The CECOM Center for Night Vision and Electro-Optics

OPTOELECTRONIC WORKSHOPS

00

OPTO-ELECTRONICS IN IIl-V SEMICONDUCTORS:

MATERIALS AND DEVICES

May 3, 1988

sponsored jointly by

ARO-URI Center for Opto-Electronic Systems ResearchThe Institute of Optics, University of Rochester

*UNCLASSIFIEDSECURITY CLASSIFICATIOD OF THIS PAGE

REPORT DOCUMENTATION PAGEIs. REPORT SECURITY CLASSIFICATION 1b. RESTRICTIVE MARKINdS

2&. SECURITY CLASSIFICATION AUTHIORITY 3. DISTR11UTIO)iIAVAjLABIuTy OF REPORT

2b. DECLASE FCATION/OOWWJ6RADING SCHEDULE Approved for public release;distribution unlimited.

4. PERFORMING ORGANIZATION REPORT NUMBER(S) 5. MONITORING ORGANIZATION REPORT KUMER(S)

6a. NAME OF PERFORMING OPCGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION(If applicable)

University of Rochester U.S ryRserhOfc6c. ADDRESS (Oty. State, a&d ZIP Code) 7b. ADDRESS (Cty, State, and ZIP Code)The Insitute of Optics P. 0. Box 12211Rochester, NY 14627 Research Triangle Park, NC 27709-2211

Be. NAME OF FUNDISPONSORING 6 b. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION O (f ipplcbe)

U. S. Army Research Office 19403-? k67Sc. ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS

P. 0. Box 12211 PROGRAM PROJECT NO.K ION UNO.

Research Triangle Park, NC 27709-2211 ELEMENT NO. INO.NOr SNN.

11. TITLE (Includ SeCurity C~alafako)Optoelectronic Workshop III: Opto-Electronics in IIl-V Semiconductors: Materials

and Devices12. PERSONAL. AUTHOR(S) Gr ik

13a. TYPE OF REPORT 13ab. TIME COVERED 14. DATE OF REPORT (YearMontADay S1. PAGE COUNTTechnical I FROM TO ____I May 3, 1988 1'

16. SUPPLEMENTARY NOTATION The view, opinions and/or findings contained in this report are those

of h auh r) an shuld nbef"A cos a n fficial Deartuent of the Army position,17. COSATI CODES 1S. SUBJECT TERMS (Condinu on rewi Nf neceaavy end ilentify by block number)

FIELD GROUP SUB-GROUWorkshop: III-V Semiconductors: Materials and Devices

'9. ABSTRACT (Contu an reve'se If necenaiy and Ident by block number)

'T his workshop on pto-Electronics in Ill-V Semiconductors: Materials and Deviceskrepresents the third of a series of intensive academic/ government interactions inthe field of advanced electro-optics, as part of the Army sponsored UniversityResearch initiative., By documenting the associated technology status and dialogueit is hoped that thi\,b aseline will serve all interested parties towards providing asolution to high priority Army requirements. Responsible for program andprogram execution areDr. Nicholas George, University of Rochester (ARO-URI) andDr. Rudy Buser, NVEOC. * A--'

I f\Lo t

20. DISTRIBUTION/ AVAELAIUY OF ABSTRACT 121. ABSTRACT SECURITY CLASSIFICATIONO3UNCLASSIFIEDJNLIMTED - 0 SAME AS RPT. C) OTIC USERS Unclassified

22a. NAME OF RESPONSIBLE INDIVIDUAL I22b. TELEPH4ONE (include Arts Code) 1 22c. OFFICE SYMBOLNicholas George I716-275-2417

DD FORM 1473, B4 MAR 83 APR edition may be used until eidtauted. --- SECURITY gLSSFICATION OF TIS PAGE'AX other editions are obsoWee UNCLASSIFIED

OPTOELECTRONIC WORKSHOP

ON

OPTO-ELECTRONICS IN IIl-V SEMICONDUCTORSMATERIALS AND DEVICES

Organizer: ARO-URI-University of Rochesterand CECOM Center for Night Vision and Electro-Optics

