EE5508 Lecture 10-12

EE5508 Semiconductor Fundamentals
Department of Electrical and Computer Engineering

National University of Singapore
Prof. Wu Yihong
Room: E4-8-03
E-mail: elewuyh@nus.edu.sg
EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

1
Coverage of Lecture 7-9
Lecture 7
Absorption Final
𝐸𝐸𝑓𝑓
state
Lecture 8
ℏ𝜔𝜔
𝐸𝐸𝑖𝑖 Initial
state
Reflection
Transmission
Introduction Absorption
Bulk
Scattered Semiconductor
Incident (Raman, Brillouin)
Luminescence
EM wave
Emission Initial
𝐸𝐸𝑖𝑖
state
𝑘𝑘 ℏ𝜔𝜔
Lecture 9
𝐸𝐸𝑓𝑓 Final
Light-charge interaction state
Lecture 10-12
Lecture 7-9
Lecture 8 Lecture 12b: Photodetector
As
Ga
P
Emission Initial
state
Lecture 9
Light-charge interaction
Final
state
Lecture 10-11: Low-D structures
Lecture 12a: Light emitter

Outline of Lecture 10-12
 Semiconductor alloys  Optical properties of semiconductor

quantum structures
‒ Epitaxial growth
‒ 2D excitons
‒ Ternary and quaternary alloy
‒ Optical absorption in QW
‒ Vegard’s law
‒ Franz-Keldysh (FK) effect
 Semiconductor heterostructures
‒ Quantum confined Stark effect
 Semiconductor quantum structures
 Applications of semiconductors in
‒Single quantum well (SQW)
optoelectronics
‒ Multiple quantum well (MQW)
‒ Light emitting diode (LED)
‒ Superlattice
‒ Laser diode (LD)
‒ Quantum wire & quantum dot
‒ Photodiode (PD)

4
III-V semiconductors and related alloys
 Number of binary semiconductor is
limited. In practical applications, it is
often necessary to form ternary and
quaternary alloy semiconductors.
 By combining three or four different
elements in these alloys, you can
manipulate the bandgap, lattice
constant, and other semiconductor
properties. This is essential for
designing materials that can efficiently
perform in various electronic devices.
 Ternary and quaternary alloys enable engineers to optimize semiconductor materials for
specific applications. For instance, you might need a material with a particular bandgap to
efficiently capture sunlight in solar cells, or to emit light of particular color in LED or laser
diode.
 The availability of substrates significantly influences the choice of alloy composition for
semiconductor materials. Compatibility with substrates, matching crystal structures and
lattice constant, and cost considerations guide engineers in selecting alloys that ensure
smooth interfaces, low-defects, cost-effectiveness, and optimal performance.

5
Semiconductor thin film deposition techniques
 The most useful form of semiconductors is thin Multilayer

film with a thickness ranging from sub-nm to 𝜇𝜇m. Thin Film
They are typically fabricated on either a same or
foreign substrate/underlayer using the following Substrate Substrate
Substrate
deposition techniques:
Chemical Vapor Deposition (CVD): This method

involves the chemical reaction of gaseous
precursors to deposit thin films on a substrate.
Molecular Beam Epitaxy (MBE): MBE allows for the
precise control of film thickness by evaporating
materials in an ultra-high-vacuum environment.
Physical Vapor Deposition (PVD): PVD techniques
include methods like sputtering and evaporation, MBE
where material is physically ejected from the
target/source material and deposited onto a
substrate.
Atomic Layer Deposition (ALD): ALD is a precise
technique that deposits thin films layer by layer
through alternating gas-phase chemical reactions.

6
Lattice mismatch and strain
 Lattice mismatch occurs when two

materials with different crystal lattice 𝑎𝑎𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 > 𝑎𝑎𝑠𝑠𝑠𝑠𝑠𝑠 Compressive in-
parameters are brought together in a plane strain
heterostructure. When these materials are
stacked or grown on top of each other, the
lattice constants may not match perfectly.
This mismatch can lead to defects, Substrate
dislocations, and strains in the material,
depending on its thickness. 𝑎𝑎𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 < 𝑎𝑎𝑠𝑠𝑠𝑠𝑠𝑠 Tensile in-plane
 Under a certain critical thickness, elastic strain
strain arises to accommodate the lattice
mismatch.
 Strain resulting from lattice mismatch can
Substrate
have significant effects (either desirable or
adverse) on the properties of materials and
devices.
𝑎𝑎𝑠𝑠𝑢𝑢𝑢𝑢 − 𝑎𝑎𝑓𝑓𝑖𝑖𝑖𝑖𝑖𝑖
 Strategies such as buffer layers or lattice- strain 𝜀𝜀 = mismatch
𝑎𝑎𝑓𝑓𝑖𝑖𝑖𝑖𝑖𝑖
matching techniques are often employed to
alleviate the impact of lattice mismatch in
heterostructures.

7
Critical thickness
 Strain is allowed only up to a particular thickness (critical thickness: tc):
0.1𝑎𝑎𝑠𝑠 Just an estimation. Actual case is more

𝑡𝑡𝑐𝑐 = complicated [e.g., APL 3, 323 (1985)]
𝜀𝜀
 Defects will be generated when the film thickness exceeds 𝑡𝑡𝑐𝑐 .

 Strained layer has the same lattice constant as the substrate on the plane of the
substrate but different along the growth direction.
John D. Cressler,
CRC Press, 2006
Structural scheme of pseudo-morphic (a) and fully relaxed (b)

SiGe layer grown on Si substrate

8
(a) HRTEM image of the
interface structure of AlN on
sapphire (Al2O3) along
AlN⟨1−100⟩. The interface is
atomically abrupt and
shows a regular
arrangement of misfit
dislocations. (b) Bragg-
filtered image of (a)
showing the misfit
dislocation network.
Phys. Rew. Applied 5, 054004

(2016)
Growth modes for heteroepitaxial systems
 Frank-van der Merwe: Pseudo-morphic films are obtained if the strain in the film
does not exceed the critical thickness or if the films are metastable, and partly relaxed
films are obtained if the strain is relaxed by misfit dislocation formation (also known
as layer-by-layer or pseudo-morphic growth).
 Volmer-Weber: It often occurs in systems where the adhesion energy between the
film and substrate is relatively weak, or the presence of impurities, defects, or other
factors on the substrate that initiate the island nucleation process. As a result, islands
form instead of a continuous film.
 Stranski–Krastanov: Initially, a thin film begins to grow on the substrate in a layer-by-
layer fashion, similar to the Frank-van der Merwe growth mode. Once the critical
thickness is surpassed, the film starts to form three-dimensional islands on top of the
initial layer
John D. Cressler,
CRC Press, 2006
Frank-van der Merwe Volmer-Weber Stranski–Krastanov (S-K)

10

quantum structures
‒ 2D excitons
‒ Vegard’s law
‒ Single quantum well (SQW)
optoelectronics
‒ Superlattice
‒ Photodiode (PD)

11
Alloy semiconductor N
Al
 Alloy semiconductors provide the flexibility Binary III-V
P
to tailor material properties, such as Semiconductors
bandgap, lattice constant, and carrier Ga
AlN, AlP, AlAs, AlSb,
mobility, for specific applications. This As GaN, GaP, GaAs, GaSb,
tunability is essential for designing high- InN, InP, InAs and InSb
In
performance electronic and optoelectronic
Sb
devices.
 Semiconductor alloys must be single phase Ternary III-V
(or single crystal) with nearly perfect mixing Al As Semiconductors
to be useful. 𝑥𝑥 𝑥𝑥
1 − 𝑥𝑥 P 1 − 𝑥𝑥 Ga
 There are two common methods for
producing high-quality alloy semiconductors: Ga P
(a) growing bulk crystals or (b) forming thin
AlxGa1-xAs GaAsxP1-x
epitaxial films on a large single-crystal
substrate.
Ga As
 The term “epitaxy” comes from the Greek
𝑥𝑥 𝑦𝑦 Quaternary III-V
root epi, meaning "above", and taxis, 1 − 𝑥𝑥 1 −y Semiconductors
meaning "in ordered manner". It can be
In P
translated "to arrange upon“.
In1-xGaxP1-yAsy

12
Selected III-V semiconductor alloys
 These alloy Material Example Remark

semiconductors cover System
almost the entire Ternary
spectrum range for light-
AxB1-xC AlxGa1-xAs, InxGa1-xAs, Constituent compounds:
emitting diode (LED),
InxGa1-xN, AlxGa1-xN AC, BC
laser diode (LD),
photodetector (PD), and ABxC1-x GaAsxP1-x, InAsxSb1-x, Constituent compounds:
photovoltaic (PV) InAsxP1-x, AlAsxSb1-x AB, AC
applications due to their Quaternary
tunable bandgaps and
ability to operate AxB1-xCyD1-y AlxGa1-xAsyP1-y Constituent compounds:
efficiently across various InxGa1-xAsyP1-y AC, AD, BC, BD
wavelengths and AlxIn1-xAsySb1-y
frequencies. AxByC1-x-yD AlxGayIn1-x-yP Constituent compounds:
 Some of them are also AlxGayIn1-x-yN AD, BD, CD, ABD, BCD, ACD
important for high-speed
electronics due to their ABxCyD1-x-y InAsxPySb1-x-y Constituent compounds:
high electron mobility. GaAsxPyN1-x-y AB, AC, AD, ABC, ABD, ACD

13

quantum structures
‒ 2D excitons
‒ Vegard’s law
optoelectronics
‒ Superlattice
‒ Photodiode (PD)

14
Vegard’s law and virtual crystal approximation (VCA)
 For ternary compound semiconductor AxB1-xC , if the constituent compound

AB and AC have the same crystal structure, then the final alloy also has the
same crystalline structure.
 The lattice constant of AxB1-xC alloy can be obtained by taking the average of
the lattice constants of the constituent compounds (also known as endpoint
compounds) weighted by the composition which is known as Vegard’s Law
(linear interpolation method):
𝐴𝐴𝐵𝐵 𝐶𝐶
𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
𝑥𝑥 1−𝑥𝑥
= 𝑥𝑥𝑎𝑎𝐴𝐴𝐴𝐴 + (1 − 𝑥𝑥)𝑎𝑎𝐴𝐴𝐴𝐴 Al P
𝑥𝑥 𝑥𝑥
1 − 𝑥𝑥 As 1 − 𝑥𝑥 Ga
E.g.: 𝑎𝑎 𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝐴𝐴𝐴𝐴 = 𝑥𝑥𝑥𝑥 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 + 1 − 𝑥𝑥 𝑎𝑎 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺
Ga As
𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺1−𝑥𝑥 𝑃𝑃𝑥𝑥 = 𝑥𝑥𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺 + 1 − 𝑥𝑥 𝑎𝑎𝐺𝐺𝐺𝐺𝐴𝐴𝐴𝐴
𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝐴𝐴𝐴𝐴 GaAs1-xPx
Example
What is the lattice constant of GaAs0.6P0.4 alloy? Given the lattice constants 𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 = 5.653 Å and
𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺 = 5.450 Å.
Solution: 𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺0.6 𝑃𝑃0.4 = 0.6𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 + 0.4𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺 = 5.572 Å

15
Vegard’s law (cont’d)
 Apart from the lattice constant, other material

parameters such as bandgap and effective
mass of semiconductor alloys can also be
obtained from those of the “endpoint
compounds" by using linear interpolation
(also known as Virtual Crystal Approximation
(VCA)) which assumes that the material is a
nearly perfect random alloy.
 The bandgap variation as a function of the
AxB1-xC alloy composition can be expressed as:
𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
𝐸𝐸𝑔𝑔 = 𝑥𝑥𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 + (1 − 𝑥𝑥)𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵
Example:
𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝐴𝐴𝐴𝐴

𝐸𝐸𝑔𝑔 = 𝑥𝑥𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 + 1 − 𝑥𝑥 𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 (GaAs and AlAs are endpoint compounds of 𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝐴𝐴𝐴𝐴)
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺1−𝑥𝑥 𝑃𝑃𝑥𝑥
𝐸𝐸𝑔𝑔 = 𝑥𝑥𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺 + 1 − 𝑥𝑥 𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 (GaP and GaAs are endpoint compounds of 𝐺𝐺𝐺𝐺𝐴𝐴𝐴𝐴1−𝑥𝑥 𝑃𝑃𝑥𝑥 )

16
 In most alloys, however, there is a bowing effect arising

from the disorder in the crystal. In this case, the bandgap
of an AxB1-xC alloy shows a quadratic dependence on x :
𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎
𝐸𝐸𝑔𝑔 = 𝑥𝑥𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 + (1 − 𝑥𝑥)𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵 − 𝑏𝑏𝑏𝑏 1 − 𝑥𝑥
= 𝑏𝑏𝑥𝑥 2 + 𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 − 𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵 − 𝑏𝑏 𝑥𝑥 + 𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵
where b is known as bowing parameter or bowing factor.

Note that b can be a positive or negative value.
 Bowing and Miscibility are directly linked. Alloys with

large bowing should have large regions of phase
separation. Thus, one may expect poor miscibility in
alloys with large bowing.
 Miscibility is the property of substances to mix in all
proportions (i.e., to fully dissolve in each other at any
concentration), forming a homogeneous solution. Bandgap of 𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝐴𝐴𝐴𝐴
Substances that do not mix are said to be immiscible
such as water and oil.  Solid-line: with bowing
 A miscibility gap is a region in a phase diagram for a  Dashed line: without
mixture of components where the mixture exists as two bowing
or more phases.

17
Example: AlxGa1-xAs alloy
𝐸𝐸𝑔𝑔Γ−Γ = 1.424 + 1.247𝑥𝑥 for 𝑥𝑥 < 0.45

= 1.985 + 1.147(𝑥𝑥 − 0.45)2 (for 𝑥𝑥 > 0.45)
𝐸𝐸𝑔𝑔Γ−𝑋𝑋 = 1.900 + 0.125 𝑥𝑥 + 0.143𝑥𝑥 2
𝐸𝐸𝑔𝑔Γ−𝐿𝐿 = 1.709 + 0.642𝑥𝑥 2
E E
E
EΓ− L EΓ− Γ EΓ− X

k k k
L Γ X L Γ X L Γ X
x < 0.45 (direct bandgap) x = 0.45(𝐸𝐸𝑔𝑔Γ−Γ = 𝐸𝐸𝑔𝑔Γ−𝑋𝑋 ) x > 0.45 (indirect bandgap)

18
Example
What is the bandgap of GaAs0.6P0.4 alloy without bowing factor? Given 𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 = 1.424 eV
and 𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝑃𝑃 = 2.75 𝑒𝑒𝑒𝑒.
𝐺𝐺𝐺𝐺𝐺𝐺𝑠𝑠0.6 𝑃𝑃0.4
Solution: 𝐸𝐸𝑔𝑔 = 0.6𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 + (1 − 0.6)𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺 = 1.954 eV
Example
𝐺𝐺𝐺𝐺𝐺𝐺𝑠𝑠 𝑃𝑃
Given the bandgap variation of GaAs1-xPx as a function of 𝑥𝑥: 𝐸𝐸𝑔𝑔 1−𝑥𝑥 𝑥𝑥 = 1.424 +
1.150𝑥𝑥 + 0.176𝑥𝑥 2 , calculate
(a) the value of bowing factor
(b) the bandgap for GaAs0.6P0.4
(c) the lattice constant for GaAs0.6P0.4 , using the bowing factor obtained in (a)
Solution:
(a) 𝑏𝑏 = 0.176
𝐺𝐺𝐺𝐺𝐴𝐴𝐴𝐴0.6 𝑃𝑃0.4
(b) 𝐸𝐸𝑔𝑔 = 1.424 + 1.15 × 0.4 + 0.176 × 0.42 = 1.912 𝑒𝑒𝑒𝑒
(c) 𝑎𝑎 𝑥𝑥 = 𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 + 𝑎𝑎𝐺𝐺𝐺𝐺𝑃𝑃 − 𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 − 𝑏𝑏 𝑥𝑥 + 𝑏𝑏𝑥𝑥 2
= 5.6535 + 5.4506 − 5.6535 − 0.176 × 0.4 + 0.176 × 0.42 = 5.530Å

19
 The effective mass of an AxB1-xC alloy is given by:
1 𝑥𝑥 1 − 𝑥𝑥
∗ = ∗ + ∗ In1-xGaxAs
𝑚𝑚𝐴𝐴𝐵𝐵1−𝑥𝑥 𝐶𝐶𝑥𝑥
𝑚𝑚𝐴𝐴𝐴𝐴 𝑚𝑚𝐴𝐴𝐵𝐵
Example
What is the electron effective mass of EΓ- X
GaAs0.6P0.4 using VCA? Given the
following electron effective masses: ∆EΓL
Γ
∗
𝑚𝑚𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 ∗
= 0.067𝑚𝑚𝑜𝑜 and 𝑚𝑚𝐺𝐺𝐺𝐺𝑃𝑃 = EΓ- L
0.82𝑚𝑚𝑜𝑜 where 𝑚𝑚𝑜𝑜 is the free electron

ΓEΓ- Γ
mass (9.1 x 10-31 kg)
Solution:
1 1 1
∗ = 0.4 ∗ + 0.6 ∗
𝑚𝑚𝐺𝐺𝐺𝐺𝐺𝐺𝑠𝑠0.6 𝑃𝑃0.4
𝑚𝑚𝐺𝐺𝐺𝐺𝐺𝐺 𝑚𝑚𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺
1 1 9.443
= 0.4 + 0.6 =
0.82𝑚𝑚𝑜𝑜 0.067𝑚𝑚𝑜𝑜 𝑚𝑚𝑜𝑜
In0.53Ga0.47As lattice matches to InP
∗
→ 𝑚𝑚𝐺𝐺𝐺𝐺𝐺𝐺𝑠𝑠0.6 𝑃𝑃0.4
= 0.106𝑚𝑚𝑜𝑜

