Outline

Optical Absorption

Principle of optical detection.

Know various optical detectors like:

photodiodes,
p-i-n diodes and
avalanche diodes.
Define figures of merit of the detectors.

Solar Cell

Know about different sources of noise in detectors

and the significance of signal to noise ratio.
1
 Optical Absorption
The energy of photon is E= hv
where h is the planks constant and v is the
frequency
The wavelength = c/v = ch/E= 1.24/E um
where c is the speed of light
Photon can interacts with
- semiconductor lattice
- impurity atoms (donors or acceptors)
- defects within the semiconductor
Most important Interaction with Valence
Electrons
After collision, electrons may be elevated into
the conduction band causes the generation of
excess electron-hole pairs. 2
 Optical Absorption
When semiconductor is illuminated with light, photon may
-be absorbed or
- be propagate through

Depends on: Photon Energy (E) and Bandgap of SC (Eg)

If E < Eg : Photon is not absorbed and SC are
transparent.
If E => Eg : Photon can elevate valence electron into
conduction band causes electron in conduction band and
hole in valence band (excess electron-hole pair generation)
E = Eg is enough but excess photon energy provides
additional kinetic energy to free carriers and will be dissipated
as heat. 3
 Absorption Coefficient
dx
Iv(x) Iv(x+dx)

Letthe intensity of an incident photon at x is: I v(x) energy/cm2-s

and the photon flux emerging at a distance x+dx is: Iv(x+dx)

The energy absorbed per unit time in the distance dx is: Iv(x)dx
Where is the absorption coefficient is the relative number of
photons absorbed per unit distance.
We get Iv(x+dx) -Iv(x) = = -Iv(x)dx
= -Iv(x)
If Iv(0) = Iv0
Then the solution is given by; Iv(x) = Iv0e-x

The intensity decreases exponentially with

4
PHOTODETECTORS

5
 Photodetectors
A photodetector is a device which absorbs light and
converts the optical energy to measurable electric
energy.

Detectors are classified as:

Thermal detectors or Photoconductor

Photon detectors

6
 Thermal detectors

When light falls on the device, it raises its

temperature, which, in turn, changes the electrical
properties of the device material, like its electrical
conductivity.
Photon generated excess electro-hole pairs change
the conductivity of a semiconductor.

Examples of thermal detectors are:

thermopile (which is a series of thermocouples),

7
 Photon detectors
Photon detectors work on the principle of conversion
of photons to electrons. Unlike the thermal detectors,
such detectors are based on the rate of absorption of
photons rather than on the rate of energy absorption.

However, a device may absorb photons only if the

energy of incident photons is above a certain minimum
threshold.

Examples of Photon detectors: Photodiodes

8
Light Detection Process

Photoconductivity

Photovoltaic effect

Photoemissive effect

9
 Photoconductivity
A consequence of small band gap (Eg) in
semiconductors is that it is possible to generate
additional carriers by illuminating a sample of
semiconductor by a light of frequency greater than Eg/h.
This leads to an increased conductivity in the sample
and the phenomenon is known as: Intrinsic
photoconductivity

The effect is not pronounced at high temperature.

At low temperatures, illumination results in excitation

of localized carriers to conduction or valence band.
10
 Photoconductivity

Even when an incident photon does not have

sufficient energy to produce an electron-hole pair,
it can still produce an excitation at the impurity
centres by creating a free electron - bound hole pair
(for excitation at donor level) or a free hole - bound
electron (for acceptor level).

If Ei is the impurity ionization energy, the radiation

frequency for extrinsic photoconductivity should
be at least Ei/h

11
Photoconductivity

Quantum efficiency is the fraction of absorbed photons that

produce electron-hole pairs. \
The number of pairs produced per unit time is given by:

The process of illumination will lead to a continued increase in

the number of carriers as the amount of energy absorbed will
increase linearly with time. However, the excited pairs have a
finite life time. This results in recombination of the pairs. The
relevant life time is that of minority carriers as a pair is
required in the process. Recombination ensures that the number
of excess carrier does not increase indefinitely but saturates. 12
 Photoconductivity

Consider an n-type semiconductor. If the recombination

life time for the minority carriers is Tp
Then the rate of change of carrier concentration is given
by:

p =

This excess hole density leads to an additional conductivity

13
 Photovoltaic Effect

Photovoltaic effect can occur in a material which has

a space charge layer, e.g. in a p-n unction.

