E7: 214 Optoelectronic Devices

January-April 2023
Instructor: Varun Raghunathan
Assistant Professor
ECE department
Indian Institute of Science
varunr@iisc.ac.in

Mode of Teaching: In-person, classroom: 1.08, ECE department

Timing: T,Th: 2-3:30 PM

TA: Lal Krishna A.S (lalkrishna@iisc.ac.in)

Course description:
The aim of this course is to provide students with an understanding of various semiconductor
based optoelectronic devices (light emitters, detectors, modulators etc.) with a perspective of
the application of such devices to building photonic systems. The expected outcomes from
the course are: (i) understand the basic working mechanism of the devices, (ii) understand the
governing equations to be able to perform calculations to characterize the performance of the
devices and, (iii) have the practical knowledge and an understanding of the trade-offs when
using these devices in their respective applications. It is expected that the students have the
required pre-requisite basic Maths background, especially under-graduate level Calculus and
have also completed an under-graduate level courses in Electro-magnetics and
semiconductors. This is an elective course offered within the microelectronics program in the
ECE department.

This course will cover the following teaching modules: introduction to optoelectronic devices
- top-down approach, quick refresher into semiconductors & electromagnetics, optical
processes in semiconductors, light sources (LEDs, semiconductor laser diodes), light
detectors (PMTs, photo-diodes, PIN, APD), sensor arrays (CCD,CMOS), Noise processes in
light generation & detection. A detailed course syllabus is given below.

Course website: Access to a Microsoft Teams class group will be provided. Notes and assignments
will be posted in the Teams group.

Textbooks:
[1] Photonics: Optical Electronics in Modern Communications, Amnon Yariv & Pochi Yeh
[2] Fundamentals of Photonics, B.E.A. Saleh & M.C. Teich
[3] Semiconductor Optoelectronic Devices, Pallab Bhattacharya
[4] Class notes

Evaluation components:
The following are the evaluation components for this course:
[1] Two assignment (early February & April): 15 % weightage each
[2] One Mid-term exam (early March): 30% weightage
[2] Final exam (June 1st and 2nd week): 40 % weightage
Detailed course Syllabus:
Lecture No. Description Text book
1 Basic overview of the course, course organization etc. --
2 Introduction to optoelectronics 4 & references
- Top-down approach - learning applications to motivate therein
learning about the devices
- Fiber-optic communications - long-haul, short-distance
and chip-scale
- Free-space optical communications - inter-satellite and
indoor Li-Fi
- Quantum communication - quantum key distribution
- Consumer electronics - cameras, TVs, smart, wearable
sensors
- Photonic components and basic terminology
3-6 Refresher into semiconductors and Electromagnetics 3, Chapters 2,
- Electronic properties of semiconductors 4
- P-N Junctions, I-V curves
- Hetero-junctions, quantum wells, quantum dots &
super-lattices
- Electromagnetic waves in media (isotopic/ anisotropic)
- Polarization property of light
- Reflection & refraction of light
- Dielectric function, Kramer-Kronig relationship
- Phase velocity, group velocity, dispersion
7,8 Optical Processes in semiconductors/ optical media 3, chapter 3
- Interaction of photons with matter 2, chapter 15
- Radiative and non-radiative processes 4
- Rates of absorption & emission
- Measurement techniques for various optical processes
9-13 Light sources: 3, chapter 7
- LEDs (working principles, spectral properties, output 1, chapter 15-16
power, modulation bandwidth) 2, chapter 16
- Basics of lasers (Einstein's relations, gain, population 1, Chapter 10
inversion, feedback mechanism, threshold, efficiency 2, Chapter 17
etc.) 4
- Semiconductor diode lasers (operating principles, gain
& population inversion, threshold current, output power,
efficiency, modulation bandwidth)
- Various types of semiconductor lasers (hetero-junctions,
quantum-well/ quantum-wire/ quantum-dot lasers, DBR/
DFB lasers, VCSEL)
- Noise processes in light generation:
Noise and its impact on measurement, various noise
processes, Laser line-width, Coherence & interference
14-17 Light detectors: 3, chapter 8
Operating principles and characteristics of: 1, chapter 11
- Photo-multipliers (PMTs) 2, chapter 17
- Photo-conductors 4
- Photo-diodes (PIN)
- Avalanche Photo-diodes (APD)
 - CCD and CMOS sensor array
18-20 Noise processes in light generation & detection 1, Chapter 10
- Noise and its impact on measurement 2, Chapter 17
- Laser linewidth, spontaneous vs. Stimulated emission 4
- Coherence and interference
- Various noise processes: Shot noise, Johnson noise, 1/f
noise, RIN noise etc.
- Deriving SNR and receiver sensitivity calculations,
- Heterodyne vs. homodyne detection
- Error probability calculations
 E7: 214
Optoelectronic Devices

Lecture: 2
Instructor: Dr. Varun Raghunathan,
ECE department

[1] Top-down approach to understanding photonic devices

[2] Electro-optic/ photonic systems
[3] Applications of optoelectronics
[4] Components & Terminology

February 27, 2021 1

 Top-down approach to understanding
photonic devices
• We draw motivation from the end applications to
learn about component level photonic devices
• Electro-optic/ photonic systems:
- Fiber optic communication
- Free-space optical communication
- Opto-electronic integrated circuits (OEICs)
- Consumer electronics:
Cameras, LED illumination, LED TV, CD players
- Your own photonics experiment set-up

February 27, 2021 2

 Fibre optic communication [1]

February 27, 2021 3

 Fibre optic communication [2]
Block diagram:
Amplifier:
Er-doped fiber amplifier (EDFA)
Raman amplifier
Semiconductor optical amplifier (SOA)

Transmitter Repeater Receiver

3R:
Semiconductor lasers Photodetectors (p-i-n or travelling wave)
Re-amplification
Modulators Re-shaping
Amplifiers
Wavelength division multiplexing Re-timing Electronic circuitry (clock & data recovery)
Channel spectral separation

Transmission windows
Transmission speeds, distance & technology

Source: wikipedia
Source: wikipedia

February 27, 2021 4

 Fibre optic communication [3]
Fibre communication at different length scales:
Fibre-to-the-home High end computing/ server farms

Verizon FiOS

Google’s server farm (11000 sq. ft) in Iowa, USA

February 27, 2021 5

 Fibre optic communication [4]

Founder Dr. K. Sivarajan, former IISc ECE Professor

Tejas Networks designs, develops and manufactures high-performance and future-ready optical
and data networking products that are sold to telecommunications service providers, web-scale
internet companies, utilities, large enterprises, defence and government entities in more than 60
countries. Tejas products are differentiated by a programmable, software-defined hardware™
architecture that provides flexibility, multi-generation support and a seamless software-enabled
network transformation to its customers.