1. INTRODUCTION

2. SUMMARY -- INCLUDING FOLLOW-UP

3. VIEWGRAPH PRESENTATIONS

A. Center for Opto-Electronic Svstems ResearchOrganizer -- Gary Wicks

Optical Properties and Technologies of III-V MaterialsGary Wicks

Superlattice DisorderingSusan Houde-Walter

Optical Interactions in Indirect Bandgap:Il-V Semiconductors and Silicon

Dennis Hall

III-V Optoelectronics for Optical Communication.Thomas Brown

B. CECOM Center for Night Vision and Electro-OpticsOrganizer-- L. N. Durvasula

4. LIST OF ATTENDEES Accesslon 'c r

NTIS GRA&IDTIC TAB

Justifioatlon

Di stribut on/

Availabllty Codes

Dist Speo lal

1. INTRODUCTION

This workshop on "Opto-Electronics in Ill-V Semiconductors: Materials andDevices" represents the third of a series of intensive academic/ governmentinteractions in the field of advanced electro-optics, as part of the Armysponsored University Research Initiative. By documenting the associatedtechnology status and dialogue it is hoped that this baseline will serve allinterested parties towards providing a solution to high priority Armyrequirements. Responsible for program and program execution areDr. Nicholas George, University of Rochester (ARO-URI) and Dr. Rudy Buser,CCNVEO.

2. SUMMARY AND FOLLOW-UP ACTION

Rudy Buser started the workshop with a short talk. He mentioned two areas ofIll-V s which are of interest at NVEOC: 'ocal plane arrays of 10 pm detectors; andmonolithic integration--sources; detectors, etc.

Ward Trussel spoke about NVEOC's interest in high power laser diode arrays forYAG pumping

Several NVEOC personnel expressed interest in high power/high efficiency visible(green) lasers with good beam quality. One way of doing this is by doubling thediode pumped YAG lasers. Dennis Hall received some intercst when hementioned the possibility of green GaP lasers,

George Simonis spoke of interest in optoelectronics at Harry Diamond Labs.Their activities include: doping superlattices for optical modulation; resonanttunneling (mainly theory); piezoreflectance studies of Ill-V heterostructures, andlightwave research for microwave radar applications.

The original plan was for L. N. Durvasula and I to get together for a short whileat the end of the workshop to decide how future interactions might proceed.Time did not allow this, however, and we agreed to discuss these items on thephone. I tried several times without success during the week following theworkshop to contact Dr. Durvasula. Finally I sent him a letter stating a fewspecific areas where there appeared to be good overlap between research areasat UR and interests at NVEOC. Also stated in the letter was my feeling that theformal presentation structure was appropriate for the first workshop, butsubsequent interactions should consist of smaller, less formal groups and moredialogue.

I would estimate that 30 NVEOC personnel attended the workshop.

L1

2o

J

w

0

Z-J

.-

0

Outfin

*Crystal growth technologies

*Bulk materials properties

*Properties of quantum wells and superlattices

*Device fabrication technologies

ENERGY GAP (in eV and /im) VERSUS LATTICECO.NSTANT AT 300K FOR COMPOUND SEMICONDUCTORS

.0-

ZnSe

AlP 0.5

.0-

:Si 1n

0

0eG~

-2

-H 105.4 5.6 5.8 6.0 6.2 6.4 6.6

a0(A)

Heterojunction Energy Bands

*~jjjujd~ConductionAEC Ooeeo Ban d

EA~~ Gaas as adsa

Quantum Well

T~na a Iz

~n=Z E

AMGaAs GaAs ANGaAs

Superlattice (type 1)

Important Epitaxiol Structures

LosersDouble Heterostructure (DH) Al concentration

p GaAs

p AlGoAs 121,m

GaAs 0.2,&Lmmbn AIGoAs 121mn GaAs ___

Graded refractive index- Al concentrationseparate confinementheterostructure (GRINSCH)

p GaAsp AIGaAs

groded AIGoAs

n AIGoAsn GaAs

Other Versionsmulti-quantum wellconstant composition, seporai confinement heterostructure