20
 Lattice constant for quaternary compounds:

Compound Lattice constant
AxB1-xCyD1-y 𝑎𝑎 = 𝑥𝑥𝑥𝑥𝑎𝑎 𝐴𝐴𝐴𝐴 + 𝑥𝑥(1 − 𝑦𝑦)𝑎𝑎 𝐴𝐴𝐴𝐴 + (1 − 𝑥𝑥)𝑦𝑦𝑎𝑎𝐵𝐵𝐵𝐵 + (1 − 𝑥𝑥)(1 − 𝑦𝑦)𝑎𝑎𝐵𝐵𝐵𝐵
ABxCyD1-x-y 𝑎𝑎 = 𝑥𝑥𝑎𝑎 𝐴𝐴𝐴𝐴 + 𝑦𝑦𝑎𝑎 𝐴𝐴𝐴𝐴 + (1 − 𝑥𝑥 − 𝑦𝑦)𝑎𝑎 𝐴𝐴𝐴𝐴
AxByC1-x-yD 𝑎𝑎 = 𝑥𝑥𝑎𝑎 𝐴𝐴𝐴𝐴 + 𝑦𝑦𝑎𝑎𝐵𝐵𝐵𝐵 + (1 − 𝑥𝑥 − 𝑦𝑦)𝑎𝑎𝐶𝐶𝐶𝐶
 Bandgap for quaternary compounds:
Compound Bandgap
𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 𝐸𝐸𝑔𝑔𝐴𝐴𝐶𝐶 1 − 𝑥𝑥 bXYZ : bowing

AxB1-xCyD1-y 𝐸𝐸𝑔𝑔 = 𝑦𝑦 𝑦𝑦(1 − 𝑦𝑦) (1 − 𝑦𝑦) 𝑏𝑏𝐵𝐵𝐵𝐵𝐵𝐵 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 𝑥𝑥(1 − 𝑥𝑥) parameter for XYZ
𝐸𝐸𝑔𝑔𝐵𝐵𝐷𝐷 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 𝑥𝑥 Keh Yung Cheng
ABxCyD1-x-y 𝐸𝐸𝑔𝑔 = 𝑥𝑥𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 + 𝑦𝑦𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 + 1 − 𝑥𝑥 − 𝑦𝑦 𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 Springer 2020
−𝑥𝑥𝑥𝑥𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 −𝑥𝑥 1 − 𝑥𝑥 − 𝑦𝑦 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 − 𝑦𝑦 1 − 𝑥𝑥 − 𝑦𝑦 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴
AxByC1-x-yD 𝐸𝐸𝑔𝑔 = 𝑥𝑥𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 + 𝑦𝑦𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵 + 1 − 𝑥𝑥 − 𝑦𝑦 𝐸𝐸𝑔𝑔𝐶𝐶𝐶𝐶
−𝑥𝑥𝑥𝑥𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 −𝑥𝑥 1 − 𝑥𝑥 − 𝑦𝑦 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 − 𝑦𝑦 1 − 𝑥𝑥 − 𝑦𝑦 𝑏𝑏𝐵𝐵𝐵𝐵𝐵𝐵
𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵 0 𝐸𝐸𝑔𝑔𝐴𝐴𝐶𝐶 1 − 𝑥𝑥
Ignore = 𝑥𝑥𝑥𝑥𝐸𝐸𝑔𝑔𝐴𝐴𝐶𝐶 + 𝑥𝑥 1 − 𝑦𝑦 𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 +
𝐸𝐸 = 𝑦𝑦 𝑦𝑦(1 − 𝑦𝑦) (1 − 𝑦𝑦) 0 0 0 𝑥𝑥(1 − 𝑥𝑥)
bowing: 𝑔𝑔 (1 − 𝑥𝑥)𝑦𝑦𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵 + (1 − 𝑥𝑥)(1 − 𝑦𝑦)𝐸𝐸𝑔𝑔𝐵𝐵𝐷𝐷
𝐸𝐸𝑔𝑔𝐵𝐵𝐷𝐷 0 𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴 𝑥𝑥

21
Example:
Consider a quaternary alloy material In1-xGaxAs1-yPy
(i) Using virtual crystal approximation, calculate the
bandgap of In1-xGaxAs1-yPy as a function of x and y
Materials Lattice Bandgap (eV)
composition (ignore bowing). constant (Å)
(ii) Determine the relationship between x and y for InP 5.8687 1.344
In1-xGaxAs1-yPy to be lattice matched to InP.
InAs 6.0580 0.354
(iii) Determine the values of x such that In1-xGaxAs is
lattice matched to InP. GaAs 5.6530 1.424
(iv) Hence determine the bandgap energy of GaP 5.4505 2.750
In1-xGaxAs in (iii)
Solution:
(i) Bandgap energy of In1-xGaxAs1-yPy
𝐸𝐸𝑔𝑔 = (1 − 𝑥𝑥)𝑦𝑦𝐸𝐸𝑔𝑔𝐼𝐼𝐼𝐼𝐼𝐼 + (1 − 𝑥𝑥)(1 − 𝑦𝑦)𝐸𝐸𝑔𝑔𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 + 𝑥𝑥(1 − 𝑦𝑦)𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 + 𝑥𝑥𝑥𝑥𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺
= 1.344(1 − 𝑥𝑥)𝑦𝑦 + 0.354(1 − 𝑥𝑥)(1 − 𝑦𝑦) + 1.424𝑥𝑥(1 − 𝑦𝑦) + 2.75𝑥𝑥𝑥𝑥 iv 𝐸𝐸𝑔𝑔 = 0.354 + 1.07 0.47
= 0.354 + 1.07𝑥𝑥 + 0.99𝑦𝑦 + 0.336𝑥𝑥𝑥𝑥
= 0.857 𝑒𝑒𝑒𝑒
(ii) Using Vegard’s law, the lattice constant of In1-xGaxAs1-yPy
𝑎𝑎 = 𝑎𝑎𝐼𝐼𝐼𝐼𝐼𝐼 (1 − 𝑥𝑥)𝑦𝑦 + 𝑎𝑎𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 (1 − 𝑥𝑥)(1 − 𝑦𝑦) + 𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑥𝑥(1 − 𝑦𝑦) + 𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺 𝑥𝑥𝑥𝑥
y = 0, → 𝑥𝑥 = 0.47
= 5.8687(1 − 𝑥𝑥)𝑦𝑦 + 6.058(1 − 𝑥𝑥)(1 − 𝑦𝑦) + 5.653𝑥𝑥(1 − 𝑦𝑦) + 5.4505𝑥𝑥𝑥𝑥
(iii) For In1-xGaxAs1-yPy to be lattice matched to InP,

0.1893 − 0.405𝑥𝑥
5.8687 = 5.8687(1 − 𝑥𝑥)𝑦𝑦 + 6.058(1 − 𝑥𝑥)(1 − 𝑦𝑦) + 5.653𝑥𝑥(1 − 𝑦𝑦) + 5.4505𝑥𝑥𝑥𝑥 ⇒ 𝑦𝑦 =
0.1893 + 0.013𝑥𝑥

22
Example: GaxIn1-xAsyP1-y
Lattice matched to GaAs
Lattice matched to InP
Keh Yung Cheng

Springer 2020
Band gap (in eV) map for GaxIn1-xAsyP1-y. The x-y coordinate of any point in the plane gives the composition.
The curved lines are constant direct bandgap energy values. The composition lattices matched to InP and
GaAs are shown as straight lines (in red) connected to InP and GaAs corners, respectively

23
Springer Handbook of Electronic and Photonic Materials, 2007 (S. Kasap and P. Capper, Editors)

quantum structures
‒ 2D excitons
‒ Vegard’s law
optoelectronics
‒ Superlattice
‒ Photodiode (PD)

25
Semiconductor heterostructure
 A semiconductor heterostructure refers to a material structure that involves the
combination of two or more different semiconductor materials with distinct
bandgaps or electronic properties.
 Semiconductor heterostructures play a crucial role in modern electronics and
optoelectronics including field-effect transistors, bipolar transistors, LEDs, lasers,
etc. They allow engineers to design devices with specific and optimized electronic
and optoelectronic properties, enabling better performance and functionality.
 Depending on the band alignment, there are three different types of
heterostructures, as illustrated below:
Straddled Staggered Broken gap

alignment alignment alignment

26
 Both theoretical models and experimental methods have been developed to obtain
band lineups between two different types of semiconductors, but the values
produced generally vary in a large range. As a general guideline, one may use the
following method to estimate the conduction and valence band offsets:
Identify material properties: Find the electron

affinity of each semiconductor material (𝝌𝝌).
Energy band diagram: Draw energy band
diagram for both semiconductors. This will help
visualize the valence and conduction bands and
Vacuum energy level
their respective energy levels.
Band alignment: Compare the energy band 𝝌𝝌𝑮𝑮𝑮𝑮𝑮𝑮𝑮𝑮 𝝌𝝌𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨
diagrams and align them to see how well they
match up. The alignment will determine the band
offset at the interface.
Calculate band offsets: Use the electron affinity
and bandgaps to calculate the band offsets
between the two materials. This will indicate
how the energy bands are aligned at the
interface.

27
Example:
Consider a CdS/InP heterostructure. The CdS and InP have direct bandgap energies
of 2.42 eV and 1.34 eV, respectively. Given the valence band-offset between CdS
and InP is 1.63 eV and the InP valence band is higher than that of CdS,
(i) find the conduction band-offset of CdS/InP heterostructure.
(ii) Is the CdS/InP heterostructure forms a type I or type II band-alignment?
(iii) Sketch a schematic diagram to show the band alignment of CdS and InP.
Solution:
(i) ΔEC = EgInP − EgCdS − ΔEV = 1.34 − ∆Ec= 0.55 eV

𝐸𝐸𝑔𝑔𝐼𝐼𝐼𝐼𝐼𝐼 = 1.34 𝑒𝑒𝑒𝑒
2.42 − 1.63 = 0.55 eV
(ii) It is Type II. 0.79 eV
(iii) See band alignment  𝐸𝐸𝑔𝑔𝐶𝐶𝐶𝐶𝐶𝐶 = 2.42 𝑒𝑒𝑒𝑒

InP
∆EV = 1.63 eV
CdS

28

quantum structures
‒ 2D excitons
‒ Vegard’s law
optoelectronics
‒ Superlattice
‒ Photodiode (PD)

29
Quantum well, wire, and box
(ℏ𝑘𝑘∥ )2 ℏ2 𝑛𝑛𝑥𝑥 𝜋𝜋 2
𝐸𝐸 = + E
E 2𝑚𝑚 2𝑚𝑚 𝐿𝐿
z
V =∞
3D (bulk) L
(free electron) (ℏ𝑘𝑘)2
𝐸𝐸 =
2𝑚𝑚
2D
k
y
𝑘𝑘 = (𝑘𝑘𝑥𝑥2 + 𝑘𝑘𝑦𝑦2 + 𝑘𝑘𝑧𝑧2 )1/2 Quantum well
x L k//
𝑘𝑘∥ = (𝑘𝑘𝑦𝑦2 + 𝑘𝑘𝑧𝑧2 )1/2
z E
z V =∞
V =∞
Quantum wire
L 0D y 1D
L
L L y
x
Quantum box L L 𝒌𝒌𝒛𝒛
x L
2 2
ℏ 𝜋𝜋 (ℏ𝑘𝑘𝑧𝑧 )2 ℏ2 𝑛𝑛𝑥𝑥 𝜋𝜋 2 ℏ2 𝑛𝑛𝑦𝑦 𝜋𝜋 2
𝐸𝐸 = (𝑛𝑛2 + 𝑛𝑛𝑦𝑦2 + 𝑛𝑛𝑧𝑧2 )
2𝑚𝑚𝐿𝐿2 𝑥𝑥 𝐸𝐸 =
2𝑚𝑚
+
2𝑚𝑚 𝐿𝐿
+
2𝑚𝑚 𝐿𝐿
30
Semiconductor low-dimensional structures
 The basic concepts of quantum well, quantum wire, and quantum box covered in
Lecture 4 still apply to semiconductor heterostructures except that
1) In semiconductors, in addition to electrons, we also have holes.
2) Electrons and holes are characterized by effective masses which are dependent on
the materials.
 Following are some typical semiconductor MQW

low-dimensional structures:
 Double heterostructure (DH)

 Single quantum well (SQW)
 Multiple quantum well (MQW)
 Superlattices (SLs)
SL
𝑊𝑊
DH SQW

31
Quantum confinement or quantum size effect
 Quantum confinement effect only becomes significant when the confinement
induced energy splitting is comparable to or larger than the kinetic energy (KE) of
the particle due to its thermal motion, which can be estimated from the Heisenberg
uncertainty principle.
 According to the uncertainty principle, if we confine a particle to a region of width
∆𝑥𝑥, the uncertainty in its momentum is given by:
ℏ ℏ 𝐾𝐾𝐵𝐵 𝑇𝑇 = 25.9 𝑚𝑚𝑚𝑚𝑚𝑚 (at RT)
Δ𝑝𝑝𝑥𝑥 Δ𝑥𝑥 ≈ → Δ𝑝𝑝𝑥𝑥 ≈ (3.3.6)
2 2Δ𝑥𝑥 𝑚𝑚𝑒𝑒 = 0.512 × 106 𝑒𝑒𝑒𝑒/𝑐𝑐 2
 This translates into a change of kinetic energy of (3.3.5)
𝑐𝑐 = 3 × 108 𝑚𝑚/𝑠𝑠
∆𝑝𝑝 2 2 ℏ = 1.054 × 10−34 𝐽𝐽 � 𝑆𝑆

𝐸𝐸≈ ≈ 2 = 6.582 × 10−16 𝑒𝑒𝑒𝑒 � 𝑠𝑠
2𝑚𝑚 8𝑚𝑚 ∆𝑥𝑥
∆𝑝𝑝 2 2 1 
 Let 𝐸𝐸≈ 2𝑚𝑚
≈
8𝑚𝑚 ∆𝑥𝑥 2 ≥ 𝑘𝑘𝐵𝐵 𝑇𝑇 → ∆𝑥𝑥 ≤
2 2 𝑚𝑚𝑘𝑘𝐵𝐵 𝑇𝑇
, meaning ∆𝑥𝑥 must be smaller than a
specific value in order for quantum confinement effects to be observed.
 The de Broglie wavelength of electron corresponding to thermal energy is given by
2
𝑝𝑝2 1 ℎ ℎ 6.28 × 6.582 × 10−16 𝑒𝑒𝑒𝑒 � 𝑠𝑠
𝐸𝐸 = = → λ𝑑𝑑𝑑𝑑 = = = 10.7 𝑛𝑛𝑛𝑛
2𝑚𝑚 2𝑚𝑚 λ𝑑𝑑𝑑𝑑 𝑚𝑚𝑘𝑘𝐵𝐵 𝑇𝑇 0.512 × 106 𝑒𝑒𝑒𝑒/𝑐𝑐 2 × 0.0259𝑒𝑒𝑒𝑒
𝜆𝜆𝑑𝑑𝑑𝑑
 Therefore, ∆𝑥𝑥 ≤ or ∆𝑥𝑥 ≤ 0.1λ𝑑𝑑𝑑𝑑 , i.e., ∆𝑥𝑥 must be comparable to or smaller than
4𝜋𝜋
0.1𝜆𝜆𝑑𝑑𝑑𝑑 for quantum size effect to be important.

32
 A semiconductor heterostructure is a type of layered material system composed of
two or more semiconductor materials with different bandgaps. These materials are
stacked together in a specific sequence, forming interfaces where the properties of
the materials abruptly change.
 Heterostructures are engineered to exploit the differences in bandgap energies and
other electronic properties between the constituent materials, allowing for precise
control of charge carriers and enabling the design of novel electronic and
optoelectronic devices with enhanced performance characteristics.
 A double heterostructure (DH) refers to a structure in which a small-bandgap region
(often called active region) is sandwiched from the two sides by a large-bandgap
barrier (or cladding) layer, particularly when the width of active region is large such
that quantum confinement effects are negligible. Note: carriers are still loosely
confined in the active region in classical sense. When it is applied to optoelectronics
devices, the large refractive index of the active region also facilitate light
confinement.
ℎ
𝑊𝑊 ≫ λ𝑑𝑑𝑑𝑑 = 10.7 𝑛𝑛𝑛𝑛
𝑚𝑚𝑘𝑘𝐵𝐵 𝑇𝑇
𝑊𝑊

33
Graph showing the relationship between

The Nobel Prize in Physics 2000 was awarded "for basic work active layer thickness (horizontal axis) of a
on information and communication technology" with one laser and threshold current (vertical axis)
half jointly to Zhores I. Alferov and Herbert Kroemer "for required for the laser to emit light. A smaller
developing semiconductor heterostructures used in high- threshold current is desirable for practical
speed- and opto-electronics" and the other half to Jack S. applications.
Kilby "for his part in the invention of the integrated circuit"

34
Semiconductor quantum well
 A semiconductor quantum well is a nanoscale structure consisting of a thin layer of
semiconductor material sandwiched between two layers of semiconductors with a
larger bandgap. Quantum confinement effects confine charge carriers (electrons and
holes) within the well, allowing for precise control of their energy levels and
enhancing their interaction, making quantum wells essential components in various
optoelectronic devices like lasers and photodetectors. When there is only one well,
it is called single quantum well (SQW).
Quantum well for electrons EC

3
x 2
E ∆EC
 2π 2
j=1 E1e =
z EgAlGaAs 2me* L2
EgGaAs
 2π 2
y J =1 E1h =
2mh* L2
∆EV 2
EV 3
Z
2 𝑗𝑗 2 𝜋𝜋2
𝐸𝐸𝑗𝑗 =
2𝑚𝑚∗ 𝐿𝐿2
j = 1, 2, 3,…
Quantum well for holes
Infinite potential well approximation

36
 The following figure shows the lowest quantised electron energy E1 as a
function of the QW layer thickness (also known as well width) L for a finite
potential well (with ∆Ec = 250 meV) and infinite potential well.
250
finite potential well
Infinite potential well

E1 (meV)
 For large well width ( L ~ 30 nm), the two curves almost coincide.
 For L < 30 nm, the E1 in the finite potential well is always less than that
in the infinite potential well approximation.
 In case of extremely small L, E1 ≈ ∆Ec = 250 meV (thus making the
electron just slightly confined in the QW).

