EE5508 Semiconductor Fundamentals

Department of Electrical and Computer Engineering

National University of Singapore

Prof. Wu Yihong

Room: E4-8-03
E-mail: [email protected]

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

1
Optical properties of semiconductors & applications

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
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect Main references for this part:
‒ Mott transition
 Prof. Teo Kie Leong’s past lecture
‒ Mott trailing notes
‒ Burstein-Moss shift  E. Fred Schubert, Light-emitting
‒ Bandgap narrowing diodes(Cambridge University
Press, 2023)
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

3
Semiconductors play crucial roles in optoelectronics / photonics

Outline of Lecture 7

Preamble of electrical conduction in metals

Electrical conduction in semiconductors

Hall effect
 light in water pipe (John Tyndall, 1870) Surface plasmon at metal / dielectric
Piping light patent using mirrors (Wheeling, 1880) interface (1980’s)
Photophone using free space light (Bell, 1880) Surface plasmon in nanostructures
Fiberscope using glass fiber (Brien and Kapany, early 1950’s)) (e.g.,Ebbesen, 1989, 2003)
Optical fiber communication Plasmonics (Atwater, 2000)
(Kao and Hockham, 1966) Metamaterials (Kock,
Low loss fiber ( 1940, Veselago, Stimulated
Maurer et al, 1960, Walser emission
Law of reflection 1999) (Einstein, 1917)
1970)
(Euclid, 300 B.C.)
Maser & Laser (Townes, 1954,
Law of refraction Maiman, 1958)
(Hero, 100 A.D.) p-n junction laser & LED
Working principle of lens (Bosov, 1961, Nasledov, 1962,
Hall, 1962, Nathan, 1962,
(Kepler, 1604)
Holonyak Jr., 1962)
Invention of telescope
(Galileo, 1609) Hetrojunction laser & LED
(Alferov, 1963, Kroemer, 1963)
Snell’s law (1652)
Prism (Newton, 1704) RT & CW DHJ lasers (Hayashi,
Wavelike nature of light 1970, Alferov, 1971)
(Huygens, 1690, Young, 1803
DFB & DBR laser (Nakamura,
Fresnel, 1819, Maxwell, 1873,
1973, Alferov, 1974, Scifres, 1974)
Hertz, 1888)
QW & superlattice (Cho, 1971, Dingle,
Quantum nature of light
1974, Esaki & Tsu, 1970, Dupuis, 1977)
(Plank 1900, Einstein, 1905, Bohr, Heisenberg, etc)
QW laser (Ziel, 1975, Dupuis, 1978, Tsang, 1982 )
Speed of light (299,792.458 km/s)
(Römer, 1676, Huygens, Maxwell, Einstein) QWR & QD (Arakawa, 1982, Alferov, 1996, …… )
Blue LED & LD (Nakamura, 1993, 1995)
Photonic crystal (Eli Yablonovitch, 1987)
Metamaterials (Veselago, 1968, Pendry, 2000) Quantum cascade laser (Faist, 1994 )
Optics / optoelectronics / photonics /
plasmonics …
Reflective / refractive Diffractive Photonic crystal
 Optics: a branch of components components
physics that studies the Optics -1 0 +1 -1 0 +1

behavior, properties, and

interactions of light with
matter, including its
reflection, refraction,
Natural light source Semiconductor Semiconductor
diffraction, and the (sun, incandescent) Thin film Micro/nanostructure
formation of images Optoelec- Gas, liquid, solid,
through lenses and tronics plasma
mirrors.
AlGaAs

 Optoelectronics: a field of AlGaAs

technology focusing on
Light propagation in Light propagation in Plasmonics
generation and detection free space fibers
of light using Photonics /
semiconductors like LED, plasmonics
LD, PD, solar cell, etc.

 Photonics is the science and technology of generating, controlling, and detecting light (photons) for
various applications, while plasmonics involves the study of surface plasmons—collective
oscillations of electrons at the interface between metals and dielectrics—to manipulate light at the
nanoscale for enhanced optical properties.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

6
Light source

 Incandescent light: light

bulbs glow due to
heating of filaments.
 Gas and vapor
discharges: examples of
lighting technologies
using gas or vapor
discharges, like neon or
fluorescent lights.
 Light emitting diode
(LED): light generated
by recombination of
electrons and holes in
semiconductors.
 Color of LED is Light emitting diode made by semiconductors

determined by bandgap
of the semiconductors.
Variety of lighting sources

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

7
Spontaneous versus stimulated emission
Spontaneous emission: Light emitted by atoms or molecules when they randomly drop from an
excited state to a lower energy state without external influence. This light is incoherent, meaning it
has no specific phase relationship or directionality.
Stimulated emission: Light emitted when an incoming photon causes an excited atom or molecule to
drop to a lower energy state, releasing a photon that is coherent with the incoming one. This results
in a focused, coherent beam of light, as seen in lasers.

To achieve lasing, you need:

 Gain medium: A substance that
amplifies light when being
excited.
 Pumping source: A way to excite
the gain medium, like a flashlamp
or electric current. Electron are excited from low
to high energy levels Spontaneous emission  normal light
 Optical cavity: Two mirrors that
reflect light back and forth
through the gain medium.
 Feedback mechanism: The
optical cavity helps amplify the
light with each pass.
 Threshold condition: The
medium needs to be pumped
enough to produce more light
Stimulated emission  laser
than it is absorbed.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

8
Different types of lasers
Semiconductor Lasers:
 Operate based on the properties of
semiconductor materials.
 Commonly used in optical communication, laser
printers, barcode scanners, etc.
 Compact and energy-efficient.
Gas Lasers:
 Active medium is a gas or mixture of gases.
 Examples include helium-neon (HeNe) lasers,
carbon dioxide (CO2) lasers, and argon ion lasers.
 Used in scientific research, medical applications,
and industrial processes.
Solid-State Lasers:
 Active medium is a solid crystalline or glass
material.
 Examples include Nd:YAG lasers, ruby lasers, and
erbium-doped lasers.
 Used in materials processing, medical
procedures, and military applications.
Dye Lasers:
 Active medium is a liquid dye solution.
 Can be tuned to emit light across a wide range of
wavelengths.
 Used in spectroscopy, medical research, and
some scientific applications
Others

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

9
Major milestones of light emitting diode
1927 - Electroluminescence Discovery: Oleg Losev
observes electroluminescence in silicon carbide.
1962 - First Practical LED: Nick Holonyak Jr. develops
the first practical red LED at General Electric.
1972 - Green LED: M. George Craford creates the first
green LED, expanding LED color range.
1989 - High-Brightness Blue LED: Shuji Nakamura
achieves a high-brightness blue LED using
gallium nitride.
1990s - White LEDs: Combination of red, green, and
blue LEDs enables the creation of white light.
2000s - High-Power LEDs: Advances lead to high-power
LEDs for diverse lighting applications.
2014 - Nobel Prize in Physics: Isamu Akasaki, Hiroshi
Amano, and Shuji Nakamura awarded for blue
LED development.
Present - Solid-State Lighting Revolution: LEDs
dominate lighting, transforming residential,
commercial, and street lighting. Injected Light Emission of Silicon Carbide Crystals
K. LEHOVEC, C. A. ACCARDO, AND E. JAMGOCHIAN
Signal Corps Engineering Laboratories, Fort Monmouth, New Jersey
Ongoing - Miniaturization and Efficiency: Continued
research focuses on improving LED efficiency, Phys Rev 83:603–608 (1951)
miniaturization, and cost-effectiveness.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

10
Major milestones of semiconductor laser
1962 - First Semiconductor Laser: Robert N. Hall creates the
inaugural semiconductor laser at General Electric using gallium
arsenide (GaAs).
1963 - Quantum Well Lasers: Quantum well laser first suggested by
H.Kroemer from the U.S. and Kazrinov and Alferov from the Soviet
Union.
1970 - Continuous-Wave Operation: Zhores Alferov and Izuo Hayashi
achieve continuous-wave operation, marking a key advancement in
semiconductor laser technology. Alferov Kroemer
1975 – Quantum well laser Operation: First quantum well laser
The 2000 Nobel Prize in Physics
operation made by J.P. Van der Ziel, R.Dingle, R.C Miller, W.
Wiegmann, and W.A. Nordland, Jr. "for basic work on information and
1980 - Distributed Feedback (DFB) Lasers: Morton Panish and Izuo communication technology, specifically for
Hayashi develop DFB lasers, crucial for stable, single-mode output in developing semiconductor heterostructures
optical communication systems. used in high-speed-and opto-electronics"
1980s - Vertical-Cavity Surface-Emitting Lasers (VCSELs): James
Harris and Kent Choquette contribute to the development of VCSELs,
with applications in optical communication, sensing, and consumer
electronics.
1994 - Quantum cascade lasers : Superlattice-based infrared laser
was developed.
2014 - Nobel Prize in Physics: Isamu Akasaki, Hiroshi Amano, and Nobel prize in physics 2014
Shuji Nakamura receive the Nobel Prize in Physics for their
contributions to efficient blue LEDs and lasers, impacting optical “for the invention of efficient blue light-
emitting diodes which has enabled bright and
storage and lighting.
energy-saving white light sources”

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

11
Key applications of semiconductor lasers

 Optical fiber communications

 Optical data storage
 Pump source for other lasers
 Metrology / instrumentation
 Medical applications
 Sensing (e.g., LIDAR)
 Laser printer
 Laser scanner, etc

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

12
Major milestones of photodetectors
Late 19th Century: Discovery of the photoelectric effect by
Heinrich Hertz.
1905: Albert Einstein explains the photoelectric effect. The Nobel Prize in Physics
1920s: Invention of the photomultiplier tube (PMT) for signal 1921 was awarded to
Albert Einstein "for his
amplification. services to Theoretical
1950s: Emergence of semiconductor-based photodetectors like Physics, and especially for
photodiodes. his discovery of the law of
the photoelectric effect"
1960s: Development of avalanche photodiodes (APDs) for
increased sensitivity.
1970s: Advancements in phototransistors and PIN photodiodes.
1980s: Introduction of charge-coupled devices (CCDs) for imaging.
1990s: Rise of quantum well photodetectors.
2000s: Evolution of quantum dot photodetectors for size-tunable
absorption.
2010s: Integration of 2D materials, like graphene, into
photodetectors. The Nobel Prize in Physics 2009 was divided,
2010s-Present: Advancements in quantum well and quantum dot one half awarded to Charles Kuen Kao "for
photodetectors for various applications. groundbreaking achievements concerning the
transmission of light in fibers for optical
Research Frontiers: communication", the other half jointly to
Willard S. Boyle and George E. Smith "for the
Ongoing research explores novel materials, such as quantum
invention of an imaging semiconductor circuit
dots and 2D materials, to push the performance boundaries - the CCD sensor"
of photodetectors.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

13
Major milestones of infrared photodetectors
Early Infrared detection: The discovery of thermocouples and bolometers in the 19th century,
which were early methods for detecting infrared radiation through thermal effects.
Photoconductors: Development of semiconducting materials like lead sulfide (PbS) and indium
antimonide (InSb) in the mid-20th century, which offered improved sensitivity and response in
the infrared spectrum.
Quantum well infrared photodetectors (QWIPs): Introduction in the 1990s
of QWIPs, which use quantum wells to enhance infrared detection
capabilities.
Infrared focal plane arrays (FPAs):
Advancements in the 2000s, where
large arrays of photodetectors
were developed for imaging
applications, providing high-
resolution infrared images.

P. Martyniuk,
Applied physics
reviews 2014

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

14
Important semiconductors for light sources

Materials System Remarks

AlGaAs • Lattice matched to GaAs.
• Red LED and Laser

InGaAsP • Lattice matched to InP.

• Wide range of emission energies can be accessed
0.8 eV (0.98 µm) to 1.35 eV (1.55 µm).
• Material technology is quite advanced and can be exploited for
communication applications.
GaAsP • With N doping, the LED can operate even if the material is indirect and
green light emission (0.55µm) can be achieved.
• The versatile material can provide red, orange and yellow emission.
AlGaInP • Red, yellow and orange LEDs
• Laser
InGaN, ZnS, ZnSe • Large bandgap materials.
and SiC • Important materials for blue light emission (displays). and applications
for optical storage.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

15
EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010
16
Important semiconductors for detectors

Materials System Remarks

InGaAs • Excellent material for long haul communications (1.55 µm).
• Can be lattice matched to InP.
InGaAsP • Suitable for long haul applications (1.55 µm and 1.3 µm).
• Can be lattice matched to InP.
AlGaSb • Excellent optical properties.
• Can be used for long haul communications (2 µm).
• Suffers from poor substrate availability, since it has to be grown on
GaSb (expensive).
HgCdTe • Excellent material for long wavelength applications in night vision and
thermal imaging (2 to 5 µm)
Si • High performance avalanche photodetectors (1.1 µm)
Ge • High performance avalanche photodetectors. (1.85 µm)
GaN • UV detectors (0.36 µm)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

17
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

18
What is light?
 Light, or visible light, commonly refers to electromagnetic radiation that can be
detected by the human eye.
 Semiconductor optical properties typically deal with light-semiconductor
interactions in the (far) infrared – ultra-violet (UV) range.

 Electromagnetic (EM) waves are governed by Maxwell’s equations

𝜕𝜕𝐵𝐵 𝜕𝜕𝐷𝐷 𝐸𝐸 : electric field vector (in V/m)

𝛻𝛻 × 𝐸𝐸 = − 𝛻𝛻 × 𝐵𝐵 = 𝜇𝜇0 𝐽𝐽⃗ + 𝜇𝜇0 𝐻𝐻 : magnetic field vector (in A/m)
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕

𝛻𝛻 ⋅ 𝐷𝐷 = 𝜌𝜌 𝛻𝛻 ⋅ 𝐵𝐵 = 0 𝐷𝐷 : electric displacement vector (in C/m2)

𝐵𝐵 : magnetic induction (in Wb/m2)
𝐷𝐷 = 𝜀𝜀0 𝐸𝐸 + 𝑃𝑃⃗ 𝐵𝐵 = 𝜇𝜇0 (𝐻𝐻 + 𝑀𝑀) 𝑃𝑃⃗ : electric polarization (in C/m2)
𝑀𝑀 : magnetic polarization (in Wb/m2 or emu/m3)
𝑃𝑃⃗ = 𝜀𝜀0 𝜒𝜒𝐸𝐸 (linear optics), 𝐽𝐽⃗ = 𝜎𝜎𝐸𝐸, 𝛻𝛻 � 𝐽𝐽⃗ = −
𝜕𝜕𝜕𝜕 𝐽𝐽⃗ : current density (in A/m2)
𝜕𝜕𝜕𝜕 𝜀𝜀0 : permittivity of a vacuum (8.854x10-12F/m)
𝜇𝜇0 : permeability of a vacuum (4px10-7H/m)
𝐷𝐷 = 𝜀𝜀0 𝐸𝐸 + 𝜀𝜀0 𝜒𝜒𝐸𝐸 = 𝜀𝜀0 1 + 𝜒𝜒 𝐸𝐸 = 𝜀𝜀0 𝜀𝜀𝑟𝑟 𝐸𝐸
𝜒𝜒 : electric susceptibility (dimensionless)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

19
EM wave in free space
 In free space, 𝐽𝐽⃗ = 0, 𝑀𝑀 = 0, 𝑃𝑃⃗ = 0

𝜕𝜕𝐵𝐵 𝜕𝜕𝐸𝐸 Energy flux density

∇ × 𝐸𝐸 = − ∇ × 𝐵𝐵 = 𝜀𝜀0 𝜇𝜇0 ⃗
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 (Poynting vector 𝑆𝑆)
∇ ⋅ 𝐸𝐸 = 0 ∇ ⋅ 𝐵𝐵 = 0
𝑆𝑆⃗ = 𝐸𝐸 × 𝐻𝐻 𝑆𝑆(𝑥𝑥, 𝑡𝑡) = 𝑐𝑐𝜀𝜀0 𝐸𝐸02 𝑐𝑐𝑐𝑐𝑐𝑐 2 (𝑘𝑘𝑘𝑘 − 𝜔𝜔𝜔𝜔)
𝐷𝐷 = 𝜀𝜀0 𝐸𝐸 𝐵𝐵 = 𝜇𝜇0 𝐻𝐻
Intensity or average 1 1
𝐼𝐼 =< 𝑆𝑆 >= 2 𝑐𝑐𝜀𝜀0 𝐸𝐸02 = 2 𝑐𝑐𝜇𝜇0 𝐻𝐻02 (unit: W/m2)
energy flux density

𝜕𝜕 2 𝐸𝐸 𝐸𝐸0
2
∇ 𝐸𝐸 − 𝜀𝜀0 𝜇𝜇0 2 = 0 ∇ ⋅ 𝐸𝐸 = 0 ⇒ 𝑒𝑒1̑ ⋅ 𝑘𝑘 = 0 𝐻𝐻0 =
𝜕𝜕𝑡𝑡 𝑍𝑍0
∇ ⋅ 𝐻𝐻 = 0 ⇒ 𝑒𝑒̑2 ⋅ 𝑘𝑘 = 0
𝜕𝜕 2 𝐻𝐻 𝑍𝑍0 = 𝜇𝜇0 /𝜀𝜀0 = 377 𝛺𝛺
∇2 𝐻𝐻 − 𝜀𝜀0 𝜇𝜇0 =0 𝜕𝜕𝐻𝐻 𝑘𝑘 × 𝑒𝑒̑1
𝜕𝜕𝑡𝑡 2 ∇ × 𝐸𝐸 = −𝜇𝜇0 ⇒ 𝑒𝑒̑2 = (impedance of free space)
𝜕𝜕𝜕𝜕 𝑘𝑘

𝑒𝑒̂1
⃗
𝐸𝐸 = 𝑒𝑒̑1 𝐸𝐸0 𝑒𝑒 𝑖𝑖(𝑘𝑘⋅𝑟𝑟−𝜔𝜔𝜔𝜔)
𝐸𝐸
⃗
𝑖𝑖(𝑘𝑘⋅𝑟𝑟−𝜔𝜔𝜔𝜔)
𝐻𝐻 = 𝑒𝑒̑2 𝐻𝐻0 𝑒𝑒  EM waves are transverse waves.
1 𝑒𝑒̂2 𝑒𝑒̂1 , 𝑒𝑒̂2 , 𝑘𝑘� are perpendicular to each
𝑐𝑐 = other and are related by right-
𝜀𝜀0 𝜇𝜇0 𝐵𝐵
handed rule.
𝜔𝜔
𝑘𝑘 = 𝑘𝑘�
𝑐𝑐

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

20
EM wave in free space
Example: Comparison:
A laser beam has an intensity of 1
Electric field in a copper wire of cross-sectional
mW/m2 . Assuming that the beam is
area 1𝑚𝑚𝑚𝑚2 and resistivity 1.7×10−8Ωm, carrying a
composed of plane waves, calculate current of 1 A:
the amplitude of the electric and
𝑗𝑗 𝐼𝐼 1
magnetic fields in the beam. 𝐸𝐸 = = 𝜌𝜌 = 1.7 × 10−8 = 0.017𝑉𝑉/𝑚𝑚
𝜎𝜎 𝑆𝑆 1 × 10−6
Solution:
Magnetic field generated by a wire carrying a 1 A
From previous slide, we have current at a distance of 1 cm:
𝜇𝜇0 𝐼𝐼 1𝐴𝐴
1 1 𝐵𝐵 = = = 2 × 10−5 𝑇𝑇
𝐼𝐼 = 𝑐𝑐𝜀𝜀0 𝐸𝐸02 = 𝑐𝑐𝜇𝜇0 𝐻𝐻02 2𝜋𝜋𝜋𝜋 2 × 3.14 × 0.01𝑚𝑚
2 2

2 2
𝐸𝐸0 = 𝐼𝐼 = × 1.0 × 10−3 W/m2 = 0.87 V/m
𝑐𝑐𝜀𝜀0 (3 × 108 m/s)(8.85 × 10−12 F/m)

𝜇𝜇0 𝐸𝐸0 𝐸𝐸0

𝐵𝐵0 = = = 2.9 nT
𝜇𝜇0 /𝜀𝜀0 𝑐𝑐

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

21
EM wave in lossless and lossy medium
 In a medium with 𝑃𝑃⃗ ≠ 0, 𝑀𝑀 = 0, 𝜌𝜌 = 0 , we have

2
𝜕𝜕 2 𝐸𝐸 𝜕𝜕 2 𝑃𝑃⃗ 𝜕𝜕⃗𝚥𝚥
𝛻𝛻 𝐸𝐸 − 𝜀𝜀0 𝜇𝜇0 2 = 𝜀𝜀0 2 + 𝜇𝜇0
𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡

When 𝚥𝚥⃗ = 0 When 𝚥𝚥⃗ ≠ 0

(or 𝜎𝜎 = 0, lossless medium) (or 𝜎𝜎 ≠ 0, lossy medium)

2
𝜕𝜕 2 𝐸𝐸 𝜕𝜕 2 𝑃𝑃⃗ 𝜕𝜕 2 𝐸𝐸 𝜕𝜕𝐸𝐸
∇ 𝐸𝐸 − 𝜀𝜀0 𝜇𝜇0 2 = 𝜇𝜇0 2 2
𝛻𝛻 𝐸𝐸 − 𝜀𝜀0 𝜀𝜀𝑟𝑟 𝜇𝜇0 2 = 𝜇𝜇0 𝜎𝜎
𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡 𝜕𝜕𝜕𝜕
𝑃𝑃⃗ = 𝜀𝜀0 𝜒𝜒𝐸𝐸
𝜀𝜀𝑟𝑟 = 1 + 𝜒𝜒 Find solution in the form of
𝜕𝜕 2 𝐸𝐸 𝐸𝐸 = 𝐸𝐸0 𝑒𝑒 𝑖𝑖(𝐾𝐾𝐾𝐾−𝜔𝜔𝜔𝜔)
2
∇ 𝐸𝐸 − 𝜀𝜀0 𝜀𝜀𝑟𝑟 𝜇𝜇0 2 = 0
𝜕𝜕𝑡𝑡
𝐾𝐾 2 = 𝜀𝜀𝑟𝑟 𝜀𝜀0 𝜇𝜇0 𝜔𝜔2 + 𝑖𝑖𝜇𝜇0 𝜎𝜎𝜎𝜎 = 𝑛𝑛� 2 𝑘𝑘02