February 27, 2021 6

 Optoelectronic integrated circuits (OEIC)

Source: Intel
• Optical communications is preferred over long distances (speed and bandwidth)
• Need for high data rates for every day applications (streaming video etc.)
• Even further: data transmission between computer boards, inter-chip and intra-chip (??)
• Advantage of optical over copper interconnect:
- Large bandwidth & high speed data transmission
- Immunity to mutual interference and cross-talk
- Freedom from capacitive loading effects
• With reduced distance, volume goes up (which is good) but,
February 27, 2021 price per unit has to be low (this is challenging) 7
 Optoelectronic integrated circuits (OEIC)
Optical PCB, IBM research

• Use light for inter-chip communication

• VCSELs at 985 nm
• Multimode polymer waveguides
• Array light source and detectors

February 27, 2021 Jeff Kash, IBM reserch 8

 Optoelectronic integrated circuits (OEIC)
CMOS technology nodes

Electronic-photonic IC

Computation capability of electronics

+
High speed of photonics

February 27, 2021 9

 Optoelectronic integrated circuits (OEIC)

February 27, 2021 10

 Optoelectronic integrated circuits (OEIC)
Individual components

Source: Intel

- Light source: Semiconductor lasers

- Guiding light: waveguides
- Modulators: Electro-optic, thermo-optic modulators
- Photo-detectors: p-i-n photodetectors, travelling-wave photodetectors
- Assembly, packaging, testing, reliability
February 27, 2021 11
 Optoelectronic integrated circuits (OEIC)
III-V OEICs

GaAs technology InP technology

Source: Yariv’s book

I. Ury, K.Y. Lau, N. Bar-Chaim, A. Yariv, “Very high frequency
GaAlAs laser-field effect transistor monolithic integrated
circuit”, Appl. Phys. Lett. 41 (1982) Source: Infinera

• Advantages of III-V OEICs:

Ability to build efficient photonic components (lasers, waveguides & detectors)
Ability to build high speed driver electronics (high carrier mobility)
• Infinera: 2.28 Tb/sec super-channel with 40 channels x 57 Gbps (PM-QPSK)

February 27, 2021 12

 Optoelectronic integrated circuits (OEIC)
III-V vs. Silicon OEICs

Source: Infinera

February 27, 2021 13

 Optoelectronic integrated circuits (OEIC)
Silicon OEICs – silicon photonics

Luxtera 100G – 2 Km reach Intel 100G – 2 Km reach

February 27, 2021 14

 Free-space optical communication [1]

• India’s MOM mission, using RF S-band communication to Indian deep space network
• Interest in free-space optical communications (between satellites, satellite-to-earth etc.)
• NASA, Moon to Earth ~ 400,000 Km, at 622 Mbps. Video rate ~ 20 Mbps.
• Components used for this:
-Telecom photonic components (around 1550 nm)
- Low power lasers with modulators and amplifiers
- Magnification optics
- Beam tracking, steering optics
- Adaptive optics for turbulence correction
February 27, 2021 - High sensitivity detectors (InGaAs APDs) 15
 Free-space optical communication [2]
Visible light communication (VLC) and Li-Fi

• Visible light communication: communication with visible light (transmitter, receiver, medium etc.)
• Li-Fi: Light fidelity, the communication protocol, atto-cells, handover etc.
• Advantages of VLC: high data rates, restricted space (privacy), use existing infrastructure
• Applications of VLC: communication, localization etc.
• Components:
- Light source (visible or Near-IR LEDs)
- Dectectors (photodiode to APDs to single-photon APDs)
- Optics (lens, scanning mirrors etc.)

February 27, 2021 16

 Quantum Communication [1]
Inter-continental QKD distribution (2017):
Quantum key distribution:

Quantum photonic link:

Single
Single Active/ passive Active/ passive
photon
photon photonics for photonics for
detectors +
sources manipulation manipulation
Electronics

Transmitter Receiver

February 27, 2021 17

 Quantum Communication [2]

Quantum dot emitters 2D material on etched pillars NV centers in diamond

 Quantum Communication [3]
Single photon detectors – PMT and APD

Super conducting single photon detector (Cryo-cooled)

 Consumer Electronics: Cameras

• 48 MP sensors have been incorporated into cell phones

• 12-18 MP are now routine in most phones
• CMOS sensors is the main driver for low cost, high resolution sensors
• CMOS sensors are replacing CCD sensors even in scientific applications

February 27, 2021 20

 Consumer Electronics: LEDs

• LEDs are slowly replacing lamps/ tube lights for lighting applications
• Key enabler: GaN LED technology
• White LED: Blue + Phosphor or RGB mixture
• Advantages of LEDs: less power consumption, longer life, compact sizes
• Issues to be solved: Not standardized, binning (variation), sharp illumination
• Metrics: Color temperature, lumen output

February 27, 2021 21

 Consumer Electronics: displays & TVs

• LEDs are used to build full scale 2D display for indoor/ outdoor applications
• Combination of R,G,B LEDs in Bayer pattern
• LEDs are also used as backlit for LCD TVs
- Better color rendering possible
February 27, 2021
- Fast refresh rates (important for gaming applications) 22
 Wearable Gadgets
Smart watch VR headset

IR & VIS sensors IR sensors

Heart rate, O2 saturation monitoring LCD display
February 27, 2021 Lenses for 3D viewing 23
 Recap: the basic components that
make a photonic system
Modulator
Light detectors
Light source +
De-modulator

Data
Options for light sources: Options for light detectors:
• Incoherent or coherent • Speed/ sensitivity
• LEDs, Lamps, Globar • PMTs, photodiodes
• Lasers: semiconductor, solid state Avalanche PDs,
fibre lasers, nonlinear optical Single photon counters

Options for modulators: The focus is mainly on

• Electro-optic semiconductor photonic devices:
• All-optical - Low cost,
• Acousto-optic - Large volume manufacturing,
• Speed, power consumption - Advanced functionality
February 27, 2021 - Widespread applications 24
 E7: 214
Optoelectronic Devices

Lecture: 3a
Instructor: Dr. Varun Raghunathan,
ECE department
Semiconductor Physics
[1] Electronic bands of solids
[2] Semiconductors – periodic table, Energy band-gap, laFce constant
[3] Crystal structure
[4] Strain effects
[5] E vs. k, direct and in-direct semiconductors

January 8, 2017 1
 Electronic bands in solids

Sea of free carriers Carriers elevated to Large band gap

(No restoring force) conducTon band due to Difficult to excite cariers
Plasmonics elevated T or doping (poor electrical conducTvity)
January 8, 2017 2
 968:()#5'(%)&$!

M.#'.&<!N2!CO>P! G!
 Some material properTes

January 8, 2017 4
 Semiconductors
La,ce constant and energy band-gap of various semiconductors

AlAs

• How to calculate laFce constant and band-gap of ternary and quaternary alloys ?
January 8, 2017 5
V.F(6!()#$%.#%!B)&!$68:()#5'(%)&!.@@)<$!

"""ER:%&:56$2!!
V.F(6!()#$%.#%!!
^!3#6&0<!A.#5E0.7!

M.#'.&<!N2!CO>P! ]!
3#6&0<!A.#5E0.7!)B!$68:()#5'(%)&!.@@)<$!

M.#'.&<!N2!CO>P! P!
 9%&.:#65!@.F(6!$%&'(%'&6!!
5>!

5C!

