CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene....
Transcript of CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene....
![Page 1: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/1.jpg)
CMOS Single Photon DetectorsCMOS Single Photon Detectors
Edoardo CharbonEdoardo CharbonAQUA - LAP - CSCTAQUA - LAP - CSCT
Ecole Polytechnique FEcole Polytechnique Fééddééralerale, Lausanne, Lausanne(EPFL)(EPFL)
![Page 2: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/2.jpg)
Cristiano Niclass
Dr. Claudio Bruschini Chantal Schneeberger
Ties Kluter Randhir K. Singh
Huan Du
Dr. Max Sergio
Abhishek Garg
AcknowledgementsAcknowledgementsThe The AQUA GroupAQUA Group was started in June 2003. It is funded by the Swiss was started in June 2003. It is funded by the SwissNational Science Foundation and LogitechNational Science Foundation and Logitech
Marek Gersbach
Claudio Favi
![Page 3: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/3.jpg)
The World Is AnalogThe World Is Analog
Physical EntityPhysical Entity
ObserverObserver
Quantitative Quantitative Result Result
OB
SE
RV
AT
ION
OB
SE
RV
AT
ION
MEASUREMENTMEASUREMENT
Is it Continuous?
![Page 4: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/4.jpg)
Typical Measurement ProcessTypical Measurement Process
Quantization in time and amplitudeQuantization in time and amplitude
![Page 5: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/5.jpg)
Quanta Quanta →→ Analog Signal Analog Signal →→ DigitalDigital
In RealityIn Reality……
•• The world is not analog but measurable byThe world is not analog but measurable byquantaquanta
•• The analog measurement is the result of aThe analog measurement is the result of aconversion of the true signalconversion of the true signal
•• Every conversion introduces errorsEvery conversion introduces errors
![Page 6: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/6.jpg)
Why not Skipping Analog?Why not Skipping Analog?
ExampleExample::
photons photons →→ digital pulsesdigital pulses
Quanta Quanta →→ DigitalDigital
![Page 7: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/7.jpg)
OutlineOutline
•• Why, when, where?Why, when, where?
•• Single Photon Avalanche Diodes (SPADs)Single Photon Avalanche Diodes (SPADs)
•• 3D Imaging3D Imaging
•• 2D Imaging2D Imaging
•• Conclusions & future workConclusions & future work
•• AcknowledgementsAcknowledgements
![Page 8: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/8.jpg)
Why, When, Where?Why, When, Where?
![Page 9: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/9.jpg)
Why Do We Care?Why Do We Care?
•• Single photon detectors (Single photon detectors (SPDsSPDs) are entering the) are entering themainstreammainstream
•• They will be found in medium to high volumeThey will be found in medium to high volumeapplicationsapplications
ExamplesExamples::–– Secure Telecom (private key distribution systems) [Gisin/Secure Telecom (private key distribution systems) [Gisin/UniGeUniGe,,
IdQuantiqueIdQuantique, , SensLSensL]]–– Optical Telecom [Optical Telecom [BoserBoser/UCB,/UCB, Pister Pister/UCB, Dust Networks]/UCB, Dust Networks]–– True Random Number Generators [Popovic/EPFL, True Random Number Generators [Popovic/EPFL, IdQuantiqueIdQuantique]]–– High sensitivity (LHigh sensitivity (L33) videocams [Cork ) videocams [Cork UnivUniv., ., QinetiQ, e2v]]–– 3D 3D videocams videocams for gaming applications [***]for gaming applications [***]–– Low dose X-ray imaging [Low dose X-ray imaging [HamamatsuHamamatsu, Fairchild Imaging], Fairchild Imaging]
![Page 10: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/10.jpg)
SPDs SPDs in Sciencein Science
•• Molecular count/detectionMolecular count/detection
•• NeuroNeuro-activity scanning (dye-activity scanning (dye11-based)-based)
•• Fluorescence correlation spectroscopyFluorescence correlation spectroscopy(FCS)(FCS)
•• Two-photon spectroscopyTwo-photon spectroscopy
•• Fluid-dynamics researchFluid-dynamics research
•• Non-ionizing imaging (Non-ionizing imaging (transilluminationtransillumination))
1) Dye is a substance composed of nanoparticles which change reflectivity depending on potential
![Page 11: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/11.jpg)
NeuroNeuro-activity Scanning-activity Scanning
[Grinvald et al., 2001]
![Page 12: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/12.jpg)
NeuroNeuro-activity Scanning (Cont.)-activity Scanning (Cont.)
