tests and performance of multi-pixel geiger mode …pd07 workshop, kobe, june 28, 2007 y. musienko...

PD07 Workshop, Kobe, June 28, 2007 Y. Musienko

Tests and performance of multi-pixel Geiger mode APD's and APD's for the CMS ECAL

Y. Musienko, INR (Moscow)/Northeastern University (Boston)


Motivation• At Beaune-05 NDIP conference several groups reported about

development of multi-pixel Geiger-mode APDs (G-APDs)• The G-APD parameters (gain, PDE, excess noise factor, timing

response) were reported to be similar or even superior to the parameters of PMTs

• During last 2 years new G-APD structures have been developed. Improved performances of these photosensors were reported by different investigators

• These results increased an interest to G-APDs from HEP, astroparticleand medical communities

• Correct evaluation of the G-APDs parameters and their influence on detector performance became very important

• However measurements of these parameters (especially QE) is not an easy task taking into account small sensitive area (typically 1 mm2) and rather high dark count rates at room temperature.


OutlineIn my talk:• I will briefly describe the experimental technique we use to

characterize G-APDs• The results of our studies of recently developed G-APDs from 3

producers will be reported:- PDE(U)- F(U)- Ndark(U) - Gain(U)- KV(U)- KT(U)- PDE(λ)• Main parameters of the G-APDs will be compared with the parameters

of an APD operated in linear mode (S8148 HPK APD)


G-APDs studied

G-APDs Producer's reference Package Protection SubstrateArea

[mm2]# of

pixels VB(T=22 C) [V]

CPTA/Photonique* SSPM_0701BG_PCB PCB No p-type 1 556 30.7

Dubna/Mikron** pMP-3d-11 TO-18 Epoxy p-type 1 1024 39.4

Hamamatsu*** S10362-11-050C Ceramic Epoxy n-type 1 400 68.8

*) http//www.photonique.ch

**) http//sunhe.jinr.ru/struct/neeo/apd/***) http//www.hamamatsu.com


Set-up•MPGM APD and XP2020 PMT were illuminated with the parallel lightfrom LED through 0.5 mm diameter collimator• Mechanical system allowed precise positioning (<50 mm) of the APD and PMT in all 3 dimensions• LEDs with the peak emission of 410 nm and 515 nm were used in this study• APD was connected to fast linear transimpedance amplifier (gain~20) • Temperature - monitored using Pt-100 resistor• Currents were measured using Kethley-487 source-meter• Amplitude spectra were measured using LeCroy 2249 W CAMAC ADC• LeCroy 623B discriminator and 250 MHz scaler were used for signal counting• “Optometrics” spectrophotometer was used for spectral response measurements• Low temperature measurements were done inside the freezer


LED spectrum (low light)

MPGM APDs have very good pixel-to-pixel signal uniformity. Pedestal is separated from the signal produced by single fired pixel Q1 .

CPTA APD (U=37 V, T=22 C)

0

2000

4000

6000

8000

10000

12000

100 150 200 250 300 350 400

ADC channels

Cou

nts


Single electron spectrum

When V-Vb>>1 V typical single pixel signal resolution is better than 10% (FWHM)). However photons produced during the pixel breakdown can penetrate another pixel and fire it. As a result more than one pixel is fired by single photoelectron.

SES CPTA APD, U=42 V, T=-28 C

1

10

100

1000

10000

200 300 400 500 600 700ADC ch.

Cou

nts

SES MEPhI/PULSAR APD, U=57.5V, T=-28 C

1

10

100

1000

10000

0 100 200 300 400 500ADC ch.

Cou

nts

(Y. Musienko et al., A 567 (2006) pp.57–61)


Parameter definition: Gain

Each pixel works as a digital device – 1,2,3... photons produce the same signal Q1=Cpixel*(V-Vb) (or Single Pixel Charge).Multi-pixel structure works as a linear device, as soon asNpe=Ng*QE<<N0 , N0 – is a total number of pixels/deviceMeasured charge :

Qoutput=Npe*Gain ,It was found by many groups that : Gain ≠ Q1 ,More than 1 pixel is fired by one primary photoelectron!

Gain=Q1*np , where np is average number of pixels fired by one primary photoelectron.


PDE measurements

Photon detection efficiency (PDE) is the probability to detect single photon when threshold is <Q1. It depends on the pixel active area quantum efficiency (QE), geometric factor and probability of primary photoelectron to trigger the pixel breakdown Pb (depends on the V-Vb !!, Vb – is a breakdown voltage) :

PDE = QE*Gf*PbFor G-APDs with low dark count rate (<3 MHz) pedestal events can be easilyseparated from the event when one or more than one pixel were fired by the incident photons. In this case we can use well known property of the Poisson distribution :

<Npe> = - ln(P(0)) This equation works even in the case of the photodetector with very high multiplication noise !!!(“Peak” counting method overestimates the <Npe>. Method which uses the width of the signal distribution underestimates the <Npe>). Number of incoming photons (Nγ) from LED pulse can measured with calibrated PMT (XP2020 PMT, for example). Then:

PDE(λ) = Npe/Nγ

LED emission spectra must be measured as well (in pulsed mode !!!)


