Schottky Junction Photodiodes

Schottky kunction type metal-

semiconductor-metal (MSM) type
photodetectors. (Courtesy of
Hamamatsu)

GaAsP Schottky junction GaP Schottky junction

AlGaN Scottky junction
photodiode for 190-680 photodiode for 190 nm
photodiode for UV
nm detection, from UV to to 550 nm detection.
detection (Courtesy of
red (Courtesy of (Courtesy of
sglux, Germany)
Hamamatsu) Hamamatsu)
 schottky .

*
ohmic
.

reverse

mn
Schottky Junction tunneling r
evereaudoan
> breakdown

(a) Metal and an n-type semiconductor before contact. The metal work function m is
greater than that of the n-type semiconductor (b) A Schottky junction forms between the
metal and the semiconductor. There is a depletion region in the semiconductor next to the
metal and a built-in field Eo (c) Typical I vs. V characteristics of a Schottky contact device.
 Schottky Junction

Reverse biased Schottky junction and the dark current due to the
injection of electrons from the metal into the semiconductor over
the barrier B.
 Schottky Junction
Ψ

LEFT: Photogeneration in the depletion region and the resulting

photocurrent. RIGHT: The Schottky junction photodetector
 Schottky Junction Photodiodes
Schottky junction based photodetectors and some of their features. R and F are the rise and fall times
of the output of the photodetector for an optical pulse input. The rise and fall times represent the times
required for the output to rise from 10% to 90% of its final steady state value and to fall from 90% to
10% of its value before the optical pulse is turned off.

 range Rpeak (at peak) Jdark

Schottky junction Features with typical values
nm (A/W) per mm2
GaAsP 190−680 0.18 (610 nm) 5 pA UV to red, R = 3.5 s. (G1126 seriesa)
GaP 190−550 0.12 (440 nm) 5 pA UV to green, R = 5 s. (G1961a)
Measurement of UV; blind to visible light.
AlGaN 220−375 0.13 (350 nm) 1 pA
(AG38Sb)
Wide bandwidth > 10 GHz, R < 30 ps. (UPD-30-
GaAs 320−900 0.2 (830 nm) ~ 1 nA
VSG-Pc)
Optical high speed measurements, R = 80 ps, F =
InGaAs MSM 850−1650 0.4 (1300 nm) 5 A
160 ps. (G7096a)
Optical high speed measurements, R = 30 ps, F =
GaAs MSM 450−870 0.3 (850 nm) 0.1 nA
30 ps. (G4176a)

aHamamatsu (Japan); bsglux (Germany); cAlphalas

 Schottky Junction Photodiodes
"

LEFT: The metal electrodes are on the surface of the

semiconductor crystal (which is grown on a suitable substrate).
RIGHT: The electrodes are configured to be interdigital and on
the surface of the crystal.
 Schottky Junction Photodiodes

LEFT: Two neighboring Schottky junctions are connected end-to-end, but in opposite
directions as shown for A and B. The energy band diagram without any bias is
symmetrical. The grey areas represent the SCL1 and SCL2 at A and B. RIGHT: Under a
sufficiently large bias, the SCL1 from A extends and meets that from B so that the whole
semiconductor between the electrodes is depleted. There is a large field in this region, and
the photogenerated EHPs become separated and then drifted, which results in a
photocurrent.
Phototransistor
Transistor action

IE  exp(eVBE/kBT)
Gain

Iph  Ipho
 Photoconductive Detectors

PbS (lead sulfide) photoconductive detectors for the detection of IR

radiation up to 2.9 m. They are typically used in such applications
as radiation thermometers, flame monitors, water content and food
ingredient analyzers, spectrophotometers etc.. (P9217 series)
(Courtesy of Hamamatsu.)
 Photoconductive Detectors
dark J
efunn Mpp )
=
: +

photo :
Jphot
.
=
efunonMpop )