37
 In bulk, the valence band (VB) at k = 0 is 4-fold degenerate.
 The heavy hole (hh) and light hole (lh) masses are degenerate in bulk GaAs
at k = 0. However, this degeneracy is lifted due to quantum confinement.
The quantum potential acts a perturbation that lifts the degeneracy of the
VB and results in the hh and lh masses.
(a) Dispersion relations for electron and hole subbands in a quantum well confining two electron states,
one light-hole and two heavy hole states. (b) Optical excitation transitions into the ground state of the
conduction band. (c) Light- and heavy hole dispersions showing the crossing at enlarged scale
K. W. Böer and U. W. Pohl, Springer, 2020
38
Example:
The Γ, X and L conduction valleys (with respect to the top of the valence band taken as
zero-reference energy) of AlxGa1-xN alloy can be expressed by the following equations:
Parameters GaN AlN
Γ 2
𝐸𝐸 = 3.38 + 2.50𝑥𝑥 + 0.05𝑥𝑥 Valence band 0.74mo 1.02mo
𝐸𝐸 𝑋𝑋 = 4.57 − 0.08𝑥𝑥 + 0.61𝑥𝑥 2 Conduction band Γ valley 0.13mo 0.21mo
Conduction band L valley 0.58mo 0.53mo
𝐸𝐸 𝐿𝐿 = 5.64 + 2.99𝑥𝑥 + 0.80𝑥𝑥 2 Conduction band X valley 0.30mo 0.31mo
a) Plot EΓ- Γ, E Γ -X and E Γ -L as a function of x in one graph.

b) From the graph, deduce if GaN and AlN is a direct or indirect gap semiconductor?
c) What is the x value for cross-over from direct to indirect gap ?
d) Is Al0.7Ga0.3N direct or indirect gap semiconductor? What is the band gap?
e) Using the linear interpolation method to calculate the effective masses of the valence
band and the lowest conduction valley of Al0.7Ga0.3N. Hence, calculates its reduced
effective mass.
f) AlxGa1-xN/AlN has a type I heterostructure. The conduction band-offset to valence
band-offset ratio of AlxGa1-xN/AlN heterojunction is 7/3. Design an AlN/AlxGa1-xN/AlN
quantum well structure with conduction band offset of 1.00 eV. Calculate the x value
such that AlxGa1-xN has a direct bandgap. Hence, calculate the AlxGa1-xN bandgap for
this x value obtained. You may use infinite potential well approximation for your
calculation.
39
Solution: Parameters GaN AlN
Valence band 0.74mo 1.02mo

𝐸𝐸 Γ = 3.38 + 2.50𝑥𝑥 + 0.05𝑥𝑥 2
Conduction band Γ valley 0.13mo 0.21mo
𝑋𝑋 2
𝐸𝐸 = 4.57 − 0.08𝑥𝑥 + 0.61𝑥𝑥 Conduction band L valley 0.58mo 0.53mo
Conduction band X valley 0.30mo 0.31mo
𝐸𝐸 𝐿𝐿 = 5.64 + 2.99𝑥𝑥 + 0.80𝑥𝑥 2
a) Plot EΓ- Γ, E Γ -X and E Γ -L as a function of x in one graph.
EΓ- L
EΓ- X 5.1 eV
3.38 eV EΓ- Γ
GaN AlN
Solution:
b) From the graph, deduce if GaN and AlN is a direct or indirect gap semiconductor?
From the graph, GaN is a direct gap and AlN is an indirect gap semiconductor.
c) What is the x value for cross-over from direct to indirect gap ?

The two curves intercept at 𝑥𝑥 = 0.52.
d) Is Al0.7Ga0.3N direct or indirect gap semiconductor? What is the band gap?

Al0.7Ga0.3N, x=0.7, X-valley, is an indirect gap with Eg = 4.813 eV
e) Using the linear interpolation method to calculate the effective masses of the valence
band and the lowest conduction valley of Al0.7Ga0.3N. Hence, calculates its reduced
effective mass.
1 𝑥𝑥 1 − 𝑥𝑥 0.7 0.3
= 𝐴𝐴𝐴𝐴𝐴𝐴 + 𝐺𝐺𝐺𝐺𝐺𝐺 = + ⇒ 𝑚𝑚𝑒𝑒∗ = 0.307𝑚𝑚𝑜𝑜
𝑚𝑚𝑒𝑒𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑚𝑚𝑒𝑒 𝑚𝑚𝑒𝑒 0.31 0.3
1 𝑥𝑥 1 − 𝑥𝑥 0.7 0.3
= 𝐴𝐴𝐴𝐴𝐴𝐴 + 𝐺𝐺𝐺𝐺𝐺𝐺 = + ⇒ 𝑚𝑚ℎ∗ = 0.916𝑚𝑚𝑜𝑜
𝑚𝑚ℎ𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑚𝑚ℎ 𝑚𝑚ℎ 1.02 0.74
1 1 1 1 1
∗ = ∗ + ∗ = + ⇒ 𝑚𝑚𝑟𝑟∗ = 0.23𝑚𝑚𝑜𝑜
𝑚𝑚𝑟𝑟 𝑚𝑚𝑒𝑒 𝑚𝑚ℎ 0.307𝑚𝑚𝑜𝑜 0.916𝑚𝑚𝑜𝑜

41
Solution:
f) AlxGa1-xN/AlN has a type I heterostructure. The conduction band-offset to valence
band-offset ratio of AlxGa1-xN/AlN heterojunction is 7/3. Design an AlN/AlxGa1-
xN/AlN quantum well structure with conduction band offset of 1.00 eV. Calculate the
x value such that AlxGa1-xN has a direct bandgap. Hence, calculate the AlxGa1-xN
bandgap for this 𝑥𝑥 value obtained
Δ𝐸𝐸𝑐𝑐 7 3
Δ𝐸𝐸𝑐𝑐 = 1.00 = ∆𝐸𝐸𝑉𝑉 = ∆𝐸𝐸𝐶𝐶 = 0.43 𝑒𝑒𝑒𝑒
Δ𝐸𝐸𝑣𝑣 3
7
∆𝐸𝐸g = ∆𝐸𝐸𝐶𝐶 + ∆𝐸𝐸𝑉𝑉 = 1.43 𝑒𝑒𝑒𝑒
Ec
𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝑁𝑁 ∆Ec= 1 eV
∆𝐸𝐸g = 𝐸𝐸g𝐴𝐴𝐴𝐴𝐴𝐴 − 𝐸𝐸g
1.43 = 5.1 − 3.38 + 2.50𝑥𝑥 + 0.05𝑥𝑥 2 𝐸𝐸𝑔𝑔𝐴𝐴𝐴𝐴𝐴𝐴 = 5.1 𝑒𝑒𝑒𝑒 𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝑁𝑁
𝐸𝐸𝑔𝑔
𝑥𝑥 2 + 50𝑥𝑥 − 5.8 = 0
∆EV= 0.43 eV
Solving gives x = 0.12 or -50.1 (reject)
Ev42
𝐸𝐸 Γ−Γ 0.12 = 3.38 + 2.5 0.12 + 0.05 0.12 2
= 3.681 eV

42
Multiple quantum well (MQW)
 Multiple quantum wells (MQWs) are
semiconductor structures composed of alternating
thin layers of two different semiconductor
materials. In a typical MQW structure, thin layers
of one material (the well) are sandwiched between
thicker layers of another material (the barrier). The
thickness of the well layers is on the order of a few
nanometers, while the barrier layers are typically
tens to hundreds of nanometers thick.
Increase interaction volume
 Compared to the SQW, the presence of multiple
quantum wells increases the volume of the active
region, which is important for optoelectronic
devices.
 The MQWs are used in various optoelectronic
devices, such as lasers, light-emitting diodes
(LEDs), photodetectors, and modulators. By
engineering the thickness and composition of the
layers, one can tailor the electronic band structure
and optical properties of the material, allowing for
the creation of devices with specific performance
characteristics, such as wavelength-tunable lasers Formation of different energy levels
or high-efficiency LEDs.
43
Superlattices (SLs)
 A superlattice is a periodic structure composed of alternating
layers of two or more different semiconductor materials. Each
layer typically has a thickness in a range of sub-nm to a few
nanometers.
 The key characteristic of a superlattice is its periodicity, which
gives rise to unique electronic and optical properties. Unlike the
MQW, the SL typically has more repetition layers, and the barrier
layer is thin enough to allow overlapping of electron wavefunction
in neighboring wells, leading to the formation of mini-bands
throughout the structure instead of discrete energy levels.
 The Sls have applications in various devices, such as quantum
cascade lasers, infrared detectors, thermoelectric generators, etc.
L. Esaki and R. Tsu, IBM J. Res. Develop. 14, 61 (1970).

44
Superlattices (SLs)
 The right figure shows the

computed widths of mini-
bands and intermittent gaps
as a function of the period
length for a symmetrical
well/barrier structure with a
barrier height of 0.4 eV.
• For 𝑙𝑙1 = 𝑙𝑙2 = 40 Å, the lowest
band is rather narrow and
lies at 100 meV above the
well bottom. The second
band extends from 320 to
380 meV. Higher bands (E3,
E4 . . .) overlap above the
top of the barrier. Increasing
the thickness or the height
of the barrier layer reduces
the tunneling through the
barriers.

45
Superlattices (SLs)
The Nobel Prize in Physics 1973 was divided, one half

jointly to Leo Esaki and Ivar Giaever "for their
experimental discoveries regarding tunneling
phenomena in semiconductors and superconductors,
respectively" and the other half to Brian David
Josephson "for his theoretical predictions of the
properties of a supercurrent through a tunnel barrier, in
particular those phenomena which are generally known
as the Josephson effects"

46
Semiconductor nanowires
ZnO nanowire
 A semiconductor nanowire is a one-
dimensional (1D) nanostructure with a
diameter on the order of sub-nm to
several tens of nanometers and lengths
extending up to several micrometers or
more.
 Semiconductor nanowires can be grown
through various techniques, including
vapor-liquid-solid (VLS) growth, vapor-
solid-solid (VSS) growth, MBE/MOCVD,
etc.
Science 2001 June 8; 292: 1897-1899
 In addition to quantum confinement
effect, the nanowires typically have a
high surface-to-volume ratio, which
makes them attractive for applications
such as sensors, where interactions with Strong confinement
the surrounding environment are leads to large
crucial. optical gain,
 Nanowires can serve as building blocks desirable for laser
for nanoelectronic circuits and devices applications
due to their small size and potential for
integration into densely packed arrays.

47
Semiconductor quantum dots
 A semiconductor quantum dot is a

InAs QDs (top-down)
zero-dimensional (0D) nanostructure
with a diameter in the range of sub-nm
to several tens of nanometers, and it is CdSe QD (bottom-up)
surrounded either by vacuum/air or a
semiconductor with a larger bandgap.
 The key characteristic of quantum dots
is their ability to confine electrons and
holes in all three dimensions, leading
to quantized energy levels, similar to
those found in atoms.
 Semiconductor quantum dots can be
fabricated using various methods,
including colloidal synthesis (bottom-
up), self-assembly (bottom-up), and
lithographic techniques (top-down).
 The unique optical and electronic
properties of the quantum dots make
them attractive for various
applications such as LED/laser, single
photon source, solar cell, display, PL spectra of CdSe QD with different
cosmetics, biomedical imaging, etc. sizes (Annu. Rev. Mater. Res. 2000 ,
Vol. 30: 545-610)

48
The Nobel Prize in Chemistry 2023

was awarded to Moungi G.
Bawendi, Louis E. Brus and Aleksey
Yekimov "for the discovery and
synthesis of quantum dots"
Measured and calculated energy of the lowest

state of electron–hole pairs in CdSe nanocrystal
quantum-dots as a function of the radius

49
https://doi.org/10.1017/9781316535868.013 Published online by Cambridge University Press

50

quantum structures
‒ 2D excitons
‒ Vegard’s law
optoelectronics
‒ Superlattice
‒ Photodiode (PD)

51
3/2
2 (ℏ𝜔𝜔−𝐸𝐸𝑔𝑔 )1/2 1 2 𝑚𝑚𝑟𝑟∗ 1/2 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔
𝛼𝛼 ∝ 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 𝐼𝐼 ℏ𝜔𝜔 ∝
𝜏𝜏𝑟𝑟 𝜋𝜋 2 ℏ2
ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 exp −
𝑘𝑘𝐵𝐵 𝑇𝑇
ℏ𝜔𝜔
Room temperature
photoluminescence of
4 μm thick GaAs crystals
grown on the deeply
patterned substrate (red)
and on the bulk GaAs
(black).
Optical absorption in quantum wells
 The procedures of calculating optical absorption spectrum in bulk

semiconductors are applicable to quantum-well materials with two
important modifications:
1) The step-like 2D DOS profile should be included in the absorption

coefficient calculation;
2) In bulk semiconductors, all vertical transitions between valence
and conduction bands are allowed as long as momentum and
energy are conserved. The optical transition in quantum wells has
to follow the quantum mechanical ‘selection rule’ such that the
quantum numbers of the initial and final states have to be the
same, i.e., 𝑛𝑛𝑓𝑓 − 𝑛𝑛𝑖𝑖 = ∆𝑛𝑛 = 0. For example, 𝑛𝑛 = 1 energy state in
the valence band is only alloyed to jump to 𝑛𝑛 = 1 energy state in
the conduction; similar rule applied to recombination.
 The absorption coefficient of a multiple quantum-well system has the

form of This is because the transition matrix
𝑚𝑚𝑟𝑟 element is proportional to |𝑀𝑀∥ ||𝑀𝑀𝑧𝑧 |,
𝛼𝛼 2𝐷𝐷 ∝ 𝑁𝑁 � 𝛩𝛩(ℏ𝜔𝜔 − 𝐸𝐸𝑚𝑚𝑚𝑚 ) 𝑓𝑓𝑣𝑣 𝐸𝐸 − 𝑓𝑓𝑐𝑐 (𝐸𝐸) where |𝑀𝑀∥ | is the usual dipole
𝜋𝜋ℏ2 𝑊𝑊 𝑤𝑤
𝑚𝑚,𝑛𝑛 transition matrix element for 𝑘𝑘∥ ,
whereas 𝑀𝑀𝑧𝑧 =< 𝜑𝜑𝑒𝑒 (𝑧𝑧)|𝜑𝜑ℎ (𝑧𝑧) > is
where 𝑊𝑊 is the width of the quantum well, 𝑁𝑁𝑤𝑤 is the total number of the overlap integral between
quantum wells, and 𝛩𝛩 is the Heaviside step function. electron and hole wavefunctions for
the quantized states.

