OutlineOutlineIR SPADs: Geiger-mode avalanche gain
Nanowire detectors: Phototransistive gain
Id l i l h t d t tIdeal single-photon detector: Both gain mechanisms in one device true PMT-like h t i ti ith hi h IR d t ti ffi icharacteristics with high IR detection efficiency
Summary
OutlineOutlineIR SPADs: Geiger-mode avalanche gain
Nanowire detectors: Phototransistive gain
Id l i l h t d t tIdeal single-photon detector: Both gain mechanisms in one device true PMT-like h t i ti ith hi h IR d t ti ffi icharacteristics with high IR detection efficiency
Summary
Challenges for Geiger Mode APDsChallenges for Geiger Mode APDs
M t b t d ith hi i itMust be operated with quenching circuitsDifficult to make large array
Dark count and after pulsing
Excess noise and timing jitter
Reliability, degradation, and property drift
Geiger Mode APDsGeiger Mode APDsPassive quenching:q g
Active quenching:
Gated mode operation: Von
Voff
Self Quenching and Self RecoverySelf Quenching and Self RecoveryUse band offsets to form energy barrier “TCB” (transient carrier buffer)Avalanche carriers collect at barrierEm is reduced below Ebr, quenching deviceCarriers slowly escape from barrier, recovering device
Implementing the Self-Quenched SPAD Using B d G E i iBand Gap Engineering
Latticed matched to InP Latticed matched to InP system
absorptio
TCB
multiplican tion
Device Structure and Bandstructure Simulation
p-InP
i-InP Multiplication region
InGaAsP Grading Layern-InP charge layer
InAlAs TCB region
InGaAsP Grading Layer
InAlGaAs Grading Layer
InGaAs Absorption Layer
InP Substrate
Process FlowProcess Flow
Step1: Pattern SiO2
Step2: Mesa Etch
Step3: Polyimide PassivationStep3: Polyimide Passivation
Step4: Metalization
I-V characteristicsI V characteristics
10-6
10-5
λ =1550nm10um Device
120KDark 160KDark200KLi h
10-7
10 10um Device 200KLight 240KDark 300KDark120KLight
10-9
10-8
rren
t [A
]
120KLight 160KLight 200KDark240KLight
10-10
10
Cu 240KLight
300KLight
40 35 30 25 20 15 10 5 010-12
10-11
-40 -35 -30 -25 -20 -15 -10 -5 0
Bias [V]
SPAD Characterization Setup
to video monitor
InGaAs
p
InGaAsFPA
1534 nmprobe
Optical Power Monitor
6 dB
to DSO
pbeam
Av=800Nf=1.1dB
6 dBsplitter
to 1 Gsa/s
Detectordie
Vbias
Nf 1.1dB2 GHz 450 MHz
low pass
digitizer
Single Photon Detection Efficiency vs Dark Count
0 12
0.14
0.16 120K 160K 200K240K
10um Deviceλ=1550nm
cien
cy
0.08
0.10
0.12 240Ket
ectio
n Ef
fi
0.04
0.06
le P
hoto
n D
0.0 2.0x106 4.0x106 6.0x106 8.0x106 1.0x1070.00
0.02Sin
gl
Dark Count Rate [Hz]
Poor device passivation causes very high after pulsing rate, limiting the device performancedevice performance.
Self-Recovery Time
70
8010um Device160Ks]
1.0
a.u]
32 0V
40
50
60
160KRecover to 63%
y Ti
me
to 6
3% [n
s
0.6
0.8
rk C
ount
Rat
e [ 32.0V
32.2V 32.4V 32.6V 32.8V
10
20
30
Dev
ice
Rec
over
y
0 0
0.2
0.4
Nor
mal
ized
Dar
31.8 32.0 32.2 32.4 32.6 32.8 33.00
Reverse Bias [V]
0 20 40 60 80 100 1200.0N
Time [ns]
Recovery time decreases rapidly with the overbias (bias aboveRecovery time decreases rapidly with the overbias (bias above breakdown).
Recovery time eventually approaches the carrier (hole) escape timeRecovery time eventually approaches the carrier (hole) escape time as determined by the (valence)band offset (80meV in current design).