ΔΔI I ≈≈ 10 10-3-3 I I
for for ΔΔI >> I >> σσ22(I) = I(I) = I1/21/2
→→ must work with ultra-highmust work with ultra-highintensitiesintensities
but, most but, most CCDs CCDs and CMOS APSand CMOS APSare are saturatedsaturated
[Grinvald et al., 2001]
![Page 13: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/13.jpg)
Fluorescence CorrelationFluorescence CorrelationSpectroscopySpectroscopy
CTI project 5814.1
Normalized Correlation
1
1.2
1.4
1.6
1.8
2
2.2
2.4
0.000010.0001 0.001 0.01 0.1 1 10 100 1000tau, ms
Cor
rela
tion
τ [ms]
FCS with CMOS SPAD
→→Picosecond time resolutionPicosecond time resolution is needed to build sufficiently is needed to build sufficiently precise characterizationprecise characterization
![Page 14: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/14.jpg)
Bioluminescence DetectionBioluminescence Detection
•• PrinciplePrinciple–– Luciferase Luciferase based assaybased assay
methods (550nm) usedmethods (550nm) usedto detect pathogens,to detect pathogens,proteins and DNAproteins and DNAsequencingsequencing
•• Features:Features:–– Sensitivity: Sensitivity: 1010-6-6LxLx
–– pixel size: 250x250umpixel size: 250x250um
[Eltoukhy et al., ISSCC 2004]
![Page 15: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/15.jpg)
Courtesy of H.M.FritzCourtesy of H.M.Fritz
Fluid-dynamics ResearchFluid-dynamics Research
•• Principle:Principle:–– Particle imageParticle image
velocimetry velocimetry (PIV)(PIV)•• Features: Features:
–– 1008x924 pixels1008x924 pixels–– Strobe illuminationStrobe illumination
1kHz1kHz–– Velocity profileVelocity profile
computation viacomputation viacross-correlationcross-correlation
![Page 16: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/16.jpg)
Single Photon ProjectsSingle Photon Projects
•• CurrentCurrent–– 3D camera3D camera–– Near-saturation cortical imaging (EPFL-LS)Near-saturation cortical imaging (EPFL-LS)–– High-speed digital camera (Industry)High-speed digital camera (Industry)–– High-speed shutter digital camera (ColumbiaHigh-speed shutter digital camera (Columbia
University)University)
•• Near futureNear future–– Ultra-high speed Calcium Flux in neurons (EPFL-LS)Ultra-high speed Calcium Flux in neurons (EPFL-LS)–– 3D long-distance camera (Industry)3D long-distance camera (Industry)–– Free space com (Academia)Free space com (Academia)–– Transillumination (EPFL-MT)Transillumination (EPFL-MT)
![Page 17: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/17.jpg)
Single Photon AvalancheSingle Photon AvalancheDiodes (SPADs)Diodes (SPADs)
![Page 18: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/18.jpg)
Φ(λ)Φ(λ)
pp
Photoelectric EffectPhotoelectric Effect
Parameters: QE, p, Parameters: QE, p, λλ
![Page 19: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/19.jpg)
Avalanche EffectAvalanche Effect
Φ(λ)Φ(λ)
ppPhoto-multiplicationPhoto-multiplication
Parameters: QE, p, Parameters: QE, p, λλ
![Page 20: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/20.jpg)
SPAD Operating in Geiger ModeSPAD Operating in Geiger Mode
•• In avalanche mode:In avalanche mode:–– <G> <G> ≅≅ 1000 1000
•• In Geiger mode:In Geiger mode:
–– <G> <G> → ∞ avalanche must be stoppedavalanche must be stopped
[Rochas, TransED, 2002]
V
V
I
conventionalavalancheGeiger
opticalgain<G>
Vbd
1
![Page 21: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/21.jpg)
SPAD Operating in Geiger ModeSPAD Operating in Geiger Mode
passive quenching:
—One photon one cycle (dead time definition)—Thermally generated carriers dark counts
operation cycle:
t
VVbdbd
VVopop
V
V photonarrival
avalanchequenching
SPADrecharge
VVopop
RRqq
VIIAA
tdead0
DEAD TIME
![Page 22: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/22.jpg)
Dark CountsDark CountsIntuitionIntuition::
Traps capture Traps capture photocharges photocharges and release them randomlyand release them randomly(by diffusion, tunneling, etc.) (by diffusion, tunneling, etc.) →→ avalanche is triggered avalanche is triggered →→ spurious spurious
pulsespulses
p+
n-well
n+p+ p+n+
p-sub
![Page 23: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/23.jpg)
CMOS SPAD (ISSCC04)CMOS SPAD (ISSCC04)
OUTOUT
GNDGND
VDD = +5V
RRqq