A photon of sufficient energy can be absorbed by the

detector material to excite an electron from the valence
band to the conduction band.

The excited electron may be observed through its

contribution to the current.

A photovoltaic detector can be operated without

application of a bias voltage.
14
Photovoltaic Effect

15
 Photoemissive Process
In a photoemissive process incident radiation
causes electron emission from photocathode which are
to be collected by an anode.

Photoemissive detectors have an advantage over

other detectors as they have faster speed, higher gain
and low noise.

However, their spectral range is somewhat limited as

the incident photon must have sufficient energy to eject
electrons from the photocathode.

Photoemissive detectors are, therefore, natural

choices in the ultraviolet range. 16
 Performance Parameters

Responsivity

Responsivity of a detector is given as the ratio of the

generated photocurrent ( I) to the amount of optical
power (Po ) incident on the detector:
R = I/ Po amperes/watt.
Quantum Efficiency
A detector is not capable of collecting all the photons
and convert them to electron-hole pairs. The number of
electrons produced per incident photon is defined as
the quantum efficiency , which is usually expressed
as a percentage
17
 Performance Parameters

The responsivity depends on the wavelength.

For ideal conductor: = 1 and R is linear with

18
 Spectral Response
The spectral response of a detector is given by the manner in
which the output signal of the detector varies with the change in
the wavelength of the incident radiation. As the quantum
efficiency depends on the wavelength, the response is not linear
as would be the case if = 1
The energy of the photon must be sufficient to excite an electron
across the energy barrier Eg. If Eg is in eV, the maximum
wavelength that the detector would respond to is

However, the response does not fall off abruptly to zero for
values of above the threshold. This is because, due to thermal
energy of the molecules, the absorption coefficent of the
material of the device is found to be given by:

E is the incident photon energy. For > max ,
and E<Eg. so that the absorption of radiance
becomes smaller. 19
Performance Parameters
Spectral Response

20
Noise Equivalent Power (NEP)
Source of noise in a detector is thermal fluctuation. Charged
particles are always in a state of motion. Even when no
radiation is incident on a device, a background current, whose
magnitude could be in nano-amperes or pico-amperes, is
generated. This is known as dark current . In order that a
detector may be able to differentiate between such random
noise and an incoming signal, the power of the signal must be
greater than the noise signal. In a detector design, one defines
signal to noise ratio (SNR) as: signal power/Noise power
Noise equivalent power (NEP) is an important figure of merit
for a detector. NEP is defined as the rms incident power which
gives rise to a current (or voltage) whose rms value is equal to
the rms value of the current (voltage) due to noise effects.
For a detector, the NEP is usually specified at particular
wavelength and temperature. The bandwidth for the incident
radiation for the measurement of NEP is generally taken as 1
Hz. 21
Detectivity and Dee Star (D* )

Both these terms are frequently used interchangibly,

D* is essentially the inverse of NEP normalized to unit
area (A) of the detector.