•!!4)16&6#%!^!:#E()16&6#%!$%&.:#!
•!!9%&.:#!8)5:_6$!%16!6@6(%&)#:(!^!
)7T(.@!7&)76&T6$!)B!$68:()#5'(%)&!

M.#'.&<!N2!CO>P! N!
 4)#(67%!)B!(&:T(.@!%1:(L#6$$!

U3a!:8.06!)B!67:E@.<6&$!Z:%1!5:$@)(.T)#$!
6/17.8-$*91.:0,44;$<.$

b)&!8)&6!6X.(%!%&6.%86#%!&6B6&!%)!&6B6&6#(6!:#!%16!()'&$6!Z6A$:%6!
M.#'.&<!N2!CO>P! `!
 4&<$%.@!$%&'(%'&6!)B!$68:()#5'(%)&$!
9(*#'/*:#;$*<=>>*$.08).80-?* @(""-0*('/()-$*<>02$.#"*5"#'-$?*

M.#'.&<!N2!CO>P! =98*$14$*9,$0>?@,/$<,041*A$+B$C1$8*+?4$10$8$C1-1.+0$D8B,/$E$ >O!

 Y6X.0)#.@!V.F(6!D!"""ER:%&:56$!
A80.B(.-*$.08).80-* !&"#0*$-%()&'/8).&0C*!(-B&-"-).0()*-D-).*

•!!"""ER:%&:56$!6X1:A:%!$7)#%.#6)'$!7)@.&:c.T)#!
!!!QA6(.'$6!)B!:)#:(!(1.&.(%6&!)B!%16!A)#5$S!!!!!!!!!
!!!.@)#0!%16!4E.X:$!
•!!9%&.:#!.@$)!&6$'@%$!:#!#6%!7)@.&:c.T)#!D!!
!!!;:6c)6@6(%&:(!6I6(%!
•!!;)@.&2!R)#E7)@.&!.#5!$68:E7)@.&!
•!!d.#5!T@%!&6$'@%$!:#!7))&!)K6&@.7!)B!(.&&:6&$!:#!;)@.&!
$%&'(%'&6!

M.#'.&<!N2!CO>P! >>!
 Crystal structure based on unit-cell
properTes

also called
trigonal

LiNbO3 GaN, AlN, InN GaAs, InP, Si, Ge

January 8, 2017 12
 Crystal structure: Linear & Nonlinear
opTcal properTes

• Linear opTcal properTes of crystal depend on

crystal symmetry
• Nonlinear opTcal properTes depend on the
crystal symmetry and the point group

January 8, 2017 13
 Band structure calculaTon for
semiconductors (techniques)
• Crystal structure: Electrons in the presence of a
periodic crystal potenTal
• Electrons in this periodic potenTal is equivalent
to free electrons with modified mas (effecTve
mass)
• Different techniques to solve:
- Tight binding approach
- Pseudo potenTal method
- Plane wave expansion method
- k.p method (popular, fast for close to band edge)

January 8, 2017 14
 3!K$,!L!&67&6$6#%.T)#!

•! 3#6&0<!A.#5!$%&'(%'&6!.@)#0!:87)&%.#%!$<886%&<!7@.#6$!
•! -.@6#(6!A.#5!%<7:(.@@<!1.$!.!7!%<76!(1.&.(%6&:$T($!
•! -.@6#(6!A.#5*!16.K<!1)@6$2!@:01%!1)@6$!.#5!$7:#E)&A:%!$7@:%!)B!
•! 4)#5'(T)#!A.#5!1.$!$!)&!7!%<76!(1.&.(%6&:$T($!
M.#'.&<!N2!CO>P! >J!
 Energy – momentum relaTonship

E vs. K diagram for direct

band-gap semiconductor

January 8, 2017 16
+:&6(%!^!:#E5:&6(%!A.#5!$68:()#5'(%)&$!

>,HC!6-!
6-

O,]]!6-!
O,]]! >,>C!6-!

•! f.g$!D!@)Z6$%!7.&%!)B!-d!.#5!4d!@:#6!'7!.%!h!7):#%!
•! 9:!^!f6!D!@)Z6$%!7.&%!)B!-d!.#5!4d!5)!#)%!@:#6!'7!
•! f6!D!:%!:$!7)$$:A@6!%)!8.L6!:%!5:&6(%!A.#5!0.7!'$:#0!$%&.:#!^!16.K<!5)7:#0!>P!
M.#'.&<!N2!CO>P!
 +:&6(%!^!:#E5:&6(%!A.#5!%&.#$:T)#!

,L$ G*
9.#.-*E*
9N$ 9.#.-*E*
,L$
9N$
9.#.-*F*
9M$
!$ 9.#.-*F*
9M$ O$

i3!j!3CE3>!j!1k! i3!j!3CE3>!j!1k!
iL!j!LC!D!L>!j!O! iL!j!LC!D!L>!l!O!
M.#'.&<!N2!CO>P! /6[':&6$!e2!5:W('@%!B)&!68:$$:)#! >N!
 3I6(%!)B!$%&.:#!)#!%16!A.#5!$%&'(%'&6!

•! 9%.:#!@)Z6&$!%16!$<886%&<!)B!%16!(&<$%.@!
•! 41.#06!:#!A.#50.72!6@6(%&)#!^!1)@6!6I6(TK6!8.$$6$!
•! 3I6(%!)B!A:.X:.@!$%&.:#!)#!K.@6#(6!A.#5*!
!!!!!!E!Y6.K<!1)@6!.#5!@:01%!1)@6!A.#5$!$1:m!
!!!!!!E!41.#06!:#!('&K.%'&6!)B!%16!A.#5$!Q:,6,!6I6(TK6!8.$$S!
•! P914$984$1?)+/*80*$1?)-1.87+04$10$4*/810,<$Q>80*>?LD,--$-84,/4$

M.#'.&<!N2!CO>P! >`!
 Topics not covered here
• Crystal/ semiconductor growth:
- Float zone, Czochralski, Bridgman technique
- Epitaxy: Molecular beam epitaxy (MBE)
Metallo-organic vapour phase epitaxy (MOVPE)
Liquid phase epitaxy (LPE)
- Chemical vapour deposiTon (CVD) – PECVD, LPCVD

• Device fabricaTon – deposiTon, litho, etching, metallizaTon etc.

• Band structure calculaTons in detail

January 8, 2017 20
 ~ 2.-) {1-)
\-:i..)
~hv f,_.,v ~ Arv~
\ \) ~ ,)
\ \)
~'""'~ #~~·dvv
~~ ~~°"--

·-, ~ -ti"""-'- ~-.J.. p= ~ l ..-..l'-"- _ ~ . ::i;,,.. ~

I I4
~~ ~ ~ Kl
c,.._,-.,,._

:~,N--- f-"""f>"~
~ . At ~
£ -tt---,t- f"-">~ u,,---

o.,,c r -j€..._ (c,-- ....._,J,; 0--... 1' c.,,,j-,,..,,__.,:,<- -"r- er~

~ ~~ ~ ~ h ~ ~ 1 n;:...._.&-<- 3
t.,._ "'-

r~ ~ ~ ti- ~ h, p~J!..
V._ / c.,._

o--,,1 'T J-,__,.J.