Ill-V Materials for Lasers

x = 2.2 A~m - GaInAsSb/AIGaAsSb DH

x' *=1.7 -1.2 Am GaInAsPAnP DH

x = 1.2 -0.9 A~m -pseudomorphic

GaInAs/AIGaAs OW's

* X = 0.9 0.68 Am - AIxGaj-xAs/AIyGai 1yAs OW's

e x = 0.65 pm - Gajswtn.48P/AIGaInP

IIl-V's for Long WavelengthPhotodetectors

@Lowest bandgap bulk Ill-V is InAs Sb6 l

eStrained InAsSb Quantum Wells

x,< 12 pm

*Intraband transitions in the conduction band ofsuperlattices

2 A/Watx = 10 pm

note: only radiation polarized along growth axis isdetected

Crystal G rowh Tech noI

*Bulk Growth Techniques

-Bridgeman

-Uquid Encapsulated Czochralski (LEC)

9Epitaxial Growth Techniques

-Uquid Phase Epitaxy (LPE)

-Vapor Phase Epitaxy (VPE)

-Molecular Beam Epitaxy (MBE)

-Metal Organic Chemical Vapor Deposition (MOCVD)

VAPOR PHASE EPITAXY

Goci

AsC 3 As2 A5. - Chloilide VPEHtECI -

2 Go H2 GoAs

T, > T

GHCIHyrdVEH-2, H C Hydride VPE

Go Ha - GAs

T constont

hot wll

_ __H3 MON MOCVD

GoAs

cold wall

0.00

Ad b

"4 P

'I 4W40

060

0o 0

P-4 C OC b0 0 U 0( 0 ae m)

Ad c to0 9-

.C 40 w Wbi 64A00 0 Ik.k 004

Wo 4' c c

0~A 0 be 0 0)0 0

01 CO 4) 9-4 u44'l 0)

00 C

. ).Cbe um

MOLECULAR SEAM EPITAXY

CRVOPANELLINO

SUBSTRATENEATING fLOCX

GGAS WAFER

EFFUUSNICEL

CELLS MUTTER

IIIcc T ePOte[je 0 0s 6 0 S 0 0

thea ~~ccoupSW'oX4 AS5P;NFL.i~tic'. mteia__________________VACUUM cucbl

Mo~ciA RBam Eotax (MBE)

MAIN ORO~T CHAMBERCHME

FUSIO CELS ILED. MASS SPEC

SMAM SOUTRCS ALMANIPLATO (LQaD NITON E

II SENSTRAjMOEFFUSION~.I GCELL

6as Source MBE SystemMOMBE

Standard MBEgrowth chamber

DiffusionPump

*Pressure control of

qroup V gases

* Crack AsHf and PH

in same cell

Computer Controlled K- Cells L PUP I I /:3 j

CIHAMBER I (a ,Al, As,

SOLID SOURCES Be, Si)

Optical PortsandRED

Residual GasAnalyzer

High Temperature

3 Waer TackCleaning Stage3" Wafer Track --- '

Syatem

Expansion o o

0 Dual e-Beame-gun Metallization

Optical Ports Residual Gas

and Analyzer

RED N

CHAMBER # 2 (s3P3GAS SURCESTEGa, TEIn,

Al, Be, Si)

Computer Controlled Gas Sources

Densityof

States 2

n-1

Energy

Electrooptic Effects in Ill-V's

Bulk materials - Franz-Keldysh Effect

EC

AEGv

A & lOKV/cm

Shift Absorption Edge * Change refractiv index:

An- 5 x1O1

AE lOKV/cm

Electrooptic Effects an HIl 'sQuantum Wells - Excitonic Effects

Case 1. Electric field // Growth axis

Ec

Ev

*Change shape of potential well 4* shift exciton energy levels 4 shiftabsorption edge =# change refracive index.

Case 2. Electric field I Growth axis

*Field ionize exciton * remove excitonic features from absorptionspectrum 4 change refracive Index.

*The excltonic absorption In quantum wells leads to very largeelectroopticeffects:

A n - 5 x 10-2A-F lOKVlcm

ha c 100 cm- ,

lOKVlcm

Edge Emitting Laser

Epiv= .Layers

Substrate 1

Dry Etching Technius optoelectronic Devi

*Chemically Assisted Ion Beam Etching

•Reactive Ion Beam Etching

eReactive Ion Etching

Applications

*Integrated lasers no cleaved mirrors

*Other integrated devices- lenses- mirrors- modulators

L97W

IAJQ

0

Iui

LU

0U.

zwj

Experimental:ref: Y. Suzuki and H. Okamoto, J. Elec. Mats. 12,397 (1983)

Compare refractive index (dispersion) of bulkAlo.3Gao.oAs to GaAs/AlAs SL's with various barrierthicknesses:38

3.7 -j LE

w 3.6, .0 - Et(e-hh) -

,,,3.5 "

3.4 -162,27)

w3.3,

1.3 1.4 1.5 1.6 1.7 IB

PHOTON ENERGY (eV)

Note: for a fixed average alloy composition, one canmake large changes in refractive index by tailoringheterostructure dimensions.