22
Photodiode

A photodiode is a p-n junction diode that can

absorb photons and generate either a
photovoltage or free carriers that can produce
photocurrent. They are used for detection of
optical signals and for conversion of optical
power to electrical power.
The figure shows a p+n junction diode. The p-
layer is very thin and is formed on the front
surface of the device by thermal diffusion or
ion implantation on an n-type silicon. The
active area is coated with an antireflection
coating of material (like silicon nitride) so that23
 Photodiode
Silicon is the most favoured material for a
photodiode. With a band gap of 1.1 eV, its peak
sensitivity is in I.R. between 800 to 950 nm. The
sensitivity drops at shorter wavelengths.
For <700 nm, the light gets absorbed in the p-layer
before reaching the junction. Thus in order to
increase sensitivity at shorter wavelengths, the
width of the p-layer should be smaller.
As p-type region has an excess of holes and n-type
region an excess of electrons, the holes diffuse
towards n-side and electrons to the p-side
resulting in a built in electric field gradient from n-
side to p-side. The built in electric field has a
strength such that there is no further movement of
charges through the depletion region. The
depletion region extends well into the lightly24
 Photodiode
If the n-side has a donor
density Nd per unit volume
and the p-side an acceptor
density of Na per unit
volume, equal amount of
mobile carriers are
annihilated from the two
sides leaving fixed charges
on the p+ and n sides. The
charge density distribution
is as shown. The condition
of charge neutrality
requires:
qNaxp = qNdxn

25
 Photodiode

The maximum field occurs at x = d + xp

The built in potential is given by:

26
 Photodiode

The photodiode can be represented by an equivalent circuit

shown. The current source is due to the drift current
produced by illumination.

The photodiode junction provides an equivalent capacitance

and a shunt resistance in the circuit. The series resistance is
due to the resistance of other circuit elements like contact
resistance, resistance of substrate etc., and may be
considered negligible. 27
 P-I-N Photodiode

The width of the depletion layer may be increased artificially,

by adding an intermediate intrinsic region.
As the intrinsic region has high resistance, a small reverse
bias is good enough to increase the width of the depletion
region so that it extends into the n-layer.
A further advantage of a p-i-n diode is that the charge
separation in the active region is larger which leads to smaller
28
junction capacitance.
Band diagram of P-I-N
Photodiode

29
 P-I-N Photodiode
Like p-n diode,the electric
field in the intrinsic region is
uniform as:

Considering VB is the built in

potential and applied reverse
bias voltage V, we get

The width of the depletion

region is:

30
 Avalanche Photodiode (APD)
A major disadvantage of a p-n or a p-i-n diode is that each photon
generates only one pair of electron and hole and there is no internal
gain. Amplifying the output current after the detector stage introduces
significant noise. One of the ways to deal with this problem is to
design a detector with an internal gain (the other is to amplify the
optical signal itself). An avalanche photodiode (APD) is a device with
internal gain which could be as high as 100. Si APDs have
sensitivities in the range 400 to 1100 nm while Ge-APDs have their
spectral sensitivities in 800 to 1550 nm. InGaAs and InP APDs
provide better sensitivity and spectral response.
Avalanche photodiode (APD) is a highly sensitive semiconductor
electronic device that can be thought of as photodetectors but provide
a built-in first stage of gain through avalanche multiplication. From a
functional standpoint, they can be regarded as the semiconductor
analog to photomultipliers. By applying a high reverse bias voltage
(typically 100-200 V in silicon), APDs show an internal current gain
effect (around 100) due to impact ionization (avalanche effect).
31
 Avalanche Photodiode (APD)
APDs are essentially p-n photodiodes operated under reverse
bias near the breakdown voltage. The configuration consists of
an n+ layer followed, in sequence, by
(i) a thin p layer,
(ii) an intrinsic layer (in reality, a lightly doped p layer) and
(iii) a heavily doped p layer.
Avalanche process occurs when carriers in the transition
region are accelerated by the electric field with high
energies that these carriers collide with other bound
carriers and free them to generate higher current.

32
 Structures for InGaAs APDs
Separate-absorption-and multiplication (SAM) APD

light

InP substrate
InP buffer layer
INGaAs Absorption layer

InP multiplication layer

Metal contact

InGaAs APD superlattice structure (The multiplication region is

composed of several layers of InAlGaAs quantum wells separated by
InAlAs barrier layers.
 Sources of Noise in Detectors
Dark Noise

Detection of optical power consists of streams of photons

which arrive randomly. The probability of arrival of number of
photons in a given time interval is usually given by a Poission
distribution. This causes a random fluctuation of the signal. As
the source of noise is statistical, there is no way to reduce or
eliminate it. Dark noise, or dark current shot noise is due to
thermal generation of e-h pair even in the absence of photo
illumination. For large values of reverse bias, an estimate of
the dark noise is given by:

where Is is the reverse bias saturation current in

dark and B is the bandwidth of operation. 34
 Sources of Noise in Detectors
Shot Noise

Shot noise is also associated with the signal

(photocurrent) itself because of statistical nature of
generation of electron hole pairs due to photon
absorption. The signal noise current is estimated by:

where I ph is the (signal) photocurrent.