~ ~ J.a- ~~- .,.,,.,-<~ l,v\~ ~ - r-J,.. -'<- rJ, ~

~=-----~._ ,·a r·,
~ \1./ J.c \ 1J ~~ ·
,---..
-__ --14 ...-..J<. 'IT ~1 N,. ~...._ '._,, ( ) ~ ,,_.,, ·.

cA N >-
~
 Cao)
~ Nu

<f ~ )s l'"l-- f ( V) fo r--

--
i--

----
0)
e
A.,__,, -t BL.-, f(rv)

___ _ 0
i

j,yv/k-T
C - .1
( ~ ~ ,~c.e..) ,.p.._ "

£1'~~/4
t
A... , f-7T l. v 3.
-- c...3,. ~- ~~
J~ i
(A) B Co*~) -
~ )7....-
&_ ~ -'9.Y 1 ~
~if
[SL.- \
-
,-

o,
~ ~ ~d,) \Sh- =- Js2- J
~ ~ ~~ k
..5 r-- R r s ~ 1' ~1/47 •-- ~u.~ °'-A.,<.._ -

~ ~ / /;,__ ~~"'-- ITtiv1-i_ s~ ~

~¥~ q A~ Arf>~C~ ~~
-II' A~' c.--Jtic:...._,.l- ~ a-e,..__ ~ tt::.. t-~1--:c-,,, f ~~ f:.-,,
5 ~~ ~~ ..___ .

~ ~v7j(,~
fr,
--e-y ~~,,fyy~Je,~ ,

~r
'
,f 71~ (; ~ /<--<--. /<JgJ,fo,__,,e..;f'0 ~~ -f.--,, ~ L ~+,v,,-,.., ~
~ ~ .i {.,.,- ~ ~____,}n, ~ ~ C
~ z---<,_,....Q -7l<,.-..._)
,,,,,; It-, ~ ~- L;, "-'- J.;:. i,,:, ~ ,...., ~ :?J"-_,__./ ~ c,__ ,i--.-...+--.

~11~ .
~ ::r.'if""~ -ft,-,._ ~ - ~J-.-1 Co_,_ ~ &,)
_?.,

\~ l °l.-- ::.. ~1.,-\ ..:. ,_C_---=

&rrh v1.-c, .!--p
-
'K ~ ~ ~~r-8 ~ --tv,,;,____.__.,.L ~ ft;;_
flu.,
~<., ~ Jv,-..J ~~ ,____ J-/::, - ~ ~ ~o...-1s·~ ;.,....
~ ~L,,;.,,,;"'""""'- ) cl- L') G>,/'- - - : - / - ~ <.-- ~ "o rt:
~~J,,-< ~J./ ,fL.J- ;,, ~ ~ ft:.. .,!; ~---=-
~ ~ . J - u . , v 4iv:-
1
cg- It::....~-~ ~-~-- ...
-k 1N- t ~ ko---,_ r ~~ 1.:.? ;:;..,. kf,:~ ra-1t•- ~ ~,
~ 91'\ ~~ ~~ ~~ lL~1)l ~ LA-!E'IZl) ~ ~
t ~ ' h ~ L . . : - ~ ~ ~ h ' ~ (to.. , ~~ ~J,
)

_,. ---it--,.. ,J,-.--rv-(h·..,., / sn->" " ~ ~rn - ~ °'"' (j::,_, ,;.,, c.:~

~ ~ )0
~ ,ff...>' ~ ~7/ ( ~ ~ -1...~
• '

* "1k :,, ....,__w.v.. h, -tk b,,

~ - f' ;;..., -ie--,.._ UV'<- -a- ~ ._,,,,.,.:-1.:.,~~) 0-.--, f,,<l,,-w--1:

'P~ ~ _-:) P~~~~·na

r ~ l cp) ( f)
v~~ ~;°""',-) ( J ;~~- Hi)

[ _____): ?,, 1
< - c_

t, ~ h f' .

s~ir~f'X'
r~ ck~~
~
~~ -c
\
~ Ai\=- -C--<
~~ \4 +~c-.v--k r~~

Q~~ X ~!{~ ~
r~~1 ~ r~~~ s ~ °' ~ w~
~ ~·J
t~k-

\,,-<I \ ~ EL- E, -- hi)_)

£c...-Ev ~ ~ <]
 -- 0

Ez., ~ E ~ ii fC.. ~c.-f:~ ~ I~ ~ ~ ' o - - ~

t1 ;. c;~ ~ ft.-( ~ ~ ft,._ ~~ ~--of

- - -<D

_)
 Op-n~ .j o--i~ ~ i 1 ~~
'D~?J "3' s~ J f ( v) ..,; ft: .,_,L;..,e,, ftA,.. ~ ~ ~, kv

~ V\,Cu,,.......H ( ~ 1t:u,. ~~ -,, ~ tk..,() ~ ~ CJ""-.,(,;+,.' ~1-

~ 9.:. ,,,.,,~ ~ v-J"'--' ~ ~ a l ' ~ ' - ~
~ ~rJ.v...~ ~ *vt- ~ j ~-vvt ~21: ~ s~ i:..., krttt:::-

~(.....L ~ C?'--cl.J.,._J-i~ /e,P, cl.. •

~:Lt· LJ>----. ~ i.,... °'-if..........l ....... -It-..,__ -,,, ~ ~ e.,-- ~ p"'--'-~•

--f Uv t,v,.u\:: ~ ~ i~----J-. ~-~ ~ ~ ft.... -f~
~1 V +,, ·v-+ol'-' .

---
ryv, C,

~ t°{v) ::
14g">/,,_, :f--_
~~:

;.: ( 2 ~ : ) 3./.,_

2. T2...-;;- 3

( '2 'h-),: ) '3.h_ • •

11
2- 1f ,_-j,(' (_ EV - £I) '--'
 f(v)

Dp+i~ J·~y-.,t ~~
d)-s~

£ ~ .s s I ev-- ~ ,h·~ : E )... L/.) ~- ~ """'tt... ~ ~

~

EI V) ~~ (f;~ ~·t~ ~

'..,.__ A-bso-v7>-n ~ c.~~~ ·. E. 2- V) ~~

-\-
£, ~ f , ~ w\'tt.. ~NY\,,

--

* 1k ~ h'o--. h, ~~ ~ ~ . S S l•~ ~ fe_ ( V) ;;,-£(VJ .

~ Vl so.::h,,s ~uJ. ~ EFc.. - . e-Fv 7 Av

u....S~~-~ ~ ~ - A A ~ ~ 7 ~~.
,h.L ~ ~ s .s:-,: ~ ~ ~s~·~ ~ [ ~ °' +v.-.-~-~ 1-
f~o) ~ Ls<- c__J~ Cl/)

Offl"·~
;x ,._ - .,_
JU'""lv'\1\.-
~"d 1-
/5 ~ ·

v)
--

7,"-- JJv) f(v)

gir~"<

g-,r LY
~ fe.. k f-c... ~N- ~ 1 ' ~ - ~ ~.> ~'- ~ E,F-v

,; Y.sjo - ~ ~ ~~ ~ "'-"" LE:!) ~s.ri~ svvt/1,1.,1.N-.. .