A semiconductor slab of length , width w and depth d

is illuminated with light of wavelength 
 Photoconductive Detectors

A photoconductor with ohmic contacts (contacts not limiting carrier entry) can
exhibit gain. As the slow hole drifts through the photoconductors, many fast
electrons enter and drift through the photoconductor because, at any instant, the
photoconductor must be neutral. Electrons drift faster which means as one
leaves, another must enter.
 Photoconductivity  and Photocurrent Density Jph

Photon flux = ph Photogeneration rate

 I  i = Internal
i  
i A ph hv  I quantum
g ph = =  = i
Ad d hcd efficiency
Steady state illumination

dn n
= gph − =0
dt 
Photoconductivity  = een + ehp = en(e + h)

eiI ( e + h ) V
 = J ph =  = E
hcd 
 Photoconductive Gain
Photon flux = ph

I ph wdJ ph iIw ( e + h )E
Rate of electron flow = = =
e e hc

iI
Rate of electron generation = ( Volume)gph = ( wd)gph = w
hc

Photoconductive gain G
Rate of electron flow in external circuit  (  e +  h )E
G= =
Rate of electron generation by light absorption 
 Photoconductive Gain
Photon flux = ph Rate of electron flow in external circuit  (  e +  h )E
G= =
Rate of electron generation by light absorption 

Electron and hole transit times (time to

cross the semiconductor) are

te =  / (eE) Electron

th =  / (hE) Hole

Photoconductive gain G

 h    
G = + = 1 + 
te t h te  e 
 Basic Photodiode Circuits

(a) The photodiode is reverse biased through RL and illuminated. Definitions of positive I and V are
shown as if the photodiode were forward biased. (b) I−V characteristics of the photodiode with
positive I and V definitions in (a). The load line represents the behavior of the load R. The operating
point is P where the current and voltage are I and V.
 Basic Photodiode Circuits: The Load Line

P is the
operating point
V = − V
I = −  A

I  Iph

The current through RL is

I = − (V + Vr) / RL

This is the load line shown in the figure. P is the intersection of the
load line with the photodiode I vs. V curve and is the operating point.
 Basic Photodiode Circuits

A simple circuit for the measurement of the photocurrent Iph by using a current-voltage
converter or a transimpedance amplifier. The reverse bias Vr is a positive number. Note
that biasing circuit for the op amp is not shown.
 Photodiode Equivalent Circuit
.

(a) A real photodiode has series and parallel resistances Rs and Rp and a SCL
capacitance Cdep. A and C represent anode and cathode terminals. (b) The equivalent
circuit of a photodiodes. For ac (or transient) signals, the battery can be shorted since ac
signals will simply pass through the battery.
 a lu (I )

Reverse Biased Photodiode Equivalent Circuit I2= I

OV = W☆ ωi
-
1

Total capacitance = ideal

'
V V √
,

=
wtoomV

photodiode SCL -
Isexphv+ )

Rp = Shunt (parallel) resistance

=
,

capacitance + terminal In loI ( =

Isexplvmg

volntIs ]
capacitance
U. -

Ideal photodiode Rs = Series resistance VE

I
=velu
(
7
)

DV :
K Wizhie hnlo

are bomv
 Cutoff Frequency fc

V(t)

The cutoff frequency or the bandwidth of the PD

1 1 1
fc =  
2ReqCt 2 ( Rs + RL )Ct 2RLCt
Req is equivalent resistance and Assumption
represents (Rs + RL) in parallel with Rp Drift time of carriers is much less than 1/fc.
Response is not limited by drift and diffusion
times of caries within the device.
 A Commercial Photoreceiver
Thermoelectric (TEC) cooler TEC Current in direction
APD bias
Temperature
sensor (Tsense)
APD Base/Collector

Emitterr
TEC Current out
direction
Op amp bias

Output

Op amp

A photoreceiver that has an InGaAs APD and peripheral electronics (ICs) to achieve
high gain and high sensitivity. There is also a thermoelectric cooler (TEC) and a
temperature sensor (TSense). Courtesy of Voxtel Inc ([Link])
 Pulsed Excitation
Po(t)

Short light pulse

Large resistor to t
bias the PD
Very fast buffer
or amplifier that
PD does not load RL.