53
Exciton in 2D systems
 Excitons are bound states of an electron and a hole in a semiconductor material. The behavior
of excitons can vary depending on the dimensionality of the semiconductor system.
 In 2D systems, such as quantum wells or monolayer materials like transition metal
dichalcogenides (TMDs), the motion of charge carriers is confined in two dimensions, leading
to a stronger interaction between the electron and hole, resulting in enhanced exciton
binding energies compared to bulk 3D materials.
 The Schrödinger equation for exciton in two dimensions is:
−ℏ2 2 ℏ2 2
∇ − ∇ + 𝑉𝑉𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 + 𝑉𝑉𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝜓𝜓 𝑟𝑟 = 𝐸𝐸𝐸𝐸 𝑟𝑟
2𝑚𝑚𝑒𝑒∗ 𝑒𝑒 2𝑚𝑚ℎ∗ ℎ
3D exciton 2D exciton
where the first two terms are the kinetic energy of the electron and the hole, and Vconfinement and
Vcoulomb are the confinement potential and electron-hole Coulomb potential, respectively.
 Solving the Schrödinger equation (out of the scope of this module) gives
2 𝑗𝑗 2𝜋𝜋2 3𝐷𝐷
𝐸𝐸𝐵𝐵
𝐸𝐸 = 𝐸𝐸𝑔𝑔 +
2𝑚𝑚𝑟𝑟∗ 𝐿𝐿2
− 1 2
, System Binging energy Radius
𝑗𝑗−
2
Hydrogen 1 1 𝑒𝑒 2 1 ℎ2 𝜀𝜀0
where j = 1,2,3,… 𝐸𝐸𝑛𝑛 = − = −13.6 𝑒𝑒𝑒𝑒 𝑎𝑎0 = = 0.0529 𝑛𝑛𝑛𝑛
2 4𝜋𝜋𝜀𝜀0 𝑛𝑛2 𝑎𝑎0 𝑛𝑛2 𝜋𝜋𝑚𝑚0 𝑒𝑒 2
3D exciton 𝑚𝑚𝑟𝑟∗ /𝑚𝑚𝑜𝑜 1 𝜀𝜀𝑟𝑟 𝑎𝑎0
𝑚𝑚𝑟𝑟∗ 13.6 𝐸𝐸𝐵𝐵3𝐷𝐷 = − × 13.6 𝑒𝑒𝑒𝑒 𝑎𝑎𝐵𝐵3𝐷𝐷 = ∗
𝐸𝐸𝐵𝐵3𝐷𝐷 = 𝑒𝑒𝑒𝑒 𝜀𝜀𝑟𝑟2 𝑛𝑛2 𝑚𝑚𝑟𝑟 /𝑚𝑚𝑜𝑜
𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟2
2D exciton ε3𝐷𝐷
𝐵𝐵 𝑎𝑎𝐵𝐵2𝐷𝐷 = 𝑎𝑎𝐵𝐵3𝐷𝐷 /2
𝐸𝐸𝐵𝐵2𝐷𝐷 = 2 = 4𝐸𝐸𝐵𝐵3𝐷𝐷
1
𝑗𝑗 −
2

54
 Quantum confinement effect is important when the well width W < 𝑎𝑎𝐵𝐵3𝐷𝐷 . In the
limit that W << 𝑎𝑎𝐵𝐵3𝐷𝐷 , the exciton has the properties of a 2-D hydrogen atom.
 The 2D exciton binding energy is enhanced (four times larger than the bulk exciton)
due to smaller distance between electrons and holes in the confinement direction.
 However, when W is too small, electron and hole wavefunctions will “leak” into the
barrier region, leading to weaker confinement, and thus smaller exciton binding
energy.
Exciton binding energy

𝐸𝐸𝐵𝐵2𝐷𝐷 as a function of well The corresponding
width for 𝐴𝐴𝐴𝐴0.4 𝐺𝐺𝐺𝐺0.6 𝐴𝐴𝐴𝐴/ exciton Bohr radius 𝑎𝑎𝐵𝐵2𝐷𝐷 Well width/𝑎𝑎𝐵𝐵3𝐷𝐷
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 QW as function of well width
Exciton binding energy in a quantum well B. Gerlach, J. Wüsthoff, M. O. Dzero, and M. A. Smondyrev
Phys. Rev. B 58, 10568 – Published 15 October 1998

55
 In naturally formed 2D systems, such as

TMDs, due to the strong quantum
confinement the Coulomb interaction
between the electrons and holes is only
weakly screened. This results in excitons
with notably high binding energies, nearly
two orders of magnitude greater than
those observed in semiconductor quantum
wells.
 The large binding energy facilitates
observation of excitonic absorption at Applied Materials Today 8 (2017) 68–103
room temperature.
 In addition to conventional excitons (or
bright excitons), the TMD also hosts spin-
forbidden and momentum-forbidden dark
excitons, which are optically non-radiative
due to selection roles. Those excitons
typically have a lower energy, and
therefore, when temperature rises, more
excitons occupy the energetically higher
bright states within the light cone resulting
in the experimentally observed increased
PL (detailed mechanisms are out of the
scope of this module). npj 2D Materials and Applications 2, (2018) 29

56
Moiré exciton in 2D systems
Moiré pattern formed by a 2D

semiconductor heterobilayer
Nano Letters 20, 8534 (2020)

57
SCIENCE
29 Jun 2023
Vol 380, Issue 6652
pp. 1367-1372
Physics Breakthrough:
Scientists Discover Rydberg
Moiré Excitons

58
Hydrogen atom Positively charged proton Wannier-Mott exciton (3D)
𝑚𝑚𝑒𝑒∗ and 𝑚𝑚ℎ∗ on 𝑚𝑚ℎ∗
Centre
- the same order +
of mass
+
Negatively
charged 𝑚𝑚𝑒𝑒∗
𝑚𝑚𝑒𝑒∗ 𝑟𝑟⃗𝑒𝑒 + 𝑚𝑚ℎ∗ 𝑟𝑟⃗ℎ 𝑅𝑅 𝑟𝑟ℎ
electron Proton mass: ~1837𝑚𝑚𝑒𝑒 - 𝑅𝑅 =
𝑚𝑚𝑒𝑒∗ + 𝑚𝑚ℎ∗
1 1 𝑒𝑒 2 1 ℎ2 𝜀𝜀0 𝑟𝑟𝑒𝑒
𝐸𝐸𝑛𝑛 = − = −13.6 𝑒𝑒𝑒𝑒 𝑎𝑎0 = = 0.0529 𝑛𝑛𝑛𝑛
𝜋𝜋𝑚𝑚0 𝑒𝑒 2
2 4𝜋𝜋𝜀𝜀0 𝑛𝑛2 𝑎𝑎0 𝑛𝑛2
1 1 1 0 (origin)
= + 𝑀𝑀 = 𝑚𝑚𝑒𝑒∗ + 𝑚𝑚ℎ∗
𝑚𝑚𝑟𝑟∗ 𝑚𝑚𝑒𝑒∗ 𝑚𝑚ℎ∗
Hydrogenic atom ℏ2 𝐾𝐾𝑐𝑐2
𝐸𝐸𝑅𝑅 = 𝐾𝐾𝑐𝑐 = 𝑘𝑘𝑒𝑒 + 𝑘𝑘ℎ
2𝑀𝑀
𝜀𝜀𝑟𝑟 =11.7 for Si 𝑚𝑚𝑟𝑟∗ 1 𝜀𝜀𝑟𝑟 𝑎𝑎0

𝐸𝐸𝑟𝑟 = − 2 𝑛𝑛2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐵𝐵 = 𝑚𝑚∗
𝜀𝜀𝑟𝑟 𝑟𝑟
𝑚𝑚𝑒𝑒∗
- - +
P 2D exciton
𝜀𝜀𝑟𝑟 𝜀𝜀0
2 𝑗𝑗 2𝜋𝜋2 3𝐷𝐷
𝐸𝐸𝐵𝐵
𝐸𝐸 = 𝐸𝐸𝑔𝑔 + − 1 2
,
𝑚𝑚𝑒𝑒∗ 1 𝜀𝜀𝑟𝑟 2𝑚𝑚𝑟𝑟∗ 𝐿𝐿2 𝑗𝑗−
𝐸𝐸𝑑𝑑 = 2 2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐷𝐷 = ∗ 𝑎𝑎0 (for donor) 2
𝜀𝜀𝑟𝑟 𝑛𝑛 𝑚𝑚𝑒𝑒 where j = 1,2,3,…
𝑚𝑚ℎ∗ 1 𝜀𝜀𝑟𝑟
𝐸𝐸𝑎𝑎 = 2 2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐴𝐴 = ∗ 𝑎𝑎0 (for acceptor) 𝑚𝑚𝑟𝑟∗ 13.6
𝜀𝜀𝑟𝑟 𝑛𝑛 𝑚𝑚ℎ 𝐸𝐸𝐵𝐵3𝐷𝐷 = 𝑒𝑒𝑒𝑒 𝑎𝑎𝐵𝐵2𝐷𝐷 = 𝑎𝑎𝐵𝐵3𝐷𝐷 /2
𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟2
Note: here 𝑚𝑚𝑒𝑒∗ and 𝑚𝑚ℎ∗ are normalized to 𝑚𝑚0
Example:
(i) 3D exciton binding energy (ground state)
Solution: 𝑚𝑚𝑟𝑟∗ −13.6 eV 0.068𝑚𝑚𝑜𝑜 −13.6 eV

𝜀𝜀𝐵𝐵3𝐷𝐷 = =
𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟2 𝑚𝑚𝑜𝑜 122
By linear interpolation: = 6.4 meV
𝐺𝐺𝐺𝐺𝑁𝑁0.2 𝐴𝐴𝐴𝐴0.8
𝜀𝜀𝑟𝑟 = 0.2 × 8 + 0.8 × 13 = 12 2D exciton binding energy (ground state)
1 0.8 0.2 𝜀𝜀𝐵𝐵2𝐷𝐷 = 4𝜀𝜀𝐵𝐵3𝐷𝐷 = 4 × 6.4 meV = 25.6 meV
= + ⇒ 𝑚𝑚𝑒𝑒∗ = 0.07𝑚𝑚𝑜𝑜
𝑚𝑚𝑒𝑒∗ 0.067𝑚𝑚𝑜𝑜 0.19𝑚𝑚𝑜𝑜
1 0.8 0.2
⇒ 𝑚𝑚ℎ∗ = 0.6𝑚𝑚𝑜𝑜
(ii) 3D exciton binding energy (ground state)
∗ = +
𝑚𝑚ℎ 0.53𝑚𝑚𝑜𝑜 1.2𝑚𝑚𝑜𝑜
1 1 1 1 1
⇒ 𝑚𝑚𝑟𝑟∗ = 0.068𝑚𝑚𝑜𝑜 25.6
∗ = ∗ + ∗ = + 𝑘𝑘𝐵𝐵 𝑇𝑇 = 25.6 meV ⇒ 𝑇𝑇 = × 300 = 296.5 𝐾𝐾
𝑚𝑚𝑟𝑟 𝑚𝑚𝑒𝑒 𝑚𝑚ℎ 0.07𝑚𝑚𝑜𝑜 0.6𝑚𝑚𝑜𝑜 25.9

60
Optical absorption in quantum wells
 The ideal absorption curve of QW contains sharp step edges due to its step-like DOS.
However, the intra-band carrier relaxation due to scattering within the quantum well can
cause the broadening of the energy states and, thus, spectral broadening.
Keh Yung Cheng

Springer 2020
Absorption spectrum of a 77-period GaAs /

Absorption spectra of an InGaAs/InAlAs
Al0:28Ga0:72As MQW structure with 10 nm QWs at
quantum well measured at various
room temperature. The absorption spectrum of
temperatures.
bulk GaAs is included for comparison.
Sharp peaks are due to exciton absorption Springer Handbook of Electronic and Photonic
Materials, 2007 (S. Kasap and P. Capper, Editors)

61
Optical absorption in QWs with different well width
 For L = 400 nm (bulk GaAs thin film), sharp exciton peak

appears.
 For L = 19.2 nm, 11.6 nm and 5 nm, multiple absorption
peaks can be observed.
 The exciton peaks in QW are blue shifted as L decreases.
The shift is larger for smaller L, clearly exhibiting the
quantum size effect.
 The splitting of heavy-hole and light-hole can be
observed, which give rise to 1e-hh, 1e-lh, 2e-hh, 2e-lh, 3e-
hh, 3e-lh transitions, etc.
 The linewidth (i.e. FWHM) of the exciton peaks in the QW
increases as L decreases due to fluctuations in the L
during the growth process. The L cannot be kept
absolutely constant, leading to Lz ± ∆Lz.
 Fluctuations in L is reflected as broadening of the exciton
absorption peak. For example, a typical broadening of 2
meV for 10 nm size of undoped GaAs well compared to
0.1meV in high quality bulk samples.
 Other broadening of the peaks can be due to interface
roughness and alloy composition fluctuation.
https://doi.org/10.1017/9781316535868.004

62
Optical absorption due to light and heavy holes in QW

63
𝑚𝑚𝑟𝑟
𝛼𝛼 2𝐷𝐷 ∝ 𝑁𝑁 � 𝛩𝛩(ℏ𝜔𝜔 − 𝐸𝐸𝑚𝑚𝑚𝑚 ) 𝑓𝑓𝑣𝑣 𝐸𝐸 − 𝑓𝑓𝑐𝑐 (𝐸𝐸) Absorption spectra from
𝜋𝜋ℏ2 𝑊𝑊 𝑤𝑤
𝑚𝑚,𝑛𝑛 a sample containing two
QWs with widths of 5
and 10 nm (shown in
lower and upper panels
respectively). The broken
curves show the PL from
these two wells. The
primes on E and H label
structures belonging to
the 5 nm well [
GaAs/AlGaAs quantum well
Example:
Consider a semiconductor single quantum well made of AlN/In0.15Ga0.85N/AlN layers.
i) Calculate the electron effective mass, hole effective mass and the band gap energy of
In0.15Ga0.85N layer using linear interpolation.
ii) Neglecting the exciton effect, calculate the minimum absorption photon energy required to
make a transition from the sublevel j = 2 in the valence band to the sublevel j = 2 in the
conduction band in the In0.15Ga0.85N quantum well of well width 50 Å with infinite barrier
heights for both the conduction band and the valence band.
iii) Calculate the lowest two-dimensional exciton binding energy for In0.15Ga0.85N. Assume that the
dielectric constant of In0.15Ga0.85N is the same as GaN.
Given GaN and InN parameters as follows:

Materials Bandgap (eV) 𝑚𝑚𝑒𝑒∗ /𝑚𝑚0 𝑚𝑚ℎ∗ /𝑚𝑚0 𝜀𝜀𝑟𝑟
GaN 3.3 0.15 0.8 8
InN 1.9 0.09 0.17 8

65
Solution:
(i) For In0.15Ga0.85N ,
1 1 1 1
= 0.85 + 0.15 ⇒ 𝑚𝑚𝑒𝑒∗ = 0.136𝑚𝑚𝑜𝑜 EC
𝑚𝑚𝑒𝑒∗ ∗
𝑚𝑚𝐺𝐺𝐺𝐺𝐺𝐺 ∗
𝑚𝑚𝐼𝐼𝐼𝐼𝐼𝐼 𝑚𝑚0 3
1 1 1 1 2
4ℏ2 𝜋𝜋 2
∗ = 0.85 ∗ + 0.15 ∗ ⇒ 𝑚𝑚ℎ∗ = 0.514𝑚𝑚𝑜𝑜 𝐸𝐸2𝑒𝑒 =
𝑚𝑚ℎ 𝑚𝑚𝐺𝐺𝐺𝐺𝐺𝐺 𝑚𝑚𝐼𝐼𝐼𝐼𝐼𝐼 𝑚𝑚0 J=1 2𝑚𝑚𝑒𝑒∗ 𝐿𝐿2
1 1 1 1 1
= + = + ⇒ 𝑚𝑚𝑟𝑟∗ = 0.108𝑚𝑚𝑜𝑜
∗ ∗ ∗
𝑚𝑚𝑟𝑟 𝑚𝑚𝑒𝑒 𝑚𝑚ℎ 0.136𝑚𝑚𝑜𝑜 0.514𝑚𝑚𝑜𝑜 Eg
J=1 4ℏ2 𝜋𝜋 2
𝐸𝐸𝑔𝑔 = 0.85𝐸𝐸𝑔𝑔𝐺𝐺𝐺𝐺𝐺𝐺 + 0.15𝐸𝐸𝑔𝑔𝐼𝐼𝐼𝐼𝐼𝐼 = 3.09 eV 𝐸𝐸2ℎ =
2 2𝑚𝑚ℎ∗ 𝐿𝐿2
EV 3
(ii) Minimum energy required:
L = 50 Å
2 2
ℏ 2𝜋𝜋
ℏ𝜔𝜔 = 𝐸𝐸𝑔𝑔 + 𝐸𝐸2𝑒𝑒 + 𝐸𝐸2ℎ = 3.09𝑒𝑒𝑒𝑒 + ℏ2 𝑛𝑛𝑥𝑥 𝜋𝜋 2 𝑚𝑚𝑒𝑒∗ = 0.136𝑚𝑚𝑜𝑜
2𝑚𝑚𝑟𝑟∗ 𝐿𝐿 𝐸𝐸𝑛𝑛𝑥𝑥 =
4 2𝑚𝑚 𝐿𝐿
= 3.09 + 0.377 × = 3.09 + 0.56 = 3.65𝑒𝑒𝑒𝑒
𝑚𝑚ℎ∗ = 0.514𝑚𝑚𝑜𝑜
0.108 × 52 1 n2x
= 0.377 ∗ 2 𝑚𝑚𝑟𝑟∗ = 0.108𝑚𝑚𝑜𝑜
(iii) 𝑚𝑚𝑟𝑟∗ 13.6 eV 0.108𝑚𝑚𝑜𝑜 13.6 eV m L
𝜀𝜀𝐵𝐵3𝐷𝐷 = = = 23 meV
𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟2 𝑚𝑚𝑜𝑜 82 𝐸𝐸𝑔𝑔 = 3.09 eV
𝜀𝜀𝐵𝐵2𝐷𝐷 = 4𝜀𝜀𝐵𝐵3𝐷𝐷 = 92 meV

66
Franz-Keldysh (FK) effect
 Franz–Keldysh effect (named after German physicist

Walter Franz and Russian physicist Leonid Keldysh) is
a change in the absorption of a bulk semiconductor
when an electric field is applied. x
 Under the effect of applied electric field, the bands CB
tilted which allows the electron and hole states to VB
tunnel into the bandgap. This leads to oscillations in
the absorption coefficient at ω > Eg and an d1
exponentially decaying tail is introduced below Eg as
a result of photon-induced tunnelling.
ξ1
 The carrier wavefunctions are no longer behaving as
plane-waves instead they can be described by an
Airy function.
 The probability of finding an electron in the energy
gap is described by exponentially decaying function
with Eg - ω for ξ > 0: x
4 2𝑚𝑚𝑒𝑒∗ 3/2
𝛼𝛼 ℏ𝜔𝜔 ∝ exp − 𝐸𝐸 − ℏ𝜔𝜔
3 𝑞𝑞 ℏ𝜉𝜉 𝑔𝑔 VB CB
𝑬𝑬𝒈𝒈
here, 𝜉𝜉 = . 𝑑𝑑 is the distance between two points d2
𝒅𝒅𝒅𝒅
at which the conduct and valence band edges are at
same energy due to band tilting caused by the ξ2
electric field and 𝜉𝜉.