'1 Y_,.l:; - ~ t,.<._ '--"<-<{ i-<> ~~ q>'-~ 1- erpt,;o.J.~'fv.,
~ ,..-~.➔ ~k~"1J ~ ~ . SJ,<--~of}-~~ -
 Af +e~ ~~b-Yi~ )

~ -:.,, ~ "'---"- ~,_;_ ~ \.,_,v.J__ .

fe.Cv) ::- -F(Ei,) (1 - f(E,) J

Ass~ ~""" - oU ""'<- J..,'.,, ,.,;, ~ " " ' £-0-,.. be-. "'-ppµ,~;~ '?J
~}~"' ~h,,~~ .
__ (£\.,,- E+)/e,_ T ( _ (E 1 - EF) /t._ ' )
f ,_ (v) ,;::; e I - e,

- lE L - ~ ,) / k T - t,...-v /~I
,-: e. ::- e

,.....__

h-v
~
~
·7·~ ~ ~ E J Lk ~ ~~ ~ ,~ ~ 1 , ~

~_rs,G""') ~ Vl 7 cJrs~l-1' ~ ~ .
.
LL, Y.st(vJ - --y°'-b(-v) > a

"o/ /\ 2- {' ( v) [ :fe ( v) - -f'.,Jv) ]fv7 °

g-TI LY
1k i" ~ ; , , . ~ ·.__ ~ fe.-,x ~ ""- ~ o!;y lA

J)
r:- "11').,. '\ "\....
W1 ~ D -- '1 2-
- h- -'e---r
--
'h')y-

2-
/1

w~ Mlt Mvi&-~t- ~ o-.b(J\N_ ~~~ ~~~

1.,-,e_ +--J..k ~ 91'ft'W ~j,c.,J.·- ...,., <,.A_ ,,, ~ ~
wb~ ~~Joy ~vs
~~ J-r~-~ r/1-,D~ ~ ~ ~~6,-,~ ,

f e. ( v) - f°'- ( v) -= fc LE '2-) - -F v ( £,) ~ O

~ ~ ~ ~ ~----- ~ ~~c.,f,ri v-/1 tt ~

~~ L:..~ .

<>< cv ) =- D C1-. v - E:o) '4.. [ f Ct,) - F ct 2-) J

~ /~ +e-..t_ °'-}o~A-Ox,'~·c-... ) f (f I J ~ j_ J

Ai:,s7o-Y> ~-~~ (oG.A,(_~ ~ - 1--o- ~ ) ; ~ " '-~ )

~ ~ ,'/\~ '>-Vl.O'Y\.~~ &,./)·~ 1'" \ ~ ,('~·

J:.+ ~c..tJJ.A ~ ~ ~ k . ~ ~ h:, _!_

yl., -w,- ---
An,1A/),._.d;_,.,_L{
A ~NL ~~ ~r~°'"c.k ~ i'""~ t ~ i<-e.-
i ,,Jr--<.v<..._,/,j = 1 ~ V: 1-C-, -f-w_ ~~ ~ c..,__ "---" o._ , '

r h,..,c~ ,,.._, ,._° r - ~ , ~ h fl.;,__ .S ~~ ~ -- ~

uJ_ J,,.A,·'c) -fk c/J..r ~ , _;_ " ~ fr-, ft;_,_ f-~ i h~

~~ ,·" ~ F~ ,, ~~ h> c , ; , . 1 _ ~
-"t ~ t,;""" ~ ,,.__,_..t ~ '-'-- ff.,..,,_), J I ~· h °"' ~
J: ,.,,.,,~ .

'Jlv.»- v, "' cJ..,A,.,,.;..~- "{ ~ °1 /" ,._,o"' J,.. o/',..,.,,,__ ;,.., ~ n,O • "

1 IJ,,J-fA ~0 ', -+"-:x j- ~

 (3) Monolithic Full-RGB LED using III-Nitride material system

From: [1] H.E. Ghoroury et. al. "Growth of monolithic full-color GaN based with
intermediate carrier blocking layer AIP Advances (2016).
[2] M.V. Kisin et. al. "Simulation of a full colour III-Nitride RGB LED", NUSOD 2016.

• InGaN with increasing Indium helps achieve longer wavelength emission

• Suitable Electron-blocking layer and intermediate electron-blocking layers required
to achieve carrier confinement
• With increasing current emission wavelength shifts from red to green to blue and
intermediate mixing colours also observed
Emission colour with increasing current
Typical band diagram of a InGaN multi-quantum well LED (single colour)

Energy band diagram of a multi-quantum well LED (RGB emission)

Dynamic Range
 Lecture #14
Theory of Epitaxy: Lattice misfit and imperfections
¾ Equilibrium Theory of Lattice Misfit (Merve, Matthews‐Blakeslee)
e.g., Epitaxial GexSi1‐x films exhibit interesting strain‐induced
modulations (of the film surface)
¾ Other manifestations in Epitaxial Ge‐rich GexSi1‐x films
(heavy compressive strain) : 2D to 3D transition
¾ Defects in Thin Films

Text Book: The Materials Science of Thin Films, by Milton Ohring

2/16/2016 PYL116‐2016‐L14 1
 Equilibrium Theory of Lattice Misfit
‰ Frank and van der Merwe: Any epitaxial layer having a lattice‐parameter
mismatch (with the substrate) < ~9% would grow pseudomorphically
Initially, very thin films strain elastically to have a0(f)=a0(S), making the
interface coherent with atoms on either side lining up
‰ Above a critical thickness, the rising total elastic strain energy will finally
exceed the energy associated with a relaxed structure consisting of an array
of misfit dislocations separated by wide regions of relatively good fit.
‰ At this point, the initially strained film ideally decomposes to this relaxed
structure where a portion of the misfit is relieved by dislocations
‰ As the film continues to grow, more misfit is relieved until at infinite
thickness the elastic strain is totally eliminated
‰ In epitaxial growth without interdiffusion, pseudomorphism exists only up
to some critical film thickness dc beyond which dislocations are introduced
2/16/2016 PYL116‐2016‐L14 2
Matthews‐Blakeslee equilibrium theory of misfit‐dislocation formation
‰ For simplicity, we assume that film and
substrate have the same Young's (Y)
modulus and same shear (μ) modulus
‰ The analysis is divided into two film
thickness (d) regimes, d < dc and d > dc

In the early stage of film growth,

elastic strain energy per unit area Ee
(J/m2) increases with d as
⎛ J ⎞ Yd ε 2
Ee⎜ 2 ⎟ =
ε = biaxial elastic strain, ⎝m ⎠ (1 − ν )
ν = Poisson's ratio

Formation of dislocations at the interface≡

a means of relieving the elastic strain
2/16/2016 PYL116‐2016‐L14 3
‰ If dislocations are assumed to arrayed in
square grid of side S on the film, the
elastic strain ε in the film is reduced from
its initial misfit value (f) to, ε = f − b
S
(b=magnitude of Burger’s vector)
‰ The quantity b/S ∝ no. of dislocations at
the substrate interface (Note: b//S)
(when b/S = f, the strain ε vanishes)
‰ Each dislocation‐plane threads the entire film and extends the
lateral film length by the Burger’s vector magnitude, b