VH= -25.5V= -25.5VMain features of the pixel• dead time : 32ns• PDP: ~ 12% @ 635nm• dark count rate typ. 350Hz• timing jitter: < 50ps
Vop = VDD + |VH| > Vbd
digitalpulse
![Page 24: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/24.jpg)
SPAD Evolution (ISSCC05)SPAD Evolution (ISSCC05)
FeaturesFeatures–– Pitch reductionPitch reduction
–– Column routing mechanismColumn routing mechanism
–– MOS Quenching resistanceMOS Quenching resistance
VDD
Vop
VDD
COLUMN
ROWSEL
ROWSEL
X
Tq
![Page 25: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/25.jpg)
3D Imaging3D Imaging
![Page 26: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/26.jpg)
GoalGoal
–– Find distance to each point in the sceneFind distance to each point in the scene((rangefindingrangefinding))
–– Build depth map of the sceneBuild depth map of the scene
![Page 27: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/27.jpg)
Specific 3D ApplicationsSpecific 3D Applications
•• SurveyingSurveying–– Geo, constructionGeo, construction
•• AutomotiveAutomotive–– Airbag deploymentAirbag deployment–– Collision detection/avoidanceCollision detection/avoidance
•• SecuritySecurity–– facial recognitionfacial recognition–– room volumetric analysisroom volumetric analysis
•• ConsumerConsumer–– Human-computer interfacesHuman-computer interfaces–– VRVR
•• Medical/ResearchMedical/Research–– Optical tomographyOptical tomography–– Bio-chemical & molecular analysisBio-chemical & molecular analysis
![Page 28: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/28.jpg)
Electronic Electronic TheodolitesTheodolites
•• Some specificationsSome specificationsSub-mm precisionSub-mm precision~100m range~100m rangeLow fps(?)Low fps(?)
![Page 29: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/29.jpg)
Face RecognitionFace Recognition
•• Some specificationsSome specifications–– 600um~1mm precision600um~1mm precision
–– 1~2m range1~2m range
–– 30 fps30 fps
![Page 30: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/30.jpg)
Virtual (Planar) KeyboardVirtual (Planar) Keyboard
•• Some specificationsSome specifications–– 2mm precision2mm precision–– 20~30cm range20~30cm range–– 50-70 fps50-70 fps
![Page 31: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/31.jpg)
3D Imaging:3D Imaging:
optical Time-of-flight (TOF)optical Time-of-flight (TOF)
![Page 32: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/32.jpg)
RangefindingRangefinding
MethodsMethods–– InterferometryInterferometry–– TriangulationTriangulation
–– Time-of-flight (TOF)Time-of-flight (TOF)PrinciplePrinciple
–– LightLight–– RFRF–– UltrasoundUltrasound–– MagnetismMagnetism
![Page 33: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/33.jpg)
Optical TOF Based RangefindingOptical TOF Based Rangefinding
pulsedlight source
TOF measurement
3D image reconstruction d
d = (c/2) TOFd = (c/2) TOF
opticalsensor
target
![Page 34: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/34.jpg)
Conventional ApproachesConventional Approaches
•• SingleSingle highly sensitive detector with mechanical scanning highly sensitive detector with mechanical scanningdevicedevice
•• ProblemsProblems::→→ high-cost high-cost→→ non-portable non-portable
•• ArrayArray of conventional CMOS detectors with high-power of conventional CMOS detectors with high-poweroptical sourceoptical source
•• ProblemsProblems::→→ 10~100W peak power 10~100W peak power→→ potentially unsafe potentially unsafe
![Page 35: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/35.jpg)
Our ApproachOur Approach
•• ArrayArray of Single Photon Avalanche Diodes of Single Photon Avalanche Diodes(SPADs) fabricated in CMOS process(SPADs) fabricated in CMOS process
•• ProsPros::–– Milliwatt Milliwatt laser source (peak power)laser source (peak power)
–– CompactCompact
–– Low costLow cost
![Page 36: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/36.jpg)
3D Imaging:3D Imaging:
System ArchitectureSystem Architecture
![Page 37: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/37.jpg)
Sensor ArchitectureSensor Architecture
SPAD Array
Multiplexer
![Page 38: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/38.jpg)
Sensor PhotomicrographSensor Photomicrograph
LFSRLFSR
1k pixel1k pixelarrayarray
Osc.Osc.
