Silicon Photomultiplier tests in LN, LAr Janicsk o J ozsef - - PowerPoint PPT Presentation

▶
silicon photomultiplier tests in ln lar
SMART_READER_LITE
LIVE PREVIEW

Silicon Photomultiplier tests in LN, LAr Janicsk o J ozsef - - PowerPoint PPT Presentation

Nov 06, 2023 135 likes •460 views

Silicon Photomultiplier tests in LN, LAr Janicsk o J ozsef February 6, 2009 GERmanium Detector Array (GERDA) Is a double beta decay experiment We operate HPGe detectors in LN/LAr. Low countrate low background experiment, extreme

slide-1
SLIDE 1

Silicon Photomultiplier tests in LN, LAr

Janicsk´

  • J´
  • zsef

February 6, 2009

slide-2
SLIDE 2

GERmanium Detector Array (GERDA)

Is a double beta decay experiment

  • We operate HPGe detectors in LN/LAr.
  • Low countrate low background experiment, extreme radiopurity required
  • LAr is a well known scintillator
  • LAr is:

  

  • 1. cooling liquid
  • 2. passive shielding
  • 3. scintillator

February 6, 2009 1/31

slide-3
SLIDE 3

February 6, 2009 2/31

slide-4
SLIDE 4

Ar scintillation spectrum

ρ(LAr) = 1.4 g/cm3 n(LAr)= 1.24, 40000 photon/MeV

February 6, 2009 3/31

slide-5
SLIDE 5

Heidelberg setup

stainless steel lead shield

* LAr

GAr

PMT Dewar WLS source tube fiber

  • ptical

LAr GAr

signal HV

LArGe@MPI-K:

Schematic system description

Internal source External source

  • Dewar Ø29 cm, h=65 cm (43 L – total volume)
  • Light detection: WLS (VM2000 + PST/TPB)

+ PMT(8“, ETL 9357-KFLB )

  • Active volume Ø20 cm, h=43 cm

§ 19 kg LAr (13,5 L)

  • Shielding:

5 cm lead (+ 10 cm BP for n) +15 mwe underground

Measurements: NaI – detector used for: 1) coincidence measurements; 2) reference measurements.

UV vis

February 6, 2009 4/31

slide-6
SLIDE 6

Goals

PMT’s, in general are:

  • ”Dirty” (radioactive), not suitable for low background experiments
  • They don’t work at cryogenic temperatures
  • Requires HV (problems in Ar atmosphere)

SiPM could be a replacement of PMT’s with higher radiopurity, no HV, UV sensitive etc. Goal is to reproduce the results of the Heidelberg (GERDA) group with SiPM’s : at least 1000 p.e. / MeV

February 6, 2009 5/31

slide-7
SLIDE 7

Hamamatsu MPPC

February 6, 2009 6/31

slide-8
SLIDE 8

Photon Detection Efficiency

February 6, 2009 7/31

slide-9
SLIDE 9

Setup

  • Bias circuit and amplifier built on one PCB
  • Preamp. works in LN
  • works with a coax. cable between the SiPM

and the PCB

  • in the final setup SiPM in LN, preamp. at RT

February 6, 2009 8/31

slide-10
SLIDE 10

SiPM properties at LN temperature

February 6, 2009 9/31

slide-11
SLIDE 11

Dark rate v. temperature

The main reason for cooling down a Si device is the thermal noise. We don’t have cryostat, I just wait until the temperature stabilizes in the dewar. A Pt-100 is attached to the SiPM

temperature [K] 100 150 200 250 300 dark rate [cps] 10

2

10

3

10

4

10

5

10

6

10 S10365-11-100

Effect of ambient light is not excluded, during overnight measurement the rate dropped below 1Hz. = ⇒ Up to 6 orders of magnitude reduction in dark rate.

