Post on 04-Jul-2020
Slide # 1
SiPM and SPAD:Emerging Applications for Single-Photon Detection
Slawomir S. Piatek, Ph.D.Senior University Lecturer of PhysicsNew Jersey Institute of Technology
Slide # 2
SiPM and SPAD: emerging applications
Slawomir Piatek
NJIT/Hamamatsu
Copyright 2019 Hamamatsu
Slide # 3
Outline
1/17/2019 SiPM and SPAD
• Introduction to PN junction devices
• Fundamentals of single-photon avalanche photodiode (SPAD)
• Fundamentals of silicon photomultiplier (SiPM)
• Key opto-electronic characteristics of SiPMs
• Applications of SiPMs and SPADs
• Summary and conclusions
Slide # 41/17/2019
Introduction to PN junction devices
SiPM and SPAD
Slide # 5
Terminology
1/17/2019
PD – Photodiode
APD – Avalanche photodiode
SPAD – Single-photon avalanche photodiode
SiPM – Silicon photomultiplier
MPPC – Multi-pixel photon counter; another name for SiPM
PMT – Photomultiplier tube
SiPM and SPAD
The same
SPPC – Single-pixel photon counter; another name for SPAD The same
Slide # 6
Generic PN junction: modes of operation
1/17/2019
I
V
VBD
linear mode APDoperates in this region
PD operates inthis region
≈≈
SPAD and SiPMoperate in this region
Breakdown voltage
V
I
Gei
ger
regi
on
SiPM and SPAD
Dark current versus bias voltage
Slide # 7
PD, APD, and SPAD
1/17/2019
hν
p
n
Vbias
PD
hν
p
n
Vbias
APD
hν
p
n
Vbias
SPAD
1 γ −> 1 pair 1 γ −> ~100 pairs 1 γ −> “∞” pairs
VBD
V
I VBD
V
IVBD
V
I
SiPM and SPAD
Slide # 81/17/2019
Single-photon avalanche photodiode
(SPAD)
SiPM and SPAD
Slide # 9
Structure of a SPAD
1/17/2019
Figures from Zappa et al. 2007
Structure of a thick SPAD Structure of a thin SPAD. This structure is used in SPAD arrays.
SiPM and SPAD
Slide # 10
Operation of a SPAD
1/17/2019
SPAD
photonVBIAS > VBD
time
curr
ent
avalanche starts here
Once triggered, the avalancheis perpetual; however, the SPADis no longer sensitive to light.
Without quenching, SPAD operates as a light switch.
SiPM and SPAD
Slide # 11
Operation of a SPAD (passive quenching)
1/17/2019
VBIAS
RQ
RS
SPAD
RQ >> RS
ISPD
VSPAD
VBIAS
RQ
VBD VBIAS
steady state
transient
quench
recharge
RQ must be large enough to ensure quenching.
quasi-stable “ready”state
photon
SiPM and SPAD
Slide # 12
Operation of SPAD (passive quenching)
1/17/2019
VBIAS
RQ
RS
SPAD
RQ >> RS
VS
time
quench
τr ~RQ.CJ (recovery characteristic time)
10s – 100s of ns
recovery
(quench resistor)
(load resistor)
SiPM and SPAD
Voltage pulse on Rs due to avalanche
Slide # 131/17/2019
Silicon photomultiplier(SiPM)
SiPM and SPAD
Slide # 14
Silicon photomultiplier: structure
1/17/2019
Anode
Cathode
=
SPAD
RQ
p+
π
n+
p+
oxide
Single microcell
Electrical equivalent circuit of a microcell
Side
vie
wTo
p vi
ew
Each microcell is a SPAD in series with a quench resistor. All microcells are connected in parallel. SiPM is not an imaging device because all microcells share a common current summing node.
SiPM and SPAD
Slide # 15
Silicon photomultiplier: operation
1/17/2019
VBIAS > VBD
K
A×G
R
time [ns]
Am
plit
ude
[mV
]
Example of single-photoelectron waveform (1 p.e.)
Gain = area under the curve in electrons
Overvoltage, ΔV = VBIAS - VBD
SiPM and SPAD
Slide # 16
Silicon photomultiplier: modes of operation
1/17/2019
VBIAS
1, 2, 3,….
VBIAS
If the pulses are distinguishable, SiPM can be operated in a photon counting mode.