>>useful degrees of freedom in device designlaserswaveguides, lenses, modulatorsdetectors

Heterostructures are grown epitaxiallymolecular beam epitaxymetallorganic chemical vapor deposition

-,ilayers(2-6000 A)

GaAs substrate

INTEGRATION: grow one structure, use same basematerial for all components

define components(e.g. ridge waveguides,heterostructure lasers) by removing surroundingmaterial (ion beam etching, cleaving, etc.)

Disadvantage:scattering, band bending, not planar, etc.

Objective: change bandgap and refractive index locallyto define components

New Method: smooth abruptness of heterostructureslocally

GALLIUM COC.T

C,, 0

Ch

fCD.1%

A9)

co 0

00

.4DI

Mixing is enhanced by impurity transport

ref: W.D. Laildig, et al, Appl. Phys. Lett. 38,776 (1981)

MASKED Zn DIFFUSION INTOSL

Zn2As3Si3N4

AlAs/GaAs SL

small bandgap large bandgaphigh index low index

Lasers:

Use layer mixing to:1) tune emission wavelength

ref.: M. Camras, et a!, App!. Phys. Lett, 54y,5637(1983).

fgl-3s EC

*,-12

2.0l

2h

W'' 0 PWl

IS

CW CURRET (A)

THIS WORK:

lID involves electrically active species>>use applied electric field

>>more accurate diffusion coefficient measurements>>improved lateral resolution for component definition

Zn diffusion

1gin

Si diffusion - ---'---_

1gm

mask edgediffusion dominated

drive in mixing profile;little time for sidewaysdiffusion.

field-dominated

Use evaporated source or gettering layer:

Zn2As3Si3N4

V

AlAs/GaAs SL

Use to study lID mechanism, fabricate components.

z0

0z

=0

CAZ

0 Lp

Liu<

0M

LU

=z0-

Z0

OPTICAL INTERACTIONS IN INDIRECT BANDGAP

Ill-V SEMICONDUCTORS AND SILICON

" Gallium Phosphide, GaPOptical Emission

* SiliconOptical EmissionWaveguides

ENERGY GAP (in eV and /m) VERSUS LATTICEONSTANT AT 300K FOR COMPOUND SEMICONDUCTORS

ZnSe

AlP CdS 0.5.0

All~b

-AISb CdTe - )

GaAs_"Si "I

' Ge GaSb" ---2--0

S_. 11I I .L~ -110

.4 5.6 5.8 6.0 6.2 6.4 6.6a,(A)

GALLIUM PHOSPHIDE, GaP

INDIRECT GAP: - 2.27 eV AT T =300K.

IMPURITY - RELATED LUMINESCENCE

B C N, X

Al Si P S

NZn Ga Ge As S e

Cd in S n Sb To

Hg TI Pb Bi, Po

ISOVALENTWITH

PHOSPHORUS

ISOVALENT ( ISOELECTRONIC) IMPURITY

GaP: N or GaP :Bi

ISOVALENT ( ISOELECTRONIC ) MOLECULE

GaPZn-O or GaP:Cd-O0

(Nearest-Neighbor Donor-Acceptor Pair)

ENERGY BAND STRUCTURE

GaAs - DIRECT ENERGY GAP

E Conduction

Band

SphotonE=,r - 1.4ev 1

ValenceBand

k 0 Crystal Momentum, k

Silicon - INDIRECT ENERGY GAP

't'I CB

EIS -1.1 OVj

k=kk -0

I I1k- CONSERVATION SELECTION RULE

*UNASSISTED , BAND - TO - BAND, RADIATIVE RECOMBINATION OF ELECTRONS

AND HOLES IS FORBIDDEN.

PHONON - ASSISTED RADIATIVE TRANSITIONS CAN AND DO OCCUR.