𝐸𝐸 = 𝐸𝐸0 𝑒𝑒 𝑖𝑖(𝑘𝑘𝑘𝑘−𝜔𝜔𝜔𝜔) in 1D case 𝐾𝐾 = 𝜀𝜀𝑟𝑟 𝜀𝜀0 𝜇𝜇0 𝜔𝜔 2 + 𝑖𝑖𝜇𝜇0 𝜎𝜎𝜎𝜎

𝑘𝑘 = 𝜔𝜔 𝜀𝜀𝑟𝑟 𝜀𝜀0 𝜇𝜇0 (wavevector)
𝐾𝐾: complex wavevector
2 2
𝑘𝑘 = 𝜀𝜀𝑟𝑟 𝜀𝜀0 𝜇𝜇0 𝜔𝜔 = 𝑛𝑛2 𝑘𝑘02 , 𝑘𝑘0 = 𝜔𝜔 𝜀𝜀0 𝜇𝜇0 � complex refractive index
𝑛𝑛:

𝜀𝜀𝑟𝑟 : permittivity or relative dielectric constant

𝑛𝑛 = 𝜀𝜀𝑟𝑟 : refractive index

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

22
EM wave in lossy medium: absorption
𝐾𝐾 2 = 𝜀𝜀𝑟𝑟 𝜀𝜀0 𝜇𝜇0 𝜔𝜔2 + 𝑖𝑖𝜇𝜇0 𝜎𝜎𝜎𝜎 = 𝑛𝑛� 2 𝑘𝑘02
1/2 1/2
𝜀𝜀1 + 𝜀𝜀12 + 𝜀𝜀22
𝑛𝑛 =
2
𝜀𝜀𝑟𝑟 𝜀𝜀0 𝜇𝜇0 𝜔𝜔2 + 𝑖𝑖𝜇𝜇0 𝜎𝜎𝜎𝜎 𝜎𝜎 1/2 1/2
2 −𝜀𝜀1 + 𝜀𝜀12 + 𝜀𝜀22
𝑛𝑛� = = 𝜀𝜀𝑟𝑟 + 𝑖𝑖 𝜅𝜅 =
𝑘𝑘02 𝜀𝜀0 𝜔𝜔 2

Let 𝜀𝜀�𝑟𝑟 𝜔𝜔, 𝑘𝑘 = 𝜀𝜀1 𝜔𝜔, 𝑘𝑘 + 𝑖𝑖𝜀𝜀2 (𝜔𝜔, 𝑘𝑘) 𝐾𝐾 = 𝑛𝑛� 𝜔𝜔 𝑘𝑘0 = (𝑛𝑛 𝜔𝜔 + 𝑖𝑖𝜅𝜅 𝜔𝜔 )𝑘𝑘0
𝑛𝑛� 𝜔𝜔 = 𝑛𝑛 𝜔𝜔 + 𝑖𝑖𝜅𝜅 𝜔𝜔 𝑛𝑛𝜔𝜔 𝜅𝜅𝜔𝜔
𝑖𝑖 𝑥𝑥−𝜔𝜔𝜔𝜔
𝐸𝐸 = 𝐸𝐸0 𝑒𝑒 𝑖𝑖(𝐾𝐾𝐾𝐾−𝜔𝜔𝜔𝜔) = 𝐸𝐸0 𝑒𝑒 𝑐𝑐 𝑒𝑒 − 𝑐𝑐
𝑥𝑥
𝑛𝑛� 2 = 𝜀𝜀�𝑟𝑟
2𝜅𝜅𝜔𝜔
𝐼𝐼 = 𝐸𝐸 2 = 𝐼𝐼0 𝑒𝑒 − 𝑐𝑐
𝑥𝑥
= 𝐼𝐼0 𝑒𝑒 −𝛼𝛼𝑥𝑥
𝜀𝜀1 𝜔𝜔 = 𝑛𝑛2 𝜔𝜔 − 𝜅𝜅 2 (𝜔𝜔)
2𝜅𝜅𝜅𝜅 4𝜋𝜋𝜅𝜅 𝜎𝜎 𝛼𝛼 : absorption
𝜀𝜀2 𝜔𝜔 = 2𝑛𝑛 𝜔𝜔 𝜅𝜅(𝜔𝜔) 𝛼𝛼 = = = = 𝜎𝜎𝑍𝑍0
𝑐𝑐 𝜆𝜆 𝜀𝜀0 𝑐𝑐 coefficient

𝜕𝜕𝐷𝐷 Let 𝐸𝐸 = 𝐸𝐸0 exp (𝑖𝑖𝜔𝜔𝜔𝜔) Non-dissipative

𝛻𝛻 × 𝐵𝐵 = 𝜇𝜇0 𝐽𝐽⃗ + 𝜇𝜇0
𝜕𝜕𝜕𝜕
dissipative
𝐷𝐷 = 𝜀𝜀0 1 + 𝜒𝜒 𝐸𝐸 = 𝜀𝜀0 𝜀𝜀𝑟𝑟 𝐸𝐸 𝜕𝜕𝐷𝐷 𝜕𝜕𝐸𝐸
𝐽𝐽⃗𝐷𝐷 = = 𝜀𝜀0 𝜀𝜀𝑟𝑟 = 𝑖𝑖𝜀𝜀0 𝜀𝜀𝑟𝑟 𝜔𝜔𝐸𝐸 = 𝑖𝑖𝜎𝜎𝐷𝐷 𝐸𝐸
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
𝜕𝜕𝐷𝐷 𝜕𝜕𝐸𝐸 𝜀𝜀𝑟𝑟 𝜀𝜀0 𝜇𝜇0 𝜔𝜔2 + 𝑖𝑖𝜇𝜇0 𝜎𝜎𝜎𝜎 𝜎𝜎 𝜎𝜎𝐷𝐷 𝜎𝜎
𝐽𝐽⃗𝐷𝐷 = = 𝜀𝜀0 𝜀𝜀𝑟𝑟 2
𝑛𝑛� = = 𝜀𝜀 + 𝑖𝑖 = + 𝑖𝑖
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕 𝑘𝑘02
𝑟𝑟
𝜀𝜀0 𝜔𝜔 𝜀𝜀0 𝜔𝜔 𝜀𝜀0 𝜔𝜔

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

23
Why is absorption related to conductivity?

 In materials with finite conductivity, there are free charge

carriers (like electrons in metals or doped semiconductors).
When an EM wave passes through such a material, the electric
field of the wave causes these free charges to oscillate.
 As the charges oscillate, they collide with atoms and other
particles in the material, converting part of the wave’s energy
into heat. This energy conversion manifests as absorption of
the EM wave.
 The loss tangent (tan𝛿𝛿) characterizes the ratio of energy lost to
energy stored in the material:
𝜎𝜎 𝜎𝜎
tan 𝛿𝛿 = =
𝜎𝜎𝐷𝐷 𝜀𝜀0 𝜀𝜀𝑟𝑟 𝜔𝜔
 A higher loss tangent indicates more significant absorption due
to conductivity.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

24
Phenomenological model of light absorption

 If a light beam is propagating in the 𝑥𝑥 direction,

and the intensity (optical power per unit area or
also known as optical power density) at position
𝑥𝑥 is I(𝑥𝑥), then the decrease of the intensity in
an incremental thickness d𝑥𝑥 is given by:

𝑑𝑑𝑑𝑑 = −𝛼𝛼𝛼𝛼𝛼𝛼 × 𝐼𝐼(𝑥𝑥) 𝛼𝛼: absorption coefficient

𝐼𝐼𝑜𝑜

 The intensity of light at position 𝑥𝑥 is given by

−𝛼𝛼𝛼𝛼 I ( x) = I 0 e −αx
𝐼𝐼(𝑥𝑥) = 𝐼𝐼𝑜𝑜 𝑒𝑒 𝐼𝐼𝑜𝑜 = 𝐼𝐼(0)

Beer-Lambert Law or Beer’s Law 𝑥𝑥

 After the light passes through a medium with
thickness 𝑙𝑙, its intensity is given by
𝐼𝐼0 𝐼𝐼 = 𝐼𝐼0 𝑒𝑒 −𝛼𝛼𝒍𝒍
𝐼𝐼 = 𝐼𝐼𝑜𝑜 𝑒𝑒 −𝛼𝛼𝛼𝛼
𝑙𝑙

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

25
Absorption coefficient v.s. penetration depth

 The penetration depth is defined as the

distance in which the light intensity
decays by a factor of 1/𝑒𝑒.

𝐼𝐼 = 𝐼𝐼0 𝑒𝑒 −𝛼𝛼𝛼𝛼

𝑒𝑒 −1 𝐼𝐼0 = 𝐼𝐼0 𝑒𝑒 −𝛼𝛼𝛼𝛼

1
𝑑𝑑 =
𝛼𝛼

where d is the penetration depth. Distance (µm)

The light intensity versus distance for a few

typical examples of absorption behavior

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

26
 Example:

Solution:
IT
𝐼𝐼 = 𝐼𝐼0 𝑒𝑒 −𝛼𝛼𝛼𝛼

𝐼𝐼𝑇𝑇 = 𝐼𝐼0 𝑒𝑒 −𝛼𝛼𝐿𝐿

Since 95% of light has been absorbed, 𝐼𝐼𝑇𝑇 = 0.05𝐼𝐼0 x=0 x=L

1 2.996
𝐿𝐿 = − 𝑙𝑙𝑙𝑙 0.05 =
𝛼𝛼 2.5 × 104 𝑐𝑐𝑐𝑐−1
= 1.198 × 10−4 𝑐𝑐𝑐𝑐 = 1.2𝜇𝜇𝜇𝜇

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

27
Reflection/transmission at interface (𝒏𝒏𝟏𝟏 / 𝒏𝒏𝟐𝟐 )
 The reflectance/transmittance of light at dielectric interface depends on light
polarization:
 S-polarized (TE) waves, or transverse electric waves, have the electric field perpendicular
to the plane of incidence (parallel to the interface surface).
 P-polarized (TM) waves, or transverse magnetic waves, have the electric field parallel to
the plane of incidence (perpendicular to the interface surface).

 The reflection and transmission of EM waves at an interface are governed by

Maxwell’s boundary conditions:
 Tangential components of the electric field and magnetic field must be continuous
across the interface.
 For S-polarized waves, the electric field is tangential to the interface.
 For P-polarized waves, the magnetic field is tangential to the interface.

𝑛𝑛1 𝑛𝑛2 𝑛𝑛1 𝑛𝑛2 s wave p wave

𝐻𝐻1𝑟𝑟 𝐸𝐸2
𝐸𝐸2 𝐻𝐻1𝑟𝑟
𝐸𝐸1𝑟𝑟 𝑛𝑛1 cos 𝜃𝜃1 − 𝑛𝑛2 cos 𝜃𝜃2 𝑛𝑛1 cos 𝜃𝜃2 − 𝑛𝑛2 cos 𝜃𝜃1
𝐻𝐻2 𝑟𝑟⊥ = 𝑟𝑟∥ =
𝐸𝐸1𝑟𝑟 𝜃𝜃2 𝐻𝐻2 𝜃𝜃2 𝑛𝑛1 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃1 + 𝑛𝑛2 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃2 𝑛𝑛1 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃2 + 𝑛𝑛2 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃1
𝜃𝜃1 𝜃𝜃1
𝐸𝐸1 𝐸𝐸1 2𝑛𝑛1 cos 𝜃𝜃1 2𝑛𝑛1 cos 𝜃𝜃1
s wave p wave 𝑡𝑡⊥ = 𝑡𝑡∥ =
𝐻𝐻1 𝑛𝑛1 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃1 + 𝑛𝑛2 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃2 𝑛𝑛1 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃2 + 𝑛𝑛2 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃1
𝐻𝐻1

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

28
Reflection/transmission at interface (𝒏𝒏𝟏𝟏 / 𝒏𝒏𝟐𝟐 )
 For S-polarized waves, when the angle of incidence is smaller than the critical angle
(for total internal reflection), the reflection is generally stronger compared to P-
polarized waves. The reflectance of S-waves approaches 100% more quickly.
 P-polarized waves, even above the critical angle, can experience some partial
transmission as an evanescent wave near the interface, altering the reflectance
behavior.

Reflectance of the GaAs/vacuum

interface (close to the band gap, n =
3.347) for radiation from vacuum/air
(left panel) and from the GaAs (right
panel), respectively, as a function of
incidence angle and polarization

 For P-polarized waves, there is a special angle of incidence called Brewster’s angle
(𝜃𝜃𝐵𝐵 ​ ), where no reflection occurs (reflected and the diffracted light are perpendicular
𝑛𝑛
to each other): tan 𝜃𝜃𝐵𝐵 = 2
𝑛𝑛1

 For normal incidence: 𝑛𝑛𝐼𝐼 − 𝑛𝑛𝐼𝐼𝐼𝐼 2𝑛𝑛𝐼𝐼 𝑛𝑛� − 1 2

𝑛𝑛 − 1 2
+ 𝜅𝜅 2
𝑟𝑟∥,⊥ 𝛼𝛼𝑖𝑖 = 0 = 𝑡𝑡∥,⊥ 𝛼𝛼𝑖𝑖 = 0 = 𝑅𝑅 = = 2 + 𝜅𝜅 2
𝑛𝑛𝐼𝐼 + 𝑛𝑛𝐼𝐼𝐼𝐼 𝑛𝑛𝐼𝐼 + 𝑛𝑛𝐼𝐼𝐼𝐼 𝑛𝑛� + 1 𝑛𝑛 + 1

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

29
Reflection/transmission of thin slab
Multiple reflection in a thin slab of medium

𝐼𝐼0
 Transmission coefficient is
−𝛼𝛼𝛼𝛼
𝐼𝐼0 (1 − 𝑅𝑅1 )(1 − 𝑅𝑅2 )𝑒𝑒
𝐼𝐼𝑇𝑇 (1−𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝛼𝛼
𝐼𝐼0 𝑅𝑅1 𝑇𝑇 = =
𝐼𝐼0 1−𝑅𝑅 2 𝑒𝑒 −2𝛼𝛼𝛼𝛼
𝐼𝐼0 (1 − 𝑅𝑅1 )2 𝑅𝑅2 𝑒𝑒 −2𝛼𝛼𝛼𝛼
When αl is large, i.e. αl >> 1

Let 𝑅𝑅1 = 𝑅𝑅2 = 𝑅𝑅

𝑇𝑇 ≈ (1 − 𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝛼𝛼
1
𝐼𝐼𝑇𝑇 = 𝐼𝐼0 (1 − 𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝛼𝛼
1 − 𝑅𝑅 2 𝑒𝑒 −2𝛼𝛼𝛼𝛼

Example:
A measurement is performed to determine the transmission coefficient T at a photon wavelength of 0.89 µm, of
two GaAs samples S1 and S2, with thickness l1 = 250 µm and l2 = 350 µm, respectively. The transmission
coefficients are found to be T1 = 1.15×10-3 and T2 = 2.31×10-4, respectively. Determine the absorption
coefficient α and the reflection coefficient R at the same wavelength (assume αl >> 1).
Solution:
𝑇𝑇1 (1−𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝑙𝑙1 1.15×10−3
= = 𝑒𝑒 −𝛼𝛼(0.025−0.035) 𝛼𝛼 = 1.61 × 102 𝑐𝑐𝑐𝑐-1
𝑇𝑇2 (1−𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝑙𝑙2 2.31×10−4

−2 ×0.025
𝑇𝑇1 = (1 − 𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝑙𝑙1 1.15 × 10−3 =(1 − 𝑅𝑅)2 𝑒𝑒 −1.61×10 𝑅𝑅 = 0.75

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

30
 Example:
For sapphire, the absorption coefficient is very small
within the transparency range. Thus, we take complex
refractive index to be real with no imaginary component (n
= 1.77). Calculate the reflectivity and transmission
coefficient for sapphire in the transparency range.

Solution:

2 2
𝑛𝑛� − 1 𝑛𝑛 − 1 + 𝜅𝜅 2
𝑅𝑅 = = 2 + 𝜅𝜅 2
𝑛𝑛� + 1 𝑛𝑛 + 1

α→ 0, and substitute n = 1.77, gives R = 0.077 (note: α= 2ωκ/c); κ = 0

𝐼𝐼𝑇𝑇 (1 − 𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝛼𝛼

𝑇𝑇 = = (Don’t need to know l since αl = 0)
𝐼𝐼0 1 − 𝑅𝑅2 𝑒𝑒 −2𝛼𝛼𝛼𝛼

α→ 0, substitute R = 0.077 gives T = 0.85

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

31
Dispersion and absorption: Kramers –
Kronig relations

 A dispersive medium is a material or substance through which different wavelengths of light or

other electromagnetic waves travel at different speeds. In other words, dispersion refers to the
phenomenon where the velocity of a wave depends on its frequency or wavelength. It is
characterized by the real part of the refractive index, electric susceptibility or permittivity.

 An absorptive medium is a material that absorbs or takes in energy, often in the form of waves
such as light, sound, or electromagnetic waves. When waves interact with an absorptive
medium, the energy of the waves is typically converted into other forms, such as heat. It is
characterized by the imaginary part of the refractive index, electric susceptibility or
permittivity.

 The real and imaginary part of the electric susceptibility 𝜒𝜒 = 𝜒𝜒1 + i𝜒𝜒2 are related by the
Kramers –Kronig relations:

∞ ′ ∞
2 𝜔𝜔 𝜒𝜒2 (𝜔𝜔′ ) ′ 2 𝜔𝜔𝜒𝜒1 (𝜔𝜔′ ) ℘ denotes the Cauchy
𝜒𝜒1 (𝜔𝜔) = ℘ � ′ 2 − 𝜔𝜔 2
𝑑𝑑𝜔𝜔 𝜒𝜒2 (𝜔𝜔) = ℘ � 2 𝑑𝑑𝜔𝜔′ principal value
𝜋𝜋 0 𝜔𝜔 𝜋𝜋 2
0 𝜔𝜔 − 𝜔𝜔
′

 Similarly, the following relations also hold for permittivity:

∞ ′
2 𝜔𝜔 𝜖𝜖2 (𝜔𝜔′ ) ′
2𝜔𝜔 ∞
𝜖𝜖1 𝜔𝜔′ − 1 ′
𝜀𝜀1 (𝜔𝜔) − 1 = ℘ � 2 𝑑𝑑𝜔𝜔 𝜖𝜖2 (𝜔𝜔) = ℘� 𝑑𝑑𝜔𝜔
𝜋𝜋 ′
0 𝜔𝜔 − 𝜔𝜔
2 𝜋𝜋 2 ′2
0 𝜔𝜔 − 𝜔𝜔

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

32
Dispersion and absorption: Kramers –
Kronig relations

 The key to understanding the connection between dispersion and absorption lies in the
causality principle: a material cannot respond to an electromagnetic wave before the wave
interacts with it. This constraint imposes a mathematical relationship between how a material
disperses light (changes its speed) and how it absorbs light (converts wave energy into heat).
 When a material absorbs strongly at certain frequencies (large imaginary part), this absorption
will cause changes in the phase velocity (large real part variations) near those frequencies.
 Likewise, when dispersion (real part) varies rapidly, it indicates that there must be strong
absorption (imaginary part) somewhere nearby in frequency.
 For example:
 Resonances: Near a material's natural resonance frequency (like in glass or water), the
material strongly absorbs energy from light. Around this same frequency, the refractive
index also changes rapidly, showing a large dispersion.
 Transparency: In regions where a material is transparent (i.e., very little absorption), the
refractive index changes more slowly with frequency, indicating less dispersion.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

33
Examples of complex permittivity in
semiconductors
Peter Y. Yu &
 Permittivity of a solid- Manuel Cardona 2003
state materials is
determined by its band
structure and its
interaction with light

A comparison between the

experimental and calculated
imaginary part of the
dielectric function in Ge. The
theoretical curve was
calculated from a band
structure obtained
by the k · p method without
spin-orbit interaction

Band structure of Ge

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

34
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

35
Typical absorption spectrum of semiconductor

Marius Grundmann (Springer, 2021)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

36
Typical absorption spectrum of semiconductor

Photon Energy Absorption Type Description

Range
10⁻³ to 10⁻² eV Free Carriers Absorption by free electrons or holes in the conduction
or valence bands.
10⁻² to 10⁻¹ eV Donor to Electrons transition from donor impurity states to the
Conduction Band conduction band.
10⁻² to 10⁻¹ eV Valence Band to Electrons transition from the valence band to acceptor
Acceptor impurity states.
10⁻² to 10⁻¹ eV Optical Phonons Absorption of photons by the lattice, causing vibrations
(phonons) rather than electronic transitions.
~1 eV (near Excitons Formation of bound electron-hole pairs just below the
bandgap) conduction band.
1 to 10 eV Valence to Direct transitions of electrons from the valence band to
Conduction Band the conduction band, crossing the bandgap.
1 to 10 eV Acceptor to Electrons transition from acceptor impurity states
Conduction band directly into the conduction band.
1 to 10 eV Valence Band to Electrons transition from the valence band into donor
Donor impurity states near the conduction band.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

37
Basics of light-matter interaction

 Light consists of oscillating electric and magnetic fields. These fields exert forces on
charged particles, primarily electrons in matter, causing them to move or oscillate.
 The movement of these charged particles in response to the oscillating electric field
of light is the basis of how light is absorbed, reflected, or transmitted through
materials.
 Therefore, understanding light-charge interaction is crucial for grasping light-matter
interaction because it is the fundamental process that governs how electromagnetic
waves (light) interact with the particles in a material, which are mostly composed of
charged entities like electrons and ions.

Reflected
Semiconductor
Transmitted
(absorbed)
Scattered
Incident (Raman, Brillouin)
Luminescence

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

38
Classical model of light-charge interaction: Lorentz
oscillator model

 A simple model to describe the optical response of bound charges is the

Lorentz oscillator model.
 The model assumes that electrons in a material are bound to their
respective nuclei by springs and that they experience a restoring force
when displaced from their equilibrium positions.
 In the presence of an external oscillating electric field (e.g., an incident
light wave), the electrons are forced to oscillate around their equilibrium
positions, which can be described by a classical harmonic oscillator.