Dark current vs. Bias (T=22 C)CPTA APD

0

1

2

3

4

5

6

7

8

30 32 34 36 38 40 42 44

Bias [V]

Dar

k C

urre

nt [ μ

A]

Dubna/Mikron APD (pMP-3d-11)

0

0.5

1

1.5

2

2.5

40 41 42 43 44 45 46 47

Bias [V]

Dar

k C

urre

nt [ μ

A]

S10362-11-050C HPK MPPC

0

0.05

0.1

0.15

0.2

0.25

68 68.5 69 69.5 70 70.5 71

Bias [V]

Dar

k C

urre

nt [ μ

A]


Dark count vs. Bias (~0.5 p.e. threshold, T=22 C)

CPTA APD

0

500

1000

1500

2000

2500

3000

3500

4000

4500

30 32 34 36 38 40 42 44Bias [V]

Dar

k C

ount

[kH

z]


0

1000

2000

3000

4000

5000

6000

40 41 42 43 44 45 46 47Bias [V]

Dar

k C

ount

[kH

z]

S10362-11-050C HPK MPPC

0

100

200

300

400

500

600

700

800

68 68.5 69 69.5 70 70.5 71

Bias [V]

Dar

k C

ount

[kH

z]


Signal shape (HPK and Dubna G-APD)

Hamamatsu G-APD Dubna/Mikron G-APD


Signal shape (CPTA G-APD)

Direct signal (RL=50 Ohm) After tail cancelation using C=0.47 nF (RL=50 Ohm)


Photon detection efficiency vs. Bias

CPTA APD

0

5

10

15

20

25

30

35

40

30 32 34 36 38 40 42 44

Bias [V]

PDE(

515

nm) [

%]


0

2

4

6

8

10

12

14

16

40 41 42 43 44 45 46 47

Bias [V]

PDE(

515

nm) [

%]

S10362-11-050C HPK MPPC

0

5

10

15

20

25

30

35

68 68.5 69 69.5 70 70.5 71

Bias [V]

PDE(

515

nm) [

%]


Gain vs. Bias (T=22 C)

CPTA APD

0

0.2

0.4

0.6

0.8

1

1.2

30 32 34 36 38 40 42 44

Bias [V]

Gai

n, 1

06


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

40 41 42 43 44 45 46 47

Bias [V]

Gai

n, 1

06

S10362-11-050C HPK MPPC

0

0.5

1

1.5

2

2.5

68 68.5 69 69.5 70 70.5 71

Bias [V]

Gai

n, 1

06


G-APD voltage coefficient

CPTA APD

0

5

10

15

20

25

30

35

40

45

34 35 36 37 38 39 40 41 42 43

Bias [V]

1/A

*dA

/dV

[%]


0

20

40

60

80

100

120

39 40 41 42 43 44 45 46 47

Bias [V]

1/A

*dA

/dV

[%]

S10362-11-050C HPK MPPC

0

50

100

150

200

250

300

350

69 69.2 69.4 69.6 69.8 70 70.2 70.4 70.6

Bias [V]

1/A

*dA

/dV

[%]

kV=dA/dV*1/A, [%/V]


Temperature sensitivity CPTA APD

0

50

100

150

200

250

300

350

400

30 32 34 36 38 40 42 44

Bias [V]

Sign

al a

mpl

itude

[AD

C c

h.]

T=-25 CT= 22 C

Dubna/Mikron APD

0

20

40

60

80

100

120

140

30 32 34 36 38 40 42 44 46 48Bias [V]

Am

plitu

de [A

DC

ch.

]

T=-25 CT= 22 C

Hamamatsu MPPC

020406080

100120140160180200

66.5 67 67.5 68 68.5 69 69.5 70 70.5 71Bias [V]

Am

plitu

de [A

DC

ch.