I ..o

EXCITONS BOUND TOISOELECTRONIC IMPURITIES

CB

electron -

UCOMPLEX gap photon

hole

PROBABILITY OF NONRADIATIVE AUGER RECOMBINATION IS LOW.

PROBABILITY OF RADIATIVE RECOMBINATION IS HIGH.

GaP:NGaP 4

- ~~NNW ~IT N- -E T ON, BOUND O

2 A

,,, -- A --o SINGLE.N"IMP' URITY

NN ' PAIRS OF/q, jb P.s

2.30 2.305 a.t 2.315 2.32PHOTON ENERGY (ev)

ATOMS/CC or -;?GEd

d * d0 @ 4 3 '

40 1.6"K F 7 o{

20 NN/ ,. Rev. p__, (toa (I a ID).

I to-1 uo

4 4- SLOPE 2

2 z

S. a

4

06-

06 6 .0L4 -

2 4 6 a 10 20 00 60 t00

ABSORPTION AT 2.32eeV ON PHONON WING OF A (cu.)

G p~ N2R

24 Lc~oUlil~LL

G'ree~ ~ 20

LED16 Absorption

02.0 2.1 2.2 2.3 2.4

Photon energy', eV

LASER /AC7)ON

V.S. Pateyit 3,'71)3'7 (1973)

Le~em , Loan NoAIor'8) and Shaklee

"Losers io. Xvidmipet0- Samdpap Sernicondvdht.

crstgas Voec WA Xsoe)ectronh-. T. r 1 ,:.

G; in GAP i= ~sre

RoM T7EMEATOT

Gap: zn Cd-O Zn-O UI

z0-.4TK

24

3 a-

0 . 18 92

StiuIiteJ ETmiSSjon and4 LASerOrverati..st (CtW, '77K) Associite4 ~.

WAt pe ISOc)eci'.n~c Tra~s.li

N. He-fornoi' +7*vies

TL~optIV Ail *MSe 1fVSedletebO fo1 ORIO1 W/cnil'l

IN.arr~vd\ at -fie FeSS~nIM_______cov"C~s;ons 5ha 16f We arcceAI(W 7.2 7.0 6.8 6.6 6.4 62

ik ' +he Nfrmiur'e r.VnctrmIA Woveength (103A)

in~rc sewikesduct-ars.controve'sia L

WHAT ABOUT SILICON ?

ISOELECTRONIC FSOEECTONCPAIRS SERIES

II IV V VI

B C N

191986

Cu Zn

1979.Ag Cd In Sn Sb Te

Au Hg ~ Pb Bi Po

W. P. DUMKEIBM, YORKTOWN HEIGHTS

"INTERBAND TRANSITIONS AND MASER ACTION"PHYS. REV. 127, 1559 (1962).

FIRST LINE OF THE PAPER

Since the initial operation of the ruby maser, there has been

considerable speculation concerning the possibility of observing maser

action in semiconductors such as Ge and Si.

ABSTRACT OF THE PAPER

The possibility of using interband transitions to achieve maser action is

considered. The criterion for maser action Is presented in a way which allows the most

direct use of optical absorption data. The absorption constant for interband transitions,

which is negative corresponding to induced emission when a population inversion

exists, is related to the normal absorption constant for direct, indirect, and indirect

exciton transitions. Using available absorption data, it is shown that in Ge (S )maser action, using either the Indirect or indirect exciton transitions, would

be prevented by absomtion due to free carriers. In GaAs, or other

materials with a direct gap, however, it is entirely possible that maser

action could be achievbd.

fU HC4-

14

0Lr

co

(sLNnl aEIV) AJISNHJLNI HONDSNIFII

~~ k

I~~ ~~~ -- _New v v

wovOw

%#Mo

IMO

'I'3~woo

a a 2 a a m I ML

(Snmn BHY) AATISNW IDNzoszNmflfI

z=0

LOZ

u0

L

Oo

W.Jzo

0 LI

Ill-V OPTOELECTRONICS FOR OPTICAL

COMMUNICATIONS

1. Overview: Communications Systems

2. Overview: Modulation/DemodulationSchemes

3. Source Stability Requirements forAdvanced Optical Communications

4. Methods of Source Stabilization

Historical Perspective:

Free Space Optical Communications AncientGreeks

Morse Telegraph 1840's

Bell Telephone 1870's

Marconi's Wireless Telegraph 1900. a. Radio

Coherent Carrier 1940's

Modulation/Demodulation

Phased Locked Receivers 1950's

First Optical Mixing Observed 1955

Semiconductor Laser Invented 1963

Low Loss Optical Fiber 1970

Free-Space Optical Heterodyne 1970'sSystems

High" speed, low loss single-mode 1980'sfiber links

Coherent Optical Fiber Communications 1980'sproposed

Communications Systems:

Free Space Optical(Fiber

Telecommunications Data Transfer

Trunk Local Area Network Subscriber

There is current interest in:

Long repeaterless links.Extremely high capacity links.

Wide area distribution.

Direct Detection Systems

Traditional:

_n-F-u1+ Dispersi~Channel

IntensityModulated PIN-FET

Diode Laseror

Avalanche PhotodiodeReceiver

Wavelength-Division Multiplexed (WDM) Systems:

X1+ rnrLDispersive X2

X3 el-anne3

X4

X4t

Pulse-Position Modulation (PPM) Systems:

Low Dispersion na Detector

(lowest fundamental limit)

* Direct Detection Systems

Average Power Ps ideal detector rl =1

Average # of photons per time slot:

Ns- PsTB

?ico

Photo-electron statistics:

p(n)= e poisson processn!

Ideal photon counting: (no dark count)

prob. of error - PE = prob(n = 0)

= E-Ns

I Ns-20.7 for PE-- 99

The Ideal System: Receiver Sensitivity

* Heterodyne/Homodyne Systems- No phase noise, perfect phase

matching

- Coherent(Phase-estimation) D{ IncoerentDemod.Incoherent (Envelope)

Photodetector RF

Es(r, t) + Eo(r, t)

i 0~t)t

i (t) = r Es(r, t) + ELo(r, t) 2

i (t) = (rI e/hw) {Ps + PLO + 2 VPTPL Cos (&t + 4(t))}

ASK

representation: a(t) = 1 "mark"

(Binary) a(t) = 0 "space"

representation: a(t) = 1

(Quaternary) a(t) = .67a(t) = .33

a(t)= 0

Analog limit (minimum bandwidth)

* We are not interested in bandwidth

compression, so use binary.

I TB

t=O

ASK IMPLEMENTATION

• Same modulator technology as direct

detection systems.

* External modulators only, since direct

modulation of semiconductor lasers

results in frequency modulation.

(GHz/mA)

* Easy detection (in analogy to RF

systems)

MACH-ZEHNDER INTERFEROMETRIC MODULATOR

4- Phase Modulator 01

0 1 = Io/2 Cos 2 (AO/2)

- _Phase Modulator 02, AO=O6 _-e

A =1 62

* Identical waveguides, equal path

lengths yield constructive interference.

* Relative phase shift AO causes

destructive interference for AO = n

• Practical design considerations require

tuning for "on" and "off' states.

• Top speed > 17 GHz

[Ref. Gee and Thurmond OFC'84

R. C. Alferness, IEEE JQE 17,946 (1981)]

Receiver Sensitivity PenaltyASK Envelope Detection with

Optimized Post-detection Filtering

3.o

im.0

Analytical (single-pole IF filter)C

1.0 .O

-. Numerical (integrating IF filter)

7

/

0.0

0.0 .2 .4 . .9 1.0

Llnewidth/Data Rate

PSK

representation: a(t) = 1 mark"

(Binary) a(t) = -1 "space"

representation: a(t) = 1

(Quaternary) a(t) = 1a(t) -. 4=Levels

a(t)

Analog limit (minimum bandwidth)* Once again, use binary.

IIIII

I/

* PSK is a supp ressed carrier modulation

technique

SSE, E

, \ /, \,I/I

* No energy at ct = Co

$ Requires nonlinear phase recovery

scheme.