𝑑𝑑 2 𝑥𝑥 𝑑𝑑𝑑𝑑
𝑚𝑚 2 + 𝑚𝑚𝛾𝛾 + 𝑚𝑚𝜔𝜔𝑜𝑜2 𝑥𝑥 = (−𝑒𝑒)𝐸𝐸𝑥𝑥 (𝑡𝑡)
𝑑𝑑𝑡𝑡 𝑑𝑑𝑑𝑑

Electron Damping force Restoring force Driving force

mass (𝛾𝛾: damping (𝜔𝜔0 = 𝑘𝑘/𝑚𝑚: From electric
constant) resonance field
frequency) 𝑘𝑘

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

39
Lorentz oscillator model
 Let 𝐸𝐸𝑥𝑥 = 𝑅𝑅𝑅𝑅 𝐸𝐸𝑥𝑥𝑥 𝑒𝑒 𝑖𝑖𝑖𝑖𝑖𝑖 and 𝑥𝑥 = 𝑅𝑅𝑅𝑅 𝑥𝑥0 𝑒𝑒 𝑖𝑖𝑖𝑖𝑖𝑖 (harmonic solution)

𝑑𝑑2 𝑥𝑥 𝑑𝑑𝑑𝑑 Driven harmonic oscillator

𝑚𝑚 2 + 𝑚𝑚𝛾𝛾 + 𝑚𝑚𝜔𝜔𝑜𝑜2 𝑥𝑥 = (−𝑒𝑒)𝐸𝐸𝑥𝑥
𝑑𝑑𝑡𝑡 𝑑𝑑𝑑𝑑

−𝑚𝑚𝑥𝑥0 𝜔𝜔2 + 𝑖𝑖𝑖𝑖𝛾𝛾𝛾𝛾 + 𝑚𝑚𝜔𝜔𝑜𝑜2 𝑥𝑥0 = (−𝑒𝑒)𝐸𝐸𝑥𝑥𝑥

𝜔𝜔 ≪ 𝜔𝜔0 𝜔𝜔 = 𝜔𝜔0 𝜔𝜔 ≫ 𝜔𝜔0

𝑥𝑥 in phase with 𝐸𝐸𝑥𝑥 𝑥𝑥 90𝑜𝑜 out of phase 𝑥𝑥 in antiphase

−𝑒𝑒 1 (small amplitude) with 𝐸𝐸𝑥𝑥 with 𝐸𝐸𝑥𝑥
𝑥𝑥0 = 𝐸𝐸
𝑚𝑚 𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 + 𝑖𝑖𝑖𝑖𝑖𝑖 𝑥𝑥𝑥 (large amplitude) (vanishing
amplitude)

 Let 𝑛𝑛 be the atomic density and assume each atom contributes one
“electron-nucleus” dipole, then the electric field induced polarization can
be written as
𝑃𝑃⃗ = 𝜀𝜀0 𝜒𝜒𝐸𝐸
2 2
𝑛𝑛𝑒𝑒 1 𝑛𝑛𝑒𝑒 1
𝑃𝑃𝑥𝑥 = −𝑛𝑛𝑛𝑛𝑥𝑥0 = 𝐸𝐸 = 𝜀𝜀 𝐸𝐸
𝑚𝑚 𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 + 𝑖𝑖𝑖𝑖𝑖𝑖 𝑥𝑥𝑥 𝜀𝜀0 𝑚𝑚 𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 + 𝑖𝑖𝑖𝑖𝑖𝑖 0 𝑥𝑥𝑥

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

40
Lorentz oscillator model  complex permittivity

𝑛𝑛𝑒𝑒 2 1 𝑛𝑛𝑒𝑒 2 𝜔𝜔𝑜𝑜2 𝜔𝜔𝑜𝑜2

𝜒𝜒 = = = 𝜒𝜒𝑜𝑜 2
𝜀𝜀0 𝑚𝑚 𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 + 𝑖𝑖𝑖𝑖𝑖𝑖 𝜀𝜀0 𝑚𝑚𝜔𝜔𝑜𝑜2 𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 + 𝑖𝑖𝑖𝑖𝑖𝑖 𝜔𝜔𝑜𝑜 − 𝜔𝜔 2 + 𝑖𝑖𝑖𝑖𝑖𝑖

𝜒𝜒 𝜔𝜔 = 𝜒𝜒1 𝜔𝜔 + 𝑖𝑖𝜒𝜒2 (𝜔𝜔) 1

𝑛𝑛𝑒𝑒 2 𝜔𝜔𝑝𝑝2 𝑛𝑛� 𝜔𝜔 = 𝑛𝑛 𝜔𝜔 + 𝑖𝑖𝜅𝜅 𝜔𝜔 = 𝜀𝜀 2 (𝜔𝜔)
𝜒𝜒𝑜𝑜 = =
𝜔𝜔𝑜𝑜2 (𝜔𝜔𝑜𝑜2 − 𝜔𝜔2 ) 𝜀𝜀0 𝑚𝑚𝜔𝜔𝑜𝑜2 𝜔𝜔𝑜𝑜2 𝜀𝜀1 𝜔𝜔 = 𝑛𝑛2 𝜔𝜔 − 𝜅𝜅 2 (𝜔𝜔)
𝜒𝜒1 𝜔𝜔 = 𝜒𝜒𝑜𝑜
𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 2 + (𝜔𝜔𝜔𝜔)2
𝜀𝜀2 𝜔𝜔 = 2𝑛𝑛 𝜔𝜔 𝜅𝜅(𝜔𝜔)
𝑛𝑛𝑒𝑒 2 Plasma
𝜔𝜔𝑝𝑝 =
𝜔𝜔𝑜𝑜2 𝜔𝜔𝛾𝛾 𝜀𝜀0 𝑚𝑚 frequency 𝜀𝜀𝑟𝑟 = 1 + 𝜒𝜒
𝜒𝜒𝟐𝟐 𝜔𝜔 = 𝜒𝜒𝑜𝑜
𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 2 + (𝜔𝜔𝜔𝜔)2

2𝜅𝜅𝜅𝜅 2𝜀𝜀2 𝜔𝜔 𝜒𝜒2 𝑘𝑘

𝛼𝛼 = = =
𝑐𝑐 2𝑐𝑐𝑛𝑛 𝑛𝑛
𝜀𝜀𝟐𝟐 𝜔𝜔

𝜔𝜔𝑜𝑜2 (𝜔𝜔𝑜𝑜2 − 𝜔𝜔2 ) 𝜀𝜀1 𝜔𝜔

𝜀𝜀1 𝜔𝜔 = 1 + 𝜒𝜒𝑜𝑜
𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 2 + (𝜔𝜔𝜔𝜔)2

𝜔𝜔𝑜𝑜2 𝜔𝜔𝛾𝛾
𝜀𝜀2 𝜔𝜔 = 𝜒𝜒𝑜𝑜
𝜔𝜔𝑜𝑜2 − 𝜔𝜔 2 2 + (𝜔𝜔𝜔𝜔)2

When 𝛾𝛾 = 0, there is no absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

41
Analogue with electric circuits

 EM wave in lossy medium

𝜕𝜕 2 𝐸𝐸 𝜕𝜕𝐸𝐸
𝛻𝛻 2 𝐸𝐸 − 𝜀𝜀0 𝜀𝜀𝑟𝑟 𝜇𝜇0 2 = 𝜇𝜇0 𝜎𝜎
𝜕𝜕𝑡𝑡 𝜕𝜕𝜕𝜕
𝜕𝜕 2 𝐸𝐸 𝜎𝜎 𝜕𝜕𝐸𝐸 𝑐𝑐 2 2
+ = 𝛻𝛻 𝐸𝐸
𝜕𝜕𝑡𝑡 2 𝜀𝜀0 𝜀𝜀𝑟𝑟 𝜕𝜕𝜕𝜕 𝑛𝑛2

 Light-charge interaction According to RCL: 𝑖𝑖𝑅𝑅 𝑡𝑡 + 𝑖𝑖𝐿𝐿 𝑡𝑡 + 𝑖𝑖𝐶𝐶 𝑡𝑡 = 𝑖𝑖𝑠𝑠 𝑡𝑡

𝑑𝑑2 𝑥𝑥 𝑑𝑑𝑑𝑑
𝑚𝑚 2 + 𝑚𝑚𝛾𝛾
𝑑𝑑𝑡𝑡 𝑑𝑑𝑑𝑑
+ 𝑚𝑚𝜔𝜔𝑜𝑜2 𝑥𝑥 = (−𝑒𝑒)𝐸𝐸𝑥𝑥 𝑑𝑑 2 𝑣𝑣(𝑡𝑡) 1 𝑑𝑑𝑑𝑑(𝑡𝑡) 1 𝑑𝑑𝑖𝑖𝑠𝑠 𝑡𝑡
𝐶𝐶 + + 𝑣𝑣 𝑡𝑡 =
𝑑𝑑𝑡𝑡 2 𝑅𝑅 𝑑𝑑𝑑𝑑 𝐿𝐿 𝑑𝑑𝑑𝑑
 2nd and 0th order time
derivative  oscillation
Damping or power consumption
 1st order time derivative
 damping/absorption/ 1
∝ 𝜎𝜎
dissipation 𝑅𝑅

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

42
Quantum mechanical model of light-matter
interaction: interaction Hamiltonian
 In quantum mechanics, light-matter interaction is
described by the following Hamiltonian: ∇ × 𝐸𝐸 = − 𝜕𝜕𝜕𝜕
𝜕𝜕𝐵𝐵 𝜕𝜕𝐷𝐷
∇ × 𝐻𝐻 = 𝐽𝐽⃗ + 𝜕𝜕𝜕𝜕
∇ ⋅ 𝐷𝐷 = 𝜌𝜌 ∇ ⋅ 𝐵𝐵 = 0
2
1 ℏ
𝐻𝐻 = ∇ − 𝑞𝑞 𝐴𝐴⃗ + 𝑉𝑉(𝑟𝑟)
⃗ 𝑞𝑞 = −𝑒𝑒 for electron 𝐷𝐷 = 𝜀𝜀0 𝐸𝐸 + 𝑃𝑃 𝐵𝐵 = 𝜇𝜇0 𝐻𝐻 + 𝑀𝑀
2𝑚𝑚 𝑖𝑖
𝐵𝐵 = ∇ × 𝐴𝐴⃗
𝜕𝜕𝐴𝐴⃗
𝑒𝑒 ℏ 𝑒𝑒 2 𝐸𝐸 = −∇𝜑𝜑 −
𝐻𝐻 = 𝐻𝐻𝑒𝑒𝑒𝑒 + ∇ � 𝐴𝐴⃗ + 𝐴𝐴⃗ � ∇ + ⃗2
|𝐴𝐴| 𝜕𝜕𝜕𝜕
2𝑚𝑚 𝑖𝑖 2𝑚𝑚
1 𝜕𝜕 2 𝜑𝜑 𝜌𝜌
2 ∇2 𝜑𝜑 − = −
1 ℏ 𝑐𝑐 2 𝜕𝜕𝑡𝑡 2 𝜀𝜀0
𝐻𝐻𝑒𝑒𝑒𝑒 = ∇ + 𝑉𝑉(𝑟𝑟)
⃗
2𝑚𝑚 𝑖𝑖 1 𝜕𝜕 2 𝐴𝐴⃗
∇ 𝐴𝐴⃗ −
2
= −𝜇𝜇0 𝐽𝐽⃗
𝑒𝑒 ℏ 𝑒𝑒 ℏ 𝑐𝑐 2 𝜕𝜕𝑡𝑡 2
𝐻𝐻𝑖𝑖𝑖𝑖𝑖𝑖 = ∇ � 𝐴𝐴⃗ + 𝐴𝐴⃗ � ∇ = 𝐴𝐴⃗ � ∇
2𝑚𝑚 𝑖𝑖 𝑚𝑚 𝑖𝑖
1 𝜕𝜕 2 𝐴𝐴⃗
∇ 𝐴𝐴⃗ −
2
= 0 when 𝐽𝐽⃗ = 0
𝑐𝑐 2 𝜕𝜕𝑡𝑡 2
Note: ∇ � 𝐴𝐴⃗ 𝜑𝜑 > = ∇ � 𝐴𝐴⃗ 𝜑𝜑 > +𝐴𝐴⃗ � ∇ 𝜑𝜑 >  ∇ ⋅ 𝐴𝐴̂ = 𝐴𝐴̂ � ∇

∇ ⋅ 𝐴𝐴⃗ = 0 Coulomb gauge 𝐴𝐴⃗ = 𝐴𝐴⃗0 𝑒𝑒 𝑖𝑖(𝑘𝑘𝑝𝑝𝑝 ⋅𝑟𝑟−𝜔𝜔𝜔𝜔)

⃗
+ 𝑐𝑐. 𝑐𝑐.
⃗
̂ 𝑖𝑖(𝑘𝑘𝑝𝑝𝑝 ⋅𝑟𝑟−𝜔𝜔𝜔𝜔)
= 𝐴𝐴0 𝑒𝑒𝑒𝑒 + 𝑐𝑐. 𝑐𝑐.
 Light intensity in a medium with refractive index n: 𝐸𝐸0 𝑖𝑖 𝑘𝑘 �𝑟𝑟−𝜔𝜔𝜔𝜔
= 𝑒𝑒𝑒𝑒
̂ 𝑝𝑝𝑝 ⃗ +𝑐𝑐.𝑐𝑐.
1 1 𝑖𝑖𝜔𝜔
𝐼𝐼 = |𝐸𝐸 × 𝐻𝐻| = 𝑐𝑐𝑐𝑐𝜀𝜀0 𝐸𝐸02 = 𝑐𝑐𝑐𝑐𝜀𝜀0 𝜔𝜔2 𝐴𝐴20
2 2
EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010
43
Light-charge interaction in semiconductors
 Wave function of electron in the conduction (f) and
valence band (i) are given by: Emission Initial
𝐸𝐸𝑖𝑖
state
𝜓𝜓𝑖𝑖,𝑓𝑓 𝑟𝑟, ⃗ 𝑖𝑖𝑘𝑘𝑖𝑖,𝑓𝑓 ⋅𝑟𝑟⃗ 𝑒𝑒 −𝑖𝑖𝑡𝑡𝐸𝐸𝑖𝑖,𝑓𝑓 /ℏ
⃗ 𝑡𝑡 = 𝑢𝑢𝑘𝑘𝑖𝑖,𝑓𝑓 (𝑟𝑟)𝑒𝑒
ℏ𝜔𝜔
𝐸𝐸𝑓𝑓 Final
state
Bloch function Time-dependent part
of wavefunction
Absorption Final
𝐸𝐸𝑓𝑓
 Light will cause change of quantum states of electrons. state
The transition probability between 𝑖𝑖 and 𝑓𝑓 states can
ℏ𝜔𝜔
be calculated using Fermi’s golden rule (applicable to 𝐸𝐸𝑖𝑖 Initial
both absorption and emission): state

2
2𝜋𝜋 ℏ𝑒𝑒𝐸𝐸0 2𝜋𝜋 2
𝑊𝑊𝑓𝑓𝑓𝑓 = 𝐻𝐻𝑖𝑖𝑖𝑖𝑖𝑖 2 𝛿𝛿 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 − ℏ𝜔𝜔 = 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 𝛿𝛿 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 − ℏ𝜔𝜔
ℏ 𝑚𝑚𝜔𝜔 ℏ

𝑒𝑒 ℏ 𝑒𝑒 ℏ 𝐸𝐸0 𝑖𝑖𝑘𝑘 �𝑟𝑟⃗

𝐻𝐻𝑖𝑖𝑖𝑖𝑖𝑖 = 𝐴𝐴⃗ � ∇= 𝑒𝑒𝑒𝑒
̂ 𝑝𝑝𝑝 � ∇
𝑚𝑚 𝑖𝑖 𝑚𝑚 𝑖𝑖 𝑖𝑖𝜔𝜔

𝑘𝑘𝑓𝑓 −𝑘𝑘𝑝𝑝𝑝 ⋅𝑟𝑟⃗ ∗

𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 =< 𝜓𝜓𝑓𝑓 𝑟𝑟, ̂ 𝑖𝑖𝑘𝑘𝑝𝑝𝑝 ⋅𝑟𝑟⃗ � ∇ 𝜓𝜓𝑖𝑖 𝑟𝑟,
⃗ 𝑡𝑡 𝑒𝑒𝑒𝑒 ⃗ 𝑡𝑡 > = ∭ 𝑒𝑒 −𝑖𝑖 𝑢𝑢𝑘𝑘𝑓𝑓 (𝑟𝑟)
⃗ 𝑒𝑒̂ ⃗ 𝑖𝑖𝑘𝑘𝑖𝑖 ⋅𝑟𝑟⃗ 𝑑𝑑3 𝑟𝑟⃗
� ∇𝑢𝑢𝑘𝑘𝑖𝑖 (𝑟𝑟)𝑒𝑒

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

44
Complex permittivity and absorption coefficient

 Add up contributions that satisfy energy conservation:

𝐼𝐼 = 𝐼𝐼0 𝑒𝑒 −𝛼𝛼𝒍𝒍 (W/m2)
2
2 ℏ𝑒𝑒𝐸𝐸0 2𝜋𝜋 2 Energy absorption per unit
𝑊𝑊 = � 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 𝛿𝛿 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 − ℏ𝜔𝜔 𝑑𝑑3 𝑘𝑘 volume per unit time:
(2𝜋𝜋)3 𝑚𝑚𝑚𝑚 ℏ
𝐼𝐼0 (1 − 𝑒𝑒 −𝛼𝛼∆𝒍𝒍 )
= 𝛼𝛼𝐼𝐼0
∆𝒍𝒍
 Photon energy loss per unit time per unit volume: 𝑊𝑊𝑊𝜔𝜔
2𝜅𝜅𝜅𝜅 1
𝛼𝛼𝐼𝐼0 = 𝑐𝑐𝑐𝑐𝜀𝜀0 𝐸𝐸02
1 𝑐𝑐 2
𝛼𝛼𝐼𝐼0 = 𝜔𝜔𝜀𝜀2 𝜀𝜀0 𝐸𝐸02 = 𝑊𝑊𝑊𝑊𝑊 𝜀𝜀2 𝜔𝜔 = 2𝑛𝑛 𝜔𝜔 𝜅𝜅(𝜔𝜔)
2
1
2𝑊𝑊𝑊 4𝜋𝜋ℏ 𝑒𝑒 2 2
2 𝛼𝛼𝐼𝐼0 = 𝜔𝜔𝜀𝜀2 𝜀𝜀0 𝐸𝐸02
2 2
𝜀𝜀2 = 2 = 2 2
� 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 𝛿𝛿 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 − ℏ𝜔𝜔 3
𝑑𝑑 3 𝑘𝑘
𝜀𝜀0 𝐸𝐸0 𝜀𝜀0 𝑚𝑚 𝜔𝜔 (2𝜋𝜋)

4𝜋𝜋ℏ2 𝑒𝑒 2 2 2 𝑑𝑑𝑆𝑆𝑘𝑘 2
𝜀𝜀2 = � 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 usually varies
𝜀𝜀0 𝑚𝑚2 𝜔𝜔 2 (2𝜋𝜋)3 |∇𝑘𝑘 (𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 )|𝐸𝐸𝑓𝑓 −𝐸𝐸𝑖𝑖 −ℏ𝜔𝜔
slowly with k

4𝜋𝜋ℏ2 𝑒𝑒 2 2 2 𝑑𝑑𝑆𝑆𝑘𝑘 4𝜋𝜋ℏ2 𝑒𝑒 2 2

≈ 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 � = 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 𝐽𝐽𝑖𝑖𝑖𝑖 (ℏ𝜔𝜔)
𝜀𝜀0 𝑚𝑚2 𝜔𝜔 2 (2𝜋𝜋)3 ∇𝑘𝑘 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 𝜀𝜀0 𝑚𝑚2 𝜔𝜔 2
𝐸𝐸𝑓𝑓 −𝐸𝐸𝑖𝑖 −ℏ𝜔𝜔

2 𝑑𝑑𝑆𝑆𝑘𝑘
𝐽𝐽𝑖𝑖𝑖𝑖 ℏ𝜔𝜔 = ∬ (Joint density of states)
(2𝜋𝜋)3 ∇𝑘𝑘 𝐸𝐸𝑓𝑓 −𝐸𝐸𝑖𝑖
𝐸𝐸𝑓𝑓 −𝐸𝐸𝑖𝑖 −ℏ𝜔𝜔

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

45
Derivation of joint density of states (JDOS)

𝑘𝑘𝑧𝑧

𝑥𝑥, 𝑦𝑦, 𝑧𝑧 → (𝑘𝑘𝑥𝑥 , 𝑘𝑘𝑦𝑦 , 𝑘𝑘𝑧𝑧 )

𝑟𝑟 → 𝑘𝑘
𝑘𝑘𝑦𝑦
𝑑𝑑𝑑𝑑 → 𝑑𝑑𝑆𝑆𝑘𝑘
𝑘𝑘𝑥𝑥
𝐸𝐸 = constant

ℏ2 𝑘𝑘 2 ℏ2 𝑘𝑘
When 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 = 𝐸𝐸𝑔𝑔 + , 𝑑𝑑𝑆𝑆𝑘𝑘 = 𝑘𝑘 2 sin 𝜃𝜃 𝑑𝑑𝜃𝜃𝜃𝜃𝜃𝜃, ∇𝑘𝑘 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 =
2𝑚𝑚𝑟𝑟 𝑚𝑚𝑟𝑟

2 𝑑𝑑𝑆𝑆𝑘𝑘 2 𝑘𝑘𝑚𝑚𝑟𝑟 2 𝑘𝑘𝑚𝑚𝑟𝑟

𝐽𝐽𝑖𝑖𝑖𝑖 ℏ𝜔𝜔 = � = � sin 𝜃𝜃 𝑑𝑑𝜃𝜃𝜃𝜃𝜃𝜃 = 4𝜋𝜋
(2𝜋𝜋)3 ∇𝑘𝑘 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 (2𝜋𝜋)3 ℏ2 (2𝜋𝜋)3 ℏ2
𝐸𝐸𝑓𝑓 −𝐸𝐸𝑖𝑖 −ℏ𝜔𝜔

2 𝑚𝑚𝑟𝑟 2𝑚𝑚𝑟𝑟 1/2 1 2𝑚𝑚𝑟𝑟 3/2 2𝑚𝑚𝑟𝑟 1/2

= 3 4𝜋𝜋 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 − 𝐸𝐸𝑔𝑔 = 𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 − 𝐸𝐸𝑔𝑔
(2𝜋𝜋) ℏ2 ℏ2 2𝜋𝜋2 ℏ2 ℏ2

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

46
Derivation of joint density of states (JDOS)

 Alternatively, JDOS can be derived as follows:

ℏ𝑘𝑘𝑓𝑓 = ℏ𝑘𝑘𝑖𝑖 + ℏ𝑘𝑘𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 ≈ ℏ𝑘𝑘𝑖𝑖 = ℏ𝑘𝑘 𝐸𝐸 𝑚𝑚𝑒𝑒∗ ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔

ℏ2 𝑘𝑘 2 ℏ2 𝑘𝑘 2 𝐸𝐸𝑓𝑓 𝑚𝑚𝑟𝑟∗
𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 = ℏ𝜔𝜔 = + ∗ + 𝐸𝐸𝑔𝑔
2𝑚𝑚𝑒𝑒∗ 2𝑚𝑚ℎℎ Reduced to
𝐸𝐸𝑔𝑔 (𝐸𝐸𝐶𝐶 )
ℏ2 𝑘𝑘 2
= 𝐸𝐸𝑔𝑔 +
2𝑚𝑚𝑟𝑟
𝑘𝑘 𝑘𝑘
𝐸𝐸𝑖𝑖
2𝑚𝑚𝑟𝑟 (ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 )
𝑘𝑘 2 = ∗
𝑚𝑚𝑙𝑙𝑙 ∗
𝑚𝑚ℎℎ
ℏ2

 Compared with DOS of free electron

𝐸𝐸
3/2
𝑚𝑚 1 2𝑚𝑚
𝐷𝐷(𝐸𝐸) = 2 𝐸𝐸
2𝜋𝜋 ℏ2
3/2
1 2𝑚𝑚𝑟𝑟
𝐽𝐽𝑖𝑖𝑖𝑖 ℏ𝜔𝜔 = 2 (ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 )1/2 DOS of free electron
2𝜋𝜋 ℏ2 𝑘𝑘
model discussed in
Lecture 3 (slide 38)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

47
Direct and indirect transitions

 In the context of semiconductor physics, direct and indirect transitions refer to the
mechanisms by which electrons move between energy bands, particularly from the valence
band to the conduction band. These transitions are fundamental to the operation of many
electronic and optoelectronic devices.
 Direct transitions occur when an electron in the valence band absorbs energy (e.g., from a
photon) and moves directly to the conduction band, conserving both energy and
momentum.
 In direct bandgap semiconductors, the minimum of the conduction band and the maximum
of the valence band are at the same point in the Brillouin zone (k-space), which means that
the electron can conserve both energy and momentum during the transition without
involving any additional particles like phonons.
 Indirect transitions occur when an electron in the valence band absorbs energy and moves
to the conduction band, but the process requires a change in momentum that is typically
provided via absorption or emission of phonons.
 In indirect bandgap semiconductors, the conduction and valence band extrema do not align
in k-space. Thus, the electron cannot conserve momentum during the transition, and a
phonon is needed to facilitate the momentum change.
 Direct transition is more efficient than indirect transition.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

48
Band-to-band transition

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

49
Absorption due to direct transition
3/2
4𝜋𝜋ℏ2 𝑒𝑒 2 2 4𝜋𝜋ℏ2 𝑒𝑒 2 1 2𝑚𝑚𝑟𝑟 2
𝜀𝜀2 = 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 𝐽𝐽𝑖𝑖𝑖𝑖 𝜔𝜔 = 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 (ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 )1/2
𝜀𝜀0 𝑚𝑚2 𝜔𝜔 2 𝜀𝜀0 𝑚𝑚2 𝜔𝜔 2 2𝜋𝜋 2 ℏ2

𝜔𝜔𝜀𝜀2 4𝜋𝜋ℏ2 𝑒𝑒 2 1 2𝑚𝑚𝑟𝑟 3/2 2 1/2 2 (ℏ𝜔𝜔−𝐸𝐸𝑔𝑔 )1/2

𝛼𝛼 = = 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 (ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 ) ∝ 𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖
𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛𝜀𝜀0 𝑚𝑚2 𝜔𝜔 2𝜋𝜋2 ℏ2 ℏ𝜔𝜔

𝑘𝑘𝑓𝑓 −𝑘𝑘𝑝𝑝𝑝 ⋅𝑟𝑟⃗ ∗

𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 = � 𝑒𝑒 −𝑖𝑖 𝑢𝑢𝑘𝑘𝑓𝑓 (𝑟𝑟)
⃗ 𝑒𝑒̂ ⃗ 𝑖𝑖𝑘𝑘𝑖𝑖 ⋅𝑟𝑟⃗ 𝑑𝑑3 𝑟𝑟⃗
� ∇𝑢𝑢𝑘𝑘𝑖𝑖 (𝑟𝑟)𝑒𝑒
𝐸𝐸

𝑘𝑘𝑓𝑓 −𝑘𝑘𝑖𝑖 −𝑘𝑘𝑝𝑝𝑝 ⋅𝑟𝑟⃗ ∗ 𝐸𝐸𝑓𝑓

= � 𝑒𝑒 −𝑖𝑖 𝑢𝑢𝑘𝑘𝑓𝑓 (𝑟𝑟)
⃗ 𝑒𝑒̂ ⃗ 𝑑𝑑3 𝑟𝑟⃗
� ∇𝑢𝑢𝑘𝑘𝑖𝑖 (𝑟𝑟) 𝐸𝐸𝑒𝑒 =
ℏ2 𝑘𝑘 2
2𝑚𝑚𝑒𝑒∗
0 ℏ𝜔𝜔
⋅𝑟𝑟⃗ ∗ 𝐸𝐸𝑔𝑔
+𝑖𝑖𝑒𝑒̂ � 𝑘𝑘𝑖𝑖 � 𝑒𝑒 −𝑖𝑖 𝑘𝑘𝑓𝑓 −𝑘𝑘𝑖𝑖 −𝑘𝑘𝑝𝑝𝑝
𝑢𝑢𝑘𝑘𝑓𝑓 (𝑟𝑟) ⃗ 𝑑𝑑 3 𝑟𝑟⃗
⃗ 𝑢𝑢𝑘𝑘𝑖𝑖 (𝑟𝑟)
𝑘𝑘

𝑢𝑢𝑘𝑘𝑖𝑖 : p − orbital; 𝑢𝑢𝑘𝑘𝑓𝑓 : 𝑠𝑠 − orbital → < 𝑢𝑢𝑘𝑘𝑖𝑖 |𝑢𝑢𝑘𝑘𝑓𝑓 > = 0 𝐸𝐸𝑖𝑖

−ℏ2 𝑘𝑘 2
𝐸𝐸ℎ =
2𝑚𝑚ℎ∗
𝑒𝑒̂ � 𝑀𝑀𝑖𝑖𝑖𝑖 will only be non-zero when 𝑘𝑘𝑓𝑓 − 𝑘𝑘𝑖𝑖 − 𝑘𝑘𝑝𝑝𝑝 = 0
or 𝐺𝐺⃗ (momentum conservation) because 𝑢𝑢𝑘𝑘∗ 𝑓𝑓 (𝑟𝑟) ⃗ 𝑒𝑒̂ �
⃗ varies rapidly over unit cell
∇𝑢𝑢𝑘𝑘𝑖𝑖 (𝑟𝑟)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

50
 Example:
A photon energy of 2.2 eV is absorbed by a valence
band electron in GaAs. If the band gap energy of
GaAs is 1.43 eV, calculate the energy of the electron
and hole produced by the photon absorption. Take
the electron effective mass (𝑚𝑚𝑒𝑒∗ ) and hole effective
mass (𝑚𝑚ℎ∗ )of GaAs to be 0.067mo and 0.45mo,
respectively. Hence, calculate the wave-vector k
measured from the conduction minimum.

Solution:
2 𝑘𝑘 2 2 𝑘𝑘 2
𝜔𝜔 = 𝐸𝐸𝑔𝑔 + or 𝜔𝜔 − 𝐸𝐸𝑔𝑔 =
2𝑚𝑚𝑟𝑟∗ 2𝑚𝑚𝑟𝑟∗

𝑚𝑚𝑟𝑟∗ 𝑚𝑚𝑟𝑟∗ 2 𝑘𝑘 2 0.058𝑚𝑚𝑜𝑜

Multiply both sides by , 𝜔𝜔 − 𝐸𝐸𝑔𝑔 = , 2.2 − 1.43 = 𝐸𝐸𝑒𝑒 , 𝐸𝐸𝑒𝑒 = 0.667 eV
𝑚𝑚𝑒𝑒∗ 𝑚𝑚𝑒𝑒∗ 2𝑚𝑚𝑒𝑒∗ 0.067𝑚𝑚𝑜𝑜

𝑚𝑚𝑟𝑟∗ −𝑚𝑚𝑟𝑟∗ −0.058𝑚𝑚𝑜𝑜

Similarly, multiply both sides by ∗ , 𝐸𝐸ℎ = ∗ 𝜔𝜔 − 𝐸𝐸𝑔𝑔 = , 𝐸𝐸ℎ = −0.099 eV
𝑚𝑚ℎ 𝑚𝑚ℎ 0.45𝑚𝑚𝑜𝑜

2𝑘𝑘 2 1.055 × 10−34 𝐽𝐽𝐽𝐽 2𝑘𝑘 2

𝐸𝐸𝑒𝑒 = = = 0.667 × 1.6 × 10−19 → 𝑘𝑘 = 1.08 × 109 m−1
2𝑚𝑚𝑒𝑒∗ 2 × 0.067 × 9.11 × 10−31 𝑘𝑘𝑘𝑘

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

51
Comparison with experimental results

1/2
𝛼𝛼 = 𝐶𝐶 ω − 𝐸𝐸𝑔𝑔
Eg 𝛼𝛼2 = 𝐶𝐶 2 (ω − 𝐸𝐸𝑔𝑔 )

photon energy ω (eV) photon energy ω (eV)

Eg

 Plots of the absorption coefficient as a function of photon energy for

GaAs and InAs. For direct transition one expects no absorption below
the energy gap, and therefore a steeply rising absorption edge. In
practice, one usually finds an exponentially increasing absorption edge.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

52
 Example:
GaAs is a direct gap semiconductor with allowed direct optical transition at the
absorption edge of 1.43 eV (band gap). The absorption coefficient α= 2.83×103 cm-1
at a photon energy hω = 1.45 eV.
(a) Calculate the absorption coefficient at 1.50 eV photon energy.
(b) With reflection coefficient R=0.25, calculate the transmission
coefficient of a 3 µm thick GaAs layer at photon energy 1.45 eV.

Solution:

1/2
(a) 𝛼𝛼 ℏ𝜔𝜔 = 𝐶𝐶 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 = 𝐶𝐶 ℏ𝜔𝜔 − 1.43 1/2

1/2
𝛼𝛼 𝐶𝐶 1.50 − 1.43
=
2.83 × 103 𝐶𝐶 1.45 − 1.43 1/2

𝛼𝛼 = 5.3 × 103 𝑐𝑐𝑚𝑚−1

(b) R = 0.25, d = 3 µm = 3×10-4 cm, 𝛼𝛼 = 2.83 × 103 cm-1, substituting into

1 − 𝑅𝑅 2 𝑒𝑒 −𝛼𝛼𝛼𝛼
𝑇𝑇 = = 0.24
1 − 𝑅𝑅 2 𝑒𝑒 −2𝛼𝛼𝛼𝛼

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

53
Absorption curves for some of the direct
and indirect semiconductors

 The absorption edge in

direct bandgap materials is
typically abrupt, showing a
sharp increase in
absorption at the bandgap
energy.

 The absorption edge in

indirect bandgap materials
is usually more gradual,
with a less defined onset.
This is because additional
momentum provided by
phonons is necessary to
make the transition
happen.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

54
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

55
Indirect absorption transition
𝐸𝐸
 Optical transition between different
valleys require the involvement of
Emit phonon
phonons.
 At low temperature, phonons can
only be generated, and therefore the
Photon Absorb
absorption occurs at energies above phonon
Energy: ℏ𝜔𝜔 Egind
the band gap (i.e., emitting phonons). momentum: ℏ𝑘𝑘𝑝𝑝𝑝
𝑘𝑘𝑓𝑓 − 𝑘𝑘𝑖𝑖
 At higher temperatures (typically 𝑘𝑘𝑓𝑓
𝑘𝑘
above 40K), acoustical phonons can
𝑘𝑘𝑓𝑓
assist the optical absorption
transition at a photon energy smaller
ℏ 𝑘𝑘𝑓𝑓 − 𝑘𝑘𝑖𝑖 = ±ℏ𝑞𝑞 + ℏ𝑘𝑘𝑝𝑝𝑝 ≈ ±ℏ𝑞𝑞
than the band gap (absorbing
(momentum conservation)
phonon).
𝐸𝐸𝑓𝑓 − 𝐸𝐸𝑖𝑖 = ±𝐸𝐸𝑝𝑝 + ℏ𝜔𝜔
 At even higher temperatures (> 200
(energy conservation)
K), optical phonons can also
(+: phonon absorption; -: phonon emission)
contribute to the optical transition.
See next slides for more explanation
EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010
56
 Indirect absorption transition

 An electron in the valence band absorbs a photon,

indicated by the orange arrow. However, the photon 𝐸𝐸
does not provide enough momentum to directly
transition the electron to the conduction band
Emit phonon
minimum (located at a different momentum).
 Since the electron needs a momentum boost to
reach the conduction band minimum, it also
interacts with a phonon (a quantized vibration of Photon Absorb
phonon
the crystal lattice). Energy: ℏ𝜔𝜔 Egind
momentum: ℏ𝑘𝑘𝑝𝑝𝑝
 The green arrow shows the electron absorbing a 𝑘𝑘𝑓𝑓 − 𝑘𝑘𝑖𝑖
phonon. This process provides the additional 𝑘𝑘
𝑘𝑘𝑓𝑓
momentum required for the electron to complete 𝑘𝑘𝑓𝑓
the transition to the conduction band minimum.
After absorbing both photon energy and phonon
momentum, the electron reaches the conduction
band.
 Alternatively, the electron may first move to a higher momentum state in the conduction
band and then emit a phonon to lose momentum (shown by the purple dashed line).

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

57
Absorption coefficient of indirect transition

 Theoretical calculations yield an absorption coefficient with a

quadratic dependence on energy
2
𝐴𝐴 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝑝𝑝𝒉𝒉
𝛼𝛼𝑎𝑎 (ℏ𝜔𝜔) = for ω > Eg - Eph
𝑒𝑒 𝐸𝐸𝑝𝑝𝒉𝒉 /𝑘𝑘𝑘𝑘
−1 Keh Yung Cheng
Springer 2020
2
𝐴𝐴 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 + 𝐸𝐸𝑝𝑝𝒉𝒉
𝛼𝛼𝑒𝑒 (ℏ𝜔𝜔) = for ω > Eg + Eph
−𝐸𝐸𝑝𝑝𝒉𝒉 /𝑘𝑘𝑘𝑘
1 − 𝑒𝑒

 Essentially, for the absorption into a specific (empty) conduction band

state (with 𝐷𝐷𝐷𝐷𝐷𝐷 ∝ 𝐸𝐸), various initial (filled) valence band states (also
with 𝐷𝐷𝐷𝐷𝐷𝐷 ∝ 𝐸𝐸) take part, resulting is a probability proportional to
the product of the two DOS, i.e., α(𝐸𝐸) ∝ 𝐸𝐸. Integrating over all energy
states with energy separation 𝐸𝐸 ± 𝐸𝐸𝑝𝑝𝑝 , yields an 𝐸𝐸 2 - dependence of
the absorption coefficient for indirect transition (interested students
may refer to the next slide for derivation).

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

58
 Theoretical derivation of 𝜶𝜶𝒂𝒂 and 𝜶𝜶𝒆𝒆
Electron-phono interaction (lattice vibration induces deformation potential):

𝑅𝑅𝑗𝑗 : 𝑗𝑗th lattice point

𝐻𝐻𝑒𝑒𝑒𝑒 = � 𝑉𝑉 𝑟𝑟⃗ − 𝑅𝑅𝑗𝑗 − 𝑢𝑢𝑗𝑗 − 𝑉𝑉(𝑟𝑟⃗ − 𝑅𝑅𝑗𝑗 ) ≈ − � 𝑢𝑢𝑗𝑗 � ∇𝑉𝑉(𝑟𝑟⃗ − 𝑅𝑅𝑗𝑗 )
𝑢𝑢𝑗𝑗 : small displacement
𝑅𝑅𝑗𝑗 𝑅𝑅𝑗𝑗

Indirection transition is a 2nd order transition, and the transition probability by absorbing
phonon 𝑞𝑞⃗ = 𝑘𝑘2 − 𝑘𝑘1 can be calculated as (derivation out of scope of this module):

2 2
ℏ𝑒𝑒𝐸𝐸0 2𝜋𝜋 < 𝜓𝜓𝑓𝑓𝑘𝑘2 |𝑉𝑉𝑝𝑝 (𝑞𝑞⃗ � 𝑟𝑟|𝜓𝜓
⃗ 𝛽𝛽𝑘𝑘1 > 𝑁𝑁𝑞𝑞 < 𝜓𝜓𝛽𝛽𝑘𝑘1 |𝑒𝑒̂ � 𝑀𝑀𝛽𝛽𝑓𝑓 |𝜓𝜓𝑖𝑖𝑘𝑘1 >
𝑊𝑊𝑓𝑓𝑘𝑘2 ,𝑖𝑖𝑘𝑘1 = 𝛿𝛿 𝐸𝐸𝑓𝑓 (𝑘𝑘2 ) − 𝐸𝐸𝑖𝑖 (𝑘𝑘1 ) − ℏ𝜔𝜔 + ℏ𝜔𝜔𝑞𝑞
𝑚𝑚𝜔𝜔 ℏ 𝐸𝐸𝛽𝛽 𝑘𝑘1 − 𝐸𝐸𝑖𝑖 𝑘𝑘1 − ℏ𝜔𝜔

1/2
ℏ
𝑉𝑉𝑝𝑝 𝑞𝑞⃗ � 𝑟𝑟⃗ = � 𝑒𝑒 𝑖𝑖𝑞𝑞�𝑅𝑅𝑗𝑗 𝑒𝑒̂𝑞𝑞 � ∇𝑉𝑉(𝑟𝑟⃗ − 𝑅𝑅𝑗𝑗 )
2𝑀𝑀𝜔𝜔𝑞𝑞
𝑅𝑅𝑗𝑗

Sum contributions from states satisfying

energy and momentum conservation
2
ℏ𝑒𝑒𝐸𝐸0 2 2𝜋𝜋 2 2 ⃗ 𝛽𝛽𝑘𝑘 > 𝑁𝑁𝑞𝑞 <𝜓𝜓𝛽𝛽𝑘𝑘 |𝑒𝑒�𝑀𝑀
<𝜓𝜓𝑓𝑓𝑘𝑘 |𝑉𝑉𝑝𝑝(𝑞𝑞�𝑟𝑟|𝜓𝜓 ̂ 𝛽𝛽𝑓𝑓 |𝜓𝜓𝑖𝑖𝑘𝑘 >
𝑊𝑊𝑓𝑓,𝑖𝑖 = 𝑚𝑚𝜔𝜔 ℏ
∭ (2𝜋𝜋)3 𝑑𝑑 3 𝑘𝑘1 ∭ (2𝜋𝜋)3 𝑑𝑑3 𝑘𝑘2 2 1 1
𝐸𝐸𝛽𝛽 𝑘𝑘1 −𝐸𝐸𝑖𝑖 𝑘𝑘1 −ℏ𝜔𝜔
1
𝛿𝛿 𝐸𝐸𝑓𝑓 (𝑘𝑘2 ) − 𝐸𝐸𝑖𝑖 (𝑘𝑘1 ) − ℏ𝜔𝜔 + ℏ𝜔𝜔𝑞𝑞

2ℏ𝑊𝑊𝑓𝑓,𝑖𝑖 𝜔𝜔𝜀𝜀2 ℏ𝜔𝜔−(𝐸𝐸𝑔𝑔 −𝐸𝐸𝑝𝑝ℎ )

2
ℏ𝜔𝜔−(𝐸𝐸𝑔𝑔 +𝐸𝐸𝑝𝑝𝑝 )
2
𝜀𝜀2 = 𝛼𝛼 = 𝛼𝛼𝑎𝑎 ∝ 𝛼𝛼𝑒𝑒 ∝
𝜀𝜀0 𝐸𝐸02 𝑛𝑛𝑛𝑛 𝑒𝑒 𝐸𝐸𝑝𝑝 /𝑘𝑘𝐵𝐵 𝑇𝑇 −1 1−𝑒𝑒 −𝐸𝐸𝑝𝑝 /𝑘𝑘𝐵𝐵 𝑇𝑇

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

59
Typical absorption spectrum due to
indirection transition

𝛼𝛼(ℏ𝜔𝜔) = 𝛼𝛼𝑎𝑎 (ℏ𝜔𝜔) + 𝛼𝛼𝑒𝑒 (ℏ𝜔𝜔) for ω > Eg + Eph

𝛼𝛼 2
𝐴𝐴 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝑝𝑝𝒉𝒉
𝛼𝛼𝑒𝑒 𝛼𝛼𝑒𝑒 𝛼𝛼𝑎𝑎 (ℏ𝜔𝜔) =
𝑒𝑒 𝐸𝐸𝑝𝑝𝒉𝒉 /𝑘𝑘𝑘𝑘 − 1

2
T2>T1 T1 𝐴𝐴 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 + 𝐸𝐸𝑝𝑝𝒉𝒉
𝛼𝛼𝑒𝑒 (ℏ𝜔𝜔) =
1 − 𝑒𝑒 −𝐸𝐸𝑝𝑝𝒉𝒉/𝑘𝑘𝑘𝑘

T→ 0, αa → 0, αe is finite
𝛼𝛼𝑎𝑎

𝛼𝛼𝑎𝑎
ω When α = 0
𝐸𝐸𝑔𝑔 − 𝐸𝐸𝑝𝑝𝒉𝒉
ω1 = 𝐸𝐸 g−𝐸𝐸𝑝𝑝
𝐸𝐸𝑔𝑔 − 𝐸𝐸𝑝𝑝𝒉𝒉
ω2 = 𝐸𝐸g + 𝐸𝐸𝑝𝑝
𝐸𝐸𝑔𝑔 + 𝐸𝐸𝑝𝑝𝒉𝒉 𝐸𝐸𝑔𝑔 + 𝐸𝐸𝑝𝑝𝒉𝒉

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

60
Direct v.s. indirect absorption spectrum

Absorption coefficient at a direct

band gap (a) and an indirect band
gap (b). In the later absorption (a)
and emission (e) of phonons
contribute to the optical absorption
process

Square root absorption coefficient as a

function of photon energy for a high-
purity Si measured at different
Temperatures

Keh Yung Cheng

Springer 2020

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

61
Example: absorption spectrum of silicon

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

62
Absorption spectrum of Si at different
temperature
𝛼𝛼𝛼𝛼𝛼 (𝑐𝑐𝑐𝑐−1/2 𝑒𝑒𝑒𝑒 1/2 )

Photon energy (eV)

Macfarlane, G. G., T. P. McLean, J. E. Quarrington, and V. Roberts, J.