February 6, 2009 10/31

slide-12
SLIDE 12

Xtalk

Simply recording the dark counts RT:

h1_tmp Entries 1906180 Mean 277 RMS 181.3

ADC counts 500 1000 1500 2000 2500 3000 1 10

2

10

3

10

4

10

h1_tmp Entries 1906180 Mean 277 RMS 181.3

Dark counts at RT

LN:

hps0 Entries 4698 Mean 0.982 RMS 1.38
  • 1

1 2 3 4 5 6 7 8 9 1 10

2

10 hps0

Entries 4698 Mean 0.982 RMS 1.38

hps0

hps1 Entries 4698 Mean 1.127 RMS 1.488
  • 1

1 2 3 4 5 6 7 8 9 1 10

2

10 hps1

Entries 4698 Mean 1.127 RMS 1.488

hps1

RT: I estimate 21% LN: 40 - 50% or is not dark rate (can not be measured)

February 6, 2009 11/31

slide-13
SLIDE 13

Relative efficiency

RT LN 850nm IR laser

h2 Entries 10000 Mean 116.7 RMS 89.23 a.u. 50 100 150 200 250 300 350 400 450 500 counts 50 100 150 200 h2 Entries 10000 Mean 116.7 RMS 89.23

IR laser at RT

2.1 p.e.

h2 Entries 10000 Mean 114.8 RMS 69.66 50 100 150 200 250 300 350 400 450 500 20 40 60 80 100 120 h2 Entries 10000 Mean 114.8 RMS 69.66 IR laser in LN

1.9 p.e.

400nm Blue LED No visible efficiency drop at LN temperature (compared to RT) Looks like we can have low dark rate and high QE in the same time

February 6, 2009 12/31

slide-14
SLIDE 14

Pulse shape in LN

RT LN

Can be explained by structure of the SiPM. The polysilicon resistor is temperature dependent.

February 6, 2009 13/31

slide-15
SLIDE 15

Photon counting

RT LN

h2 Entries 10000 Mean 0.01105 RMS 0.006868

Amplitude (V) 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 a.u. 50 100 150 200 250

h2 Entries 10000 Mean 0.01105 RMS 0.006868

Photon counting

h2 Entries 10000 Mean 0.007811 RMS 0.00307

Amplitude (V) 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02 a.u. 10 20 30 40 50 60 70 80 90

h2 Entries 10000 Mean 0.007811 RMS 0.00307

Photon counting

February 6, 2009 14/31

slide-16
SLIDE 16

Photon counting

  • The peak amplitude cannot be measured accurately
  • By integrating the area under the pulse (offline analysis) I could restore the resolution

RT LN

h2 Entries 10000 Mean 0.01105 RMS 0.006868

Amplitude (V) 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 a.u. 50 100 150 200 250

h2 Entries 10000 Mean 0.01105 RMS 0.006868

Photon counting

h3 Entries 10000 Mean 0.941 RMS 0.3588

Area (a.u.) 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 20 40 60 80 100 120 140 160 180

h3 Entries 10000 Mean 0.941 RMS 0.3588

Photon counting

February 6, 2009 15/31

slide-17
SLIDE 17

DAQ and SiPM

  • PIXIE4 DAQ (75 MHz, 14 bit) can record the pulse-shape without any

“charge amplifier”

  • Cooling the SiPM gives longer pulse, better for the DAQ

Typical pulse shapes for the 1600 and 100 pixel Si PM in LN. Recorded with the DAQ.

February 6, 2009 16/31

slide-18
SLIDE 18

DAQ and SiPM

  • The resolution is not so good as with the oscilloscope, but I still can

distinguish 20 photon peaks.

  • Amplitude calculated with trapezoidal energy filter.

hpx Entries 4.161728e+07 Mean 20.17 RMS 11.17

  • Nb. of Photons

5 10 15 20 25 30 35 40 1000 2000 3000 4000 5000 6000 7000 8000 9000

hpx Entries 4.161728e+07 Mean 20.17 RMS 11.17

Photon number

MCA spectrum recorded with the DAQ. The light intensity (LED) was increased in more (4) steps.

February 6, 2009 17/31

slide-19
SLIDE 19

DAQ and SiPM

Charge-sensitive preamplifier

February 6, 2009 18/31

slide-20
SLIDE 20

DAQ and SiPM

Resolution with the charge-sensitive preamplifier

Photon spectrum with charge preamplifier and DAQ

February 6, 2009 19/31

slide-21
SLIDE 21

... but the resolution deteriorates too fast

htmp Entries 20000 Mean 3801 RMS 3036

2000 4000 6000 8000 10000 12000 14000

  • 2

10

  • 1

10 1 10

2

10

3

10

htmp Entries 20000 Mean 3801 RMS 3036

Cha_MCAEnergy[0]

February 6, 2009 20/31

slide-22
SLIDE 22

Pulseshape analysis

  • Time resolution is important to distinguish slow and fast scintillation.