Pulse Height Histogram
The avalanche pulse height has very narrow distribution.p g y
The excess noise factor measured by signal intensity at a gain of 5x105-106: 1.001 – 1.003
K.Zhao, A.Zhang,Y.Lo and W.Farr,Applied physics letter 91,2007
The excess noise (by total charge contained in the pulse): ~1.007
OutlineOutlineIR SPADs: Geiger-mode avalanche gain
Nanowire detectors: Phototransistive gain
Id l i l h t d t tIdeal single-photon detector: Both gain mechanisms in one device true PMT-like h t i ti ith hi h IR d t ti ffi icharacteristics with high IR detection efficiency
Summary
Phototransistive Gain (P-Type)Phototransistive Gain (P Type)
hν1. Without light, NW is depleted of majority carriers due pppp
pppp
to large number of surface states
2. When illuminated, e/h pairs generated in body of NWp p
3. Due to the lateral electric field, minority carriers move to surface and recombine with trapped majority carrier
4 Photogenerated majority carriers are confined to4. Photogenerated majority carriers are confined to the center of the wire and generate photocurrent with applied bias
5 Photocurrent stops when majority carrier recaptured5. Photocurrent stops when majority carrier recaptured at the surface
6. Gain occurs when the lifetime of majority carriers is h l th th i t it ti Gaindc =
τcmuch longer than their transit time. Gaindc tt
Gain vs Doping in the WiresGain vs Doping in the Wires100000
Nga=1e11
1e-6 W/cm2 1e-4 W/cm2
10000 Nga=3e10
1e-6 W/cm21 4 W/ 2
1000
10000 1e-2 W/cm2
Gai
n
1000
1e-4 W/cm2 1e-2 W/cm2
Gai
n
1E14 1E15 1E16 1E17 1E18 1E1910
100
1E14 1E15 1E16 1E17 1E18 1E1910
100
P d i ( 3)P-doping (cm3)
P-doping (cm3)
10000
100000
1000000 Nga=3e11
1e-6 W/cm2 1e-4 W/cm2 1e-2 W/cm2
100
1000
10000
Gai
n
1E14 1E15 1E16 1E17 1E18 1E191
10
P-doping (cm3)
Gain SaturationGain SaturationNga=1e11, Na=4e16
100000g ,
1000010000
Gai
n
10001E-7 1E-6 1E-5 1E-4 1E-3 0.01 0.1 1
Intensity (W/cm2)
Advantages of Top-Down FabricationAdvantages of Top Down FabricationTop-down approach for etching NW’s advantageous for:
Control of geometry (diameter, height, pitch)
Exact placement of wire
Control of exact number per contact
Single crystalline NW
Doping through conventional methods
NW Fabrication Using Nanoimprintingg p gNanowire diameter: 200 nmPitch: 1 umPhysical fill factor: ~ 3%y
Before etch After etch
Vertical Device FabricationVertical Device Fabrication
2. E-beam pattern Ni dots and RIE/ICP etch (SF6/C4F8)
3. Spin on dielectric(PMGI, PI, SOG)
1. Epi Si wafer
5. Pattern and deposit ITO transparent contact
window
4. Plasma etch to expose wire tips
6. Pattern and deposit Ti/Au contacts
1μm1μm 5μm
Room Temperature PhotoresponseRoom Temperature Photoresponse
50.0n
60.0n
10 NWElectrode spacing 8um390+-50nm Filter
100000
20 0n
30.0n
40.0n
Cur
rent
(A)
Dark 8uW/cm2
80uW/cm2
800uW/cm2
10000
Gai
n0 0 0 1 0 2 0 3 0 4 0 5
0.0
10.0n
20.0n 800uW/cm 8mW/cm2
100
1000
0.0 0.1 0.2 0.3 0.4 0.5Bias (V)
1E7 1E8 1E9100
Absorbed Photon Flux (#/sec)
G@0 5V>35,000G =τc
h
th =I e
η P hν @0.5V ,tth ηQEPopt hν
Photoresponsep
10-10 b
Visible (650nm) response
IR (1550 nm) response
10-11
A)
13
10-12
urre
nt (A
10-14
10-13
84 K 110 K130 K
Pho
tocu
10-16 10-14 10-12 10-10 10-810-15
130 K 150 K
Incident Power (W)Power: net incident power over 1um2 area.
Gain vs Incident PowerGain vs Incident PowerVisible (650nm)
responseIR (1550 nm)
response
104
105
90K 110K130K
W)
aresponse response
10
100
(A/W
)
103
104
150K
nsiv
ity (A
/W
0.1
1
espo
nsiv
ity
82 K110K
101
102
Res
pon
1E-3
0.01Ph
oto
R 110K 130K 150K 170K 200K
10-16 10-15 10-14 10-13
Incident Power (W)
1E-12 1E-11 1E-10 1E-9 1E-81E 3
Incident Light Power per Nanowire (W)
OutlineOutlineIR SPADs: Geiger-mode avalanche gain
Nanowire detectors: Phototransistive gain
Id l i l h t d t tIdeal single-photon detector: Both gain mechanisms in one device true PMT-like h t i ti ith hi h IR d t ti ffi icharacteristics with high IR detection efficiency
Summary
Comparisons Between SPADs and NW Detectors
Geiger-mode SPADs NW DetectorgAvalanche gain Phototransistor gainBiased above breakdown Low bias √√E t l iti t bi M h l iti t biExtremely sensitive to bias and temperature
Much less sensitive to bias and temperature √√
High detection efficiency √√ Low detection efficiencyHi h i l h d L i l h dHigh single-photon data rate (GHz) √√
Low single-photon data rate (< MHz)
Dark count/after pulsing Less dark count √√√√Lower reliability Higher reliability √√
Can we combine the designs to achieve the merits of both?
That is, to achieve a multiplication gain of 100-1000 and a phototransistor gain of 10-1000 for a net gain of 105-6. This will behave like a PMT with high IR detection efficiency.
Device ConceptDevice Concept
0 V 5 V
InAlAs TCB
NW transistor
InPmultiplication
InGaAs
multiplicationp-i-n
15 V
absorption