PSK Implementation:

" Simplest possible modulator

configuration.- Electro-optic (Pockels) effect- Free carrier effect (semiconductors)

* External modulators only

" Detection is difficult

as(t) Cos (cost + 4(t))

weak signal/XI

Strong Local OscillatorCOS(COLOt - ( )

for error-free detection,4(t) = )(t)

PSK M -ER-oP'(I0E to51I-c-TIc

4.0

3.0-

A. 2.0-

1.0-

O0.0 , . , , , ,I , , ,, g , , , , t,*

0.0 0.05 0.1 0.15 0.2 0.25

NORMALIZED SPECTRAL WIDTH (/)

FSK

Binary Representation: a(t) = e i "mark"

(Ac channel spacing a(t) e cspace"

M-ary a(t) = ei(m+I)()ct

representation a(t) e iIct

#Channels ]M=2m .#e ( a(t) = e - i(m+ )Oct

Bandwidth.Expansion

technique

I Power Spectrum SE E ()

I( C

CIL) = o

Receiver Sensitivity PenaltyFSK Envelope Detection with

Optimized Post-detection Filtering3.0

Analytical (single-pole IF filter)

4D - Numerical (integrating IF filter)

1.0 .4

Linewidth/Data Rate

Ill-V Optoelectronic Devices: Limitations to SystemPerformance

Semiconductor Lasers:

Mode Partitioning(Direct Detection)

Phase Fluctuations/Line Broadening

(Coherent Detection)

Noise Mechanisms

Thermal Instabilities (Phase ShiftsIn

Quantum Fluctuations Optical Signal

How does this happen?

AX. Fabry-Perot Modes

n It

, I- -- I

- ,, e -Gain ProfileIIj

I I !I I

I I I IS

Wavelength

Large changes ..... Mode "partitioning"Small changes ..... Line Broadening

Mode Partitioning

1) Time Averaged Laser Spectruma0

a a+1

a 2 a+2

2) "Instantaneous" Measurementa-1

a0 a

+1a 2 a +2

0

Assuming the population adiabatically follows thequantum fluctuations:

+00XIa. = Constant

Constant flux Is dynamically "partitioned" among themodes.

What happens In a dispersive channel?

---- a° Dispersive W -W.. ao

a Channel .

Nearly Single Mode][SourceSolutions:

- Use the minimum dispersion point In fiber.- Stabilize the diode laser:

"Single-mode" lasers require a side modesuppression ratio of about 100:1.

Distributed Feedback (DFB) Lasers provide modestabilization.

Active Region

DFB Laser Structure

10-6 - --

k=0.18- - -(MODE RATIO= -31)

10-8 -------

o k m0. 15410 (MODE RATIO= -42)

0-1 V:ODE RATIO-Si) __

10-16L- - - -

0.6 .0.9 1.2 1.6 2.2INTERMODAL DELAY/BIT PERIOD

LINE BROADENING

Take a closer look at what Is driving the phasefluctuations:

IM[C] Stimulated SpontaneousEmission o°°° Emission

Re[s]

AO= A~irst + A0 relax

Al -- Im[An] -> Re[Anj -*A4

Figure of (de)merit for Gain medium:

cRe[An] - 5 for GaInAsIm[An]

$ 4) exhibits a "random walk"

o,2=Dt

t=0 p(;t)= 2/22

//- ",, t=tl

/ \ 4a/ t=t2>tl

R-Sp. emission rateD (1 +a 2) R./21

a increases o2 by -25 to 50.

Av = D/2n

- 0 SolutionsDFB Lasers (MHz)External Cavities (10-100 KHz)

Hew do we narrow the laser Iinewidth?

In an Ideal system, we can:

-Increase the pump power (% above threshold).

-Decrease the spontaneous emission rate.(External/Coupled Cavities)

-Introduce active stabilization.

-Decrease a

How can (x be reduced?

Bulk III-V Gain Medium

G Gai 71q" -- refractiveELos Index

Discrete Transitions (Multiple Quantum Well)

RefractiveIndex GaIn

Energy ..-

S a eS p

Reducing the Strong amplitude-phase coupling can

improve linewidths by 25 - 50.

Options for discrete transitions:

- Multiple Quantum Well (MOW) diode lasers.

- Rare-eerth resonant diode lasers.

Modulation Sensitivity Linewidth LaserTechnique (Photons/bit) Requirements of choice

Direct 400-4000 100:1 Good FPDetection side mode or

DFB

ASK Het. 80 0.1 x Data Rate DFB

FSK Het. 40 0.1 x Data Rate DFB

PSK Het. 18 10-3 x Data Rate ?

PSK Horn. 9 10-4 x Data Rate ?