If the pulses overlap, the SiPM can be operated in an analog mode. The measured output is voltage or current.
counter
SiPM V or I
0 100
light
SiPM and SPAD
Slide # 171/17/2019
Key opto-electronic characteristics of SiPMs
SiPM and SPAD
Slide # 18
Structural and geometrical considerations
1/17/2019
Active area: w×h; ranges from 1×1 mm2 to 6×6 mm2
Microcell area: ranges from 10×10 μm2 to 100×100 μm2
Pitch: center-to-center distance
Geometrical fill factor: fraction of a μ-cell sensitive to light, ~30% to ~ 80%
w
h
μ-cell
pitch
w
h
μ-cellsensitive to light
SiPM and SPAD
Slide # 19
Packaging
1/17/2019
Ceramic Surface mount Metal, TE-cooled Array of SiPMs
SiPM and SPAD
Slide # 20
Gain
1/17/2019
Gai
n×
105
1
2
3
4
5
0
Overvoltage [V]
0 2 4 6 8
Gai
n×
106
0
4
8
12
16
Overvoltage [V]
0 2 4 6 8 10
75×75 μm210×10 & 15×15 μm2 S13360-6075PE
SiPM and SPAD
Gain depends linearly on overvoltage. For a given overvoltage, gain depends on the μ-cell size.
Slide # 21
Photon detection efficiency (PDE)
1/17/2019
PD
E [
%]
0
10
20
30
40
50
Wavelength [nm]
200 400 600 800 1000
Overvoltage [V]P
DE
[%
]
0
10
20
30
40
0 2 4 6 8 10
SiPM and SPAD
Probability that the incident photon is detected. PDE is a function of wavelength and overvoltage.For a given overvoltage, PDE depends on the μ-cell area.
Slide # 22
Dark current and dark count rate
1/17/2019
-5 0 5 10 1510-11
10-10
10-9
10-8
10-7
10-6
10-5
10-4S13720
1.3×1.3 mm2
25×25 μm2
2668 μ-cells
Recommended overvoltage: 7 V
ID = 10-7 A
DCR = ID/(μe) ≈ 570,000 cps
Gain (μ) = 1.1×106
DCR/μ-cell = 213 cps
Overvoltage [V]
Dar
k cu
rren
t [A
]
SiPM and SPAD
Slide # 23
Crosstalk
1/17/2019
primary avalancheor discharge
secondary avalanchesor crosstalk
μ-cell
Photon emitted during an avalanche.
~30 photons with Eγ > 1.14 eV (μ=106)
time
1 p.e.
3 p.e.
3A
AVol
tage
SiPM and SPAD
Slide # 24
Crosstalk probability
1/17/2019
0 2 4 6 8 10
Overvoltage [V]
0
10
20
30
40
Cro
ssta
lk p
roba
bili
ty [
%]
S13360
25×25 μm2
0 2 4 6 8 10
Overvoltage [V]
0
20
40
60
8075×75 μm2
S13360
Cro
ssta
lk p
roba
bili
ty [
%]
Crosstalk probability depends on overvoltage and microcell size (gain).
SiPM and SPAD
Slide # 25
Linearity and dynamic range
1/17/2019
10-12 10-11 10-10 10-9 10-8 10-7 10-6 10-510-8
10-7
10-6
10-5
10-4
10-3
10-2
10-1
10-0
Incident light level [W]
Out
put c
urre
nt [
A]
25×25 μm2
2668 μ-cells
S13720
10-8 W implies 3.7×10-12 W per μ-cell
which is ~16×106 photons per second per μ-cell
which is ~1 photon per 63 ns
SiPM and SPAD
For a given active area, the number of μ-cells determines the upper limit of the dynamic range.