Phys. Chem. Solids 8, (1959) 388-392.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

63
Example: absorption spectrum of germanium

𝛼𝛼 (𝑐𝑐𝑐𝑐−1/2 )

Photon energy (eV)

Absorption spectrum of high purity Ge at various temperatures

(excitons are observed in both direct and indirection regions)

Macfarlane, G. G., T. P. McLean, J. E. Quarrington, and V. Roberts, J. Phys. Chem. Solids 8, (1959) 388-392.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

64
Example: absorption spectrum of GaP

Absorption edge of GaP ( 𝛼𝛼 versus E) at various temperatures. The index ‘e’

(‘a’) indicates phonon emission (absorption) during the optical absorption
process. The theoretical excitonic gap (Egx) at T = 7K is indicated.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

65
Measurement of absorption spectrum

ℎ𝑐𝑐
ω =
𝜆𝜆
(4.13 eV to 0.35 eV

Lock-in amplifier Power supply λ= 0.3 to 3.5 µm

Light source

Xenon lamp

Chopper controller

Sample ω =
ℎ𝑐𝑐
Filter Diffraction
IT 𝜆𝜆 grating

Photodetector Chopper Monochromator

A schematic diagram of an experimental set up used in an optical absorption measurement. The light
source is a Xenon lamp that emits radiation in the UV to near IR part of the electromagnetic spectrum.
The monochromator is used to select the range individual wavelength (monochromatic) from the
Xenon lamp.
EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010
66
Analysis of absorption spectrum

1
𝐼𝐼𝑇𝑇 = 𝐼𝐼𝑜𝑜 1 − 𝑅𝑅 2𝑒𝑒 −𝛼𝛼𝛼𝛼 t: is thickness of the sample
1 − 𝑅𝑅2𝑒𝑒 −2𝛼𝛼𝛼𝛼
1/2 ℎ𝑐𝑐
α is related to bandgap by 𝛼𝛼 = 𝐶𝐶 ω − 𝐸𝐸𝑔𝑔 where ω =
λ
R, IT , Io , t and λ are measured quantity

A plot of α against ω to obtain Eg

𝐼𝐼0
2 −α l
Let say, αl >> 1 T ≈ (1 − R ) e
𝐼𝐼0 (1 − 𝑅𝑅1 )(1 − 𝑅𝑅2 )𝑒𝑒 −𝛼𝛼𝛼𝛼
𝐼𝐼0 𝑅𝑅1
−1 𝑇𝑇
𝛼𝛼 = 𝑙𝑙𝑙𝑙 2 𝐼𝐼0 (1 − 𝑅𝑅1 )2 𝑅𝑅2 𝑒𝑒 −2𝛼𝛼𝛼𝛼
𝑙𝑙 1 − 𝑅𝑅

−1 𝑇𝑇 1/2
𝛼𝛼 = 𝑙𝑙𝑙𝑙 = 𝐶𝐶 ω − 𝐸𝐸𝑔𝑔 Let 𝑅𝑅1 = 𝑅𝑅2 = 𝑅𝑅
𝑙𝑙 1−𝑅𝑅 2
1
𝐼𝐼𝑇𝑇 = 𝐼𝐼0 (1 − 𝑅𝑅)2 𝑒𝑒 −𝛼𝛼𝛼𝛼
𝐼𝐼
𝑇𝑇� 1 − 𝑅𝑅 2 𝑒𝑒 −2𝛼𝛼𝛼𝛼
−1 𝐼𝐼𝑜𝑜 1/2
𝛼𝛼 = 𝑙𝑙𝑙𝑙 = 𝐶𝐶 ω − 𝐸𝐸𝑔𝑔
𝑙𝑙 1−𝑅𝑅 2

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

67
 Example:

The absorption coefficient of an indirect gap semiconductor Ge at temperature 77K

can be expressed as
2
𝛼𝛼 ℏ𝜔𝜔 = 𝐶𝐶 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 + 𝐸𝐸𝑝𝑝

where Eg is the energy gap, Ep=0.04 eV is the energy of phonon involved in the
indirect optical transition and C is the constant. Given ℏ𝜔𝜔 = 0.74 𝑒𝑒𝑒𝑒 for 𝛼𝛼 =
0.2 𝑐𝑐𝑐𝑐−1 and ℏ𝜔𝜔 = 0.78 𝑒𝑒𝑒𝑒 for 𝛼𝛼 = 2 𝑐𝑐𝑐𝑐−1 , calculate
(i) the energy gap of Ge at 77 K
(ii) the penetration depth for the photon energy ℏ𝜔𝜔 = 0.8 𝑒𝑒𝑒𝑒

Solution:
2
0.2 𝐶𝐶 0.74− 𝐸𝐸𝑔𝑔 +0.04
i = 2 → 𝐸𝐸𝑔𝑔 = 0.683 𝑒𝑒𝑒𝑒
2 𝐶𝐶 0.78− 𝐸𝐸𝑔𝑔 +0.04

2
𝛼𝛼 𝐶𝐶 0.8 − 0.683 + 0.04 1
ii = 2
→ 𝛼𝛼 = 3.56 𝑐𝑐𝑐𝑐−1 → 𝑑𝑑 = = 0.281 𝑛𝑛𝑛𝑛
2 𝐶𝐶 0.78 − 0.683 + 0.04 𝛼𝛼

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

68
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

69
Exciton absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

70
Excitons
 An electron in the conduction band and a hole in the valence
band form a hydrogen-like state due to the mutual Coulomb
interaction. Such a state is called an exciton.
 Wannier-Mott (WM) excitons (weakly bound electron-hole
pair) can be observed in semiconductors in which an electron
and a hole which are attracted to each other by the
electrostatic Coulomb force. WM exciton is also known as free
exciton.
 Frenkel excitons (tightly bound exciton) have a much smaller
radius which is comparable to the size of the unit cell. This Wannier-Mott exciton
makes them localized at states which are tightly bound to (Free exciton)
specific atoms or molecules (less mobile).
 Frenkel excitons are typically found in alkali halide crystals (LiF,
NaF, NaCl, KCl, LiBr, NaBr, etc) and in organic molecular
crystals. In these materials which has small dielectric constant,
the Coulomb interaction between an electron and a hole is
strong and the excitons thus tend to be small, of the same
order as the size of the unit cell.
 The exciton is stable only if the attractive potential is strong
enough to prevent the exciton from dissociating into free
electron and free hole owing to collisions with phonons. Frenkel exciton

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

71
Free excitons: binging energy and radius
 We can model the free exciton as a hydrogenic
system. Unlike the pair of a light electron and a
heavy proton, the exciton is composed of 2 light
quasi-particles with comparable masses, which
entails a lower stability of the exciton in
comparison with the hydrogen atom.
 To obtain the exciton transition energy and the
binding energy, we solve the Schrödinger
equation: Centre 𝑚𝑚ℎ∗
of mass +
−ℏ2 2 ℏ2 2 𝑒𝑒 2
𝛻𝛻 − 𝛻𝛻 − 𝛷𝛷 = 𝐸𝐸𝐸𝐸
2𝑚𝑚𝑒𝑒∗ 𝑒𝑒 2𝑚𝑚ℎ∗ ℎ 4𝜋𝜋𝜀𝜀𝑟𝑟 𝜀𝜀𝑜𝑜 𝑟𝑟𝑒𝑒 − 𝑟𝑟ℎ 𝑚𝑚𝑒𝑒∗
- 𝑅𝑅 𝑟𝑟⃗ℎ

which is mathematically identical to that of the 𝑟𝑟⃗𝑒𝑒

hydrogen atom except that both holes and
electrons are moving.
 Separation of center of mass and relative motion 𝑚𝑚𝑒𝑒∗ 𝑟𝑟⃗𝑒𝑒 + 𝑚𝑚ℎ∗ 𝑟𝑟⃗ℎ 𝑟𝑟⃑ = 𝑟𝑟⃗𝑒𝑒 − 𝑟𝑟⃗ℎ
𝑅𝑅 =
leads to −2 𝑚𝑚𝑒𝑒∗ + 𝑚𝑚ℎ∗
𝛻𝛻𝑅𝑅2 𝜑𝜑 𝑅𝑅 = 𝐸𝐸𝑅𝑅 𝜑𝜑 𝑅𝑅
2𝑀𝑀 1 1 1
−2 2 𝑒𝑒 2
𝑀𝑀 = 𝑚𝑚𝑒𝑒∗ + 𝑚𝑚ℎ∗ = +
𝑚𝑚𝑟𝑟∗ 𝑚𝑚𝑒𝑒∗ 𝑚𝑚ℎ∗
− 𝛻𝛻 − φ 𝑟𝑟 = 𝐸𝐸𝑟𝑟 φ 𝑟𝑟
2𝑚𝑚𝑟𝑟∗ 𝑟𝑟 4𝜋𝜋𝜀𝜀𝑟𝑟 𝜀𝜀𝑜𝑜 𝑟𝑟

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

72
Free excitons: binging energy and radius
 Solution of the two Schrödinger equations are as follows:

−2 2 ℏ2 𝐾𝐾𝑐𝑐2
𝛻𝛻 𝜑𝜑 𝑅𝑅 = 𝐸𝐸𝑅𝑅 𝜑𝜑 𝑅𝑅 → 𝜑𝜑 𝑅𝑅 = 𝐶𝐶𝑒𝑒 𝑖𝑖𝐾𝐾𝐶𝐶 �𝑅𝑅
and 𝐸𝐸𝑅𝑅 = 𝐾𝐾𝑐𝑐 = 𝑘𝑘𝑒𝑒 + 𝑘𝑘ℎ
2𝑀𝑀 𝑅𝑅 2𝑀𝑀

−2 2 𝑒𝑒 2 𝑚𝑚𝑟𝑟∗ 𝑅𝑅𝐻𝐻 1 1

− ∗ 𝛻𝛻𝑟𝑟 − φ 𝑟𝑟 = 𝐸𝐸𝑟𝑟 φ 𝑟𝑟 → 𝐸𝐸𝑟𝑟 = − = −𝐸𝐸𝐵𝐵3𝐷𝐷
2𝑚𝑚𝑟𝑟 4𝜋𝜋𝜀𝜀𝑟𝑟 𝜀𝜀𝑜𝑜 𝑟𝑟 𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟 𝑛𝑛2 𝑛𝑛2

𝐸𝐸𝐵𝐵3𝐷𝐷 is the exciton binding energy (for n =1), which is written as

𝑚𝑚𝑟𝑟∗ 𝑅𝑅𝐻𝐻
𝐸𝐸𝐵𝐵3𝐷𝐷 =
𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟2

𝑒𝑒 4 𝑚𝑚𝑜𝑜
where 𝑅𝑅𝐻𝐻 = = 13.6 𝑒𝑒𝑒𝑒 (Rydberg energy)
2ℏ2 4𝜋𝜋∈𝑜𝑜 2
𝑚𝑚𝑜𝑜
The exciton radius is given by: 𝑟𝑟𝑛𝑛 = 𝜀𝜀 𝑎𝑎 𝑛𝑛2 = 𝑎𝑎𝐵𝐵 𝑛𝑛2 , where aH and
𝑚𝑚𝑟𝑟∗ 𝑟𝑟 𝐻𝐻
aB are hydrogen Bohr radius and exciton Bohr radius, respectively.

4𝜋𝜋𝜀𝜀𝑜𝑜 ℏ2 𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟 𝑎𝑎𝐻𝐻

𝑎𝑎𝐻𝐻 = = 5.29 × 10−11 m = 0.0529 nm 𝑎𝑎𝐵𝐵 =
𝑚𝑚𝑜𝑜 𝑒𝑒 2 𝑚𝑚𝑟𝑟∗

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

73
Free exciton transition energy

 Total energy of exciton:

1 ℏ2 𝐾𝐾𝑐𝑐2 ℏ2 𝐾𝐾𝑐𝑐2
𝐸𝐸 = 𝐸𝐸𝑔𝑔 + 𝐸𝐸𝑟𝑟 + 𝐸𝐸𝑅𝑅 = 𝐸𝐸𝑔𝑔 − 𝜀𝜀𝐵𝐵 + = 𝐸𝐸𝑋𝑋 +
𝑛𝑛2 2𝑀𝑀 2𝑀𝑀

For 𝐾𝐾𝑐𝑐 = 𝑘𝑘𝑒𝑒 + 𝑘𝑘ℎ = 0,

𝐸𝐸𝐵𝐵3𝐷𝐷
1
𝐸𝐸x = 𝐸𝐸g − 𝐸𝐸𝐵𝐵3𝐷𝐷 2 Free exciton transition energy
𝑛𝑛 𝐸𝐸g

0 𝐾𝐾c
Example:
Two particle picture
Calculate 𝐸𝐸𝐵𝐵3𝐷𝐷 and 𝑎𝑎𝐵𝐵 for GaAs.
Given 𝜀𝜀 = 13, 𝑚𝑚𝑒𝑒∗ = 0.067𝑚𝑚𝑜𝑜 , 𝑚𝑚ℎ∗ = 0.53𝑚𝑚𝑜𝑜 , 𝑚𝑚𝑟𝑟∗ = 0.059𝑚𝑚𝑜𝑜

Solution:
𝑚𝑚𝑟𝑟∗ 𝑅𝑅𝐻𝐻 0.059𝑚𝑚𝑜𝑜 13.6
𝐸𝐸𝐵𝐵3𝐷𝐷 = = = 0.0047 𝑒𝑒𝑒𝑒 = 4.7 𝑚𝑚𝑚𝑚𝑚𝑚
𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟2 𝑚𝑚𝑜𝑜 132

𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟 𝑎𝑎𝐻𝐻 𝑚𝑚𝑜𝑜 × 13𝑥𝑥𝑥.0529

𝑎𝑎𝐵𝐵 = = = 11.7 𝑛𝑛𝑛𝑛
𝑚𝑚𝑟𝑟∗ 0.059𝑚𝑚𝑜𝑜

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

74
 K. W. Böer and U. W. Pohl,
Springer, 2020

Giant Rydberg excitons in the

copper oxide Cu2O

Nature 514, 343-7 (2014)

Bound excitons

 Impurity can attract excitons that become localized at the impurity site,
becoming bound excitons. For example, a free hole can combine with a neutral
donor to form a positively charged exciton ion.
 In this case, the electron bound to donor still travels in a wide orbit about the
donor. The associated hole which moves in the electrostatic field of the “fixed”
dipole, determined by the instantaneous position of the electron, then also
travels about this donor; for this reason, this complex is called a bound exciton.
The same also applies to exciton bound to acceptors.

The bound exciton transition energy is

𝐸𝐸𝐵𝐵𝐵𝐵 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐵𝐵3𝐷𝐷 − 𝐸𝐸𝐵𝐵𝐵𝐵

where 𝐸𝐸𝐵𝐵𝐵𝐵 is the bound exciton EBX

binding energy.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

76
Hydrogen atom Positively charged proton Wannier-Mott exciton
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

𝐸𝐸𝐵𝐵3𝐷𝐷 = − 2 𝑛𝑛2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐵𝐵 = 𝑚𝑚∗
𝜀𝜀𝑟𝑟 𝑟𝑟
𝑚𝑚𝑒𝑒∗
- - +
P Note: here 𝑚𝑚𝑟𝑟∗ is normalized to 𝑚𝑚0
𝜀𝜀𝑟𝑟 𝜀𝜀0

Bound exciton
𝑚𝑚𝑒𝑒∗ 1 𝜀𝜀𝑟𝑟 EBX
𝐸𝐸𝑑𝑑 = × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎 𝐷𝐷 = 𝑎𝑎 (for donor)
𝜀𝜀𝑟𝑟2 𝑛𝑛2 𝑚𝑚𝑒𝑒∗ 0

𝑚𝑚ℎ∗ 1 𝜀𝜀𝑟𝑟
𝐸𝐸𝑎𝑎 = 2 2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐴𝐴 = ∗ 𝑎𝑎0 (for acceptor)
𝜀𝜀𝑟𝑟 𝑛𝑛 𝑚𝑚ℎ

Note: here 𝑚𝑚𝑒𝑒∗ and 𝑚𝑚ℎ∗ are normalized to 𝑚𝑚0 𝐸𝐸𝐵𝐵𝐵𝐵 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐵𝐵3𝐷𝐷 − 𝐸𝐸𝐵𝐵𝐵𝐵
77
Exciton absorption spectrum

 The exciton absorption consists of discrete

absorption peaks, with peak separation energy
decreases as index 𝑛𝑛 increases, which
eventually merges with the band edge
absorption.
𝜀𝜀𝐵𝐵
𝛼𝛼 ∝ � 𝛿𝛿 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 +
𝑛𝑛2

 Experimentally, the exciton transitions have

linewidth broadening, due to scattering from
impurities and phonons, where for high
temperature, the exciton peaks will blur into
the absorption band edge. i.e., when
𝑘𝑘𝐵𝐵 𝑇𝑇 > 𝐸𝐸𝐵𝐵3𝐷𝐷 .
High-resolution absorption spectrum of
highly pure GaAs at T = 1.2K in the
 Therefore, exciton absorption is typically
exciton region. Dashed line is theory
observed at low temperature.
without excitonic correlation.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

78
Example: exciton absorption in GaAs

 The exciton absorption is more prominent at low temperature. As

temperature increases, it weakens and eventually disappears due to
dissociation of exciton by thermal agitation.

At 185K , 𝑘𝑘𝐵𝐵 𝑇𝑇 = 16 𝑚𝑚𝑚𝑚𝑚𝑚, the absorption

spectrum shows weak exciton peak at the
band edge even though 𝑘𝑘𝐵𝐵 𝑇𝑇 is ~ 3 times
greater than the exciton binding energy
𝐸𝐸𝐵𝐵3𝐷𝐷 = 4.7 𝑚𝑚𝑚𝑚𝑚𝑚.
This indicate the criteria used 𝐸𝐸𝐵𝐵3𝐷𝐷 > 𝑘𝑘𝐵𝐵 𝑇𝑇
should not be taken strictly.

4.7 𝑚𝑚𝑚𝑚𝑚𝑚 → 𝑇𝑇 = 54.5 𝐾𝐾

Keh Yung Cheng

Springer 2020

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

79
Example: exciton absorption in GaN

exciton

Exciton absorption in GaN

measured at 300 K

The binding energy of WM exciton in GaAs is 5 meV, GaN (28 meV) and ZnO (60
meV), while Frenkel excitons have larger binding energy 0.1 – 1 eV.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

80
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

81
Absorption by carrier bound to shallow
donors and acceptors

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

82
Absorption by carrier bound to shallow
donors and acceptors
 Shallow donors and acceptors in semiconductors behave somewhat
like “hydrogen atoms embedded in a solid”, leading to formation of
discrete energy levels near the conduction and valence band edge.
-
 When a semiconductor with shallow impurity levels is exposed to P
light, the impurity atoms can absorb photons, promoting electrons
from the valence band to the shallow donor levels or from the
shallow acceptor levels to the conduction band.
 Optical transition also occur between impurity energy levers. The 𝜀𝜀𝑟𝑟 =11.7 for Si
absorption peaks corresponding to these shallow impurity levels can
provide information about the concentration and type of dopants 𝑚𝑚𝑒𝑒∗
present in the material. -

𝜀𝜀𝑟𝑟 𝜀𝜀0

Hydrogenic
atom

𝑚𝑚∗ 13.6
𝐸𝐸𝑛𝑛 = − 2 2 𝑒𝑒𝑒𝑒
𝜀𝜀𝑟𝑟 𝑛𝑛
Note: 𝑚𝑚∗ is dimensionless

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

83
Absorption by carrier bound to shallow
donors and acceptors
 The transition between an impurity and a band involves the whole band of
levels. Hence the transitions between an impurity and a band should manifest
themselves by a shoulder in the absorption edge at a threshold lower than the
Eg, i.e., ℏ𝜔𝜔 ≥ 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝑖𝑖

Ec
Ei

Impurity absorption
in InSb, T~10K.
ω =
ω

Ei = 7.9 meV EV
ω

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

84
Example: impurity absorption in Si:P

Absorption spectrum of n-type Si doped with P at a density of 1.2×1020 m-3 at 4.2 K. The
absorption lines correspond to transitions exciting electrons from n = 1 shell to higher
shells. In the language of atomic physics, 1s → np transition.
These transitions converge at higher n to the donor ionization energy of P in Si, which is
45 meV.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

85
Phonon absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

86
Phonon absorption

Due to the inversion

symmetry of the
elemental covalent
semiconductors one-
phonon infrared
absorption is forbidden,
and the absorption is
dominated by two
phonon processes, the
photon-phonon
coupling constants
being the second-order
dipole moments

PHYSICAL REVIEW B 69,

045205 ~2004!

1 𝑐𝑐𝑐𝑐−1 = 0.12398 meV

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

87
Donor and acceptor absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

88
Donor and acceptor absorption

J. Phys. Chem. Solids

1956. Vol. 1. pp. 65-74.

J. Phys. Chem. Solids

1956. Vol. 1. pp. 75-81.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

89
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

90
Mott transition in doped semiconductors
 As discussed in Lecture 5, doping is an effective way to introduce movable carriers to
semiconductors including both electrons and holes. The doping level significantly affects
the conductivity of semiconductors. Take n-type doping as an example, when
Nd << nc: states of adjacent impurities are isolated and do not interact. The
impurity ground states are discrete and energetically well-defined levels.
Nd ≤ nc: states of adjacent impurities interact, split, and finally form an impurity band.
Nd ≥ nc: the impurity band widens and merges with the conduction band. This is
known as Mott transition (𝑛𝑛𝑐𝑐 is the critical concentration defined in next slide).