Delayed signal expected up to 2-3 µs

50 100 150 200 250 300 7000 8000 9000 10000 11000 12000 13000 14000 15000 50 100 150 200 250 300 100 200 300 400 500 600

February 6, 2009 21/31

slide-23
SLIDE 23

Damage caused by LN

Si PM is held in place by soft epoxy resin, which doesn’t like LN temperatures. One with 100 pixels is gone ... Still, survived many (∼ 100) cooling cycles

February 6, 2009 22/31

slide-24
SLIDE 24

Light detection in LAr (preliminary)

February 6, 2009 23/31

slide-25
SLIDE 25

Direct light detection

178 nm, -95oC

February 6, 2009 24/31

slide-26
SLIDE 26

Direct light detection

2 SiPM’s in LAr at 5 cm distance from a Th228 source. Dark counts removed by coincidence trigger. Overnight measurement (about 14 h)

h2 Entries 2165 Mean 434.4 RMS 985.4

ADC counts 5000 10000 15000 20000 25000 30000 2 4 6 8 10 12

h2 Entries 2165 Mean 434.4 RMS 985.4 h2 Entries 258 Mean 404.8 RMS 363.3

ADC counts 5000 10000 15000 20000 25000 30000 0.5 1 1.5 2 2.5 3 3.5 4

h2 Entries 258 Mean 404.8 RMS 363.3

LAr + Th source left, right LN with no source

compared to LN

  • LN + Th source 2-3 fold increase
  • LAr, no source 2 fold increase
  • LAr + Th source ∼10 fold increase in the countrate with coincidence trigger

February 6, 2009 25/31

slide-27
SLIDE 27

Two step WL shifting

Idea: 128 nm = ⇒ Blue scintillator (∼400 nm) = ⇒ Green WLS fiber

  • Blue scintillator: VM2000 foil coated with TPB → high efficiency
  • Green WLS fiber → 3% trapping efficiency 3.5 m attenuation length

February 6, 2009 26/31

slide-28
SLIDE 28

Principle

Number of photons detected: N = Y F (1 − S)Ce−l/λR1/SQ where:

  • Y = 50000, photon yield / MeV in LAr
  • F = 1.35 TPB fluor efficiency
  • S = 0.1 - 0.05 surface covered by the fibers
  • C = 0.034 - 0.07 WLS trapping efficiency
  • l = 10 m (one fiber)
  • λ > 3.5m, attenuation length
  • R = 0.2 - 0.99 reflectivity of TPB or VM2000
  • Q = 0.4 QE of the SiPM at 500 nm

February 6, 2009 27/31

slide-29
SLIDE 29

My Setup

  • 2 x 10m WLS fiber total surface:

2 x 314cm2

  • effective area: 4 x 110cm2
  • 4 x SiPM
  • linear preamps

integration done by the DAQ

  • 1) Al foil with TPB coating (home-made)
  • 2) VM2000 foil from CREST

February 6, 2009 28/31

slide-30
SLIDE 30

Th228 spectrum + VM2000

Preliminary:

  • Nb. of Photons

10 20 30 40 50 a.u. 0.001 0.002 0.003 0.004 0.005 0.006 0.007

th228 + VM2000 + WLS Background February 6, 2009 29/31

slide-31
SLIDE 31

Conclusion

SiPM’s:

  • they are easy to use
  • excellent resolution
  • they do work in LN
  • I could see some light in LAr

A more serious test-setup is under construction:

  • Gas tight dewar is under construction
  • mechanical parts being prepared in the workshop
  • electronics under development

February 6, 2009 30/31

slide-32
SLIDE 32

?

  • Joint project for development of the read-out electronics ?
  • Could we have SiPM optimized for LN temperature ?
  • improoved UV sensitivity ?
  • LARGE area SiPMs ?

February 6, 2009 31/31