]

T=-25 CT= 22 C

LED signal was measured in dependence on bias at 2 temperatures. During low temperature measurements (T=-25 C) G-APDs were placed inside commercial freezer (LED was kept at room temperature)

CPTA/Photnique:dVB/dT=-20 mV/CDubna/Micron:dVB/dT=-122 mV/CHamamatsu:dVB/dT=-50 mV/C


G-APD temperature coefficient CPTA APD

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

34 35 36 37 38 39 40 41 42 43

Bias [V]

-1/A

*dA

/dT

[%]


0

2

4

6

8

10

12

14

16

39 40 41 42 43 44 45 46 47

Bias [V]

-1/A

*dA

/dT

[%]

S10362-11-050C HPK MPPC

0

2

4

6

8

10

12

14

16

69 69.2 69.4 69.6 69.8 70 70.2 70.4 70.6

Bias [V]

-1/A

*dA

/dT

[%]kT=dA/dT*1/A, [%/°C]


Excess noise factor (I)

Excess noise factor can be measured from the width of single electron spectra and calculated using:

F=1+var1/<A1>2 (2),where <A1> is average amplitude produced by single electron and var1 its variance.Another way (gives the same result) is to compare the Npecalculated from equation (1) and from Nw from the width of the measured spectra (number of measured photoelectrons should be small (P(0) should not be very low) . This method is easier to use:

F=Npe/Nw (3),


Excess noise factor (II)CPTA APD

0.9

0.95

1

1.05

1.1

1.15

1.2

30 32 34 36 38 40 42 44

Bias [V]

F


0.9

0.95

1

1.05

1.1

1.15

1.2

1.25

1.3

1.35

1.4

40 41 42 43 44 45 46 47

Bias [V]

F

S10362-11-050C HPK MPPC

0.9

0.95

1

1.05

1.1

1.15

1.2

1.25

1.3

1.35

1.4

68 68.5 69 69.5 70 70.5 71

Bias [V]

F


G-APDs spectral response (I)

For the spectral response measurements “Optometrics”SDMC1-03 spectrophotometer was used. We also used a calibrated PIN photodiode as a reference. Spectrophotometer light intensity was significantly reduced using gray filters to the level when the maximum current measured with the APD was only ~30% higher than its dark current. This was done to avoid the non-linearity effects caused by high pixel illumination. Photocurrent measured with the APD was compared with the photocurrent measured with the PIN photodiode. In addition the measurements with the LED pulsed light were used for absolute spectral response calibration (at least 2 different LED measurements were done for each APD).


G-APDs spectral response (II)

T=22 C

0

10

20

30

40

50

60

350 400 450 500 550 600 650 700 750 800Wavelength [nm]

PDE

[%]

CPTA/Photonique APDHPK S10362-11-050CDubna/Mikron APD


S8148 HPK APD developed for the CMS experiment

20

window

e

conversionγ

e driftπ (i)

e acceleration

e multiplicationn

E

�� p

p

n

++

++ collection

��

��

��

��

��

��

��

��

��

��

��

Si N3 4

��

��

��

��

The CMS electromagnetic calorimeter was built with PbWO4 crystals.61,200 barrel crystals to read out.

Two APD’s per crystal mounted in capsules

Schematic structure of APD

Capsule with 2 APD’s


S8148 performances (I)

0.1

1

10

100

1000

10000

0 50 100 150 200 250 300 350 400 450

Voltage [V]

Gai

n

0

10

20

30

40

50

60

70

80

90

100

300 400 500 600 700 800 900 1000Wavelength [nm]

Qua

ntum

Effi

cien

cy [%

]

K. Deiters et al.,NIM, A461 (2001) 574

Active Area 5x5 mm2

Operating Voltage @ M=50 ~380 V

Capacitance @ M=50 80 pF

Serial Resistance < 10 Ω

Dark Current @ M=50 < 10 nA

Excess Noise Factor @ M=50 2.1

Quantum Efficiency @ 420 nm 73 %

dM/dV x 1/M @ M=50 3.1 %

dM/dT x 1/M @ M=50 -2.4 %

Summary of APD parameters


S8148 performances (II)

0

10

20

30

40

50

60

0 500 1000 1500 2000

Gain

1/M

*dM

/dV

[%]

0

10

20

30

40

50

60

350 400 450 500 550 600 650 700 750 800

Wavelength [nm]

Gai

n

-30

-25

-20

-15

-10

-5

0

0 500 1000 1500 2000

Gain

1/M

*dM

/dT

[%]

1

3

5

7

9

11

13

15

0 500 1000 1500 2000

Gain

Exce

ss N

oise

Fac

tor


Summary• Recently developed G-APDs from 3 producers were studied

in CERN APD Lab using technique developed by our group• Such G-APD parameters as photon detection efficiency,

excess noise factor, dark count, gain, voltage coefficient of the gain, temperature coefficient of the gain and their dependence on the bias voltage were measured for 3 G-APDs at room temperature

• Dependences of the G-APDs PDEs on the wavelength of light were also measured

• Main parameters of the S8148 HPK APD operated in linear mode were presented and compared with the parameters of the G-APDs

tests and performance of multi-pixel geiger mode …pd07 workshop, kobe, june 28, 2007 y. musienko...

Documents