation)

dependence of described by model for 10 11; 6 ;5⋅1012 cm -2

allows to estimate # and DCR (if no strong saturation effects )

Φ=5E14

Φ=0

Φ=E9

…poor

fits

PHdark – irradiated SiPM – -$'


-$'-results agree within ≈ 30 % with results from

-method more complicated and range of validity limitedDifferences may help to better understand rad.damage effec ts on SiPMs ???

Var ⋅!⋅)!⋅>.⋅ABC → -$'

∙!∙)!∙(

–#(.=D/#))

Comparison of and method

Comparison:


Typical agreement E(30%), however also larger differences

Detailed comparison + understanding of differences still to be done

PHlight : ) determinationCan we determine ) ?

)

∙

.

∙


Yes, up to ~10 12 cm -2; No for higher fluences – saturation!

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

0.1

0.2

0.3

0.4

var/(mean x ENF x ESF)Cpix x (V-Voff)

V [V]

Gai

n[1

06 ]

Φ=060 ns gate

/

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

0.1

0.2

0.3

0.4


V [V]

Gai

n [1

06 ]

Φ=5⋅1011cm -2

60 ns gate/

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

0.1

0.2

0.3


V [V]

Gai

n[1

06 ]

Φ=5⋅1012cm -2

60 ns gate

/

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

0.2

0.4


V [V]G

ain

[10

6 ]

Φ=1013cm -2

60 ns gate

/

PHlight: µ-determination

Can we determine F ? (µ = no. of photons initiating a Geiger discharge)


0 3 gate 60 10 14 10

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

100

200

mean/(G x ESF)var/(G^2 x ENF x ESF^2)mean^2 x ENF/var

V [V]

mu

3 gate 60 10 14 10

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

100

200


V [V]

mu

5 gate 60 10 14 10

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

100

200


V [V]m

u

Φ=060 ns gate

Φ=5⋅1011cm -2

60 ns gate

Φ=5⋅1012cm -2

60 ns gate

27 27.5 28 28.5 29 29.5 30 30.5 31 31.5 320

200

400

600mean/(G x ESF)var/(G^2 x ENF x ESF^2)mean^2 x ENF/var

V [V]

mu

Φ=1013cm -2

60 ns gate

Yes, ; !G. 3 V + up to ~10 12 cm -2; No for higher fluences – saturation!

Most reliable:

)∙

3 ways to determine µ

μ

) ∙

μ

)! ∙ ∙ !

μ ! ∙

with )

$%&⋅( )⋅

PHlight : µ (∝ PDE) photo -detection


1 108× 1 109

× 1 1010× 1 1011

× 1 1012× 1 1013

× 1 1014× 1 1015

×

0

100

200

300V - Vbd = 1VV - Vbd = 2VV - Vbd = 3VV - Vbd = 3.5V

Fluence [cm^-2]

mu

0

6 ; ..: PDE essentially unaffected by irradiation6 .. 3⋅..: irradiation affects PDE (low 1 still ok)

6 3⋅..J

: SiPM not a useful photo-detector „ all pixels occupied “

( „ “ optimisation under way)

T = -30°C60 ns gateμ

) ∙

μ

)! ∙ ∙ !

μ ! ∙

Conclusions + next steps1. C-V, Idark , PHdark , PHlight measured at -30 °C for Φ = 0 … 5⋅1014 n/cm 2

2. Formulae presented, which relate SiPM parameters to me asured quantities

3. Cpix and Rq little affected by Φ

4. Idark DCR increase with V and Φ high DC pixel occupancy for Φ > 1013

expect a significant PDE reduction

5. PHdark DCR from variance ~consistent with DCR from I dark

6. PHlight different methods indicate Gain not affected by Φ

7. PHlight approximate V- and Φ-dependence of „relative“ PDE (µ)

for Φ=5⋅1014 cm -2 and –30°C SiPM investigated not a useful photo-detector

Next steps:1. DCR-“PDE“ as a function of ( 6,, ) + noise vs. threshold as function of ( 6,, )

2. Optimisation of pulse-shaping how much can one gain?

3. Include fast component from $( in analysis

4. Implement saturation effects due to dark counts + photo ns

5. Implement what we learn in Schwetzingen


Most methods of SiPM characterisation fail after high Φ irradiation??? Can we find + agree on the most suitable methods ???

Documents

Characterisation of radiation-damaged SiPMs · C-V measurements and $% &,’(,$(- Upper limit $(< 5 fF (~1 fF-pulse shape) - Significant differences for SiPM samples ICASP - Schwetzingen