35
 Sources of Noise in Detectors
Johnson Noise

Johnson noise arises due to thermal motion of electrons in any

resistive element regardless of applied voltage. The noise is
intrinsic to all resistors and is not due to poor design or
manufacture. The rms value of Johnson noise at a temperature
T is given by:

where R is the shunt resistance.

Combining the above the noise the signal to noise
ratio (Power SNR) is given by:

If the photocurrent is very high, the effect of dark

noise and thermal noise may be neglected and
the SNR is given by:
36
Solar Cell
Solar Cell Goal: $1/W by 2017

A module efficiency of 25 %
is wanted to enable the
installation cost reductions.
 Solar Cells (P-N)
Is a PN junction device without any applied voltage across
the junction.
It converts photon power into electrical power and delivers
to load.
hv

P E- Field
N
IF IL
I + V -

R 39
 Solar Cells
Without any bias, there is a space charge region and an
electric field exists.
Electron-hole pairs generated due to incident photon are
swept away by this electric field and generates photocurrent IL
is the reverse bias direction.
This photocurrent produces a voltage drop across the
resistive load.
This voltage forward biases the P-N junction and produces a
forward bias current IF.
The net P-N junction current in the reverse bias direction is
given by:
I = IL-IF= IL-Is [ exp(]

So the net solar current in the reverse bias direction.

40
 I-V Char. of Solar
If
Cells
load R = 0, then V = 0:
Then the Current is known as Short circuit Current Isc = IL

If load R =, then Current I = 0: The voltage V is then called

open circuit voltage Voc, which is given by:
Voc = Vtln( 1 +

ISC

V VOC 41
 Power of Solar
Cells
Power delivered to the load P= VI
P =VI = ILV - Is [ exp(] .V
For maximum power
0= IL-Is [ exp(] IsVm () exp(

(1 - exp( = 1+
where Vm is the voltage at ISC
maximum power delivered Im
and Vt =KT/e
I

V Vm VOC
Maximum power Rectengle of Solar cell
42
 Efficiency of Solar
Cell
The conversion efficiency of solar cell is given by:
= * 100 % = 100 %
For Silicon solar cell, Maximum Efficiency is about 28%.
But
Series resistance and reflection from the surface reduce to 15%
More Losses: Junction and contact voltage loss,
Recombination loss
Large optical lens can be used to increase light intensity
Fill Factor = (ranges between 0.7 and 0.8)
For a single semiconductor solar cell
If Photon energy E<Eg, then there is no effect of Photon E.
And if E >Eg, then extra energy dissipated as heat.
As the photon energy increases, more energy will be absorbed
in a very narrow region of the surface of the SC than the
deeper area of SC.
So there will be non-uniform carrier generation. 43
 Air Mass

Amount of air mass through which light

pass
Atmosphere can cut solar energy reaching
earth by 50% and more
Efficient cells using
Nanocones

Enable complete light absorption in thin Si

Source: J-H Lai, IEEE PVSC, June 2011
19.6% efficient cells on
silicon

Source: J-H Lai, IEEE PVSC, June 2011

 Heterojunction Solar
Cell
Let a P-N junction solar cell is formed with a wide bandgap
Egn for N type and narrow Egp for P type material.
Assume, photons are incident on the wide bandgap material.
Photons with energy E > Egn will be absorbed in wide bandgap
material and excess electron-hole pairs will be generated.
Photons with energy E < Egn will not be absorbed but will pass
through this wideband gap material. This wide bandgap
material acts like an optical window for those photons.
Photons with energy E > Egp will be absorbed in the narrow
bandgap material.
So there will be an excess average free carriers in the cell
contribute to photocurrent.
Another variation of heterojunction solar cell is formed by
forming a p-n homojuntion and then a wide bandgap material
is grown on top of the homejunction . 47
 Cadmium Telluride Solar Cells
(CdTe/CdS)

CdTe : Bandgap 1.5 eV; Absorption coefficient 10 times

that of Si . CdS : Bandgap 2.5 eV; Acts as window layer
High module production speed
Very inexpensive
20.4 % efficiency
 CdTe: Industrial Status

First Solar is the leader. It takes them 2.5 hours to make a 13.4 % module.