Reverse biases
the PD

Coupling capacitor that

allows ac/transient signal
coupling
Load resistor for
Bias or shorting capacitor to short RB
developing a
and the battery for the transient
photocurrent. It is a short for voltage signal
ac/transient signals
The Experiment
Pulsed Excitation

Assume: The buffer is extremely fast

and does not limit the response

Rise time Fall time

Are these related to fc?

 Rise and Fall Times, and Bandwidth
 = (Rs+ RL)Ct  RLCt

V(t)  V100exp(−t/)
Measured from toff

Rise time Very roughly, R  F

Fall time

F = 2.2  = (Rs+ RL)Ct  RLCt

1 1 0.35 350 MHz

fc  = = =
2RLCt 2 F  F ( ns)
 Pulsed Excitation
Non-RLCt response
Response due to the diffusion and drift of photogenerated carriers
Assume Rs + RL is very small so that (Rs + RL)Ct is negligible

Photocurrent
Diffusion of
Slow
carriers in the
neutral region Fast Drift of carriers in the
depletion region
Drift of carriers Diffusion of
in the depletion Fast carriers in the
region Slow neutral region
t
 Noise in Photodiodes
Consider a receiver with a photodiode and a sampling resistor RL
The amplifier A is assumed noiseless
i(t) Constant illumination
Noise current = Total
What is the RMS of fluctuations? RMS current fluctuations

Consider constant illumination Po

Total current without noise = Dark current (Id) + Photocurrent (Iph) = “Constant”
Observed Current = Dark current + Photocurrent and Fluctuations (Noise)
What is this “Noise” ?

We can represent the “noise current” RMS of fluctuatio ns = i (t )2

by the RMS of fluctuations
 Noise in Photodiodes
i(t)
Constant illumination
Noise current = Total
RMS current fluctuations
What is the RMS of fluctuations?

The dark current has shot noise or fluctuations about Id,

in-dark = (2eIdB)1/2 B = Bandwidth

Quantum noise is due to the photon nature of light and its effects are the same
as shot noise. Photocurrent has quantum noise or shot noise

in-quantum = (2eIphB)1/2
Noise in Photodiodes

Total shot noise current, in

i =i
2
n
2
n −dark +i 2
n −quantum Vout =
Avin

Vout AimR

in = [2e(Id + Iph)B]1/2
=
.
,
noise

We can conceptually view the photodetector current as

Id + Iph + in
This flows through a load resistor RL and voltage across RL
is amplified by A to give Vout
The noise voltage (RMS) due to shot noise in PD = inRLA
Noise in Photodiodes

Total current flowing into RL has three components:

Id = Dark current. In principle, we can subtract this or block it with a

capacitor if Iph is an ac (transient) signal.

Iph = Photocurrent. This is the signal. We need this. It could be a steady

or varying (ac or transient) signal.

in = Total shot noise. Due to shot noise from Id and Iph. We cannot
eliminate this.
 Noise in Photodiodes

The resistor RL exhibits thermal noise (Johnson noise)

Power in thermal fluctuations in RL = 4kBTB

2  RL i 2 = 4k BTB i = Current in RL
i
1/2
 4k BTB 
ith = Thermal noise current from RL =  
R
 L 
 shot noise

Summary of Noise in PD and RL

:

in Vk :
inin R .

?
=
in .