‰ Now, dislocation energy per unit length E μb 2 ⎛ R0 ⎞

= ln⎜ ⎟
of dislocation line is given by (theory): length 4π (1 − v) ⎝ b ⎠
R0 = radius about the dislocation outside which the strain field vanishes
‰ Since R0 scales as d, R0 = βd, where β = constant of the order of unity
2/16/2016 PYL116‐2016‐L14 4
(β ~ 0.70‐2.72)
 ⎛R ⎞ ⎛ βd ⎞
∴ ln⎜ 0 ⎟ = ln⎜ ⎟ Thus, the dislocation energy per unit film area (E ) is,
⎝ b ⎠ ⎝ b ⎠ d

⎛ J ⎞ ⎛ E ⎞ ⎛ disloc. length ⎞ μb 2 ⎛ βd ⎞ ⎛ 2 ⎞
∴ Ed ⎜ 2 ⎟ = ⎜⎜ ⎟⎟ × ⎜ ⎟= ln⎜ ⎟×⎜ ⎟
⎝ m ⎠ ⎝ disloc. length ⎠ ⎝ Film Area ⎠ 4π (1 − v) ⎝ b ⎠ ⎝ S ⎠
(Since, dislocation length per unit square area of grid~2S/S2 = 2/S)
(Note: b//S)
‰ Total strain energy per unit area (ET) is now sum of two energies;

ET (J/m2)= elastic strain energy/area (Ee ) + dislocation energy/area (Ed)

Yd ( f − b / S ) 2 μb 2 ⎛ 2 ⎞ ⎛ βd ⎞
∴ ET = + ⎜ ⎟ ln⎜ ⎟
(1 − v) 4π (1 − v) ⎝ S ⎠ ⎝ b ⎠
‰ Physically, elastic strain energy/area (1st term) increases linearly with d
‰ In contrast, dislocation energy/area (2nd term) is nearly constant with
only a weak logarithmic dependence on d (arising from R0)
2/16/2016 PYL116‐2016‐L14 5
Re‐writing, 2
Yd ⎡ ⎛ b ⎞⎤ μb ⎛ b ⎞ ⎛ βd ⎞
slightly ET = ⎢ f − ⎜ ⎟⎥ + ⎜ ⎟ ln⎜ ⎟
differently, (1 − v) ⎣ ⎝ S ⎠⎦ 2π (1 − v) ⎝ S ⎠ ⎝ b ⎠
‰ With regards to no. of dislocations (∝b/S) the 2 terms vary differently
‰ Only above a critical thickness, dc, the ET
introduction of misfit dislocations lead to a Ed
decrease in total energy ET
Ee
‰ The equilibrium state of the system can be
determined by finding the conditions for
which ET reaches an absolute minimum with
respect to the number of misfit dislocations ET
per unit length, 1/S.
Ed
‰ The fact that there is a minimum in ET at a Ee
nonzero 1/S value reveals that the structure
is in mechanical equilibrium only if ∂ET
=0
dislocations are present ⇒ ⎛1⎞
2/16/2016 PYL116‐2016‐L14 ∂⎜ ⎟ 6
⎝S⎠
 ∂ET
=
(− 2b )Yd ⎛ b⎞ μb 2 ⎛ βd ⎞
⎜ f − ⎟+ log⎜ ⎟=0
⎛1⎞
∂⎜ ⎟
(1 −ν ) ⎝ S ⎠ 2π (1 −ν ) ⎝ b ⎠
⎝S⎠
‰ Now imposing ‘no dislocation’ condition, i.e., b/S = 0 & putting d=dc
⎛ Y ⎞
μb ⎛ βd c ⎞ b ⎛ β d c ⎞ ⎜Q μ = ⎟
dc = ln⎜ ⎟ or dc = ln⎜ ⎟ ⎜ 2(1 + v ) ⎟
4πYf ⎝ b ⎠ 8π (1 + v) f ⎝ b ⎠ ⎜ ⇒ Y = 2μ (1 + v) ⎟
⎝ ⎠
‰ Film will be pseudomorphic (coherent) until a critical thickness dc
‰ For films with thickness d > dc, misfit dislocations appear
‰ The dc can be controlled by varying misfit parameter f
‰ By doping Si wafers with B (aB < aSi slightly) the lattice parameter of
the substrate can be controllably reduced. This provides opportunity
to study defect generation in subsequently deposited epitaxial films
‰ An increase in f resulted in an increase in misfit dislocation density
2/16/2016 PYL116‐2016‐L14 7
 Equilibrium Theory of Lattice Misfit
‰ Defect free (but lattice GexSi1_x/Si epitaxial films
strained) epitaxial films
dislocation‐relaxed
(commensurate) can be
(discommensurate)
obtained with thickness d Lattice‐ epitaxy
which is lesser than dc as per strained
the Mattews theory defect‐
free‐
‰ It is also seen that extended (commen
surate)
dislocation arrays do not form epitaxy
instantaneously with well‐
defined spacings; rather,
dislocations nucleate
individually
(The reason being : the strained
films are not in equilibrium)

2/16/2016 PYL116‐2016‐L14
f ∝ x (Ge − fraction) 8
 Epitaxial GexSi1‐x films exhibit interesting strain‐induced
modulations in surface morphology Actual
Model ⇓ ⇓ ⇓ ⇓
↓ ↓ ↓

‰ Surface ripples arise because the film is under compressive stress,

since a0(Ge‐Si solid solution film) > a0(Si‐substrate)
‰ Thus the lattice‐plane spacing of the film shrinks near the cusplike
troughs (shown as ↓ )and expands at the rounded peaks (⇓)
(Note: Flat film‐surface expected corr. to the minimum energy
configuration)
‰ However, when relieval of film strain‐energy outweighs the
tendency of surface energy γ (which tend to smooth/flatten out
the film) the roughening occurs
2/16/2016 PYL116‐2016‐L14 9
 Other manifestations in Epitaxial GexSi1‐x films
(heavy compressive strain) : 2D to 3D transition
‰ When significant compressive stress exist,
since a0(Ge70Si30 solid solution film) >> a0(Si‐substrate)
: Another manifestation of strain relaxation occurs in compressed
epitaxial film

SEM’s top‐view of faceted

pyramid and dome‐
shaped islands formed
during growth of a 40 nm
thick Si0.7Ge0.3 alloy film
on (001) Si

‰ This is in the form of transition from a planar film to 3D‐ morphology

‰ The result is faceted dome‐shaped islands (undesirable for devices)
2/16/2016 PYL116‐2016‐L14 10
 Schematic illustration of crystal defects in epitaxial films
1, Threading edge
dislocations;
2, interfacial misfit
dislocations;
3, threading screw
dislocation;
4, growth spiral;
5, stacking fault in film;
6, stacking fault in
substrate;
7, oval defect;
8, hillock;
‰ Minimization of the defects in increasingly
9, precipitate or void
densely packed ICs will require enormous
efforts, first to find the origin of defect and
then to devise methods for their elimination
2/16/2016 PYL116‐2016‐L14 11
2/16/2016 PYL116‐2016‐L14 12
2/16/2016 PYL116‐2016‐L14 13
 Stacking Faults

A‐layer

2/16/2016 PYL116‐2016‐L14 14
Background Story of
the Invention of
Efficient Blue InGaN
Light Emitting Diodes
SHUJI NAKAMURA
S O L I D S TAT E L I G H T I N G A N D E N E R G Y E L E C T R O N I C S
CENTER
M AT E R I A L S A N D E C E D E PA RT M E N T S
U N I V E R S I T Y O F C A L I F O R N I A , S A N TA B A R B A R A ,
U.S.A.