67
Franz-Keldysh (FK) effect
 The barrier thickness (d) can be reduced with the assistance of incoming photon ω, as the
electron energy is raised to a high energy level at which the barrier becomes much thinner.
(𝐸𝐸𝑔𝑔 − ℏ𝜔𝜔)
𝑑𝑑 ′ =
𝑞𝑞𝑞𝑞 photon energy (eV)
1.51 1.48 1.44 1.41 1.38
 For ℏ𝜔𝜔 > 𝐸𝐸𝑔𝑔 , 𝛼𝛼 exhibits Franz-Keldysh oscillations
(derivation out of scope of this module).
Eg
Bandgap of GaAs
ω
VB CB
4 2𝑚𝑚𝑒𝑒∗ 3/2
(𝐸𝐸𝑔𝑔 − ℏ𝜔𝜔) 𝛼𝛼 ℏ𝜔𝜔 ∝ exp − 𝐸𝐸 − ℏ𝜔𝜔
3 𝑞𝑞 ℏ𝜉𝜉 𝑔𝑔
𝑑𝑑 ′ =
d’ 𝑞𝑞𝑞𝑞
Bulk semiconductor under the applied ξ with assistance of incoming photon (b) Absorption coefficient
of GaAs for various electric field strengths.
68
Franz-Keldysh (FK) effect-based modulator
 The FK effect can be used to modulate light. In this type of electroabsorption modulation, the
wavelength is typically chosen to be slightly smaller than the bandgap wavelength so that
absorption is negligible. When a field is applied, the absorption is enhanced by the Franz–
Keldysh effect.
+
photon energy (eV)
1.51 1.48 1.44 1.41 1.38

Incident Modulated
beam beam
λop
Semiconductor
= 877.5 nm
T _
0.125
t
λξ=0 = 875 nm(Eg=1.417 eV)
λop = 877.5 nm(1.413 eV)

69
Quantum confined Stark effect
 Electric-field induced changes in the optical

spectra near the band edge is significant and
results in FK effect. In a QW such effect depends
critically upon the direction of the applied field
with respect to the growth direction.
 A static electric field modifies the crystal
potential from its band value 𝑉𝑉𝑐𝑐 𝑟𝑟⃑ to 𝑉𝑉𝑐𝑐 𝑟𝑟⃑ −
⃑ This tilts the conduction and valence band
𝑒𝑒ξ𝑟𝑟.
edges such that electron’s potential energy
increases and the hole’s energy decreases by the
same amount. The net effect is to pull electron
∆α = 103 −104 cm-1

and hole in the opposite direction.
 The electric field lowered the energy of the
individual bound states of both electrons and
holes. Electron and hole shifts combine to reduce
the inter-band transition energy in the QW. This
red-shift of the absorption line is known as the
quantum confined Stark effect (QCSE).
Absorption spectra for a 94 Å AlGaAs/GaAs QW as a function of normal electric field (i) 0 KV/cm; (ii) 6
KV/cm, (iii) 110 KV/cm, (iv) 150 KV/cm and (v) 200 KV/cm. The excitons effects are preserved up to very
high fields because the confining well suppresses the dissociation process; i.e. the exciton cannot ionize
because the electron and hole are trapped in the quantum well.

70
Quantum confined Stark effect
Measured
photocurrent spectra
at 2K of an
InGaAs/InGaAlAs (a)
and an
InGaAsP/InGaAsP (b)
Electro-absorption
modulator at several
bias voltages

71

quantum structures
‒ 2D excitons
‒ Vegard’s law
optoelectronics
‒ Superlattice
‒ Photodiode (PD)

72
Application of semiconductors in optoelectronics
 Light emitting diode

 Semiconductor Lasers
 Optical Modulators
 Photodetectors and Photovoltaic Devices
 Quantum Cascade Lasers
 Quantum Well Infrared Photodetectors (QWIPs)
 Quantum Dot Infrared Photodetectors (QDIPs)
 Single-Photon Sources

73
LED applications
AlInGaP
AlInGaP
InGaN
First goggle with integrated white LEDs used for

illuminated during medical surgery.
Full-color display used in the The Brandenburg Gate in

2008 Beijing Olympic Games Berlin, Germany, illuminated
with LEDs.
LEDs for indoor farming First automotive daytime running light

based on LEDs (Audi car company)

74
Light emitting diode (LED)
 A light-emitting diode, commonly known as an LED, is a semiconductor device that emits
light when an electric current passes through it. LEDs are widely used in various applications
such as indicator lights, displays, automotive lighting, and general lighting due to their high
energy efficiency, long lifetime, and compact size. They come in different colors and can be
found in a wide range of industrial and consumer products, and lighting fixtures.
 The simplest LED structure is a pn junction. When the junction is forwarded, electrons and
holes are injected into p and n regions, respectively. The recombination of electrons and
holes leads to light emission, if the recombination is dominated by radiative process.
Lehovec K., Accardo C. A. and Jamgochian E.

“Injected light emission in silicon carbide crystals”
Physical Review 83, 603 (1951)
Note: the main reference for the LED part is the

book by E. F. Schubert including some of the
illustrations. For sake of clarity, we may not cite
the reference in each slide.
LIGHT-EMITTING
DIODES
SECOND EDITION
E. FRED SCHUBERT
Rensselaer Polytechnic Institute,
Troy, New York
E. Fred Schubert, 2023, ISBN: 978-0-9 863826-6-6

75
Evolution of LED performance

76
LED performance versus cost

78
Factors that affect LED efficiency
 The efficiency of LED is limited by a variety of factors which include but are not limited to:
a) Contact resistance (forward-voltage efficiency)
b) Carrier overflow (injection efficiency)
c) Material quality of the active region (radiative efficiency)
d) Ability to extract light from the LED chip (light-extraction efficiency)
e) Emission spectrum (luminous efficacy of radiation).
E. Fred Schubert, 2023,

ISBN: 978-0-9 863826-6-6
Optical
Electrical
output
input
power
power
1 − 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 1 − 𝜂𝜂𝐼𝐼𝐼𝐼 1 − 𝜂𝜂𝑅𝑅𝑅𝑅 1 − 𝜂𝜂𝐿𝐿𝐿𝐿𝐿𝐿
Definition of 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 : forward-voltage efficiency 𝜂𝜂𝑅𝑅𝑅𝑅 : Radiation efficiency

efficiencies 𝜂𝜂𝐼𝐼𝐼𝐼 : injection efficiency 𝜂𝜂𝐿𝐿𝐿𝐿𝐿𝐿 : Light-extraction efficiency
𝜂𝜂𝐼𝐼𝐼𝐼𝐼𝐼 : internal quantum efficiency 𝜂𝜂𝐸𝐸𝐸𝐸𝐸𝐸 : external quantum efficiency
𝜂𝜂𝑃𝑃𝑃𝑃𝐸𝐸 : power conversion efficiency
𝜂𝜂𝑃𝑃𝑃𝑃𝐸𝐸 = 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 × 𝜂𝜂𝐼𝐼𝐼𝐼 × 𝜂𝜂𝑅𝑅𝑅𝑅 × 𝜂𝜂𝐿𝐿𝐿𝐿𝐿𝐿 = 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 × 𝜂𝜂𝐼𝐼𝑄𝑄𝐸𝐸 × 𝜂𝜂𝐿𝐿𝐿𝐿𝐿𝐿 = 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 × 𝜂𝜂𝐸𝐸𝑄𝑄𝑄𝑄

79
Forward-voltage efficiency (𝜼𝜼𝑭𝑭𝑭𝑭𝑭𝑭 )
 The energy of photons emitted from a semiconductor with
bandgap 𝐸𝐸𝑔𝑔 is roughly given by the bandgap energy, i.e.,
ℏ𝜔𝜔 ≈ 𝐸𝐸𝑔𝑔
 Consider an ideal diode in which every electron injected into the
active region will generate a photon with energy ≈ 𝐸𝐸𝑔𝑔 . Then,
energy conservation requires the bias voltage to be given by
𝑉𝑉𝑓𝑓 = 𝐸𝐸𝑔𝑔 /𝑒𝑒
i.e., the voltage applied to the LED is equal to the photon energy
divided by the elementary charge.
 In actual devices, however, there is always an added parasitic series
resistance due to i) contact resistance, (ii) resistances caused by
abrupt junctions, and (iii) bulk resistance of the constituent layers.
This will lead to a bias voltage higher than 𝐸𝐸𝑔𝑔 /𝑒𝑒.
 The power consumed by the parasitic resistance is given by
𝑃𝑃𝑅𝑅 = 𝐼𝐼(𝑉𝑉𝑓𝑓 − ℏ𝜔𝜔/𝑒𝑒)
 Therefore, the forward-voltage efficiency is given by
𝐼𝐼𝑉𝑉𝑓𝑓 − 𝑃𝑃𝑅𝑅 𝐼𝐼𝑉𝑉𝑓𝑓 − 𝐼𝐼(𝑉𝑉𝑓𝑓 − ℏ𝜔𝜔/𝑒𝑒) ℏ𝜔𝜔

𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 = = =
𝐼𝐼𝑉𝑉𝑓𝑓 𝐼𝐼𝑉𝑉𝑓𝑓 𝑒𝑒𝑉𝑉𝑓𝑓

80
Injection efficiency (𝜼𝜼𝑰𝑰𝑬𝑬 )
 The light emission in LED is mainly from the
active region. The actual width or volume
ratio of the active region over other regions
depends on the LED design.
 Most but not all carries are injected into the
active region. Some carriers may leak out of
the active region or transverse the active
region without being captured.
 Although it is mostly the electrons that leak
out (due to their usually lighter mass and
higher mobility), in the steady state, the
number of electrons injected into the active
region is equal to the number of holes
injected into the active region; therefore, the
injection efficiency of holes always is equal to
the injection efficiency of electrons.
 The injection efficiency can be defined as
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑝𝑝𝑝𝑝𝑝𝑝 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

𝜂𝜂𝐼𝐼𝐸𝐸 =
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝐿𝐿𝐿𝐿𝐿𝐿 𝑝𝑝𝑝𝑝𝑝𝑝 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

81
Radiation efficiency (𝜼𝜼𝑹𝑹𝑬𝑬 )
 Not all the carriers injected into the active
region will recombine and emit light.
 The radiative efficiency is defined as the ratio
between radiative rate and total
recombination rate:
1/𝜏𝜏𝑅𝑅𝑅𝑅𝑅𝑅
𝜂𝜂𝑅𝑅𝐸𝐸 =
1/𝜏𝜏𝑅𝑅𝑅𝑅𝑅𝑅 + 1/𝜏𝜏𝑁𝑁𝑁𝑁𝑁𝑁−𝑟𝑟𝑟𝑟𝑟𝑟
Internal quantum efficiency (𝜼𝜼𝑰𝑰𝑰𝑰𝑰𝑰 )

 The internal quantum efficiency (IQE) is defined as
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑝𝑝𝑝𝑝𝑝𝑝 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

𝜂𝜂𝐼𝐼𝐼𝐼𝐸𝐸 =
 Based on the definitions of injection and radiative

efficiencies, the internal quantum efficiency can also
be written as
𝜂𝜂𝐼𝐼𝐼𝐼𝐸𝐸 = 𝜂𝜂𝐼𝐼𝐼𝐼 × 𝜂𝜂𝑅𝑅𝐸𝐸
𝜂𝜂𝐼𝐼𝐼𝐼𝐸𝐸 = 𝜂𝜂𝐼𝐼𝐼𝐼 × 𝜂𝜂𝑅𝑅𝐸𝐸

82
Low 𝜼𝜼𝑰𝑰𝑰𝑰𝑬𝑬 in homojunction LED
 The basic structure of LED is a pn junction of the same type of
semiconductor material (homojunction). Under forward bias,
minority carriers are injected from both sides of the junction.
When the injected minority carriers recombine with the
majority carriers in the neutral regions, photons are emitted
by the radiative process.
 The internal quantum efficiency of such kind of homojunction
LED is generally low because of Forward-biased pn junction
i) large diffusion length of injected minority carriers in
the neutral region, leading to low recombination rate;
ii) re-absorption of emitted photons as ℏ𝜔𝜔 ≈ 𝐸𝐸𝑔𝑔 ;
iii) non-radiative recombination in a large volume. Light output
 As a mitigation measure, the pn junction LED may be designed
so that the photons are emitted close to the top layer and not p𝒑𝒑 Epitaxial
in the buried layer (substrate). n𝒏𝒏++ layers
 Usually, the top layer is p type. For photons to be emitted n +
𝒏𝒏+ 𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒𝐒
from the p layer, we can make the n layer heavily doped such Substrate
that the diode current is dominated by the electron current,
i.e., Jn >> Jp .

83
I-V characteristics of a pn junction
𝐷𝐷𝑛𝑛 𝑛𝑛𝑖𝑖2 𝐷𝐷𝑝𝑝 𝑛𝑛𝑖𝑖2 𝑒𝑒𝑒𝑒 Dn : electron diffusion coefficient

𝐼𝐼 = 𝑒𝑒𝑒𝑒 + 𝑒𝑒𝑒𝑒𝑒𝑒 −1 Dp : hole diffusion coefficient
𝐿𝐿𝑛𝑛 𝑁𝑁𝑎𝑎 𝐿𝐿𝑝𝑝 𝑁𝑁𝑑𝑑 𝑘𝑘𝐵𝐵 𝑇𝑇
Ln : electron diffusion length
(Shockley diode equation (valid for low injection) Lp : hole diffusion length
Nd : donor concentration in n-region
𝐷𝐷𝑛𝑛 𝑛𝑛𝑖𝑖2 𝑒𝑒𝑒𝑒
𝐼𝐼 ≈ 𝑒𝑒𝐴𝐴 𝑒𝑒𝑒𝑒𝑒𝑒 −1 for 𝑝𝑝𝑝𝑝+ junction Na : acceptor concentration in p-region
𝐿𝐿𝑛𝑛 𝑁𝑁𝑎𝑎 𝑘𝑘𝐵𝐵 𝑇𝑇
ni : intrinsic carrier density
A: cross section area
𝑒𝑒𝑒𝑒𝑛𝑛𝑖𝑖 𝑊𝑊𝑝𝑝 𝑊𝑊𝑛𝑛

𝐼𝐼𝐺𝐺𝐺𝐺 = +
2 τ𝑒𝑒 τℎ
𝑒𝑒𝑒𝑒
× 𝑒𝑒𝑒𝑒𝑒𝑒 −1
2𝑘𝑘𝐵𝐵 𝑇𝑇
Trap assisted generation-

recombination current
density in the depletion region
Wn : depletion width in n-region

Wp : depletion width in p-region
τℎ : hole recombination time
τ𝑒𝑒 : electron recombination time

84
I-V characteristics of a pn junction
 𝜂𝜂𝑅𝑅𝐸𝐸 is high in this region as the current is dominated In the high injection region, the injection
by the radiative recombination current outside the efficient will decease due to the increased
depletion region (or bulk region); potential drop across the bulk region,
 Trap assisted recombination is still present in both particularly in the low doped region. This
the depletion and bulk regions, but it is insignificant region is also called the series-resistance
compared to radiative recombinations. limited regime.
𝜂𝜂𝑅𝑅𝐸𝐸 is low in this region due to

relatively large trap assisted
current (SRH-recombination
current) in the depletion
region (large depletion width
at very low forward bias)

85
Example:
Consider a GaAs p-n LED with an area of 1mm2 and is operated at a forward bias of 1 V at 300K.
(a) Calculate
(i) electron current injected into the p-side. (b)(i)
(ii) hole current injected into the n-side.
𝐼𝐼 0.9×10−3
(b) If the internal quantum efficiency ηIQE = 0.5, calculate 𝑛𝑛𝑝𝑝𝑝 = η𝐼𝐼𝐼𝐼𝐼𝐼 = 0.5
𝑒𝑒 1.6×10−19
(i) the rate of photon generated from the active region of the LED.
(ii) optical power generated from the active region of the LED. = 2.8125 × 1015 s-1
Parameters: Electron diffusion coefficient: Dn = 200 cm2/s
(b)(ii)
Hole diffusion coefficient: Dp = 10 cm2/s
𝐼𝐼
p-side doping: Na = 5× 1016 cm-3 𝑃𝑃𝑖𝑖𝑖𝑖𝑖𝑖 = η𝐼𝐼𝐼𝐼𝐼𝐼 ω
𝑒𝑒
n-side doping: Nd = 5× 1017 cm-3
Electron minority carrier lifetime: τn = 10-8 s = 2.8125 × 1015 × 1.43 ×
Hole minority carrier lifetime: τp = 10-7 s 1.6 × 10−19
Intrinsic carrier concentration: ni = 1.84 × 106 cm-3
Bandgap Eg = 1.43 eV = 643.5 × 10-6 W = 0.64 mW
Solution:
𝐿𝐿𝑛𝑛 = 𝐷𝐷𝑛𝑛 𝜏𝜏𝑛𝑛 = 200 × 10−8 = 14.14 𝜇𝜇𝜇𝜇; 𝐿𝐿𝑝𝑝 = 𝐷𝐷𝑝𝑝 𝜏𝜏𝑝𝑝 = 10 × 10−7 = 10 𝜇𝜇𝜇𝜇
2
𝐷𝐷𝑛𝑛 𝑛𝑛𝑖𝑖2 𝑒𝑒𝑒𝑒 10−2 𝑐𝑐𝑐𝑐2 1.6×10−19 𝐶𝐶 200𝑐𝑐𝑐𝑐2 /𝑠𝑠 1.48×10−5 𝑐𝑐𝑐𝑐−3 1
(a)(i) 𝐼𝐼𝑛𝑛 = 𝐴𝐴𝐴𝐴 𝑒𝑒𝑒𝑒𝑒𝑒 −1 = 𝑒𝑒𝑒𝑒𝑒𝑒 − 1 = 0.9 𝑚𝑚𝑚𝑚
𝐿𝐿𝑛𝑛 𝑁𝑁𝑎𝑎 𝑘𝑘𝐵𝐵 𝑇𝑇 (1.41×10−3 𝑐𝑐𝑐𝑐) (5×1016 )2 0.0259
2
𝐷𝐷𝑝𝑝 𝑛𝑛𝑖𝑖2 𝑒𝑒𝑒𝑒 10−2 𝑐𝑐𝑐𝑐2 1.6×10−19 𝐶𝐶 10𝑐𝑐𝑐𝑐2 /𝑠𝑠 1.48×10−5 𝑐𝑐𝑐𝑐−3 1
(a)(ii) 𝐼𝐼𝑝𝑝 = 𝐴𝐴𝐴𝐴 𝑒𝑒𝑒𝑒𝑒𝑒 −1 = 𝑒𝑒𝑒𝑒𝑒𝑒 − 1 = 6.4 𝜇𝜇𝐴𝐴
𝐿𝐿𝑝𝑝 𝑁𝑁𝑑𝑑 𝑘𝑘𝐵𝐵 𝑇𝑇 (1×10−3 𝑐𝑐𝑐𝑐)(5×1017 ) 0.0259