RROOWW
DDEECCOODDEERR
COL. DECOL. DECODERCODER
2.5mm2.5mm
2.8m
m2.
8mm
0.8um0.8umCMOSCMOStechnologytechnology
![Page 39: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/39.jpg)
PULSED LASER
CMOS SPAD SENSOR
CMOSTDC
TARGETSTANDARDLENS
STARTSTOP
Electro-Optical SystemElectro-Optical System
• uncollimated 635nm laser• pulse width : 100ps• peak power : 100~250mW• fR = 40~50MHz
![Page 40: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/40.jpg)
TOF ComputingTOF Computing
TOF
laser pulse
SPAD signal
targethistogram
reference histogram
differentialTOF
start
stop
300ps
t
t
t
t
(M measurements)
![Page 41: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/41.jpg)
3D Imaging:3D Imaging:
MeasurementsMeasurements
![Page 42: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/42.jpg)
Time ResolutionTime Resolution
0 500 1000 1500 20000.0
0.2
0.4
0.6
0.8
1.0
FWHM=72ps
Nor
mal
ized
Cou
nts
Time[ps]
Ti:Ti:Saphire fs Saphire fs pulsed laser pulsed laser (710nm)(710nm)
PP QQ-- SS
PP AADD
11
TACSTARTSTART STOPSTOP
DISCRDISCR.DISCRDISCR.
AnalysisAnalysis
Attenuation FiltersAttenuation Filters
PQ-SPAD Timing Resolution: 50psPQ-SPAD Timing Resolution: 50ps
PQ-SPAD2PQ-SPAD2
![Page 43: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/43.jpg)
DistanceDistance
![Page 44: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/44.jpg)
ModelModel
High resolution Camera:High resolution Camera:20cm x 20cm view20cm x 20cm view
![Page 45: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/45.jpg)
Depth Map ExampleDepth Map Example
•• Lateral resolution:Lateral resolution:–– 64x64 pixels64x64 pixels
•• Depth resolution:Depth resolution:–– 1.3mm (1.3mm (wcwc))
•• Range:Range:–– 3.75m3.75m
[Niclass and Charbon, ISSCC 2005][Niclass and Charbon, ISSCC 2005]
![Page 46: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/46.jpg)
Companion TechniquesCompanion Techniques
•• Multi-exposure super-resolution techniquesMulti-exposure super-resolution techniques–– Microscanning Microscanning (x-/y- (x-/y- piezo piezo elements)elements)–– S/R algorithm (LCAV, Prof. Vetterli)S/R algorithm (LCAV, Prof. Vetterli)
•• CMOS time-to-digital-converter (TDC)CMOS time-to-digital-converter (TDC)IntegrationIntegration–– Current resolution: 100ps (0.8Current resolution: 100ps (0.8µµm)m)–– Target (under test): 25ps (0.18Target (under test): 25ps (0.18µµm)m)
![Page 47: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/47.jpg)
Microscanning Microscanning MechanismMechanism
58µm
7µm
58µm(+29µm, 0)
(0,-29µm)
(-29µm, 0)
(0,+29µm)
![Page 48: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/48.jpg)
CMOS Time-to-Digital-ConverterCMOS Time-to-Digital-Converter
•• CMOS TDCCMOS TDC(RSUT)(RSUT)
•• Resolution:Resolution:100ps100ps
•• Accuracy:Accuracy:<30ps<30ps
•• Clock:Clock:40MHz40MHz
![Page 49: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/49.jpg)
3D Performance Summary3D Performance Summary
<1mW<1mW
100ps100ps
50MHz50MHz
100mW/0.5mW100mW/0.5mW
268ps268ps
0.15 ~ 4m0.15 ~ 4m
1.8mm1.8mm
32x3232x32
250mW/0.7mW250mW/0.7mWOpt. peak/average powerOpt. peak/average power
100ps100psLight pulse widthLight pulse width
350ps350psSingle measurementSingle measurementuncertainty uncertainty σ(τ)σ(τ)
40MHz40MHzRepetition rateRepetition rate
< 6mW< 6mWChip power dissipationChip power dissipation
0.15 ~ 3.75m0.15 ~ 3.75mAxial rangeAxial range
1.3mm1.3mmW.C. axial uncertainty W.C. axial uncertainty σ(σ(dd))
(10k (10k measmeas.).)