2014 NOBEL LECTURE IN PHYSICS

 Outline

1) Introduction: What is an LED?

2) Material of Choice: ZnSe vs. GaN

3) The Beginning: GaN on Sapphire

4) Enabling the LED: InGaN

5) Historical Perspective
The LED
ENERGY EFFICIENT WHITE LIGHT
 What is an LED?

A Light Emitting Diode (LED) produces light of a single color by

combining holes and electrons in a semiconductor.

Source of Holes
(p-type Layer)

External Light Out

Source of
Electrons
(Battery) Combining of
Holes and Electrons
(Active / Emitting Layer)

Substrate Source of Electrons

(Foundation) (n-type Layer)
 What is an LED?

A Light Emitting Diode (LED) produces light of a single color by

combining holes and electrons in a semiconductor.
Actual Blue LED Packaged Blue LED

Size: 0.4 mm x 0.4 mm

 White LED: Combining Colors

White Light: Blue + Other colors (red, yellow, green)

Other Colors: Convert Blue LED Light to Yellow using Phosphor.

Blue LED Phosphor White Light White LED

Convert: = Blue + Yellow
Blue → Yellow

S. Pimputkar et al., Nature Photonics 3 (2009) 180—182

 Applications for InGaN-Based LEDs

Solid State Lighting Decorative Lighting Automobile Lighting

Displays Agriculture Indoor Lighting

 Energy Savings Impact
~ 40 % Electricity Savings (261 TWh) in USA in 2030 due to LEDs
Eliminates the need for 30+ 1000 MW Power Plants by 2030
Avoids Generating ~ 185 million tons of CO2

Sources: www.nobelprize.org, US Department of Energy

1980s:
ZnSe vs. GaN
II-VI VS . III-N IN THE LATE ‘80S
 Candidates for Blue LEDs: ZnSe vs.
GaN
Semiconductors that possess the required properties
to efficiently generate blue light: ZnSe and GaN

BUT … How does one create ZnSe / GaN?

Single crystal growth of material on top of different, available single crystal:

Dislocation /
Defect

ZnSe GaN

Al2O3
GaAs
(Sapphire)

0 % Lattice Mismatch 16 % Lattice Mismatch

Few Dislocations (Defects) Significant Dislocations (Defects)
 GaN on Sapphire: Heavily Defected

Too many
Dislocations/
Defects

GaN

Sapphire
1 µm (Al2O3)
Cross section Transmission Electron Microscope (TEM) of GaN on Sapphire, F. Wu et al., UCSB
 1989: ZnSe vs. GaN for Blue LED

ZnSe on GaAs Substrate

◦ High Crystal Quality: Dislocation density < 1x103 cm-2
◦ Very Active Research: > 99 % of researchers

GaN on Sapphire Substrate

◦ Poor Crystal Quality: Dislocation density > 1x109 cm-2
◦ Little Research: < 1 % of researchers

Interest at 1992 JSAP Conference:

◦ ZnSe – Great Interest: ~ 500 Audience
◦ GaN – Little Interest: < 10 Audience
◦ GaN Actively Discouraged:
◦ “GaN has no future”
◦ “GaN people have to move to ZnSe material”
 1989: Starting Point of Research

Seeking to get Ph.D. by writing papers

◦ Very few papers written for GaN
◦ Great topic to publish lots of papers!

Working at a small company:

◦ Small Budget
◦ One Researcher

Commonly accepted in 1970s—1980s:

◦ LEDs need dislocation density < 1x103 cm-2

Never thought I could invent blue LED using GaN…

Development of
GaN
G A N MATURES
 MOCVD GaN before 1990s

MOCVD Reactor MOCVD System:

◦ High carrier gas velocity:
~ 4.25 m/s
◦ Poor uniformity
◦ Poor scalability
◦ Poor reproducibility
◦ Poor control

AlN Buffer Layers:

◦ Crack free GaN growth
◦ High Structural Quality GaN

But …
◦ Al causes significant problems
in MOCVD reactor, undesired
H. Amano, N. Sawaki, I. Akasaki, Y. Toyoda,
Appl. Phys. Lett., 48 (1986) 353—355
 Invention: Two-Flow MOCVD
1991: S. Nakamura et al., Appl. Phys. Lett., 58 (1991) 2021—2023

Invention of Two-Flow MOCVD System

(MOCVD: Metal-Organic Chemical Vapor Deposition)
Reproducible, uniform, high quality GaN growth possible
Low carrier gas velocity: ~ 1 m/s

Schematic of Two-Flow MOCVD Main Breakthrough:

Subflow to gently “push” gases down
and improve thermal boundary layer
 First MOCVD GaN Buffer Layer
1991: S. Nakamura, Jpn. J. Appl. Phys., 30 (1991) L1705—L1707

GaN Buffer Layer on Hall Mobility vs. GaN Thickness

Sapphire substrate:
High Quality GaN Growth
Smooth and Flat Surface
over 2” Substrate

Highest Hall mobilities

reported to date:
No Buffer: 50 cm2/V s
AlN Buffer: 450 cm2/V s
Two- No Buffer: 200 cm2/V s
Flow GaN Buffer: 600 cm2/V s
 Passivation of p-type GaN
1992: S. Nakamura et al., Jpn. J. Appl. Phys., 31 (1992) L139—L142
1992: S. Nakamura et al., Jpn. J. Appl. Phys., 31 (1992) 1258—1266

Discovery: Hydrogen (H+) is source of passivation of p-type GaN

As grown MOCVD GaN contains significant hydrogen concentrations:

MOCVD Growth Gases

NH3 contains NH3

H+

Mg GaN:Mg
H with Mg-H Complex
(not p-type, highly resistive)
 Thermal Annealing of p-type GaN
Prior: Everyone annealed in H+ containing environment: no p-type GaN
Thermal Annealing in H+ free environment: p-type GaN, Industrial Process Compatible

Thermal Annealing in N2 Resistivity of MOCVD GaN:Mg vs. T

N2 Not p-type GaN

H2

Mg
HH

p-type GaN
 GaN Based Diodes
p-n GaN Homojunction
Needed
p-GaN
◦ Tunable Colors
n-GaN
Buffer
◦ Efficient Device Structure
Layer Sapphire ◦ Output Power > mW

p-n GaN Homojunction

(as developed by Akasaki & Amano)