E E E

k
k k
Ec Ec Ec
∆ED
Nd 𝑒𝑒 2 1/3
𝛥𝛥𝐸𝐸D ≈ 𝑁𝑁
4𝜋𝜋𝜋𝜋𝜀𝜀𝑜𝑜 𝐷𝐷
Ev Ev Ev

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

91
Mott transition in doped semiconductors

 The Mott transition occurs at the impurity concentration at which the distance between
impurities becomes comparable to Bohr radius of the impurity. It can be shown that the
Bohr orbital of an impurity is likely to overlap with the orbitals of one, two, or three
neighboring impurities if
𝜀𝜀: permittivity
𝜀𝜀𝑟𝑟 𝑎𝑎𝐻𝐻 𝑚𝑚∗ : nomalized effective mass
1/3
𝑎𝑎𝐵𝐵∗ 𝑛𝑛𝑐𝑐 ≈ 0.24 𝑎𝑎𝐵𝐵∗ =
𝑚𝑚∗ 𝑎𝑎𝐻𝐻 = 0.0529 𝑛𝑛𝑛𝑛

Example:

GaAs is heavily doped with donor impurity. Calculate 𝑛𝑛𝑐𝑐 based on the
Mott Criterion. Given 𝜀𝜀𝑟𝑟 = 13, 𝑚𝑚𝑒𝑒∗ = 0.067𝑚𝑚𝑜𝑜

Solution:
𝜀𝜀𝑟𝑟 𝑎𝑎𝐻𝐻 13×0.529
𝑎𝑎𝐵𝐵∗ = = = 10.26 𝑛𝑛𝑛𝑛
𝑚𝑚𝑒𝑒∗ 0.067
3
0.24
𝑛𝑛𝑐𝑐 = = 1.3 × 1016 𝑐𝑐𝑐𝑐−3
10.26 × 10−7 𝑐𝑐𝑐𝑐

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

92
Band-tailing effect in semiconductors

 The band tailing effect in semiconductors is caused by the presence of localized states
within the bandgap of the material, which can arise due to various factors, and their
presence affects the optical and electronic properties of the semiconductor.
 The major factors include disorder/defects, alloying and composition inhomogeneity,
doping effects, strain, interface, surface, etc.

-- --
2 𝑘𝑘 2
𝐸𝐸 =
2𝑚𝑚𝑒𝑒∗
(𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝)

Eg 2 𝑘𝑘 2 An ionized donor exerts an attractive

𝐸𝐸 =
2𝑚𝑚ℎ∗ force on the CB electrons and
(𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝) repulsive force on the VB hole
(acceptors act conversely).

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

93
Urbach’s Rule

 According to the Urbach’s rule, the absorption coefficient below the bandgap energy
can be expressed as:

ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔
𝛼𝛼 = 𝛼𝛼𝑜𝑜 exp for ω < 𝐸𝐸𝑔𝑔
Δ𝐸𝐸

where αo : the experimentally determined absorption coefficient

at the bandgap energy; i.e., when ω = Eg
∆E : the characteristic energy known as Urbach energy

In a number of materials, it is found that the slope of the absorption edge on a

semi-logarithmic plot gives
−1
𝑑𝑑 ln 𝛼𝛼
Δ𝐸𝐸 =
𝑑𝑑 ℏ𝜔𝜔

The Urbach rule was originally reported for alkali halides. It has been observed for
many ionic crystals, degenerately doped crystalline semiconductors, almost all
amorphous semiconductors. Typical values of ∆E range from 5 meV for highly quality
GaAs to > 42 meV for amorphous Si.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

94
Urbach’s Rule

(Log scale)

(
 ω − E g 
α = α o exp 
)
 𝛼𝛼 ∝ ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔
1/2
 ∆E 

(ω)
absorption edge
(due to band tailing effect) absorption edge
(ideal semiconductor)

 Absorption coefficient of a semiconductor. No light is absorbed in an idealized

semiconductor for photon energies below the bandgap. The below-bandgap
absorption in a real semiconductor can be described by “Urbach tail”.

 Urbach measured the absorption tail for different temperature and shows that the
∆E ~ kBT (thermal energy): phonon assisted transition. Another mechanism is due to
potential fluctuation.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

95
Burstein-Moss Shift (BMS)

 The Burstein-Moss shift (BMS) effect (or

sometimes known as “band filling”) occurs
in all heavily doped bulk semiconductors.
In the case of strongly degenerate n+ E
doped sample, this has the effect of
prohibiting any interband transition into
electron states below the Fermi energy
leading to an upward shift in the EF
absorption edge determined by the Ee
Ec EE’
g
fundamental energy gap 𝐸𝐸𝑔𝑔𝑔𝑔 . g
 The BMS ( ∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 ) is defined as the EEgog ω
difference between the effective
absorption edge (𝐸𝐸𝑔𝑔 ) and the fundamental Eh
energy gap (𝐸𝐸𝑔𝑔𝑔𝑔 ) of the material,
ko
k
2 𝑘𝑘𝐹𝐹2 2 𝑘𝑘𝐹𝐹2
∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝑔𝑔𝑔𝑔 = +
2𝑚𝑚𝑒𝑒∗ 2𝑚𝑚ℎ∗
1
2
𝑘𝑘𝐹𝐹 = 3𝜋𝜋 𝑛𝑛 3 is the Fermi-wave-vector

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

96
Burstein-Moss Shift (BMS)

 At high doping concentrations the bandgap of semiconductors decreases due to increase in

energy of valence-band maximum (VBM) and decrease in energy of the conduction-band
minimum(CBM).
 The magnitude of bandgap reduction increases with the doping concentration. This is
usually referred to as bandgap narrowing (sometimes known as bandgap shrinkage, or
bandgap renormalization).
 The most important reasons for bandgap narrowing are many-body effects which lower the
electron energies as compared to a non-interacting carrier system.
 Many-body effects describe the interaction of free carriers such as 𝑒𝑒 − 𝑒𝑒 interactions, ℎ −
ℎ interactions, 𝑒𝑒 − ℎ interactions, 𝑒𝑒 −donor interactions and ℎ −acceptor interactions. The
interaction becomes important at high free carrier concentrations.

The effect of charge carriers on the lowest

energy transition from the valence to
𝐸𝐸𝑔𝑔 = 𝐸𝐸𝑔𝑔𝑔 = conductions of a highly doped n-type
𝐸𝐸𝑔𝑔𝑔𝑔 𝐸𝐸𝑔𝑔𝑜𝑜 + ∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 𝐸𝐸𝑔𝑔 − ∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 semiconductor where ∆EBMS is the
Burstein-Moss shift and ∆EBGN represents
band-gap narrowing.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

97
Competition between BMS and BGN

 BMS causes the absorption curve to shift to

higher energy (blue shift).
 BNG causes the absorption curve to shift to
lower energy (red-shift).
 In n-type GaAs, BMS shift predominates over
bandgap narrowing, while in p-type GaAs,
bandgap narrowing prevails. The strongly
different characteristics of n-type and p-type
GaAs are due to the heavier hole effective mass
as compared to the electron effective mass.
 High-purity GaAs exhibits a rapidly increasing α
at the fundamental gap. As concentration
increases, the α below the gap energy increases
gradually and forms band tailing states.
 The absorption curve of high purity GaAs has the
steepest slope. The slope becomes smaller with
increasing carrier concentration.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

98
 Example:

Calculate the Burstein-Moss shift (∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 ) for n-type GaAs electron concentration of
1×1018 cm-3. Take the electron and hole effective masses to be me* = 0.067mo and
mh* = 0.45mo, respectively, where mo is the electron rest mass.

Solution:

2 𝑘𝑘𝐹𝐹2 2 𝑘𝑘𝐹𝐹2 ℏ2
∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝑔𝑔𝑔𝑔 = + = 3.807 × 10−20 𝑒𝑒𝑒𝑒𝑚𝑚2
2𝑚𝑚𝑒𝑒∗ 2𝑚𝑚ℎ∗ 𝟐𝟐𝑚𝑚𝑒𝑒
1
2 𝑘𝑘𝐹𝐹2 1 1 𝑘𝑘𝐹𝐹 = 3𝜋𝜋 2 𝑛𝑛 3
∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 = + = 3𝜋𝜋 2 × 1 × 1024 𝑚𝑚−3 1/3
2 𝑚𝑚𝑒𝑒∗ 𝑚𝑚ℎ∗
= 3.094 × 108 𝑚𝑚−1
−1
1.055 × 10−34 𝐽𝐽𝐽𝐽 2 1 1
∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 = 3𝜋𝜋 2 × 1 × 1024 𝑚𝑚−3 2/3 𝑚𝑚𝑟𝑟 = + = 0.0583
2 0.067 0.45
1 1 1
× +
0.067 0.45 9.11 × 10−31 𝑘𝑘𝑘𝑘 ∆𝐸𝐸𝐵𝐵𝐵𝐵𝐵𝐵 = 3.846 × 10−20 ×
3.094 × 108 2 /0.0583
= 1.0021 × 1020 𝐽𝐽 = 62.5 𝑚𝑚𝑚𝑚𝑚𝑚 = 62.5 𝑚𝑚𝑚𝑚𝑚𝑚

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

99
 Free carrier absorption

𝛼𝛼 ∝ 𝐸𝐸 −2

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

100
 Free carrier absorption
 Free carrier absorption (FCA) is a process in
semiconductors and other materials where the
absorption of photons leads to the excitation
of free electrons or holes within the
conduction or valence bands, respectively. This
phenomenon is distinct from band-to-band
absorption, where photons are absorbed to
promote electrons across the bandgap from
the valence band to the conduction band.
 Free carrier absorption is commonly
recognized as a parasitic phenomenon,
adversely affecting the performance of devices
like silicon solar cells, thermophotovoltaic
cells, and infrared photodetectors. However,
when its connection to free carrier density and
wavelength is accurately defined, free carrier
FCA
absorption can serve as a method to probe
electron-hole plasma (i.e., free carrier density). D.K. Schroder (1978)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

101
 Free carrier absorption
 The free carrier
absorption coefficient in
semiconductors is given
Intrinsic and free
by
carrier absorption
coefficients for a
𝑞𝑞 3 𝜆𝜆2 𝑝𝑝 function of
𝛼𝛼 = 2 3 ∗2 wavelength
4𝜋𝜋 𝜀𝜀0 𝑐𝑐 𝑛𝑛𝑚𝑚 𝜇𝜇

1.54 × 10−17 𝜆𝜆2 𝑝𝑝 −1

= 𝑐𝑐𝑐𝑐
𝑚𝑚∗2 𝜇𝜇

𝜆𝜆 = wavelength Free carrier

𝑝𝑝 = density of free carriers absorption
𝑛𝑛 = refractive index coefficient as a
function of hole
𝑚𝑚∗ = effective mass
concentration
𝜇𝜇 = mobility and wavelength
for p-Si.
H. Hara and Y. Nishi, “Free carrier absorption in p-type silicon,”
J. Phys. Sot. Japan, vol. 21, p. 1222, June 1966.
D.K. Schroder et al., “Free carrier absorption in silicon,”
IEEE J. Solid-state Circuits, vol. SC-13, p. 180, 1978.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

102
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

103
Photon density of states
 How many photon states are there in a box of
vacuum?
 Apply periodic boundary conditions to a cavity
with size 𝐿𝐿 × 𝐿𝐿 × 𝐿𝐿

𝟐𝟐𝜋𝜋 𝟐𝟐𝜋𝜋 𝟐𝟐𝜋𝜋 𝐿𝐿

𝑘𝑘𝑥𝑥 = 𝑛𝑛𝑥𝑥 , 𝑘𝑘𝑦𝑦 = 𝑛𝑛𝑦𝑦 , 𝑘𝑘𝑧𝑧 = 𝑛𝑛 𝐿𝐿
𝐿𝐿 𝐿𝐿 𝐿𝐿 𝑧𝑧

(𝑛𝑛𝑥𝑥 , 𝑛𝑛𝑦𝑦 , 𝑛𝑛𝑧𝑧 : integer) 𝐿𝐿

 States per unit volume per unit frequency

 Number of states between 𝑘𝑘 and 𝑘𝑘 + 𝑑𝑑𝑑𝑑 is
(𝜈𝜈 = 𝜔𝜔/2𝜋𝜋):
3 8𝜋𝜋𝜈𝜈 2 8𝜋𝜋
𝐿𝐿 𝑁𝑁 𝜈𝜈 = 3 = 3
𝑁𝑁 𝑘𝑘 𝑑𝑑𝑑𝑑 = 4𝜋𝜋𝑘𝑘 2 𝑑𝑑𝑑𝑑 ×2 𝑐𝑐 𝜈𝜈𝜆𝜆
2𝜋𝜋
 Similarly, for 2D cavity, we have
(2 accounts for two opposite polarization)
4𝜋𝜋𝜋𝜋 4𝜋𝜋
𝑁𝑁 𝜈𝜈 = = 2
 Since 𝜔𝜔 = 𝑐𝑐𝑐𝑐, we can derive that 𝑐𝑐 2 𝜈𝜈𝜆𝜆
 For the 1D case, it is given by
3
𝜔𝜔2
𝑁𝑁 𝜔𝜔 𝑑𝑑𝜔𝜔 = 𝐿𝐿 2 3 𝑑𝑑𝜔𝜔 4 4
𝜋𝜋 𝑐𝑐
𝑁𝑁 𝜈𝜈 = =
𝑐𝑐 𝜈𝜈𝜆𝜆

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

104
Photon density of states
 Number of states of photon per unit volume and per unit energy (3D):
8𝜋𝜋𝜈𝜈 2 8𝜋𝜋𝐸𝐸 2
𝑁𝑁 𝜈𝜈 = 3 → 𝑁𝑁 𝐸𝐸 = 3 3 3 𝐿𝐿3 𝑘𝑘 2
𝑐𝑐 ℎ 𝑐𝑐 4𝜋𝜋𝑘𝑘 2 𝑑𝑑𝑑𝑑
𝐿𝐿
×2 = 𝑑𝑑𝑑𝑑
2𝜋𝜋 𝜋𝜋2
 In a non-dispersive media with refractive index 𝑛𝑛: 𝑐𝑐
𝜔𝜔 = 𝑘𝑘
𝑛𝑛
8𝜋𝜋𝑛𝑛3 𝜈𝜈 2 8𝜋𝜋𝑛𝑛3 𝐸𝐸 2 𝑐𝑐 𝑐𝑐
𝑁𝑁 𝜈𝜈 = → 𝑁𝑁 𝐸𝐸 = 𝑑𝑑𝜔𝜔 = − 𝑘𝑘𝑘𝑘𝑘𝑘 + 𝑑𝑑𝑘𝑘
𝑐𝑐 3 ℎ3 𝑐𝑐 3 𝑛𝑛2 𝑛𝑛
 In a dispersive media with refractive index 𝑛𝑛(E): 𝑛𝑛 𝑘𝑘 𝑑𝑑𝑑𝑑
𝑑𝑑𝑑𝑑 = + 𝑑𝑑𝑑𝑑
𝑐𝑐 𝑛𝑛 𝑑𝑑𝑑𝑑
8𝜋𝜋𝑛𝑛3 𝜈𝜈 2 𝜈𝜈 𝑑𝑑𝑑𝑑 8𝜋𝜋𝑛𝑛3 𝐸𝐸 2 𝐸𝐸 𝑑𝑑𝑑𝑑
𝑁𝑁 𝜈𝜈 = 1 + → 𝑁𝑁 𝐸𝐸 = 1 +
𝑐𝑐 3 𝑛𝑛 𝑑𝑑𝜈𝜈 ℎ3 𝑐𝑐 3 𝑛𝑛 𝑑𝑑𝐸𝐸 𝑛𝑛 𝜔𝜔 𝑑𝑑𝑑𝑑
= 1+ 𝑑𝑑𝑑𝑑
𝑐𝑐 𝑛𝑛 𝑑𝑑𝑑𝑑
 Photon distribution function: 2𝜋𝜋𝑛𝑛 𝜈𝜈 𝑑𝑑𝑑𝑑
= 1+ 𝑑𝑑𝜈𝜈
1 𝑐𝑐 𝑛𝑛 𝑑𝑑𝜈𝜈
𝑛𝑛𝑝𝑝 𝐸𝐸, 𝑇𝑇 = (Bose–Einstein distribution)
𝑒𝑒 𝐸𝐸/𝑘𝑘𝐵𝐵 𝑇𝑇 −1

 Photon density distribution (number of photons per unit volume and per unit energy):

8𝜋𝜋𝑛𝑛3 𝐸𝐸 2 𝐸𝐸 𝑑𝑑𝑑𝑑 1 19 3 2
𝐸𝐸 𝑑𝑑𝑑𝑑 1
𝑁𝑁𝑝𝑝 = 3 3
1 + 𝐸𝐸/𝑘𝑘 𝑇𝑇
= 1.316 × 10 𝑛𝑛 𝐸𝐸 1 + 𝐸𝐸/𝑘𝑘 𝑇𝑇
𝑒𝑒𝑒𝑒 −3 𝑐𝑐𝑐𝑐−3
ℎ 𝑐𝑐 𝑛𝑛 𝑑𝑑𝐸𝐸 𝑒𝑒 𝐵𝐵 −1 𝑛𝑛 𝑑𝑑𝑑𝑑 𝑒𝑒 𝐵𝐵 −1

Comparison: electron density of states 𝐷𝐷 𝐸𝐸 = 6.812 × 1021 𝐸𝐸 𝑒𝑒𝑒𝑒 −1 𝑐𝑐𝑐𝑐−3

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

105
Optical transition in two-level system
 Electrons in atoms, molecules, and solids experience upward and downward transitions between
states with different energies, 𝐸𝐸𝑖𝑖 𝑖𝑖 = 1, 2, 3 …
 During the transition, energy is conserved: 𝐸𝐸𝑖𝑖 − 𝐸𝐸𝑗𝑗 = ℏ𝜔𝜔, 𝑖𝑖 > 𝑗𝑗
 There are two types of transitions, namely stimulated and spontaneous ones.
 The rate of stimulated transitions is proportional to the electron distributions in the given state and
to photon density 𝑁𝑁𝑝𝑝 , while the rate of spontaneous emission is only proportional to the excited
state population at high energy and unoccupied state at low energy.
𝐸𝐸2
Stimulated 𝐸𝐸2 − 𝐸𝐸1 = ℏ𝜔𝜔
(Stimulated) emission Spontaneous
1
absorption emission 𝑓𝑓1 =
1 + exp[(𝐸𝐸1 − 𝐸𝐸𝐹𝐹1 )/𝑘𝑘𝐵𝐵 𝑇𝑇]
𝐵𝐵21 𝑁𝑁𝑝𝑝 𝑓𝑓2 (1 − 𝑓𝑓1 ) 1
𝐵𝐵12 𝑁𝑁𝑝𝑝 𝑓𝑓1 (1 − 𝑓𝑓2 ) 𝐴𝐴21 𝑓𝑓2 (1 − 𝑓𝑓1 ) 𝑓𝑓2 =
1 + exp[(𝐸𝐸2 − 𝐸𝐸𝐹𝐹𝐹 )/𝑘𝑘𝐵𝐵 𝑇𝑇]
𝐸𝐸1

 At thermal equilibrium: 𝐵𝐵12 𝑁𝑁𝑝𝑝 𝑓𝑓1 1 − 𝑓𝑓2 = 𝐵𝐵21 𝑁𝑁𝑝𝑝 𝑓𝑓2 1 − 𝑓𝑓1 + 𝐴𝐴21 𝑓𝑓2 (1 − 𝑓𝑓1 ), and 𝐸𝐸𝐹𝐹1 = 𝐸𝐸𝐹𝐹𝐹
 This leads to
𝐴𝐴 and 𝐵𝐵 are Einstein coefficients for
𝐴𝐴21 8𝜋𝜋𝑛𝑛3 (𝐸𝐸2 −𝐸𝐸1 )2 𝐵𝐵12 spontaneous and stimulated transitions,
𝐵𝐵21
= ℎ3 𝑐𝑐 3 𝐵𝐵21
=1 respectively. They relate the spontaneous
emission rate to the absorption rate and
the stimulated emission rate.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

106
Conditions for stimulated emission
 Necessary conditions for stimulated emission to occur (spontaneous emission is
negligible) :
𝐵𝐵21 𝑁𝑁𝑝𝑝 𝑓𝑓2 1 − 𝑓𝑓1 > 𝐵𝐵12 𝑁𝑁𝑝𝑝 𝑓𝑓1 1 − 𝑓𝑓2 → 𝑓𝑓2 1 − 𝑓𝑓1 > 𝑓𝑓1 1 − 𝑓𝑓2 → 𝐸𝐸𝐹𝐹𝐹 − 𝐸𝐸𝐹𝐹1 > 𝐸𝐸2 − 𝐸𝐸1

 The above condition indicates that, for stimulated emission to occur efficiently, you
need more particles at higher energy states than in lower energy states. This situation is
called population inversion.
 For example, one can achieve population inversion in a three-level system by pumping
electrons from the ground state to an excited state. Rapid relaxation to a metastable
state with a longer lifetime allows accumulation of electrons, facilitating population
inversion which is essential for laser operation. Continuous pumping sustains this state,
enabling the coherent and intense light emission from the system.
 Population inversion can also be realized in semiconductors by either optical or
electrical pumping to have more electrons at the conduction band than holes at the
valence band (see next slides).