64x6464x64ParameterParameter
![Page 50: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/50.jpg)
2D Imaging2D Imaging
![Page 51: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/51.jpg)
CCDs CCDs (e.g.)(e.g.)
RO
W D
EC
OD
ER
A/D
AMPLIFIERS, S/H
aMUX
CH
AR
GE
CO
UP
LE
D P
AT
HC
HA
RG
E C
OU
PL
ED
PA
TH
CO
L SEL
CO
L SELDD
Three conversions:Three conversions:
Photons → q → V → D
Five sources of noise:Five sources of noise:
Shot, 1/f, kT/C, read-out, ADC
q
V
![Page 52: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/52.jpg)
Conventional CMOS APS (e.g.)Conventional CMOS APS (e.g.)
RO
W D
EC
OD
ER
A/D
AMPLIFIERS, S/H
aMUX
ROW SELECTROW SELECT
PIX
EL
RE
AD
OU
TP
IXE
L R
EA
DO
UT
CO
L SEL
CO
L SELDD
R
VV22
VV11
Three conversions:Three conversions:
Photons → q → V1,V2 → D
Five sources of noise:Five sources of noise:
Shot, 1/f, kT/C, read-out, ADC
![Page 53: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/53.jpg)
Performance of InterestPerformance of Interest
•• Dynamic rangeDynamic range–– Ratio between the smallest detectable and maximumRatio between the smallest detectable and maximum
detectable intensitydetectable intensity
•• SNRSNR–– Ratio between signal due to intensity and noiseRatio between signal due to intensity and noise
•• FPSFPS–– Frames-per-secondFrames-per-second
•• PitchPitch–– Distance between pixelsDistance between pixels
•• …… and FPN, smearing, blooming, color, etc. and FPN, smearing, blooming, color, etc.
![Page 54: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/54.jpg)
Limiting FactorsLimiting Factors
•• Miminum Miminum # detectable photons# detectable photons
•• NoiseNoise–– Read-out noiseRead-out noise
–– kTkT/C noise/C noise
–– 1/f noise1/f noise
•• SaturationSaturation–– Limited number of available chargesLimited number of available charges
•• Speed (10kfps)Speed (10kfps)
![Page 55: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/55.jpg)
Imagers Based on SPADsImagers Based on SPADs
RO
W D
EC
OD
ER
dMUX
ROW SELECTROW SELECT
PIX
EL
RE
AD
OU
TP
IXE
L R
EA
DO
UT
CO
L SEL
CO
L SELDD
One conversion:One conversion:
Photons → D
SPAD
One and 1/2 source of noise:One and 1/2 source of noise:
Shot, DCR, (digital noise)
![Page 56: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/56.jpg)
SPAD Arch Limiting FactorsSPAD Arch Limiting Factors
•• Miminum Miminum # detectable photons# detectable photons–– Limited by DCRLimited by DCR
•• NoiseNoise–– Read-out noiseRead-out noise–– kTkT/C noise/C noise–– 1/f noise1/f noise
•• SaturationSaturation–– Limited by dead time and read-out bandwidthLimited by dead time and read-out bandwidth
•• SpeedSpeed–– Shown 250kppsShown 250kpps
![Page 57: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/57.jpg)
New ChallengesNew Challenges
Departure from amplitude based imagerDeparture from amplitude based imager Therefore:Therefore:
–– Local evaluation of intensityLocal evaluation of intensity–– Intensity storageIntensity storage–– Digital noiseDigital noise–– Low-power operationLow-power operation
ExampleExample::–– R rowsR rows–– share column of bandwidth B share column of bandwidth B ≤≤ 1/ 1/ttDD
Then, in average every pixel has B/R bandwidth, thusThen, in average every pixel has B/R bandwidth, thusdynamic range dynamic range ≈≈ 20{Log(B/R)-Log(DCR)} 20{Log(B/R)-Log(DCR)}
Architecture directly influences Performance Architecture directly influences Performance
![Page 58: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/58.jpg)
Importance of Pixel TimingImportance of Pixel Timing
(FWHM)
Performance:Performance:•• Pertiming Pertiming jitter: < 50psjitter: < 50ps•• dead time : 32nsdead time : 32ns
Potential impact:Potential impact:•• SaturationSaturation•• SmearingSmearing
![Page 59: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/59.jpg)
Experimental Setup for ImagerExperimental Setup for Imager
SPAD Array
Multiplexer
Counter
![Page 60: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/60.jpg)