R(

total noise power

: shot noise power

t Thermal noise power

'
= in RL + 4 kBTB

= ( 2e [ Ia p
1 ] R.
)B + 4 kBTB

Bowear
SINR( sigmal amoise

ignal
" porer

R
Power in shot noise in PD = in2RL = [2e(Id + Iph)B]RL
=

.+ 4 KBTB

Power in thermal fluctuations in RL = 4kBTB

Important Note: Total noise is always found by first summing the average powers involved
in individual fluctuations e.g. power in shot noise + power in thermal noise

Noise in the amplifier A must also be included

See advanced textbooks
 Signal to Noise Ratio

Signal Power
SNR = |

Noise Power
2 2
I ph RL I ph
SNR = =
i RL + 4k BTB 2e( I + I ph ) B  +
4k BTB
2
n
d
RL
3
^

Important Note: Total noise is always found by first summing the average powers involved
in individual fluctuations e.g. power in shot noise + power in thermal noise
 [ A /w )
doesponsivity
R P
IiBn
=
.

sisghal Roower
I

Iph :
RP .
= RP , =
[ Ze ( In tIpm ) B 34 =
in

√
moisepoaver

NEP = :
[
[ Re

P ,
:

Rn

SNR =

Ipm
in =
heEIphtIa3B

P ,

= R

-
2 e ( m +☆
NEP ≡
 Noise Equivalent Power: NEP
Definition
Input power for SNR = 1 P1
NEP = = 1/ 2
Bandwidth B
NEP is defined as the required optical input power
to achieve a SNR of 1 within a bandwidth of 1 Hz

NEP = 1 / 2 = 2e( I d + I ph )
P1 1 1/ 2

B R

Units for NEP are W Hz–1/2

 Detectivity, D
Definition
1
Detectivity =
NEP
Specific detectivity D*
1/ 2
A
D* =
NEP
Units for D* are cm Hz-1/2 W-1, or Jones
 NEP and Detectivity of Photodetectors
Typical noise characteristics of a few selected commercial photodetectors. PC means a
photoconductive detector, whose photoconductivity is used to detect light. For PC detectors,
what is important is the dark resistance Rd, which depends on the temperature.

GaP Si Ge InGaAs PbS (PC) PbSe (PC) InSb (PC)

Photodiode
Schottky pin pin pin −10C −10 C −10C

peak (m) 0.44 0.96 1.5 1.55 2.4 4.1 5.5

Id or Rd 10 pA 0.4 nA 3 A 5 nA 0.1−1 M 0.1−1 M 1−10 k

NEP W Hz-1/2 5.4×10-15 1.6×10-14 1×10-12 4×10-14 - -

D* cm Hz1/2/ W 1×1013 1×1012 1×1011 5×1012 1×109 5×109 1×109

NEP = 1 / 2 = 2e( I d + I ph )
P1 1 1/ 2

B R
 NEP and Dark Current

Ra λ

The dependence of NEP (W Hz-1/2) on the photodetector dark current Id for Si and
InGaAs pin, Ge pn junction, and GaP Schottky photodiodes. Dashed lines indicate
observed trends. Filled circle, Si pin; open circle, InGaAs pin at 25 C, open
diamond at −10 C, open square, −20 C; inverted triangle, Ge pn; triangle, GaAsP
Schottky. (Data extracted from datasheets of 35 commercial photodiodes)
 Noise in Avalanche Photodiode (APD)
Ideally the shot noise is simply multiplied so that we
should expect
in-APD = Min = M[2e(Ido + Ipho)B]1/2
in-APD = [2e(Ido + Ipho)M2B]1/2
But, we observe excess noise above this shot noise
Avalanche Noise

in-APD = [2e(Ido + Ipho 2

)M FB]1/2

Excess Noise Factor

 Noise in Avalanche Photodiode (APD)
APDs exhibit excess avalanche noise due to the randomness of the impact
ionization process in the multiplication region. Some carriers travel far
and some short distances within this zone before they cause impact
ionization
Excess Avalanche Noise Factor F
in-APD = [2e(Ido + Ipho)M2FB]1/2
Excess Noise Factor
F  Mx where x is an index that depends on the
semiconductor, the APD structure and the type of carrier
that initiates the avalanche (electron or hole)
For Si APDs, x is 0.3−0.5 whereas for Ge and III-V (such as InGaAs)
alloys it is 0.7−1