Slide # 26
Tradeoffs
1/17/2019 SiPM and SPAD
μ-cell size
Fixed: active area & overvoltage; change μ-cell size
Slide # 27
Tradeoffs
1/17/2019 SiPM and SPAD
Overvoltage
Fixed: active area & μ-cell size; change overvoltage
Slide # 281/17/2019
Applications
• Time-of-flight LiDAR
• Flash LiDAR
• Flow cytometry
• Bio- and chemiluminescence
• Dark matter detection
SiPM and SPAD
Slide # 29
Photodetector comparison
1/17/2019
PMT APD SiPM
Quantum efficiency Up to ~40% Up to ~85% Up to ~50%
Gain 105 - 107 Up to ~100 105 - 106
Excess noise ~1.2 3 − 4 ~1.1 – 1.2
Minimum detectablepower
Best BetterGood
Dynamic range Dependent on the voltage divider
Largest (high saturation level)
Dependent on thenumber of microcells
Temperatureeffects
Weak Strong Strong
SiPM and SPAD
Slide # 30
Additional attributes of SiPMs
1/17/2019 SiPM and SPAD
• Operation immune to electric and magnetic fields
• Low bias voltage (10s of volts)
• No damage or lasting effects when exposed to full light
• Ability to make arrays and to mount on a circuit board
• Ruggedness
• Simple biasing circuit
Slide # 311/17/2019
Dark matter detection
SiPM and SPAD
Slide # 32
Background
1/17/2019 SiPM and SPAD
• A galactic mass deduced from stellar kinematics is much larger than that implied by its luminous matter. The difference is known as dark matter.
• Dark matter interacts gravitationally but emits little or no light.
• The nature of dark matter is still not known.
• Among a number of possibilities, dark matter could be in the form of particles that can interact with “ordinary” matter.
• Active research exists in identifying the nature of dark matter.
Slide # 33
Detection concept for WIMPs
1/17/2019 SiPM and SPAD
−Cathode
+Anode
Ground
Electric field
175-nm light from liquid Xenon scintillation
-95 °C
Xenon (L)
Xenon (G)
Photodetectors
PhotodetectorsDark matter particle (WIMP)
Electron
Light from ionized Xenon gas
Arb
.
S1
S2
Electron drift timeThese detect mostly S1
These detect mostly S2 Legend:
Slide # 34
Characteristics of received light
1/17/2019 SiPM and SPAD
• Two pulses of light: S1 contains ~few tens of photons, while S2 ~few thousand
• Wavelength: UV and V
• Pulse duration: ~ns for S1 and ~μs for S2
Slide # 35
Photodetector requirements
1/17/2019 SiPM and SPAD
• Photosensitivity in UV and V range
• High intrinsic gain (suppresses importance of electronic noise)
• High bandwidth (to avoid distortion of S1 pulse)
• Able to operate at low temperature (- 95 °C or 178 K)
Slide # 36
Product recommendation for liquid Xenon detection
1/17/2019 SiPM and SPAD
VUV3 and VUV4
Array: Four 5.95×5.85 mm2 SiPMs
Each SiPM has 13,923 μ-cells, 50 μm pitch
Fill factor: 60%
From Bonesini et al. 2018
Slide # 37
Effect of temperature on the gain
1/17/2019 SiPM and SPAD
From Bonesini et al. 2018
VUV4 • Breakdown voltage decreases with decreasing temperature.
• Gain depends on VBIAS - VBD
• If VBIAS – VBD is held constant, the gain is constant too.
Slide # 381/17/2019
Flow cytometry
SiPM and SPAD
Slide # 39
Flow cytometry: study of cells and micro-particles
1/17/2019
Interrogation oranalysis point
~100 μm~20 μm
Flow
SiPM and SPAD
• Used in medicine, biology, and engineering to study and sort cells and micro-particles
• Cells scatter interrogation light. The manner of scatter depends on the cell properties.
• The rate of interrogation is on the order of 1,000 cells per second or more.
Slide # 40
Flow cytometry basic concept
1/17/2019
FD
SCD
FSD
Laser 1
Las
er 2
Data collection & analysis
Beam mixer
DM
DM
Flow
Legend:FSD – forward-scatter detectorSCD – side-scatter detector
FD – fluorescence detectorDM – dichroic mirror
Fluorescence-taggedcell
SiPM and SPAD
Slide # 41
Flow cytometer: what is measured?
1/17/2019
time
FWHM
Area
Peak
Out
put v
olta
ge f
rom
pre
ampl
ifie
r
The front-end electronics can be set up to measure the peak value of the pulse, its FWHM, and/or area under the curve. These different measurements provide specific information about the cell.