DCRDCR
![Page 61: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/61.jpg)
Sensitivity and Dynamic RangeSensitivity and Dynamic Range
120dB
Cou
nt [H
z]
Optical Intensity [W/cm2]
71dB
# e-
[-]
103
104
105
106
102
CMOS APS
101
107
108
10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3
![Page 62: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/62.jpg)
LinearityLinearity
![Page 63: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/63.jpg)
Photon Detection ProbabilityPhoton Detection Probability
400 500 600 700 800 900 1000
10-3
10-2
10-1
Pho
ton
Det
ectio
n P
roba
bilit
y [1
]
Wavelength [nm]
12% @ 635nmPeak: 26% @ 460nm
![Page 64: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/64.jpg)
ExampleExample
High resolution Camera:High resolution Camera:20cm x 20cm view20cm x 20cm view
SPAD Camera:SPAD Camera:32x32pixels32x32pixels
![Page 65: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/65.jpg)
High Speed ExampleHigh Speed Example
FeaturesFeatures–– No measurable thermal noise (Poisson noiseNo measurable thermal noise (Poisson noise
dominates)dominates)–– No measurable cross-talk, blooming, smearingNo measurable cross-talk, blooming, smearing
4µs 10µs 25µs 100µs 1ms
[Niclass, Rochas, Besse, Popovic, and Charbon, Transducers 2005][Niclass, Rochas, Besse, Popovic, and Charbon, Transducers 2005]
![Page 66: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/66.jpg)
High Sensitivity ExampleHigh Sensitivity Example
Deep TwilightDeep Twilight Full MoonFull Moon Quarter MoonQuarter Moon StarlightStarlight
2 lx2 lx 22××1010-1-1 lx lx 22××1010-2-2 lx lx 22××1010-3-3 lx lx
![Page 67: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/67.jpg)
Intensity Performance SummaryIntensity Performance Summary
44µµssMinimum Integration TimeMinimum Integration Time
350Hz/75Hz350Hz/75HzDark Count Rate (DCR) T=room/0Dark Count Rate (DCR) T=room/0ooCC
1.1%1.1%Pixel Pixel FillfactorFillfactor
1-26%1-26%Photon Detection Probability (PDP)Photon Detection Probability (PDP)
1.3x101.3x10-3 -3 LxLxOptical Intensity @ SNR=0dBOptical Intensity @ SNR=0dB
120dB120dBMaximum Pixel Dynamic RangeMaximum Pixel Dynamic Range
ValueValueParameterParameter
![Page 68: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/68.jpg)
ConclusionsConclusions
•• Significance and applications of 3DSignificance and applications of 3Dimagingimaging
•• Fabricated 1k pixel SPAD array inFabricated 1k pixel SPAD array instandard CMOS technologystandard CMOS technology
•• psec (psec (µµm) accuracy reached with mWm) accuracy reached with mWlaser sourcelaser source
•• Chip power dissipation <1mWChip power dissipation <1mW•• High sensitivity, dynamic range, andHigh sensitivity, dynamic range, and
linearity for 2D applicationslinearity for 2D applications
![Page 69: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/69.jpg)
Future WorkFuture Work
•• VGA pixel arraysVGA pixel arrays
•• Count / Read-out parallelizationCount / Read-out parallelization
•• Pixel-level and column-level time-to-digitalPixel-level and column-level time-to-digitalconverters (TDCs)converters (TDCs)
•• Pitch reductionPitch reduction
•• On-chip basic computationOn-chip basic computation
•• ColorColor
![Page 70: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/70.jpg)
Future DirectionsFuture Directions
•• qCADqCAD
•• Single photon telecommunicationsSingle photon telecommunications
•• Accelerated DNA sequencingAccelerated DNA sequencing
•• Laserless Laserless ultra-short range quantumultra-short range quantumcommunication and computingcommunication and computing
•• Data managementData management
![Page 71: CMOS Single Photon Detectorsdemichel/si.epfl.ch/files... · –Build depth map of the scene. Specific 3D Applications •Surveying –Geo, construction •Automotive –Airbag deployment](https://reader034.fdocuments.in/reader034/viewer/2022042106/5e8561b54c70236078601067/html5/thumbnails/71.jpg)
http://aqua.http://aqua.epflepfl..chch