Average Manufacturing
Cost
2006: $1.40/watt
2007: $1.23/watt
2008: $1.08/watt
2009: $0.87/watt
2010: $0.77/watt
2011: $0.74/watt
2012: $0.64/watt
2013: $0.53/watt

www.firstsolar.com; greentech media 49

 Inverted Thin Film Cell

p-diamond (Bandgap 5.5 eV) as a window

layer
n-CdTe layer as an absorption layer
GaAs cell with 28.8 % Efficiency
 Photon recycling
GaAs cell with 28.8 % Efficiency

Si GaAs

Eli Yablonovitch et al., IEEE J. of Photovoltaics, 2 (2012) p. 303.

Why thin film GaAs is better

Remitted photons are weakly absorbed and can easily

travel more than a carrier diffusion length away from the
junction in a wafer-based device.
In a thin cell, a mirror keeps photons near the pn
junction.

Eli Yablonovitch et al., IEEE J. of Photovoltaics, 2 (2012) p. 303.

 Multijunctions: for Higher Efficiencies
Higher-efficiency MJ cells require new materials that divide the
solar spectrum equally to provide current match
Ge provides lattice match but the bandgap is too small

G a In P G a Gn aP I n P G a IGn
I
G a In P G a In P G a In P
11 . .88 e eV V 1 . 81 .e8 V1 e. 8V e 1 V. 8 1e V. 8 1 e
GG aaA As s G a A Gs a AGs a A
11 . .44 e eV V 1 . 12 5. 2 e5 1V e. 4V e 1V. 4 1e V. 41 .e
Ge GaInNAs
GG e e
0.7 eV G e 1 . 0 1e V. 0 e
00 . .77 e eV V 0 . 7 e V01.0 eV
0 . 7 e V 0 V. 7 0e .V7 0e
. 7 e
4 5 6 7 8 9
1
2 3 4
in New Solar
fu tu r e g e n e r a tio n
E n er g y (eV ) pConventional l er ei ns tgr ue cn t ue rr ea stf iu
r o d ui nc t i o n e x af mu t puJunction
p r o d u c t i o n e px ar omd pu lce t i so tnr u c et ux
MJ Cell
 Multi Junction multi band cell

Ideal
efficiency for infinite stack is 86.8%
GaInP/GaAs/Ge tandem cells (efficiency 40%)
 Multijunction Cells are Very Expensive
Efficiency =>50%

These complex
structures are grown
very slowly under high
Ga0.50In0.50P: vacuum.

Top Cell 37 % cells can be

purchased for
$50,000/m2
Ga0.99In0.01As:
Concentrating the light
Middle Cell is essential.

Ge substrate:
Bottom Cell R.R. King; Spectrolab Inc., AVS 54th
International Symposium, Seattle 2007
 References
1. Device Electronics for Integrated Circuits, by Muller and Kamins,
Wiley, New York, 1986.
2. GaAs Semiconductor Manufacturing,
http://www.mse.vt.edu/faculty/hendricks/mse4206/GaAsTEK/
default.htm
(Nov. 26, 2007)

3. Semiconductor Device Fundamentals, by Pierret, Addison-Wesley,

4. Physics of Semiconductor Devices, by S. M. Sze, Wiley, New York
5. Semiconductor Physics & Devices, 2nd ed., by Neamen,
6. Lecture notes of Prof. Gerhard Klimeck of Purdue University
7. Fundamentals of Semiconductor Theory and Device Physics, by
Shyh Wang, Prentice Hall

57