86
Improve 𝜼𝜼𝑰𝑰𝑰𝑰𝑬𝑬 using hetero- or quantum structures
 As shown in the previous example, in typical
semiconductors, the diffusion length is of the order of AlGaAs GaAs AlGaAs
several micrometers. In other words, in a typical pn junction,
the recombination of minority carriers occurs in a large
Cladding
region with varying carrier density, which is undesirable for
layer Cladding
efficient LED operation (re-absorption, non-radiative
layer
recombination, etc.). Reducing the thickness of p-region
may help to suppress re-absorption, but it may lead to more
non-radiative recombinations due to surface defects. Active region
 Therefore, all energy-efficient LEDs employ the
 In addition, the DH structure can also be
heterojunction design, particularly the double
used to confine light to the active region
heterostructure (DH), which has several advantages over the
to produce edge emitting LEDs.
homojunction designs including
 Although the DH structure helps to
 confinement of carriers in a narrow active religion improve the LED efficiency, there are also
(𝑊𝑊𝐷𝐷𝐷𝐷 ≈ 0.01 − 1𝜇𝜇𝜇𝜇 ≪ 𝐿𝐿𝑛𝑛 or 𝐿𝐿𝑝𝑝 )  high carrier problems associated with the design such
density  large radiative recombination rate (𝑅𝑅 = as the resistance caused by the hetero
𝐵𝐵𝐵𝐵𝐵𝐵). interface and the trade-off between
carrier overflow and diode voltage.
 less exposure of carriers to deep levels (due to small
volume of the active region)  reduced non-
radiative recombination
 reduced re-absorption as the energy of emitted
photons is smaller than the bandgap of cladding (or
confinement) layers

87
p-electrode
p-GaN
p-electrode p-Al0.15Ga0.85N
n-electrode n-In0.15Ga0.85N
p-GaN n-Al0.15Ga0.85N n-electrode
n-GaN n-GaN
GaN buffer layer GaN buffer layer
Sapphire Substrate Sapphire Substrate
(a) (b)
(a) Structure of a GaN p-n homojunction LED and (b) Structure of a blue InGaN/AlGaN DH-LED

88
 When the active region width is reduced to the regime

where quantum confinement effect becomes prominent
at room temperature, the DH becomes essentially a
quantum well (QW).
 Compared to the DH structure, the QW-based LED can
achieve even higher efficiency due to the tight
confinement of carriers in a very narrow region, leading
to more efficient radiative recombination. In addition, it
also offers the following advantages such as
DH SQW
 tunable wavelength (it can be adjusted though the
well width)
 faster response (due to small active volume)
 better temperature stability (due to step-like DOS)
 However, the strong confinement will enhance the

overflow of carriers due to higher electron/hole energy.
The reduced volume will also lead to a decrease in
brightness.
 Therefore, in actual design, multiple quantum wells
(MQW) instead of single quantum wells (SQW) are used.
MQW

89
Experimental results of an LED structure with one, four, six and eight QWs. The light intensity for the
single QW structure saturates at low current level. As the number of QWs is increased, the current
level at which saturation occurs increases, and the optical saturation intensity increases as well. The
saturation of the light intensity is caused by the overflow of carriers.

90
Light extraction efficiency (𝜼𝜼𝑳𝑳𝑳𝑳𝑬𝑬 )
 Light-extraction efficiency in LED refers to the effectiveness with which light

generated inside the device is extracted and emitted into the surrounding
medium, typically air or another material. Light generated within the device
may undergo multiple reflections at interfaces between materials with
different refractive indices. These reflections can cause some of the light to be
trapped inside the device, reducing the overall efficiency.
 Therefore, virtually all LED chips are encapsulated with optically transparent
material such as epoxy resin so as to reduce the light reflection at the outer
surface, particularly the total internal reflection (TIR).

91
 The TIR occurs when the incidence angle is larger than the critical angle (𝜃𝜃𝑐𝑐 )
1
defined by 𝜃𝜃𝑐𝑐 = 𝑠𝑠𝑠𝑠𝑠𝑠−1 . Note: here the subscript (𝑠𝑠𝑠𝑠b) means the topmost
𝑛𝑛𝑠𝑠𝑠𝑠𝑠𝑠
layer of the device not necessarily the substrate.
 Suppose we have a point-light source inside the semiconductor with a total
emission power of 𝑃𝑃𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 . The percentage of power within the so-called “escape
cone” (0 ≤ 𝜃𝜃 ≤ 𝜃𝜃𝑐𝑐 ) is
2𝜋𝜋 𝜃𝜃
𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 ∫0 ∫0 𝑟𝑟 2 sin 𝜃𝜃 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 2𝜋𝜋𝑟𝑟 2 (1 − cos 𝜃𝜃𝑐𝑐 )
𝑐𝑐
(1 − cos 𝜃𝜃𝑐𝑐 )
= 2𝜋𝜋 𝜋𝜋 = =
𝑃𝑃𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 ∫0 ∫0 𝑟𝑟 2 sin 𝜃𝜃 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 4𝜋𝜋𝑟𝑟 2 2
𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 1 𝜃𝜃𝑐𝑐2 1 1
 When 𝑛𝑛𝑠𝑠𝑠𝑠𝑠𝑠 ≫ 1, 𝜃𝜃𝑐𝑐 is small, ≈ 1− 1− = 2
𝑃𝑃𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 2 2 4 𝑛𝑛𝑠𝑠𝑠𝑠𝑠𝑠
Example:
Light
θc escape For GaAs, 𝑛𝑛𝑠𝑠𝑠𝑠𝑠𝑠 = 3.6358 at λ = 867 𝑛𝑛𝑛𝑛
cone
1
Point light source θ𝑐𝑐 = 𝑠𝑠𝑠𝑠𝑠𝑠−1 = 16o
3.6358

92
 The use of a hemispherical epoxy dome ensures

that the angle critical to the air/epoxy interface
approaches 90o, and at the same time, the critical
angle at the epoxy/semiconductor interface is
significantly enhanced due to the much smaller 𝑛𝑛 = 1.4 − 1.8
difference between the refractive indices at
epoxy/semiconductor interface in contrast to the
air/semiconductor interface.
 Take GaAs as an example, 𝑛𝑛𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 = 3.6358 at λ =
867 𝑛𝑛𝑛𝑛. The critical angle at the air/GaAs interface
is given by
1 1
θ𝑐𝑐 = 𝑠𝑠𝑠𝑠𝑠𝑠−1 = 𝑠𝑠𝑠𝑠𝑠𝑠−1 = 16𝑜𝑜
𝑛𝑛𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 3.6358
 When it is encapsulated in a hemispherical epoxy

dome with refractive index of 𝑛𝑛𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 = 1.8. The
critical angle at the air/epoxy interface is 90o, while
the critical angle at the epoxy/GaAs interface is
𝑛𝑛𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 1.8
θ𝑐𝑐 = 𝑠𝑠𝑠𝑠𝑠𝑠−1 = 𝑠𝑠𝑠𝑠𝑠𝑠−1 = 29.7𝑜𝑜
𝑛𝑛𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 3.6358

93
 Alternatively, one may also use a textured surface to improve the light extraction
efficiency. In this case, Light emitting from the active region below the base of the
cone undergo multiple reflections until they eventually have near-normal incidence
at the semiconductor-air interface and then escape from the chip.
 Another way is to use truncated-inverted pyramid (TIP) geometry. The TIP structure
can improve light extraction by redirecting the TIR photons from the sidewall
interfaces towards the top surface of the chip near normal incidence, allowing
them to escape. In addition, photons that are TIR at the top surface are redirected
for escape through the sidewalls. These two processes provide the TIP device with
a significant reduction in photon path length for extraction compared to a
conventional shaped chip.
n GaP
p GaP
AlGaInP
Textured surface AlGaInP/GaP TIP Photomicrograph of a TIP

structure LED under forward bias

94
External quantum efficiency (𝜼𝜼𝑬𝑬𝑬𝑬𝑬𝑬 )
 The external quantum efficiency is the ratio of the number of
useable photons to the number of injected charge carriers
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝𝑝𝑝 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

𝜂𝜂𝐸𝐸𝑄𝑄𝑄𝑄 =
𝑃𝑃𝑝𝑝𝑝 /ℏ𝜔𝜔
= = 𝜂𝜂𝐼𝐼𝐼𝐼𝐼𝐼 × 𝜂𝜂𝐿𝐿𝐿𝐿𝐿𝐿
𝐼𝐼/𝑒𝑒
Power-conversion efficiency (𝜼𝜼𝑷𝑷𝑷𝑷𝑬𝑬 )

 The power-conversion efficiency (PCE) of a light-emitting device is defined as the
optical output power emitted by the LED, 𝑃𝑃optical, divided by the electrical input
power, i.e.,
𝑃𝑃𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
𝜂𝜂𝑃𝑃𝑃𝑃𝐸𝐸 = = 𝜂𝜂𝑉𝑉𝑉𝑉𝑉𝑉 × 𝜂𝜂𝐸𝐸𝐸𝐸𝐸𝐸
𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒
𝜂𝜂𝑃𝑃𝑃𝑃𝐸𝐸 = 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 × 𝜂𝜂𝐼𝐼𝐼𝐼 × 𝜂𝜂𝑅𝑅𝑅𝑅 × 𝜂𝜂𝐿𝐿𝐿𝐿𝐿𝐿 = 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 × 𝜂𝜂𝐼𝐼𝑄𝑄𝐸𝐸 × 𝜂𝜂𝐿𝐿𝐿𝐿𝐿𝐿 = 𝜂𝜂𝐹𝐹𝐹𝐹𝐹𝐹 × 𝜂𝜂𝐸𝐸𝑄𝑄𝑄𝑄

95
White LED
 The LEDs are inherently monochromatic emitters. However, there are several
ways to generate white light using the LEDs.
 One of the approaches is to use multiple LEDs with different colors, which
again can be done in different ways, as illustrated below.

96
White LED
 High efficiency devices are available in the red
wavelength region (AlGaInP) and blue wavelength
region (InGaN). However, in the center of the visible
spectrum, particularly near 550 nm (the green
wavelength region), the efficiency of the devices is
significantly lower, commonly referred to as the “green
gap”.
 Despite the lower efficiency of green LEDs, they can
still be used to generate white light by controlling the
driving the current and emission wavelength. The
trichromatic emission spectrum of a white-light source
made out of three types of LEDs emitting at 455, 525,
and 605 nm is shown below.
External quantum efficiency for high-power
visible-spectrum LEDs made from the
nitride and phosphide materials. The
dashed lines are guides to the eye.

97
White LED
 The second approach is to use LED to
excite phosphors to generate the white
light.
 The combination of blue light from the
LED and yellow light from the phosphor
results in a mixture of blue and yellow
light. When properly balanced, this
mixture appears as white light to the
human eye.
 The color temperature of the resulting
white light can be adjusted by varying
the composition and concentration of
the phosphor material. Different
phosphor formulations produce white
light with varying color temperatures,
ranging from warm white (lower color
temperature, more yellowish) to cool
white (higher color temperature,
bluish).
 Phosphor-converted white LEDs can (a) Structure of the first white LED. (b) InGaN LED chip emitting blue
electroluminescence (EL) and YAG:Ce phosphor, excited by blue EL,
achieve high levels of efficiency and emitting broadband yellow fluorescence ranging from green to red
efficacy, especially when using thereby generating white light. (c) Emission spectrum of the first white
advanced phosphor materials and LED LED showing distinct blue EL line at 460 nm and broad fluorescence
chip designs. band ranging from green to red
Laser Photonics Rev., 1600147 (2017)

98
https://www.meetoptics.com/academy/super-luminescent-diodes

99
Semiconductor laser (or laser diode)
 Recall that the threshold current for lasing is
‒ Defect reduction
1 1 1 1 ‒ Strain effect
𝑖𝑖𝐷𝐷𝑡𝑡𝑡 = 𝛼𝛼𝑖𝑖 + 𝑙𝑙𝑙𝑙 + 𝑛𝑛𝑡𝑡𝑡𝑡 𝐺𝐺 = 𝐺𝐺𝑖𝑖 Γ
𝐾𝐾 𝐺𝐺 2𝐿𝐿 𝑅𝑅1 𝑅𝑅2 ‒ Quantum confinement
 To reduce 𝑖𝑖𝐷𝐷𝑡𝑡𝑡 , we have to

‒ Double heterostructure
‒ Quantum confinement (SQW,
 increase 𝐺𝐺𝑖𝑖  improve material properties MQW, MMQW, GSCH (graded
 reduce 𝛼𝛼𝑖𝑖  reduce internal loss index separate confinement
heterostructure)
 increase 𝐾𝐾  increase injection efficiency
 increase Γ  enhance optical confinement
 reduce𝑛𝑛𝑡𝑡𝑡𝑡  reduce DOS of active region
 reduce 𝛼𝛼𝑚𝑚  improve cavity design

100
Basic characteristics of LD
 When the laser diode is forward biased, electrons

and holes are injected into the active region.
Gain vs. photon energy
These electrons and holes recombine to emit
curves for a variety of
photons. carrier injections for
 When the forward bias current is small, the GaAs at 300 K. The
number of electrons and holes injected are small. electron and hole
injections are the same.
As a result, the photons that are emitted are The injected carrier
either absorbed in the cavity or lost to the densities are increased
outside. in steps of 0.25 ×
1018 𝑐𝑐𝑐𝑐−3 from the
 As the forward bias increases, more carriers are
lowest value shown.
injected into the device until eventually the
threshold condition is satisfied for some photon
energy, leading to building up of photon numbers
in the cavity.
 At this point, the gain in the active region turns
from negative to positive over some region of Typical optical
power output vs.
energy, as shown in the right figure for GaAs. The
forward current
gain only becomes positive when the condition for a LED and a
𝐸𝐸𝐹𝐹𝐹𝐹 − 𝐸𝐸𝐹𝐹𝐹𝐹 > ℏ𝜔𝜔 ≥ 𝐸𝐸𝑔𝑔 is satisfied. laser diode.
 As the bias is further increased, stimulated
emission starts to dominate the spontaneous
emission, leading to lasing from the device.

101
Basic characteristics of LD
 The threshold current is temperature dependent:

𝐼𝐼𝑡𝑡𝑡 = 𝐼𝐼𝑡𝑡𝑡𝑡 exp(𝑇𝑇/𝑇𝑇0 ) 0𝑜𝑜 𝐶𝐶 50𝑜𝑜 𝐶𝐶
25𝑜𝑜 𝐶𝐶
here 𝑇𝑇0 represents a characteristic temperature. The
larger the 𝑇𝑇0 , the better the thermal stability of the
laser.
 The optical cavity is essentially a resonant cavity in
which the photons undergo multiple reflections. The
cavity is designed such that when photons are emitted,
only a small fraction is allowed to leave the cavity. As a
result, the photon density starts to build up in the
cavity.
 For semiconductor lasers, the most widely used cavity
is the Fabry-Perot cavity formed by cleaved or polished
facets of the semiconductor that forms the pn juncion.
 The resonant modes are those that satisfy the
condition
𝜆𝜆0
𝐿𝐿 = 𝑚𝑚 (𝑚𝑚 = 1, 2,3 … )
2𝑛𝑛
where 𝐿𝐿 is the cavity length, 𝜆𝜆0 is the light
wavelength in air and is the refractive index of the
active region. From F. Grillot

102
Double heterostructure LD
 Like LED, a semiconductor laser, or laser diode (LD), can

be made from a forward-biased pn junction. But the
efficiency of homojunction LD is typically low, therefore
most LDs employ either DH or QW structures.
 Shown in the right is the schematic of a typical edge-
emitting laser which consists of a DH formed by an
undoped (or lightly p-doped) active region surrounded by
higher bandgap p and n AlxGa1-xAs cladding layers.
 Under forward bias, electrons and holes are injected from
the AlGaAs cladding layers to the GaAs active region. The
cladding layers provide an energy barrier to confine
carriers to the active region.
 The confinement of electrons and holes in the active
layer facilitates creating a state of population inversion,
allowing the amplification of light by stimulated emission.
 Since the active region has a smaller bandgap than the
cladding layers, its refractive index will be slightly larger
than that of the surrounding layers, leading to light
confinement in the active region as well, desirable for
obtaining stimulated emission.

103
Quantum well lasers
 Quantum confinement in semiconductors leads to quantized energy levels, varying
bandgaps, and significant changes in the density of states. When applied to lasers, the
quantum well brings several advantages:
 Reduced threshold current: Quantum well lasers have lower
threshold currents due to the quantum confinement effect,
where electrons and holes are confined in narrow potential
wells, resulting in higher carrier density and reduced lasing
threshold.
 Higher efficiency: The confinement of charge carriers within
quantum wells enhances carrier recombination efficiency,
leading to higher overall efficiency compared to DH lasers.
 Wavelength flexibility: Quantum well lasers offer flexibility in
controlling the emission wavelength by adjusting the thickness
and composition of the quantum wells, allowing for a wide range
of emission wavelengths across the spectrum.
 Higher modulation speed: Due to their reduced carrier transit
time and capacitance, quantum well lasers can achieve higher
modulation speeds, making them suitable for high-speed optical
communication applications.
 Temperature stability: Quantum well lasers exhibit improved
temperature stability compared to bulk lasers due to the
modified DOS.
 Single-Mode Operation: Quantum well lasers can be designed
for single-mode operation more easily than bulk lasers, resulting
in a narrower linewidth and improved spectral purity.