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

107
Stimulated emission in semiconductors
 For semiconductors, the respective transition rates are:
 Absorption rate 𝑁𝑁𝑣𝑣 : valence band density of states at 𝐸𝐸1
𝑅𝑅𝑎𝑎𝑎𝑎 = 𝐵𝐵12 𝑁𝑁𝑝𝑝 (𝐸𝐸2 − 𝐸𝐸1 )𝑁𝑁𝑣𝑣 (𝐸𝐸1 )𝑁𝑁𝑐𝑐 (𝐸𝐸2 )𝑓𝑓1 1 − 𝑓𝑓2 𝑁𝑁𝑐𝑐 : conduction band density of states at 𝐸𝐸2
 Spontaneous emission rate: 𝑓𝑓𝑖𝑖 : Fermi function evaluated at 𝐸𝐸𝑖𝑖

𝑅𝑅𝑠𝑠𝑠𝑠 = 𝐴𝐴21 𝑁𝑁𝑣𝑣 (𝐸𝐸1 )𝑁𝑁𝑐𝑐 (𝐸𝐸2 )𝑓𝑓2 1 − 𝑓𝑓1 1

𝑓𝑓1 =
1 + exp[(𝐸𝐸1 − 𝐸𝐸𝐹𝐹𝐹𝐹 )/𝑘𝑘𝐵𝐵 𝑇𝑇]
 Stimulated emission rate: 𝑓𝑓2 =
1
1 + exp[(𝐸𝐸2 − 𝐸𝐸𝐹𝐹𝐹𝐹 )/𝑘𝑘𝐵𝐵 𝑇𝑇]
𝑅𝑅𝑠𝑠𝑠𝑠 = 𝐵𝐵21 𝑁𝑁𝑝𝑝 (𝐸𝐸2 − 𝐸𝐸1 )𝑁𝑁𝑣𝑣 (𝐸𝐸1 )𝑁𝑁𝑐𝑐 (𝐸𝐸2 )𝑓𝑓2 1 − 𝑓𝑓1
8𝜋𝜋𝑛𝑛3 𝐸𝐸𝑖𝑖2 1
 At thermal equilibrium: 𝑅𝑅𝑎𝑎𝑎𝑎 = 𝑅𝑅𝑠𝑠𝑠𝑠 + 𝑅𝑅𝑠𝑠𝑠𝑠 , 𝐸𝐸𝐹𝐹𝐹𝐹 = 𝐸𝐸𝐹𝐹𝐹𝐹 , 𝑁𝑁𝑝𝑝 (𝐸𝐸𝑖𝑖 ) =
ℎ3 𝑐𝑐 3 𝑒𝑒 𝐸𝐸𝑖𝑖 /𝑘𝑘𝐵𝐵 𝑇𝑇 −1

𝐴𝐴21 8𝜋𝜋𝑛𝑛3 (𝐸𝐸2 −𝐸𝐸1 )2 𝐵𝐵12

 = , and =1
𝐵𝐵21 ℎ3 𝑐𝑐 3 𝐵𝐵21

 Conditions for stimulated emission to occur:

𝐵𝐵21 𝑁𝑁𝑝𝑝 (𝐸𝐸2 − 𝐸𝐸1 )𝑁𝑁𝑣𝑣 (𝐸𝐸1 )𝑁𝑁𝑐𝑐 (𝐸𝐸2 )𝑓𝑓2 1 − 𝑓𝑓1 > 𝐵𝐵12 𝑁𝑁𝑝𝑝 (𝐸𝐸2 − 𝐸𝐸1 )𝑁𝑁𝑣𝑣 (𝐸𝐸1 )𝑁𝑁𝑐𝑐 (𝐸𝐸2 )𝑓𝑓1 1 − 𝑓𝑓2
 𝑓𝑓2 1 − 𝑓𝑓1 > 𝑓𝑓1 1 − 𝑓𝑓2  𝐸𝐸𝐹𝐹𝐹𝐹 − 𝐸𝐸𝐹𝐹𝑣𝑣 > 𝐸𝐸2 − 𝐸𝐸1 = ℏ𝜔𝜔 ≥ 𝐸𝐸𝑔𝑔
 Therefore, we will have net optical gain, i.e., more stimulated emission than absorption, when
we have the quasi-Fermi levels separated by more than the band gap, which can be achieved
by either optical pumping or electrical injection. Under this condition, we will have more
electrons in the conduction band than electrons in the valence band (population inversion).
EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010
108
Stimulated emission in semiconductors
 Once population inversion is achieved, The absorption
coefficient will become negative, i.e., light will be amplified
with gain 𝑔𝑔 𝐸𝐸 = −𝛼𝛼 𝐸𝐸 .
 To the first order, the gain will be proportional to the excess
carrier density that exceeds the transparency level, i.e.,
𝑔𝑔 𝐸𝐸 ≅ 𝐺𝐺(𝑛𝑛 − 𝑛𝑛𝑡𝑡𝑡𝑡 ) (here, 𝑛𝑛𝑡𝑡𝑡𝑡 is the threshold carrier density
at which net absorption becomes zero)
 Apart from the gain in the active medium, we also need an
optical cavity to provide “positive optical feedback” so that
lasing can occur.
𝐿𝐿
 When light travels through the optical cavity with length 𝐿𝐿 and
two mirrors with a reflectivity of 𝑅𝑅1 and 𝑅𝑅2 , respectively, the
𝑅𝑅1
total optical loss will be: 𝑅𝑅2
1 1
𝛼𝛼𝑙𝑙 = 𝛼𝛼𝑖𝑖 + 𝑙𝑙𝑙𝑙 = 𝛼𝛼𝑖𝑖 + 𝛼𝛼𝑚𝑚
2𝐿𝐿 𝑅𝑅1 𝑅𝑅2
cleaved facets act as partially
reflecting mirrors
where 𝛼𝛼𝑙𝑙 is the internal loss in the active layer due to free-
carrier absorption and scattering and 𝛼𝛼𝑚𝑚 is the mirror loss.
 To achieve sustained oscillation (lasing), 𝑔𝑔 𝐸𝐸 must be larger 1 1
𝑔𝑔𝑡𝑡𝑡 = 𝛼𝛼𝑖𝑖 + 𝑙𝑙𝑙𝑙
than 𝛼𝛼𝑙𝑙 called threshold gain. 2𝐿𝐿 𝑅𝑅1 𝑅𝑅2

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

109
Stimulated emission in semiconductors
 Consider the fact that the carrier population is
in general proportional to the current, we may
write 𝑛𝑛 ≈ 𝐾𝐾𝑖𝑖𝐷𝐷 , where 𝑖𝑖𝐷𝐷 is the diode current.
 Based on the above equations, we may derive
the threshold current as
1 1 1 1
𝑖𝑖𝐷𝐷𝑡𝑡𝑡 = 𝛼𝛼𝑖𝑖 + 𝑙𝑙𝑙𝑙 + 𝑛𝑛𝑡𝑡𝑡𝑡
𝐾𝐾 𝐺𝐺 2𝐿𝐿 𝑅𝑅1 𝑅𝑅2

 Here, 𝐺𝐺 may be written as 𝐺𝐺 = 𝐺𝐺𝑖𝑖 Γ, where 𝐺𝐺𝑖𝑖

is determined by the material properties,
while Γ is the portion of G due to geometrical
factors (i.e., the overlap of the optical mode
and the active medium).

 To reduce 𝑖𝑖𝐷𝐷𝑡𝑡𝑡 , we have to

 increase 𝐾𝐾  increase injection efficiency
 increase 𝐺𝐺𝑖𝑖  improve material properties
 increase Γ  enhance optical confinement
 reduce 𝛼𝛼𝑖𝑖  reduce internal loss LD LED
 reduce 𝛼𝛼𝑚𝑚  improve cavity design
 reduce𝑛𝑛𝑡𝑡𝑡𝑡  reduce DOS of active region

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

110
Carrier recombination in semiconductors

 Carrier recombination is a process in which an

Electron
electron and a hole annihilate each other through Before
one or multiple steps and involving the creation of ∆𝐸𝐸
recombination
other particles (photon, phonon, or another Hole
electron/hole)

 Before recombination: Electron and hole Electron

occupy energy levels with an energy Recombination
difference of ∆𝐸𝐸. taking place
Hole
 Recombination: Electron and hole move
towards and annihilate each other.
Unoccupied
 After recombination: Both electron and state
hole disappear. The energy difference ∆𝐸𝐸 After
recombination
is released in the process, leading to the
Electron
emission of either photon (radiative
recombination), one or more phonons  Photon (radiative)
(non-radiative recombination), or another  One or more phonons
electron/hole with higher kinetic energy ∆𝐸𝐸 → (non-radiative)
(Auger recombination, also of non-  Another electron or hole
radiative nature). with higher energy
(Auger)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

111
Radiative recombination
 Band-to-band transition: the process in
which an electron in the conduction band
recombines with a hole in the valence
band, emitting light.
 Free-to-bound transition: it occurs when
an electron (hole) from the conduction
(a) (b) (c) (d) (e)
(valence) band is captured by localized
energy levels associated with an acceptor
(donor), leading to subsequent
recombination with a hole (electron) and
light emission.
 Donor-acceptor pair (DAP) transition: it (a): Band-to-Band Transitions
involves optical excitation of electrons
and holes, re-capture of electrons and (b) & (c): Free-to-bound transitions
holes by ionized donors and acceptors, (c): Donor-acceptor pair Transitions
and subsequent recombination of bound
electron and hole pairs, leading to light (d): Band-trap transitions
emission.
 Band-trap transition: recombination of Note: Not all transition occur in the same
trapped and free carriers material or under the same conditions.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

112
Non-radiative recombination
 Non-radiative recombination in
semiconductors refers to processes where
the excess energy of charge carriers is
𝐸𝐸𝐶𝐶
dissipated without the emission of a
photon. Major non-radiative processes 𝐸𝐸𝑇𝑇 X
include:
𝐸𝐸𝑉𝑉
 Shockley-Read-Hall (SRH) recombination: X
recombination through deep traps.
SRH
 Auger recombination: a high-energy
electron recombines with a hole,
transferring the excess energy to another
electron rather than emitting a photon.
 Surface recombination: charge carriers
recombine at the semiconductor surface,
where the presence of surface states or
traps facilitates non-radiative processes.
 Interface recombination: charge carriers
recombine at interfaces, where the
presence of interface states or traps
facilitates non-radiative processes.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

113
Low-level excitation / injection
 When excess carriers are injected into or optically excited in a semiconductor at
thermal equilibrium, the excess carriers will recombine with a recombination
rate that depends on the recombination mechanism and injection level.
Low-level injection
in p-type sample:
1018 𝑝𝑝𝑝𝑝𝑝 + ∆𝑝𝑝
𝑝𝑝𝑝𝑝𝑝 + ∆𝑝𝑝 𝑛𝑛𝑖𝑖 = 1010 𝑐𝑐𝑐𝑐−3
1016
𝑝𝑝𝑝𝑝𝑝 𝑝𝑝𝑝𝑝𝑝 = 1016 𝑐𝑐𝑐𝑐−3
𝑝𝑝𝑝𝑝𝑝 𝑛𝑛𝑝𝑝𝑝 + ∆𝑛𝑛
Majority carrier density 𝑛𝑛𝑝𝑝𝑝 = 104 𝑐𝑐𝑐𝑐−3
1014 remains almost constant
∆𝑛𝑛 = 1012 𝑐𝑐𝑐𝑐−3
12
10
𝑛𝑛𝑝𝑝𝑝 + ∆𝑛𝑛 ∆𝑝𝑝 = 1012 𝑐𝑐𝑐𝑐−3
1010 Minority carrier density
increases to be
108
comparable with High-level injection
Minority carrier density majority carrier density in p-type sample:
increases significantly,
106 but still much lower than 𝑛𝑛𝑖𝑖 = 1010 𝑐𝑐𝑐𝑐−3
majority carrier density
104 𝑝𝑝𝑝𝑝𝑝 = 1016 𝑐𝑐𝑐𝑐−3
𝑛𝑛𝑝𝑝𝑝 𝑛𝑛𝑝𝑝𝑝
2
𝑛𝑛𝑝𝑝𝑝 = 104 𝑐𝑐𝑐𝑐−3
10
Low-level injection High-level injection ∆𝑛𝑛 = 1016 𝑐𝑐𝑐𝑐−3
in p-type sample: in p-type sample: ∆𝑝𝑝 = 1016 𝑐𝑐𝑐𝑐−3

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

114
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

115
Band-to-band recombination
 At thermal equilibrium: 𝑅𝑅0 = 𝐺𝐺0 = 𝐵𝐵0 𝑛𝑛0 𝑝𝑝0
 Net recombination rate: 𝑅𝑅0 − 𝐺𝐺0 = 0 𝑅𝑅0 (𝐺𝐺0 ): Recombination
(generation) rate at
𝜕𝜕𝑛𝑛 𝜕𝜕𝑝𝑝 thermal equilibrium
 Under excitation, net recombination rate − or − is
𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡

𝜕𝜕𝑛𝑛 𝜕𝜕𝑝𝑝 𝐵𝐵: Recombination

= = −𝐵𝐵 𝑛𝑛𝑛𝑛 − 𝑛𝑛𝑖𝑖2 = −𝐵𝐵 𝑛𝑛0 + ∆𝑛𝑛 𝑝𝑝0 + ∆𝑝𝑝 − 𝑛𝑛𝑖𝑖2
𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡 coefficient under low
level excitation (≈ 𝐵𝐵0 )
 Under low-level injection/excitation, ∆𝑛𝑛∆𝑝𝑝 can be neglected:

i) p-type, 𝑛𝑛𝑝𝑝𝑝 ≪ ∆𝑛𝑛 ≈ ∆𝑝𝑝, 𝑝𝑝𝑝𝑝0 ≫ 𝑛𝑛𝑝𝑝0 → 𝑝𝑝𝑝𝑝 ≈ 𝑝𝑝𝑝𝑝𝑝 𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑝𝑝 ≈ ∆𝑛𝑛
1
𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 =
𝜕𝜕∆𝑛𝑛 ∆𝑛𝑛 𝐵𝐵𝑝𝑝𝑝𝑝0
= −𝐵𝐵 𝑛𝑛𝑝𝑝𝑝 + ∆𝑛𝑛 𝑝𝑝𝑝𝑝𝑝 + ∆𝑝𝑝 − 𝑛𝑛𝑖𝑖2 ≈ −𝐵𝐵𝑝𝑝𝑝𝑝0 ∆𝑛𝑛 = −
𝜕𝜕𝑡𝑡 𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟
𝑡𝑡 Minority carrier
−
∆𝑛𝑛 = ∆𝑛𝑛(𝑡𝑡 = 0)𝑒𝑒 𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 (electron) lifetime

ii) n-type, 𝑝𝑝𝑛𝑛𝑛 ≪ ∆𝑝𝑝 ≈ ∆𝑛𝑛, 𝑛𝑛𝑛𝑛0 ≫ 𝑝𝑝𝑛𝑛0 → 𝑛𝑛𝑛𝑛 ≈ 𝑛𝑛𝑛𝑛𝑛 𝑎𝑎𝑎𝑎𝑎𝑎 𝑝𝑝𝑛𝑛 ≈ ∆𝑝𝑝 1
𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 =
𝐵𝐵𝑛𝑛𝑛𝑛0
𝜕𝜕∆𝑝𝑝 ∆𝑝𝑝
= −𝐵𝐵 𝑛𝑛𝑛𝑛𝑛 + ∆𝑛𝑛 𝑝𝑝𝑛𝑛𝑛 + ∆𝑝𝑝 − 𝑛𝑛𝑖𝑖2 ≈ −𝐵𝐵𝑛𝑛𝑛𝑛0 ∆𝑝𝑝 = −
𝜕𝜕𝑡𝑡 𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 Minority carrier
−
𝑡𝑡 (hole) lifetime
∆𝑝𝑝 = ∆𝑝𝑝(𝑡𝑡 = 0)𝑒𝑒 𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

116
Band-to-band recombination

p-type, 𝑛𝑛𝑝𝑝𝑝 ≪ ∆𝑛𝑛 ≈ ∆𝑝𝑝, 𝑝𝑝𝑝𝑝0 ≫ 𝑛𝑛𝑝𝑝0 → 𝑝𝑝𝑝𝑝 ≈ 𝑝𝑝𝑝𝑝𝑝 𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑝𝑝 ≈ ∆𝑛𝑛

Source: E. Fred Schubert, 2023 (ISBN: 978-0-9 863826-6-6)

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

117
Shockley-Read-Hall (SRH) recombination
In general, the SRH recombination process is described by

𝜕𝜕𝑛𝑛 𝜕𝜕𝑝𝑝 − 𝑛𝑛𝑛𝑛 − 𝑛𝑛𝑖𝑖2

= =
𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡 𝜏𝜏𝑝𝑝 𝑛𝑛 + 𝑛𝑛1 + 𝜏𝜏𝑛𝑛 𝑝𝑝 + 𝑝𝑝1

Under low-level injection/excitation, ∆𝑛𝑛∆𝑝𝑝 can be neglected:

1
𝜏𝜏𝑛𝑛 =
i) p-type, 𝑝𝑝𝑝𝑝 ≈ 𝑝𝑝𝑝𝑝𝑝 , 𝑛𝑛𝑝𝑝 ≈ ∆𝑛𝑛, 𝑝𝑝 + 𝑝𝑝1 ≫ 𝑛𝑛 + 𝑛𝑛1 , 𝑝𝑝 ≫ 𝑝𝑝1 𝐶𝐶𝑛𝑛 𝑁𝑁𝑇𝑇

𝜕𝜕∆𝑛𝑛 ∆𝑛𝑛 1
=− 𝜏𝜏𝑝𝑝 =
𝜕𝜕𝑡𝑡 𝜏𝜏𝑛𝑛 𝐶𝐶𝑝𝑝 𝑁𝑁𝑇𝑇
𝑡𝑡
−
∆𝑛𝑛 = ∆𝑛𝑛(𝑡𝑡 = 0)𝑒𝑒 𝜏𝜏𝑛𝑛
(𝐸𝐸𝑇𝑇 −𝐸𝐸𝑖𝑖 )
𝑛𝑛1 = 𝑛𝑛𝑖𝑖 𝑒𝑒 𝑘𝑘𝐵𝐵 𝑇𝑇
ii) n-type, 𝑛𝑛𝑛𝑛 ≈ 𝑛𝑛𝑛𝑛𝑛 , 𝑝𝑝𝑛𝑛 ≈ ∆𝑝𝑝, 𝑛𝑛 + 𝑛𝑛1 ≫ p + 𝑝𝑝1 , 𝑛𝑛 ≫ 𝑛𝑛1
−(𝐸𝐸𝑇𝑇 −𝐸𝐸𝑖𝑖 )
𝑝𝑝1 = 𝑛𝑛𝑖𝑖 𝑒𝑒 𝑘𝑘𝐵𝐵 𝑇𝑇
𝜕𝜕∆𝑝𝑝 ∆𝑝𝑝
=−
𝜕𝜕𝑡𝑡 𝜏𝜏𝑝𝑝
−
𝑡𝑡 𝑁𝑁𝑇𝑇 : trap density
𝜏𝜏𝑝𝑝
∆𝑝𝑝 = ∆𝑝𝑝(𝑡𝑡 = 0)𝑒𝑒 𝐶𝐶𝑛𝑛(𝑝𝑝) : recombination
coefficient

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

118
Auger recombination
For Auger recombination, we have

𝜕𝜕𝑛𝑛 𝜕𝜕𝑝𝑝
= = −𝐴𝐴𝑛𝑛 𝑛𝑛 𝑛𝑛𝑛𝑛 − 𝑛𝑛𝑖𝑖2 − 𝐴𝐴𝑝𝑝 𝑝𝑝 𝑛𝑛𝑛𝑛 − 𝑛𝑛𝑖𝑖2
𝜕𝜕𝑡𝑡 𝜕𝜕𝑡𝑡

Under low-level injection/excitation, ∆𝑛𝑛∆𝑝𝑝 can be neglected:

i) p-type, 𝑛𝑛𝑝𝑝𝑝 ≪ ∆𝑛𝑛 ≈ ∆𝑝𝑝, 𝑝𝑝𝑝𝑝0 ≫ 𝑛𝑛𝑝𝑝0 → 𝑝𝑝𝑝𝑝 ≈ 𝑝𝑝𝑝𝑝𝑝 𝑎𝑎𝑎𝑎𝑎𝑎 𝑛𝑛𝑝𝑝 ≈ ∆𝑛𝑛

𝜕𝜕∆𝑛𝑛 ∆𝑛𝑛 1
= −𝐴𝐴𝑝𝑝 𝑝𝑝𝑝𝑝𝑝 𝑛𝑛𝑝𝑝𝑝 + ∆𝑛𝑛 𝑝𝑝𝑝𝑝𝑝 + ∆𝑝𝑝 − 𝑛𝑛𝑖𝑖2 ≈ −𝐴𝐴𝑝𝑝 𝑝𝑝𝑝𝑝𝑝
2
∆𝑛𝑛 = − 𝜏𝜏𝑝𝑝 = 2
𝜕𝜕𝑡𝑡 𝜏𝜏𝑝𝑝 𝐴𝐴𝑝𝑝 𝑝𝑝𝑝𝑝𝑝
𝑡𝑡
−
𝜏𝜏𝑝𝑝
∆𝑛𝑛 = ∆𝑛𝑛(𝑡𝑡 = 0)𝑒𝑒

ii) n-type, 𝑝𝑝𝑛𝑛𝑛 ≪ ∆𝑝𝑝 ≈ ∆𝑛𝑛, 𝑛𝑛𝑛𝑛0 ≫ 𝑝𝑝𝑛𝑛0 → 𝑛𝑛𝑛𝑛 ≈ 𝑛𝑛𝑛𝑛𝑛 𝑎𝑎𝑎𝑎𝑎𝑎 𝑝𝑝𝑛𝑛 ≈ ∆𝑝𝑝
1
𝜕𝜕∆𝑝𝑝 ∆𝑝𝑝 𝜏𝜏𝑛𝑛 = 2
= −𝐴𝐴𝑛𝑛 𝑛𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛𝑛 + ∆𝑛𝑛 𝑝𝑝𝑛𝑛𝑛 + ∆𝑝𝑝 − 𝑛𝑛𝑖𝑖2 ≈ −𝐴𝐴𝑛𝑛 𝑛𝑛𝑛𝑛0
2
∆𝑝𝑝 = − 𝐴𝐴𝑛𝑛 𝑛𝑛𝑛𝑛𝑛
𝜕𝜕𝑡𝑡 𝜏𝜏𝑛𝑛
𝑡𝑡
−
∆𝑝𝑝 = ∆𝑝𝑝(𝑡𝑡 = 𝜏𝜏
0)𝑒𝑒 𝑛𝑛

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

119
Recombination lifetime

 The overall recombination lifetime is determined by all three recombination

processes, i.e.,
1
𝜏𝜏𝑟𝑟 = −1 −1 −1
𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 + 𝜏𝜏𝑆𝑆𝑆𝑆𝑆𝑆 + 𝜏𝜏𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴

 For ultrathin films, both surface and interface recombinations have to be

included.