Forward-scatter signal
Sid
e-s
catt
er
sig
nal
SiPM and SPAD
Slide # 42
Characteristics of received light
1/17/2019
• Wavelength: can be selected depending on cell sizes and fluorescence
• Pulses: duration dependent on the sheath flow speed and cell size and is on the order of μs
• Number of photons per pulse varies from a few to thousands
• Rate of pulses is in kHz
SiPM and SPAD
Slide # 43
Photodetector requirement
1/17/2019
PD E
Photosensitivity/gain
Can we detect the red pulse?
PD E
saturation
Linearity & dynamic range
Is the output proportional to the input?
light input electricalsignal
SiPM and SPAD
Slide # 44
Photodetector requirement
1/17/2019
PD E
Gain variation
PD E
increasing temperature
Temperature drift
How much excess noise is there? Does the gain depend on temperature?
SiPM and SPAD
Slide # 45
Photodetector requirement
1/17/2019
PD E
Stability
PD E
Bandwidth
How stable are the photodetector’s characteristics such as gain?
Does the photodetector and detection electronics have enough bandwidth?
SiPM and SPAD
Slide # 46
Product recommendation for flow cytometry
1/17/2019 SiPM and SPAD
Hamamatsu S14420 SiPM (MPPC) series
Slide # 471/17/2019
Bio- and chemiluminescenceluminometer
SiPM and SPAD
Slide # 48
Basic concept of a luminometer
1/17/2019
swab
tube
reagent
bacteria
photodetector
• Used to measure surface cleanliness
• The reagent reacts with the bacteria ATP, causing emission of light.
• The amount of light is proportional to the amount of ATP.
ATP – adenosine triphosphate
SiPM and SPAD
Legend:
Slide # 49
Characteristics of the received light
1/17/2019 SiPM and SPAD
1. Wavelength in the 500 nm – 550 nm range
2. Diffuse (diffuse light cannot be focused)
3. Weak and slowly varying
diffuse light in
diffuselight out
lens
Inte
nsit
y
time~ 15 − 20 s
Slide # 50
Photodetector requirement
1/17/2019 SiPM and SPAD
1. Large active area
4. Low dark current
2. Photosensitivity in 500 – 550 nm range
3. Gain
5. Small detection bandwidth (limited by the amplifier)
frequency
Pow
er s
pect
rum
white noise spectrum
B1
B2
Wider bandwidth reduces S/N
Signal ~ A Noise ~ √A so S/N ~ √A
Slide # 51
Product recommendation for luminescence
1/17/2019 SiPM and SPAD
Hamamatsu S14420 SiPM (MPPC) series
Slide # 521/17/2019
Radiation monitoring
SiPM and SPAD
Slide # 53
Radiation monitoring basic concept
1/17/2019 SiPM and SPAD
Scintillator Photodetector
Ionizing radiation time
E
trigger
Detection of radiation such as α, β, γ, and others. Applications in, for example, environmental monitoring, radiation safety, or security.
Output pulses
Slide # 54
Characteristics of received light
1/17/2019 SiPM and SPAD
• Pulses
• Number of photons per pulse depends on energy of the ionizing radiation and type of scintillator
• Duration of the pulse depends on the size and type of the scintillator (decay time constants range from ns to μs)
• Frequency of pulses depends on the rate of incoming radiation
Slide # 55
Photodetector requirements
1/17/2019 SiPM and SPAD
• High intrinsic gain
• Ability to couple to a scintillator
• Suitable for portable hand-held devices
• Large active area
• High photon detection efficiency
Slide # 56
Product recommendation for radiation detection
1/17/2019 SiPM and SPAD
S14160 series
Slide # 571/17/2019
Time-of-flight LiDAR
SiPM and SPAD
Slide # 58
Time-of-flight (ToF) automotive LiDAR
1/17/2019 SiPM and SPAD
• Automotive LiDAR is likely needed on a self-driving car to provide high-resolution 3D imaging of its surroundings.
• There is a great interest in developing self-driving cars.
• Photodetection and beam-steering are two most outstanding challenges in the development of an automotive LiDAR.