104
Some early works on quantum well lasers
Appl. Phys. Lett. 33, 596 (1978)
Appl. Phys. Lett. 40, 939 (1982)
• Low threshold
• Better temperature sensitivity
• Ease control of wavelength

105
Different types of confinement structures
 The QW laser was first suggested by Dingle et al. in 1974 and

implemented for the first time in 1978 by N. Holonyak and
co-authors (see previous slide). The advantages of quantum
well lasers are so considerable that this type of laser
nowadays comprises essentially the entire commercial
market for laser diodes.
 Due to the significantly smaller de Broglie wavelengths of
electrons and holes in semiconductors compared to laser
radiation wavelengths, quantum well-based structures can
be integrated into a graded-index design as suggested by
Tsang in 1981, leading to further reduction in threshold
current density.
 Building on this concept, Z. I. Alferov and co-workers (1988)
proposed and successfully implemented a laser structure
featuring a single quantum well surrounded by few-period
superlattices on both sides. This innovative design not only
establishes a graded refractive index profile but also acts as a
barrier, effectively minimizing undesirable dislocation
motions within the active quantum well region. As a result,
achieving threshold currents below 50 A/cm2 has become
feasible.
Sergey V. Gaponenko, Hilmi Volkan Demir.
Tsang, W. T. (1981). Appl Phys Lett, 39 134–137. (2019)
Alferov, Z. I. (1998). Semiconductors, 32, 1–14.

106
Distributed feedback lasers
 Many laser applications require a narrow-band
emission spectrum and low noise. A simple Fabry–
Pérot edge-emitting laser is good for many
applications, but it is difficult to produce single
emission line centering on a specific wavelength.
 In 1971 H. Kogelnik and C. V. Shank proposed
spatially periodic refractive index of the gain
medium or gain itself to promote generation at a
given mode of a cavity, and they called it distributed
feed back (DFB) laser.
 In a DFB laser, a grating structure is typically
incorporated into the semiconductor waveguide,
forming a one-dimensional distributed Bragg
reflector. This grating creates a periodic modulation
of the refractive index along the length of the laser
cavity, allowing it to act as both a wavelength- A DFB laser diode has a grating on top. Positive feedback
selective filter and a high-reflectivity mirror. provided by diffraction of radiation enables single-mode
 The periodicity of the grating determines the emission with narrow bandwidth and high signal-to-noise
ratio. (a) A general layout. A metal stripe is used on top (not
wavelength of the emitted light, and by controlling shown) for current injection. (b) Scanning electron
the grating parameters during fabrication, specific micrograph of a real DFB laser diode showing the lateral
wavelengths can be achieved. metal Bragg grating and the stripe on its top. (c) Lasing
spectrum of a GaAlInSbAs device.
 The DFB laser is ideal for applications requiring
precise wavelength control, such as Sergey V. Gaponenko, Hilmi Volkan Demir. (2019)
telecommunications, optical sensing, spectroscopy, Kogelnik, H., and Shank, C. V. (1971). Appl Phys
and metrology. Lett,18, 152–154.

107
Surface-emitting semiconductor laser
 A surface-emitting semiconductor laser, also known

as a VCSEL (Vertical-Cavity Surface-Emitting Laser), is
a type of semiconductor laser diode where light is
emitted perpendicular to the semiconductor surface.
 Unlike traditional edge-emitting lasers, which emit
light from the edges of the semiconductor chip,
VCSELs emit light from the top surface of the chip.
This vertical emission allows for a highly directional
output beam, simplified integration with optical
systems, and efficient coupling with optical fibers.
 The VCSEL was first proposed and realized by Prof. K.
Iga’s group (Tokyo Institute of Technology) in 1979,
using metal mirrors. Since then, great progress has
been made in research and development, resulting
in a multitude of commercial VCSELs including single
devices, arrays and complicated systems.
 The first commercial VCSEL appeared in 1990 and by
2015, VCSELs ranked second in production volume
among semiconductor lasers. The VCSELs are highly
versatile and widely used in applications such as
data communication, optical sensing, and laser
printing.
Sergey V. Gaponenko, Hilmi Volkan Demir. (2019)

108
Springer Handbook of
Electronic and Photonic
Materials, 2007 (S. Kasap
and P. Capper, Editors)
Quantum dot laser
 A quantum dot laser is a type of semiconductor laser that utilizes quantum
dots as the active gain medium. The advantages of quantum dot laser include
 Low threshold current: Quantum dot lasers typically

exhibit lower threshold currents compared to
conventional quantum well lasers due to the unique
electronic properties of quantum dots, which allow for
efficient carrier confinement and reduced non-radiative
recombination.
 Temperature stability: Quantum dot lasers show
improved temperature stability due to discrete energy
levels in QDs.
 High efficiency: Quantum dot lasers have high efficiency
due to reduced Auger recombination in QDs.
 Wavelength flexibility: Quantum dots offer tunable
emission wavelengths over a wide spectral range by
adjusting the size and composition of the quantum dots. Temperature dependencies of threshold
current for different dimensionalities
 Small linewidth: Quantum dot lasers based on single
calculated by Arakawa and Sakaki (1982).
quantum can achieve narrow linewidths, resulting in high
spectral purity and coherence.
 Broad spectrum: Quantum dot lasers based on QD Arakawa Y. and Sakaki H., Appl Phys
assembly can cover a wide spectrum range due to size Lett, 40, 939–941(1982).
variation of the quantum dots.

110
111
Quantum wire laser

112
Quantum wire laser
ZnO nanowire
Illustration of multiple color laser
Science 2001 June 8; 292: 1897-1899
(A) Emission spectra from nanowire arrays below (line a) and above (line b and inset) the lasing threshold. The pump power for these
spectra are 20, 100, and 150 kW/cm2, respectively. The spectra are offset for easy comparison. (B) Integrated emission intensity from
nanowires as a function of optical pumping energy intensity. (C) Schematic illustration of a nanowire as a resonance cavity with two
naturally faceted hexagonal end faces acting as reflecting mirrors. Stimulated emission from the nanowires was collected in the
direction along the nanowire's end-plane normal (the symmetric axis) with a monochromator (ISA, Edison, New Jersey) combined
with a Peltier-cooled charge-coupled device (EG&G, Gaithersburg, Maryland). The 266-nm pump beam was focused to the nanowire
array at an angle 10° to the end-plane normal. All experiments were carried out at room temperature.

113
Quantum cascade laser
 A quantum cascade laser (QCL) is a type of
semiconductor laser that operates based on the mini-
band formed in a superlattice.
 Unlike traditional semiconductor lasers where light
emission occurs due to electron-hole recombination,
in a QCL, electrons are sequentially transported
through mini-bands of superlattice, releasing energy
in the form of photons at each stage when electrons
at different mini-bands recombine in the active region
(intraband transition).
 The concept of intraband optical gain within a
multiple quantum well system was initially proposed This diagram depicts a quantum cascade laser's
by R. Kazarinov and R. Suris at the Ioffe Institute in conduction band energy. When an external
1971. It was later experimentally demonstrated in voltage is applied, electrons transit from level 2
1994 at Bell Labs by Federico Capasso's group, who to level 1 in the left active region, accelerating in
also introduced the term "quantum cascade lasers" the injector region, and then undergo similar
(Faist et al., 1994). transitions in the right active region.
 The QCL provides powerful coherent radiation in the
mid- to far-infrared range, with operation
Kazarinov, R. F. and Suris, R. A. (1971), Sov Phys
wavelengths controlled by material properties and
Semicond, 5, 707–709.
quantum well width. Typical materials like
AlGaAs/GaAs and GaInAs/AlInAs offer wavelengths Faist, J., Capasso, F., Sivco, D. L., et al. (1994). Quantum
from 2.6 to 14 μm and 70 to 250 μm, respectively. The cascade laser. Science, 264(5158), 553–556.
QLCs have important applications in the
environmental and medical fields.

114
Lasers with photonic crystal
 A photonic crystal is a periodic arrangement of
dielectric or metallic materials that creates a
structure with a spatially modulated refractive
index. It includes 1D, 2D and 3D structures.
 The periodicity results in unique optical properties,
such as the inhibition or enhancement of certain
wavelengths of light, similar to how the atomic
structure of a crystal affects the behavior of
electrons in a semiconductor.
 A photonic crystal laser is a type of laser that
incorporates a photonic crystal structure into its
design. The grating used in DFB laser can be
considered as a 1D photonic crystal.
 Photonic crystal lasers utilize the unique properties A photonic crystal microcavity laser with
of photonic crystals to control the emission optical pumping: the emission spectrum,
wavelength, mode properties, and directionality of output power versus pump power, and
the laser output. They offer advantages such as low the design.
threshold currents, narrow linewidths, and high-
quality output beams.
 The 2D photonic crystal laser was first put forward Painter, O., Lee, R. K., Scherer, A., et al. (1999).
and implemented for a semiconductor quantum Two-dimensional photonic band-gap defect
well laser with optical pumping in 1999 by O. mode laser. Science, 284, 1819–1821.
Painter and colleagues at Caltech, USA.

115
116
Photodetector
Appl. Sci. 2021,
 A photodetector is a device used to detect and 11(2), 501
measure light, typically converting light energy
into an electrical signal. These devices are crucial
in various applications, including
telecommunications, optical communication,
remote sensing, image sensing, biomedical
device, photography, etc.,
 Photodetectors can come in various forms, such
as photodiodes, phototransistors, photomultiplier
tubes, and charge-coupled devices (CCDs).
 They work based on the principle of the
photoelectric effect, where incident photons
(particles of light) release electrons/holes in a
material, generating an electrical current or
voltage proportional to the intensity of the light.
 Photodetectors cover a wide wavelength range
from very long wavelength infrared (> 12 µm) to
UV.
 In this module, we only cover the basics of
semiconductor based-photodetectors.

117
Photodiode
 Photodiode (PD) is a pn junction of a semiconductor material
that produces electron hole pairs in response to light.
Photodiodes can be operated in one of two modes:
photoconductive (reverse bias) or photovoltaic (zero bias).
Mode selection depends upon the application's speed
requirements and the amount of tolerable dark current
(leakage current).
 In the photoconductive mode, a photodiode is operated
under reverse bias voltage. Incident photons generate
electron-hole pairs in the depletion region, causing a drop in
resistance at the diode junction. The generated pairs are
separated and swept to opposite terminals by the electric
field, producing a photocurrent proportional to incident light
intensity.
 Increasing reverse bias leads to a sharp rise in photocurrent Bias voltage
until the breakdown voltage, the maximum reverse bias, is
reached. This mode is suited for high-speed applications like
optical communication (due to wide depletion region and high 𝑒𝑒𝑉𝑉𝑑𝑑
𝑖𝑖 = 𝑖𝑖𝑠𝑠 𝑒𝑒𝑒𝑒𝑒𝑒 − 1 − 𝑖𝑖𝑝𝑝𝑝
electricc field) and precise light intensity measurements due 𝛽𝛽𝑘𝑘𝐵𝐵 𝑇𝑇
to improved linearity. 𝑉𝑉𝑑𝑑 + 𝑉𝑉𝐵𝐵 𝛽𝛽: ideal
𝑖𝑖 = −
 However, the PD in this mode exhibits higher noise levels due 𝑅𝑅 factor
to amplified dark current, impacting the signal-to-noise ratio, 𝑖𝑖𝑝𝑝𝑝 = 𝑁𝑁 � 𝜂𝜂𝑎𝑎𝑎𝑎𝑎𝑎 � 𝑒𝑒
and requires more power due to constant reverse bias.
# 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 # 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 Electron charge
Sensitivity is low because of the low output level, typically in
𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 # 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
the fraction of mA.

118
Photodiode
 In the photovoltaic mode, photodiodes operate

without an external bias voltage. The flow of current
out of the device is restricted and a voltage builds up.
This mode of operation exploits the photovoltaic
effect , which is the basis for solar cells. The amount of
dark current is kept at a minimum when operating in
photovoltaic mode.
 In photovoltaic operation, there's lower noise due to
the absence of an external bias voltage, making it
suitable for low-frequency and low-light applications,
and where a high signal-to-noise ratio is crucial. It also
consumes less power since no external bias is needed.
 However, it has a slower response speed due to
maximum photodiode capacitance without bias, and
its non-linear response can affect precision in light 𝑒𝑒𝑉𝑉𝑑𝑑
𝑖𝑖 = 𝑖𝑖𝑠𝑠 𝑒𝑒𝑒𝑒𝑒𝑒 − 1 − 𝑖𝑖𝑝𝑝𝑝
intensity measurements. Photovoltaic mode is 𝛽𝛽𝑘𝑘𝐵𝐵 𝑇𝑇
common in power generation detectors and 𝑉𝑉𝑑𝑑
𝑖𝑖 = −
applications needing high efficiency, weak signal 𝑅𝑅
𝑖𝑖𝑝𝑝𝑝 = 𝑁𝑁 � 𝜂𝜂𝑎𝑎𝑎𝑎𝑎𝑎 � 𝑒𝑒
detection, and low noise, like wireless sensor
networks and optical communication systems. 𝑒𝑒𝑉𝑉𝑑𝑑
𝑖𝑖𝑝𝑝𝑝 ≫ 𝑖𝑖𝑠𝑠 → 𝑖𝑖 = 𝑖𝑖𝑠𝑠 𝑒𝑒𝑒𝑒𝑒𝑒 − 𝑖𝑖𝑝𝑝𝑝
𝛽𝛽𝑘𝑘𝐵𝐵 𝑇𝑇

119
Characteristics of photodiode
 External quantum efficiency is PD’s capability to convert

photon to electron-hole pair. # 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑖𝑖𝑝𝑝𝑝
𝑠𝑠 𝑒𝑒
 Responsivity indicates the sensitivity of the photodiode 𝜂𝜂 =
𝑃𝑃
to incident light and is defined as the ratio of generated # 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝/𝑠𝑠𝑠𝑠𝑠𝑠 𝑖𝑖𝑖𝑖
ℏ𝜔𝜔
photocurrent to incident optical power. It is often
specified in amps per watt (A/W) or volts per watt (V/W) 𝑖𝑖𝑝𝑝𝑝 A 𝑖𝑖𝑝𝑝𝑝 𝑖𝑖𝑝𝑝𝑝 /𝑒𝑒
depending on whether the output is current or voltage. ℜ= = =
𝑃𝑃𝑖𝑖𝑖𝑖 W 𝐴𝐴𝛷𝛷𝑝𝑝𝑝 ℏ𝜔𝜔 𝐴𝐴𝛷𝛷𝑝𝑝𝑝 ℏ𝜔𝜔/𝑒𝑒
 Detectivity is a figure of merit used to quantify PD’s
ability to detect weak optical signals while considering = 𝜂𝜂
𝑒𝑒
= 𝜂𝜂𝜆𝜆/1.24
A
𝜆𝜆: in unit of 𝜇𝜇𝜇𝜇
ℏ𝜔𝜔 W
the influence of noise. It provides a measure of the
signal-to-noise performance of the photodiode. 𝑃𝑃𝑖𝑖𝑖𝑖 = 𝐴𝐴𝛷𝛷𝑝𝑝𝑝 ℏ𝜔𝜔 (𝛷𝛷𝑝𝑝𝑝 : photons/[𝑚𝑚2 𝑠𝑠1 ], A: area)
 Major noise sources include Johnson noise, shot noise,
1/f noise, background photon noise. 𝑟𝑟𝑟𝑟𝑟𝑟(𝑖𝑖𝑛𝑛 )
𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 (𝑁𝑁𝑁𝑁𝑁𝑁) = (𝑊𝑊)
 Dark current refers to the small current that flows ℜ
through the photodiode in the absence of light. It sets 𝐷𝐷 =
1
(𝑊𝑊 −1 )
the noise floor for PD. 𝑁𝑁𝑁𝑁𝑁𝑁
 Rise time is the time required for the PD to increase its

4𝑘𝑘𝐵𝐵 𝑇𝑇
output from 10% to 90%) of the steady-state value. 𝑟𝑟𝑟𝑟𝑟𝑟 𝑖𝑖𝑛𝑛 = 2𝑒𝑒𝑖𝑖𝑠𝑠 + 2𝑒𝑒𝑖𝑖𝑏𝑏 + 2𝑒𝑒𝑖𝑖𝑑𝑑 + 𝐵𝐵
𝑅𝑅
 Response time refers to the time required for the PD's
output to settle within a certain percentage of its steady- 𝑖𝑖𝑠𝑠 : short noise current
state value after a step change in incident light intensity. 𝑖𝑖𝑏𝑏 : background photon generated current
It includes both rise and fall times. 𝑖𝑖𝑑𝑑 : PD dark current

120
p-i-n photodiode
 A p-i-n photodiode comprises a p-n junction diode

with an undoped intrinsic (i) layer sandwiched
between the p-type and n-type regions.
 The high resistivity of the i-layer causes nearly all
applied bias to drop across it, leading to full
depletion under low reverse bias. The p-i-n diode
has a “controlled” depletion layer width, which can
be tailored to meet the requirements of
photoresponse and frequency bandwidth.
 By material and device structure optimization, p-i-n
PDs with large bandwidths can be achieved. The
response speed and bandwidth are ultimately
limited either by the transit time or by circuit
parameters. The transit time of carriers across the
depletion or i-layer can be reduced by reducing the
thickness of the i-layer.
 For practical use, photoexcitation is achieved via
either an etched opening in the top contact or a
hole in the substrate, which reduces the diode's
active area to match the incident light beam size.
S.O. Kasap (Prentice Hall 1999)