 As the SRH and Auger are non-radiative processes, the radiative efficiency
may be written as

−1
𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 1
𝜂𝜂𝑟𝑟𝑟𝑟𝑟𝑟 = −1 −1 = 𝜏𝜏
𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 + 𝜏𝜏𝑛𝑛𝑛𝑛 1 + 𝑟𝑟𝑟𝑟𝑟𝑟
𝜏𝜏𝑛𝑛𝑛𝑛

1
𝜏𝜏𝑛𝑛𝑛𝑛 = −1 −1
𝜏𝜏𝑆𝑆𝑆𝑆𝑆𝑆 + 𝜏𝜏𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

120
 Example:
A very short laser pulse at 780 nm is incident on a thick GaAs sample which has
an absorption coefficient of α = 1.5×106 m-1 at this wavelength. The pulse has
an energy of E = 10 nJ and is focused to a circular spot of radius 100 µm. 𝜆𝜆 = 780 𝑛𝑛𝑛𝑛
(a) Calculate the initial carrier density at the front of the sample.
(b) If the radiative and non-radiative lifetimes of the samples are 1 ns and GaAs
8 ns, respectively, calculate the time taken for the carrier density to drop
to 50% of the initial value.
(c) Calculate the total number of luminescent photons generated by each laser
pulse.
𝑥𝑥
Solution:
10×10−9 ×6.24×1018
(a) Number of photons per unit area: 𝜎𝜎𝑝𝑝𝑝 = 1.24 = 1.25 × 1013 /𝑐𝑐𝑐𝑐2
0.78
×3.14×(100×10−4 )2

𝐼𝐼 1−𝑒𝑒 −𝛼𝛼∆𝑥𝑥
𝐼𝐼 𝑥𝑥 = 𝐼𝐼(0)𝑒𝑒 −𝛼𝛼𝛼𝛼  light absorbed per unit length near the surface = 0 = 𝛼𝛼 (when ∆𝑥𝑥 → 0)
∆𝑥𝑥

Number of photons absorbed per unit volume near the surface: 𝑛𝑛𝑝𝑝𝑝 = 𝛼𝛼𝛼𝛼𝑝𝑝𝑝 = 1.875 × 1018 /𝑐𝑐𝑐𝑐3
Carrier density at the front of the sample: 𝑛𝑛 = 𝑝𝑝 = 1.875 × 1018 /𝑐𝑐𝑐𝑐3
1 1 −1 1 1 −1
(b) Effective recombination time: 𝜏𝜏 = + = −9 + = 0.9 × 10−9 s = 0.9 ns
𝜏𝜏𝑟𝑟 𝜏𝜏𝑛𝑛𝑛𝑛 1×10 8×10−9
−𝑡𝑡 𝑛𝑛 𝑡𝑡 1 −𝑡𝑡
𝑛𝑛 𝑡𝑡 = 𝑛𝑛𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒 → = = 𝑒𝑒𝑒𝑒𝑒𝑒 → 𝑡𝑡 = 0.62 ns
𝜏𝜏 𝑛𝑛𝑜𝑜 2 0.9×10−9
10×10−9 ×6.24×1018 1
(c) 𝑁𝑁𝑝𝑝𝑝 = 1.24 × = 3.50 × 1010 photon
1+1/8
0.78

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

121
Luminescence
 Luminescence occurs when an object absorbs energy from an external source, such
as light or electricity, and then re-emits that energy in the form of (visible) light.

 Electroluminescence:
Electroluminescence Photoluminescence
Electroluminescence is the process
in which a material emits light
when it is excited by an electric
current.
 Photoluminescence:
Photoluminescence is the process
in which a material absorbs
photons (light) and re-emits the
absorbed energy in the form of
light (typically with smaller photon Thermoluminescence Cathodoluminescence

energy).
Ec
Thermal
 Thermoluminescence: Emission of trapping
stimulation
light when a material is heated. electron trap
Recombination ω
 Cathodoluminescence: Emission of centre
light when a material is bombarded Ev
by an electron beam.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

122
Photoluminescence (PL)
 The PL involves the following three main processes:
E
 Excitation: Absorption of
Conduction
photons by a semiconductor
band
excites electrons from the
valence band to conduction
band, creating electron-hole
pairs.
 Thermalization: The excited
electrons and holes lose excess ω ≥ E g

Thermalization
energy and reach the respective PL
edges of the conduction and
valence bands in a very short Recombination Excitation
time (~ fs – ps range).
k
 Recombination: Electrons and
holes will recombine to release
excess energy in the form of
light (radiative recombination) Valence
or other forms such as heat band
(non-radiative recombination).

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

123
Photoluminescence measurement setup

 A typical experimental setup for photoluminescence

measurement involves:

 Light Source: A laser or lamp for controlled irradiation.

 Chopper/modulator: modulate excitation source.
 Optical Components: Lenses, mirrors, and filters for
precise light control.
 Sample Holder: Holds the material in the path of the
excitation light.
 Detector: Captures emitted light (photoluminescence).
 Monochromator/Grating: Separates emitted light into
different wavelengths.
 Data Acquisition System: Processes and records the
detected signal (including lock-in amplifier).
 Temperature Control: Maintains a controlled
temperature, if necessary.
 Shielding: Light-tight chamber to prevent ambient
interference.
 Software: Controls parameters, acquires data, and
facilitates analysis.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

124
Band-to-band transition
 Photoluminescence is a phenomenon where a material absorbs photons
and then re-emits them as light. PL spectrum refers to the distribution
of emitted light intensity across different wavelengths or energy levels.
 The luminescent intensity as a function of photon energy can written as
(a) (b) (c) (d) (e)

𝐼𝐼 ℏ𝜔𝜔 ∝ M 2 g ℏ𝜔𝜔 𝑓𝑓𝑒𝑒 𝑓𝑓ℎ

M 2 is the Matrix 𝑔𝑔 ℏ𝜔𝜔 is the joint density of 𝑓𝑓𝑒𝑒 and 𝑓𝑓ℎ represent the quasi-
element that describes states. For a direct bandgap equilibrium Fermi-Dirac distribution
the electric field dipole semiconductor with gap 𝐸𝐸𝑔𝑔 , function for the electrons and holes,
interaction between it is given by respectively. For low photoexcitation
the external light wave 3/2
density (quasi-Fermi levels are far away
and electron. As a 2 𝑚𝑚𝑟𝑟∗ 1/2 from the conduction/valance band
𝑔𝑔 ℏ𝜔𝜔 = 2 2 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔
crude approximation, 𝜋𝜋 ℏ edge), 𝑓𝑓𝑒𝑒 and 𝑓𝑓ℎ can be approximated by
it is given by Boltzmann distributions:

2
1 −𝐸𝐸𝑒𝑒 −𝐸𝐸ℎ
M ≈ 𝑓𝑓𝑒𝑒 ≈ 𝑒𝑒𝑒𝑒𝑒𝑒 , 𝑓𝑓ℎ ≈ 𝑒𝑒𝑒𝑒𝑒𝑒
𝜏𝜏𝑟𝑟 𝑓𝑓𝑒𝑒 𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇

𝐸𝐸𝑒𝑒 + 𝐸𝐸ℎ ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔

𝑓𝑓ℎ 𝑓𝑓𝑒𝑒 𝑓𝑓ℎ ≈ 𝑒𝑒𝑒𝑒𝑒𝑒 − = 𝑒𝑒𝑒𝑒𝑒𝑒 −
𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

125
Band-to-band transition
 Therefore, the PL spectrum is approximately given by

Example
3/2
1 2 𝑚𝑚𝑟𝑟∗ 1/2 ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔
𝐼𝐼 ℏ𝜔𝜔 ∝ ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔 exp −
𝜏𝜏𝑟𝑟 𝜋𝜋 2 ℏ2 𝑘𝑘𝐵𝐵 𝑇𝑇

Boltzmann 1/2
~ ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔
distribution
PL intensity

Theoretical
emission
spectrum
2.5kBT ~ 3kBT

𝐸𝐸𝑔𝑔 𝐸𝐸𝑔𝑔 + 𝑘𝑘𝐵𝐵 𝑇𝑇/2 ω (eV)

Experimentally measured
band-to-Band PL spectrum for
 Full Width at Half Maximum (FWHM): 2.5𝑘𝑘𝐵𝐵 𝑇𝑇~3𝑘𝑘𝐵𝐵 𝑇𝑇 n-type InSb at 4.2 K.
 Peak energy: 𝐸𝐸𝑔𝑔 + 𝑘𝑘𝐵𝐵 𝑇𝑇/2

E. Fred Schubert, 2023, ISBN: 978-0-9 863826-6-6

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

126
Example:
The intensity of a photoluminescence (PL) emission spectrum is given as

1/2 ω − 𝐸𝐸𝑔𝑔
𝐼𝐼 ω = 𝐴𝐴 ω − 𝐸𝐸𝑔𝑔 𝑒𝑒𝑒𝑒𝑒𝑒 −
𝑘𝑘𝐵𝐵 𝑇𝑇

where A is a constant, ω is the photon energy, Eg is the bandgap energy, kB is the Boltzmann constant and T
is temperature.

(i) Find an expression for ω at the maximum of the intensity spectrum.

(ii) Given the bandgap of GaAs is Eg = 1.42 eV, calculate the ω (in eV) at its maximum intensity at 300K.
(iii) Given that the FHWM of the PL emission spectrum is ΔE= 1.8kBT. Calculate Δλ (in nm) of the emission
spectrum for GaAs at the wavelength λ = 870 nm at T = 300 K.

Solution:
1 1
𝑑𝑑𝑑𝑑 1 − ω−𝐸𝐸𝑔𝑔 −1 ω−𝐸𝐸𝑔𝑔
(i) At maximum, = 𝐴𝐴 ω − 𝐸𝐸𝑔𝑔 2 𝑒𝑒𝑒𝑒𝑒𝑒 − + A ω − 𝐸𝐸𝑔𝑔 2 𝑒𝑒𝑒𝑒𝑒𝑒 − =0
𝑑𝑑 ω 2 𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇 𝑘𝑘𝐵𝐵 𝑇𝑇

1 1
1 − 1
 ω − 𝐸𝐸𝑔𝑔 2 = ω − 𝐸𝐸𝑔𝑔 2  ω = 𝐸𝐸𝑔𝑔 + 𝑘𝑘𝐵𝐵 𝑇𝑇/2
2 𝑘𝑘𝐵𝐵 𝑇𝑇

(ii) At 300K, 𝑘𝑘𝐵𝐵 𝑇𝑇 =0.0259 eV, ω =1.42+0.0259/2 =1.433 eV.

1.24 𝑑𝑑 ω 1.24 𝜆𝜆2 0.872

(iii) ω =  = − 2  ∆𝜆𝜆 = ∆ ω × = 1.8 × 0.0259 × = 0.02834𝜇𝜇𝜇𝜇 = 28.34 𝑛𝑛𝑛𝑛
𝜆𝜆 𝑑𝑑λ 𝜆𝜆 1.24 1.24

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

127
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

128
Temperature dependence of PL spectra
1/2
 When temperature rises, bandgap will decrease ℏ𝜔𝜔 − 𝐸𝐸𝑔𝑔  red shift in PL peak
energy (here, red shift means PL line moving to lower energy or longer wavelength).
ℏ𝜔𝜔−𝐸𝐸𝑔𝑔
 In addition, PL peak also broadens due to wider energy distribution exp −
𝑘𝑘𝐵𝐵 𝑇𝑇

ω (eV) ω (eV)
1.55 1.50 1.46 1.42 1.38 1.63 1.59 1.55 1.51 1.48 1.44 1.41 1.38

𝑘𝑘𝐵𝐵 𝑇𝑇 (233K)
PL Intenisty (arb. unit)

𝐸𝐸𝑔𝑔 𝑇𝑇 + 300 K 400 K

2
(298K)
200 K

(358K)

775 800 825 850 875 900

λ (nm) 𝜆𝜆(𝑛𝑛𝑛𝑛)

Theoretical PL emission spectra of GaAs Measured PL emission spectra of an AlGaAs alloy

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

129
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

130
Free-to-bound transitions

 Consider the PL from a p-type sample containing acceptors.

At low photoexcitation the density number of free electrons
created in the conduction band is much smaller than
acceptor density 𝑁𝑁𝐴𝐴 . These free electrons can recombine
radiatively with the holes trapped on the acceptors. (a) (b) (c) (d) (e)

 Such a transition, involving a free carrier (an electron in this

case) and a charge (a hole in this case) bound to an impurity,
are known as free-to-bound transitions. The emitted photon
energy is:
ℏ𝜔𝜔 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐴𝐴
Ec
Here, 𝐸𝐸𝐴𝐴 is the acceptor binging energy.
ℏ𝜔𝜔 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐴𝐴
 The emission due to free-to-bound transitions is a simple
way of measuring impurity binding energies.
EA
 The radiative recombination rates for free-to-bound Ev
transitions is proportional to 𝑁𝑁𝐴𝐴 .
 The peak in the emission spectra red shifts with increase in
dopant concentration due to bandgap narrowing.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

131
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

132
Donor-acceptor pair transition
 Some semiconductors may contain both donors and
acceptors. Such semiconductors are said to be compensated
because, under equilibrium conditions, some of the
electrons from the donors will be captured (or compensated)
by the acceptors. As a result, a compensated sample
contains both ionized donors (D+) and acceptors (A-).
(a) (b) (c) (d) (e)
 By optical excitation, electrons and holes can be created in
the CB and VB, respectively. These carriers can then be
trapped at the D+ and A- sites to produce neutral Do and Ao
centers. In returning to equilibrium some of the electrons on
the neutral donors will recombine radiatively with holes on
the neutral acceptors. This process is known as donor- Ec
acceptor pairs (DAP) transition. It can be represented by the 𝐷𝐷 𝑜𝑜

reaction:
ℏ𝜔𝜔
𝐷𝐷𝑜𝑜 + 𝐴𝐴𝑜𝑜 → ℏ𝜔𝜔 + 𝐷𝐷+ + 𝐴𝐴−
 The photon energy due to DAP transition is 𝐴𝐴𝑜𝑜
𝑒𝑒 2 Ev
ℏ𝜔𝜔 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐷𝐷 − 𝐸𝐸𝐴𝐴 +
4𝜋𝜋𝜋𝜋𝜋𝜋
𝑒𝑒 2
Here, is the Coulomb energy (the Coulomb interaction between the ionized donors
4𝜋𝜋𝜋𝜋𝜋𝜋
and acceptors. R is the distance between D+ and A-, and 𝜀𝜀 is the dielectric constant.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

133
Donor-acceptor pair transition
 Since R between ionized impurities is
determined by the lattice constant, which
takes discrete values, the DAP transitions
produce a series of sharp peaks converging
towards the photon energy 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐷𝐷 − 𝐸𝐸𝐴𝐴
(𝑅𝑅 → ∞).
 The best and most carefully studied
examples of DAP transitions are found in
GaP. Since this is a binary compound there
should be two different ways to distribute
substitutional donors and acceptors on its
sublattices.
 Type I: donors and acceptors are located on
 Type I DAP spectra in GaP due to SP-SiP, TeP-SiP
the same sublattice (e.g., SP-SiP, TeP-SiP, SeP-
pairs measured at 1.6K. One can determine the
SiP, or SiGa-ZnGa)
energy EA+ED (taking the known low temperature
 Type II: donors and acceptors are located on values of the indirect bandgap energy of GaP to
different sublattice (e.g., SP-ZnGa or OP-CdGa). be 2.339 ± 0.001 eV and 𝜀𝜀 = 10.75𝜀𝜀0 )
 The integers above the discrete peaks are shell
Here, XY means atom X replaces atom Y
numbers of the pairs which have been identified
e.g., SP: S replaces P to form a donor
by comparison with theoretical plots.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

134
Hydrogen atom Positively charged proton Wannier-Mott exciton
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

𝐸𝐸𝐵𝐵3𝐷𝐷 = − 2 𝑛𝑛2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐵𝐵 = 𝑚𝑚∗
𝜀𝜀𝑟𝑟 𝑟𝑟
𝑚𝑚𝑒𝑒∗
- - Note:𝑚𝑚𝑟𝑟∗ is dimensionless
+
P
𝜀𝜀𝑟𝑟 𝜀𝜀0
D-A pair

𝑚𝑚𝑒𝑒∗ 1 𝜀𝜀𝑟𝑟
𝐸𝐸𝑑𝑑 = 2 2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐷𝐷 = ∗ 𝑎𝑎0 (for donor) 𝑅𝑅
𝜀𝜀𝑟𝑟 𝑛𝑛 𝑚𝑚𝑒𝑒 + -
𝑚𝑚ℎ∗ 1 𝜀𝜀𝑟𝑟
𝐸𝐸𝑎𝑎 = 2 2 × 13.6 𝑒𝑒𝑒𝑒 and 𝑎𝑎𝐴𝐴 = ∗ 𝑎𝑎0 (for acceptor)
𝜀𝜀𝑟𝑟 𝑛𝑛 𝑚𝑚ℎ
𝑒𝑒 2
Note: here 𝑚𝑚𝑒𝑒∗ and 𝑚𝑚ℎ∗ are normalized to 𝑚𝑚0 ℏ𝜔𝜔 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐴𝐴 − 𝐸𝐸𝐷𝐷 +
4𝜋𝜋𝜀𝜀𝑟𝑟 𝜀𝜀0 𝑅𝑅
135
Example:
Calculate the photon energy of a shallow donor to shallow acceptor transition in Ge
semiconductor if the donor and acceptor impurity separation is 40 Å. The donor and
acceptor energy levels at ground states of Ge are 9.3 meV and 10 meV, respectively.
The bandgap of Ge is 0.66 eV and its relative dielectric constant 𝜀𝜀𝑟𝑟 of Ge is 16.2.

Solution:
The photon energy due to DAP transition is Alternatively, we can use vacuum
permittivity in different unit to simplify
𝑒𝑒 2 the calculation:
ℏ𝜔𝜔 = 𝐸𝐸𝑔𝑔 − 𝐸𝐸𝐷𝐷 − 𝐸𝐸𝐴𝐴 +
4𝜋𝜋𝜀𝜀𝑟𝑟 𝜀𝜀0 𝑅𝑅 𝜀𝜀0 = 8.854 × 10−12 𝐹𝐹 � 𝑚𝑚−1
= 55.263 𝑒𝑒 2 (𝑒𝑒𝑒𝑒)−1 � 𝜇𝜇𝜇𝜇−1
𝑒𝑒 2 (1.6 × 10−19 )2
= J
4𝜋𝜋𝜀𝜀𝑟𝑟 𝜀𝜀0 𝑅𝑅 4 × 𝜋𝜋 × 16.2 × 8.854 × 10−12 × 4 × 10−9
𝑒𝑒 2 1
=
4𝜋𝜋𝜀𝜀𝑟𝑟 𝜀𝜀0 𝑅𝑅 4 × 3.14 × 16.2 × 55.263 × 40 × 10−4
3.551 × 10−21
= eV
1.6 × 10−19
= 0.0222 eV = 22.2 meV

= 0.0222 eV = 22.2 meV

ℏ𝜔𝜔 = 0.66 − 0.01 − 0.0093 + 0.022 = 0.6627 eV
ℏ𝜔𝜔 = 0.66 − 0.01 − 0.0093 + 0.022 = 0.6627 eV

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

136
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

137
Optical band-trap transition
 Sometimes, electrons and holes captured by
traps can lead to emission of light. Such kind (a) (b) (c) (d) (e)

of transition is called optical band-trap

transition (or radiative capture of carriers by
localized traps).
 A representative example is N-doped GaP in E small 𝛥𝛥𝛥𝛥
which the N atoms form isoelectronic traps Ec
50 meV
50 meV below the conduction band edge.
Real
ℏ
 Due to uncertainty principle, Δ𝑥𝑥𝑥𝑥𝑥 ≥ a , Eg = 2.26 eV ω space
2
localized trap provide a widespread of Ev
momentum in k-space, which helps to
promote indirect transition. x
E
 GaP is an indirect semiconductor with a
bandgap of 2.26 eV. The band-trap large 𝛥𝛥𝛥𝛥
transition produces photons with an energy
of ℏ𝜔𝜔 = 2.26 − 0.05 = 2.21 eV, Momentum
ω space
corresponding to a wavelength of 561 nm
(green light). k

 This has applications in green LED.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

138
 Outline of Lecture 7-9

 Introduction  Radiative transition

 Light and light ‒ Simple two-level system
absorption/reflection/transmission ‒ Band-to-band transition
 Absorption in pure/undoped ‒ Temperature dependence of
semiconductors bandgap
‒ Direct transition ‒ Free-to-bound transitions
‒ Indirect transition ‒ Donor-acceptor pair transitions
‒ Excitonic effect ‒ Optical band-trap transition
‒ Absorption by Carrier Bound to ‒ Exciton recombination
Shallow donors and Acceptors
 Doping and disordering effect
‒ Mott transition
‒ Mott trailing
‒ Burstein-Moss shift
‒ Bandgap narrowing
‒ Free carrier absorption

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

139
Excitonic emission
 Excitonic emission in
semiconductors refers to the Luminescence decay processes stimulated by above-
process by which excitons, which bandgap photoexcitation in a semiconductor
are bound electron-hole pairs,
recombine and emit energy in the
form of light. In PL experiments on
high purity and high-quality
semiconductors at low
temperatures, we expect the
photoexcited electrons and holes
to be attracted to each other by
Coulomb interaction and to form
excitons. As a results, the
emission spectra should be
dominated by free exciton
recombination.
 When the sample contains a small
number of donors or acceptors in
their neutral state the excitons
will be attracted to these
impurities. This attraction lowers
the exciton energy, neutral
impurities are very efficient at
trapping excitons to form bound
excitons at low temperature.

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

140
Example of excitonic emission

 PL of GaAs at 2K. The inset is

an expansion of the spectra Expansion
within the rectangle labelled
X. It contains the part of the
emission spectrum associated
with free excitons.
 The spectrum in the inset
labelled (a) S-62 and those
labelled (b) S-39, (c)S-39
correspond to two different
samples.
 The spectrum (c) was excited
by light intensity ten times
higher than that used for
spectrum (b).
 The peak labelled (Do, X) is
attributed to recombination
of excitons bound to neutral
donors

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

141
Example:
A photoluminescence experiment is performed on a GaN sample at 10 K. A free
exciton peak and a bound exciton peak are observed at 3.484 eV and 3.476 eV,
respectively. Find the binding energies for the free exciton and the bound exciton.
Comapre free exciton binding energy with the calculated value. The band gap of
GaN is 3.504 eV at 10 K. The electron and hole effective masses of GaN are
0.15m0 and 0.8m0, respectively, where m0 is the rest mass of an electron. The
dielectric constant of GaN is 𝜀𝜀 = 8𝜀𝜀0 .
E
Solution:
The free exciton binding energy: εB = (3.504 –
3.484) eV = 20 meV

The bound exciton binding energy:

εBX = (3.484 – 3.476) eV = 8 meV εB

Eg = 3.504 eV
εBX
Calculated free exciton binding energy

3.484 eV

3.476 eV
𝑚𝑚𝑟𝑟∗ 1 0.126𝑚𝑚𝑜𝑜
𝜀𝜀𝐵𝐵 = −𝑅𝑅𝐻𝐻 = 13.6 = 26.8 meV
𝑚𝑚𝑜𝑜 𝜀𝜀𝑟𝑟2 𝑛𝑛2 𝑚𝑚𝑜𝑜 8 2

1 1 1
where ∗ = ∗ + ∗
k
m r me mh 0

EE5508 Semiconductor Fundamentals Lecture 7 - 9 Wu Yihong 09/09/2010

142