Slide # 59
Time-of-flight (ToF) LiDAR
1/17/2019
start pulse
stop pulse
laser
target
photodetector
timer
beam steeringsystem
Distance R
Measure time of flight Δt R = cΔt/2
λ = 905 nm or 1550 nm
SiPM and SPAD
Slide # 60
Characteristics of the received light
1/17/2019
Wavelength – 905 nm or 1550 nm
Pulses
Repetition period~μs
Duration~few ns
Peak power – few to 100’s of nW
SiPM and SPAD
Slide # 61
Photodetector requirements
1/17/2019
• High quantum efficiency at 905 nm and/or 1550 nm (affects detection range)
• High detector (intrinsic) gain (reduces importance of electronic noise)
• Small excess noise (affects timing error)
• Small time jitter (affects distance error)
• High dynamic range
SiPM and SPAD
Slide # 621/17/2019
time
arb.
trig. level
Different trigger times: time walk effect
time
Histogram of responsetimes
Num
ber
FWHM
time
trig. leveljitter causes timing errors
SiPM and SPAD
Importance of excess noise Importance of photodetector jitter
Slide # 63
Importance of dynamic range
1/17/2019
arb.
time
saturation
trigger
dark currentand
background
• If fixed trigger is used, the changing background will affect timing.
• If constant fraction trigger is used, saturation will affect timing.
SiPM and SPAD
Slide # 64
Product recommendation for ToF LiDAR
1/17/2019 SiPM and SPAD
Hamamatsu S13720 SiPM (MPPC) series
Slide # 651/17/2019
Flash LiDAR
SiPM and SPAD
Slide # 66
Flash LiDAR
1/17/2019
projector
to the targetθ
The projector illuminates the scene with multiple divergent pulses of short duration (few ns).
SiPM and SPAD
Slide # 67
Flash LiDAR
1/17/2019
θ from target
FOV matched to theillumination FOV
2D array of detectors
timing circuit (all elements in parallel)
The returned light is imaged on an array of sensors. Each pixel measures distance to the imaged element of the scene.
SiPM and SPAD
Slide # 68
Flash LiDAR
1/17/2019
SPAD array
Repetitionperiod
timeN
o.2R/c
Histogram of trigger times
A histogram like the one above must be obtained for each pixel.
SiPM and SPAD
SPAD array by Hamamatsu is being developed. More information will be released at the Photonics West 2019 meeting.
Slide # 69
Summary and conclusions
1/17/2019 SiPM and SPAD
Dark matterdetection ToF LiDAR
Flow cytometry
Radiation detection
ATP luminometer
Slide # 70
Summary and conclusions
1/17/2019 SiPM and SPAD
• SiPMs as a family have a spectral response in the 120 nm – 1000 nm range.
• SiPMs enjoy an ever increasing adoption as the detector of choice in a wide range of applications.
• SiPM has an intrinsic gain comparable to that of a PMT, and on the order of 106.
• SPAD arrays open the possibility of single-photon imaging and may resolve engineering challenges in designing automotive LiDAR.
Slide # 71
Upcoming Hamamatsu Promotional Activities:
1/17/2019 SiPM and SPAD
SPIE Photonics West 2019 (San Francisco), Wednesday, February 6, 2019 (8 hours)Mr. Koei Yamamoto (Director of Solid State and Laser Division from HQ Japan) is presenting two key topics highlighted in red:
• Part 1 (8:10 AM – 10:00 AM) – “Introduction to Photodetectors” (Slawomir Piatek)• Part 2 (10:15 AM − 11:45 AM) – “Evolution of Geiger Mode APD Technologies –
SiPM & SPAD” (Mr. Koei Yamamoto)• Part 3A (1:00 PM – 2:15 PM) – “Spectroscopy and Spectrometer Concepts” (Slawomir
Piatek)• Part 3B (2:15 PM – 3:00 PM) – “Spectroscopy and Spectrometer – What’s Next” (John
Gilmore)• Part 4A (3:15 PM – 4:00 PM) – “Automotive LiDAR: Design Concepts and
Challenges” (Jake Li)• Part 4B (4:00 PM – 5:00 PM) – “Photonics Technology Improvements that Drive
the Future of LiDAR Designs” (Mr. Koei Yamamoto)
Slide # 72Slide # 72
Microscopy Images
1/17/2019
Thank you for listening!
Contact: Slawomir Piatekpiatek@njit.edu
SiPM and SPAD
For more info, see our past SiPM webinars linked below.
SiPM: Operation, performance, and possible applicationshttps://www.youtube.com/watch?v=HHS1qYfMDfk&t=430s
Low light detection: PMT vs. SiPMhttps://www.youtube.com/watch?v=I5dUB6LT4T4&t=2s