121
Avalanche photodiode (APD)
 The APD is a specialized type of photodiode that
operates under reverse bias conditions in the
avalanche breakdown region of a p-n junction.
 When an electron-hole pair is generated by incident
photons in the depletion region of the APD, the
electric field accelerates the carriers. If the electric
field is sufficiently strong, these carriers gain enough
kinetic energy to ionize other atoms through impact Space
ionization, creating additional electron-hole pairs. charge
This process results in an exponential increase in the
number of charge carriers, leading to avalanche
multiplication and amplification of the photocurrent
(primarily in the p-region).
 The avalanche multiplication process provides
internal gain to the APD, enhancing its sensitivity
compared to conventional photodiodes. This gain Electric
allows APDs to achieve higher signal-to-noise ratios field
and improve detection sensitivity, especially in low-
light conditions.
 Avalanche photodiodes can be sensitive to
temperature variations, which may affect their
performance and stability. Temperature control
measures or temperature-compensated designs may
be employed to mitigate these effects. S.O. Kasap (Prentice Hall 1999)

122
Typical characteristics of different types of PDs

123
Example:
The right figure shows a schematic diagram of a photodiode. The photodiode has an
active length L illuminated by a light beam of optical power P and photon energy ℏ𝜔𝜔. The
diode is operated in reverse bias and a strong electric field across the i-region. Assume
each absorbed photon generates one electron-hole pair and define the quantum
efficiency 𝜂𝜂 as the fraction of these charge carriers that flow into the external circuit.
Assume the absorption coefficient in the active region is 𝛼𝛼.
a) Provide an expression of (i) photocurrent 𝑖𝑖𝑝𝑝𝑝 and (ii) responsivity of the device.
b) Calculate the maximum possible responsivity for a photon of wavelength 620 nm.
c) List a few criteria on the design of practical photodiodes.
Solution:
a) Power absorbed by the active layer: 𝑃𝑃𝑎𝑎𝑎𝑎 = 𝑃𝑃(1 − 𝑒𝑒 −𝛼𝛼𝛼𝛼 )  # of photon per second =𝑃𝑃(1 − 𝑒𝑒 −𝛼𝛼𝛼𝛼 )/ℏ𝜔𝜔
𝑒𝑒𝑃𝑃𝜂𝜂 1−𝑒𝑒 −𝛼𝛼𝛼𝛼 𝑖𝑖𝑝𝑝𝑝
𝑖𝑖𝑝𝑝𝑝 = ℜ= = 𝑒𝑒𝜂𝜂(1 − 𝑒𝑒 −𝛼𝛼𝛼𝛼 )/ℏ𝜔𝜔
ℏ𝜔𝜔 𝑃𝑃
𝑒𝑒 𝜆𝜆 0.62
b) Maximum responsivity is obtained when 𝐿𝐿 → ∝, 𝜂𝜂 = 1, → ℜ𝑚𝑚𝑚𝑚𝑚𝑚 = = = = 0.5 𝐴𝐴/𝑊𝑊
ℏ𝜔𝜔 1.24 1.24
c) The design of practical photodiodes is based on the choice of the semiconductor to optimize the
responsivity while ensuring a fast response and low noise:
(i) Eg < ω
(ii) After satisfying (i), we want Eg to be as large as possible to minimize the dark current.
(iii) We want a material in which electron and hole mobilities are high.
(iv) Materials with direct band gaps are better than those with indirect gaps.
(v) The top contact should be made thin to transmit as much of the light into the i-region as possible

124
Quantum well infrared detector
 The conventional photodiode is based on inter-band optical transitions in

which electrons are excited from the valence band to conduction to generate
photo current. Thus, the upper limit of detectable wavelength is determined
by the bandgap of the material. To detect far infrared light, we need to use
semiconductors with very small bandgap.
 In this regards, technologies based on HgCdTe and InSb have been developed
for IR detection and imaging in the wavelength region of about 3–14 µm.
Nevertheless, quantum well infrared detectors (QWIP) are appealing because
they can be made by more mature materials like GaAs. In addition, it has the
following advantages:
 By adjusting the thickness and composition of the quantum well layers,
the wavelength response of the detector can be tailored to specific
wavelengths of interest.
 Quantum well detectors can have fast response due to the small active
region.
 Quantum well detectors can exhibit lower dark current compared to some
other types of photodetectors, leading to improved signal-to-noise ratio
and detection sensitivity.
 The quantum confinement effect can enhance the temperature stability of
quantum well detectors, making them less susceptible to performance
degradation at elevated temperatures compared to bulk semiconductor
devices.

125
Quantum well infrared detector
(a) Bound-to-Bound (B-B)
Assuming infinite potential well approximation,
The IR radiation absorption due to the
and allowed transition ∆𝑗𝑗 = 1,3, 5, …
intraband transition from the ground state
to the excited state has to be accompanied
by the photoexcited carriers tunneling out of 𝑗𝑗 2 2 𝜋𝜋 2 32 𝜋𝜋 2
𝐸𝐸𝑗𝑗 = 𝐸𝐸2 − 𝐸𝐸1 =
the well. These electrons would be 2𝑚𝑚𝑒𝑒∗ 𝐿𝐿2 2𝑚𝑚𝑒𝑒∗ 𝐿𝐿2
transported by an external electric field in
the continuum above the barriers thereby
producing a photocurrent.
(b) Bound-to-Continuum(B-C)
E2
By making the well width smaller, it can be
E2 Eb1
seen that the excited state can be pushed
Eb2
into the conduction band above the barrier Eb3
∆Ec
and the photocarriers need not tunnel E1
through any barriers before they are E1
E1
collected.
L3 L1 L2
(c) Bound-to-Quasibound (B-QB) (a) Bound-to-Bound (B-B) (b) Bound-to-Continuum (B-C) (c) Bound-to-Quasi-Bound (B-QB)
The excited state is located exactly at the top
of the barrier. The goal is to increase the
energy barrier height Eb (compared to B-C) 𝐿𝐿1 < 𝐿𝐿2 < 𝐿𝐿3 (well width)
as to reduce the thermionic emission while
efficiently collecting the photoexcited 𝐸𝐸𝑏𝑏𝑏 < 𝐸𝐸𝑏𝑏𝑏 < 𝐸𝐸𝑏𝑏𝑏 (energy barrier height)
carriers.

126
Dark current in QWIP
The dark current is the current that flows through a biased detector in the dark (i.e., with no
photons impinging on it). It happens in all the three types of QWIP.
The following is an example of B-QB QWIP.
(3)
(2)
Eb
(1)
(1) Ground state quantum

mechanical tunneling
(2) Thermally assisted tunneling
(3) Thermionic emission

127
Dark current due to
thermionic emission
𝑗𝑗𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑
Dark current characteristics at 77K for a set of n-type QWIPs. Solid

curves are experimental results, while dashed and dash-dotted are
3 calculated using the simple 3D drift model for low field. The three
𝑚𝑚𝑏𝑏 𝑘𝑘𝐵𝐵 𝑇𝑇 2 −𝐸𝐸𝑎𝑎𝑎𝑎𝑎𝑎 samples differ mainly by the doping in the well
𝑗𝑗𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 ≈ 2 exp
2𝜋𝜋ℏ2 𝑘𝑘𝐵𝐵 𝑇𝑇
𝐸𝐸𝑎𝑎𝑎𝑎𝑎𝑎 is the thermal activation energy

which equals the energy difference
between the top of the barrier and the
top of the Fermi sea in the well.
Dark current characteristics at 77K for a set of p-type QWIPs. Solid curves are
experimental results, while dashed and dash-dotted are calculated using the
simple 3D drift model for low field. The three samples differ mainly by the
barrier Al fraction (x) which determines the barrier height.

128
Dark current due to
tunneling At temperatures higher than 70𝐾𝐾 (𝜆𝜆 =
10 𝜇𝜇𝜇𝜇), thermally
excited electrons are thermionically
emitted and transported
above the barriers.
Electrons are thermally

excited and tunnel through the
barriers with assistance
from defects and the triangle part of
the barrier at high bias (T = 40 −
70𝐾𝐾, 𝜆𝜆 = 10 𝜇𝜇𝜇𝜇)
Dominant at low temperature due

DT: direct tunneling to defect-related DT (𝑇𝑇 < 40𝐾𝐾, 𝜆𝜆 =
TAT: thermally assisted tunneling 10 𝜇𝜇𝜇𝜇)
TE: thermionic emission
PC: photo current
J. Appl. Phys., Vol. 93, No. 8, 15 April 2003

129
QWIP based on different optical transitions
Dark current mechanisms

1 Ground state quantum mechanical tunneling
2 Thermally assisted tunneling
3 Thermionic emission
B-QB QWIPs
50 Å
B-B QWIPs
B-C QWIPs

130
QWIP versus HgCdTe photodetector
J. Appl. Phys., Vol. 93, No. 8, 15 April 2003

131
QWIP array
 QWIP can be made into large scale
arrays for infrared imaging applications.
QWIP camera
(QWIP RADIANCETM)
Images taken using our MWIR
InGaAs/InAlAs/InP QWIP 320x256
FPA camera.

132
133
Example of applications of LED and PD
 Photoplethysmograph (PPG) is a non-

invasive medical device used to detect
changes in blood volume in peripheral
blood vessels.
 It typically works by shining a light
source, often a LED, onto the skin and
measuring the variations in light
absorption caused by blood flow.
 The PPGs are commonly used to monitor
heart rate, blood pressure, and oxygen
saturation levels. They are frequently
found in wearable fitness trackers,
medical monitors, and hospital
equipment for continuous patient
monitoring.

134
Optical properties of blood
HbO2
SpO 2 =
 Absorption: Blood absorbs light across a wide HbO2 + Hb
spectrum, particularly in the visible and near-
infrared ranges. This absorption is primarily due
to hemoglobin, the protein in red blood cells
responsible for carrying oxygen. Oxygenated
hemoglobin absorbs more red light, while
deoxygenated hemoglobin absorbs more
infrared light.
 Scattering: Blood contains various components,

including red blood cells, white blood cells,
platelets, and plasma, which can scatter light as
it passes through. This scattering contributes to
the diffuse reflection of light from blood vessels
and tissues.
 Transmission: Blood is partially transparent to

certain wavelengths of light, particularly in the PPG
red and near-infrared regions. This property is
utilized in techniques like pulse oximetry and
near-infrared spectroscopy to measure oxygen
saturation levels and tissue oxygenation.
Transmission Reflection

135
From basic study to applications
Takuo Aoyagi Ear oximeter OLV-5100 Pulse oximeter OLV-1100 Pocket SpO2 monitor Wearable PPG Transmission-type PPG Reflection-type PPG
(1936-2020) (1975) (1988)

136
Recap of topics covered in each lecture
Lattice, Reciprocal Lattice, Brillouin Zone

Energy Band Lattice vibration, Phonons
Lecture 1&2
Single atom Crystals
Metal
V
Semiconductor
nx=3
Insulator
nx=2
Discrete
Energy band
energy levels L
nx=1
Lecture 3 Lecture 3 Lecture 4
Duality of matter Particle in one- Bloch waves, Kronig-Penney Model,

Schrodinger equation dimensional Nearly free electron model, Direct & indirect gap,
Hydrogen model quantum wells Effective mass, Density of States
Tunneling effect
Lecture 6
Lecture 5
5 2 1
Ec 4
Ed 6 3 n

Donor level
ED
 ∑i
v
vd = i =1
E n
Acceptor level
EA - - -
Ea electrons
Ev
Intrinsic semiconductors
Extrinsic semiconductors Basic electrical transport
properties of semiconductors
Lecture 7
Absorption Final
𝐸𝐸𝑓𝑓
state
Lecture 8
ℏ𝜔𝜔
𝐸𝐸𝑖𝑖 Initial
state
Reflection
Transmission
Introduction Absorption
Bulk
Scattered Semiconductor
Incident (Raman, Brillouin)
Luminescence
EM wave
Emission Initial
𝐸𝐸𝑖𝑖
state
𝑘𝑘 ℏ𝜔𝜔
Lecture 9
𝐸𝐸𝑓𝑓 Final
Light-charge interaction state
Recap of topics covered Lecture 10-12
in each lecture
Lecture 7-9
Lecture 8 Lecture 12b: Photodetector
As
Ga
P
Emission Initial
state
Lecture 9
Light-charge interaction
Final
state
Lecture 10-11: Low-D structures
Lecture 12a: Light emitter

Overall summary
Electron in
free space Light in free
space
ℎ2 2
− 𝛻𝛻 𝛹𝛹 𝑥𝑥, 𝑦𝑦, 𝑧𝑧 Si 𝜕𝜕 2 𝐸𝐸
2𝑚𝑚 ∇2 𝐸𝐸 − 𝜀𝜀0 𝜇𝜇0 =0
𝜕𝜕𝑡𝑡 2
= 𝐸𝐸 𝛹𝛹(𝑥𝑥, 𝑦𝑦, 𝑧𝑧) 𝜕𝜕 2 𝐻𝐻
2
∇ 𝐻𝐻 − 𝜀𝜀0 𝜇𝜇0 2 = 0
𝜕𝜕𝑡𝑡
Intrinsic
Sound wave
in continuum
medium Light absorption
K
𝜕𝜕 2 𝑢𝑢 𝐸𝐸 𝜕𝜕 2 𝑢𝑢
= Extrinsic
𝜕𝜕𝑡𝑡 2 𝜌𝜌 𝜕𝜕𝑥𝑥 2 1st BZ
Ed
Ec
ED
Donor level Light emission
Acceptor level
EA
Ea
Ev
Other
modules
Electronic devices

EE5508 Lecture 10-12

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

EE5508 Lecture 10-12

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EE5508 Lecture 10-12

Uploaded by

Copyright:

Available Formats

EE5508 Semiconductor Fundamentals

Department of Electrical and Computer Engineering

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

Lecture 8 Lecture 12b: Photodetector

Lecture 12a: Light emitter

 Semiconductor alloys  Optical properties of semiconductor

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 The most useful form of semiconductors is thin Multilayer

Chemical Vapor Deposition (CVD): This method

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 Lattice mismatch occurs when two

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

0.1𝑎𝑎𝑠𝑠 Just an estimation. Actual case is more

 Defects will be generated when the film thickness exceeds 𝑡𝑡𝑐𝑐 .

Structural scheme of pseudo-morphic (a) and fully relaxed (b)

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

Phys. Rew. Applied 5, 054004

Frank-van der Merwe Volmer-Weber Stranski–Krastanov (S-K)

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 Semiconductor alloys  Optical properties of semiconductor

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 These alloy Material Example Remark

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 Semiconductor alloys  Optical properties of semiconductor

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 For ternary compound semiconductor AxB1-xC , if the constituent compound

Solution: 𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺0.6 𝑃𝑃0.4 = 0.6𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 + 0.4𝑎𝑎𝐺𝐺𝐺𝐺𝐺𝐺 = 5.572 Å

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 Apart from the lattice constant, other material

𝐴𝐴𝐴𝐴𝑥𝑥 𝐺𝐺𝐺𝐺1−𝑥𝑥 𝐴𝐴𝐴𝐴

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 In most alloys, however, there is a bowing effect arising

where b is known as bowing parameter or bowing factor.

 Bowing and Miscibility are directly linked. Alloys with

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

𝐸𝐸𝑔𝑔Γ−Γ = 1.424 + 1.247𝑥𝑥 for 𝑥𝑥 < 0.45

𝐸𝐸𝑔𝑔Γ−𝑋𝑋 = 1.900 + 0.125 𝑥𝑥 + 0.143𝑥𝑥 2

𝐸𝐸𝑔𝑔Γ−𝐿𝐿 = 1.709 + 0.642𝑥𝑥 2

EΓ− L EΓ− Γ EΓ− X

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

0.82𝑚𝑚𝑜𝑜 where 𝑚𝑚𝑜𝑜 is the free electron

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

 Lattice constant for quaternary compounds:

AxB1-xCyD1-y 𝑎𝑎 = 𝑥𝑥𝑥𝑥𝑎𝑎 𝐴𝐴𝐴𝐴 + 𝑥𝑥(1 − 𝑦𝑦)𝑎𝑎 𝐴𝐴𝐴𝐴 + (1 − 𝑥𝑥)𝑦𝑦𝑎𝑎𝐵𝐵𝐵𝐵 + (1 − 𝑥𝑥)(1 − 𝑦𝑦)𝑎𝑎𝐵𝐵𝐵𝐵

ABxCyD1-x-y 𝑎𝑎 = 𝑥𝑥𝑎𝑎 𝐴𝐴𝐴𝐴 + 𝑦𝑦𝑎𝑎 𝐴𝐴𝐴𝐴 + (1 − 𝑥𝑥 − 𝑦𝑦)𝑎𝑎 𝐴𝐴𝐴𝐴

AxByC1-x-yD 𝑎𝑎 = 𝑥𝑥𝑎𝑎 𝐴𝐴𝐴𝐴 + 𝑦𝑦𝑎𝑎𝐵𝐵𝐵𝐵 + (1 − 𝑥𝑥 − 𝑦𝑦)𝑎𝑎𝐶𝐶𝐶𝐶

 Bandgap for quaternary compounds:

𝐸𝐸𝑔𝑔𝐵𝐵𝐵𝐵 𝑏𝑏𝐴𝐴𝐴𝐴𝐴𝐴 𝐸𝐸𝑔𝑔𝐴𝐴𝐶𝐶 1 − 𝑥𝑥 bXYZ : bowing

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

(iii) For In1-xGaxAs1-yPy to be lattice matched to InP,

EE5508 Semiconductor Fundamentals Lecture 10-12 Wu Yihong 09/09/2010

Lattice matched to GaAs

Lattice matched to InP

Keh Yung Cheng