◦ Good Crystal Quality

◦ Very Dim Light Production
◦ Very Inefficient
◦ Output power << mW
◦ Cannot tune color
Not Suitable for LEDs
Double Heterostructure
(Z.I. Alferov & H. Kroemer,
2000 Nobel Prize in Physics)
Confines carriers, yielding
higher Quantum Efficiencies
 Homojunction vs. Double
Heterostructure
Energy Band Diagrams
Homojunction LED Double Heterostructure LED

p-type Active n-type

p-type n-type
Layer

Internal Quantum Efficiency

Auger

Shockley-Read-Hall (SRH) Spontaneous Emission

Double heterostructures increase carrier concentrations (n) in the active layer

and enhance radiative recombination rates (more light generated).
Development of
InGaN
ENABLING THE HIGH-EFFICIENCY LED
 InGaN: At the Heart of the LED
GaN Double Heterojunction (DH) InGaN meets DH requirements
Needed
Active Smaller, Tunable Band Gap / Color by
Layer changing Indium in InxGa1-xN Alloy

Sapphire
Significant Challenges though …
◦ Hard to incorporate Indium as high
vapor pressure (Indium boils off)
GaN DH-LED: Band Diagram ◦ Growth at substantially lower T:
◦ Poor Crystal Quality
n-GaN p-GaN
◦ More Defects, Impurities
◦ Grow thin Layer (“Quantum Well”)
Light
InGaN

◦ Need fine Control over Growth Conditions

◦ High quality interfaces / surface morphology
◦ Introduces Strain in Crystal
◦ Indium ~ 20 % bigger than Gallium
 InGaN growth in 1991
Despite numerous attempts by researchers in the 1970s—1980s, high quality InGaN
films with room temperature band-to-band emission had not been achieved.

Indium Incorporation
InGaN Growth:
◦ Poor quality at low T
◦ Low incorporation at high T
◦ Hard to control In concentration
◦ High impurity incorporation
◦ Heavily defected
Photoluminescence

InGaN Luminescence:
◦ No band-to-band light emission
at room temperature
(fundamental for any LED device)
◦ Significant defect emission
N. Yoshimoto, T. Matsuoka, T. Sasaki, A. Katsui,
Appl. Phys. Lett., 59 (1991) 2251—2253
 High Quality InGaN Layers
1992: S. Nakamura et al., Jpn. J. Appl. Phys., 31 (1992) L1457—L1459

Enabling Technology: Two-Flow MOCVD

High Quality InGaN Growth with Band-to-Band Emission
Controllably vary Indium Concentration and hence color

Wavelength vs. Indium Fraction

Photoluminescence Spectra of InGaN
Indigo
Lower In

Higher In

Violet
 First High Brightness InGaN LED
1994: S. Nakamura et al., Appl. Phys. Lett., 64 (1994) 1687—1689

Breakthrough Device with Exceptional Brightness

(2.5 mW Output Power @ 450 nm (Blue))

Optimization of thin InGaN Active Layer

InGaN/AlGaN Double Output Power vs. Current

Heterostructure LED

2.5 mW
The Blue LED is born

Source: www.nobelprize.org
1st InGaN QW Blue/Green/Yellow LEDs
1995: S. Nakamura et al., Jpn. J. Appl. Phys., 34 (1995) L797—L799

High Brightness LEDs of varying colors by increasing Indium content.

Demonstration of Quantum Wells (QWs).

Green SQW LED Electroluminescence

Quantum

43%
20%

70%
Indium
Wells Content

yellow
green
blue
 1st Violet InGaN MQW Laser Diode
1996: S. Nakamura et al., Jpn. J. Appl. Phys., 35 (1996) L74—L76

First Demonstration of a Violet Laser using multiple QWs.

Laser Structure using InGaN Light Output vs. Current

Starts to lase
Comparison InGaN vs. other LEDs

Inhomogeneous: (InGaN)
Bright (!) despite high defects
Higher currents mask
inhomogeneity effects
(valleys fill up)

Homogeneous: (GaN,AlGaN)
Dim as defects “swallow”
electrons without producing light

After: Lester et al., Appl. Phys. Lett., 66, (1995) 1249

 Possible Origins of High Efficiency
Indium Fluctuations form localized states:
Separate electrons from defects
Indium in Active Layer
Random Binomial Distribution Side View in Energy Landscape
% In

Valleys

No In Defects Light
Atom Probe Tomography, D. Browne et al., UCSB

Chichibu, Nakamura et al., Appl. Phys. Lett., 69 (1996) 4188; Nakamura, Science, 281 (1998) 956.
Historical
Perspective
PAST, PRESENT, FUTURE
Historical: LED Efficiency

InGaN DH-LED by
Nakamura et al., 1993

After: G. Craford, Philips Lumileds Lighting Company

 Contributions towards efficient blue
LED
p-type GaN activated by thermal p-type GaN activated by
annealing by Nakamura, 1991 Electron Beam Irradiation
Hydrogen passivation was clarified by Akasaki & Amano, 1989
as an origin of hole compensation

InGaN Emitting
(Active) Layer
n-type GaN
by Nakamura, 1992

Sapphire substrate

GaN Buffer by Nakamura, 1991 AlN Buffer by Akasaki & Amano, 1985
 GaN/InGaN on Sapphire Research
Year Researcher(s) Achievement
1969 Maruska & Tietjen GaN epitaxial layer by HVPE
1973 Maruska et al. 1st blue Mg-doped GaN MIS LED
1983 Yoshida et al. High quality GaN using AlN buffer by MBE
1985 Akasaki & Amano et al. High quality GaN using AlN buffer by MOCVD
GaN

1989 Akasaki & Amano et al. p-type GaN using LEEBI (p is too low to fabricate devices)
1991 Nakamura Invention of Two-Flow MOCVD
1991 Moustakas et al. High quality GaN using GaN buffer by MBE
1991 Nakamura High quality GaN using GaN buffer by MOCVD
p-type GaN using thermal annealing,
1992 Nakamura et al.
Discovery hydrogen passivation (p is high enough for devices)
1992 Nakamura et al. InGaN layers with RT Band to Band emission
InGaN Double Heterostructure (DH) Bright Blue LED (1
1994 Nakamura et al.
InGaN

Candela)
1995 Nakamura et al. InGaN DH Bright Green LED
1996 Nakamura et al. 1st Pulsed Violet InGaN DH MQW LDs
1996 Nakamura et al. 1st CW Violet InGaN DH MQW LDs
1996 Nichia Corp. Commercialization White LED using InGaN DH blue LED
 UCSB’s Vision
LED based White Light is great, Laser based is even better!

Device 60 W Incandescent External Quantum Efficiency

Equivalent LED/Laser vs. Current Density

Commercial LED & Laser

LED
LED Laser
Sapphire
28 mm2

Laser Bulk GaN 0.3 mm2

M. Cantore et al., UCSB

Phosphor
Strip
 Acknowledgements
Nichia:
Nobuo Ogawa, Founder of Nichia Chemical Corp.
Eiji Ogawa, President
Colleagues of R&D Departments in 1989—1999
All employees of Nichia Chemical Corporation

UCSB:
Chancellor Henry Yang
Dean Rod Alferness, Matthew Tirrell

Profs. Steve DenBaars, Jim Speck, Umesh Mishra